摘要
随着信号处理技术的发展,参数模型的自适应算法在滤波、辨识、控制等方面得到越来越广泛的应用。本文介绍四种比较有效的算法——递归最小二乘法、格形随机梯度法、格形最小二乘法、快速横向最小二乘法,并对它们在实用中必须注意的一些主要特性——计算工作量、收敛速度、鲁棒性等进行了比较分析,提出选择算法的参考意见。
With the development of the signal processing technique, the adaptive algorithms for parametric model are widely used in filtering, identification, control etc.. In this paper, we introduce four effective algorithms: the Sequential Least Squares Algorithm, the Fast Sequential Lattice Gradient Algorithm, the S-equential Lattice Least Squares Algorithm and the Fast Recursive Least Squares Transversal Algorithm, and comparatively analysis their important performances, suchas computational requirement, convergence rate and robustness, which should be noticed in practice.Finally, some views on choosing algorithm are introduced.