Abstract: The base is moved and the detect time is varied when the acoustic homing weapon detecting and tracking target. It causes the Kalman filter hard in application to track target immediately. The parameters of location and attitude provided by the acoustic homing weapon navigation system and the target information detected by the acoustic homing weapon are used to resolve the problem. The target coordinates in acoustic homing weapon coordinate systems are transformed to terrestrial coordinate systems by coordinate conversion to resolve the problem of base moving. Every detecting time is used to calculate the sampling time to resolve problem of detecting time-varying. The model of acoustic homing weapon target tracking based on adaptive memory attenuate Kalman filter is proposed. The application method of choosing filter initialization values is presented. Simulation results show that the model is accurate. The tracking algorithm can achieve target tracking in ten times detection and provides better effect on convergence and mobile target tracking.

Key words: adaptive memory attenuate, Kalman filter, target tracking, base moving, detecting time-varying

摘要： 水下声自导武器采用主动方式检测、跟踪目标时，目标跟踪过程表现为一个动基座时变过程，难以直接运用卡尔曼滤波进行目标跟踪。利用水下声自导武器导航定位及航行姿态参数和水下声自导武器检测到的目标信息，通过坐标变换将目标坐标从水下声自导武器坐标系变换到大地坐标系解决了观测基座运动的问题，通过每次观测时间实时计算每次观测的采样时间解决了观测时变问题，建立了基于自适应衰减记忆卡尔曼滤波的水下声自导武器目标跟踪模型，给出了滤波初值选取的工程方法，仿真实验证明该模型正确，跟踪算法能够在10次观测内实现对目标的跟踪，具有较强的机动目标跟踪能力，算法收敛速度较快。

关键词: 自适应衰减记忆, 卡尔曼滤波, 目标跟踪, 基座运动, 观测时变