Signal Detection And Signal Classification Based On Compressive Sensing

Compressive sensing(CS)theory can recover the original signal based on a small amount of sampled,which unifies the sampling process and the compression process,and simultaneously compresses the signal during sampling.Compared with the traditional Nyquist sampling method,the compressive sensing method greatly reduces the cost of signal processing,which greatly alleviate the hardware pressure.First of all,this paper gives an overview of the basic theory of compressive sensing,describes the source,advantages and processing steps of the compressed sensing theory,and describes the three core issues of compressed sensing theory in detail,including sparse representation of signals,measurement matrix design principle and signal reconstruction algorithm.Among them,the core issue of compressive sensing,namely,the design of reconstruction algorithm is deeply studied,and the design of reconstruction algorithm under disturbances and special sparse structure is introduced.Secondly,in the aspect of compressed signal detection,based on CS,this paper makes a decision on the existence of signal based on mathematical statistics and signal processing without reconstructing the signal,that is,the binary detection of the signal.In the field of binary detection of compressed signals,the current research only considers the signal detection under the influence of white noise,and does not consider the compression detection when the signal is subjected to sparse noise.In this paper,the sparse noise is divided into deterministic sparse noise and random sparse noise,and binary signals of compressed signals affected by different types of sparse noise are detected.The results of the compressed signal detection algorithm are deduced under the influence of different types of sparse noise.In addition,this paper simulates the proposed algorithm and compares the simulation results with the theoretical results to verify the correctness of the algorithm.Finally,in the aspect of compressive signal classification,this paper classifies and identifies the human behavior based on the compressive sensing and sparse representation theory.At present,no consideration is given to the existence of disturbances in compressive sensing in behavioral classification.In this paper,we propose a new compression classification algorithm to improve the anti-disturbance ability of the classifier when there is perturbation in the training data of the sensor.In addition,we simulate the compression algorithm in the presence of disturbance.The simulation results show that the proposed algorithm has higher classification accuracy than the traditional algorithm.