水下声纳信号目标定位优化探测研究

Optimal Detection of Underwater Sonar Signals

在复杂的水声条件下,海底和海面混响等噪声因素严重降低了传统声纳的探测性能。为了提高强噪声背景下声纳的实际检测能力,综合采用盲信号恢复和盲源分离方法。首先利用特征向量(EVA)算法得到的信道均衡器进行多信号源恢复,克服声纳信道的影响,去除非加性噪声。再根据峰度自然对数最大化算法按峰度减少的顺序对信号源逐个进行分离。利用Matlab进行的计算机仿真结果表明改进算法能够在强噪声(SNR<0dB)背景下对多个统计独立且四阶累计量不为零的窄带非高斯信号源进行分离。仿真结果验证了算法的有效性、强稳健性和快速收敛性,为水下目标定位优化提供了参考。

As the ocean environment becomes more and more complex, the detection performance of traditional sonar is severely reduced because of the noise such as bottom and surface reverberation. In order to improve the actual detection performance of sonar in high noise environment, blind source restoration and blind source separation are used. Firstly, a channel equalizer is derived based on EVA algorithm for signal restoration. It can overcome the influ- ence of the acoustic channel and eliminate the non-additive noise. Then these sources can be separated one by one through Logarithm-Kurtosis maximization. The computer simulation based on Matlab shows that the improved algorithm can separate narrowband non-gaussian signal sources, which are statistically independent and the non-zero fourth-order accumulation under strong noise background （SNR 〈 0 dB）. Simulation Results verify its effectiveness, great robust and high convergence rate.