融合机动检测的交互多模型目标跟踪

An Interacting Multiple Algorithm Incorporating Maneuver Detection for Target Tracking

传统交互多模型算法所采用的基于“软切换”思想的模式概率估计方法 ,大大减少了计算量 ,但同时只能实现对目标的次优跟踪。为了进一步提高跟踪系统的性能 ,文中讨论了一种融合机动检测的IMM跟踪算法 ,通过采用点过程滤波器对机动目标的机动进行实时检测 ,从而获得各模式的当前概率 ,并将其与传统IMM算法相结合 ,实现跟踪系统性能的提高。仿真结果表明 。

In the traditional interacting multiple algorithm, the method of mode estimation based on the idea of “soft switching” reduces the quantity of computation greatly, but leads to suboptimal tracking performance. To improve the performance, we propose a new IMM algorithm incorporating maneuver detection. This algorithm uses the point process filter in real time to obtatin the current probability of system modes, and realizes the improvement of the performance. The simulation results show excellent performance of the new algorithm.