Metaheuristic Optimization Methods For Optimal Supertwisting Sliding Mode Controller

Designing optimal controllers aims to improve the performance of a control system by minimizing tracking errors,control effort, or other system-specific objectives. However, obtaining analytical solutions for optimal controllers is often intractable for highly nonlinear, coupled multi-input multi-output (MIMO) systems. While gradient-based optimization methods can be employed, they may converge to suboptimal solutions due to the presence of multiple local minima in the cost function. Metaheuristic algorithms, on the other hand, offer a way to search for global optima even in complex, nonlinear systems. This article considers a nonlinear, coupled 2-DOF SCARA robot manipulator and proposes a super-twisting sliding mode controller that generates smooth control inputs compared to conventional sliding mode controllers, which typically produce discontinuous signals. We explore the use of genetic algorithms, particle swarm optimization, and simulated annealing to automatically tune the controller parameters, with the objective of minimizing trajectory and velocity tracking errors along a predefined path.