基于衰减因子的时变增益KALMAN滤波

Kalman Filter Based Adaptive Fading Memory

常用的卡尔曼滤波器(KF,EKF)一般只适用于平稳的状态过程,当目标处于变化不大的过程时,目标状态突然变化时,KF和EKF将产生较大的估计偏差,甚至丢失目标.文中根据正交性原理,采用时变衰减因子,使得残差达到正交或近似正交,从而使滤波器在不确定模型下保持较好的鲁棒性.通过实时调节增益,促使测量残差近似正交,从而提高对机动目标的跟踪性能.仿真实验表明,应用该方法可以快速、稳定地完成目标运动要素的估计.

The common Kalman filters generally apply to steady status process,when the target′s status changed suddenly KF and EKF′s filter effect becomes deteriorate.This article introduced adopting attenuation factor,made the measure remnants to hit decussate or close to decussate,make the filter keep better rubust under the uncertainty model though real-time adjusting gain,urged the measure remnants close to decussate,raise the on the trail of performance of mobile target.Simulation shows that the method can estimate the target′s status rapidly and stability.