2025Activity reportProject-Team​​AIO

[A1.1] Fast Calibration​​ and Standardization

The behavior​​​‌ of the oscillators within​ the SCuM chip is​‌ not fully understood. We​​ have shown in 98​​​‌ that the oscillators drift​ by up to 16,000​‌ ppm over temperature. This​​ is much more than​​​‌ the maximum 40 ppm​ required by the IEEE​‌ 802.15.4 standards. The challenge​​ is hence to turn​​​‌ an unstable 16,000 ppm​ oscillator into a stable​‌ 40 ppm oscillator so​​ SCuM chips can reliably​​​‌ communicate with off-the-shelf IEEE​ 802.15.4 and BLE transceivers,​‌ and with other SCuM​​ chips. This may be​​​‌ done in the following​ steps:

• The first question​‌ to answer is: can​​ we model the behavior​​ of these oscillators? Specifically,​​​‌ we want to know‌ how much of the‌​‌ variation of these oscillators​​ is a function of​​​‌ temperature, voltage or other‌ parameters, and how much‌​‌ is due to thermal​​ noise, i.e. a random​​​‌ variation which cannot be‌ controlled. A secondary goal‌​‌ is to turn these​​ lessons learnt into the​​​‌ simplest possible model (e.g.‌ some form of a‌​‌ decision tree) so it​​ can be implemented on​​​‌ individual SCuM chips. We‌ know that simple curve‌​‌ fitting on 2-point of​​ 3-point calibration over temperature​​​‌ is usually done for‌ crystal-oscillators, but this will‌​‌ most likely not work​​ here. We adopt the​​​‌ approach of Oroza 86‌ and use Machine Learning‌​‌ to answer both questions.​​ Random Forest, LDA or​​​‌ PCA indicate the contribution‌ of each feature to‌​‌ the overall model. One​​ preliminary step is hence​​​‌ to gather a more‌ complete and more annotated‌​‌ dataset as in 98​​. We may reach​​​‌ out to the scikit-learn‌ team, as well as‌​‌ Prof. Carlos Oroza from​​ the University of Utah​​​‌ with whom we have‌ been collaborating on ML‌​‌ approaches for wireless communication​​ systems.
• There is a​​​‌ good possibility that the‌ model from the study‌​‌ above is not able​​ to fully characterize drift​​​‌ in all cases. Do‌ we need reinforcement learning‌​‌ to allow micro-motes to​​ calibrate against regular motes?​​​‌ In practice, we use‌ motes from our OpenTestbed‌​‌ 82 as a “calibration​​ box”: they are programmed​​​‌ to listen on all‌ frequencies, and acknowledge any‌​‌ calibration probe frames they​​ receive. The goal is​​​‌ to create an algorithm‌ by which a micro-mote‌​‌ transmits such probes as​​ it sweeps through calibration​​​‌ settings until it receives‌ an acknowledgement and thereby‌​‌ knows that setting is​​ valid for a particular​​​‌ frequency. The challenge is‌ that tuning the oscillators‌​‌ needs to happen continuously,​​ as any temperature change​​​‌ will cause SCuM to‌ lose connectivity to the‌​‌ calibration box. If reinforcement​​ learning is needed, similar​​​‌ to our approach in‌ 82, one option‌​‌ is to use game​​ theory to model as​​​‌ a “Multi-Armed Bandit”, and‌ use an $ϵ$-greedy‌​‌ algorithm to balance using​​ the tuning parameters that​​​‌ worked in the past,‌ and exploring other parameters‌​‌ in case the temperature​​ has changed. The target​​​‌ is for uncalibrated micro-motes‌ to self-calibrate quickly, and‌​‌ stay calibrated as temperature​​ changes.
• The study above​​​‌ results in a protocol‌ between a micro-mote and‌​‌ the calibration box. We​​ propose this as a​​​‌ candidate for standardization at‌ the IETF, possibly in‌​‌ the 6TiSCH working group.​​

Results: From a​​​‌ scientific point of view,‌ A1.1 generates the world's‌​‌ largest open and annotated​​ dataset of micro-mote drift,​​​‌ and use Machine Learning‌ to derive a model.‌​‌ In case that drift​​ model does not capture​​​‌ the drift fully and‌ hence cannot ensure a‌​‌ micro-mote can always communicate,​​ A1.1 also develops a​​​‌ dynamic fast calibration protocol‌ against regular motes using‌​‌ game theory, and standardize​​ that protocol. From a​​​‌ project point of view,‌ A1.1 allows micro-motes to‌​‌ communicate, albeit with the​​​‌ help of regular motes,​ a necessary stepping stone​‌ for the remainder of​​ the project.

[A1.2] The​​​‌ Network as a Time​ Source

A1.1 allows micro-motes​‌ to communicate, but does​​ require having regular motes​​​‌ close to micro-motes, which​ puts a burden on​‌ the deployment strategy. The​​ goal of this area​​​‌ is bold and ambitious:​ can a micro-mote use​‌ the network as a​​ relative time reference, instead​​​‌ of an absolute time​ reference such as a​‌ crystal oscillator? This entails​​ having a micro-mote calibrate​​​‌ against another micro-mote –​ which itself is drifting​‌ – and repeating this​​ over a multi-hop network​​​‌ of micro-robots.

This goal​ translates into the following​‌ three studies:

• The challenge​​ is that both neighbor​​​‌ nodes drift: if we​ were to use the​‌ approach from A1.1, neighbor​​ nodes would lose connection.​​​‌ The question is: can​ we develop a fast-tracking​‌ algorithm which allows micro-motes​​ to calibrate against one​​​‌ another? For that, we​ can use control theory:​‌ consider the time offset​​ between two micro-motes as​​​‌ the variable, offset 0​ as equilibrium, and tune​‌ the clock calibration. Similar​​ to our approach in​​​‌ 63, we can​ use a PID controller​‌ to balance reactivity with​​ resilience to short timing​​​‌ glitches. We can develop​ the controller in simulation​‌ by replaying its behavior​​ against the datasets from​​​‌ A1.1, before implementing it​ on micro-motes and validating​‌ experimentally.
• In a scenario​​ in which the micro-robots​​​‌ form a multi-hop mesh​ network, we want to​‌ avoid forming cliques of​​ synchronized structures, given the​​​‌ important variation in drift​ between nodes. The question​‌ becomes: can we have​​ a micro-mote calibrate against​​​‌ multiple neighbors at the​ same time? The complexity​‌ is that these neighbors​​ can be far apart​​​‌ in time, so the​ micro-mote may need to​‌ “jump” in time depending​​ on which neighbor it​​​‌ is communicating with, while​ implementing some averaging function​‌ that causes the network​​ to eventually converge to​​​‌ a fully-synchronized state. We​ build upon the work​‌ published in 99 by​​ integrating the datasets from​​​‌ A1.1 into the 6TiSCH​ simulator, and expand the​‌ controller from the study​​ above to support multiple​​​‌ neighbors.
• All TSCH networks,​ including 6TiSCH, build a​‌ synchronization structure inside a​​ multi-hop mesh network rooted​​​‌ at a single time​ master, which never changes.​‌ This does not match​​ the micro-robot application, in​​​‌ which all micro-robots play​ a similar role. The​‌ question is hence: can​​ we build a TSCH​​​‌ network in which the​ role of time master​‌ changes from one micro-mote​​ to another, without any​​​‌ disruption to the network?​ A corollary question is:​‌ can we imagine having​​ multiple time masters, for​​​‌ example nodes equipped with​ crystals? We will use​‌ our previous work on​​ the flooding-based approach from​​​‌ 66, and standardize​ the behavioral/protocol changes through​‌ IETF 6TiSCH.

Results:​​ From a scientific point​​​‌ of view, A1.2 results​ in deep changes to​‌ the base behavior of​​ a TSCH protocol stack​​​‌ such as 6TiSCH, with​ a radically new way​‌ of tracking a time​​ source neighbor, based on​​ control theory, as well​​​‌ as the ability to‌ track multiple neighbors, and‌​‌ have dynamic time masters.​​ The protocol changes are​​​‌ standardized at the IETF.‌ From a project point‌​‌ of view, A1.2 allows​​ neighbor nodes to communicate:​​​‌ choose the communication frequency‌ and the modulation/demodulation rate,‌​‌ and stay synchronized.

[A1.3]​​ Wireless Enablers

Localization of​​​‌ individual robots in a‌ swarm is critical for‌​‌ their coordination and control.​​ Accurate localization, often implemented​​​‌ using visual markers and‌ multiple cameras, is used‌​‌ as a ground truth​​ to determine the accuracy​​​‌ of robotic control algorithms.‌ To enable inexpensive and‌​‌ massive deployments of robots,​​ we investigate lighthouse localization​​​‌ in conjunction with RF‌ localization, both using angle-of-arrival‌​‌ techniques using antenna arrays,​​ and ultra-wide band techniques.​​​‌

The goal is summarized‌ in the following studies:‌​‌

• The first challenge involves​​ the use of angle-of-arrival​​​‌ estimation using BLE transmission‌ from the Single Chip‌​‌ Mote. A number of​​ questions arise: How does​​​‌ the time and frequency‌ uncertainty of the Single‌​‌ Chip Mote affect the​​ accuracy of angle-of-arrival estimate?​​​‌Can inaccuracies and errors‌ caused by multipath be‌​‌ mitigated by using frequency​​ diversity, both with BLE​​​‌ standard channels and potentially‌ with operation outside of‌​‌ the ISM band?
• Lighthouse​​ localization has already been​​​‌ established as a relatively‌ accurate method to localize‌​‌ Single Chip Motes. However,​​ with the addition of​​​‌ RF angle-of-arrival and RF‌ localization techniques, can sensor‌​‌ fusion algorithms be applied​​ to multiple measurements to​​​‌ both improve accuracy of‌ the position estimate, and‌​‌ discard erroneous estimates caused​​ by the mote's clock​​​‌ drift or unpredictable RF‌ fading during robot movement?‌​‌
• Millimeter wave (mm-wave) radios,​​ operating at much higher​​​‌ frequencies than the 2.4‌ GHz ISM band, benefit‌​‌ from greater attenuation of​​ environmental reflections, narrower beam​​​‌ widths when operating as‌ phased arrays, and inherently‌​‌ higher spacial accuracy due​​ to the smaller wavelength.​​​‌ These radios, both narrowband‌ and wideband, have previously‌​‌ been used for automotive​​ radar, range finding, and​​​‌ imaging. However, these solutions‌ are rarely power constrained,‌​‌ unlike the devices and​​ robots of interest to​​​‌ the research program. The‌ question we contribute to‌​‌ address is: can low-power​​ mm-wave radios, duty-cycled to​​​‌ conserve energy, be used‌ for robotic localization? Two‌​‌ sub-questions arise: First, to​​ what degree will the​​​‌ performance limitations set by‌ the time and frequency‌​‌ imprecision of crystal-free radios​​ affect mm-wave localization accuracy,​​​‌ both using narrowband and‌ UWB rangefinding. Second,‌​‌ given that mm-wave radios​​ typically burn more power​​​‌ when active, how deeply‌ would such devices need‌​‌ to be duty-cycled to​​ minimize the effect on​​​‌ battery lifetime while still‌ improving the location estimates‌​‌ of large numbers of​​ robots? The use of​​​‌ mm-wave radios on tiny,‌ low-power robots also has‌​‌ implications for communication. Small​​ antennas are significantly more​​​‌ efficient at higher frequencies.‌ In the future, as‌​‌ micro-robot sizes become smaller​​ than cm- or even​​​‌ mm-scales, maintaining optimal communication‌ range will require the‌​‌ use of higher operating​​ frequencies.
• Many of these​​​‌ studies could benefit from‌ non-standards compliant and non-commercially‌​‌ available chips. To determine​​​‌ the effectiveness of custom​ RF localization solutions, we​‌ will use the Xilinx​​ RFSoC software defined radios​​​‌ (SDRs) to rapidly prototype​ algorithms in both the​‌ 2.4 GHz and the​​ mm-wave ISM bands. With​​​‌ external frequency conversion and​ amplification using off-the-shelf components,​‌ the same SDR can​​ be used as a​​​‌ back-end for prototyping localization​ and communication algorithms with​‌ millimeter carriers.

Results:​​ The results of A1.3​​​‌ determine, both theoretically and​ practically, the performance limitations​‌ of low-power wireless devices​​ in RF-based localization, both​​​‌ in the low (2.4​ GHz, 5 GHz) ISM​‌ bands, and in the​​ mm-wave ISM bands (24​​​‌ GHz and 60 GHz).​ Furthermore, there are implications​‌ on the quality of​​ low-power wireless networking in​​​‌ both power- and volume-constrained​ wireless devices. The extension​‌ of both localization and​​ communication to millimeter wavelengths​​​‌ also generate results in​ small-scale ad-hoc mesh networking,​‌ which, combined with results​​ from A2, have implications​​​‌ for the future of​ massive-scale communication for mobile​‌ wireless nodes.

[A2.1]​​​‌ Swarm Behavior and Mobility​

The network topology of​‌ a swarm of micro-robots​​ is dynamic because the​​​‌ robots move and have​ a short communication range.​‌ This is in stark​​ contrast to traditional TSCH​​​‌ networks, which are static​ and stable. The question​‌ we want to answer​​ is hence: can we​​​‌ use TSCH for networking​ a swarm of micro-robots​‌ where each node in​​ the network is mobile?​​​‌ Because the state-of-the-art is​ very limited, answering this​‌ question requires us to​​ reinvent both the scheduling​​ and routing approach in​​​‌ a TSCH-based networking stack‌ to support mobility.

This‌​‌ leads to the following​​ three studies:

• We introduced​​​‌ the concept of “autonomous‌ cells” in 64 as‌​‌ a bootstrapping mechanism for​​ a 6TiSCH network: each​​​‌ node has a “rendezvous”‌ cell in the schedule,‌​‌ the position of which​​ is computed by applying​​​‌ a hash function to‌ its address. Neighbor nodes‌​‌ thereby know at what​​ time and on which​​​‌ frequency it is listening.‌ Can we extend the‌​‌ concept of autonomous cells​​ to support mobility? Since​​​‌ no signaling is needed‌ to set up this‌​‌ cell, nodes can move​​ without communication overhead from​​​‌ re-scheduling. We can for‌ example allow nodes to‌​‌ change the number of​​ autonomous cells it has​​​‌ (adapting for changes in‌ traffic), and communicate that‌​‌ number for example in​​ Enhanced Beacons. We can​​​‌ evaluate this extension of‌ the 6TiSCH standard on‌​‌ the 6TiSCH simulator, and​​ answer the question: what​​​‌ is the speed limit‌ of the micro-robots at‌​‌ which point the network​​ cannot adapt to the​​​‌ resulting topological changes? That‌ speed limit depends on‌​‌ the communication range. Furthermore,​​ we need to look​​​‌ at the overhead associated‌ with a robot losing‌​‌ connectivity to the swarm:​​ how long can it​​​‌ lose connectivity while staying‌ synchronized?How long does‌​‌ it take for it​​ to reconnect?
• Micro-robots exchange​​​‌ data in a peer-to-peer‌ fashion, rather than all‌​‌ sending data to a​​ single collection point as​​​‌ in a sensor network.‌ The first question we‌​‌ want to answer is:​​ does the peer-to-peer mode​​​‌ of an IoT routing‌ protocols such as RPL‌​‌ apply to such a​​ dynamic network? Given our​​​‌ previous work on this‌ protocol the answer is‌​‌ most likely “no”, and​​ we will need to​​​‌ take a new approach.‌ We worked on the‌​‌ concept of “virtual coordinates”​​ 102, some of​​​‌ which we have used‌ to create RPL. The‌​‌ question is: can we​​ re-purpose the concept of​​​‌ virtual coordinates to support‌ peer-to-peer communication in a‌​‌ micro-robotics context? We will​​ extend RPL to have​​​‌ multiple DODAGs: each potential‌ destination is the root‌​‌ of a new DODAG,​​ which it announces by​​​‌ sending RPL DIOs. We‌ then use reinforcement-based learning‌​‌ to maintain only the​​ DODAGs which are actively​​​‌ used.
• The IoT-lab testbed‌ Inria manages includes robots.‌​‌ The Lille deployment contains​​ 64 Turtlebot2 robots. These​​​‌ are large robots approximately‌ 40 cm high, but‌​‌ are an ideal ready-to-use​​ platform for experimentally evaluating​​​‌ the scheduling and routing‌ approaches described above. We‌​‌ will equip these with​​ micro-motes, and verify the​​​‌ performance of the micro-mote-based‌ mesh network as we‌​‌ have the robots move​​ on a hardcoded track.​​​‌ The question we want‌ to answer is: what‌​‌ are the mobility patterns​​ which stress the mobile​​​‌ mesh network the most?‌ There is a trade-off‌​‌ between the mobility algorithm​​ in the swarm which​​​‌ stretches the robots away‌ from one another to‌​‌ progress fast, and the​​ TSCH network which operates​​​‌ best when each robot‌ has many neighbors.

Results‌​‌: From a scientific​​​‌ point of view, a​ completely new approach to​‌ scheduling and routing in​​ TSCH networks, which are​​​‌ evaluated both in simulation​ and experimentally on a​‌ testbed. These new scheduling​​ and routing protocols are​​​‌ standardized in the 6TiSCH​ and ROLL IETF working​‌ groups. From a project​​ point of view, A2.1​​​‌ develops the networking stack​ which is the stepping​‌ stone for A3.1 (which​​ adds localization) and the​​​‌ experimental validation.

[A2.2] Wireless​ Control Loops and Latency​‌ Predictability

As we have​​ shown in 61,​​​‌ implementations of TSCH networks​ such as Analog Devices'​‌ SmartMesh IP guarantee delivery.​​ That is, a TSCH​​​‌ network guarantees that data​ reaches the destination. The​‌ catch is that it​​ does not guarantee when​​​‌. Given the unreliable​ nature of wireless, this​‌ makes sense: if my​​ neighbor did not get​​​‌ my frame, I retransmit​ until it does. One​‌ can even argue that,​​ given an infinite amount​​​‌ of time and a​ connected network, only an​‌ implementation bug can justify​​ not having 100% reliability.​​​‌ The next question is:​ can a TSCH network​‌ guarantee latency? The answer​​ to that is “no”,​​​‌ since there is always​ a non-zero probability of​‌ an infinite amount of​​ retries to happen on​​​‌ a link that has​ a packet delivery ratio​‌ strictly below 100%. The​​ bold and ambitious question​​​‌ this work package aims​ to answer is: can​‌ TSCH network offer predicable​​ latency and be used​​​‌ to run control loops?​ We have anecdotally shown​‌ in 94 that the​​ control loop of an​​​‌ inverted pendulum can run​ through a TSCH network,​‌ but without rigorous proof.​​

This objective leads to​​​‌ the following 3 studies:​

• For the latency of​‌ a network to be​​ predictable, the easiest is​​​‌ that it does not​ depend on the amount​‌ of traffic in the​​ network. In a TSCH​​​‌ context, this means that​ the schedule is collision-free.​‌ While collision-free schedule is​​ straightforward when using a​​​‌ centralized scheduler, the question​ is: can we achieve​‌ collision free scheduling in​​ a distributed setting such​​​‌ as a swarm of​ micro-robots? One way of​‌ answering is to solve​​ the following mathematical challenge:​​​‌ find a whitening function​ which turns a small​‌ number into a set​​ of cells in a​​​‌ schedule, in such a​ way that any two​‌ different numbers result in​​ disjoint sets of cells.​​​‌ If we can solve​ this, we can assign​‌ a unique number to​​ each node in the​​​‌ network (possibly during secure​ join), which the whitening​‌ function turns in a​​ collision-free schedule.
• Minet et​​​‌ al. 80 have shown​ on very simple topologies​‌ that it is possible​​ to turn a connectivity​​​‌ graph and a schedule​ into a distribution of​‌ latencies, for stable topologies​​ and convergecast traffic. This​​​‌ is similar to the​ repetition strategy used by​‌ 3GPP in the NB-IoT​​ protocol. Can we extend​​​‌ that work to take​ into account a changing​‌ topology, changing schedules and​​ any-to-any traffic? While the​​​‌ approach involves probabilistic analysis,​ it most likely results​‌ in a computational approach.​​ This tool can then​​ be used to explore​​​‌ trade-offs between throughput and‌ power consumption on the‌​‌ one hand, and average​​ latency and latency distribution​​​‌ on the other.
• The‌ two studies above are‌​‌ a necessary mathematical foundation,​​ but can we turn​​​‌ that mathematical foundation into‌ working scheduling approach? Consistency‌​‌ cannot be guaranteed in​​ a practical setting, i.e.​​​‌ nodes do not all‌ have the same information‌​‌ at the same time,​​ which we need to​​​‌ take into account using‌ protocol engineering. To evaluate‌​‌ this protocol, we can​​ define a control loop​​​‌ on the IoT-lab robots,‌ for example a maximum‌​‌ round-trip time between two​​ mobile nodes, and verify​​​‌ that our scheduling approach‌ successfully closes that loop,‌​‌ while we control the​​ movement of the robots.​​​‌

Results: From a‌ scientific point of view,‌​‌ a whitening function that​​ serves as the cornerstone​​​‌ for a collision-free scheduling‌ algorithm with a predictable‌​‌ latency distribution. This whitening​​ function is formally proven,​​​‌ while the overall scheduling‌ approach is implemented and‌​‌ exercised on different scenarios​​ on the IoT-lab robots.​​​‌ From a project point‌ of view, A2.2 develops‌​‌ the scheduling aspect of​​ the protocol stack which​​​‌ allows micro-robots to communicate‌ with predictable latency.

[A2.3]‌​‌ Agile Networking

Today's low-power​​ wireless devices typically consist​​​‌ of a micro-controller and‌ a radio. The most‌​‌ commonly used radios are​​ IEEE802.15.4 2.4 GHz, IEEE802.15.4g​​​‌ sub-GHz and LoRA (SemTech)‌ compliant. Radios offer a‌​‌ different trade-off between range​​ and data-rate, given some​​​‌ energy budget. To make‌ things more complex, standards‌​‌ such IEEE802.15.4g include different​​ modulations schemes (2-FSK, 4-FSK,​​​‌ O-QPSK, OFDM), further expanding‌ the number of options.‌​‌

“Agile Networking” is the​​ concept we are developing​​​‌ which redefines a low-power‌ wireless device as having‌​‌ multiple radios, which it​​ can possibly use at​​​‌ the same time. That‌ is, in a TSCH‌​‌ context, for each frame​​ a node sends, it​​​‌ can change the radio‌ it is using, and‌​‌ its setting. If the​​ next hop is close,​​​‌ it sends the frame‌ at a fast data‌​‌ rate, thereby reducing the​​ radio on-time and the​​​‌ energy consumption. If the‌ next hop is far,‌​‌ it uses a slower​​ data rate.

