Recent Metaheuristics on Control Parameter Determination

Abstract

Metaheuristics have been widely used in recent years for tuning control parameters since they have a simple structure, are easy to apply, and provide efficient solutions. In this study, control of a two-wheeled mobile robot using the inverted pendulum principle is proposed. The performances of nine recent metaheuristics (Political Optimizer, Equilibrium Optimizer, Aquila Optimizer, Flow Directional Algorithm, Cheetah Optimizer, Golden Jackal Optimizer, Artificial Rabbit Optimization, Gazelle Optimizer, and Pelican Optimization) have been investigated for the balancing and speed control of a two-wheeled vehicle. In this context, a framework consisting of two cascaded PI controllers is designed to provide balance and speed control of the two-wheeled vehicle. The performances of the recent metaheuristics are also compared with previously introduced effective metaheuristic algorithms for further evaluation. The parameters of the controllers are tuned by using these metaheuristics. In experimental studies, quantitative and qualitative analyses are performed for evaluation of the metaheuristics. The dynamic system properties, convergence curves, computational times, and statistical results are provided to prove optimal control performances. The results show that 11 out of 14 compared algorithms produce similar optimal results in speed and balance control of the two-wheeled vehicle. The rest of them do not provide satisfactory results for the tuning of optimum control parameters of the two-wheeled vehicle.