A Suboptimal Approach to Real-time Model Predictive Control of Nonlinear Systems

被引:10
|
作者
Graichen, Knut [1 ]
Egretzberger, Markus [2 ]
Kugi, Andreas [2 ]
机构
[1] Univ Ulm, Inst Mess Regel & Mikrotech, D-89081 Ulm, Germany
[2] TU Wien, Inst Automatisierungs & Regelungstech, A-1040 Vienna, Austria
关键词
Model Predictive Control; nonlinear systems; real-time feasibility; time and memory efficient computation; suboptimality; incremental improvement; RECEDING-HORIZON CONTROL; STABILITY; OPTIMALITY; SCHEME;
D O I
10.1524/auto.2010.0860
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
The paper describes a fast Model Predictive Control (MPC) scheme for nonlinear systems subject to control constraints. The MPC scheme computes a suboptimal solution of the underlying optimal control problem that is incrementally refined over the runtime of the process. The MPC scheme uses a fixed number of iterations in each sampling step and allows for a time and memory efficient computation of the single iterations. The control performance, the real-time feasibility as well as the incremental improvement of the MPC scheme are demonstrated for a laboratory overhead crane with a sampling time of 2 ms.
引用
收藏
页码:447 / 456
页数:10
相关论文
共 50 条