Advancing early forest fire detection utilizing smart wireless sensor networks

Abstract

Forests are an important part of the ecosystem and play a crucial role in preserving and maintaining the global environment. Forest fires around the world cost billions of dollars and priceless human lives. Earlier detection of forest fires might mitigate their threat. In this paper, a smart forest fire detection system which combines Wireless Sensor network (WSN) and Artificial Neural Network (ANN) technologies (henceforth SWSN) is proposed. A small scale experimental emulation of controlled fire is carried out with a deployed SWSN collecting data for many simulated scenarios, including the crucial fire-about-to-start scenario. The sensed data are used to train different models of ANNs, while measuring scenario detection success rates. Obtained experimental results are very promising and set the model in a competitive placement among peer-related work, which encourages the utilization of a SWSN approach in a range of other civil and military applications.