SLIDING MODE NEURO-ADAPTIVE CONTROLLER DESIGNED IN DISCRETE TIME FOR MOBILE ROBOTS
Francisco Rossomando, Carlos Miguel Soria, Eduardo Oliveira Freire, Ricardo Carelli · Control and Intelligent Systems · 2018
It is well known that all robotic systems have additional dynamics and disturbances beyond those that are modelled.Hence, in this paper, the performance of a neuro-adaptive sliding mode control (NA-SMC) controller is analysed in the presence of disturbances and unmodelled dynamics.The proposed control strategy has two stages: first, a feedback linearization controller for a kinematic model; second, a neuro-adaptive SMC controller for a dynamic model.The entire control strategy is designed in discrete time using Lyapunov's criterion, and the stability problems caused by direct implementation in discrete time to a system designed in a continuous domain are thereby avoided.The unmodelled dynamics introduce tracking errors in the closed-loop system; however, experiments made using the proposed approach to control mobile robots show that output tracking error tends to zero.