Adaptive Elite Ant Lion Optimizer tuned optimal controller for underactuated systems

Abstract

The issue of stabilization and ensuring the correct path is followed in the cart-inverted pendulum underactuated system is being addressed. In the construction of a robust Linear Quadratic Regulator (LQR) optimum controller, the optimization algorithms are best suited for tuning the LQR weighing matrices which are usually obtained by cumbersome trial and error methods. In this study, a novel Adaptive Elite Ant Lion Optimizer (AE-ALO) is suggested to tune the weighing matrices of the optimum controller. Aiming at the foible of ALO's imbalanced exploration for some intricate optimization problems, and influenced by Adaptive Particle Swarm Optimization (APSO), the amended location of antlions in ALO's elitism operator is enhanced, yielding the Adaptive Elite-ALO (AE-ALO). The proposed AE-ALO tested in 12 standard benchmark functions outperforms the existing algorithms interms of global exploration and is then used to tune the LQR weighing matrices such that the trajectory tracking error is minimized. The suggested controller's viability is demonstrated in Quanser's IP02 benchmark Cart-inverted pendulum system. The study found that using the method proposed resulted in a 10.97 % decrease in the ISE of trajectory tracking while stabilizing the pendulum in the unstable upright position, as compared to using ALO and APSO tuned LQR control schemes.