Thesis: Reliable Communications for the Industrial Internet of Things

PhD defense: Vasileios KOTSIOU

Title:  Reliable Communications for the Industrial Internet of Things

Team: Networks

Abstract: The Industrial Internet of Things (IIoT) is an emerging concept aiming at re-using the IoT mechanisms to make the production chains more profitable by maximizing flexibility and adaptability in the factories. However, industrial applications require often deterministic communications as well as end-to-end reliability close to 100\%. Unfortunately, wireless communications mean also contention for the medium access. Moreover, a plethora of wireless devices may use the same unlicensed band, generating a large volume of interference. We still have to tackle several challenges to provide high reliability in any condition and to respect strict end-to-end delay constraints in multi-hop topologies. Thus, the main scope of this thesis was to propose the mechanisms to achieve to achieve the previously mentioned goals. In this thesis, we first conducted a series of experiments to characterize the radio channels in an indoor testbed. We demonstrated in particular the existence of specific per-link characteristics, where external interference may be locally high for some radio channels. Thus, we proposed to improve the efficiency of the slow channel hopping technique with blacklisting techniques. The objective is to exclude from the channel hopping sequence the low-quality channels. First, we proposed a distributed blacklisting technique, that adopts a pseudo-random approach to avoid using the worst radio channels. While this approach allows each radio link to decide autonomously the best radio channels to use, some collisions may still arise pseudo-randomly. Therefore, we then proposed a centralized blacklisting scheme, able to adapt the blacklists for each radio link, while still making the full behavior deterministic, by re-arranging the conflicting blacklists. We also extended an hybrid blacklisting scheme that exploits the full radio spectrum, assigning all the channel offsets to increase the network efficiency when handling long blacklists. Finally, we proposed a scheduling function that aims to meet the requirements of IIoT for low latency and high reliability in a network with radio links subjected to external interference.

The jury is composed by Fabrice Theoleyre (CNRS), Georgios Papapadopoulos (IMT Atlantique), Periklis Chatzimisios (International Hellenic University, Thessaloniki), Martin Heusse - Professeur (Grenoble INP - Rapporteur), Nathalie Mitton (Directrice de Recherche, Inria Nord Europe - Rapporteur), Nicolas Montavont (Professeur, IMT Atlantique - Examinateur), Thomas Watteyne (Directeur de Recherche, Inria Paris - Examinateur).

The defense will be held in English, on September 25th at 2pm in "PAM" amphitheater.