Research On Compressed Sensing Technology Based On Intelligent Optimization Algorithm

Compared with traditional signal acquisition and processing technology,compressed sensing(CS)technology realizes parallel processing of signal sampling and compression at a lower sampling rate,reconstructs the original signal with high precision,and saves the storage and transmission resources of the system.The reconstruction algorithm is the key technology of compressed sensing,and the performance of the algorithm affects the reconstruction effect.Although the most widely used matching pursuit class reconstruction algorithm at present has low complexity and simple structure,it is easy to fall into local optimum and greatly affected by the signal sparsity,resulting in poor reconstruction effect.In this paper,the fireworks algorithm(FWA)with global optimization is introduced into the compressed sensing technology,and an improved fireworks reconstruction algorithm(IFWRA)is proposed to improve the performance of the algorithm.The main research contents are summarized as follows:(1)Based on the reconstruction method of matching pursuit class algorithm,this paper uses the intelligent optimization algorithm to optimize the atomic set containing reconstruction information and designs an intelligent optimization reconstruction algorithm.Among them,the algorithm dimension is determined by the sparsity estimation of the sparsity adaptive matching pursuit(SAMP)algorithm.Aiming at the shortcoming of the SAMP algorithm that the sparsity estimation is easy to be too large,this paper proposes an improved SAMP(DBCSAMP)algorithm.The double threshold variable step-size method and adaptive backtracking mechanism can reduce the estimation error.The candidate set reduction method can eliminate some redundant atoms.The performance of the algorithm is tested by Gaussian sparse signals.The results show that the DBCSAMP algorithm can effectively improve the reconstruction probability of the algorithm,and accurately estimate the sparsity when the signal length is small.(2)Based on the intelligent optimization reconstruction algorithm structure,this paper combines FWA and DBCSAMP pre-algorithm to propose the fireworks reconstruction algorithm(FWRA),which can improve the accuracy of the algorithm.Aiming at the problem that the optimization capability is weakened when the solution space range and dimension increase,this paper proposes IFWRA.The algorithm uses the non-numerical optimization solution method to fix the solution elements,simplifies the explosion intensity to reduce redundant sparks,adapts the explosion amplitude to prevent falling into local optimum,and uses the elite-fitness value selection strategy to save running time.The performance of the algorithm is tested by Gaussian sparse signals.The results show that the reconstruction probability of IFWRA remains unchanged,the reconstruction error of the signal is significantly reduced,and the operational efficiency is higher than that of the same type of reconstruction algorithm.(3)To verify the practicability of IFWRA,this paper uses a shock wave signal measured by a 50 psi range sensor during the firing of a gun as the test signal.The application results show that the IFWRA can successfully reconstruct the shock wave signal,and in the same type of reconstruction algorithm,the reconstruction error of the signal is the smallest and the algorithm operational efficiency is the highest.In this paper,IFWRA is proposed based on the matching pursuit class algorithm and FWA,which optimizes the performance of the reconstruction algorithm and proves its practicability in the field of shock wave testing.