The first​​​‌ challenge was hardware support.‌ With our input, the‌​‌ OpenMote company designed the​​ OpenMote B, which contains​​​‌ both a CC2538 IEEE802.15.4‌ radio, and an AT86RF215‌​‌ IEEE802.15.4g radio, offering communication​​ on both 2.4 GHz​​​‌ and sub-GHz frequency bands,‌ 4 modulation schemes, and‌​‌ data rates from 50​​ kbps to 800 kbps.​​​‌

The second challenge is‌ to redesign the protocol‌​‌ stack in a standards-compliant​​ way. We are working​​​‌ on a 6TiSCH design‌ in which neighbor discovery‌​‌ happens independently on each​​ radio, and the same​​​‌ neighbor node can appear‌ as many times in‌​‌ the neighbor table as​​ it has radios. The​​​‌ goal is to standardize‌ an “Agile 6TiSCH” profile,‌​‌ without having to touch​​ the core specifications. Jonathan​​​‌ Munoz has co-authored an‌ Internet Draft which details‌​‌ the impact agile networking​​ has on the IETF​​​‌ 6TiSCH protocol stack. This‌ is being implemented in‌​‌ OpenWSN by Mina Rady.​​​‌ The next step is​ to evaluate the performance​‌ of the solution.

[A3.1] Lightweight Protocols​ and System-level Integration

The​‌ last couple of years​​ have witnessed a significant​​​‌ progress in secure communication​ protocols for the IoT.​‌ The IETF has taken​​ steps in standardizing new​​​‌ solutions for protecting the​ communication channel (e.g. OSCORE,​‌ TLS 1.3) and 3-party​​ authorization protocols (e.g. ACE​​​‌ framework). These new solutions​ have been demonstrated as​‌ much more efficient than​​ their predecessors (e.g. TLS​​​‌ 1.2, OAuth 2 as​ used in the Web),​‌ and are expected to​​ be deployed with the​​​‌ next generation of IoT​ products 88, 68​‌. There are a​​ couple of remaining pieces​​​‌ to complete the IoT​ puzzle. One of those​‌ pieces is the LAKE​​ protocol – to be​​​‌ standardized by the group​ we co-chair in the​‌ IETF – a lightweight​​ authenticated key exchange protocol​​​‌ for IoT. As an​ important building block, the​‌ LAKE protocol is expected​​ to enable key exchange​​​‌ in the most constrained​ Internet-of-Things use cases 100​‌.

A common assumption​​ for these communication security​​​‌ solutions is that the​ trust relationship between the​‌ different entities involved in​​ the communication has already​​​‌ been established through for​ example common keying material,​‌ root trust certificates. At​​ manufacturing time, the trust​​​‌ relationship is typically established​ between the IoT device​‌ and the manufacturer. However,​​ the domain where the​​​‌ IoT device will be​ installed is not known​‌ at manufacturing time, and​​ before the IoT device​​​‌ can join a given​ domain, it needs to​‌ be provisioned with domain-specific​​ credentials. Bootstrapping this trust​​​‌ relationship between the IoT​ device and the domain​‌ owner is typically considered​​ out-of-scope for the standards​​​‌ bodies, yet it is​ a non trivial task​‌ as IoT devices lack​​ a user interface. Companies​​ typically resort to out-of-band​​​‌ channels (e.g. NFC, ad-hoc‌ wireless network, pre-shared keys‌​‌ printed on the back​​ of a device, serial​​​‌ port) or proximity-based authentication,‌ requiring the user to‌​‌ go through a cumbersome​​ process of installing a​​​‌ new IoT device. This‌ opens up various vulnerabilities‌​‌ as the “bootstrapping” protocol​​ ends up being designed​​​‌ in-house, without a thorough‌ review of the community‌​‌ and security experts.

One​​ challenge is to enable​​​‌ a solution that allows‌ an IoT device to‌​‌ join (mutually authenticate, authorize,​​ be provisioned with parameters)​​​‌ a network in a‌ new administrative domain, with‌​‌ zero pre-configuration of the​​ IoT device required by​​​‌ the user 97.‌ One cornerstone component of‌​‌ such a solution is​​ the LAKE protocol 96​​​‌. The open research‌ questions include the provisioning‌​‌ of network bandwidth for​​ initial bootstrapping in a​​​‌ zero-touch manner, efficient but‌ flexible transport of public-key‌​‌ certificates.

With LAKE standardization​​ under way, expectations are​​​‌ high in that the‌ working group will provide‌​‌ an efficient key exchange​​ solution for IoT that​​​‌ has been missing. This‌ opens up questions on‌​‌ how the LAKE protocol​​ compares to TLS 1.3​​​‌ in terms of security‌ and performance, which we‌​‌ plan on answering. Also,​​ in collaboration with the​​​‌ Inria PROSECCO team, we‌ work on a formally‌​‌ verified implementation of the​​ LAKE protocol in the​​​‌ OpenWSN environment, similarly to‌ what has been done‌​‌ during the standardization of​​ TLS 1.3 54.​​​‌

Software update mechanisms are‌ being standardized by the‌​‌ IETF 81. Their​​ use to patch vulnerabilities​​​‌ is primordial in constrained‌ environments to improve the‌​‌ reputation IoT products have​​ in terms of security.​​​‌ We plan to study‌ these mechanisms 104 in‌​‌ the context of 6TiSCH​​ networks to improve their​​​‌ performance and make it‌ approachable to product designers.‌​‌

Results: From a​​ scientific point of view,​​​‌ we plan on delivering‌ a turn-key, open-source solution‌​‌ for network access of​​ constrained devices, which does​​​‌ not require user input‌ at deployment time. We‌​‌ plan on publishing a​​ comparative study focusing on​​​‌ performance in constrained environments‌ between the LAKE protocol‌​‌ and TLS 1.3. We​​ also plan on implementing​​​‌ the software update mechanism‌ in 6TiSCH networks and‌​‌ improving it to support​​ software updates of large​​​‌ networks. The envisioned work‌ includes both the derivation‌​‌ of new algorithms and​​ protocols, as well as​​​‌ the optimization of existing‌ solutions.

[A3.2] Microrobot Swarm‌​‌ Security

We plan on​​ exploring the applicability of​​​‌ security mechanisms developed and‌ standardized as part of‌​‌ A3.1 with the swarms​​ of constrained micro-robots. There​​​‌ are several challenges that‌ we envision on such‌​‌ a path.

Highly dynamic​​ logical topologies: Considering the​​​‌ structure such as a‌ swarm of micro-robots moving‌​‌ through space, the wireless​​ links between them are​​​‌ expected to have time-variant‌ quality. As a consequence,‌​‌ we can expect highly​​ dynamic logical topologies between​​​‌ nodes in the network.‌ In such conditions, where‌​‌ a node constantly discovers​​ and communicates with new​​​‌ neighbors, how do we‌ ensure that adequate pairwise‌​‌ cryptographic keys are in​​​‌ place? This problem is​ similar to what is​‌ encountered in Mobile Ad-hoc​​ Networks (MANETs), but the​​​‌ constraints of SCuM nodes​ are much higher. Can​‌ we use the protocol(s)​​ standardized by the LAKE​​​‌ standardization working group, or​ channel-anonymity-based solutions 95?​‌

Securing Localization: Being able​​ to localize nodes within​​​‌ a swarm is an​ essential feature from the​‌ application standpoint. To that​​ end, as outlined in​​​‌ A4.2, we plan on​ using an approach with​‌ mobile lighthouses 76,​​ each equipped with lasers​​​‌ that periodically sweep the​ surrounding space. How to​‌ protect such a system​​ from (accidental) Denial of​​​‌ Service attacks where the​ attacker randomly points a​‌ laser towards the network?​​Can we use the​​​‌ fact that mobile lighthouses​ are equipped with radio​‌ transceivers and can communicate​​ with the SCuM nodes?​​​‌ We plan on studying​ whether lighthouses can use​‌ the radio channel to​​ authenticate their broadcasts (e.g.​​​‌ using TESLA-like solutions 87​) and exchange the​‌ supplementary information in order​​ to pseudo-randomly change the​​​‌ sweep pattern, such that​ the presence of the​‌ attacker does not disturb​​ the network's localization feature.​​​‌

Absence of a stable​ clock source on SCuM:​‌ The fact that SCuM​​ has no external components​​​‌ means that it also​ has no crystal oscillator​‌ to use as a​​ stable clock source. This​​​‌ changes the very basics​ of how wireless networking​‌ is done 62.​​ The work laid out​​​‌ in in A1.2 aims​ at using the network​‌ of micro-robots to provide​​ a relatively stable clock​​​‌ source to each individual​ SCuM. Instead of trusting​‌ its local clock, the​​ nodes in the IoT​​​‌ network must now additionally​ use the network-provided information.​‌ However, this opens up​​ an attack vector where​​​‌ the attacker can disturb​ the network by simply​‌ heating up some of​​ the nodes and changing​​​‌ their clock drift. We​ plan on exploring the​‌ use of machine learning​​ techniques on network drift​​​‌ patterns in order to​ design an Intrusion Detection​‌ System (IDS) to detect​​ nodes under attack.

Results​​​‌: This research axis​ plans on closely following​‌ the team developments on​​ micro-robot swarms and ensuring​​​‌ that the appropriate mechanisms​ are secure-by-design. To that​‌ end, the challenges presented​​ serve as an example​​​‌ of the scientific studies​ we plan on pursuing.​‌

[A4.1] Mathematical​​​‌ Framework for Constrained Localization‌

Localization is key to‌​‌ any robotic application, and​​ many solutions have been​​​‌ developed. Out of those,‌ lighthouse bearing measurement and‌​‌ ultrasonic range measurements are​​ simple enough sensors that​​​‌ they can be integrated‌ in a subset of‌​‌ micro-robots relatively easily. Wheeler​​ et al. 103 have​​​‌ shown that SCuM can‌ detect the laser from‌​‌ a commercial lighthouse. In​​ parallel to a more​​​‌ experimental work, this research‌ area looks at the‌​‌ mathematical framework for constrained​​ localization.

In this work,​​​‌ we assume all robots‌ can be equipped with‌​‌ a lighthouse and/or an​​ ultra-sonic transducer, which allows​​​‌ them to measure relative‌ bearings and distance. The‌​‌ goal is to localize​​ each robot, possibly in​​​‌ a coordinate system which‌ is relative to the‌​‌ swarm. Yet, what is​​ the mathematical framework for​​​‌ turning local bearing and‌ distance measurements into localization,‌​‌ and what is the​​ resulting localization accuracy? We​​​‌ combine mathematical modeling and‌ simulation to answer:

• Assuming‌​‌ all nodes are equipped​​ with a lighthouse, they​​​‌ can measure the relative‌ bearing to one-another. It‌​‌ is well understood that​​ having all relative bearing​​​‌ measurements is enough information‌ to localize all nodes‌​‌ to one another 73​​. The challenge is​​​‌ that, in any practical‌ scenario, each bearing measurement‌​‌ comes with some error,​​ and not all measurements​​​‌ happen at the same‌ time. We first consider‌​‌ a simple case where​​ all micro-robots are within​​​‌ lighthouse range of one‌ another: given the exact‌​‌ position of two “global​​ anchor” micro-robots, how accurate​​​‌ can we get the‌ location of all other‌​‌ robots, and how does​​ that change with having​​​‌ more measurements? This study‌ involves state estimation. We‌​‌ formulate the state estimation​​ model using an Extended​​​‌ Kalman Filter (EKF) to‌ answer the following questions:‌​‌ How many bearing measurements​​ do I need to​​​‌ have a localization measurement‌ better than X mm?‌​‌How does the inaccuracy​​ of the bearing measurement​​​‌ impact the location? We‌ introduce this mathematical model‌​‌ in a simulation with​​ robots far enough from​​​‌ each other than they‌ form a “multi-hop” topology‌​‌ rooted in these two​​ global anchors.
• We add​​​‌ local distance measurements from‌ ultrasonic transducers to our‌​‌ model, and use our​​ EKF for sensor fusion.​​​‌ This allows us to‌ answer the following question:‌​‌ how much more accurate​​​‌ is the localization if​ we combine bearing with​‌ distance measurements, compared to​​ bearing alone, or distance​​​‌ alone? This helps us​ navigate the cost/accuracy trade-off,​‌ and compare our EKF-based​​ methodology with well-established literature​​​‌ on localization using ultra-sonic​ measurements 83, 67​‌.
• Robots move, which​​ limits the number of​​​‌ bearing measurements for each​ location. By introducing movement​‌ in the simulator, we​​ can answer the following​​​‌ questions: what is the​ mobility pattern (maximum velocity,​‌ pause period, etc.) which​​ yields an appropriate trade-off​​​‌ between speed of progression​ of the swarm and​‌ localization accuracy?
• It is​​ unnecessary to equip each​​​‌ robot with a lighthouse​ (the laser transmitter); a​‌ heterogeneous swarm is possible,​​ in which a portion​​​‌ of robots are equipped​ with a laser transmitter,​‌ the others with only​​ a photodiode. The question​​​‌ becomes: what is the​ trade-off between the portion​‌ of laser-equipped robots and​​ localization accuracy?

Results:​​​‌ From a scientific point​ of view, A4.1 creates​‌ the mathematical framework for​​ a localization solution which​​​‌ combines lighthouse and ultrasonic​ range measurements, in an​‌ infrastructure-free, distributed and mobile​​ context. A state estimation​​​‌ and sensor fusion approach​ allows us to explore​‌ trade-offs between accuracy and​​ cost, and understand the​​​‌ impact of robot movement,​ and of the portion​‌ of lighthouse robots on​​ localization accuracy. From a​​​‌ project point of view,​ A4.1 allows us to​‌ decide what hardware to​​ build for experimentation. The​​​‌ model we develop in​ A4.1 is used as-is​‌ in A4.2.

[A4.2] Localization​​ and Network Stack Co-Design​​​‌

The goal of A4.2​ is to co-design the​‌ localization solution (the model​​ is built in A4.1)​​​‌ and the networking stack.​ There are two aspects​‌ to this. On the​​ one hand, the network​​​‌ puts constraints on the​ localization system, in particular​‌ on the amount of​​ data that can be​​​‌ exchanged per period of​ time, and the associated​‌ latency. On the other​​ hand, the synchronized and​​​‌ scheduled nature of the​ networking stack presents a​‌ tremendous opportunity for the​​ localization solution: coordinating when​​​‌ the different sensors are​ on, yielding a better​‌ coexistence and power savings.​​ Similar system-level studies can​​​‌ be found 84,​ 83, 67,​‌ yet none to be​​ best of our knowledge​​​‌ focuses on extremely constrained​ micro-robots. The system-level questions​‌ we want to answer​​ is: What is the​​​‌ overhead of localization on​ the network?How scalable​‌ is a lighthouse and​​ ultrasound-based localization?How low-power​​​‌ can a mote participating​ in the network and​‌ the localization be? We​​ use simulation and modeling.​​​‌

• For a lighthouse, using​ the laser consumes power,​‌ and makes it harder​​ for another lighthouse to​​​‌ also have its laser​ on. The same holds​‌ for the ultra-sonic transducers.​​ Can we use the​​​‌ synchronized nature of the​ network to schedule when​‌ each lasers and ultrasonic​​ transducers are on, in​​​‌ such a way that​ only one pair of​‌ close nodes measures their​​ relative bearing and distance​​​‌ at any given time?​ The scheduled nature of​‌ these measurements has two​​ immediate advantages. First, the​​ swarm scales to more​​​‌ robots, as a collision-free‌ localization schedule can be‌​‌ injected. Second, the robots​​ switch their lighthouse and​​​‌ ultrasound only exactly when‌ they know they will‌​‌ be ranging, resulting in​​ ultra-low power operation, key​​​‌ for immobile robots that‌ want to extend their‌​‌ battery lifetime.
• Each node​​ knows a relative bearing​​​‌ and distance to each‌ or its neighbors. Can‌​‌ this information be shared​​ in such a way​​​‌ that all nodes can‌ compute their location? This‌​‌ location can be computed​​ in a coordinate system​​​‌ that is local to‌ the swarm. The goal‌​‌ of this task is​​ to extend the 6TiSCH​​​‌ protocols with a mechanism‌ to share local measurements,‌​‌ and a distributed localization​​ algorithm.
• The protocol resulting​​​‌ from the previous task‌ necessarily comes with some‌​‌ latency, which results in​​ inconsistency between the view​​​‌ of that information. What‌ is the impact of‌​‌ this inconsistency on location​​ accuracy?

Each of these​​​‌ studies results in a‌ new algorithm or protocol,‌​‌ which is first analyzed​​ then evaluated through simulation.​​​‌

Results: From a‌ scientific point of view,‌​‌ this research has the​​ potential of deeply changing​​​‌ indoor localization as it‌ develops a full RTLS‌​‌ using micro-robots with extreme​​ constraints, in particular in​​​‌ a heterogeneous setup. The‌ result is a method‌​‌ by which a swarm​​ of micro-robots localizes as​​​‌ it progresses through some‌ space. From a project‌​‌ point of view, A4.2​​ is the last building​​​‌ block to realize the‌ exploration and mapping expedition,‌​‌ including experimentally.

[A4.3] Coordination​​ & Control of a​​​‌ Robotic Swarm

Two important‌ considerations when programming large‌​‌ numbers of tiny cyber-physical​​ agents is: what is​​​‌ the easiest way to‌ program them, either individually,‌​‌ or en masse, and,​​ perhaps more importantly, how​​​‌ can a programmer debug‌ them? When the platforms‌​‌ themselves are on the​​ mm or cm scale,​​​‌ the mass and volume‌ requirements to make physical‌​‌ contact are prohibitive. The​​ fact that the robots​​​‌ could move during live‌ programming or debugging.

Four‌​‌ different approaches to contact-free​​ programming will be compared​​​‌ for various swarm robotic‌ applications. The first two‌​‌ are optical: both focused​​ optical communication, and large-scale​​​‌ optical communication. The second‌ two are electromagnetic: near-field‌​‌ communication (NFC) and far-field​​ communication over a wireless​​​‌ data link. All of‌ these have been used‌​‌ in the past to​​ program cyber-physical systems, but​​​‌ we propose to perform‌ a comprehensive survey on‌​‌ their reliability (effectively, program​​ error rate). Furthermore, there​​​‌ is little effort on‌ how these communication systems‌​‌ can be used for​​ debugging, along the lines​​​‌ of a wireless JTAG‌ interface. And, most importantly,‌​‌ we propose to study​​ how these communication interfaces​​​‌ scale in performance with‌ severe volume limitations. As‌​‌ an example, received optical​​ power scales linearly with​​​‌ diode area, assuming uniform‌ illumination.

For contact-free debugging,‌​‌ the problem is less​​ constrained, as it is​​​‌ difficult to quantify how‌ “easy” a system is‌​‌ to work with. Because​​ physical access is impossible,​​​‌ not all on-board voltages‌ and logic levels are‌​‌ accessible. But, it is​​​‌ feasible to create a​ back-and-forth communication between the​‌ robot's on-board processor and​​ the programmer where certain​​​‌ logic levels and registers​ can be observed remotely,​‌ aiming for a wireless​​ JTAG. This may not​​​‌ be realistic in the​ en-masse optical programming, although​‌ these robots could use​​ diagnostic LEDs that can​​​‌ be read by the​ user (at low data-rate).​‌ The point-to-point laser programmer​​ can also receive data​​​‌ from an on-board LED.​ The far-field RF communication​‌ presumably has a link​​ already established, so debugging​​​‌ can be performed over​ any communication standard that​‌ the robot normally uses​​ (earlier examples of Bluetooth​​​‌ or IEEE 802.15.4 are​ both valid). The near-field​‌ programming could be modified​​ to incorporate an RFID​​​‌ style backscattering to send​ data from the device​‌ to the programmer.

Metrics,​​ like power consumption, programming​​​‌ time, reliability (program error​ rate) and debug latency​‌ will all be concretely​​ measured and compared. User​​​‌ safety, particularly in the​ case of IR programming​‌ for point-to-point optical, or​​ heating due to the​​​‌ potentially low efficiency of​ near-field capacitive or inductive​‌ programming, will also be​​ considered. A more holistic​​​‌ survey of experienced embedded​ systems engineers will also​‌ be performed to determine​​ which programming strategy is​​​‌ most desirable, from the​ user's perspective, in various​‌ applications (single robot, two​​ robot, and many robot,​​​‌ with either a uniform​ code-base, or a diverse​‌ and heterogeneous code-base).

[A5.1] Improvement​‌ of communications protocols

The​​ standard IEEE 802.11p protocol​​​‌ has been shown no​ to scale properly the​‌ density of vehicles rapidly​​ varies. In EVA, we​​​‌ have proposed enhancement of​ the IEEE 802.11p access​‌ scheme by considering an​​ adaptive carrier sense level​​​‌ 56, 50.​ The idea is to​‌ create local communication and​​ to allow the network​​​‌ traffic to scale with​ the density of the​‌ vehicle. Detailed proposals have​​ been developed during Younes​​​‌ Bouchaala thesis 55.​ Another approach is to​‌ use a TDMA approach.​​ The main concept is​​​‌ to use the position​ of vehicles on the​‌ roads to control the​​ slots allocations 70,​​​‌ 72. This technique​ allows to drastically reduce​‌ packets collision. We can​​ use a decentralized (possibly​​​‌ using cluster heads) or​ a centralized approach assisted​‌ by roadside units 71​​. These studies have​​​‌ been carried during Mohamed​ Hadded thesis 69.​‌ We are currently working​​ to improve these approaches.​​​‌ The idea is to​ use active signaling techniques​‌ in combination with TDMA​​ approaches 58. The​​​‌ active signaling techniques work​ as an advanced CSMA​‌ scheme and thus bring​​ to the protocol the​​ advantage of random access​​​‌ scheme. We can thus‌ benefit from the stability‌​‌ of the TDMA approaches​​ whereas the active signaling​​​‌ scheme allows the protocol‌ to reduce the collision‌​‌ and offer low latency​​ access when required 60​​​‌, 57, 59‌. This present work‌​‌ will Fouzi's Boukhalfa thesis​​ whose defense is scheduled​​​‌ in October 2021. We‌ can probably even improve‌​‌ your protocol if we​​ use another communication medium​​​‌ such as the visible‌ light. We have started‌​‌ to propose a new​​ architecture which uses simultaneously​​​‌ visible light and radio‌ communication. The smart combination‌​‌ of these two media​​ will be on the​​​‌ focus of our work‌ during the next research‌​‌ period.

We have to​​ study if the visible​​​‌ light communication is mature‌ enough to be used‌​‌ in VANETs in replacement​​ of radio links. The​​​‌ question of the performance‌ and the stability of‌​‌ the visible link is​​ not satisfactorily answered. According​​​‌ to the present state‌ of the art vlc,‌​‌ there are significant problems​​ with interference (natural light,​​​‌ car headlight) 75,‌ and beam propagation due‌​‌ to vehicles trajectories and​​ their movements. Thus vlc​​​‌ appears more to be‌ a complementary technology to‌​‌ radio communication than a​​ technology that can be​​​‌ used alone. So we‌ do not know yet‌​‌ if the visible light​​ links can be used​​​‌ reliably or if we‌ can use them only‌​‌ as backup liaison to​​ increase our protocols reliability​​​‌ .

