Résumé
In this chapter, we address a new control strategy for parallel manipulators based on L 1 adaptive control. This latter is known for its decoupled control and estimation loops, enabling fast adaptation and guaranteed robustness. To improve the tracking performance of parallel manipulators, we propose in this work to include the dynamic model of the robot in the control loop of L 1 adaptive control. The additional dynamics-based term partially compensates for inherent nonlinear dynamics of the robot in order to reduce the impact of uncertainties on the closed-loop system and enhance the overall control performance. Real-time experiments, conducted on a 4-DOF fast parallel ma-nipulator developed in our laboratory, demonstrate the effectiveness of the proposed controller and its superiority compared to standard L 1 adaptive control in terms of tracking performance.