一种可以用于小卫星跟踪的平滑粒子滤波器

针对小卫星的地面机动目标跟踪环境日趋复杂,跟踪精度要求日益提高的现状,引入平滑粒子滤波器。算法结合粒子滤波器和吉布斯采样器(Gibbs Sampler),通过对系统的机动性和测量与目标关联问题的平滑估计,很好地解决了在杂波环境下具有非高斯非线性特性机动目标的跟踪问题。在对杂波环境下机动目标跟踪问题的仿真研究中,对比了该算法与BMM PDA算法(BMM,靴带多模型算法)的跟踪性能,结果证明了新的平滑粒子滤波器算法以计算量为代价获得了更好的跟踪性能。

A Smoothing Particle Filter for Small Satellite Tracking

A smoothing particle filter is proposed to deal with the problem of small satellite tracking of ground target under complicated background and to improve tracking accuracy.The proposed algorithm can commendably solve the tracking problem of complicatedly disturbed Gausian nonlinear moving ground target,through the smoothing estimation of system measurement and target by combination of smoothing particle filter and Gibbs sampler.Comparing the tracking performance of the proposed algorithm with those of BMM PDA and BMM algorithm,it shows that the new proposed algorithm achieves better tracking performance with the cost of increasing calculation amount.