摘要
传统滤波算法,如卡尔曼滤波,通常对系统模型具有较高的依赖性,需要精确建模才能达到较高的估计精度.而现实场景中由于未知环境因素与建模误差的存在,往往使得估计品质不尽人意.因此针对离散系统同时受有界功率扰动和高斯白噪声影响的滤波问题,提出了一种新的线性时不变的鲁棒最优滤波方法.该方法在保证鲁棒性的同时还能够保证最优均方估计.为了保证鲁棒最优滤波,基于系统级综合方法,并根据误差动力学的系统响应以及干扰和噪声的参数来描述估计性能的上界.并在此基础上,提出了一种数值可处理的新的滤波器设计算法.最后借助测速算例验证了结果的有效性,仿真表明采用该方法得出的滤波算法相比于其它现有方法,能够实现理想的估计品质.
Traditional filtering algorithms,like Kalman filtering,are highly model dependent,and are able to achieve acceptable estimation performance under precise model.However,unknown environments and modeling mismatch always occur in real applications,which unavoidably leads to poor estimation performance.Therefore,a new linear time invariant robust optimal filter design method is proposed to solve the problem that discrete systems are affected by both bounded-power disturbance and white Gaussian noise.This method is capable of ensuring the robustness and the optimal mean square estimation,simultaneously.In order to ensure robust optimal filtering,this is based on system level synthesis method,and conducts the upper bound of estimation performance according to the system response of error dynamics and the parameters of disturbance and noise.On this basis,a computationally tractable algorithm is proposed for filter’s gain design.Finally,a velocity estimation example is used to verify the effectiveness of the current results,and it is shown that the current method outperforms some existing filtering algorithms,and achieves better estimate.
作者
孙洪达
冯宇
SUN Hongda;FENG Yu(School of Information Engineering,Zhejiang University of Technology,Hangzhou 310023)
出处
《系统科学与数学》
CSCD
北大核心
2022年第12期3163-3172,共10页
Journal of Systems Science and Mathematical Sciences
基金
国家自然科学基金(61973276)资助课题。
关键词
鲁棒最优滤波
多目标设计
系统级综合
有界功率扰动
速度估计
Robust optimal filtering
multi-objective design
system level synthesis
bounded-power disturbance
speed estimation
