双基阵纯方位目标跟踪无偏最小二乘估计算法

Unbiased Least Square Estimation for Bearings-Only Target Tracking Algorithm Using Bistatic Sonar

纯方位目标跟踪是非线性估计问题,伪线性估计算法是主要的纯方位目标跟踪算法之一,其计算简单,但对测量方程进行线性化处理后其运动要素的估计会出现发散或偏置现象。为了克服伪线性估计算法的估计有偏性,本文提出一种带约束条件的最小二乘估计算法,使得伪线性化后的均方等效误差最小来实现目标运动要素的无偏估计。通过仿真说明了利用该方法所得到的位置、速度跟踪误差曲线能很好快速地逼近CRLB,比双基阵EKF滤波器及伪线性估计算法有着更好的收敛速度和跟踪精度,证明了该方法的有效性。

Bearings-only target tracking is a nonlinear estimation problem because of the non- linearity of its measure equation. The pseudo-linear estimation （PLS） algorithm is one of the primary bearings-only target tracking algorithms. Although the algorithm can be simply imple- mented, it results in biased or divergent solution according to the pseudo linear equation de- rived from the measure equation. This paper presents an unbiased least square estimation algo- rithm with restrictive conditions. The algorithm realize the unbiased estimation by minimizing the expectation of equivalence MSE. The position and velocity root mean square errors are ana- lyzed. Simulation results show that the algorithm is faster and more accurate compared with extended Kalman filter （EKF） and PLS algorithms. It achieves the Cramer-Rao low bound （CRLB） for Gaussian noise around small error region. Simulations validate the effectiveness of the method.