Robust model reference adaptive controller for 3-DOF planar manipulator

Abstract

The three-link robotic manipulator plays an important role in industrial automation, where productivity and efficiency are directly related to precise motion control. However, its dynamics are highly nonlinear and strongly coupled, making the control problem challenging. In this work, a novel decoupling technique is introduced, where each joint acceleration is modeled as being primarily influenced by its own torque and velocity, while the remaining coupling dynamics are considered as uncertainties. Following this approach, simplified decoupled state—space equations are derived for all three links. A conventional Model Reference Adaptive Controller (MRAC) is then designed to track the desired reference model. While effective under ideal conditions, the conventional MRAC suffers from parametric divergence in the presence of external disturbances. To overcome this issue, a robust MRAC is proposed, which explicitly accounts for parametric uncertainties and disturbance effects. Simulation results demonstrate that the robust MRAC achieves superior performance compared to the conventional MRAC, maintaining stability and accurate trajectory tracking even under uncertainty and external disturbances.