Intelligent Control System for Ground Vehicles

This paper explores a hybrid control system for the reactive autonomous navigation of ground vehicles tested on a differential drive robot, in an unknown environment featuring both static and dynamic scenarios, including a maze-like environment. The task comprises two primary functions: navigation and obstacle avoidance, managed by two sub-controllers. Neural networks, trained via supervised learning on lookup tables generated from Type-2 Sugeno fuzzy logic controllers, are implemented for real time experiments due to constraints on processing capability. This hybrid approach capitalises on the computational efficiency and potential generalisation of neural networks while preserving the interpretability inherent to fuzzy logic controls. The efficacy of the proposed controllers is demonstrated in both simulation and real-world experiments as well as with comparison to other methods.