TO PROPOSE A TECHNIQUE FOR FAULT TOLERANCE IN WIRELESS SENSOR NETWORKS

The wireless sensor network is a sort of Ad hoc network which is self configuring which means, any sensor node can join or leave the network when they want. The wireless sensor nodes are responsible for data routing in the network due to the unavailability of central controller in them. Wireless sensor network is used to monitor the environmental conditions like temperature, pressure etc. and thus, is deployed in the far spots like forests, deserts and so on. Wireless Sensor nodes are very small in size and have constrained resources. In such far spots it is exceptionally hard to energize or supplant the battery of the sensor nodes. In such conditions, we concentrate on lessening the battery consumption of the sensor nodes. In this work, a new technique is proposed to reduce battery consumption which is based on the dynamic clustering using neural network. Before data transmission sensor nodes form the cluster dynamically using Boltzmann learning of the neural network and weights are adjust according to the conditions and it also enhance the efficiency of the dynamic clustering. Experimental results show that new proposed technique is more efficient, reliable and provides more throughputs as compared to the existing technique.