The standard IEEE‌ 802.11p protocol has slowly‌​‌ started its deployment in​​ the real life leaving​​​‌ the door open to‌ operator initiative. 5G has‌​‌ developed an approach for​​ vehicles promising a very​​​‌ low latency access for‌ vehicles. We need to‌​‌ better understand what are​​ the 5G services for​​​‌ vehicular networks and what‌ are their strength and‌​‌ limitations We plan a​​ collaboration with the RITS​​​‌ team to deploy a‌ 5G vehicular network in‌​‌ Rocquencourt. More specifically we​​ plan to use 5G​​​‌ network to send Cooperative‌ Awareness Messages CAMs and‌​‌ Decentralized Emergency Notification Messages​​ and to develop a​​​‌ 5G assisted intersection crossing‌ application. We will try‌​‌ to take benefit of​​ this deployment to evaluate​​​‌ how 5G could scale‌ on a real VANET‌​‌ network.

[A5.2] Towards Autonomous​​ Cars

Autonomous driving is​​​‌ a target followed by‌ many new companies such‌​‌ as Google, Uber, Telsa,​​ and even by older​​​‌ players in the field‌ such BMW, Mercedes, etc.‌​‌ Recent progress has been​​ accomplished but it is​​​‌ still unclear whether whether‌ full autonomous vehicles can‌​‌ be obtained in large​​ amount of different cases​​​‌ or if we have‌ to treat only special‌​‌ case such as driving​​ in platoon. The exact​​​‌ role of communication in‌ an autonomous car also‌​‌ remains to be further​​ studied even the importance​​​‌ of the vehicular communication‌ networks has been acknowledge‌​‌ 78, 101.​​

We wish to start​​​‌ by the study of‌ platoons of vehicle and‌​‌ to design the suitable​​ communication network to ensure​​​‌ a high degree of‌ safety. The idea is‌​‌ to use the concepts​​​‌ of 77 and to​ adapt them to the​‌ protocol AS-DTMAC that we​​ have recently designed. Given​​​‌ the nature and the​ probability of hazards and​‌ assumptions on packet transmission​​ errors, we plan to​​​‌ compute the probability that​ our platoon of vehicles​‌ communicating with our protocol,​​ and according to strict​​​‌ rules, can safely progress.​ The use of VANETs​‌ can be studied in​​ other simple situation such​​​‌ as keeping safe distance​ between vehicles, changing lanes​‌ or inserting in a​​ lane at the entrance​​​‌ of an highway. We​ plan to combine the​‌ use of radio VANETs​​ with other sensing technologies​​​‌ (RADAR, LIDAR, Video) or​ even with visible light​‌ communication to increase the​​ reliability of the system.​​​‌ We believe that as​ any safety system, autonomous​‌ cars have to rely​​ on many different and​​​‌ independent sensing systems to​ be able to ensure​‌ a high degree of​​ reliability. We will have​​​‌ to

We also have​ to study if the​‌ visible light communication is​​ mature enough to be​​​‌ used in VANETs. According​ to the present state​‌ of the art vlc,​​ there are significant problems​​​‌ with interference (natural light,​ car headlight) and beam​‌ propagation due to vehicle​​ trajectories and their movements.​​​‌ Thus vlc appears more​ to be a complementary​‌ technology to radio communication​​ than a technology that​​​‌ can be used alone.​

[A5.3] Machine learning and​‌ VANETs

Vehicular networks can​​ generate a lot of​​​‌ data; the vehicles have​ positioning capabilities (e.g. GPS),​‌ they also have communication​​ devices and computing power.​​​‌ We have shown that​ the power received from​‌ packets transmitted by road​​ side units can be​​​‌ used by machine learning​ algorithms such as Random​‌ Forest (RF) , K​​ Nearer Neighbors (KNN), Neural​​​‌ Networks (NN) to predict​ the position of the​‌ vehicle and performance of​​ the wireless network (e.g.​​​‌ packet delivery ratio) see​ 89, 90,​‌ 93. We have​​ shown that these prediction​​​‌ can obtained even a​ significant portion of the​‌ measurements are lost and​​ that the predictions still​​​‌ remain exploitable. We believe​ that these results remain​‌ to be improved, for​​ instance the use communication​​​‌ data with input of​ other sensor appear to​‌ be very promising. These​​ studies will depend on​​​‌ the availability of large​ amount of vehicle network​‌ data.

It is also​​ possible to use machine​​​‌ learning to forecast accidents.​ Urban traffic forecasting models​‌ generally follow either a​​ Gaussian Mixture Model (GMM)​​​‌ or a Support Vector​ Classifier (SVC) to estimate​‌ the features of potential​​ road accidents. Although SVC​​​‌ can provide good performances​ with less data than​‌ GMM, it incurs a​​ higher computational cost. We​​​‌ have proposed framework that​ combines the descriptive strength​‌ of the Gaussian Mixture​​ Model with the high-performance​​​‌ classification capabilities of the​ Support Vector Classifier. A​‌ new approach is presented​​ that uses the mean​​​‌ vectors obtained from the​ GMM model as input​‌ to the SVC. Experimental​​ results show that the​​​‌ approach compares very favorably​ with baseline statistical methods,​‌ see 92. Advances​​ are possible in forecasting​​ accidents, these progresses depend​​​‌ on the availability of‌ data, in particular covering‌​‌ a wide variety of​​ problems from simple incidents​​​‌ to accidents with injuries‌ to fatalities. It is‌​‌ clear that such a​​ system could be very​​​‌ interesting for a driver‌ who could in dangerous‌​‌ conditions increase his attention​​ and even activate driving​​​‌ aids.

The positioning of‌ the AIO team in‌​‌ machine learning for VANETs​​ consists of using and​​​‌ combining techniques already available‌ and exploiting these techniques‌​‌ in open data sets.​​ This positioning is different​​​‌ from that of the‌ SIERRA team which seeks‌​‌ the design of a​​ new algorithm preferably to​​​‌ solve fundamental problems in‌ networks. For example in‌​‌ AIO we plan to​​ use a customized Deep​​​‌ Learning mechanism-based congestion control‌ identification approach that does‌​‌ not need any enriched​​ domain knowledge other than​​​‌ training traffic of a‌ congestion control variant. By‌​‌ only using packet arrival​​ data, it is also​​​‌ directly applicable to encrypted‌ (transport header) traffic. At‌​‌ the same time, during​​ the customization phase, we​​​‌ will also use deep‌ reinforcement learning to consolidate‌​‌ the congestion control. Trust​​ Region Policy Optimization (TRPO)​​​‌ and proximal policy optimization‌ will be adopted in‌​‌ the proposed customized approach​​ as the measure of​​​‌ optimization.

[A5.4] Security and‌ Privacy in VANETs

Security‌​‌ in VANETs has already​​ been the subject of​​​‌ numerous studies 74,‌ 79. Attacks can‌​‌ be carried out in​​ several places: on the​​​‌ air interface of the‌ network, in the hardware‌​‌ or software of vehicle​​ transmissions, in the vehicle​​​‌ sensors whose information is‌ sent over the network,‌​‌ in the infrastructure of​​ the VANET network. All​​​‌ types of security attacks‌ can be found in‌​‌ vehicular networks and the​​ dynamics of network links​​​‌ add further complexity to‌ the problem. Faced with‌​‌ the difficulty of the​​ problem, VANET networks still​​​‌ have an advantage, the‌ vehicles and infrastructure elements‌​‌ have significant computing power​​ and energy resources. The​​​‌ classic security approach in‌ VANETs is the deployment‌​‌ of a PKI. This​​ approach has been standardized​​​‌ in Europe at ETSI‌ and in the US.‌​‌ This approach does not​​ solve all the issues​​​‌ and gives rise to‌ problems in particular of‌​‌ Privacy. To remedy this,​​ the technique of pseudonyms​​​‌ has been proposed 51‌.

The approach of‌​‌ the AIR project is​​ not to study security​​​‌ in VANETs networks in‌ general but to propose‌​‌ punctual improvements of the​​ state-of-the-art on precise and​​​‌ well-defined security problems.

VLC‌ links could be used‌​‌ between vehicles for communications.​​ We plan to study​​​‌ how the use of‌ such a link could‌​‌ advance security in VANETs.​​ With the nature of​​​‌ VLC links, it is‌ clear that capturing messages‌​‌ or sending of fraudulent​​ packets is almost impossible​​​‌ in the context of‌ point to point VLC‌​‌ links. On the privacy​​ side, the use of​​​‌ VLC links can be‌ very beneficial. We plan‌​‌ to quantify this benefit​​ compared to existing solutions.​​​‌

Another area we want‌ to tackle is the‌​‌ security of routing protocols​​​‌ in VANETs. We have​ started to study security​‌ attacks on cross-layer routing​​ and the benefit of​​​‌ trust against these attacks​ 53, 52.​‌ Ismael Tayssir in her​​ PhD plans to development​​​‌ of a new intelligent​ routing protocol which uses​‌ information from the MAC​​ layer to find an​​​‌ optimized path between the​ transmitter and the receiver​‌ and which takes into​​ account the specificities of​​​‌ the transmission medium while​ minimizing the time of​‌ transmission. In her PhD,​​ detection of malicious behavior​​​‌ at the MAC and​ routing layer will be​‌ proposed to secure the​​ routing protocol developed.

We​​​‌ also have started very​ preliminary studies to use​‌ blockchains in VANETs 91​​. We will continue​​​‌ on this topic and​ will try to determine​‌ if there are cases​​ of applications in VANETs​​​‌ where this technology can​ find its best application.​‌

The RITS team works​​ primarily on security issues​​​‌ in VANETs caused by​ the sending of fraudulent​‌ or erroneous data coming​​ from vehicle sensors. Collaboration​​​‌ between our team and​ RITS in this area​‌ is possible.

[A5.5] FANETs​​

The field of FANETs​​​‌ (Flying Area NETworks) is​ experiencing significant growth in​‌ the academic community and​​ also among manufacturers, particularly​​​‌ the arms manufacturers. This​ area poses many interesting​‌ problems such as maintaining​​ connectivity, multiple access, satisfying​​​‌ quality of service, etc.​ The problems obtained are​‌ often very combinatorial in​​ nature and the arrival​​​‌ of artificial intelligence opens​ a new horizon of​‌ research in this field.​​ AIO has Cifre PhD​​​‌ with Thales.

7.1.1 embedded-cal

• Name:​​
  Embedded Cryptographic Abstraction Layer​​​‌
• Keywords:
  Rust, Cryptography, Embedded​ systems
• Functional Description:
  Embedded-cal​‌ develops a verified implementation​​ of the cryptographic provider​​​‌ in Rust which is​ compatible with popular embedded​‌ platforms. This cryptographic provider​​ will be 1) fast​​​‌ on popular embedded platforms,​ 2) resistant to certain​‌ classes of side-channel attacks,​​ 3) usable without the​​​‌ Rust standard library. The​ module will lever the​‌ available hardware acceleration support​​ of popular microcontroller units​​​‌ for embedded systems and​ fill in the gaps​‌ in hardware support through​​ software implementations. The module​​​‌ will be formally verified​ for secret independence using​‌ the hax framework, a​​ verification tool for high​​​‌ assurance code.
• URL:
  https://­github.­com/­lake-rs/­embedded-cal​
• Contact:
  Malisa Vucinic
• Participants:​‌
  Malisa Vucinic, Elsa Lopez​​ Perez, William Pereira
• Partners:​​​‌
  Cryspen, Christian Amsüss

7.1.2​ lakers

• Name:
  EDHOC implemented​‌ in Rust, optimized for​​ microcontrollers, with bindings for​​​‌ C and Python
• Keyword:​
  EDHOC
• Functional Description:
  EDHOC​‌ is a lightweight authenticated​​ key exchange protocol targeting​​​‌ constrained environments and Internet​ of Things use cases.​‌ This is a Rust​​ implementation of the protocol,​​​‌ adapted for use on​ microcontrollers.
• URL:
  https://­github.­com/­lake-rs/­lakers
• Contact:​‌
  Malisa Vucinic
• Participants:
  Malisa​​ Vucinic, Geovane Fedrecheski, Elsa​​​‌ Lopez Perez

7.1.3 DotBot-hardware​

• Keywords:
  3D printing, Electrical​‌ circuit, Robotics
• Functional Description:​​
  PCB and Mechanical Parts​​​‌ for DotBots
• URL:
  https://­github.­com/­DotBots/­DotBot-hardware​
• Contact:
  Thomas Watteyne

7.1.4​‌ DotBot-firmware

• Keywords:
  Robotics, Embedded,​​ Microcontroller
• Functional Description:
  Source​​​‌ code for firmwares usable​ with the DotBot hardware.​‌
• Release Contributions:
  https://github.com/DotBots/DotBot-firmware/releases/tag/REL-1.14
• URL:​​
  https://­dotbot-firmware.­readthedocs.­io/­en/­latest
• Contact:
  Thomas Watteyne​​

7.1.5 PyDotBot

• Keywords:
  Robotics,​​​‌ Python, Web Application
• Functional‌ Description:
  A complete environment‌​‌ for controlling and visualizing​​ DotBots.
• URL:
  https://­pydotbot.­readthedocs.­io/­en/­latest/
• Contact:​​​‌
  Thomas Watteyne

7.1.6 SwarmIT‌

• Name:
  SwarmIT
• Keywords:
  Swarms,‌​‌ Trusted software, Embedded, ARM,​​ Testbeds
• Functional Description:
  SwarmIT​​​‌ provides a C port‌ for nRF53 of ARM‌​‌ TrustZone for Cortex-M as​​ well as Python based​​​‌ services to easily build‌ and deploy a robotic‌​‌ swarm infrastructure testbed.
• URL:​​
  https://­github.­com/­DotBots/­SwarmIT
• Contact:
  Alexandre Abadie​​​‌

7.1.7 QrKey

• Keywords:
  MQTT,‌ KDF, Middleware protocol interoperability,‌​‌ Swarms
• Functional Description:
  Qrkey​​ is a library implementing​​​‌ a protocol designed to‌ simply and securely faciliate‌​‌ the deployment of robotic​​ swarms
• URL:
  https://­github.­com/­dotbots/­qrkey
• Contact:​​​‌
  Alexandre Abadie

7.2.1 DotBot v3‌​‌

Participants: Alexandre Abadie,​​ Said Alvarado-Marin, Filip​​​‌ Maksimovic, Martina Balbi‌, Trifun Savic,‌​‌ Thomas Watteyne.

The​​ DotBot v3 swarm robot​​​‌ (left) and a prototyping‌ of its drive-through fast‌​‌ charger (right).

The​​ DotBot v3 swarm robot​​​‌ (left) and a prototyping‌ of its drive-through fast‌​‌ charger (right).

Large,​​ coordinated “swarms” of small,​​​‌ resource-constrained robots have the‌ potential to complete complex‌​‌ tasks that single monolithic​​ robots cannot. However, while​​​‌ there is ongoing research,‌ little progress has been‌​‌ made in successfully deploying​​ these swarms in the​​​‌ real world. To help‌ further the field, we‌​‌ are building a research​​ platform called DotBot, shown​​​‌ in Fig. 3:‌ a low-price, versatile laser‌​‌ cut robot that can​​ inexpensively act as an​​​‌ agent in a swarm‌ of robots. Each DotBot‌​‌ has two small motors​​ for mobility, accurate localization​​​‌ using laser lighthouses, and‌ can communicate using off-the-shelf‌​‌ radios in either time-synchronized​​ channel-hopping mesh networks originally​​​‌ designed for reliable transmission‌ in crowded IoT networks,‌​‌ or with BLE so​​ that the robots can​​​‌ be programmed from a‌ cell phone or other‌​‌ Bluetooth-enabled device. We see​​ the DotBot platform as​​​‌ an ideal tool for‌ introducing robotics and embedded‌​‌ programming in education. We​​ target three levels. First,​​​‌ in primary school, DotBot‌ serves as a basic‌​‌ introduction to robotics, using​​ simple interaction and remote-control​​​‌ scenarios. In high school,‌ DotBot is used as‌​‌ an introduction to embedded​​ programming, with a focus​​​‌ on the interaction with‌ the real world. Finally,‌​‌ in university, a DotBot​​ swarm is used to​​​‌ introduce the concepts of‌ distributed algorithms, task assignment‌​‌ as well as planning​​ and scheduling.

We have​​​‌ continued to work hard‌ on the DotBot in‌​‌ 2025. The DotBot remains​​ at the heart of​​​‌ the Horizon Europe OpenSwarm‌ project. More than 100‌​‌ DotBots were manufactured in​​ 2025, and the focus​​​‌ now shifted into making‌ the software reliable and‌​‌ usable at scale. With​​ the help of Alexandre​​​‌ Abadie and Filip Maksimovic‌ , the lighthouse localization‌​‌ has been optimized to​​ run embedded in the​​​‌ DotBots, improving localization responsiveness‌ at scale. This allows‌​‌ more complex and more​​ secure applications to be​​​‌ used on DotBots. Thanks‌ to Alexandre Abadie ,‌​‌ we kept improving the​​​‌ code base and making​ it more reliable, from​‌ low-level drivers to high​​ level web based controls.​​​‌ The version 3 of​ the robot, now using​‌ supercapacitors for fast charge​​ and ecological reasons, was​​​‌ reliably used for a​ combined amount of time​‌ of more than 500​​ hours. This demonstrates that​​​‌ the shift towards design-for-manufacturing​ and system reliability improvements​‌ was a success, making​​ the DotBot platform a​​​‌ useful and reliable tool​ for the team. We​‌ continued the collaboration in​​ terms of hardware design​​​‌ with Prof. Danny Hughes​ from KU Leuven in​‌ Belgium.

More information at​​ www.dotbots.org.

7.2.2 OpenSwarm​​​‌ Testbed

Participants: Geovane Fedrecheski​, Said Alvarado-Marin,​‌ Filip Maksimovic, Thomas​​ Watteyne.

The 1,000​​​‌ DotBot testbed will initially​ be deployed in a​‌ large warehouse, as part​​ of the OpenSwarm project.​​​‌

The 1,000 DotBot testbed​ will initially be deployed​‌ in a large warehouse,​​ as part of the​​​‌ OpenSwarm project.

As​‌ part of the OpenSwarm​​ project, we are building​​​‌ testbed with 1,000 DotBots,​ which enables researchers to​‌ study robotic Swarms in​​ a real and scalable​​​‌ environment. Users are able​ to remotely program the​‌ DotBot Swarm, as well​​ as monitor its status​​​‌ (including robot localization) and​ control the robots in​‌ a real time fashion​​ (individually or collectively). All​​​‌ the software in the​ testbed is designed to​‌ be simple to use​​ and have low overhead,​​​‌ while at the same​ time offering high reliability.​‌ For example, the firmware​​ update mechanism leverages industry-grade​​​‌ resource protection, ensuring that​ the robots always fallback​‌ into a working state.​​ Similarly, by leveraging the​​​‌ team expertise on reliable​ TSCH-based networks, we developed​‌ the Micro-robot Access Radio​​ Infrastructure (Mari), a custom​​​‌ link layer dedicated to​ connecting the 1,000 DotBot​‌ testbed. So far, we​​ used Mari to connect​​​‌ and flash up to​ 200 nodes. In September​‌ 2025 we did the​​ first deployment of the​​​‌ testbed with a first​ batch of 50 robots.​‌ The deployment took place​​ in the Catalyst collaborative​​​‌ space at ADI's campus​ in Limerick, Ireland, where​‌ all consortium members have​​ had the chance to​​​‌ use the testbed for​ the first time. The​‌ next step is to​​ scale the testbed to​​​‌ 1,000 robots, which are​ currently under production.

7.2.3​‌ Espace Experimentation at Inria-Paris​​

Research in the AIO​​​‌ team involves hardware design​ and prototyping. The team​‌ therefore has equipment necessary​​ to assemble PCBs, create​​​‌ mechanical structures and deploy​ and test in real​‌ environments. Fig. 5 shows​​ pictures of some of​​​‌ the equipment of the​ team.

Some of the​‌ equipment in the “espace​​ experimentation” operated by the​​​‌ AIO team.

Some of​ the equipment in the​‌ “espace experimentation” operated by​​ the AIO team.

As a result, the​​ team also develops hardware.​​ A representative example is​​​‌ the AIOT Systems solution.‌ AIOT Systems is a‌​‌ one-stop shop for learning​​ embedded low-power wireless. The​​​‌ AIOT Play board is‌ a ready-to-learn-on platform, designed‌​‌ specifically to be both​​ easy to learn with,​​​‌ and close to an‌ production system. It features‌​‌ a prototyping area allowing​​ a student to build​​​‌ little circuits directly on‌ the board. They then‌​‌ write firmware directly on​​ the fully programmable microcontroller​​​‌ to interact with the‌ circuit, and hand the‌​‌ data to the true​​ mesh networking module. The​​​‌ source code consists of‌ Python code that runs‌​‌ on a computer, and​​ C code that runs​​​‌ on a microcontroller. The‌ source code is developed‌​‌ under an open-source license​​ so students can really​​​‌ see how things are‌ working, and use it‌​‌ beyond learning. We crafted​​ the AIOT Systems Academy​​​‌ so it is a‌ completely self-contained set of‌​‌ course material.

8.1.1​​ GAIA: Bridging Sustainability and​​​‌ Ubiquity in Next-Generation IoT‌

Participants: Filip Maksimovic.‌​‌

In 2025, the GAIA​​ project was submitted and​​​‌ accepted to the highly‌ competitive EIC-PATHFINDER funding program‌​‌ as part of a​​ consortium of five Universities​​​‌ and Research Institutes. GAIA's‌ goal is to leverage‌​‌ major advances in biodegradable​​ electronics and ambient communication​​​‌ to offer a new‌ vision of the Internet‌​‌ of Things (IoT): a​​ transparent, sustainable, and circular​​​‌ IoT ecosystem.

It introduces‌ a new category of‌​‌ microelectronic systems with very​​ low energy consumption that​​​‌ combine sensing, computing, and‌ communication in a unique‌​‌ biodegradable architecture, without batteries,​​ radio transmitters, or persistent​​​‌ microcontrollers. Using 6G ambient‌ IoT backscatter concepts, GAIA‌​‌ devices operate autonomously and​​ interface natively with existing​​​‌ cellular infrastructure. Designed to‌ degrade naturally after use,‌​‌ they offer a battery-free,​​ waste-free solution that combines​​​‌ infrastructure compatibility with environmental‌ responsibility. This marks a‌​‌ key step toward an​​ infrastructure-aligned IoT that avoids​​​‌ long-term ecological impact while‌ maintaining advanced digital functionality.‌​‌ Over three years, GAIA​​ will follow a phased​​​‌ strategy, from component validation‌ to system-level verification, culminating‌​‌ in a proof of​​ concept.

This proof of​​​‌ concept, focused on smart‌ logistics and cold supply‌​‌ chain tracking, will present​​ a fully transient IoT​​​‌ system that is very‌ low cost, low power‌​‌ consumption, and offers seamless​​ cellular interoperability.

This project​​​‌ will enable applications in‌ human and industrial contexts‌​‌ requiring (near) real-time tracking,​​ detection, and identification without​​​‌ compromising sustainability. The project‌ takes a vertically integrated‌​‌ approach, ranging from biodegradable​​ materials to system design​​​‌ and telecommunications infrastructure, paving‌ the way for digital‌​‌ objects at very low​​ cost and with minimal​​​‌ environmental impact.

