Trajectory tracking of a wheeled mobile robot with uncertainties and disturbances: Proposed adaptive neural control

• Autor(a) principal: Martins, Nardênio Almeida
• Data de Publicação: 2015
• Outros Autores: De Alencar, Maycol, Lombardi, Warody Claudinei, Bertol, Douglas Wildgrube, De Pieri, Edson Roberto, Filho, Humberto Ferasoli [UNESP]
• Tipo de documento: Artigo
• Idioma: eng
• Título da fonte: Repositório Institucional da UNESP
• Texto Completo: http://hdl.handle.net/11449/228309
• topic: Dynamic control; Kinematic control; Lyapunov theory; Neural networks; Sliding mode theory; Trajectory tracking; Variable structure control; Wheeled mobile robot

This paper analyses a trajectory tracking control problem for a wheeled mobile robot, using integration of a kinematic neural controller (KNC) and a torque neural controller (TNC), in which both the kinematic and dynamic models contain uncertainties and disturbances. The proposed adaptive neural controller (PANC) is composed of the KNC and the TNC and is designed with use of a modeling technique of Gaussian radial basis function neural networks (RBFNNs). The KNC is a variable structure controller, based on the sliding mode theory and is applied to compensate for the disturbances of the wheeled mobile robot kinematics. The TNC is an inertia-based controller composed of a dynamic neural controller (DNC) and a robust neural compensator (RNC) applied to compensate for the wheeled mobile robot dynamics, bounded unknown disturbances, and neural network modeling errors. To minimize the problems found in practical implementations of the classical variable structure controllers (VSC) and sliding mode controllers (SMC), and to eliminate the chattering phenomenon, the nonlinear and continuous KNC and RNC of the TNC are applied in lieu of the discontinuous components of the control signals that are present in classical forms. Additionally, the PANC neither requires the knowledge of the wheeled mobile robot kinematics and dynamics nor the timeconsuming training process. Stability analysis, convergence of the tracking errors to zero, and the learning algorithms for the weights are guaranteed based on the Lyapunov method. Simulation results are provided to demonstrate the effectiveness of the proposed approach.