一种面向机动目标跟踪的交互式多模型算法

摘要

针对在传感器可能偏差的情况下进行机动目标跟踪,提出一种新的交互式多模型算法,即IMM-TS-EV算法.该算法同时考虑传感器测量模型与目标运动模型这两类模型的多样性,以此处理传感器偏差性与目标机动情况.但由于考虑两类模型可能会造成模型过多而导致性能衰退,故结合扩展维特比(EV)算法以期望能有效缓解该问题,即新算法属于一种同时考虑两类模型的改进交互式多模型扩展维特比(IMM-EV)算法.最后以仿真结果验证了算法的有效性,能够利用多模型特点同时解决传感器的偏差性与目标的机动性所带来的匹配问题.%Aiming at the possibility of maneuvering target trackingin the case of sensor biases, a new interactive multi-model algorithm, IMM-TS-EV algorithm, is proposed.The algorithm takes into account the diversity of the sensor measurement model and the target motion model in order to deal with sensor deviation and target maneuver.However,considering the two models may cause the model is too much and lead to performance degradation, it is expected to combine extended Viterbi (EV) algorithm to effectively alleviate the problem, that is, the new algorithm belongs to an improved interactive multi-model extended Viterbi algorithm (IMM-EV) which considers both types of models simultaneously.Finally, the simulation results verify the effectiveness of the algorithm, can use the characteristics of multiple models simultaneously to solve the deviation of the sensor and the target mobility caused by the matching problem.