Format

Send to

Choose Destination
ISA Trans. 2016 May;62:120-8. doi: 10.1016/j.isatra.2016.01.008. Epub 2016 Feb 2.

Iterated non-linear model predictive control based on tubes and contractive constraints.

Author information

1
Research Institute for Signals, Systems and Computational Intelligence (sinc(i)), National Scientific and Technical Research Council (CONICET), Ciudad Universitaria UNL, 4° piso FICH, (S3000) Santa Fe, Argentina. Electronic address: mmurillo@sinc.unl.edu.ar.
2
Research Institute for Signals, Systems and Computational Intelligence (sinc(i)), National Scientific and Technical Research Council (CONICET), Ciudad Universitaria UNL, 4° piso FICH, (S3000) Santa Fe, Argentina. Electronic address: gsanchez@sinc.unl.edu.ar.
3
Research Institute for Signals, Systems and Computational Intelligence (sinc(i)), National Scientific and Technical Research Council (CONICET), Ciudad Universitaria UNL, 4° piso FICH, (S3000) Santa Fe, Argentina. Electronic address: lgiovanini@sinc.unl.edu.ar.

Abstract

This paper presents a predictive control algorithm for non-linear systems based on successive linearizations of the non-linear dynamic around a given trajectory. A linear time varying model is obtained and the non-convex constrained optimization problem is transformed into a sequence of locally convex ones. The robustness of the proposed algorithm is addressed adding a convex contractive constraint. To account for linearization errors and to obtain more accurate results an inner iteration loop is added to the algorithm. A simple methodology to obtain an outer bounding-tube for state trajectories is also presented. The convergence of the iterative process and the stability of the closed-loop system are analyzed. The simulation results show the effectiveness of the proposed algorithm in controlling a quadcopter type unmanned aerial vehicle.

KEYWORDS:

Contractive constraint; Iterative process; Non-linear model predictive control; Robust model predictive control

Supplemental Content

Loading ...
Support Center