Robust Neuro-sliding Mode Control for Position Servo-actuated Robot Manipulators

Artificial neural networks (ANN) are well known for their powerful function approximation capabilities, which have been extensively utilized in control systems to model complex dynamic behaviors. Specifically, in robotics, neurocontrol schemes have been developed to address model nonlinearities. However, these models typically assume that robots receive torque as control inputs to the actuators. This assumption does not always hold true, as in applications requiring high-precision movements, most robotic manipulators are equipped with position servo actuators. These actuators do not receive torque inputs; instead, they receive control signals directed to the actuators. This discrepancy creates a gap between the analysis and the implementation of control schemes for position-commanded robots. To address this gap, in this study, we propose a neurocontroller designed for trajectory tracking based in a position servo-actuated robot manipulator model. Lyapunov analysis is employed to guarantee the stability of the closed-loop trajectories, while the convergence of tracking errors is ensured by the underlying integral sliding mode. Finally, simulations validate the effectiveness of our proposed method. © ICROS, KIEE and Springer 2025.

Robust Neuro-sliding Mode Control for Position Servo-actuated Robot Manipulators Academic Article in Scopus