基于PF的EKF算法及应用

针对拦截弹与目标遭遇时间短,制导精度要求高的特点,提出了一种基于预测滤波(PF)的扩展卡尔曼滤波(EKF)算法.PF是在综合了模型预测控制理论和最小模型误差估计法(MME)的基础上提出的一种新的估计方法.该方法最大特点是能实时地估计出校正项(模型误差项)并加到估计模型中,使产生的新模型更能准确地描述实际系统的行为.PF的这一特点,恰恰可以解决EKF无法处理由于模型不准而导致估计精度下降的问题.因此,可以利用PF估计出的模型误差项实时地修正系统模型,令原来的系统模型更加准确,然后再利用EKF对新的、能更准确描述系统行为的模型进行滤波,最后得到更为精确的制导信息估计值.仿真结果表明,该算法与扩展卡尔曼滤波相比,增加的计算量并不明显,却在系统模型不准确的情况下,能够显著提高制导信息的估计精度,有效减少了拦截弹的脱靶量.

The EKF Based on PF and Appliction

A new EKF algorithm based on predictive filter (PF) aiming the characteristic of the short encounter time and high guidance accuracy with interceptor is presented.The PF is a new algorithm which is developed based on the theory of model predictive control and minimum model error (MME).The algorithm determines the corrections added to the assumed model so that the model with corrections yield an accurate representation of the system behavior.As a result of such characteristic,the problem of less accurate resulted in inaccurate model is efficiently solved,which can not be solved by EKF.Thus,the corrections can be used to correct the system model that become more accurate than before.And then,the more accurate model is filtered by EKF.By this means,the higher accurate estimates are determined.Results using this EKF algorithm based on PF indicate that the new way slightly increases computational complexity,and provides more accurate guidance information than EKF without using of precise system and effectively decrease the miss distance.