单矢量水听器多目标方位估计的算法研究

矢量水听器可有效抑制各向同性的背景噪声干扰,能同时共点地拾取声场的声压标量和振速矢量信息。这些信息的偶次阶矩组成了以各个目标的声强和方位为未知数的非线性方程组。为了求解该方程组,研究了具有全局最优化概率搜索的快速方法:遗传算法和小生境遗传算法,并对两者的性能进行比较。仿真试验的结果证明了算法的可行性,即基于小生境遗传算法能够实现单个矢量水听器进行多个非相干源目标的方位估计。为了提高弱目标的方位估计精度,又提出了二次小生境搜索方法。

Direction estimation of multi-sources received by a single vector sensor

The vector sensor may effectively depress the isotropic noise in the sea; it can measure sound pressure and orthogonal components of particle velocity co-locately in space and simultaneously in time. Nonlinear systems of even order moments of wide band receiving signals include direction azimuth and intensity of every target. Two adaptive heuristic probabilistic searching algorithms: genetic algorithm and niche genetic algorithm are researched. Performance between them is compared. The simulation results show that single vector sensor can estimate directions of multi-incoherent sources. Second search is made for increasing estimated accuracy of weak target using niche genetic algorithm.