Research On Data Compression Algorithm Of Wireless Sensor Networks Based On Compressive Sensing

The application of wireless sensor networks provides extremely convenience for the people’s life. Compared with the traditional networks, the wireless sensor networks is very limited in energy. How to ensure the accuracy of data collection and to reduce energy consumption is the primary to be considered of designing a wireless sensor network. As a way to save the energy consumption, data compression can effectively prolong the lifetime of the network and attract more and more attention in recent years.Based on the analysis of data compression algorithm in compressive sensing theory, this thesis gives corresponding improvements on distributed sparse projections algorithm and distributed compressive sensing algorithm with the target of maximum energy saving. The main work is as follows:Firstly, to reduce the redundant information during the encoding of distributed sparse projections algorithm, a centralized compressive sensing algorithm is proposed. In this algorithm, the cluster head gets the sensory data projected to a pseudo-random matrix, and adds the projection of data to the received projection of the data, and then transmitted the summation to the next cluster head until the sink node. The algorithm cuts down the amount of data of transmission within the network, so it can reduce energy of transmission consumption.Secondly, to slove the problem that distributed compressive sensing algorithm can not determine whether the sensory data has a joint sparse structure, this thesis proposes a twice compressive sensing algorithm. As a prerequisite, to ensure the accuracy of data collection, the algorithm gets rid of the maximum redundant information of original data through two times data compression of the nodes within the cluster and the cluster head, which reduces energy of transmission consumption in wireless sensor networks.Finally, an analysis and simulations are done on the two algorithm proposed in this thesis. Simulation results show that:for a signal with sparse structure, the centralized compressive sensing algorithm can effectively reduce energy of transmission consumption in wireless sensor networks. For a signal with joint sparsity structure, the twice compressive sensing algorithm can reduce energy of transmission consumption in wireless sensor networks. Then we can draw the conclusion that the two algorithm proposed in this thesis is effective.