基于模糊控制交互式多模型粒子滤波的静电机动目标跟踪

针对交互式多模型粒子滤波算法（IMMPF）的精度不高，算法更新时间长，难以满足静电机动目标跟踪要求的问题，提出了一种新的基于模糊控制的交互式多模型粒子滤波算法（FIMMPF）。该算法先利用模糊控制方法实现实时调整交互式多模型算法中的转换概率矩阵，使与目标当前运动状态最接近的运动模型在混合产生这一采样时刻的初始状态向量里占有更大的比重。同时，为了提高基本粒子滤波算法的精度，减小算法更新时间，再利用中心差分扩展卡尔曼滤波算法产生基本粒子滤波的建议分布函数，实现对目标运动状态的更新。理论分析和仿真结果表明，所提出的算法能够以更高的定位精度，更小的计算量实现对静电机动目标的跟踪。

Electrostatic Maneuvering Target Tracking Based on Fuzzy Interacting Multiple Model Particle Filter

Interacting multiple model particle filter (IMMPF) has the defects of low precision and long updated time, which is not able to meet the requirement of the electrostatic maneuvering target tracking. To solve the problem, a new fuzzy interacting multiple model particle filter (FIMMPF) is proposed. This algorithm can adjust the transition probability matrix of IMM by fuzzy control in real-time and make the motion model that is closet to the current actual motion model occupy larger proportion in the initial state vector. In order to improve the precision of PF algorithm and reduce the updated time of algorithm, central difference expand Kalman filter algorithm is used to generate a proposed distribution and update the target state. The theory analysis and simulation results show that FIMMPF can track the electrostatic maneuvering target with higher positioning precision and smaller updated time.