The project‌ will officially begin in‌​‌ February of 2026, and​​ the full consortium is​​​‌ made up of the‌ following partners:

Partners:

• UNIVERSITAT‌​‌ OBERTA DE CATALUNYA, Spain​​ (Coordinators)
• INSTITUT NATIONAL DE​​​‌ RECHERCHE EN INFORMATIQUE ET‌ AUTOMATIQUE (INRIA), France
• NOVA‌​‌ ID FCT - ASSOCIACAO​​​‌ PARA A INOVACAO E​ DESENVOLVIMENTO DA FCT, Portugal​‌
• CENTRE TECNOLOGIC DE TELECOMUNICACIONS​​ DE CATALUNYA, Spain
• TECHNISCHE​​​‌ UNIVERSITAET BRAUNSCHWEIG (TUB), Germany​
• TAMPEREEN KORKEAKOULUSAATIO SR, Finland​‌

8.1.2 Smart Dust Security​​

Participants: Sara Faour,​​​‌ Filip Maksimovic, Mališa​ Vučinić.

A significant​‌ concern in small-scale low-power​​ embedded systems is security​​​‌ and privacy. This is​ exacerbated by the lack​‌ of non-volatile memory available​​ on chip-scale systems, which​​​‌ make it challenging to​ both generate and store​‌ reliable secret information. This​​ can be mitigated by​​​‌ SRAM (static random access​ memory) PUFs (physically uncloneable​‌ functions). In this work,​​ we develop techniques designed​​​‌ for usage on the​ single-chip mote platform.

Physically​‌ Unclonable Functions (PUFs) extract​​ secret keys from intrinsic​​​‌ hardware, offering strong resistance​ to cloning and physical​‌ tampering. SRAM PUFs reuse​​ memory at start-up but​​​‌ are sensitive to environmental​ noise. We present a​‌ two-stage scheme to extract​​ reliable cryptographic keys from​​​‌ SRAM PUFs. The scheme​ extends our prior single-stage​‌ TMVS approach to reduce​​ memory overhead while maintaining​​​‌ reliability. The method first​ extracts intermediate reliable bits​‌ from raw SRAM responses,​​ then reapplies TMVS to​​​‌ achieve the target key​ error probability 10 -6​‌ with 128-bit keys. The​​ two-stage cascade maintains all​​​‌ the benefits of the​ original design (low-complexity majority​‌ voting decoder, single SRAM​​ measurement, low entropy loss)​​​‌ while reducing the SRAM​ memory overhead by 3.5×​‌ for 7.5% raw bit​​ error rates. Our analysis​​​‌ reveals new error-memory trade-offs,​ demonstrating that the TS-TMVS​‌ maintains practical computational and​​ memory requirements for resource-constrained​​​‌ devices.

In order to​ improve on current methods,​‌ which use complex error​​ correction with NVM-stored helper​​​‌ data or by pre-selecting​ stable cells through repeated​‌ measurements, we propose novel​​ methods for PUF extraction.​​​‌ Avoiding NVM reduces leakage​ risks and manufacturing costs,​‌ benefiting low-end devices without​​ NVM. This strategy, called​​​‌ ODHD (on-demand helper data​ generation) is a self-contained​‌ approach for stabilizing SRAM​​ PUFs without helper data​​​‌ stored in non-volatile memory.​ It uses a simple​‌ decoder and a small​​ number of SRAM readouts​​​‌ to improve reliability.

The​ two-stage TMVS scheme is​‌ published in: 24.​​ ODHD is published in​​​‌ 25.

8.1.3 The​ Single-Chip Mote Software Development​‌ Kit

Participants: Filip Maksimovic​​, Alexandre Abadie.​​​‌

In order to improve​ the single-chip mote build​‌ and programming infrastructure, we​​ introduce the SCuM SDK,​​​‌ an open source and​ multi-platform software development kit​‌ that can be used​​ to build and load​​​‌ a firmware on the​ SCuM chip. The SCuM​‌ SDK is designed to​​ help newcomers, especially those​​​‌ with prior experience with​ ARM, to start programming​‌ SCuM. The SCuM SDK​​ is fully open-source and​​​‌ can program, boot and​ calibrate SCuM much faster​‌ than the existing tools.​​

As an example of​​​‌ the usage of this​ new software development kit,​‌ we demonstrate the first​​ port of a real-time​​​‌ operating system, RIOT, for​ the single-chip mote (SCuM).​‌ RIOT provides SCuM all​​ its features to develop​​​‌ multithreaded embedded applications for​ the Internet of Things.​‌ This demonstration showcases RIOT​​ real-time capabilities using threads,​​ embedded AI to detect​​​‌ hand-written digits in an‌ image, and some cryptographic‌​‌ primitives, all running on​​ SCuM.

The software development​​​‌ kit is published in‌ 16. The port‌​‌ of RIOT-OS is published​​ in 17.

8.1.4​​​‌ Studying Interference Effects in‌ High-Density IoT Networks

Participants:‌​‌ Filip Maksimovic, Diego​​ Badillo.

Bluetooth Low​​​‌ Energy and IEEE 802.15.4‌ are commonly used wireless‌​‌ communication protocols in IoT​​ applications. Both operate in​​​‌ the 2.4 GHz ISM‌ band, where their coexistence‌​‌ can lead to interference.​​ We study the bit​​​‌ error rate (BER) under‌ controlled timed interferences within‌​‌ the physical payload, varying​​ interference power and frequency​​​‌ offsets. Measurements with a‌ co-located software-defined-radio antenna are‌​‌ captured and shown to​​ validate the experimental setup.​​​‌ IEEE 802.15.4 proves more‌ robust to BLE interference‌​‌ than vice versa, consistent​​ with results in literature.​​​‌ Interference at lower relative‌ frequencies is found to‌​‌ have a greater impact​​ on BER. Controlled interference​​​‌ causes persistent bit errors‌ after it ends, suggesting‌​‌ receiver desynchronisation.

These results​​ are published in 21​​​‌.

8.1.5 Interference Mitigation‌ in Low-Power Networks Using‌​‌ Successive Interference Cancellation

Bluetooth​​ Low Energy (BLE) and​​​‌ IEEE 802.15.4 are two‌ widely adopted wireless communication‌​‌ protocols in Internet of​​ Things (IoT) applications. Both​​​‌ operate in the 2.4‌ GHz ISM band, where‌​‌ their coexistence can cause​​ mutual interference. In this​​​‌ demo, we present a‌ proof-of-concept post-processing prototype that‌​‌ recovers colliding packets by​​ employing an effective algorithm​​​‌ based on a coarse-to-fine‌ exhaustive search and cross-correlation‌​‌ for Successive Interference Cancellation​​ (SIC). First, the stronger​​​‌ signal is demodulated, then‌ its reconstructed waveform is‌​‌ subtracted from the captured​​ interfered signal in order​​​‌ to demodulate the weaker‌ signal. This approach allows‌​‌ the demodulation of both​​ protocols despite overlaps in​​​‌ frequency and time. The‌ experimental setup includes three‌​‌ software-defined radios (two transmitters,​​ one per protocol, and​​​‌ one receiver), all connected‌ to the same host‌​‌ computer. This solution is​​ most suitable for gateways,​​​‌ where power consumption and‌ computational resources are less‌​‌ constrained. Its main limitation​​ is that it assumes​​​‌ direct line of sight‌ between transmitters and receiver,‌​‌ so a dominant high-power​​ signal exceeds any multipath​​​‌ reflections. It further requires‌ a sufficiently high SNR‌​‌ to recover the low-power​​ signal after subtraction, and​​​‌ the ADC's quantisation limits‌ the power difference between‌​‌ the two protocols. It​​ demonstrates the potential to​​​‌ increase coexistence throughput in‌ dense IoT deployments.

These‌​‌ results are published in​​ 20.

The GNU​​​‌ Radio–based implementation of the‌ SIC scheme as well‌​‌ as the BLE and​​ IEEE 802.15.4 demodulators is​​​‌ available in 49.‌

8.1.6 Development of Ultra‌​‌ Low-Power Parameterizable RISC-V Processors​​

Participants: Filip Maksimovic,​​​‌ Alfonso Cortes.

There‌ are currently two barriers‌​‌ to the widespread adoption​​ and development of customized​​​‌ chip-scale hardware. The first‌ is the lack of‌​‌ eaisly parameterizable hardwave description​​ (HDL) for open-source processors,​​​‌ such as RISC-V. The‌ standard is public, and‌​‌ many singular open-source solutions​​ exist, but paramterizability is​​​‌ limited. The second, and‌ more significant barrier, is‌​‌ the closed ecosystem around​​​‌ integrated circuit design, verification,​ and manufacturing. In order​‌ to mitigate both of​​ these barriers, we have​​​‌ worked on an easily​ customizable RISC-V processor (the​‌ Rocket core using a​​ high-level generation system called​​​‌ ChipYard), to which we​ are currently adding a​‌ root-of-trust. The RISC-V core​​ is then synthesized and​​​‌ passed through a place​ and route tool (librelane).​‌ This entire process is​​ open-source, resulting in a​​​‌ clean mask set design​ that can be manufactured​‌ in an open-source semiconductor​​ process.

The development of​​​‌ these parameterizable RISC-V cores​ also lays the groundwork​‌ for future endeavors in​​ digital development for the​​​‌ European Project GAIA.

The​ register transfer layer logic​‌ as well as forks​​ of all digital flows​​​‌ (librelane) and RISC-V cores​ (chipyard/rocket) are currently available,​‌ open-source, on GitHub.​​ Publication of results is​​​‌ currently pending.

8.1.7 Wireless​ Over-the-Air Hardware Updates for​‌ Embedded Devices

Participants: Alexandre​​ Abadie, Filip Maksimovic​​​‌.

Attaching field-programmable gate​ arrays to microcontrollers deployed​‌ in the field can​​ increase their potential applications.​​​‌ Making these systems reconfigurable​ post-deployment effectively makes them​‌ future-proof, allowing maintenance and​​ updating of firmware in​​​‌ distributed computing applications. This​ paper presents UpGate, an​‌ embedded software that can​​ securely perform an over-the-air​​​‌ update of an FPGA's​ function with minimal communication​‌ overhead, enabling complete flexibility​​ of the device's functionality.​​​‌ Several benchmarks are performed​ using simple to complex​‌ hardware designs for in-depth​​ analysis of memory footprint,​​​‌ timing, and power consumption.​ The results show that​‌ regular lossless compression techniques,​​ such as GZip and​​​‌ LZ4, with 80% compression​ rate in the worst​‌ case, are reducing the​​ global binary update duration​​​‌ by at least 4​ times. This represents 15​‌ times less 128 B​​ packets transmitted, and approximately​​​‌ 40 s to transfer​ a RISC-V core bitstream​‌ of 579 kB. Surprisingly,​​ what takes the most​​​‌ time and energy is​ writing to flash.

This​‌ work has been published​​ in 15.

8.1.8​​​‌ Smart-Dust Localization

Participants: Filip​ Maksimovic.

The Single-Chip​‌ Micro Mote (SCuM) is​​ a 2 by 3​​​‌ mm${}^2$ single die,​ featuring a 2.4 GHz​‌ radio supporting IEEE 802.15.4​​ and Bluetooth Low Energy​​​‌ (BLE). Because of its​ small size, it is​‌ an ideal platform for​​ tracking and localizing wild​​​‌ animal or insect. The​ existing localization method for​‌ SCuM is based on​​ lighthouse laser, and uses​​​‌ the chip's built in​ light sensor. While this​‌ technique works, the chip​​ needs to be less​​​‌ than 5 m from​ a lighthouse base station,​‌ and this type of​​ localization can only be​​​‌ used indoors. We develop​ a localization method for​‌ SCuM based on the​​ direction-finding feature of BLE​​​‌ 5.1. SCuM is flexible​ enough that we can,​‌ in software, add a​​ Constant Tone Extension (CTE)​​​‌ portion to a BLE​ frame, which is key​‌ to use BLE's direction​​ finding feature. With an​​​‌ antenna array and CTE,​ frequency offset can be​‌ estimated through the phase​​ samples, assisting to improve​​​‌ the direction finding performance​ over SCuM. We built​‌ a prototype Angle of​​ Arrival (AoA) localization system​​ and achieves a median​​​‌ angle estimation error of‌ $\pm 5^o$ inside‌​‌ of an anechoic chamber.​​

These results are published​​​‌ in: 22.

8.2.1 A​​ Comprehensive Survey on Channel​​​‌ Hopping and Scheduling Enhancements‌ for TSCH Networks

Participants:‌​‌ Martina Balbi, Thomas​​ Watteyne.

Time-Synchronized Channel​​​‌ Hopping (TSCH) is playing‌ an essential role in‌​‌ enabling reliable and energy-efficient​​ communication in low-power wireless​​​‌ applications, thanks to its‌ scheduling and adaptive channel‌​‌ hopping capabilities. Advancements in​​ these areas are vital​​​‌ for further improving TSCH‌ networks performance. Enhanced scheduling‌​‌ algorithms can reduce energy​​ consumption and increase network​​​‌ capacity, while adaptive channel‌ hopping strategies dynamically respond‌​‌ to changing network conditions​​ and interference patterns, ensuring​​​‌ robust communication in complex‌ environments. This survey provides‌​‌ a comprehensive review of​​ existing research on scheduling​​​‌ and adaptive channel hopping‌ enhancements for TSCH networks,‌​‌ categorizing, analyzing, and classifying​​ them to reveal current​​​‌ trends. Furthermore, it highlights‌ open challenges that have‌​‌ the potential to shape​​ the future of TSCH​​​‌ networks.

These results are‌ published in 8.‌​‌

8.2.2 Embedded Artificial Intelligence​​ for IoT Applications Using​​​‌ the MAX78000

Participants: Martina‌ Balbi, Thomas Watteyne‌​‌.

Recent advances in​​ embedded AI and IoT​​​‌ have revolutionized the development‌ of intelligent edge devices.‌​‌ This work provides a​​ tutorial on developing and​​​‌ deploying AI models on‌ the MAX78000, a low-power‌​‌ microcontroller designed specifically for​​ AI applications. Starting with​​​‌ the foundational understanding of‌ neural networks and machine‌​‌ learning, this tutorial explores​​ the architecture and capabilities​​​‌ of the MAX78000, which‌ integrates a CNN accelerator‌​‌ with an ARM Cortex-M4​​ core. We give practical​​​‌ guidance on creating, training,‌ and quantizing AI models,‌​‌ detailing essential tools, frameworks,​​ and the deployment process.​​​‌ Real-world examples illustrate the‌ versatility of AI microcontrollers‌​‌ and their performance in​​ various IoT applications. Emphasizing​​​‌ the importance of accessible‌ development tools, this tutorial‌​‌ aims to increase awareness​​ within the IoT community​​​‌ about low-power accelerators. This‌ enables developers to create‌​‌ efficient, real-time AI solutions,​​ highlighting the transformative potential​​​‌ of embedded AI in‌ IoT.

These results are‌​‌ published in 9.​​

8.2.3 Management of 6TiSCH​​​‌ Networks Using CORECONF: A‌ Clustering Use Case

Participants:‌​‌ Fabian Graf, Thomas​​ Watteyne.

Industrial low-power​​​‌ wireless sensor networks demand‌ high reliability and adaptability‌​‌ to cope with dynamic​​ environments and evolving network​​​‌ requirements. While the 6TiSCH‌ protocol stack provides reliable‌​‌ low-power communication, the CoAP​​ Management Interface (CORECONF) for​​​‌ runtime management remains underutilized.‌ In this work, we‌​‌ implement CORECONF and introduce​​ clustering as a practical​​​‌ use case. We implement‌ a cluster formation mechanism‌​‌ aligned with the Routing​​ Protocol for Low-Power and​​​‌ Lossy Networks (RPL) and‌ adjust the TSCH channel-hopping‌​‌ sequence within the established​​ clusters. Two use cases​​​‌ are presented. First, CORECONF‌ is used to mitigate‌​‌ external Wi-Fi interference by​​ forming a cluster with​​​‌ a modified channel set‌ that excludes the affected‌​‌ frequencies. Second, CORECONF is​​ employed to create a​​​‌ priority cluster of sensor‌ nodes that require higher‌​‌ reliability and reduced latency,​​​‌ such as those monitoring​ critical infrastructure in industrial​‌ settings. Simulation results show​​ significant improvements in latency,​​​‌ while practical experiments demonstrate​ a reduction in overall​‌ network charge consumption from​​ approximately 50 mC per​​​‌ hour to 23 mC​ per hour, by adapting​‌ the channel set within​​ the interference-affected cluster.

These​​​‌ results are published in​ 11.

8.2.4 What​‌ to Expect when Using​​ DECT NR+

Participants: Fabian​​​‌ Graf, Thomas Watteyne​.

DECT NR+ is​‌ the new kid on​​ the block in wireless​​​‌ technologies: it re-purposes the​ 1.9 GHz Digital Enhanced​‌ Cordless Telecommunications (DECT) standard​​ for IoT-type applications. How​​​‌ does DECT NR+ perform​ in practice and where​‌ should we be using​​ it? The ambition of​​​‌ this article is to​ provide answers to those​‌ questions. We start by​​ an overview of the​​​‌ fundamental principles of the​ physical layer. We then​‌ survey the DECT NR+​​ products on the market​​​‌ today. Using the nRF91​ series from Nordic Semiconductor,​‌ we conduct a comprehensive​​ set of hands-on power​​​‌ consumption and communication range​ measurements. Our results place​‌ DECT NR+ in a​​ gap in-between existing technologies.​​​‌ Its range is comparable​ to long range standards​‌ such as IEEE 802.15.4g​​ under certain parameter choices:​​​‌ 200 m in an​ urban setting, 6 km​‌ in the most favorable​​ line-of-sight conditions. For higher​​​‌ order Modulation and Coding​ Schemes, range drops and​‌ is rather comparable to​​ Wi-Fi. The nRF9161 draws​​​‌ significant power (at 3.7​ V, 220 mA transmitting​‌ at +19 dBm, 45​​ mA receiving), approximately 10×​​​‌ higher than BLE radios.​ This is likely primarily​‌ due to earlystage design​​ inefficiencies and the inherent​​​‌ complexity of the DECT​ NR+ physical layer, limiting​‌ its adoption in batterypowered​​ applications. We conclude that​​​‌ DECT NR+ is particularly​ appropriate for applications that​‌ require a dynamic tradeoff​​ between communication range and​​​‌ data rate, but are​ not cost-sensitive.

These results​‌ are published in 12​​.

8.2.5 HyPM: Hybrid​​​‌ Performance Metric Transmission in​ Low-Power Wireless Network

Participants:​‌ Fabian Graf, Thomas​​ Watteyne.

Low-power wireless​​​‌ networks are increasingly deployed​ in Industrial Internet of​‌ Things (IIoT) environments to​​ ensure reliable and secure​​​‌ operation of machinery. One​ of the key requirements​‌ for these networks is​​ reliability, which hinges on​​​‌ the ability to continuously​ assess the network's healtha​‌ challenge addressed by Application​​ Performance Monitoring (APM). Within​​​‌ APM, transmitting performance metrics,​ particularly in low-power networks,​‌ must be efficient to​​ preserve battery life. This​​​‌ paper addresses various APM​ approaches, including In-band Network​‌ Telemetry (INT), which leverages​​ multi-hop topologies to append​​​‌ telemetry data to existing​ packets. We propose HyPM,​‌ a novel hybrid method​​ that combines active monitoring​​​‌ with INT to meet​ user requirements for telemetry​‌ scope and frequency while​​ improving energy efficiency by​​​‌ 31%. Simulations in the​ Contiki-NG network simulator demonstrate​‌ the superiority of HyPM​​ in balancing energy consumption​​​‌ and data reporting accuracy.​

These results are published​‌ in 29.

8.2.6​​ Zephyr and SmartMesh IP​​​‌ - Happy Together

Participants:​ Fabian Graf, Thomas​‌ Watteyne.

The Zephyr​​ RTOS has seen significant​​ traction in recent years,​​​‌ particularly in embedded and‌ IoT applications. This paper‌​‌ demonstrates the use of​​ the SmartMesh IP networking​​​‌ functionality in Zephyr through‌ a dedicated networking chip‌​‌ connected via UART. Our​​ analysis across different platforms​​​‌ shows minimal resource requirements,‌ with the complete solution‌​‌ requiring about 30 kB​​ of ROM and 8​​​‌ kB of RAM.

These‌ results are published in‌​‌ 30.

8.2.7 Application​​ Performance Monitoring and Management​​​‌ of Low-power Wireless Networks‌

Participants: Fabian Graf,‌​‌ Thomas Watteyne.

In​​ this thesis, we investigate​​​‌ concepts for Application Performance‌ Monitoring and Management (APM)‌​‌ of low-power wireless IoT​​ networks. The preconditions of​​​‌ our work follow typical‌ embedded system constraints. The‌​‌ hardware architecture comprises microcontroller​​ units with low-memory footprint​​​‌ and the network architecture‌ relies on low-power radios‌​‌ and protocols like 6TiSCH​​ built upon IEEE 802.15.4.​​​‌ APM can be modeled‌ as a feedback-driven control‌​‌ loop, which continuously aims​​ to optimize reliability and​​​‌ power efficiency. First, we‌ define a broad set‌​‌ of performance metrics that​​ indicate system health and​​​‌ explore efficient methods for‌ metric collection on resource-constrained‌​‌ nodes. For metric transmission,​​ we demonstrate the advantages​​​‌ of lightweight approaches such‌ as In-Band Network Telemetry.‌​‌ Metric analysis focuses particularly​​ on network-related KPIs, emphasizing​​​‌ wireless link quality metrics.‌ The management component implements‌​‌ countermeasures based on critical​​ trends and bottlenecks identified​​​‌ during metric analysis. We‌ investigate lightweight deployment protocols‌​‌ including CORECONF, which integrates​​ seamlessly with the 6TiSCH​​​‌ standard. This dissertation presents‌ three novel applications demonstrating‌​‌ the practical use of​​ the control loop. First,​​​‌ an adaptive Forward Error‌ Correction (FEC) scheme is‌​‌ presented for IEEE 802.15.4​​ O-QPSK that responds to​​​‌ monitored error patterns in‌ packets transmitted over the‌​‌ wireless link on different​​ channels used with in​​​‌ time slotted channel hopping‌ (TSCH). Second, a dynamic‌​‌ clustering algorithm is proposed,​​ which forms priority clusters​​​‌ and mitigates external interference‌ based on environmental changes‌​‌ and user requirements. Third,​​ the APM concepts are​​​‌ applied to the modern‌ DECT NR+ standard by‌​‌ deriving a link adaptation​​ mechanism for that optimized​​​‌ transmit parameter selection.

These‌ results are published in‌​‌ the PhD thesis of​​ Fabian Graf  40.​​​‌

8.3.1 IETF LAKE Standardization​​

Participants: Mališa Vučinić,​​​‌ Geovane Fedrecheski, Yuxuan‌ Song, Elsa Lopez-Perez‌​‌.

