软边界鲁棒控制器设计新方法*

A New Method of Soft-Bound Robust Controller Design

中文关键词  软边界  概率鲁棒设计  截尾分布  参数摄动  性能折衷

英文关键词  soft-bound  probabilistic robust design  truncating distribution  parametric perturbation  performance  trade-offs.

中文摘要

      本文基于对系统模误差的软边界描述，提出了一种新的鲁棒设计方法.采用对应不同参数摄动状态的概率加权性能指标作为优化性能指标，对参数域进行一体化设计.与传统的硬边界方法相比，本文结果保守性低，较好地兼顾了系统的鲁棒稳定性和鲁棒性能，并使系统性能在参数域内得到良好的概率折衷.

英文摘要

      A new method of robust controller design has been developed on the basis of soft-bound description of modeling errors. The parameter space has been designed in its entirety with probabilistic weighted cost functions as the optimization performance. Compared with the generally used hard-bound method, this new one is less conservative and compromises robust stability and robust performance better. It also makes probabilistic trade-off s for system performances in the whole parametric space.