The Research Of Self-adaptive Reconstruction Algorithm About Compressed Sensing Based On Wireless Sensor Network

As one of the core technologys of Internet of things, Wireless sensor network attracts more and more attention. How to overcome the limitations of the hardware resources and realize large-scale data acquisition and transmission becomes a bottleneck in the rapid development and wide application of WSN. Compressed Sensing is a new way to acquisit and process data by eliminating redundant of the original information and reducing the amount of transmitted signal with non-adaptive measurement, which breaks the tradional data processing technology.This paper is dedicated to applying CS in WSN and based on the in-depth study of the both theories especially focused on exploring the appropriate signal reconstruction algorithms which match the characteristics for WSN.On the basis of theoretical analysis and experimental simulation comparison on the the matching pursuit Series algorithm and threshold iterative algorithm, we propose a sparsity adaptive compressive sampling matching pursuit algorithm, which adopts compressive sampling, has improvements in Support Set atoms picking method and iterative terminal condition and overcomes the shortcoming that traditional Matching Pursuit can’t realize accurate reconstruction under the situation of signal with noise, achieving the sparsity self-adaptive reconstruction signal with noise. It not only improvs the accuracy and efficiency, and has a good noise robustness, but also realizes the sparsity adaptive characteristic eventually in reconstruction process. MATLAB simulation results show that, this algorithm has great advantages in accuracy, efficiency and stability, furthermore, its robustness to noise makes it have strong application potential, which could promote the step of applying compressive sensing theory to practice.Based on CSAMP, combine the iterative threshold thought with CSAMP, we propose a novel iterative thresholding compressive sampling matching pursuit algorithm focused on the charateristics of WSN data. It not only improves the reconstruction accuracy and efficiency with good noise robustness, but also has strong stability when measurement value is small, which saves the WSN hardware resources and is more suitable for WSN application environment. Analysis the characteristics of the algorithm intuitively based on MATLAB simulation experiments, and compared with other algorithms, it realizes the high precision, high efficiency and strong stability original signal reconstruction. It promotes compressive sensing to be applied in WSN and will also have great significance in the large-scale application of WSN.