不完全量测下长基线系统的水下目标跟踪算法

针对不完全量测情况下长基线系统对水下目标跟踪精度会下降的问题,提出了最小二乘-容积卡尔曼滤波(Least Squares-Cubature Kalman Filter,LS-CKF)算法。选取容积卡尔曼滤波(Cubature Kalman Filter,CKF)为基本跟踪算法并将其改进为两步滤波模式.增加的第1步滤波使用最小二乘估计优化时间更新阶段的容积点,提高了第2步滤波中量测更新的精度。进一步推导了量测信息为距离时新算法的简化形式,降低了运算复杂度,使其能更好地应用于水下跟踪系统.仿真实验和湖试数据的处理结果表明,在丢失量测数据较多且初始状态误差很大的恶劣情况下,LS-CKF收敛速度比标准CKF算法提升了1倍,且跟踪误差降低10%以上。 

Target tracking algorithm for underwater ranges-only long baseline system with incomplete measurements

A Least Squares-Cubature Kalman Filter(LS-CKF) algorithm is proposed,aiming at increasing the tracking accuracy of underwater maneuvering target with incomplete measurements of the long baseline system.Firstly,a practical two-dimensional mathematical model of underwater target tracking with incomplete measurements is established.Secondly,a two-layer CKF is introduced where the expression of the cubature point under the latest measurement constraint is derived via least square estimation in the first layer to improve the accuracy of the measurement updating process.Finally,considering the underwater tracking system is power-limited,a simplified form of the new algorithm with ranges-only measurements is derived.The accuracy of the proposed algorithm is improved because the latest measurement is effectively integrated into each cubature point with the Least-Squares method,which allows the cubature points to move from the a prior area to the high likelihood region.The results of the simulation and the lake trial both show that the new algorithm significantly improves the tracking accuracy on condition that severe incomplete measurements and large initial error.Compared with the standard CKF algorithm,the convergence speed of LS-CKF is double while the tracking error is reduced by more than 10%.