有界扰动系统高效鲁棒预测控制器设计

Design of efficient robust model-predictive controller for systems with bounded disturbances

具有有界扰动的有约束线性系统是一类常见的不确定系统.针对此类系统,本文借鉴扰动不变集方法,通过离线设计两个椭圆不变集以降低以往设计的保守性,进而提出一种有界扰动系统的高效鲁棒预测控制器(SD-ERPC)的设计方法.该方法能够较好地处理扰动对系统的影响,在减小控制器在线计算量的同时,扩大原ERPC设计的初始可行域,且具有较好的控制性能.文中给出了SD-ERPC控制器可行性和鲁棒稳定性的理论证明,并通过仿真算例验证了该控制器的有效性.

The constrained linear system with bounded disturbance is a familiar type of uncertain systems. For this type of uncertain systems, an efficient robust model-predictive controller for systems with bounded disturbances（SD-ERPC） is developed. By making use of the invariant set of bounded disturbances, we release the conservativeness in the design by offline designing two ellipsoidal invariant sets. This approach preferably deals with the system effect from the bounded disturbances, reduces the on-line computational work, enlarges the feasible initial-region in the design of ERPC, and guarantees a better control quality. The feasibility and robustly stability of SD-ERPC are also proved. Finally, a simulation example is presented to verify its validity.