基于扩张状态观测器的分散模型预测控制

针对复杂关联系统中分散控制方法无法有效解决子系统间的耦合和干扰问题, 提出一种基于扩张状态观测器的分散模型预测控制算法. 首先将复杂关联系统分解为多个状态维数较低、控制变量较少的子系统, 并为每个子系统设计本地预测控制器; 然后, 采用扩张状态观测器对子系统的耦合项以及干扰项进行估计, 进而利用估计值对子系统进行前馈补偿, 从而降低复杂关联系统的计算复杂度, 提高系统的稳定性和抗干扰能力; 最后, 利用液位控制系统验证了所提出算法的有效性.

Extended state observer-based decentralized model predictive control

A decentralized model predictive control scheme is proposed for complex interconnected systems by using extended state observer (ESO) to deal with the coupling between subsystems and disturbance. Firstly, the considered system is decomposed into several lower-order subsystems with the local model predictive controller. In order to improve the stability and disturbance rejection performance, the ESO is introduced into the local model predictive controller. The coupling and interference of the subsystem are estimated by the proposed ESO and compensated by using a feedforward compensation strategy. Finally, a liquid level control system is given to demonstrate the effectiveness of the proposed method.