自适应转移概率交互式多模型跟踪算法

摘要

针对标准的交互式多模型算法(Interacting Multiple Model,IMM)存在模型集设计困难和采用固定转移概率矩阵导致模型切换缓慢、跟踪精度下降的不足,提出一种自适应转移概率IMM算法.首先,提出了一种新的模型集设计方法,将强跟踪修正输入估计(Strong Tracking Modified Input Estimation,STMIE)模型和匀速运动(Constant Ve-locity,Cv)模型作为IMM算法的模型集,利用STMIE算法对高机动目标的跟踪能力以及CV模型对非机动目标跟踪的高精度,实现对目标的全面自适应跟踪.其次,提出一种依据模型似然函数值对Markov转移概率进行实时修正的方法,增强匹配模型的作用,削弱不匹配模型的影响.仿真结果表明,依据模型似然函数修正转移概率的方法使IMM算法的模型切换速度和跟踪精度都得到提高,提出的IMM-STMIECV算法的跟踪精度高于IMM-CVCA、IMM-CVCACT以及IMM-CVCS算法.%There are two shortcomings in the standard interacting multiple model (IMM) algorithm:one is that designing models is difficult,the other is that the application of constant transition probability matrices makes the model switching speed slow and tracking accuracy decreased.To overcome these shortcomings,an IMM algorithm with adaptive transition probability is proposed.Firstly,a new model-set design method is proposed,and the strong tracking modified input estimation (STMIE) model and constant velocity (CV) model are adopted as the model sets of the IMM algorithm.By using the capability of STMIE model to track high maneuvering targets and the precision of CV model to track non-maneuvering targets,this algorithm can be comprehensively adaptive in target tracking.Secondly,a new method is proposed to modify the Markov transition probability in real time based on the likelihood values of the models,which enhances the effect of the matching model,and weakens the influence of the mismatched model.Simulation results show that the new method improves model switching speed and tracking precision of IMM algorithm,and the tracking precision of IMM-STM1ECV algorithm is higher than that of IMM-CVCA,IMM-CVCACT and IMM-CVCS algorithms.