Reliability assessment of photovoltaic wireless sensor networks for forest fire propagation detection
Résumé
The objective of this paper is to assess the reliability of a wireless sensor network (WSN) equipped with mini photovoltaic cells (PV-WSN), simulating forest fire conditions. Our assessment considers the hardware specifications of WSN sensors, photovoltaic (PV) cells, rechargeable batteries, communication protocols, and elements required for efficient fire detection. Our study’s main assumptions are that: the PV-WSN is structured randomly with one sink per cluster and identical sensors used, random losses due to environmental interference to the network and variability of PV-cell orientation are accounted for, the shape of forest fire propagation is uniform and elliptical, the surrounding vegetation is homogeneous, and packets lost during transmission are retransmitted with the use of a dynamic source routing protocol. Physical hardware specifications were included in our reliability assessments for each network type and a realistic battery re-charge/depletion scenario based on sensor states, sun availability, and PV efficiency was considered for a PV-WSN. We developed a simulator to study and to compare how the PV-WSN behaved differently under different forest fire conditions. Our results support the use of a PV-WSN, as network lifetime can be efficiently prolonged in the context of large-scale fire detection with this setup.