We propose a method to realize trajectory tracking control of a three-wheeled omnidirectional mobile robot by using radial basis function (RBF) neural network.
In the proposed method, the parameter uncertainty, structural uncertainty and unknown disturbance present in the dynamic model of a three-wheeled omnidirectional mobile robot are considered as a total disturbance and estimated and compensated using RBF neural network.
First, based on the dynamic model of the three-wheeled omnidirectional mobile robot, a computed torque(CT) control law is derived and a disturbance tracking compensator using RBF neural network is designed to overcome the shortcomings of CT control in the presence of model uncertainties and unknown disturbances.
Next, we derive the training rules of the proposed neural network for disturbance estimation compensation by using gradient descent method, with the square form of the filter tracking error as the objective function.
Finally, the effectiveness of the proposed method is verified by simulation and hardware experiments.
Experimental results show that the proposed method is effective compared to other trajectory tracking methods in the presence of model uncertainties and unknown disturbances.
Our work has been published in "Journal of the Brazilian Society of Mechanical Sciences and Engineering" under the title of "Trajectory Tracking Control of a Three-wheeled Omnidirectional Mobile Robot Using Disturbance Estimation Compensator by RBF Neural Network" ((2023) 45:432) (https://doi.org/10.1007/s40430-023-04340-5).