2025 Standardization Highlights​​

• Adoption and formal analysis​​​‌ call of the draft‌ on PSK-based authentication with‌​‌ EDHOC led by Elsa​​ Lopez-Perez . 7 versions​​​‌ of the draft published‌ and presented in 2025‌​‌ in IETF LAKE.
• Adoption​​ of the draft on​​​‌ Remote attestation over EDHOC‌, led by Yuxuan‌​‌ Song , getting it​​ in a state ready​​​‌ to be called for‌ formal analysis. 4 versions‌​‌ of the draft published​​ and presented in 2025​​​‌ in IETF LAKE.
• Working‌ Group Last Call in‌​‌ IETF ACE issued on​​ the draft on lightweight​​​‌ certificate enrollment, with‌ Mališa Vučinić as the‌​‌ Editor. 3 versions of​​ the draft published and​​​‌ presented in 2025.
• 3‌ versions of the draft‌​‌ on lightweight authorization over​​​‌ EDHOC co-authored by Geovane​ Fedrecheski and Mališa Vučinić​‌ published and presented in​​ 2025 in IETF LAKE.​​​‌

Context and goal. The​ Inria AIO team plays​‌ a crucial role in​​ the standardization activity on​​​‌ lightweight security for the​ Internet-of-Things within the Internet​‌ Engineering Task Force (IETF).​​ This is a cross-working-group​​​‌ effort, spanning different working​ groups of the IETF.​‌ In that context, we​​ co-chair the IETF LAKE​​​‌ working group (Mališa​ Vučinić ), and also​‌ participate in the standardization​​ activities in the IETF​​​‌ ACE and IETF IOTOPS​ working groups.

Summary of​‌ LAKE Activities in 2025​​

The current IETF LAKE​​​‌ charter scopes the working​ group to work on​‌ the extensions to the​​ base EDHOC protocol standardized​​​‌ in RFC 9528 and​ RFC 9529. These​‌ include lightweight authorization, remote​​ attestation, pre-shared key-based authentication,​​​‌ and other extensions to​ the EDHOC protocol.

These​‌ extensions to the EDHOC​​ protocol are particularly important​​​‌ for the work of​ the Inria AIO team,​‌ as they form a​​ backbone of several important​​​‌ contributions in the scope​ of Horizon Europe OpenSwarm​‌ project and the French​​ national PEPR Networks of​​​‌ the Future HiSec project.​

• "Lightweight Authorization using EDHOC"​‌, co-authored by Geovane​​ Fedrecheski and Mališa Vučinić​​​‌ , is the main​ component of the zero-touch​‌ joining scheme, which the​​ team has been promoting​​​‌ in the scope of​ the Horizon Europe OpenSwarm​‌ project. The solution​​ specified in the draft​​​‌ enables an IoT product​ to be deployed in​‌ its final site without​​ requiring any provisioning of​​​‌ deployment-specific configuration. Our main​ use case is the​‌ DotBot platform. In​​ 2025, we published a​​​‌ paper in IEEE DCOSS-IoT​ conference on this work​‌27 and the extended​​ version of the paper​​​‌ is under a revision​ for Elsevier Computer Networks​‌ journal. In 2025, we​​ have submitted 3 versions​​​‌ of this IETF-adopted draft,​ presenting each at IETF​‌ meetings and getting it​​ closer to publication as​​​‌ an RFC.
• The PhD​ thesis of Yuxuan Song​‌ , is on remote​​ attestation for Internet-of-Things swarms.​​​‌ Yuxuan Song leads a​ draft in the LAKE​‌ working group which describes​​ how to perform remote​​​‌ attestation over EDHOC.​ In 2025, the draft​‌ has been formally adopted​​ by the LAKE working​​​‌ group and Yuxuan Song​ published a total of​‌ 4 versions, presenting each​​ at IETF meetings, and​​​‌ adopting the reviews by​ different actors in the​‌ group. In 2025, we​​ published a paper at​​​‌ IEEE ISCC conference35​ and the extended version​‌ of the paper is​​ under a revision for​​​‌ IEEE Transactions on Computers.​
• The PhD thesis of​‌ Elsa Lopez-Perez , funded​​ by the PEPR 5G​​​‌ NF-HiSec project, is on​ “Backwards-compatible Next-Generation Security for​‌ the Internet-of-Things Infrastructure”. The​​ PhD thesis explores the​​​‌ integration of the EDHOC​ protocol in Next-G networks.​‌ EDHOC, as standardized in​​ RFC 9528 and RFC​​​‌ 9529, gave priority to​ asymmetric authentication credentials, leaving​‌ authentication based on symmetric​​ credentials (e.g. pre-shared keys)​​​‌ as future work. Elsa​ Lopez-Perez leads a draft​‌ in the LAKE working​​ group which fills this​​ gap and specifies an​​​‌ extension to EDHOC that‌ supports authentication based on‌​‌ pre-shared keys. This​​ draft has seen a​​​‌ total of 7 versions‌ published in 2025. It‌​‌ was called for formal​​ analysis in November 2025,​​​‌ meaning that it is‌ currently frozen from changes‌​‌ until February 2026 1​​. In parallel with​​​‌ the IETF call for‌ analysis, our team is‌​‌ working in collaboration with​​ cryptographers from University of​​​‌ Limoges and University of‌ Clermont Auvergne on a‌​‌ symbolic analysis of the​​ protocol.

Implementation Efforts

We​​​‌ implement the protocols that‌ we standardize in IETF‌​‌ LAKE within the lakers​​ repository in Rust programming​​​‌ language (See 7.1.2).‌ For the moment, the‌​‌ official repository gathers implementations​​ of the EDHOC and​​​‌ its lightweight authorization extension.‌ An implementation of PSK-based‌​‌ authentication method with EDHOC​​ is available as a​​​‌ pull request and is‌ currently undergoing reviews before‌​‌ being merged with the​​ official lakers repository.

At​​​‌ the time of the‌ writing, the lakers library‌​‌ has been downloaded more​​ than 60,000 times on​​​‌ crates.io!

The work‌ on remote attestation is‌​‌ implemented within a fork​​ of the DotBot-firmware repository​​​‌, and uses the‌ lakers library as a‌​‌ transport mechanism (See 7.1.4​​).

Summary of Activities​​​‌ in IETF ACE and‌ IOTOPS

Our work on‌​‌ lightweight certificate enrollment has​​ been called for Working​​​‌ Group Last call in‌ IETF ACE working group.‌​‌ We published a total​​ of three revisions of​​​‌ this document in 2025,‌ progressing it, adopting the‌​‌ input of the working​​ group, and each time​​​‌ presenting the improvements during‌ the IETF meeting of‌​‌ the working group. We​​ are currently incorporating reviews​​​‌ from the last call‌ before shipping the document‌​‌ for publication as an​​ RFC.

Another informational document​​​‌ that compares the overhead‌ of different security protocols‌​‌ standardized in the IETF​​ is an official IETF​​​‌ document in the IETF‌ IOTOPS working group. The‌​‌ document underwent several reviews​​ in the group and​​​‌ is now considered ready,‌ but is parked because‌​‌ it waits for the​​ dependencies to be published.​​​‌

Conclusion The standardization activity‌ of the AIO team‌​‌ remains a high priority​​ track in the team.​​​‌ The AIO team currently‌ participates in the standardization‌​‌ of 5 work-in-progress documents,​​ each on a well-defined​​​‌ path to becoming an‌ RFC. The proposed standards‌​‌ documents are implemented within​​ the lakers repository.

8.3.2​​​‌ Pre-Shared Key Authentication with‌ EDHOC: the Security-Performance Tradeoff‌​‌

Participants: Elsa Lopez-Perez,​​ Thomas Watteyne, Mališa​​​‌ Vučinić.

The rapid‌ growth of the Internet‌​‌ of Things ecosystem has​​ intensified the need for​​​‌ secure, resource-efficient communication protocols.‌ The EDHOC protocol is‌​‌ a lightweight authenticated keyexchange​​ protocol, recently developed by​​​‌ the Internet Engineering Task‌ Force. EDHOC addresses the‌​‌ challenges of transport over​​ constrained radio technologies and​​​‌ execution on constrained microcontroller‌ units. In its standardized‌​‌ version, the key-exchange can​​ be authenticated using signatures​​​‌ or static Diffie-Hellman keys.‌ However, many Internet of‌​‌ Things deployments in the​​ wild rely on Pre-Shared​​​‌ Keys. As such, the‌ potential use of EDHOC‌​‌ in those deployments requires​​​‌ a new authentication method​ for this protocol, based​‌ on Pre-Shared Keys. Two​​ variants of Pre-Shared Keys​​​‌ authentication in EDHOC are​ currently under consideration in​‌ the Internet Engineering Task​​ Force LAKE working group.​​​‌ This paper presents a​ comprehensive analysis of these​‌ variants, examining their performance​​ metrics, implementation complexity, and​​​‌ security and privacy considerations.​ Our evaluation focuses on​‌ computational time, memory usage,​​ and deployment challenges in​​​‌ diverse Internet of Things​ ecosystems. Based on our​‌ analysis, we have formulated​​ a recommendation to the​​​‌ Working Group, which has​ opted to adopt and​‌ standardize PSK2.

These results​​ are published in 14​​​‌.

8.3.3 ELA: Secure,​ Lightweight, and Zero-Touch Enrollment​‌ for IoT Devices

Participants:​​ Geovane Fedrecheski, Thomas​​​‌ Watteyne, Mališa Vučinić​.

When deploying large​‌ numbers of IoT devices,​​ an enrollment protocol takes​​​‌ care of admitting each​ device into their target​‌ network for the first​​ time. The protocol must​​​‌ be secure to block​ malicious actors, easy to​‌ operate to reduce cost,​​ and lightweight due to​​​‌ bandwidth constraints. Solutions in​ literature either involve use​‌ of pre-shared keys, require​​ perdevice user input, or​​​‌ have been designed for​ non-constrained environments. This paper​‌ introduces EDHOC with Lightweight​​ Authorization (ELA), a protocol​​​‌ for securely authorizing enrollment​ of devices in constrained​‌ networks with support for​​ zerotouch deployments. We define​​​‌ ELA as an extension​ to Ephemeral Diffie-Hellman Over​‌ COSE (EDHOC), a key​​ exchange protocol with extremely​​​‌ low message footprint. We​ evaluate ELA on DotBot,​‌ a platform for research​​ in swarm micro-robotics. We​​​‌ find that enrolling a​ DotBot with ELA takes​‌ 2.52 s and consumes​​ 39.31 mC. When compared​​​‌ to a baseline EDHOC​ version, flash and RAM​‌ have an overhead of​​ 10.67% and 22.63%, respectively,​​​‌ and message footprint increases​ by only 49 B.​‌ ELA is being standardized​​ in the Internet Engineering​​​‌ Task Force (IETF).

These​ results are published in​‌ 27.

8.3.4 Fine-grained,​​ privacy-augmenting LI-compliance in the​​​‌ LAKE standard

Participants: Elsa​ Lopez-Perez, Mališa Vučinić​‌.

The Internet Engineering​​ Task Force and its​​​‌ LAKE working group standardized​ the Ephemeral Diffie-Hellman over​‌ COSE (EDHOC) authenticated key-exchange​​ protocol for use in​​​‌ constrained Internet of Things​ deployments. The use cases​‌ include cellular networks, such​​ as NB-IoT, but also​​​‌ non-cellular networks such as​ 6TiSCH, and LoRaWAN. As​‌ a result of its​​ use in cellular networks,​​​‌ EDHOC will be subject​ to Lawful Interception (LI),​‌ which allows a group​​ of authorities to break,​​​‌ if equipped with a​ warrant, the end-to-end (E2E)​‌ security of the channel​​ established through EDHOC. Current​​​‌ implementations of EDHOC would​ only allow lawful interception​‌ by using the cellular​​ network operator as a​​​‌ legitimate endpoint, essentially running​ a Person-in-the-Middle attack against​‌ the protocol. In this​​ work, we focus on​​​‌ a privacy-preserving, finegrained LI-compliant​ modification of EDHOC for​‌ all four authentication methods​​ that this protocol currently​​​‌ supports. We achieve this​ via a careful white-box​‌ composition of EDHOC with​​ the Lawful Interception Key-Exchange​​​‌ approach of Arfaoui et​ al. (ESORICS 2021) and​‌ Bultel and Onete (SAC​​ 2022). Our resulting construction​​ not only achieves strong​​​‌ key-security, but also non-frameability,‌ and LI-compliance, without breaking‌​‌ the identity-protection property of​​ EDHOC. Our implementation results​​​‌ show that, while LIKE‌ adds an overhead to‌​‌ a standard EDHOC implementation​​ in Rust, the resulting​​​‌ protocol remains practical while‌ achieving much better privacy‌​‌ and LI-compliance.

These results​​ are published in 31​​​‌.

8.3.5 AuRA: Remote‌ Attestation over EDHOC for‌​‌ Constrained Internet-of-Things Use Cases​​

Participants: Yuxuan Song,​​​‌ Geovane Fedrecheski, Thomas‌ Watteyne, Mališa Vučinić‌​‌.

Remote Attestation (RA)​​ is a security process​​​‌ that verifies the integrity‌ and trustworthiness of a‌​‌ remote device's software and​​ hardware. While RA for​​​‌ high-end devices is well-developed,‌ RA in constrained IoT‌​‌ environments remains incomplete. Existing​​ embedded RA mechanisms focus​​​‌ on local evidence generation‌ and verification, but lack‌​‌ a complete process that​​ includes a secure attestation​​​‌ channel. This paper introduces‌ AuRA, a lightweight RA‌​‌ solution that builds upon​​ the newly standardized Ephemeral​​​‌ Diffie-Hellman over COSE (EDHOC)‌ protocol. AuRA specifies how‌​‌ to transport existing attestation​​ mechanisms in parallel with​​​‌ network authentication. We evaluate‌ AuRA on the nRF5340‌​‌ microcontroller running at 64​​ MHz. This implementation has​​​‌ a memory footprint of‌ 6,665 B of RAM‌​‌ and 17,163 B of​​ flash. The device completes​​​‌ Remote Attestation by exchanging‌ three EDHOC messages with‌​‌ a verifier entity, of​​ sizes 42 B, 59​​​‌ B and 223 B.‌ This allows the device‌​‌ to prove that it​​ is running the right​​​‌ hardware and software in‌ only 5.51 s, consuming‌​‌ as little as 88​​ mC of charge.

These​​​‌ results are published in‌ 35.

8.3.6 When‌​‌ to Attest? Intra-and Post-Handshake​​ Attestation for IoT Swarms​​​‌

Participants: Yuxuan Song,‌ Geovane Fedrecheski, Thomas‌​‌ Watteyne, Mališa Vučinić​​.

Remote attestation is​​​‌ a security mechanism that‌ allows a device to‌​‌ prove its integrity and​​ trustworthiness by generating fresh​​​‌ verifiable evidence to be‌ assessed by a verifier.‌​‌ It is gaining increasing​​ attention in the context​​​‌ of IoT security for‌ both IoT devices and‌​‌ services. Within the ongoing​​ standardization efforts at the​​​‌ IETF, two distinct approaches‌ have emerged and are‌​‌ actively discussed by different​​ working groups and protocol​​​‌ designers: (1) intra-handshake attestation,‌ where attestation is performed‌​‌ during the handshake process;​​ (2) posthandshake attestation, where​​​‌ it occurs after the‌ handshake is complete. This‌​‌ position paper analyzes the​​ respective security properties and​​​‌ discusses their applicability across‌ different IoT deployment scenarios.‌​‌ We highlight the key​​ trade-off: intrahandshake attestation enables​​​‌ early trust establishment prior‌ to session setup, making‌​‌ it suitable for onboarding​​ scenarios, while post-handshake attestation​​​‌ provides continuous assurance and‌ supports runtime integrity validation.‌​‌

These results are published​​ in 36.

8.3.7​​​‌ Lightweight Solutions for a‌ Secure Internet of Things‌​‌

Participants: Mališa Vučinić.​​

The Internet of Things​​​‌ enables the interconnection of‌ everyday objects with the‌​‌ global communication network, the​​ Internet. The research field​​​‌ brings together the notions‌ of computer science, telecommunications‌​‌ and electronics. This manuscript​​ summarizes the main focus​​​‌ of my research on‌ enabling a secure and‌​‌ efficient Internet of Things.​​​‌ My research is positioned​ at the intersection of​‌ low-power wireless networking and​​ communication security. The manuscript​​​‌ describes my main contributions​ to the field and​‌ is organized around four​​ technical chapters. The first​​​‌ chapter describes my contributions​ to the area of​‌ time-slotted channel hopping networks​​ and their secure integration​​​‌ with the Internet. The​ second chapter describes the​‌ process of designing and​​ standardizing a lightweight authenticated​​​‌ key exchange protocol for​ Internet-of-Things use cases.The third​‌ chapter discusses my contributions​​ to the field of​​​‌ constrained swarm robotics, where​ each robot is a​‌ node in a low-power​​ and constrained network. The​​​‌ fourth chapter discusses the​ Smart Dust vision -​‌ computation, communication and sensing​​ in a cubic millimeter​​​‌ - and my contributions​ to the area.

These​‌ results are published as​​ part of the HDR​​​‌ thesis of Mališa Vučinić​  42.

8.4.1 Mari: Connecting Large​​​‌ Scale Robot Swarms using​ TSCH over BLE

Participants:​‌ Geovane Fedrecheski, Thomas​​ Watteyne, Yinghao Gao​​​‌, Filip Maksimovic,​ Mališa Vučinić, Thomas​‌ Watteyne, Alexandre Abadie​​.

Instrumenting a large-scale​​​‌ micro-robot swarm is challenging​ due to limitations in​‌ existing communication systems. We​​ propose Mari, a TSCH-based​​​‌ link layer over BLE​ that supports multiple non-coordinated​‌ gateways with fast node-controlled​​ handovers. Simulations show a​​​‌ single Mari gateway can​ handle 100 nodes with​‌ latency below 100 ms​​ and best-case node join​​​‌ time under 40 ms.​ We implement Mari on​‌ constrained hardware and find​​ that it uses 33.9​​​‌ kB of Flash memory​ and 24.1 kB of​‌ RAM. We validate it​​ on a 57-node setup,​​​‌ where full network formation​ takes 5.8 s. The​‌ demo will exhibit a​​ portable network-focused testbed with​​​‌ at least 40 nodes​ and an user interface​‌ showing live network metrics.​​

This work has been​​​‌ demonstrated at 26,​ and is currently in-submission​‌ for publication as a​​ full paper.

8.4.2 Supercapacitor-Powered​​​‌ Robotic Platforms for Research​ and Experimentation

Participants: Said​‌ Alvarado-Marin, Geovane Fedrecheski​​, Filip Maksimovic,​​​‌ Thomas Watteyne.

Swarm​ robotics focuses on designing​‌ and coordinating large groups​​ of relatively simple robots​​​‌ to perform tasks in​ a decentralised and collective​‌ manner. The swarm provides​​ a resilient and flexible​​​‌ solution for many applications.​ However, contemporary swarm robots​‌ have a significant power​​ problem in that secondary​​​‌ (i.e. rechargeable) batteries are​ slow to charge and​‌ offer lifetimes of only​​ a few years, increasing​​​‌ maintenance costs and pollution​ due to battery replacement.​‌ We imagine a different​​ future, wherein battery-free robots​​​‌ powered by supercapacitors can​ be recharged in seconds,​‌ offer long-life autonomous operation​​ and can rapidly pass​​​‌ charge between one another​ using trophallaxis. In pursuit​‌ of this vision, we​​ contribute the CapBot, a​​​‌ battery-free swarm robot equipped​ with Mecanum wheels, a​‌ Cortex M4F application processor​​ and Bluetooth Low Energy​​​‌ networking. The CapBot fully​ recharges in 16 s,​‌ offers 51 min of​​ autonomous operation at top​​​‌ speed, and can transfer​ up to 50% of​‌ its available charge to​​ a peer via trophallaxis​​ in under 20 s.​​​‌ The CapBot is fully‌ open source and all‌​‌ software and hardware source​​ is available online.

A​​​‌ major segment of the‌ Swarm Robotics market will‌​‌ operate indoors; patrolling secure​​ facilities, sorting packages in​​​‌ a warehouse or cleaning‌ commercial buildings. As the‌​‌ swarm focuses on collaboration,​​ effective communication is essential.​​​‌ The cost of installing‌ power and network cables‌​‌ thus poses a significant​​ barrier to entry for​​​‌ facilities lacking wireless networking.‌ We tackle this problem‌​‌ by introducing a “peel​​ and stick”' backhaul for​​​‌ swarm robot telemetry that‌ operates for several years‌​‌ on a D-cell battery.​​ Our solution is tiered,​​​‌ using a time-synchronized mesh‌ network as its backhaul‌​‌ and near-field communication between​​ the robots and the​​​‌ mesh, with uW-scale listening‌ power. Our evaluation demonstrates‌​‌ that this architecture achieves​​ over 90% reliability, low​​​‌ power consumption and long‌ battery life.

These results‌​‌ are published in 32​​ and in 33.​​​‌

8.4.3 Automatic Indoor Localization‌ and Mapping of Robotic‌​‌ Swarms

Participants: Filip Maksimovic​​, Said Alvarado-Marin.​​​‌

In this work, we‌ apply lighthouse localization, originally‌​‌ designed for virtual reality​​ motion tracking, to positioning​​​‌ and localization of indoor‌ robots. We first present‌​‌ a lighthouse decoding and​​ tracking algorithm on a​​​‌ low-power wireless microcontroller with‌ hardware implemented in a‌​‌ cmscale form factor. One-time​​ scene solving is performed​​​‌ on a computer using‌ a variety of standard‌​‌ computer vision techniques. Three​​ different robotic localization scenarios​​​‌ are analyzed in this‌ work. The first is‌​‌ a planar scene with​​ a single lighthouse with​​​‌ a four-point pre-calibration. The‌ second is a planar‌​‌ scene with two lighthouses​​ that self calibrates with​​​‌ either multiple robots in‌ the experiment or a‌​‌ single robot in motion.​​ The third extends to​​​‌ a 3D scene with‌ two lighthouses and a‌​‌ self-calibration algorithm. The absolute​​ accuracy, measured against a​​​‌ camerabased tracking system, was‌ found to be 7.25‌​‌ mm RMS for the​​ 2D case and 11.2​​​‌ mm RMS for the‌ 3D case, respectively. This‌​‌ demonstrates the viability of​​ lighthouse tracking both for​​​‌ smallscale robotics and as‌ an inexpensive and compact‌​‌ alternative to camera-based setups.​​

These results are published​​​‌ in 18 and in‌ 19.

8.4.4 The‌​‌ Drawing and Use of​​ Conics for Automatic Indoor​​​‌ Swarm Localization

Participants: Filip‌ Maksimovic, Said Alvarado-Marin‌​‌.

