Intelligent adaptive fractional order controller for mobile robot trajectory tracking

Recently, the ability of autonomous wheeled mobile robots to handle complex tasks, such as autonomous delivery systems in various environments has drawn researchers' interest. However, trajectory tracking control is considered a critical component in these autonomous delivery systems. This work proposes an intelligent adaptive Fractional Order Full State Feedback Controller (FOFSC) using Gray Wolf Optimization (GWO) to enhance a Differential Drive Robot's (DDR) trajectory tracking performance. Fractional-order control has different advantages over typical integer-order controllers due to its nonlocality and memory characteristics. These qualities enable fractional controllers to better capture system dynamics and achieve more flexible tuning by utilizing fractional exponents. The optimization process is deployed in adaptive and non-adaptive approaches. In the non-adaptive approach, GWO iteratively determines optimal gain values for the closed-loop FOFSC (offline). These values remain fixed for future runs, requiring re-optimization if the trajectory changes. On the other hand, the adaptive approach continuously updates gain values during DDR operation, self-adjusting without pre-tuning for changes in trajectory (online). The fractional order controller is compared to its integer form Integer Order Full State Feedback Controller (IOFSC) for DDR trajectory tracking. Circular, infinity-shaped, and square-shaped trajectories are used to test the controllers. Simulation and experimental validation on a QBot 2e mobile robot platform with an OptiTrack motion capture system demonstrate that the intelligent adaptive FOFSC outperforms other controllers in terms of convergence speed, tracking error minimization, and disturbance rejection.