基于稀疏度自适应算法的压缩感知

Adaptive Algorithm for Sparsity Compressive Sensing

针对目前压缩感知的重构算法需预知信号稀疏度和重构时间较长的问题,提出基于稀疏度自适应算法的压缩感知,该方法基于分段和回溯思想,通过增大步长、合并上次原子来选择最匹配原子,能显著减小计算复杂度,从而减少信号重构时间。以语音信号为处理对象,对SAMP算法进行了仿真比较,仿真结果表明,在未知信号稀疏度的情况下,与基追踪(BP)算法和正交匹配追踪(OMP)算法比较,SAMP算法的重构信号运行时间明显降低,并且在不同的信号压缩比的条件下重构信号性能得以保证,验证了SAMP算法在稀疏度方面的自适应性以及重构效率高等优点。

The current compressive sensing needs signal sparsity in advance and the reconstruct process sustained over long time.To solve the problem,a sparsity adaptive algorithm(SAMP)was put forward in this paper.The method was based on segmentation and retrospective thinking.By increasing step and merging last atom to select the best match,this algorithm could significantly reduce the computational complexity,thereby reduced the time signal reconstruction.The speech signal was processed in comparison with SAMP algorithm simulation.Simulation results showed that,in the case of unknown signals sparsity,the reconstructed signal quality of SAMP algorithm reduced the reconstruction time evidently comparing with the base pursuit(BP)algorithm and the orthogonal matching pursuit(OMP)algorithm.At the same time,the reconstructed signal quality of SAMP had a good performance under different signal compression ratio.