Communications protocols for wireless sensor networks in perturbed environment

Overview & Synopsis

This thesis is mainly in the Smart Grid domain. SGs improve the safety of electrical networks and allow a more adapted use of electricity storage, available in a limited way. SGs also increase overall energy efficiency by reducing peak consumption. The use of this technology is the most appropriate solution because it allows more efficient energy management. In this context, manufacturers such as Hydro-Quebec deploy sensor networks in the nerve centers to control major equipment. To reduce deployment costs and cabling complexity, the option of a wireless sensor network seems the most obvious solution. However, deploying a sensor network requires in-depth knowledge of the environment. High voltages substations are strategic points in the power grid and generate impulse noise that can degrade the performance of wireless communications. The works in this thesis are focused on the development of high performance communication protocols for the profoundly disturbed environments. For this purpose, we have proposed an approach based on the concatenation of rank metric and convolutional coding with orthogonal frequency division multiplexing. This technique is very efficient in reducing the bursty nature of impulsive noise while having a quite low level of complexity. Another solution based on a multi-antenna system is also designed. We have proposed a cooperative closed-loop coded MIMO system based on rank metric code and max-dmin precoder. The second technique is also an optimal solution for both improving the reliability of the system and energy saving in wireless sensor networks.