This repository contains Simulink models and Matlab scripts and functions to run the experiments presented in:
@article{comelli2024application,
author = {Comelli, Román and Olaru, Sorin and Seron, María M. and Kofman, Ernesto},
title = {Application of a stabilizing model predictive controller to path following for a car-like agricultural robot},
journal = {Optimal Control Applications and Methods},
volume = {},
number = {},
pages = {},
year = {2024},
doi = {10.1002/OCA.3126},
url = {https://onlinelibrary.wiley.com/doi/abs/10.1002/oca.3126},
eprint = {https://onlinelibrary.wiley.com/doi/pdf/10.1002/oca.3126}
}If something from this repository or the article is used in an academic work, please cite it.
The paper introduces a theoretical extension to use different control and prediction horizons in a recently developed MPC scheme that instead of a control invariant set, utilizes a pair of inner and outer sets with certain properties to guarantee finite-time convergence and ultimate boundedness to a target set. A controller for a car-like robot in a path following application is then designed, based on this strategy, and simulations and a comparison with another NMPC approach are presented.
The previously developed ideas were presented in:
-
R. Comelli, S. Olaru and E. Kofman, Inner-Outer Approximation of Robust Control Invariant Sets, Automatica, vol. 159, p. 111350, 2024.
-
R. Comelli, A. H. González, A. Ferramosca, S. Olaru, M. M. Seron and E. Kofman, Simplified Design of Practically Stable MPC Schemes, Systems & Control Letters, vol. 180, p. 105626, 2023.
By executing the script run_experiments.m, you can try the proposed controller. It will perform many simulations with different combinations of prediction and control horizons and it will save results in folders within directories fcs_mpc and nmpc. To obtain the figure with trajectories given in the paper, go to figures and execute plot_trajectories.m.
Take into account that this was run in Matlab R2021b with Ubuntu 20.04.