Research

The Internet of Things (IoT) offers opportunities for connected objects to combine artificial intelligence (AI) and signal processing in low-cost on-board devices. Being part of a constantly growing sensor networks with low data rates, the IoT is generally associated with a simple digital device that takes advantage of the AI resources offered by service providers (Cloud). By bringing AI to the edge of the sensor (IoT-edge), it then becomes possible to give the IoT more flexibility in energy management and decision-making. One natural solution for energy autonomy is energy harvesting, another is ultra low consumption, using transistors in weak inversion. Two research projects are undergoing:

1. Smart IoT – One way to act on the autonomy of connected objects is by optimizing consumption management by an IoT clock radio (WuR). This WuR could also promote intelligent communication by taking advantage of information processing solutions such as sparce approximations (compressive sensing) and neuromorphic solutions (Spiking Neural Network) for decision making. This intelligent radio (Smart IoT) would surpass Mitola's software radio by bringing energy and decision autonomy through the hardware implementation of a neuromorphic radio.
   Details are avaliable in this <seminar> 
2. Reliable IoT – IoT solutions, often associated with large public applications, are diversifying into electronics known as beyond Moore's Law (More-Than-Moore). They become relevant for much more specific applications operating, in particular, in harsh environments. These applications may relate to intelligent vehicles and their high temperature constraints, nanosatellite constellations and their radiation constraints, among others. These harsh environments naturally condition the reliability, robustness and operational safety of the IoT, which remain real scientific and technological obstacles. My research actions plan to respond to Smart IoT challenges while diversifying IoT solutions towards applications in harsh environments.