Particle Swarm Optimisation for Model Predictive Control Adaptation

0574863 - ÚTIA 2024 RIV US eng C - Conference Paper (international conference)
Belda, Květoslav - Kuklišová Pavelková, Lenka
Particle Swarm Optimisation for Model Predictive Control Adaptation.
Proceedings of the 27th International Conference on Circuits, Systems, Communications and Computers - CSCC 2023. Piscataway: IEEE, 2023 - (Mastorakis, N.), s. 144-149. ISBN 979-8-3503-3760-0.
[International Conference on Circuits, Systems, Communications and Computers (CSCC 2023) /27./. Rodos (GR), 19.07.2023-22.07.2023]
R&D Projects: GA ČR(CZ) GC23-04676J
Institutional support: RVO:67985556
Keywords : data-driven modelling * parameter estimation * particle swarm optimisation * predictive control
OECD category: Robotics and automatic control
http://library.utia.cas.cz/separaty/2023/AS/belda-0574863.pdf

This paper is focused on parameter identification for Model Predictive Control (MPC). Two identification techniques for parameters of Auto Regressive model with eXogenous input (ARX model) are considered: namely the identification based on Particle Swarm Optimisation (PSO) and Least Square (LS) method. PSO is investigated and LS is presented in square-root form as a reference method for comparison, respectively. The following points are elaborated and discussed: i) parameters’ estimation of ARX model, ii) design of PSO and LS procedures, iii) design of data-driven MPC algorithm in square-root form, iv) concept of possible use of PSO for semiautomatic fine tuning or retuning of MPC parameters. The proposed theoretical procedures are demonstrated using simply reproducible simulation experiments. Application possibilities are discussed towards robotics and mechatronics.
Permanent Link: https://hdl.handle.net/11104/0344802