Energy efficient wireless sensor network using neural network based smart sampling and reliable routing protocol

The potential growth and successful convergence of different technologies like Wireless communications, micro electro Mechanical systems (MEMS) and Digital electronics have leveraged to realise the concept of sensor networks with the collaborative efforts of large number of low powered, low cost and small sized sensor nodes used for short distance, multihop communications. Efficient utilization of node's battery energy to improve the network lifetime is the fundamental challenge in Wireless Sensor Network (WSN) [5]. The periodic control signal overheads exchanged between the base station & the sensor nodes add substantial amount of burden in energy expenditure. The other major energy issues lies in packet losses and the subsequent retransmissions, while minimizing the control signal overhead charges. To mitigate this problem, we propose the combined design of non uniform sampling and reliable routing for efficient energy saving and fast data communication in the system while minimizing control signal overhead charges. Our work enhances the reliability in packet transmission by predicting energy robust and near-by nodes in the data forwarding path towards the destination, using back propagation neural network algorithm. We have demonstrated through simulation results the improvement in battery energy consumption without trading off with the speed of data communication which is achieved at the cost of minimal overhead charges.