Utilisation de la Récolte Énergétique pour l’Amélioration de la Duréede Vie des Réseaux de Capteurs Sans Fil.

Abstract

The energy harvesting is the process of obtaining energy from the external environment (e.g., solar and artificial light, vibratory motion, heat, electromagnetic radiation, etc.) and transforming this energy into electrical energy, directly used or stored for use in the operation of small devices. The harvested energy is generally very low (in the mJ range) compared to large-scale energy harvesting systems such as solar farms or wind farms. This harvest of ambient energy can be used to power the small wireless autonomous sensors that have small batteries and are deployed in remote or hostile locations for event detection. Therefore, the integration of energy harvesting systems is a promising solution to extend the sensor lifetime and improve its performances. In this thesis, we are interested in this problematic. The use of the energy harvesting technique for wireless sensor networks. More specifically, the objective is to adapt the operation of the network to the recharging/discharging cycles of the sensors. It's about a general way of switching between two states (sleep and wake-up) based on the duty-cycle technique (i.e., a sensor with a lower energy level selects a longer sleep period to harvest as much energy as possible before going to the unloading state (wake-up) where the deposited energy is consumed). Indeed, since the wake-up of a sensor cannot be accurately estimated, the fact that the exact rate of harvested energy by the sensor fluctuates over time, it is very difficult to ensure the routing of the packets. In addition, the uncertainty over the time it takes for a node to harvest enough energy before it can work again means that a sensor's sleep/wake-up planning solutions are proposed in the literature are unusable. By introducing an energy threshold policy and using the remaining energy in the battery of a sensor, the wake-up and sleep periods are regulated to decrease the duty cycle of each sensor in the network while ensuring a balance between energy consumption and the energy harvesting capacity. The proposed energy threshold policy comprises two phases (firstly, the switching phase between three possible states for each sensor to regulate its wake-up period and ensure load balance in the network and the second, is the calculation of the duration of the sleep period to optimize the duty cycle and make it dynamic). The main results obtained through the integration of this policy in the different types of communications used between the sensors in this type of network, are the possibility of minimizing the number of collisions, the maximum number of retransmissions and the contention on the channel in the network. As a result, the use of this policy has allowed us to extend the life of the network and improve its performance. Abstract Consequently, increasing the lifetime of a wireless sensor network is a complicated heavy work, and that takes a lot of time. Although the experience obtained by proposing the implementing the improvement of energy management in energy harvesting wireless sensor networks is clearly beneficial in overcoming some of the shortcomings of several existing works and making this management more fundamental and more practical.