马尔可夫修正的IMM-CKF目标跟踪算法

Research on Adaptive Markov Transition Probability Matrix IMM-CKF Target Tracking Algorithm

针对非线性情况下的机动目标跟踪问题,提出一种马尔可夫转移概率矩阵修正的交互多模型容积卡尔曼滤波(IMMCKF)算法。修正后验信息,使马尔可夫转移概率矩阵在线更新,缩短模型之间的切换时间,提高机动目标的跟踪精度。结合加速度模型(CA)和匀速模型(CV)在MATLAB软件上进行仿真,结果表明跟踪精度明显高于模型转移概率固定下的交互多模型容积卡尔曼滤波算法。验证了算法的可行性和有效性,具有一定的理论意义。

To track manoeuvering target in nonlinear condition, an interacting multiple model eubature Kalman filter（IMM-CKF） algorithm with adaptive Markov matrix was proposed in the paper. By modifying posterior information, the Markov matrix was online updated, the switching time was shortened and the tracking accuracy was improved. Constant acceleration model （CA） and constant velocity model （CV） were combined for simulation. Results demonstrate that although the improved algorithm increases time cost, it has obviously higher tracking accuracy than the IMM-CKF with the model of stationary transition probalities, and the feasibility and effectiveness of the algorithm were also verified.