Prescribed-Time Trajectory Tracking Control of Wheeled Mobile Robots Using Neural Networks and Robust Control Techniques

Abstract

This research presents a novel trajectory generation algorithm and the design of a prescribed time controller for trajectory tracking tasks for autonomous vehicles. The trajectory generation algorithm uses a hybrid combination of computer vision techniques and intelligent rail detection methods using an on-board camera. Based on the previous information, a possible trajectory is then generated that the vehicle should follow. A time-prescribed controller is then developed and implemented to track the trajectory generated by the proposed methodology. The controller uses a hybrid structure in which a time-varying feedback controller transitions into a fixed-time controller. This approach achieves stabilization in the prescribed time despite the initial conditions. To address the trajectory design, a scaled autonomous vehicle simulator was used to then evaluate the prescribed time controller compared to a finite time controller and a dynamic feedback controller. The simulation results demonstrate the effectiveness of trajectory generation and trajectory tracking control algorithms in addressing these challenges in real-world scenarios by examining two situations: unperturbed and perturbed cases.