一种鲁棒的战场声目标自适应时延估计算法

A Robust Time Delay Estimation Algorithm for Acoustic Target on Battlefield

提出了一种可用于战场声目标定向的鲁棒自适应时延估计算法。该算法应用最优变步长自适应仿射投影技术和鲁棒M估计理论,无须事先假定信号和噪声的统计特性,自适应调整自身参数,抵御战场环境中的噪声干扰。野外坦克声数据实测表明,在无干扰环境下,算法性能与高阶统计量法、广义互相关法相当,但在武器噪声干扰下表现了更强的鲁棒性和估计精度;在各种条件下,文中算法都比最小均方自适应法表现出更好的性能。

Robust time delay estimation algorithm for acoustic target on battlefield was proposed. The novel algorithm, using the structure of optimal variable step-size affine projection algorithm and robust-M estimators, is free of the assumption that the statistical characteristic of signal and noise is known. Experiments under real environments were conducted to compare the proposed TDE method with the existing higher-order spectra （HOS）, least mean square （LMS） and generalized cross-correlation （GCC）. The result shows that the propose m d TDE method is of higher robustness and precision,