Offset-free MPC for resource sharing on a nonlinear SCARA robot

M. Bianchi, A. van der Maas, E. Maljaars, W.P.M.H. Heemels

Research output: Contribution to journalConference articlepeer-review

1 Citation (Scopus)
1 Downloads (Pure)

Abstract

High-precision motion industrial systems must satisfy tight performance requirements. Both positioning accuracy and throughput demands are typically achieved through improvements in hardware, thereby raising the bill of materials. A cost saving alternative could be to strive for a reduction in the hardware components needed, in combination with advanced motion control, to still meet the desired specifications. Particularly, in this paper, the possibility is analyzed to allow for resource sharing among several actuators. This results in a switched system, for which we develop a real-time MPC algorithm for optimization of both the input and the switching signals. This implementation applies to a fairly general class of nonlinear systems and uses a novel offset-free formulation in velocity form for LTV prediction models, to realize good tracking performance under the resource sharing constraints. We provide a proof of concept for this MPC solution on a high fidelity model of an industrial SCARA robot, where it is proposed to use a single amplifier to serve two actuators. The MPC solution is compared to heuristically switched LTI controllers, and the potential of the proposed approach is shown in simulations.

Original languageEnglish
Pages (from-to)265-272
Number of pages8
JournalIFAC-PapersOnLine
Volume51
Issue number20
DOIs
Publication statusPublished - 1 Jan 2018
Event6th IFAC Conference on Nonlinear Model Predictive Control NMPC 2018 - Madison, United States
Duration: 19 Aug 201822 Aug 2018

Keywords

  • Control applications
  • hybrid systems
  • mixed-integer quadratic programming
  • motion control
  • offset-free MPC
  • robotics
  • switching systems

Fingerprint

Dive into the research topics of 'Offset-free MPC for resource sharing on a nonlinear SCARA robot'. Together they form a unique fingerprint.

Cite this