基于预报器的综合LQG自校正控制器*

Predictor-Based Synthetic LQG Self-Tuning Controller

中文关键词  自校正  最优控制  谱分解

英文关键词  self-tuning  optimal control  spectral factorization

中文摘要

      本文推到了一类多项式LQG自校正控制算法，采用了增量和非增量两种模型形式。全面考虑了权P（z-1），Q（z-1）和滤波多项式T（z-1）的作用，能控制存在有色噪声和前馈干扰的复杂过程，与广义预测控制仿真比较，结果表明这两种控制器均有良好的适应能力和强的鲁棒性。

英文摘要

      A sort of polynomial LQG self-tuning control algorithm is expanded in this paper. Both increment model and non-increment model are adopted. It has considered the function of the weighed P(z-1),Q(z-1) and the filter polynomial T(z-1). It can control the complicated plant with the colour noise and feedforward. Compared with the recursive synthetic generalized predictive control (RSGPC), The simulations show that the two kinds of controllers have a good adaptability and strong robustness. The LQG self-tuning controller needs less adjustive time than RSGPC, but it needs more calculation time.