Description
Date depot: 8 février 2023
Titre: Augmenting Low-power Wireless Network Management through Embedded AI
Directeur de thèse:
Thomas WATTEYNE (Inria-Paris (ED-130))
Domaine scientifique: Sciences et technologies de l'information et de la communication
Thématique CNRS : Systèmes et réseaux
Resumé: In a Time Synchronized Channel Hopping (TSCH) network, all communication is orchestrated: a communication schedule tells each node what to do in each of the timeslots, including whether to transmit, listen or sleep, with which neighbor node to communicate, and on which frequency. How to build that schedule is the critical question as it determines the throughput, battery lifetime and reliability of the network. Several approaches have been taken, including purely centralized and distributed approaches, using algorithms crafted by networking experts. In the last years, micro-controllers with embedded Artificial Intelligence (AI) capabilities have become commercially available. These can execute a model (e.g. a Convolutional Neural Network, CNN) directly at the device, at low power. This redefines what capabilities low-power wireless devices have, as today AI algorithms are typically purely run in the cloud.
Doctorant.e: Balbi Antunes Martina Maria