多模型粒子滤波在机动目标跟踪中的应用

针对在非线性机动目标跟踪中存在的滤波器易发散,跟踪误差大等问题,在双机协同跟踪的基础上,提出了利用交互式多模型粒子滤波(IMMPF)对空中机动目标进行跟踪的算法。该算法将粒子滤波和交互多模型有效结合,基本解决了非线性机动目标跟踪中存在的问题。通过仿真表明,与扩展卡尔曼滤波(EKF)和交互式模型扩展卡尔曼滤波(IMMEKF)相比,IMMPF能够降低跟踪误差,提高收敛速度,且有很强的鲁棒性。

Application of Multiple Model Particle Filter in Maneuvering Target Tracking

In non linear maneuvering target tracking,filters are liable to diverge or have large tracking errors.To solve this problem,a new Interactive Multiple Model Particle Filter (IMMPF),which can efficiently track the aerial maneuvering target,was proposed based on dual plane cooperated tracking.The algorithm combined the particle filtering with interactive multiple mode effectively,and thus could solve the problem existed in maneuvering target tracking.The computer simulation results showed that:compared with Extended Kalman Filter(EKF) and Interactive Multiple Model Extended Kalman Filter (IMMEKF),IMMPF has a lower tracking error,higher convergence speed and higher robustness.