随机时变系统最小二乘辨识的收敛速度

CONVERGENCE RATE OF LEAST SQUARES IDENTIFICATION FOR STOCHASTIC TIME-VARYING SYSTEMS

对不相关噪声的多维时变系统，在相当一般的非持续激励条件下，建立了最小二乘算法的收敛速度，并通过引入衰减激励控制，在可辨识条件下，获得了该非持续激励条件．

Under the more general non-persistent excitation condition convergence rate of least squares algorithm for multidimensional time-varying systems with uncorrelated noise is established. By introducing attenuating excitation control the non-persistent excitation condition is obtained under identifying condition.