FPGA implementation of artificial Neural Network for forest fire detection in wireless Sensor Network

Remote Sensor Network (WSN) screens dynamic environment that progressions quickly after some time and is utilized by outer components. In WSN, the sensor hubs are to screen the natural parameters, for example, carbon monoxide, stickiness, smoke etc. The aim is to identify firestorm in forest and by predicting the firestorm in forest the sensing ability of the sensor node becomes limited which leads to delay in the alert signal or fail to report and it is difficult to deduce the occurrence of fire. In order to overcome the above problem a Feed forward Neural Network (FNN) was proposed which gives the prediction of firestorm when it occurs without any delay. The neural systems have low power with higher accuracy and it decreases the few bogus recognition of firestorm in timberland. This model recognizes the flame fire with higher accuracy and it controls the alarm delay. FNN is composed with a few hubs N, and made an examination with single and numerous concealed layers to anticipate the higher accuracy. The recreation consequence of the proposed framework is checked and is actualized utilizing Virtex-5 and the RTL schematic was planned utilizing Xilinx ISE 14.6.