用于机动目标跟踪的自适应交互式多模型算法

针对当前统计模型对弱机动或非机动目标跟踪效果不理想等问题,提出了一种修正当前统计模型与匀速模型的自适应交互式多模型算法,可在线修正当前统计模型的加速度极限值,调整过程噪声方差,提高了当前统计模型的自适应性。同时,通过在常规匀速模型中引入机动检测机制,抑制了常规匀速模型对机动目标跟踪的滤波发散,通过引入强跟踪算法,增强了模型对目标突发机动的自适应跟踪能力。仿真结果表明,该算法充分发挥了当前统计模型和交互式多模型算法的优势,对强机动和弱机动目标都具有很好的效果。

Adaptive Interacting Multiple Model Algorithm for Maneuvering Target Tracking

Aiming at the problem that the current statistical model has not a good performance on tracking non- maneuvering targets,the adaptive interacting multiple model algorithm based on modified current statistical model and constant velocity model is proposed.The algorithm could online modify the extreme value of acceleration and the variance of process noise in current statistical model,and the adaptability of current statistical model is improved. Meanwhile,maneuvering detection is introduced to restrain the filter divergence of the normal the constant velocity model,and tracking performance is enhanced for sudden maneuvering targets by introducing a strong track filter algorithm. The simulation results show that the algorithm takes full advantage of the current statistical model and the interactive multiple model algorithm,and has a good performance both on weak and strong maneuvering targets.