一种基于线性近似的非线性系统模型预测控制方法  

Nonlinear Model Predictive Control Scheme Based on Linearization Approximation

作　　者：盖俊峰[1] 赵国荣[1] 周大旺[1] 

机构地区：[1]海军航空工程学院,烟台264001

出　　处：《航天控制》2015年第5期3-7,共5页Aerospace Control

基　　金：国家自然科学基金(61004002);航空科学基金(20110184001)

摘　　要：针对一类系统模型的输入输出函数不满足连续可导条件的非线性系统,设计了一种基于线性近似的模型预测控制方案。运用Stirling插值公式对非线性函数进行线性近似处理,并将模型预测控制性能指标重构为一个二次型最优化问题,通过对二次型最优化问题的求解得到了模型预测控制的最优控制序列。为了减少计算量,忽略了线性近似过程中产生的非线性高阶项。仿真结果表明,该模型预测控制方案控制效果满意,且具有降低控制能量消耗和缩短控制时间的优点。In this paper,a model predictive control scheme based on linearization approximation method is proposed for a kind of nonlinear system. The object nonlinear system is linearized by using Stirling interpolation formula method and the control performance index is reformulated to a quadratic optimization problem,and then the optimization control sequences are obtained by solving the quadratic optimization problem. In order to reduce the complexity of computation,the high-order terms associated with the linearization are ignored. The simulation results demonstrate the effectiveness of the proposed method.

关 键 词：线性近似 模型预测控制 二次型最优化 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]