White Shark Optimizer for LQR tuning applied to position control of a 2-DOF helicopter

Resumen

The design of optimal controllers for complex dynamic systems is a key application area of Artificial Intelligence. This work proposes the optimization of a Linear Quadratic Regulator (LQR) applied to the control of a 2 degree-of-freedom (2-DOF) helicopter using the bio-inspired White Shark Optimizer (WSO) algorithm. Conventional LQR parameters tuning is typically performed using analytical methods, which may be inefficient given the complexity and nonlinearity inherent in these systems. As an alternative, WSO is implemented to adjust these parameters and is compared against an LQR designed using the traditional analytical method and another optimized with Particle Swarm Optimization (PSO). The evaluation is conducted through simulations and an inferential statistical analysis of both heuristics. The results show that WSO enables a more efficient LQR tuning in terms of performance indices, including the Integral Absolute Error (IAE), the Integral Squared Error (ISE), and their time-weighted variants (ITAE and ITSE), compared to other tuning approaches.