Dynamic Voltage Scaling Based On Wavelet Decomposition In Wireless Sensor Network

Wireless Sensor Network is a Mult-leap network system, that a good many sensor nodes which are subminiature, sixpenny and low-consumption are disposed by wireless communications in the observation circumstance or around. Generally, sensor nodes are provided electric energy by batteries, and replacing battery is very difficult, In order to extend Me of wireless sensor network by making economical use of energy, An effective method of dynamic voltage scaling which is based on wavelet, aimed at the node of wireless sensor network in running state, is put forward to resolve it.Based on wavelet, Kalman and autoregressive model, a method for DVS is presented in the first time, compartmentalizing frequency-voltage states based on the dispersion of processor frequency, then the workloads in the next period are forecasted, based on wavelet, Kalman and AutoRegressive model, cumulating probability can be calculated by forecast-result, and the frequency-voltage state that sensor should enter is decided according the calculating result. The more accurate the forecasting result is, the better frequency-voltage state the processor enters, the better the effect of DVS is, and the lower energy the wireless sensor use.The data firstly is decomposed and reconstructed in each scale (frequency). Then, high frequency coefficient is predicted by Kalman filter and low frequency coefficient by AR module. The original forecast-result is received by colligated each frequency forecast-result. According the forecasting result, wireless sensor may turn into some low frequency-voltage state. the smaller task, the smaller operating frequency using, thereby reducing the energy consumption of nodes in one period.Experimental results show that the system energy consumption of sensor, by using this method. It is achieved to maximize the sensor node's lifetime.