基于模型概率估计的两层交互多模型算法

摘要

在机动目标跟踪问题的研究中，针对转弯机动目标跟踪精度差的问题，设计了一种利用模型概率对模型集合进行实时计算的两层交互多模型算法。该算法由第一层目标转弯速率的粗估计和第二层目标状态向量的精估计构成，它利用第一层模型计算出的目标运动估计转弯速率，构建与当前目标运动状态匹配程度较高的第二层转弯模型集合，并将该转弯模型集合运用于状态估计中。仿真结果证明了该算法在针对转弯机动目标时仍有较好的跟踪精度。%Aimming at the problom of the poor precision of tracking turn maneuvering target, a two-stage IMM algorithm which uses model probability to calculate model set is proposed. The algorithm contains the rough estimation of turn rate and the accurate estimation of the target�s state, it estimates the target�s state by a well matched model set which formed by the estimated turn rate. The simulation result proves that the algorithm has a high tracking precision dealing with turn maneuve-ring target.