Experimental Synthesis of PSO-Tuned Enhanced Optimal Controllers for Visual Tracking System in Unmanned Aerial Vehicles

The success of aerial surveillance missions relies on rapid and precise visual tracking systems that can withstand various motion disturbances. Existing control strategies often fail to optimize simultaneously for tracking accuracy, disturbance robustness, and energy efficiency. To address this challenge, this work proposes two novel enhanced optimal controllers: an enhanced Linear Quadratic Integral (enhanced-LQI) controller and an enhanced Linear Quadratic Gaussian Integral (enhanced-LQGI) controller. These controllers are derived from a novel augmented state-space formulation that incorporates a modified integral action with a derivative component, which significantly improves transient response and damping characteristics. A comprehensive mathematical derivation is provided, and the asymptotic stability of the enhanced-LQI controller is rigorously proven using Lyapunov theory and LaSalle’s invariance principle. However, tuning the effective parameters of these multiobjective controllers is a challenging process. Therefore, this work employs a particle swarm optimization to automatically and objectively tune the controller and estimator parameters, ensuring a fair, performance-optimized comparison across a suite of six controllers: LQR, LQG, LQI, enhanced-LQI, LQGI, and enhanced-LQGI. The proposed designs are rigorously tested through high-fidelity nonlinear simulations and Monte Carlo analyses under various disturbance scenarios. The results demonstrate that each controller offers distinct advantages. Finally, the real-time feasibility is confirmed via Processor-in-the-Loop experiments within the computational limits of standard embedded processors. To the best of our knowledge, this is the first work to propose, mathematically derive, provide a stability proof for, and validate the enhanced-LQGI architecture in aerial visual tracking systems. This study provides a crucial comparative synthesis, offering clear guidance on selecting the optimal control strategy based on specific operational priorities for aerial visual tracking systems and systematically highlights the performance trade-offs, robustness, and energy efficiency of each controller in realistic aerial operations.