A Neuro-Fuzzy Controller for a Trajectory Following Mobile Robot

The design of a motion controller for a mobile robot can be a very difficult and tedious task, especially for robots with a complex kinematic structure. Even though several types of motion controllers have been proposed in literature, they are not always applicable on car-like mobile robots equipped with conventional steering wheels. The reason is, that it is generally not possible to derive an inverse kinematic model for such robots. In this paper, a self-tuning intelligent controller for a quasi-omnidirectional mobile robot is presented. The controller is used to control the robot following a desired trajectory. It is implemented as a neuro-fuzzy controller, which can adapt its parameters by a self-learning process in such manner, that the mobile robot can follow a desired trajectory with required accuracy and speed. The process of tuning controller parameters tuning is demonstrated on experiments with a quasi-omnidirectional mobile robot F.A.A.K.