Premium
Embedded MPC Controller Based on Interior‐Point Method with Convergence Depth Control
Author(s) -
Ding Yi,
Xu Zuhua,
Zhao Jun,
Wang Kexin,
Shao Zhijiang
Publication year - 2016
Publication title -
asian journal of control
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.769
H-Index - 53
eISSN - 1934-6093
pISSN - 1561-8625
DOI - 10.1002/asjc.1299
Subject(s) - quadratic programming , model predictive control , controller (irrigation) , solver , computer science , convergence (economics) , interior point method , computation , control theory (sociology) , digital signal processor , mathematical optimization , quadratic equation , algorithm , digital signal processing , computer hardware , mathematics , control (management) , artificial intelligence , geometry , agronomy , economics , biology , economic growth
To allow the implementation of model predictive control on the chip, we first propose a primal–dual interior point method with convergence depth control to solve the quadratic programming problem of model predictive control. Compared with algorithms based on traditional termination criterion, the proposed method can significantly reduce the computation cost while obtaining an approximate solution of the quadratic programming problem with acceptable optimality and precision. Thereafter, an embedded model predictive controller based on the quadratic programming solver is designed and implemented on a digital signal processor chip and a prototype system is built on a TMDSEVM6678LE digital signal processor chip. The controller is verified on two models by using the hardware in loop frame to mimic real applications. The comparison shows that the whole design is competitive in real‐time applications. The typical computation time for quadratic programming problems with 5 decision variables and 110 constraints can be reduced to less than 2 ms on an embedded platform.