摘要
随着步数的增加,传统递推最小二乘法增益矩阵将逐渐收敛到零,使得新的有效测量数据不能够被用来修正估计输出,出现数据饱和现象。为解决该问题,引入包含过程噪声的状态变量更新方程,该过程噪声的方差影响算法的滤波效果和新的测量数据的更新快慢。Matlab/Simulink仿真实验和冲压发动机实验验证该算法的有效性,取得了很好的滤波能力,同时准确表示出新的测量数据的动态过程。
With the increase of step number, gain matrix of traditional recursive least squares method will gradually converge to zero, which causes that new effective measurement data can’t be used for modifying estimation out put, with emergency of data saturation phenomenon. In order to solve this problem, this paper introduces state variable update equation containing process noise. The variance of process noise has influence on the noise filtering effectof algorithm and the update speed of new measurement data. MATLAB／SIMULINK simulation experiment and ramjet test experiment verified the validity of algorithm, with good ability of noise filtering and accurate show of dynamic process of new measurement data.
出处
《中国仪器仪表》
2018年第6期57-62,共6页
China Instrumentation
关键词
数据饱和
过程噪声
最小二乘法
递推最小二乘法
Data saturation
Process noise
Least squares method
Recursive least squares method
