改进的基于隐马尔可夫模型的自适应IMM算法

Improved Adaptive IMM Algorithm Based on Hidden Markov Model

针对机动目标跟踪中交互式多模型算法(IMM)的马尔可夫转移概率矩阵固定不变造成跟踪精度降低的问题,在已有的基于隐马尔科夫模型(HMM)的自适应IMM算法的基础上,对隐马尔可夫链的长度和Baum-Welch算法迭代次数的2个参数对该算法跟踪性能的影响,进行了深入研究分析,进一步明确了这2个参数选择的依据;并针对该算法在目标机动转换时峰值误差增大的问题,给出了2种修正方法,从而提出了改进的基于HMM的自适应IMM算法。最后,通过仿真分析了算法的参数和修正方法对跟踪性能的影响,并与传统IMM算法进行对比,证明了文章提出算法的有效性。

Aiming at the problem that the Markov transition probability matrix of interactive multiple model algorithm(IMM)was fixed and invariant in maneuvering target tracking,based on the existing adaptive IMM algorithm based on hidden Markov model(HMM),the length of hidden Markov chain and the number of iterations of Baum-Welch algorithm were followed by the algorithm.The influence of tracking performance was deeply studied and analyzed,and the basis for selecting these two parameters was further clarified.Aiming at the regression of peak error increasing during target maneuver conversion,two correction methods were given,and an improved adaptive IMM algorithm based on HMM was proposed.Finally,the influence of the parameters and correction methods of the algorithm on the tracking performance was analyzed by simulation,and compared with the traditional IMM algorithm,the effectiveness of the proposed algorithm was proved.