Research On Compressed Sensing In Wireless Sensor Networks Based On Internet Of Things

With the continuous development of sensor, RFID tags and embedded systems management technology, Internet of things has become a major landmark of information technology. Owing to its characteristic of self-organizing ability, convenient arrangement and fault-tolerant ability, wireless sensor network, as the Internet of things’ key technology of the underlying network, has become the core of networking information access method. At the same time, the wireless sensor network is facing the bottleneck of technology development because of its poor ability to compute nodes, small storage space and low battery life.Compressed sensing (CS) is a recently proposed front-end signal processing technology, it has broken the Shannon theoretical data recovery for the signal bandwidth restrictions with much less than twice the size of frequency of the signal sampling rate to complete the signal acquisition, which would substantially save the operation, storage space and energy consumption when the wireless sensor network is collecting signal. The application of compressed sensing technology to wireless sensor networks based on Internet of things has gradually become a hot research topic.In this paper, we focus on the problem of multi-signal processing in the Internet of things based on the analysis of traditional compressed sensing methods.Firstly, this paper introduces the architecture of wireless sensor network and other significant knowledge. An improved traditional compressed sensing and distributed compressed sensing method is proposed for properties of the multi-signal processing in the Internet of things.Secondly, a new adaptive compressed sensing(ACS) is proposed in this paper. one-dimensional search method is applied to find the appropriate value to construct an adaptive observation matrix, which is related to the experiment signal. Then the existing stagewise orthogonal matching pursuit method is improved to reconstruct the signal, and the adaptive method is adopted to select the threshold and the number of iterations.Thirdly, for the multi-signal processing ability, this paper presents a distributed compressed sensing based on temporal correlation:first of all, with the temporal correlation between the data of signals, the one-dimensional linear regression method is used in signals’ compressed sensing. Then, the Joint Sparse Model of distributed compressed sensing is improved and a preliminary compressive matrix is designed to extract the linear part of signals’ data. Finally, the adaptive compressed sensing is adopted to compress the signal processed by the compressed matrix and a complete framework of compressed sensing signal processing system comes into being.In the end of this paper, the content and research results of this paper are summarized. The problems of further research in this direction are discussed as well.