杂波环境下机动目标跟踪算法研究

杂波环境下的机动目标跟踪问题具有非线性、非高斯、不完全观测的特点,其难点在于观测值与目标的对应关系及每一时刻的运动模式均呈现高度不确定性。文中将多模型理论和辅助粒子滤波算法相结合,提出了一种新的杂波环境下机动目标跟踪算法——多模型辅助粒子滤波算法(MMAPF)。仿真结果表明,该算法与传统的交互多模型——扩展卡尔曼滤波算法、辅助粒子滤波算法相比,在相同的情况下,具有更高的滤波精度和较好总体性能。

Study on maneuvering target tracking algorithm in clutter

The maneuvering target tracking problem in clutter is always characterized by nonlinear,non-Gaussian and incompletely observed,and the main difficulty lies in the fact that the measurements may not all have originated from the target-of-interest,and maneuvering model at every step is of highly uncertainties,so observations.The paper combines multiple model theory and auxiliary particle filter algorithm,and presents a new maneuvering target tracking algorithm called MMAPF(multi-model auxiliary particle filter).Simulation results demonstrate that the MMAPF algorithm improves filtering accuracy,and has satisfactory integrated performance,compared with IMM-EKF(interacting multi-model extended Kalman filter)and APF(auxiliary particle filter) algorithm.