Forest fire monitoring based on mixed wireless sensor networks

Abstract

Forest fires can be fatal threats. Motivated by the need to detect fire early and locate forest fires clearly, in this paper, we propose a paradigm called the forest fire monitoring paradigm (FFMP). The purpose of the FFMP is for forest fire early detection and locating based on mixed wireless sensor networks (WSNs). Different from pure static and mobile WSNs, mixed WSNs are composed of both mobile sensor nodes and static sensor nodes. Mixed WSNs are a tradeoff between cost and coverage. In the FFMP, the mobile sensor nodes perform as cluster heads and they will construct a backbone network in which the mobile sensor nodes can connect with their neighbors and be capable of transmitting data to the base station. Each static sensor node chooses one neighboring mobile sensor node as its cluster head and uploads the generated messages to the cluster head. The mobile sensor nodes then fuses the information and transfers the fusion results to the base station, where the data were further processed to obtain the temperature distribution graph and locate the fires. The simulation illustrates that our approach performs well in early forest fire detection and locating. In addition, our approach can significantly prolong the lifetime of WSNs.