ROBUST ADAPTIVE NEURAL TRACKING CONTROL FOR UNDERACTUATED NONLINEAR SYSTEMS: A ROTARY INVERTED PENDULUM CASE STUDY

Abstract

The underactuated system demonstrates significant coupling and highly nonlinear dynamics, posing challenges for precise control. This paper introduces an adaptive control approach for the underactuated rotary inverted pendulum (RIP) system. The objective is to enable the manipulator arm to track a desired trajectory and to maintain the pendulum in an upright position concurrently. The proposed method employs two neural networks: the first focuses on tracking the desired trajectory of the manipulator arm, while the second stabilizes the pendulum in its upright position. Stabilize the system using Lyapunov theory. To validate the effectiveness of the proposed control strategy, experiments are conducted using the NI-PCI 6221 data acquisition card with the RIP system. Both simulation and experimental results underscore the robustness of the proposed control method against system uncertainties and external disturbances, achieving stable operational performance.