We extend the​​ prior work for calibrating​​​‌ Lighthouse localization systems using‌ a single view of‌​‌ two or more coplanar​​ circles traced by a​​​‌ moving robot. The calibration‌ method leverages conic algebra‌​‌ to compute the homography​​ between the Lighthouse view​​​‌ and the world plane,‌ up to similarity. This‌​‌ approach requires minimal user​​ intervention and is particularly​​​‌ suited for automatically calibrating‌ large-scale deployments involving hundreds‌​‌ of mobile robots. We​​ validate our method using​​​‌ a centimeter-scale differential drive‌ robot, utilizing 5 cm‌​‌ circles to calibrate a​​ 2×2m 2 area. The​​​‌ proposed technique achieved a‌ mean positional accuracy of‌​‌ 7.77 mm, compared to​​ the 5.37 mm accuracy​​​‌ of a previous calibration‌ method based on manual‌​‌ measurements and known correspondences.​​​‌ We demonstrate that the​ conics traced by the​‌ robot are accurate enough​​ for reliable homography estimation,​​​‌ even under varying conditions​ of tire material and​‌ surface type. A camera-based​​ motion capture system served​​​‌ as the ground truth​ for all experiments. This​‌ work represents a step​​ toward scalable and decentralized​​​‌ lighthouse calibration, enabling efficient​ 2D localization in large-scale​‌ robotic systems.

These results​​ are published in 7​​​‌.

8.4.5 Energy-Aware Swarm​ Algorithms for Recharging and​‌ Charge Sharing

Participants: Thomas​​ Watteyne.

Unmanned Aerial​​​‌ Vehicles (UAVs) are expected​ to transform logistics, reducing​‌ delivery time, costs, and​​ emissions. This study addresses​​​‌ an on-demand delivery ,​ in which fleets o​‌ f UAVs are deployed​​ to fulfil orders that​​​‌ arrive stochastically. Unlike previous​ work, it considers UAVs​‌ with heterogeneous, unknown energy​​ storage capacities and assumes​​​‌ no knowledge of the​ energy consumption models. We​‌ propose a decentralised deployment​​ strategy that combines auction-based​​​‌ task allocation with online​ learning. Each UAV independently​‌ decides whether to bid​​ for orders based on​​​‌ its energy storage charge​ level, the parcel mass,​‌ and delivery distance. Over​​ time, it refines its​​​‌ policy to bid only​ for orders within its​‌ capability. Simulations using realistic​​ UAV energy models reveal​​​‌ that, counter-intuitively, assigning orders​ to the least confident​‌ bidders reduces delivery times​​ and increases the number​​​‌ of successfully fulfilled o​ rders. T his s​‌ trategy i s s​​ hown t o outperform​​​‌ threshold-based methods which require​ UAVs to exceed specific​‌ charge levels at deployment.​​ We propose a variant​​​‌ of the strategy which​ uses learned policies for​‌ forecasting. This enables UAVs​​ with insufficient charge levels​​​‌ to commit to fulfilling​ orders at specific future​‌ times, helping to prioritise​​ early orders. Our work​​​‌ provides new insights into​ long-term deployment of UAV​‌ swarms, highlighting the advantages​​ of decentralised energy-aware decision-making​​​‌ coupled with online learning​ in real-world dynamic environments.​‌

The utility of swarms​​ of robots would greatly​​​‌ increase if they could​ operate over extended periods​‌ of time. Here, we​​ consider two strategies for​​​‌ swarms of robots to​ replenish their energy while​‌ performing work in a​​ remote location. In the​​​‌ first, each robot commutes​ to work and replenishes​‌ at its base. In​​ the second, some robots​​​‌ perform work, whereas others​ commute to provide them​‌ with energy. We present​​ results from extensive physics-based​​​‌ simulations. The first strategy​ performs 92.8% of the​‌ work at only 12.6%​​ lower energy efficiency than​​​‌ an optimal strategy. The​ second strategy is beneficial​‌ for low charging rates​​ or if the robots​​​‌ providing energy are permitted​ increased amounts of storage.​‌ We provide proof-of-concept validation​​ using the CapBot swarm​​​‌ robot platform.

These results​ are published in 37​‌ and in 38.​​

8.4.6 Communication-Aware Localization of​​​‌ Robotic Swarms

Participants: Said​ Alvarado-Marin, Filip Maksimovic​‌.

A significant challenge​​ in large-scale robotic swarms​​​‌ is distributed localization with​ minimal overhead. In this​‌ work, we aim to​​ quickly and reliably find​​​‌ the location of every​ member in a fleet​‌ of 1,000 miniature mobile​​ robots in an indoor​​ testbed environment. After a​​​‌ review of the state‌ of the art, we‌​‌ identified that most localization​​ technologies cannot achieve the​​​‌ cm-level accuracy and high‌ update-rate required for our‌​‌ work, and those who​​ can are prohibitively expensive.​​​‌ This work contributes to‌ addressing these drawbacks by‌​‌ proposing a high performance,​​ inexpensive localization system based​​​‌ on the Lighthouse v2‌ Localization system. We further‌​‌ propose Metronome, a dynamic,​​ low-latency TDMA wireless network​​​‌ for multi-robot systems. The‌ Lighthouse Localization system (LH2)‌​‌ is a infrared laser-based​​ motion capture system, originally​​​‌ developed by Valve and‌ HTC Vive for virtual‌​‌ reality motion tracking. This​​ system uses base stations​​​‌ that project laser sweeps‌ onto photodiodes mounted on‌​‌ robots, with a small​​ front-end chip decoding the​​​‌ signal. In this work‌ we propose a series‌​‌ of LH2 compatible algorithms​​ that achieve 7.25 mm​​​‌ RMS accuracy for 2D‌ tracking and 11.2 mm‌​‌ RMS for 3D tracking,​​ both at 50 Hz​​​‌ refresh rate. Metronome is‌ a dynamic Time Division‌​‌ Multiple Access (TDMA) network​​ protocol that adjusts its​​​‌ slotframe size in real-time,‌ responding to nodes joining‌​‌ or leaving the network.​​ Using Metronome, a central​​​‌ gateway can coordinate 100‌ robots on a single‌​‌ channel, achieving a downlink​​ latency of 20 ms​​​‌ and a maximum uplink‌ latency of 290 ms.‌​‌ This thesis further proposes​​ an automatic calibration technique​​​‌ for Lighthouse localization systems‌ in which a robot‌​‌ calibrates the system by​​ following a set of​​​‌ circular patterns. When tested‌ on a set of‌​‌ mobile robots, the algorithm​​ achieved an accuracy of​​​‌ 9.27 mm RMS. This‌ system minimizes the need‌​‌ for user intervention and​​ enables efficient localization in​​​‌ large-scale robotic systems. This‌ thesis contributes to the‌​‌ growing field of research​​ on multi-robot localization and​​​‌ communication. It demonstrates the‌ effectiveness of lighthouse based‌​‌ localization for multi-robot systems,​​ and highlights the potential​​​‌ of off-the-shelf hardware for‌ the development of high-accuracy‌​‌ robot motion-capture systems.

These​​ results are published in​​​‌ the thesis of Said‌ Alvarado-Marin  39.

8.4.7‌​‌ Porting A Real-Time Operating​​ System for Embedded IoT​​​‌ Devices to Linux

Participants:‌ Thomas Watteyne.

In‌​‌ the context of the​​ Horizons Europe OpenSwarm project,​​​‌ we aim to develop‌ the next generation swarm‌​‌ devices; those are deeply​​ embedded and use real-time​​​‌ operating systems. These real-time‌ operating systems fulfill strong‌​‌ real-time requirements needed by​​ protocol stacks. However, as​​​‌ the OpenSwarm use cases‌ are complex, development cycles‌​‌ become longer and more​​ costly, and application developers​​​‌ require better tools to‌ increase productivity. This paper‌​‌ introduces the port of​​ uC/OS-II to Linux, enabling​​​‌ quick prototyping of RTOS‌ applications on a controlled‌​‌ environment without the need​​ to continuously reflash microcontrollers​​​‌ to test and debug.‌ Additionally, it allows applications‌​‌ to leverage existing tools​​ in a Linux environment.​​​‌ We present an improved‌ interrupt context switch implementation‌​‌ than the default port,​​ ensuring that interrupts on​​​‌ Linux run to completion,‌ mimicking bare metal behavior.‌​‌ Moreover, we measure the​​ interrupt response, recovery and​​​‌ context switch latencies for‌ both bare metal and‌​‌ Linux version of uC/OS-II.​​​‌ The worst-case time for​ context switch on Linux​‌ is 396.5 us, the​​ mean value is 391.1​​​‌ us with a standard​ deviation of 4.68 us.​‌

The results are published​​ in 34.

8.4.8​​​‌ Vega – Turning a​ Toy into a Ready-to-Use​‌ Robotic Platform

Participants: Narmin​​ Elkilani, Baptiste Carbillet​​​‌, Thomas Watteyne.​

Robotic platforms that are​‌ customizable, mobile, and compact​​ are crucial for testing​​​‌ algorithms in dynamic settings​ and enabling small-scale swarm​‌ robotics. In this paper​​ we introduce Vega, a​​​‌ robotic platform that augments​ the DJI RoboMaster S1​‌ with a Raspberry Pi-based​​ computer, allowing full control​​​‌ of its hardware capabilities​ and immediate deployment for​‌ various swarm applications.

These​​ results were published at​​​‌ a workshop: 23.​

8.5.1 Topology optimization in​ mobile wireless networks using​‌ machine learning

Participants: Felix​​ Marcoccia, Paul Mühlethaler​​​‌.

Mobile aerial networks​ have emerged as compelling​‌ technologies due to their​​ capacity to deliver autonomous,​​​‌ infrastructure-free communication in dynamic​ environments. Their growing relevance​‌ is driven by a​​ wide range of practical​​​‌ applications, ranging from UAVs​ to planes and satellites.​‌ In order to overcome​​ the need for a​​​‌ centralized proxy, to achieve​ higher resilience and capacity,​‌ such networks can leverage​​ ad hoc, multi-hop communications​​​‌ between nodes. However, they​ generally suffer from theoretical​‌ limitations, particularly when using​​ omnidirectional antennas. To overcome​​​‌ these limitations and leverage​ directional antennas, it becomes​‌ necessary to orchestrate all​​ antenna steering decisions, transmissions​​​‌ and receptions in real​ time, ensuring a viable​‌ and efficient network topology.​​ Given the highly combinatorial​​​‌ nature of this problem,​ this thesis proposes to​‌ address it using artificial​​ intelligence techniques, including supervised​​​‌ learning and generative models.​ In the course of​‌ this thesis, we experiment​​ with various deep learning​​​‌ methods to solve our​ problem and develop several​‌ solution architectures. By adapting​​ and extending state-of-the-art deep​​​‌ learning methods, we propose​ a data-driven method which​‌ generates high-performance network configurations​​ in real time. Furthermore,​​​‌ leveraging advanced generative approaches,​ we propose a learning​‌ architecture capable of jointly​​ generating the network links​​​‌ and a compatible transmission​ schedule, while accounting for​‌ the network's dynamic behavior.​​ The resulting models yield​​​‌ a substantial theoretical throughput​ improvement over conventional omnidirectional​‌ protocols, with even better​​ scalability as the number​​​‌ of nodes increases.

These​ results are published in​‌ the PhD thesis of​​ Felix Marcoccia  41.​​​‌

8.5.2 Delay performance in​ ITS-G5-based Vehicular Ad Hoc​‌ Networks

Participants: Paul Mühlethaler​​.

This work provides​​​‌ a comprehensive survey of​ the delay performance in​‌ ITS-G5-based Vehicular Ad Hoc​​ Networks (VANETs), highlighting its​​​‌ critical role in the​ reliability of safety-critical applications.​‌

This study explores the​​ challenges of low-latency communication​​​‌ in decentralized environments where​ high mobility and frequent​‌ topology changes often compromise​​ performance. It details various​​​‌ methodologies for delay modeling,​ including stochastic geometry and​‌ MAC contention models, to​​ better understand how interference​​​‌ and node density impact​ packet delivery. The author​‌ examines the specific constraints​​ of safety-related messages, such​​ as collision warnings, which​​​‌ require end-to-end delays below‌ 100 ms to be‌​‌ effective. Key mitigation strategies​​ are discussed, such as​​​‌ the use of Enhanced‌ Distributed Channel Access (EDCA)‌​‌ for traffic prioritization, opportunistic​​ routing, and dynamic power​​​‌ adjustment to manage congestion.‌ The survey also highlights‌​‌ emerging techniques like network​​ coding and machine learning​​​‌ for predictive resource management‌ and traffic load balancing.‌​‌ Experimental results from field​​ trials are analyzed to​​​‌ show how environmental factors‌ like urban canyons and‌​‌ tunnels affect signal propagation​​ and latency. Finally, the​​​‌ paper concludes that while‌ ITS-G5 provides a solid‌​‌ foundation, further integration of​​ hybrid models and advanced​​​‌ AI is necessary to‌ meet the ultra-reliable low-latency‌​‌ requirements of future autonomous​​ driving systems.

This work​​​‌ is published in 47‌.

8.5.3 IEEE 802.11p‌​‌ and IEEE 802.11bd for​​ Vehicular Communication

Participants: Paul​​​‌ Mühlethaler.

This study‌ provides an in-depth comparative‌​‌ analysis between the legacy​​ IEEE 802.11p standard and​​​‌ its evolutionary successor, IEEE‌ 802.11bd, for vehicular communications.‌​‌ While 802.11p enabled the​​ first wave of safety​​​‌ services, it faces significant‌ limitations regarding data rate,‌​‌ latency, and scalability in​​ dense traffic scenarios. IEEE​​​‌ 802.11bd addresses these shortcomings‌ by integrating advanced features‌​‌ from Wi-Fi 6 (802.11ax),​​ such as OFDMA, MU-MIMO,​​​‌ and 256-QAM modulation. These‌ enhancements allow for data‌​‌ rates exceeding 100 Mbps​​ and support for wider​​​‌ 20/40 MHz channels while‌ ensuring backward compatibility with‌​‌ existing 802.11p hardware. The​​ new standard introduces Dual​​​‌ Carrier Modulation (DCM) and‌ increased subcarrier spacing to‌​‌ specifically combat the severe​​ Doppler effects associated with​​​‌ high-speed mobility. Beyond basic‌ safety messages, 802.11bd is‌​‌ designed for advanced use​​ cases like cooperative perception,​​​‌ platooning, and high-definition sensor‌ sharing. The implementation of‌​‌ OFDMA provides more deterministic​​ scheduling, reducing collision probabilities​​​‌ and access delays compared‌ to traditional CSMA/CA methods.‌​‌ Furthermore, the standard introduces​​ Hybrid Automatic Repeat Request​​​‌ (HARQ) mechanisms, which significantly‌ improve communication reliability in‌​‌ challenging propagation environments by​​ allowing for efficient retransmissions.​​​‌ This resilience is bolstered‌ by the addition of‌​‌ midambles, which enable more​​ frequent and accurate channel​​​‌ estimation during high-speed movement.‌ Ultimately, IEEE 802.11bd stands‌​‌ as a strategic evolution​​ essential for supporting the​​​‌ increasing data demands and‌ ultra-reliability required by future‌​‌ autonomous driving ecosystems.

This​​ work is published in​​​‌ 46.

8.5.4 IEEE‌ 802.11p and IEEE 802.11bd‌​‌ Based on an Extended​​ Bianchi Model

Participants: Paul​​​‌ Mühlethaler.

This study‌ provides an analytical comparison‌​‌ between IEEE 802.11p and​​ IEEE 802.11bd for vehicular​​​‌ communications (V2X), focusing on‌ throughput and MAC-layer delay‌​‌ under saturated traffic conditions.​​ While IEEE 802.11p has​​​‌ been widely adopted in‌ early V2X deployments, its‌​‌ performance degrades significantly in​​ dense scenarios, motivating the​​​‌ development of IEEE 802.11bd‌ with enhanced physical-layer capabilities‌​‌ and backward compatibility. The​​ analysis builds on Bianchi’s​​​‌ classical Markov chain model‌ for IEEE 802.11 DCF‌​‌ and extends it to​​ account for the improved​​​‌ decoding and reliability features‌ of IEEE 802.11bd. A‌​‌ single parameter is introduced​​ to capture the probability​​​‌ of successful packet reception‌ in the presence of‌​‌ collisions, allowing both standards​​​‌ to be analyzed within​ a unified framework. Closed-form​‌ expressions for saturated throughput​​ are derived as a​​​‌ function of network density​ and this reliability parameter.​‌ Results show that IEEE​​ 802.11bd achieves higher throughput​​​‌ than IEEE 802.11p, with​ performance gains that increase​‌ in dense vehicular environments.​​ The paper also extends​​​‌ a renewal-process-based approach to​ characterize MAC-layer delay. While​‌ improved decoding enhances throughput,​​ the analysis reveals a​​​‌ trade-off whereby increased reliability​ can lead to longer​‌ delays under fully saturated​​ conditions due to higher​​​‌ channel occupancy.

Overall, this​ study clarifies the mechanisms​‌ through which IEEE 802.11bd​​ improves performance over IEEE​​​‌ 802.11p and highlights the​ throughput–delay trade-offs inherent to​‌ contention-based V2X systems. The​​ results provide useful insights​​​‌ for understanding the expected​ behavior of next-generation vehicular​‌ networks under high load.​​

This work is published​​​‌ in 45.

8.5.5​ Inferring Posted Speed Limits​‌ from Cooperative Awareness Messages:​​ A Neural Speed Limit​​​‌ Estimator

Participants: Paul Mühlethaler​.

This study introduces​‌ an innovative AI-driven approach​​ to determine road speed​​​‌ regulations by leveraging real-time​ data shared between vehicles.​‌

This research work proposes​​ a neural network-based framework​​​‌ to infer posted speed​ limits by analyzing Cooperative​‌ Awareness Messages (CAMs) broadcast​​ by neighboring vehicles in​​​‌ an ITS-G5 environment. In​ many regions, digital maps​‌ and physical signage are​​ incomplete or outdated, creating​​​‌ significant challenges for autonomous​ systems that require precise​‌ knowledge of local speed​​ laws. By treating surrounding​​​‌ vehicles as mobile probes,​ a "Neural Speed Limit​‌ Estimator" uses the collective​​ speed and positioning data​​​‌ of traffic to predict​ the current limit with​‌ high accuracy. The model​​ is trained and validated​​​‌ using real-world datasets, demonstrating​ its robustness against variations​‌ in traffic density and​​ diverse driver behaviors. .​​​‌ This architecture is designed​ to complement computer vision​‌ systems, acting as a​​ secondary layer of verification​​​‌ when cameras are obstructed​ by weather, lighting, or​‌ large obstacles. The study​​ also explores the trade-off​​​‌ between the number of​ required CAM samples and​‌ the speed of the​​ estimation convergence. Ultimately, this​​​‌ approach offers a scalable,​ decentralized solution that enhances​‌ environmental awareness for connected​​ vehicles without requiring expensive​​​‌ infrastructure upgrades. It effectively​ reduces uncertainty in unmarked​‌ road segments, paving the​​ way for safer and​​​‌ more compliant next-generation intelligent​ transportation systems.

This work​‌ is published in 48​​.

8.5.6 Secure and​​​‌ Decentralized Networking

Participants: Paul​ Mühlethaler.

Work in​‌ 2025 placed a strong​​ emphasis on leveraging decentralized​​​‌ technologies like Blockchain and​ IPFS to enhance security​‌ and transparency in cloud​​ and data sharing environments.​​​‌

Cloud Auditing and Transparency:​

The study "Towards Secure​‌ and Transparent Cloud Auditing:​​ A Blockchain and IPFS-Driven​​​‌ Framework with Batch Verification"​ outlines a new framework​‌ to ensure secure and​​ transparent auditing within cloud​​​‌ environments. This study proposes​ a decentralized framework to​‌ enhance the security and​​ transparency of data integrity​​​‌ auditing in cloud storage.​ Traditional auditing often relies​‌ on a Third-Party Auditor​​ (TPA), which introduces risks​​​‌ of centralization, corruption, or​ single points of failure.​‌ To eliminate the need​​ for a TPA, the​​ authors integrate Blockchain technology​​​‌ to record auditing logs‌ and ensure non-repudiation of‌​‌ results. The framework utilizes​​ IPFS (InterPlanetary File System)​​​‌ to store large auditing‌ metadata efficiently while keeping‌​‌ the blockchain overhead low.​​ A key feature is​​​‌ the use of identity-based‌ signatures, which simplifies key‌​‌ management by avoiding the​​ complexity of traditional certificates.​​​‌ To improve performance, the‌ system implements a batch‌​‌ verification mechanism, allowing multiple​​ auditing tasks to be​​​‌ processed simultaneously. This batching‌ significantly reduces the computational‌​‌ burden on the auditor​​ and lowers communication latency​​​‌ between parties. The architecture‌ ensures data privacy, as‌​‌ the actual file content​​ is never revealed to​​​‌ the auditors during the‌ verification process. By leveraging‌​‌ smart contracts, the framework​​ automates the auditing process,​​​‌ ensuring that results are‌ immutable and publicly verifiable.‌​‌ Security analysis demonstrates that​​ the scheme is resistant​​​‌ to data substitution and‌ replay attacks from malicious‌​‌ cloud service providers. Experimental​​ results show that the​​​‌ framework maintains high efficiency‌ even as the number‌​‌ of data blocks and​​ auditing tasks increases. Overall,​​​‌ it provides a robust‌ solution for trustless cloud‌​‌ environments, balancing decentralization with​​ high-performance verification. This work​​​‌ has been presented in‌ the 21st International Conference‌​‌ on Wireless and Mobile​​ Computing, Networking and Communications​​​‌ (WIMoB) in October 2025.‌

This work is published‌​‌ in 28.

Secure​​ Data Sharing:

The study​​​‌ "A blockchain-based framework for‌ secure data sharing and‌​‌ access control using IPFS​​ and ECC" details a​​​‌ robust mechanism for data‌ security, access control, and‌​‌ sharing using Blockchain, IPFS,​​ and Elliptic Curve Cryptography​​​‌ (ECC).

This paper presents‌ a decentralized framework designed‌​‌ to provide secure and​​ fine-grained access control for​​​‌ data sharing in untrusted‌ environments. It addresses the‌​‌ privacy concerns of centralized​​ cloud storage by combining​​​‌ Blockchain, IPFS, and Elliptic‌ Curve Cryptography (ECC). Data‌​‌ owners encrypt their files​​ using ECC, which offers​​​‌ high security with smaller‌ key sizes and lower‌​‌ computational overhead compared to​​ RSA. Instead of storing​​​‌ large encrypted files directly‌ on the blockchain, they‌​‌ are uploaded to IPFS,​​ ensuring efficient and distributed​​​‌ storage. The resulting IPFS‌ hash (the file's unique‌​‌ address) is then stored​​ on the blockchain to​​​‌ ensure data integrity and‌ traceability. Access control is‌​‌ managed via Smart Contracts,​​ which automatically enforce sharing​​​‌ rules and verify the‌ identity of data requesters.‌​‌ The framework ensures that​​ only authorized users with​​​‌ the correct cryptographic keys‌ can retrieve and decrypt‌​‌ the content from IPFS.​​ By using blockchain as​​​‌ a tamper-proof ledger, all‌ access requests and data‌​‌ transactions are recorded for​​ transparent auditing. The decentralized​​​‌ nature of the system‌ eliminates the single point‌​‌ of failure typically found​​ in traditional database management​​​‌ systems. Security analysis shows‌ that the integration of‌​‌ ECC provides robust protection​​ against unauthorized access and​​​‌ man-in-the-middle attacks. Performance evaluations‌ indicate that the framework‌​‌ is computationally efficient, making​​ it suitable for resource-constrained​​​‌ environments. Overall, this approach‌ provides a scalable and‌​‌ trustless ecosystem for sharing​​ sensitive information across various​​​‌ industries, such as healthcare‌ or finance.

