改进SA-SP算法的路面不平度信号压缩感知采集研究

为大幅度减少采集路面不平度信号的存储空间，提高采集速度，基于压缩感知理论针对标准路面的不平度信号进行压缩采样和重构。首先验证了B级路面不定度信号在频域下的近似稀疏性，并进行了信号的压缩采样。针对现阶段凸优化方法和常用的三种贪婪算法的不足，提出一种改进的模拟退火算法与子空间追踪算法相结合的稀疏度自适应匹配追踪算法，利用改进的模拟退火算法快速搜索匹配最优的稀疏度，并采用子空间追踪算法快速重构信号。仿真实验对比五种重构方法，结果表明，凸优化方法精度较高，耗时过长；OMP算法和SP算法耗时极短，但需要预先进行实验来估测信号的稀疏度，实用性低；SAMP算法能实现稀疏度的自适应匹配，但匹配的误差较大，且耗时较长；提的新方法具有良好的精度和较快的执行速度，R-squares和耗时的均值分别为0.9837和2.77 s，稀疏度估测效果较好，且采样点数的增加不影响算法重构信号的速度。

Research on compression sensing acquisition of road roughness based on improved SA-SP algorithm

In order to reduce the storage space and improve the acquisition speed of the road roughness signal, this paper carried out the compressive sampling and reconstruction for the roughness signals of standard roads based on the compressed sensing theory. Firstly, it verified B grade road roughness signals in the frequency domain with approximate sparsity, and carried out the signals compression sampling. For shortcomings of convex optimization method and three kinds of greedy algorithms at this stage, an algorithm which combined improved simulated annealing algorithm and subspace tracking algorithm could make sparse adaptive matching and complete signal reconstruction. This algorithm used the improved simulated annealing algorithm to get the optimal sparse degree and used subspace tracking algorithm to reconstruct signals fast. The simulation experiments compared 5 kinds of reconstruction methods. The results show that, convex optimization method has higher accuracy, but it takes longer time. OMP and SP algorithm take shorter time, but they need to do pre-experiment in order to estimate the signal''s sparse degree and have lower practicality. SAMP algorithm can achieve the adaptive matching of sparse degree, but the matching error is larger, and it takes longer time. The new proposed method has good accuracy and fast execution speed, R-squares and time consuming means are 0.9837, 2.77 s, sparse estimation is better, and the increase of sampling points does not affect the speed of the algorithm.