一类模型不确定非线性系统的反演预测控制  

Backstepping predictive control for one type of nonlinear systems with model uncertainty

作　　者：周卫东[1] 郑兰[1] 廖成毅[2] 

机构地区：[1]哈尔滨工程大学自动化学院,黑龙江哈尔滨150001 [2]中国船舶工业集团公司第七0八研究所,上海200011

出　　处：《电机与控制学报》2015年第10期87-92,共6页Electric Machines and Control

基　　金：国家自然科学基金(61102107;61374208)

摘　　要：针对一类模型不确定的单输入单输出仿射非线性系统,设计一种使得闭环系统稳定且滚动时域性能指标在线最小化的预测控制器。运用反演(Backstepping)设计思想获得具有待定参数的控制器表达式,其误差导数中的未知函数采用模糊逻辑系统来逼近,通过直接估计模糊系统最优参数向量的范数上界来设计控制器和自适应律,大大降低了在线计算量。理论证明该方法设计的控制器保证了闭环系统所有信号是半全局有界的,并且跟踪误差收敛于零的某一邻域。仿真算例验证了所提出算法的有效性。Aiming at a class of uncertain SISO affine nonlinear system, a predictive controller which can make the system stable and can minimize an on-line receding horizon guaranteed cost was designed. By using backstepping technique the controller expression with undetermined parameters was acquired. With-in the design procedure, the fuzzy logic system （ FLS） was utilized to approximate the unknown function in error derivative. By directly estimating the norm bound of the FLS optimal parameter vector the control-ler and the adaptive laws were obtained, then the online computation complexity was reduced and the cal-culation was overcome. Theoretically, applying the controller proposed can guarantee that all signals in closed-loop system are semi-globally bounded and the tracking error converges to a neighbourhood of ori-gin. Finally, the simulation results demonstrate the effectiveness of the proposed control scheme.

关 键 词：模型不确定 滚动时域 模糊逻辑系统 反演设计 预测控制 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]