This work‌​‌ is published in 10​​​‌.

NDN Security:

In​ the domain of Named​‌ Data Networking (NDN), the​​ study "Deep Q-ICAN: A​​​‌ Deep Reinforcement Learning-based Approach​ for Real-time CPA Attack​‌ Detection and Mitigation in​​ NDN Architecture" proposes a​​​‌ Deep Reinforcement Learning (DRL)​ solution for real-time detection​‌ and mitigation of Content​​ Poisoning Attacks (CPA).

Deep​​​‌ Q-ICAN addresses the vulnerability​ of Named Data Networking​‌ (NDN) to Content Poisoning​​ Attacks (CPA). In a​​​‌ CPA, malicious nodes inject​ fake or corrupted data​‌ into the network, polluting​​ caches and wasting resources.​​​‌ The framework introduces a​ Deep Reinforcement Learning (DRL)​‌ approach to detect and​​ mitigate these attacks in​​​‌ real-time. It specifically utilizes​ a Deep Q-Network (DQN)​‌ to learn optimal defense​​ strategies based on dynamic​​​‌ network states. The "ICAN"​ component focuses on Intelligent​‌ Cache and Network management​​ to isolate suspicious content​​​‌ efficiently. The model monitors​ key metrics such as​‌ Cache Hit Ratio (CHR)​​ and interest satisfaction rates​​​‌ to identify anomalies. Once​ an attack is detected,​‌ the agent takes actions​​ to bypass malicious producers​​​‌ or purge poisoned cache​ entries. Unlike static threshold-based​‌ methods, Deep Q-ICAN adapts​​ to evolving attack patterns​​​‌ and traffic fluctuations. The​ architecture is designed to​‌ minimize computational overhead, ensuring​​ low latency for real-time​​​‌ traffic processing. Experimental results​ demonstrate that the approach​‌ significantly restores the data​​ delivery performance during attacks.​​​‌ It effectively protects legitimate​ users from receiving fake​‌ content while maintaining high​​ network throughput. Overall, Deep​​​‌ Q-ICAN provides a robust,​ self-learning security layer for​‌ next-generation information-centric architectures.

This​​ work is published in​​​‌ 13.

8.5.7 Network​ Topology Generation and Prediction​‌

Participants: Felix Marcoccia,​​ Paul Mühlethaler, Thomas​​​‌ Watteyne.

Significant effort​ was also directed towards​‌ developing advanced methods for​​ generating and updating complex​​​‌ network topologies, particularly for​ ad hoc and mobile​‌ networks, utilizing graph-based and​​ deep learning techniques.

Ad​​​‌ Hoc Topology Generation:

"TopoFormer​ An Efficient Link-Set Prediction​‌ Architecture for Ad Hoc​​ Network Topology Generation" proposes​​​‌ an efficient architecture based​ on link-set prediction for​‌ generating topologies in ad​​ hoc networks. TopoFormer is​​​‌ an innovative architecture designed​ to generate realistic topologies​‌ for Mobile Ad Hoc​​ Networks (MANETs) by predicting​​​‌ link sets. Unlike traditional​ models that focus on​‌ fixed infrastructures, TopoFormer addresses​​ the dynamic and decentralized​​​‌ nature of ad hoc​ connectivity. The core of​‌ the system is based​​ on a Transformer-inspired architecture,​​​‌ which is highly efficient​ at processing the relative​‌ positions of nodes. It​​ treats topology generation as​​​‌ a link-set prediction problem,​ determining which nodes should​‌ be connected based on​​ spatial and mobility features.​​​‌ By leveraging self-attention mechanisms,​ TopoFormer captures the global​‌ context of the network,​​ ensuring that the generated​​​‌ graphs are globally coherent.​ The model is specifically​‌ optimized for computational efficiency,​​ making it much faster​​​‌ than complex iterative or​ diffusion-based methods. It excels​‌ at maintaining graph-theoretic properties,​​ such as connectivity, average​​​‌ path length, and robustness​ against node failure. This​‌ approach allows researchers to​​ create large-scale, synthetic datasets​​​‌ that accurately mimic the​ behavior of real-world moving​‌ nodes. TopoFormer is particularly​​ useful for evaluating routing​​ protocols and network performance​​​‌ in scenarios where real‌ data is scarce. It‌​‌ outperforms existing heuristic and​​ deep learning baselines in​​​‌ terms of both generation‌ speed and structural accuracy.‌​‌ Overall, TopoFormer provides a​​ scalable solution for simulating​​​‌ the evolving structures of‌ next-generation tactical and emergency‌​‌ networks. The architecture ensures​​ that the generated topologies​​​‌ are not just random‌ graphs, but functional networks‌​‌ ready for protocol benchmarking.​​

This study is presented​​​‌ in 43.

Mobile‌ Topology Updates:

The working‌​‌ paper "NetDiff: Graph Diffusion​​ with Improved Global Capabilities​​​‌ to Generate and Update‌ Mobile Network Topologies" introduces‌​‌ a Graph Diffusion approach​​ with enhanced global capabilities​​​‌ specifically designed to generate‌ and update topologies for‌​‌ mobile networks. NetDiff is​​ a specialized Graph Diffusion​​​‌ Model designed to solve‌ the challenges of generating‌​‌ and updating mobile network​​ topologies. While traditional generative​​​‌ models like GANs often‌ fail to capture long-range‌​‌ spatial relationships, NetDiff introduces​​ a cross-attention mechanism to​​​‌ enhance its global capabilities.‌ This allows the model‌​‌ to understand the complex​​ dependencies between base stations,​​​‌ ensuring that generated layouts‌ respect real-world constraints and‌​‌ density patterns. A key​​ innovation is its dual-purpose​​​‌ design: it can generate‌ entirely new synthetic networks‌​‌ or incrementally update existing​​ ones to reflect urban​​​‌ growth. The diffusion process‌ works by gradually denoising‌​‌ graph structures, which leads​​ to higher statistical fidelity​​​‌ compared to older methods.‌ By accurately preserving graph‌​‌ metrics such as degree​​ distribution and clustering coefficients,​​​‌ NetDiff provides a robust‌ foundation for 6G network‌​‌ planning. Ultimately, it offers​​ telecommunication researchers a scalable​​​‌ tool to simulate realistic‌ infrastructures for testing resource‌​‌ management and coverage optimization.​​

This study is presented​​​‌ in 44.

9.3.1 Alfred​​ Audio

The team spun​​​‌ off its second startup​ called Alfred Audio in​‌ 2025. The Chief Technical​​ Officer (CTO) and  co-founder​​​‌ of the company is​ Said Alvarado-Marin who defended​‌ his PhD thesis in​​ the group in 2025.​​​‌ The company works on​ wireless solutions for live​‌ audio transmission. The company​​ leverages the expertise of​​​‌ the team in wireless​ embedded systems to create​‌ a low-latency network stack.​​ The solution is currently​​​‌ under development with open​ interest from stakeholders. The​‌ company won the i-PhD​​ award from Bpifrance. The​​​‌ relationship remains very tight​ with the AIO team,​‌ as the company is​​ hosted at the Inria​​​‌ Startup Studio two floors​ above AIO, and uses​‌ the AIO Experimentation Space​​ for its experiments.

10.1.1 Inria​‌ associate team not involved​​ in an IIL or​​​‌ an international program

The​ SWARM2 associate team is​‌ with Prof. Pister's team​​ at UC Berkeley and​​​‌ Prof. Burnett's team at​ Villanova University. The SWARM2​‌ team, and its predecessor,​​ SWARM, have been running​​​‌ since 2021. The Grand​ Challenge of this joint​‌ research is to empower​​ swarms of micro-robots built​​​‌ around an advanced crystal-free​ micro-mote with standards-compliant networking.​‌ This is a key​​ step towards enabling swarms​​​‌ of coordinated micro-robots to​ carry out exploration and​‌ mapping expeditions in hard-to-reach​​ locations, such as a​​​‌ collapsed building after an​ earthquake.

10.2.1 Visits of​​ international scientists

10.2.2 Visits​​ to international teams

10.3.1‌ Horizon Europe

10.3.2 Other​​​‌ european programs/initiatives

• Title:
  embedded-cal:‌ Embedded Cryptographic Abstraction Layer‌​‌
• Duration:
  From October 1,​​​‌ 2025 to September 30,​ 2026
• Partners:
  Inria, Cryspen,​‌ Christian Amsüss
• Inria contact:​​
  Mališa Vučinić
• Coordinator:
  Mališa​​​‌ Vučinić
• Summary:
  Funded by​ NLnet grant, embedded-cal develops​‌ a verified implementation of​​ the cryptographic provider in​​​‌ Rust which is compatible​ with popular embedded platforms.​‌ This cryptographic provider will​​ be 1) fast on​​​‌ popular embedded platforms; 2)​ resistant to certain classes​‌ of side-channel attacks; 3)​​ usable without the Rust​​​‌ standard library. The module​ will lever the available​‌ hardware acceleration support of​​ popular microcontroller units for​​​‌ embedded systems and fill​ in the gaps in​‌ hardware support through software​​ implementations. The module will​​​‌ be formally verified for​ secret independence using the​‌ hax framework, a verification​​ tool for high assurance​​​‌ code.

10.4.1 Inria Exploratory Research​‌ (AEx) SDMote

• lead: Filip​​ Maksimovic
• period: 2021-2024

The​​​‌ goal of the SDMote​ project is to develop​‌ a software-reconfigurable wireless hardware​​ platform, consisting of a​​​‌ low-power FPGA running a​ RISC-V soft core and​‌ a wide-band wireless transceiver.​​ This entire battery-powered embedded​​​‌ platform is open-source. SDMote​ is the next-generation IoT​‌ hardware that empowers the​​ research community to design​​​‌ custom digital peripherals and​ radio configurations, giving it​‌ the ultimate flexibility to​​ address applications that cannot​​​‌ be addressed with today's​ off-the-shelf motes. Filip Maksimovic​‌ leads.

The current state​​ of the project is​​​‌ primarily development in software.​ Customizable RISC-V soft cores​‌ have been synthesized and​​ ported to FPGAs with​​​‌ both memory mapped and​ peripherally accessible interfaces to​‌ custom digital receiver hardware.​​ At the moment, this​​​‌ still uses closed-source and​ proprietary synthesis tools, but​‌ work is progressing on​​ an open-source flow. The​​​‌ wireless front-end that is​ used is a fairly​‌ limited custom integrated circuit.​​ Work with a multi-application​​​‌ off-the-shelf transceiver is underway.​

10.4.2 PEPR 5G -​‌ FITNESS

• AIO lead: Filip​​ Maksimovic
• period: 01-May-2023 –​​​‌ 31-Dec-2027

The goal of​ this subject is an​‌ investigation of both the​​ limitations of extremely size-limited​​​‌ (millimeter-scale) wireless communication devices​ as well as their​‌ ability to maintain reliable​​ communication in unfavorable and​​​‌ changing wireless environments. These​ tiny devices have short​‌ range due to low​​ antenna gain, low transmit​​​‌ power, and low receive​ sesitivity. However, networks of​‌ these devices create new​​ opportunities with applications in​​​‌ micro-robotics, high resolution sensing,​ and smart medicine.

Experiments​‌ will be performed both​​ with an academic research​​​‌ platform and with off-the-shelf​ hardware designed for wireless​‌ mesh networking. A number​​ of performance metrics will​​​‌ be evaluated on these​ experimental hardware platforms including​‌ individual and network-level energy​​ consumption, packet delivery rate,​​​‌ and latency. Furthermore, the​ coexistence of crystal-free and​‌ crystal-enabled de17 vices will​​ be investigated. A crystal-free​​​‌ network has the potential​ benefit of further miniaturization​‌ and reduced energy consumption,​​ but requires overhead due​​​‌ to consistent requirement to​ maintain time and channel.​‌ Additional concerns that merit​​ investigation in networks with​​​‌ large numbers of elements​ are the tradeoffs between​‌ reliability and latency and​​ the effects of of​​​‌ receive and transmit linearity​ on packet delivery. There​‌ exists a fundamental tradeoff​​ between network traffic and​​ energy consumption. For crystal-free​​​‌ devices, this tradeoff is‌ exacerbated by requiring more‌​‌ keepalive packets if there​​ is a faster change​​​‌ in their environment (like‌ a rapid temperature change).‌​‌ Because of their limited​​ resources, an investigation of​​​‌ the security vulnerabilities of‌ both the devices and‌​‌ their networks is necessary,​​ as proven by recent​​​‌ IoT BotNet denial-of-service attacks.‌ The final portion of‌​‌ the project will be​​ deployments of these large​​​‌ scale sensor networks for‌ agricultural monitoring, early-warning wildfire‌​‌ detection with temperature sensing,​​ and an evaluation of​​​‌ network performance and adaptation‌ in a changing wireless‌​‌ channel by building a​​ time synchronized mesh network​​​‌ on a swarm of‌ robots.

The expected outcomes‌​‌ are: a simulated evaluation​​ of energy optimization of​​​‌ crystal-free mesh networks with‌ experimental validation, an evaluation‌​‌ of the security requirements​​ for miniature wireless devices,​​​‌ and real-world deployments that‌ stress the reliability of‌​‌ the network when latency​​ is critical (by testing​​​‌ the control of a‌ large number of connected‌​‌ swarm robots) and when​​ packet delivery and minimal​​​‌ energy consumption are critical‌ (in a smart-agriculture monitoring‌​‌ application).

10.4.3 PEPR 5G​​ - NF-HiSec

• AIO lead:​​​‌ Mališa Vučinić
• period: 01-May-2023‌ – 31-Dec-2027

The HiSec‌​‌ project focuses on cyber-security​​ issues in future networks.​​​‌ These networks have played‌ a key role in‌​‌ service delivery for digital​​ infrastructures. These new networking​​​‌ technologies have also penetrated‌ essential and critical services‌​‌ for our daily lives,​​ such as energy, transportation​​​‌ or healthcare. The pervasive‌ use of digital services‌​‌ and networks to control​​ these critical infrastructures significantly​​​‌ increases the attack surface‌ and the opportunities for‌​‌ attackers. We regularly observe​​ attacks against these infrastructures,​​​‌ leading to successful compromise‌ and very significant impacts.‌​‌ The objective of the​​ NF-HiSec project is thus​​​‌ to handle cybersecurity issues‌ in these environments, and‌​‌ propose new mechanisms to​​ protect these networks and​​​‌ detect attacks, attacks against‌ the networking infrastructure itself,‌​‌ or against the services​​ hosted or the users​​​‌ of the network. The‌ project is organized in‌​‌ five work-packages. The first​​ work-package addresses the definition​​​‌ and deployment of security‌ policies that are specific‌​‌ to these future networks,​​ taking into account the​​​‌ specificities of virtual environments,‌ the requirements of endpoint‌​‌ security, and the deployment​​ of end-to-end security properties.​​​‌ The second work-package deals‌ with operational security for‌​‌ these networks, around specific​​ mechanisms for attack detection​​​‌ in virtual or decentralized‌ environments, and taking into‌​‌ account the specificities of​​ the Internet of Things.​​​‌ The third work-package deals‌ with personal information protection,‌​‌ to provide new tools​​ enabling legal interception. The​​​‌ fourth work-package works on‌ modeling security properties in‌​‌ future networks, to ensure​​ that these networks will​​​‌ provide to end users‌ the security services that‌​‌ they require. The fifth​​ work-package focuses on the​​​‌ link between hardware and‌ software on one hand,‌​‌ and cybersecurity properties on​​ the other hand, to​​​‌ ensure a strong integration‌ of these properties in‌​‌ foundational network fabric.

11.1.1 Scientific events: organisation‌

11.1.2 Scientific events: selection​​

11.1.3 Journal

11.1.4 Invited talks​​

• Mališa Vučinić gave a​​​‌ talk at the Swarm​ Workshop in Brussels on​‌ 26 November 2025. The​​ talk title was “Lightweight​​​‌ Authenticated Key Exchange for​ Internet-of-Things Use Cases”.
• Mališa​‌ Vučinić gave a talk​​ at MIRES 2025 gathering​​​‌ in Poitiers, France on​ 8 July 2025 on​‌ “Lightweight Authenticated Key Exchange​​ for Internet-of-Things Use Cases”.​​​‌
• Mališa Vučinić gave a​ talk at PEPR 5G​‌ days in Bordeaux, France​​ on 25 June 2025​​​‌ on “Lightweight Authenticated Key​ Exchange for Internet-of-Things Use​‌ Cases”.
• Filip Maksimovic was​​ invited to give a​​​‌ presentation at the CANELOS​ seminar in Valparaiso, Chile.​‌
• Filip Maksimovic gave a​​ talk at an online​​​‌ PEPR workshop in September​ 2025 on "Mitigating Interference​‌ in High-Density Multi-Standard Wireless​​ Area Networks".
• Geovane Fedrecheski​​​‌ gave a talk entitled​ “Mari: connecting a 1,000​‌ robot swarm testbed” at​​ Hong Kong University of​​​‌ Science and Technology campus​ Guangzhou (HKUST-GZ).

11.1.5 Standardization​‌

The Internet Engineering Task​​ Force (IETF) is one​​​‌ of the major standardization​ bodies for networking technology,​‌ and is behind protocols​​ such as TCP, IP​​​‌ and HTTP.

Mališa Vučinić​ is co-chair of the​‌ IETF LAKE working group.​​ This is a very​​​‌ significant scientific responsibility. The​ two co-chairs (the other​‌ is Renzo Navas from​​ IMT Atlantique) steer and​​​‌ trigger the work of​ the working group (WG).​‌ The activity of the​​ LAKE group is followed​​ by 116 people, with​​​‌ a healthy mix of‌ industrial and academia contributors.‌​‌ Geovane Fedrecheski , Yuxuan​​ Song , and Elsa​​​‌ Lopez-Perez participate in the‌ IETF LAKE standardization effort.‌​‌

11.2.1‌​‌ Supervision

Mališa Vučinić supervises​​ the PhD theses of:​​​‌

• Sara Faour (graduates in‌ April 2026),
• Yuxuan Song‌​‌ ,
• Elsa Lopez-Perez .​​

Paul Mühlethaler supervised the​​​‌ PhD thesis of:

• Felix‌ Marcoccia (graduated in October‌​‌ 2025).

Filip Maksimovic supervised​​ the PhD thesis of:​​​‌

• Said Alvarado-Marin (with Thomas‌ Watteyne , graduated in‌​‌ June 2025).

Thomas Watteyne​​ supervised the PhD theses​​​‌ of:

• Martina Balbi (graduates‌ in March 2026),
• Fabian‌​‌ Graf (graduated in December​​ 2025).

11.2.2 Juries

• Mališa​​​‌ Vučinić served as a‌ reviewer (rapporteur) for a‌​‌ PhD thesis of Jiali​​ XU. The PhD thesis​​​‌ is on Characterisation of‌ Anomalous Behaviour for Security‌​‌ in Deep-Edge Wireless Systems.​​
• Mališa Vučinić served on​​​‌ the Comité de Suivi‌ (PhD progress monitoring committee)‌​‌ of Elias Maharmeh, a​​ PhD candidate in the​​​‌ Inria ASTRA team, under‌ the supervision of Fawzi‌​‌ Nashashibi (Inria) and Paulo​​​‌ Resende (Valeo). The PhD​ is on “Integrity and​‌ Robustness of Algorithms for​​ Localization and Autonomous Driving”.​​​‌
• Paul Mühlethaler served as​ a member of Felix​‌ Marcoccia ’s PhD defense​​ committee at Inria Paris​​​‌ (October 13). This PhD​ thesis (Topology Optimization in​‌ Mobile Wireless Networks Using​​ Machine Learning) investigates how​​​‌ machine learning can optimize​ network topology in mobile​‌ wireless and aerial ad​​ hoc networks using directional​​​‌ antennas. It proposes efficient​ neural architectures, including Transformer-based​‌ models and diffusion models,​​ to jointly generate network​​​‌ links and transmission schedules​ in real time. The​‌ results show significant throughput​​ and scalability improvements over​​​‌ traditional omnidirectional protocols, especially​ as network size increases.​‌
• Paul Mühlethaler served as​​ President of the PhD​​​‌ defense committee for Najoua​ Ben Alaya at Inria​‌ Saclay on December 1.​​ This PhD thesis (UAV​​​‌ Search Path Planning for​ Cattle Monitoring: From Linear​‌ Programming to Learning-Based Approaches)​​ addresses UAV search path​​​‌ planning for cattle monitoring​ under uncertainty, with the​‌ goal of minimizing expected​​ search time in large,​​​‌ dynamic environments. It introduces​ the UAV Cattle Search​‌ (UCS) problem and combines​​ exact linear programming models​​​‌ with learning-based approaches (reinforcement​ learning and Transformer architectures)​‌ to balance optimality and​​ scalability. The proposed methods​​​‌ demonstrate that AI-driven path​ planning can efficiently generalize​‌ to large-scale scenarios, beyond​​ livestock monitoring, to broader​​​‌ combinatorial search problems.
• Paul​ Mühlethaler served as an​‌ examiner for the HDR​​ defense of Abdelwahab Boualouache​​​‌ at the University of​ Burgundy on December 1​‌ (remote participation). This HDR​​ thesis (AI-Driven Security for​​​‌ Next-Generation 5G and V2X​ Networks) explores AI-driven security​‌ frameworks for next-generation 5G​​ and V2X networks, addressing​​​‌ intrusion detection, zero-day attacks,​ and proactive cyber defense.​‌ It proposes advanced AI​​ architectures (RNNs, federated learning,​​​‌ self-supervised learning, and multi-agent​ reinforcement learning) to enable​‌ early detection, privacy-preserving collaboration,​​ and adaptive threat mitigation.​​​‌ The work demonstrates how​ scalable, lightweight, and intelligent​‌ security systems can be​​ deployed across core, edge,​​​‌ and vehicular networks to​ protect critical digital infrastructures.​‌
• Filip Maksimovic is currently​​ on the dissertation committee​​​‌ of Jacob Louie at​ the Hong Kong University​‌ of Science and Technology:​​ Guangzhou.

11.2.3 Educational and​​​‌ pedagogical outreach

• Alexandre Abadie​ and Yuxuan Song gave​‌ the DotBot demo at​​ RJMI in November 2025.​​​‌
• Mališa Vučinić gave a​ presentation “Dependable Networking, Low-power​‌ Wireless and Micro-robotics”. Geovane​​ Fedrecheski and Yuxuan Song​​​‌ gave the DotBot demo​ at CHICHE on 8​‌ December 2025.
• Alfonso Cortes​​ served as supervisor for​​​‌ the digital design track​ of the in-person Workshop​‌ at CAD and Nanoelectronics​​ Seminar (Canelos) 2025, Valparaíso,​​​‌ Chile, from August 20​ to August 22.

