一种自适应模糊模型滤波算法

An Adaptive Fuzzy Model Filtering Algorithm

针对使用单一模型滤波在滤波过程中模型结构及参数不再变化,难以适配目标不同机动阶段问题,提出一种自适应模糊模型滤波算法。该算法首先选取一个CV模型和CA模型,其次每时刻通过计算CV模型和CA模型滤波模型似然函数,计算一种代表目标真实运动模式期望值的期望模型,最后利用期望模型进行目标跟踪。蒙特卡洛仿真表明,该算法能有效跟踪目标不同的机动阶段,相比较单一CV和CA模型位置和速度的跟踪精度更高。

An adaptive fuzzy model filtering algorithm is proposed to solve the problem that the model structure and parameters do not change during the filtering process, and then it is difficult to adapt to different maneuvering stages of the targets. The algorithm firstly selects a CV model and CA model, then calculates an expectation model representing the expected value of the real motion mode of the target by calculating the likelihood function of the CV model and CA model filtering model at every moment, and final y uses the expectation model to track the target. Monte Carlo simulation shows that the proposed algorithm can effectively track different maneuvering stages of the target, and the tracking accuracy of position and velocity is higher than that of single CV and CA models.