强机动目标自适应变结构多模型跟踪算法

Adaptive Variable Structure Multiple Model Algorithm for High Maneuvering Target Tracking

在强机动目标跟踪领域,采用传统基于固定模型集合的交互式多模型算法需要大量模型来描述目标机动,需要巨大的计算量,并且过多模型会带来不必要的模型竞争反而降低跟踪性能.为解决它所带来的问题,提出一种自适应变结构多模型算法,采用少量与目标运动模式相关的模型,在不同时刻根据目标当前机动水平自适应调整模型参数建立新的模型集合,并对其进行滤波估计.仿真结果显示该方法能更好的匹配目标运动规律,有效降低计算复杂度,提高跟踪精度.

In the region of high maneuvering target tracking, the traditional interacting multiple model（IMM） algorithm has a fixed structure, which needs lots of models to describe the target maneuver, brings calculation burden and unnecessary inter-competition among the models maybe reduce the performance. To resolve this problem, an adaptive variable structure multiple model （AVSMM） algorithm is presented. Using several models which related to the target motion, according to the current target maneuvering level, at different moments, the algorithm can adjust the model＇s parameters and build a different model set, then begin to filter estimate. Simulation results show that this algorithm can match actual target trajectory with less computational complexity and better accuracy.