基于S修正RBUKF的自适应网格交互式多模型算法

摘要

针对观测方程为非线性,状态方程为线性,且噪声为加性情况下的机动目标跟踪问题,应用Rao-Black-wellised UKF(RBUKF)算法滤波并对其进行了s修正,在此基础上,采用自适应网格(AG)方法对模型集进行自适应调整,得到一种基于S修正RBUKF的自适应网格交互式多模型(SRBUKF-AGIMM)算法.对二维蛇形机动目标跟踪的仿真结果表明,该算法与固定结构多模型(FSMM)算法相比,可在计算量相当的情况下大幅提高跟踪精度,大幅提高算法的费效比.%Devoted to the problem of maneuvering target tracking under nonlinear observation, linear state and added noise,an adaptive grid interacting multiple model algorithm based on S-amended RBUKF(SRBUKF-AGIMM)is developed,which uses Rao-Blackwellised UKF(RBUKF) for filtering and S-amended for anti-divergent work,adapts adaptive grid method to adjust adaptively the model sets. Two-dimensional snake maneuvering target tracking simulation results show that this algorithm can largely improve the tracking accuracy and cost-efficiency ratio of the algorithm under the similar computation.