多普勒辅助测量IMM-PF机动目标跟踪

Maneuvering Target Tracking Using IMM-PF with Doppler-Aided Measurement

针对机动目标跟踪问题,提出了一种辅助多普勒测量的交互多模型粒子滤波算法。该算法利用粒子滤波解决系统观测与目标状态之间的非线性问题,并辅助多普勒测量信息在线估计目标角速率,作为目标转弯模型的参数。采用两个可内部切换的运动模型(常速度/逆时针转弯模型、常速度/顺时针转弯模型)作为交互模型,有效地描述了目标机动并减少了模型个数,进一步提高了算法的效率。对一个以不同角速率多次转弯的机动目标进行跟踪仿真,结果验证了算法的可行性和优越性。

To cope with the problem of maneuvering target tracking,an improved interactive multiple model particle filtering（IMM-PF） algorithm is proposed on the basis of Doppler-aided measurement.The algorithm adopts the PF to deal with the nonlinear relation of the measurements and states of target.The angular rate of the target used in the CT model is estimated by the Doppler-aided measurement on-line.The two switching model（CV/counter clockwise CT model and CV/clockwise CT model） are used as the interactive multiple model.This method could describe the target maneuvering accurately,and the filtering efficiency is improved by using less models.The tracking simulation results of the maneuvering target which turned at different angular for many times rates demonstrate feasibility and superiority of the novel algorithm.