Projet de recherche doctoral numero :8301

Description

Date depot: 6 avril 2022
Titre: STUDY OF MANET SOLUTIONS FOR A RADIO COMMUNICATION SYSTEM BASED ON ARTIFICIAL INTELLIGENCE ALGORITHMS
Directeur de thèse: Paul MUHLETHALER (Inria-Paris (ED-130))
Encadrant : Cedric ADJIH (INRIA-Saclay)
Domaine scientifique: Sciences et technologies de l'information et de la communication
Thématique CNRS : Intelligence artificielle

Resumé: Context: The upraising of fifth generation fighters and unmanned combat aircrafts, designed to reach a high level of stealth and to follow a highly collaborative concept of operation, presents new challenges for military radio communication systems to meet expectations in terms of low probability of detection and interception while offering high throughput and low latency networking capabilities. To face this challenge, new communications solutions have been envisioned to operate in high spectrum bands (>10 GHz) where large bandwidth can be used despite constraints on spectrum allocated to military purposes. Such bands, particularly susceptible to signal degradation, impose the use of directive antenna to meet range and throughput requirements. The current challenge is to design an efficient ad-hoc networking system based on individual point-to-point communication links. During the last few years, military aeronautical communications have already evolved to support not only voice and tactical data, but also a growing part of machine-to-machine communications (e.g. sensor collaboration), making use of MANET principles at the routing protocol level. Applying such principles in the context of directive antenna has to be performed not only at the routing level, but also at the network configuration level, where each point-to-point link has to be chosen, setup, maintained or disrupted to continuously adapt the network to fit with propagation, topology and technological constraints as well as operational connectivity needs. Objectives: A project of a new ad-hoc radio communication system based on directional antenna is under study. As explained above, new algorithms are required to perform link selection and transmission resource allocation on each node of the network. We believe as a promising direction for a PhD thesis to carry an original research work pertaining to the design and performance evaluation of the above algorithms. Evaluation of artificial intelligence techniques and in particular machine learning algorithms is encouraged. Conducted in parallel with the definition of the communication system, this research work will contribute to propose and investigate different solutions, evaluate performances, and confirm soundness of other design choices such as media access method, type of duplexing, antennas technology, spectrum usage... The work is expected to mix analytical techniques and proof of concept simulations to confirm the performance under realistic scenarios and prototyping of interesting solutions for demonstration.

Doctorant.e: Marcoccia Félix