基于声散射统计特性的压缩感知水下目标定位与观测矩阵性能

Compressed sensing underwater target localization and performance of measurement matrix based on acoustic scattering statistical characteristics

近年来,压缩感知理论在水下目标定位领域得到了广泛的应用,压缩感知定位算法将水下目标定位问题转化为稀疏模型重构问题,具有较高的定位准确性,但定位结果受散射环境影响。该文开展水下目标定位的水池试验,对水听器阵列接收的试验数据进行处理,通过定位误差、可定位声源数量和定位信噪比分析定位结果,研究压缩感知定位算法的宽带性能、散射环境对定位结果的影响,并使用Spark系数和声散射统计特性近似分析了观测矩阵性能。水池试验结果表明,压缩感知定位算法在较宽的频率范围内有较好的适用性;随着环境散射增强,水声定位观测矩阵的列相干性减小,稀疏重构性能提升,使得定位结果更加准确;在近似分析水声定位观测矩阵性能时,声散射统计特性与Spark系数的变化趋势相似,并具有直观且计算简单的优点。

In recent years,compressed sensing theory has been widely applicated in the field of underwater target localization.Compressive sensing localization algorithm transforms the underwater target localization problem into sparse model reconstruction problem,which has high positioning accuracy.But localization results are affected by the scattering environment.In this paper,pool experiments of underwater target localization are carried out,and experimental data received by hydrophone array are processed.Localization results are analyzed by localization error,number of localizable sound sources and localization signal-to-noise ratio.On this basis,broadband performance of compressed sensing localization algorithm and influence of scattering environment on localization results are studied.Then,performance of measurement matrix is approximately analyzed using Spark coefficient and acoustic scattering statistical characteristics.Results of pool experiments show that compressed sensing localization algorithm is well applicable in a wide frequency range.With enhancement of the ambient scattering,column coherence of measurement matrix decreases and sparse reconstruction performance improves,which make localization results more accurate.When approximately analyzing performance of hydroacoustic localization measurement matrix,acoustic scattering statistical characteristics are similar to Spark coefficient in variation trend,and have advantages of intuitiveness and simple calculation.