An experimental dynamic identification & control of an overconstrained 3-DOF parallel mechanism in presence of variable friction and feedback delay

The closed-loop dynamic control of parallel mechanisms is a challenging field due to the high complexity of their dynamic behavior, especially those with prismatic actuators. Prismatic joints usually provide considerable amount of friction which could vary along its axis. This paper addresses the application of different control algorithms in order to tackle the challenges in closed-loop control of an overconstrained 3-DOF decoupled parallel mechanism which compromises three prismatic actuators with variable frictions along each axis. In addition, a Kinect vision sensor, as the position feedback, is installed. Since the Kinect RGB sensor performs at 30 frame per second, there is a 33 ms delay in the feedback of the control unit which restricts the control loop frequency to maximum value of 30 Hz. Then, based on the models obtained from the identifications of step response, kinetic friction and inverse dynamic model, various attempts have been made in order to obtain a controller with a reasonably performance. First, the conventional PID and sliding mode controllers are applied. Then, the position-velocity controller based on the obtained experience of the mechanism performance is proposed, in which a feedforward unit as the friction compensator is added to the latter feedback-based control units. Eventually, a feedback-feedforward controller based on PID controller and a compensator based on the identified inverse dynamic model is applied to the mechanism which was able to improve the performance of the control unit to a sufficient level.