根据卡尔曼滤波器池思想。设计一组有不同初始滤波参数组成的具有并行关系的卡尔曼滤波器。通过浮点编码的遗传算法离线对这些滤波参数寻优,找到最佳滤波参数。仿真结果说明机动多目标跟踪的遗传算法具有较强的全局最优搜索能力,自适应好且跟踪精确度较高,这证实了此新方法在自适应机动多目标跟踪下的可应用性。
According to the idea of Kalman filter pool, a group of the different initial filter parameters of Kalman filters with parallel relationship was designed. Through the float-encoded genetic algorithm offline on optimization of these filtering parameters, the best filtering parameters could be found. Simulation results show that the multi-maneuvering target tracking genetic algorithm search the global optimum capacity strong, having good adaptive effect and tracking precision, which has confirmed this new method can be applied in the adaptive multi-maneuvering target trackingI.