The Application Of Compressed Sensing In Wireless Sensor Network

This paper mainly study the distributed compressed sensing reconstruction algorithm. The purpose is to make better use of the temporal and spatial similarity for joint reconstruction in the wireless sensor network(WSN) whose structure is general irregular and large.First of all, we use the node location information with irregular network topology to model large-scale wireless sensor network into an undirected weighted geometric graph. Then assign sensor nodes numbers or colors to determine the communication order of the nodes. Considering the disadvantage of without utilizing the similarity between adjacent nodes in a iteration soft threshold algorithm, add the similarity between adjacent nodes of the original signal to the original signal reconstruction standards, then reconstruct the minimal l1 norm function. A distributed iterative soft threshold algorithm based on color is proposed, and the experimental results show that the effectiveness of the algorithm.Second, 1-bit compressed sensing is used to WSN. Each sensor node get the original acquisition signal projection to the observation matrix and quantify the observed value through 1-bit quantizer, only send the symbol of the observed value to the base station lastly. Using the similarity between the adjacent sensor nodes, a distributed binary iteration hard threshold algorithm is put forward to improve the hard threshold binary iteration algorithm, and the experimental results show that the effectiveness of the algorithm.Finally, a distributed clustering method is used for WSN and sensors are organized into some clusters. To improve the disadvantage of the iteration soft threshold algorithm which solves the question of reconstruction of each cluster data in the wireless sensor network, the similarity of the original data of sensors within the same cluster is used to reconstruct the question of the least l1 norm. Then a distributed iterative soft threshold based on clustering algorithm is proposed to solve the problem of the joint reconstruction the sensors within the same cluster. In the experiment section we use the algorithm to reconstruct the temperature signals of multiple clusters jointly, and the experimental results show that the algorithm get good results.