基于模糊逻辑和机动检测的AGIMM跟踪算法

AGIMM Tracking Algorithm Based on Fuzzy Logic and Maneuvering Detection

针对机动目标高机动和多阶段运动的特性,单模型跟踪算法难以实现精确跟踪的问题,提出一种基于模糊逻辑和机动检测的AGIMM跟踪算法。考虑现有的AGIMM算法网格模型收敛速度慢和机动检测方法过于依赖模型后验概率的问题,首先引入模糊逻辑算法计算机动检测的可信度,然后利用机动检测可信度信息、目标的机动信息和模型的后验概率调整跟踪模型集的结构,克服了AGIMM算法模型收敛速度慢的不足,实现了模型与目标运动模式的快速匹配。仿真结果表明:与AGIMM算法和固定结构交互式多模型算法相比,文中提出的算法使得不同条件下的位置跟踪误差分别至少降低5.6%和15.1%,说明该算法具有更快的模型收敛速度和更高的目标跟踪精度。

Due to the high mobility and multistage mobility of the maneuvering targets,the single-model tracking algorithm is difficult to achieve accurate tracking.To solve this problem,AGIMM tracking algorithm based on fuzzy logic and maneuvering detection is proposed.Considering the slow convergence speed of the mesh model of adaptive grid interacting multiple model algorithm and the problem that the maneuvering method is too dependent on the posterior probability of the model,the algorithm introduces fuzzy control to detect the credibility of the maneuver decision.Then,the adjustment method of the model mesh is re-given by the information of maneuver detection credibility target maneuver information and the model posterior probability,which overcomes the shortcoming of the slow convergence speed of the AGIMM algorithm and achieves a fast matching between the model and the target motion mode.Simulation results show that,under different observation conditions,the position tracking error of the proposed algorithm is at least 5.6%lower than that of the AGIMM algorithm,and the position tracking error of the proposed algorithm is at least 15.1%lower than that of the fixed-structure interactive multi-model algorithm.Experimental results show thatthe proposed algorithm guarantees faster model convergence speed and large target tracking accuracy.