Detection of forest fire in wireless sensor network using evidence combination

In this paper, we propose an intelligent and robust system for forest fire detection using wireless sensor network. The intelligent system consists of three different neural network classifiers to overcome the problem of training and testing under conditions of insufficient and noisy data. The aim of using three classifiers in the intelligent system is that classification in this forest fire context is extremely important as the cost of misclassification using a single classifier is very high. Hence, a combination of their beliefs by Dempster– Shafer evidence combination which provides a representation of epistemic plausibility overcomes weaknesses exhibited by anyone classifier to a particular data set and helps to detect the forest fire accurately. The combination approach provides a higher accuracy in detecting and forecasting forest fire more promptly. The experimental results show the combination approach yield better accuracy in predicting the forest fire.