11.3.1 Specific official​​ responsibilities in science outreach​​​‌ structures

• Alfonso Cortes founded​ the Chilean Microelectronics Foundation​‌ in August 2025 and​​ served as Secretary General.​​​‌

11.3.2 Inria Museum Demo​

An installation of musicians​‌ of the embedded orchestra​​ that plays at the​​​‌ Inria museum.

An installation​ of musicians of the​‌ embedded orchestra that plays​​ at the Inria museum.​​​‌

We installed at​​​‌ the Inria museum a‌ variant of the AIOT‌​‌ Play board that is​​ able to play music​​​‌ (see 6). We‌ deployed 20 of these‌​‌ motes all across the​​ museum. When visitors solve​​​‌ a little riddle, they‌ collectively play an orchestral‌​‌ arrangement of a popular​​ song.

11.3.3 Videos

• Video:​​​‌ Demo "Remote Attestation over‌ EDHOC"
• Video: Demo: zero-touch‌​‌ onboarding of DotBots with​​ draft-ietf-lake-authz
• Video: OpenSwarm hackathon​​​‌ at Inria May 2025‌
• Video: Mari is a‌​‌ new link layer for​​ connecting robot swarms

11.3.4​​​‌ Others science outreach relevant‌ activities

• Mališa Vučinić gave‌​‌ a lecture on “Dependable​​ Networking, Low-power Wireless and​​​‌ Micro-robotics” on 25 September‌ 2025 for a Montenegrin‌​‌ Science Promotion Hub EPISTEME.​​ He also gave a​​​‌ series of mentorship sessions‌ to EPISTEME participants.
• Mališa‌​‌ Vučinić gave a talk​​ “Dependable Networking, Low-power Wireless​​​‌ and Micro-robotics” for a‌ group of high-school students‌​‌ visiting Inria on 25​​ June 2025.
• Alexandre Abadie​​​‌ and Geovane Fedrecheski gave‌ a presentation on the‌​‌ OpenSwarm project and the​​ DotBot demo at JNRR’25​​​‌ in Rennes.
• Geovane Fedrecheski‌ gave a presentation on‌​‌ the OpenSwarm project and​​ the DotBot demo in​​​‌ Salvador Arena Foundation, São‌ Paulo, Brazil.
• Geovane Fedrecheski‌​‌ gave a presentation on​​ the OpenSwarm project and​​​‌ the DotBot demo in‌ Federal Technological University, Paraná,‌​‌ in Brazil.
• Geovane Fedrecheski​​ gave a presentation of​​​‌ the OpenSwarm project and‌ the DotBot demo at‌​‌ Latinoware, the 22nd Latin​​ American Congress on Free​​​‌ Software and Open Technologies‌ in Foz do Iguaçu,‌​‌ Brazil.
• Geovane Fedrecheski ,​​ Martina Balbi and Thomas​​​‌ Watteyne gave a 2-day‌ DotBot Academy course at‌​‌ L'École supérieure d'ingénieurs Réunion​​ océan Indien (ESIROI), in​​​‌ Saint Piere, island of‌ La Réunion/
• Alfonso Cortes‌​‌ gave an in-person talk​​ to first year electronic​​​‌ engineering students at UTFSM,‌ Valparaíso, Chile, entitled “Working‌​‌ as a Research &​​ Development Engineer at Inria”,​​​‌ on August 22 2025.‌
• Alfonso Cortes gave an‌​‌ online talk in the​​ context of a Computer​​​‌ Architecture course at UTFSM‌ entitled “Open-source implementation of‌​‌ RISC-V systems”, on November​​ 27.

International journals

• 7​‌ articleS.Said Alvarado-Marin​​, A.Alexandre Abadie​​​‌, M.Martina Balbi​, T.Thomas Watteyne​‌ and F.Filip Maksimovic​​. Automatic Lighthouse Calibration​​​‌ Using Conics for Indoor​ Robot Localization.IEEE​‌ Robotics and Automation Letters​​107July 2025​​​‌, 7532-7538HALDOI​back to text
• 8​‌ articleM. M.Martina​​ Maria Balbi Antunes,​​​‌ L.Lance Doherty and​ T.Thomas Watteyne.​‌ A Comprehensive Survey on​​ Channel Hopping and Scheduling​​​‌ Enhancements for TSCH Networks​.Journal of Network​‌ and Computer Applications (JNCA)​​238June 2025HAL​​​‌back to text
• 9​ articleM. M.Martina​‌ Maria Balbi Antunes,​​ L.Lance Doherty and​​​‌ T.Thomas Watteyne.​ Embedded Artificial Intelligence for​‌ IoT Applications Using the​​ MAX78000.IEEE Access​​​‌132025, 38979-39005​HALDOIback to​‌ text
• 10 articleH.​​Houaida Ghanmi, N.​​​‌Nasreddine Hajlaoui, H.​Haifa Touati, S.​‌Saadi Boudjit, M.​​Mohamed Hadded and P.​​​‌Paul Muhlethaler. A​ blockchain-based framework for secure​‌ data sharing and access​​ control using IPFS and​​​‌ ECC.Cluster Computing​2815October 2025​‌, 1007HALDOI​​back to text
• 11​​​‌ articleF.Fabian Graf​, D.David Pauli​‌, M.Michael Villnow​​ and T.Thomas Watteyne​​​‌. Management of 6TiSCH​ Networks Using CORECONF: A​‌ Clustering Use Case.​​IEEE Transactions on Network​​​‌ and Service Management2025​, 1-1HALDOI​‌back to text
• 12​​ articleF.Fabian Graf​​​‌, T.Thomas Watteyne​ and M.Michael Villnow​‌. What to Expect​​ when Using DECT NR+​​​‌.IEEE Communications Magazine​2025. In press.​‌ HALback to text​​
• 13 articleA.Abdelhak​​​‌ Hidouri, H.Haifa​ Touati, M.Mohamed​‌ Hadded, M. A.​​Mohamed Amin Asri,​​​‌ N.Nasreddine Hajlaoui,​ P.Paul Muhlethaler and​‌ S.Samia Bouzefrane.​​ Deep Q-ICAN: A Deep​​ Reinforcement Learning-based Approach for​​​‌ Real-time CPA Attack Detection‌ and Mitigation in NDN‌​‌ Architecture.Computer Networks​​August 2025, 111604​​​‌HALDOIback to‌ text
• 14 articleE.‌​‌ L.Elsa Lopez Perez​​, T.Thomas Watteyne​​​‌, R.Rafael Marin-Lopez‌, C.Cristina Onete‌​‌, C.Clément Papon​​ and M.Mališa Vučinić​​​‌. Pre-Shared Key Authentication‌ with EDHOC: the Security-Performance‌​‌ Tradeoff.IEEE Access​​2025, 1-1HAL​​​‌DOIback to text‌

International peer-reviewed conferences

• 15‌​‌ inproceedingsA.Alexandre Abadie​​, A.Alfonso Cortes-Neira​​​‌, T.Thomas Watteyne‌ and F.Filip Maksimovic‌​‌. UpGate: Low-Power and​​ Secure Over-the-Air Hardware Reconfiguration​​​‌ for Deployed FPGAs.‌ICIN 2025 - 28th‌​‌ Conference on Innovation in​​ Clouds, Internet and Networks​​​‌Paris, FranceIEEEMarch‌ 2025, 28-35HAL‌​‌DOIback to text​​
• 16 inproceedingsA.Alexandre​​​‌ Abadie, T.Titan‌ Yuan, T.Thomas‌​‌ Watteyne and F.Filip​​ Maksimovic. An Open-Source​​​‌ Multi-Platform SDK for Programming‌ SCuM.The 2025‌​‌ Workshop on Crystal-Free/-Less Radio​​ and System-based Research for​​​‌ IoT (CrystalFreeIoT 2025) -‌ EWSN 2025Leuven (BE),‌​‌ BelgiumSeptember 2025HAL​​back to text
• 17​​​‌ inproceedingsA.Alexandre Abadie‌, K.Koen Zandberg‌​‌, T.Thomas Watteyne​​ and F.Filip Maksimovic​​​‌. A RIOT Port‌ for the Single-Chip Mote‌​‌.The 2025 Workshop​​ on Crystal-Free/-Less Radio and​​​‌ System-based Research for IoT‌ (CrystalFreeIoT 2025) - EWSN‌​‌ 2025Leuven, BelgiumSeptember​​ 2025HALback to​​​‌ text
• 18 inproceedingsS.‌Said Alvarado-Marin, C.‌​‌Cristobal Huidobro-Marin, M.​​Martina Balbi, T.​​​‌Trifun Savić, T.‌Thomas Watteyne and F.‌​‌Filip Maksimovic. Lighthouse​​ Localization of Miniature Wireless​​​‌ Robots.ICRA 2025‌ - IEEE International Conference‌​‌ on Robotics and Automation​​Atlanta, United StatesMay​​​‌ 2025HALback to‌ text
• 19 inproceedingsS.‌​‌Said Alvarado-Marin, A.​​Arnaud Taffanel, M.​​​‌Marcus Eliasson, F.‌Filip Maksimovic and T.‌​‌Thomas Watteyne. Demo:​​ Multi-sensor Lighthouse v2 Decoding​​​‌ Optimized for Low-Power CPU‌.EWSN 2025 -‌​‌ 22nd International conference on​​ Embedded Wireless Systems and​​​‌ NetworkLeuven, BelgiumSeptember‌ 2025HALback to‌​‌ text
• 20 inproceedingsD.​​Diego Badillo-San-Juan, S.​​​‌Said Alvarado-Marin, T.‌Thomas Watteyne and F.‌​‌Filip Maksimovic. Successive​​ Interference Cancellation Prototype for​​​‌ BLE and IEEE 802.15.4‌ in the Physical Layer‌​‌.European Signal Processing​​ Conference (EUSIPCO)Palermo, Italy​​​‌September 2025HALback‌ to text
• 21 inproceedings‌​‌D.Diego Badillo-San-Juan,​​ A.Alfonso Cortes,​​​‌ S.Said Alvarado-Marin,‌ A.Alexandre Abadie,‌​‌ F.Fabian Graf,​​ T.Thomas Watteyne and​​​‌ F.Filip Maksimovic.‌ Experimental Investigation of Bit‌​‌ Errors in Coexisting BLE​​ and IEEE 802.15.4 Channels​​​‌.Performance Evaluation and‌ Modeling in Wired and‌​‌ Wireless Networks (PEMWN 2025)​​Paris, FranceNovember 2025​​​‌HALback to text‌
• 22 inproceedingsM.Manjiang‌​‌ Cao, F.Filip​​ Maksimovic, D.David​​​‌ Burnett, T.Thomas‌ Watteyne, K. S.‌​‌Kris S J Pister​​ and T.Tengfei Chang​​​‌. Angle of Arrival‌ Localization for Single-Chip Micro‌​‌ Mote.CrystalFreeIoT 2025​​​‌ - The 2025 Workshop​ on Crystal-Free/-Less Radio and​‌ System-based Research for IoT​​Leuven, BelgiumSeptember 2025​​​‌HALback to text​
• 23 inproceedingsN.Narmin​‌ Elkilani, B.Baptiste​​ Carbillet, G.Geovane​​​‌ Fedrecheski, T.Trifun​ Savić and T.Thomas​‌ Watteyne. Demo: Vega​​ – Turning a Toy​​​‌ into a Ready-to-Use Robotic​ Platform.EWSN 2025​‌ - European Conference on​​ Wireless Sensor NetworksLeuven,​​​‌ BelgiumSeptember 2025HAL​back to text
• 24​‌ inproceedingsS.Sara Faour​​, M.Mališa Vučinić​​​‌, F.Filip Maksimovic​, T.Thomas Watteyne​‌ and K.Kristofer Pister​​. ODHD: On-Demand Helper​​​‌ Data Generation for Reliable​ NVM-Free Key Derivation from​‌ SRAM PUF.ISC​​ 2025 - IEEE International​​​‌ Conference on Information Security​ And CryptologyAnkara, Turkey​‌October 2025HALback​​ to text
• 25 inproceedings​​​‌S.Sara Faour,​ M.Mališa Vučinić,​‌ T.Thomas Watteyne and​​ K. S.Kristofer S.J.​​​‌ Pister. Two-Stage Threshold-based​ Majority Voting Scheme (TS-TMVS)​‌ for Robust SRAM PUFs​​.Workshop on Crystal-Free/-Less​​​‌ Radio and System-based Research​ for IoT (CrystalFreeIoT), International​‌ Conference on Embedded Wireless​​ Systems and Networks (EWSN)​​​‌Leuven (Belgique), BelgiumSeptember​ 2025HALback to​‌ text
• 26 inproceedingsG.​​Geovane Fedrecheski, Y.​​​‌Yinghao Gao, A.​Alexandre Abadie, S.​‌Said Alvarado-Marin, M.​​Mališa Vučinić, F.​​​‌Filip Maksimovic and T.​Thomas Watteyne. Demo:​‌ Mari Allows Connecting Large​​ Scale Robot Swarms using​​​‌ TSCH over BLE and​ Multiple Independent Gateways.​‌International Conference on Embedded​​ Wireless Systems and Networks​​​‌ - EWSN 2025Leuven,​ BelgiumSeptember 2025HAL​‌back to text
• 27​​ inproceedingsG.Geovane Fedrecheski​​​‌, G.Göran Selander​, T.Thomas Watteyne​‌ and M.Mališa Vučinić​​. ELA: Secure, Lightweight,​​​‌ and Zero-Touch Enrollment for​ IoT Devices.DCOSS-IoT​‌ 2025 - The 21th​​ Annual International Conference on​​​‌ Distributed Computing in Smart​ Systems and the Internet​‌ of ThingsLucca, Italy​​June 2025HALback​​​‌ to textback to​ text
• 28 inproceedingsH.​‌Houaida Ghanmi, N.​​Nasreddine Hajlaoui, H.​​​‌Haifa Touati, S.​Saadi Boudjit, M.​‌Mohamed Hadded, M.-Y.​​Mohand-Yazid Saidi and P.​​​‌Paul Mühlethaler. Towards​ Secure and Transparent Cloud​‌ Auditing: A Blockchain and​​ IPFS-Driven Framework with Batch​​​‌ Verification.IEEE WiMob​ 2025 - 21st International​‌ Conference on Wireless and​​ Mobile Computing, Networking and​​​‌ CommunicationsMarrachech, MoroccoOctober​ 2025HALback to​‌ text
• 29 inproceedingsF.​​Fabian Graf, E.​​​‌Esteban Norena Arroyave,​ T.Thomas Watteyne and​‌ M.Michael Villnow.​​ HyPM: Hybrid Performance Metric​​​‌ Transmission in Low-Power Wireless​ Networks.ISCC 2025​‌ - 30th IEEE Symposium​​ on Computers and Communications​​​‌Bologna, ItalyJuly 2025​HALback to text​‌
• 30 inproceedingsF.Fabian​​ Graf, M.Michael​​​‌ Villnow and T.Thomas​ Watteyne. Zephyr and​‌ SmartMesh IP - Happy​​ Together.EWSN 2025​​​‌ - 22nd International Conference​ on Embedded Wireless Systems​‌ and NetworksLeuven, Belgium​​September 2025HALback​​​‌ to text
• 31 inproceedings​P.Pascal Lafourcade,​‌ E.Elsa Lopez Perez​​, C.Charles Olivier-Anclin​​, C.Cristina Onete​​​‌, C.Clément Papon‌ and M.Mališa Vučinić‌​‌. Fine-grained, privacy-augmenting LI-compliance​​ in the LAKE standard​​​‌ - Extended version.‌Lecture Notes in Computer‌​‌ Science LNCSESORICS 2025​​ - 30th European Symposium​​​‌ on Research in Computer‌ SecurityToulouse, FranceSpringer‌​‌September 2025HALback​​ to text
• 32 inproceedings​​​‌M.Mengyao Liu,‌ L.Lowie Deferme,‌​‌ T.Tom Van Eyck​​, S.Sam Michiels​​​‌, A.Alexandre Abadie‌, S.Said Alvarado-Marin‌​‌, F.Filip Maksimovic​​, G.Genki Miyauchi​​​‌, J.Jessica Jayakumar‌, M. S.Mohamed‌​‌ S. Talamali, T.​​Thomas Watteyne, R.​​​‌Roderich Gross and D.‌Danny Hughes. CapBot:‌​‌ Enabling Battery-Free Swarm Robotics​​.ICRA 2025 -​​​‌ IEEE International Conference on‌ Robotics & AutomationAtlanta‌​‌ (GA), United StatesMay​​ 2025HALback to​​​‌ text
• 33 inproceedingsM.‌Mengyao Liu, L.‌​‌Luiz Sampaio, J.​​Jonathan Oostvogels, K.​​​‌Kate O'Riordan, S.‌Sam Michiels, T.‌​‌Thomas Watteyne and D.​​Danny Hughes. SC-Canopy:​​​‌ A Tiered Network Architecture‌ for Indoor Robot Swarms‌​‌.Workshop on Crystal-Free/-Less​​ Radio and System-based Research​​​‌ for IoT (CrystalFreeIoT), International‌ Conference on Embedded Wireless‌​‌ Systems and Networks (EWSN)​​Leuven, BelgiumSeptember 2025​​​‌HALback to text‌
• 34 inproceedingsL.Luiz‌​‌ Sampaio, K.Kate​​ O'Riordan, D.Dara​​​‌ O'Sullivan, B.Brian‌ Coffey and T.Thomas‌​‌ Watteyne. Porting A​​ Real-Time Operating System for​​​‌ Embedded IoT Devices to‌ Linux as Part of‌​‌ the OpenSwarm Project.​​MECC 2025 - ACM​​​‌ International Workshop on MetaOS‌ for the Cloud-Edge-IoT Continuum‌​‌Rotterdam, NetherlandsMarch 2025​​HALback to text​​​‌
• 35 inproceedingsY.Yuxuan‌ Song, G.Geovane‌​‌ Fedrecheski, M.Mališa​​ Vučinić and T.Thomas​​​‌ Watteyne. AuRA: Remote‌ Attestation over EDHOC for‌​‌ Constrained Internet-of-Things Use Cases​​.ISCC 2025 -​​​‌ IEEE Symposium on Computers‌ and CommunicationsBologne, Italy‌​‌July 2025HALback​​ to textback to​​​‌ text
• 36 inproceedingsY.‌Yuxuan Song, M.‌​‌Muhammad Usama Sardar,​​ G.Geovane Fedrecheski,​​​‌ M.Mališa Vučinić and‌ T.Thomas Watteyne.‌​‌ When to Attest? Intra-and​​ Post-Handshake Attestation for IoT​​​‌ Swarms.CSCN 2025‌ - IEEE Conference on‌​‌ Standards for Communications and​​ NetworkingBologne, ItalySeptember​​​‌ 2025HALback to‌ text
• 37 inproceedingsM.‌​‌ S.Mohamed S. Talamali​​, G.Genki Miyauchi​​​‌, T.Thomas Watteyne‌, M. S.Micael‌​‌ S. Couceiro and R.​​Roderich Gross. Ready,​​​‌ Bid, Go! On-Demand Delivery‌ Using Fleets of Drones‌​‌ with Unknown, Heterogeneous Energy​​ Storage Constraints.AAMAS​​​‌ 2025 - ACM International‌ Conference on Autonomous Agents‌​‌ and Multi-Agent SystemsDetroit,​​ MI, United StatesMay​​​‌ 2025HALback to‌ text

Conferences without proceedings‌​‌

• 38 inproceedingsG.Genki​​ Miyauchi, M. S.​​​‌Mohamed S Talamali,‌ M.Mengyao Liu,‌​‌ L.Lowie Deferme,​​ T.Tom Van Eyck​​​‌, S.Sam Michiels‌, J.Jessica Jayakumar‌​‌, A.Alexandre Abadie​​, G.Geovane Fedrecheski​​​‌, M.Martina Balbi‌, S.Said Alvarado-Marin‌​‌, F.Felix Matzdorf​​​‌, J.Julian Rau​, D.Danny Hughes​‌, T.Thomas Watteyne​​ and R.Roderich Groß​​​‌. Energy Replenishment Strategies​ for Robot Swarms.​‌GRC 2025 - 1st​​ German Robotics ConferenceNuremberg,​​​‌ GermanyMarch 2025HAL​DOIback to text​‌

Doctoral dissertations and habilitation​​ theses

• 39 thesisS.​​​‌Said Alvarado-Marin. Communication-Aware​ Coordination of Multi-Robot Systems​‌.Sorbonne UniversitéJune​​ 2025HALback to​​​‌ text
• 40 thesisF.​Fabian Graf. Application​‌ Performance Monitoring and Management​​ of Low-power Wireless Networks​​​‌.Sorbonne UniversiteDecember​ 2025HALback to​‌ text
• 41 thesisF.​​Félix Marcoccia. Topology​​​‌ optimization in mobile wireless​ networks using machine learning​‌.Sorbonne UniversitéOctober​​ 2025HALback to​​​‌ text
• 42 thesisM.​Mališa Vučinić. Lightweight​‌ Solutions for a Secure​​ Internet of Things.​​​‌Ecole Normale Superieure de​ Paris - ENS Paris​‌October 2025HALback​​ to text

Reports &​​​‌ preprints

• 43 miscF.​Félix Marcoccia, V.​‌Victor Fagoo, G.​​Gilles Monzat, C.​​​‌Cédric Adjih, T.​Thomas Watteyne and P.​‌Paul Mühlethaler. NetDiff:​​ Graph Diffusion with Improved​​​‌ Global Capabilities to Generate​ and Update Mobile Network​‌ Topologies.December 2025​​HALback to text​​​‌
• 44 miscF.Félix​ Marcoccia, V.Victor​‌ Fagoo, G.Gilles​​ Monzat de Saint Julien​​​‌, C.Cédric Adjih​, T.Thomas Watteyne​‌ and P.Paul Mühlethaler​​. TopoFormer: An Efficient​​​‌ Link-Set Prediction Architecture for​ Ad Hoc Network Topology​‌ Generation.September 2025​​HALback to text​​​‌
• 45 reportP.Paul​ Muhlethaler. Analytical Comparison​‌ Between IEEE 802.11p and​​ IEEE 802.11bd Based on​​​‌ an Extended BianchiModel.​AIO Inria ParisDecember​‌ 2025HALback to​​ text
• 46 reportP.​​​‌Paul Muhlethaler. Comparative​ Survey of IEEE 802.11p​‌ and IEEE 802.11bd for​​ Vehicular Communication.AIO​​​‌ Inria ParisJuly 2025​HALback to text​‌
• 47 reportP.Paul​​ Mühlethaler. Delay Performance​​​‌ in ITS-G5-Based Vehicular Ad​ Hoc Networks (VANETs).​‌AIO Inria ParisJune​​ 2025HALback to​​​‌ text
• 48 reportP.​Paul Muhlethaler. Inferring​‌ Posted Speed Limits from​​ Cooperative Awareness Messages: A​​​‌ Neural Speed Limit Estimator​.AIO Inria Paris​‌October 2025HALback​​ to text

Software

• 49​​​‌ softwareD.Diego Badillo-San-Juan​ and F.Filip Maksimovic​‌. gr-sic.September​​ 2025 lic: GNU General​​​‌ Public License v3.0 or​ later.HALSoftware​‌ HeritageVCSback to​​ text