Section: Research Program

Vertical Perspective

As mentioned, future ubiquitous networks evolve in dynamic and unpredictable environments. Also, they can be used in a large scope of applications that have several expectations in terms of performance and different contextual limitations. In this heterogeneous context, IoT devices must support multiple applications and relay traffic with non-deterministic pattern.

To make our solutions practical and efficient in real conditions, we will adopt the dual approach both top-down and bottom-up. The top-down approach will ensure that we consider the application (such as throughput, delay, energy consumption, etc) and environmental limitations (such as deployment constraints, etc). The bottom-up approach will ensure that we take account of the physical and hardware characteristics such as memory, CPU, energy capacities but also physical interferences and obstacles. With this integrated perpective, we will be in capacity to design cross-layer integrated protocols well adapted [39]. We will design jointly routing and MAC layers by taking dynamics occurring at the physical layer into account with a constant concern for energy and security. We will investigate new adaptive frequency hopping techniques combined with routing protocols [41], [50], [24]. Also, we will work on new scheduling techniques for TSCH (a MAC layer of IEEE 802.15.4e) that are able work under the above-mentionned assumptions and bring the robustness of TSCH to IoT scenarios. We will investigate the performance boundaries of TSCH in particular in terms of energy-efficiency of time synchronization [63], and will propose alternative approaches such as capture effect-based time synchronization in TSCH or opportunistic routing. Another technology we will consider is IEEE 802.15.4g, which provides communication ranges in the order of tens of kilometers. We will propose mechanisms to support scaling up to networks with a density of hundreds of nodes, at the MAC layer and above. We will also consider dual-technology networks where both long and short-range communication cooperate for increased robustness.

This vision will also allow us to integrate external factors by design in our protocols, in an opportunistic way. Yet, we will leverage on the occurrence of any of these phenomena rather than perceiving them as obstacles or limitations. As an example, we will rely on node undergone mobility to enhance routing performance as we have started to investigate in [74], [59]. On the same idea, when specific features are available like controlled mobility, we will exploit it to improve connectivity or coverage quality like in [46] [67].