Research On Wireless Sensor Network's Data Compression Algorithm Based On Compressed Sensing Algorithm

In recent years,wireless sensor network technology has become the focus between scholars at home and abroad.However,large-scale wireless sensor networks exists a large number of technical shortcomings.Such as network energy consumption and data that collected usually has strong spatial and temporal redundancy.On the one hand,traditional network data compression technology lacks definite compression method.On the other hand,it needs a large mount of calculations or communication resources while the wireless sensor networks with limited resources.Nonetheless,the compression perception theory that proposed in recent years not only has good data compression performance,but also makes encoding very simple which conforms to the characteristics of the wireless sensor network.Thus helps to improve the network life cycle and reduces the network energy consumption.This thesis based on compressed sensing's large-scale wireless sensor network data collection and studied deeply on it.The joint sparse model's distributed compressed awareness framework is used in this thesis which aiming at the research of routing algorithm and compression perception reconstruction algorithm.The main research content is as follows:First of all,on the basis of in-depth study of LEACH routing algorithm proposed its shortcomings.According to LEACH routing algorithm,this thesis puts forward an improvement multiple hops routing algorithm based on the node residual energy and sparse choosing cluster heads.The thought of this algorithm is become clusters first and then choose cluster heads.According to the occupancy of clusters to identify the amount of cluster heads and then on the basis of cluster heads' number to divide wireless sensor network monitoring area on average.Afterwards,by residual energy and the sparse degree of node in specific area to choose cluster heads.Compared with LEACH routing algorithm,the algorithm used in this thesis is helpful to balance the network energy and prolong the network life cycle.Some algorithms must be known sparse data,which thereby determine the reconstruction iterations of the algorithm.This study found that if sparse value estimation is inaccuracy,the precision of reconstruction accuracy will decrease.Meanwhile,when estimate is large,the resource of operation is waste.This thesis proposed the improvement algorithm based on thecombination of ROMP and SAMP which is an algorithm can reconstruct the signal without knowing the sparse degree.But SAMP runs for a long time.The improvement algorithm brings in regularization thought and set the threshold of regularization's energy division.Compared with the contrast algorithm,improvement algorithm in this thesis can improve the reconstruction accuracy and reduce the time complexity.