Particle swarm optimization-based reduced-order controller for balancing control issues with a two-wheeled mobile robot

Abstract

Abstract Existing controllers for a two-wheel mobile robot (TWMR) are based on higher-order controller (HOC) structures, which are complex and challenging to analyze, synthesize, and implement on hardware. To address this issue, the authors propose a reduced-order controller (ROC) for the pre-specified HOC that can effectively handle unpredictable dynamics. The proposed approach involves two phases. In phase 1, the ROC is made using dominant poles and moment matching. In phase 2, the particle swarm optimization (PSO) method improves the ROC numerator coefficients by minimizing the root mean square error and integral square error between the step responses and Bode magnitude plots of the HOC and ROC. One benefit of this technique is that the PSO algorithm’s search space bounds are not entirely arbitrary. They are instead picked based on the numerator coefficients calculated in Phase 1. It overcomes issues with evolutionary algorithms, such as random search space selection, optimization of additional decision variables, and longer simulation duration. Comparisons are made between the performance of the proposed controller and controllers built using other approaches. The determinations reveal that the proposed ROC outperforms these current techniques, resulting in a more straightforward and effective solution for regulating the unstable TWMR system. The efficacy of higher- and lower-order controllers is evaluated using a perturbed two-wheeled mobile robot. This evaluation focuses on analyzing various time response parameters and performance indices, including integral time absolute error (ITAE), integral square error (ISE), and integral absolute error (IAE). The MATLAB environment is employed as the preferred tool to carry out these simulations.