一类非自衡系统的简化模型鲁棒预测控制

预测控制是基于模型的控制 ,从模型着手来提高预测控制系统的鲁棒性不失为一种简单易行的办法。鉴于脉冲响应模型具有简单易得 ,鲁棒性强的特点 ,利用阶梯化脉冲响应模型来作为预测控制系统的预测模型 ,对一类非自衡对象进行鲁棒预测控制器的设计。用这种阶梯化脉冲响应模型来拟和实际脉冲响应模型的办法 ,不仅减少了模型的参数 ,而且由于这种简化模型与对象响应的细节无关 ,进一步提高了模型的鲁棒性能。通过仿真研究 ,证明了本文设计的两种鲁棒预测控制器 (单步鲁棒预测控制器和扩展时域鲁棒预测控制器 )对对象模型失配和干扰有较强的性能鲁棒性和稳定鲁棒性。该模型的参数少 ,控制器设计简单 ,为实现自适应预测控制提供了基础

Robust Predictive Control Based on Simplified Model to Kind of Nonself-Regulating Systems

Predictive control is based on model. Beginning with model is a simple and easy-to-apply method to improve the robust of the predictive control system. Since the impulse-response model is easy to be obtained from the experiments and has strong robust capability, a kind of cascade impulse-response model is used as the predictive model of the predictive control system, and two robust predictive controllers are designed via a sort of nonself-regulating systems. This kind of simplification method not only can reduce the model's parameters, but also can enhance the robust capability of the model because the simplified model does not associate much to the details of the response. Simulations proved that the two proposed robust predictive controllers in this paper have strong robust stability and robust performance against the model error and disturbances. The model has a few parameters, which makes controller design simple. And this model privides a basis of self-adaptive predictive control.