一类递推辨识算法的评价和比较

Evaluation and Comparison ot A Class ot Recursive Identification Algorithms

本文对递推伪线性回归辨识算法结构进行了描述和比较。分析了各种输出值估计法对递推算法行为的影响,从而得到了使算法县有良好鲁棒性的一个条件以及算法一致收敛的充分条件。文中还列出了9个算法的仿真比较结果,验证了分析结论。

This paper describes a uniform structure of recursive pseudolinear regression identification algorithems. The effects of different output estimators on the behavior of the recursive algorithems are analysed, a condition which ensures the robust and the sufficient condition of consistency convergence for recursive algorithems are found. To validate the results of theoretic analysis, nine algorithems formulated in the structure are compared by simulation.