非线性粒子滤波目标跟踪应用仿真

Application of nonlinear particle filtering for target tracking and simulation

针对非线性环境中存在的机动目标跟踪问题,对基于贝叶斯估计的粒子滤波器进行研究,为解决混合退火粒子滤波重要密度函数构造的问题,在混合退火粒子滤波的基础上,通过对系统状态和观测粒子方差的研究,提出了非线性环境下动态退火参数粒子滤波的改进算法,在混合退火粒子滤波中引入动态退火参数来构造高效的重要密度函数,提高了混合退火粒子滤波的跟踪精度,应用该滤波方法对机动目标模型进行仿真,并对多种滤波跟踪算法进行性能测试和比较,仿真实验结果表明,在非线性环境下该粒子滤波方法可行有效。

A novel method for nonlinear filtering based on the research of Bayesian filtering is proposed. The particle filtering is a Bayesian nonlinear filtering that works by the recursion algorithm of posterior distributions, and is widely applied in many fields. The dynamic hybrid annealed parameter is presented to construct a better important density function by the covariance of the state-space and measurement particles. A discrete time dynamic model and simulation smoothing of the target tracking of this model in a nonlinear environment is described. The result of this simulation displays that the particle filtering is more efficient to deal with the nonlinear target tracking.