Constellation, le dépôt institutionnel de l'Université du Québec à Chicoutimi

Résumé

Optimal tension control in parallel mechanisms for balancing wrench at the end‐effector introduces a complex model of the overall system. Moreover, actuator saturations such as motor current and angular velocity limits impose that a non‐linear system be controlled. Such a system is usually controlled sub‐optimally due to the complexity of its model. In haptic applications where transparency and stability are both critical parameters, optimal performance should be achieved. This paper introduces an enhanced FPID auto‐tuning based on Extremum Seeking‐Tuning (ES‐Tuning) which uses a different cost function to find the local optimal solution for haptic rendering and a correction on noise measurement for a high order complex system. The solution of this optimization gives a direct measure of the reel dynamic transparency evaluated by the proposed cost function.