Wavelet Neural Network Approach for Dynamic Power Management in Wireless Sensor Networks

Abstract

Energy is a limited resource in wireless sensor networks. The reduction of energy consumption is crucial to prolong the lifetime of wireless sensor networks. Dynamic power management (DPM), which is to reduce power dissipation by putting the sensor node into different states, should be carefully taken into account in wireless sensor networks. In this paper, a new method of DPM is proposed. In this method, the next eventpsilas time which is a non-stationary series is predicted as accurate as possible by wavelet neural networks. Nodes in deeper sleep states consume lower energy while asleep, but incur a longer delay and higher energy cost to awaken. So the nodes state is decided to move through the predictable time associated with the threshold time and residual power. The simulation results show that the energy consumption is significantly reduced and the whole lifetime of the wireless sensor networks is greatly prolonged with the proposed method.