摘要
提出了一种改进的加权最小二乘法估计噪声统计特性Q和R的方法。该方法在最小二乘法的基础上,通过建立次优估计新息自相关函数来确定权值并递推出最优增益。进而推出了噪声的统计特性,并且简化了Q的推导方法。仿真结果表明,加权后的估计值比不加权的估计值收敛更快,精度更高。
An Improved weighted least square method is proposed to estimate the noise statistics Q and R from the sub-optimal estimation by incorporating the autocorrelation functions of the innovation, determine the weight, deduce the optimal gain and finally get the noise statistics. The deduction of Q is also simplified. The simulation result shows the estimation has better convergence rate and precision.
出处
《华东理工大学学报（自然科学版）》
CAS
CSCD
北大核心
2004年第4期433-436,共4页
Journal of East China University of Science and Technology
基金
国防"十五"预研项目(41306030102)
关键词
卡尔曼滤波
新息
自相关函数
噪声估计
Kalman filter
innovation
autocorrelation function
noise estimation
