The Optimum Scheme Of Wireless Sensor Network With Low Power Consumption Based On Program Analysis

Wireless Sensor Network is usually applied to hostile environment,and there’s no way to arrange people to maintain the sensors. The sensors’power is supplied with limited batteries while its lifetime should keepseveral months as long as possible. So it’s a great challenge for designersto save energy without degradation of performance.This research has two innovative points compared with others in thisfield:(1) It combines program analysis and wireless sensor network, usingthe analysis tool to collect the profile of the whole network program inorder to improve the energy utilization on the level of software.(2) It combines dynamic and static program analysis. At first, weshould add some parameters to the makefile to obtain the programstatistics; then we focus on the key programs and use the programoptimizer corresponding to structure analysis to optimize the code fordecreasing the network consumption.The optimized scheme is started with building a wireless sensornetwork. We write the NesC program and burn it to all of the Telos nodesto implement the communication during the network. Secondly, it usesdynamic program analysis method to get the profile while the program isrunning and focus on the functions with high accumulated running timewith structure analysis. Thirdly, we change the scheduling mechanism forthe functions with long running time while we use program optimizer todivide the functions into some blocks and reorganize them for the functions with high use frequency.After we deploy the whole scheme, we use NS-2emulator to test andmanifest it in true sensor network. By taking experiments in some kinds ofwireless sensor network with different protocols, the result suggests thatthis scheme is able to decrease8%-15%of the energy consumption. Sinceit can work well with most network protocols, we can see its university andfeasibility in prolonging the lifetime of the sensors with the samecommunication performance.