基于滚动时域估计的带约束运动目标跟踪

摘要

针时实际的运动目标跟踪问题中存在的各种物理约束,采用基于在线滚动优化原理的滚动时域估计方法,将跟踪滤波问题转换为带约束的有限时域优化问题,并通过引入到达代价函数,有效减少了优化问题求解所需的计算量.最后,对实际的目标跟踪问题进行了滚动时域估计仿真研究.Monte Carlo仿真结果表明,滚动时域估计能有效提高跟踪精度,并且能在采样周期之内完成求解,满足在线估计的需要.%This paper introduced a method of moving horizon estimation (MHE) to solve the problem of physical constraints in actual moving target tracking.Based on moving optimization online, MHE transformed the problem of target tracking into the process of constrained optimal state estimation on finite horizon.Meanwhile, MHE reduced the computation burden by introducing the conception of arrival cost.And it also gave a Monte Carlo simulation of moving target tracking based on MHE, sufficiently considering the physical constraints.The simulation results show that, MHE can improve the tracking accuracy and complete the optimization within the period of sampling.It meets the demands of estimation-online.