Data Aggregation Research Based On Sparse Compressive Sensing In Wireless Sensor Networks

With the development of big data and artificial intelligence,the Internet of Things has become the focus of the researchers again.Internet of Things not only has brought great convenience to our ordinary life,but also it has made great achievements in health care,environmental monitoring,military and industrial fields.Wireless sensor networks process and analyze these information by collecting a large number of randomly distributed sensor nodes,and then pass valid information to the users.However,due to the limited power and limited storage space of the nodes,it is meaningful that researchers study how to integrate these large amounts of data,reduce the energy consumption and extend the life of the network in the transmission process.In addition,for some real-time demanding scenes,how to collect data quickly is another hot topic in WSNs.In WSNs,the traditional ways of data fusion aim at reducing redundant data,but cannot reduce the data packets in transmission and communication consumption to large extent.Therefore,this thesis adopted the spatio-temporal correlation of data and the characteristics of the sensor networks.By designing the measurement matrix,this thesis proposed a data aggregation algorithm based on sparse compressive sensing.Our algorithm can reduce the throughout and comminication consumption in WSNs.The main content of this paper is as follows:Firstly,comparing with traditional collection methods based on compressive sensing,we design a measurement matrix based on deterministic binary matrix.The construction process is simple and fast.In this thesis,each row of the measurement matrix represents a measurement process.Each measurement process is independent.Owing to the sparsity characteristic of the matrix,the nodes whose corresponding elements in the matrix are non-zero take part in each measurement.The readings of the nodes participating in the same measurement are merged into one packet to the aggregation node.After the sink node collects all the measurement values,the raw data can be accurately reconstructed.Secondly,aiming at the problems of time delay and energy imbalance in sensor networks,we have proposed an aggregation algorithm based on sparse random measurement matrix.The algorithm decomposes the measurement process into multiple aggregation trees.A single fusion tree is composed of a few nodes whose elements are non-zero in the corresponding matrix.In the process of transmission,a new low-delay router algorithm is proposed.In addition,due to randomness of the matrix,consumption of networks can achieve a balance.At the same time,the lifetime of networks is prolonged.Finally,this thesis systematically analyzes the proposed algorithm and simulates the theory on the same platform.For sparse signals in the frequency domain,the results show we can effectively reduce the energy consumption by designing sparse measurement matrix.Our algorithm can reduce the communication consumption,balance the network's consumption,and shorten the delay.