Experimental Evaluation of Centralized Control Strategies on a 5R Robot

Parallel kinematics robots are often used in high-speed pick and place applications, since their mechanical characteristics result in superior dynamic performance. Achievable speed and accelerations are also affected by the control algorithm, with centralized controllers typically leading to substantial improvements over decentralized ones. Several centralized control algorithms are therefore described in the literature, and often tested in simulation or through the execution of simple and slow movements. This article compares two centralized inverse dynamics motion control systems applied to a 5R parallel robot: one featuring integral contributions, the other characterized by sliding mode terms. The test trajectory is representative of a fast pick and place operation on two moving conveyor belts; the experimental findings concern, for the two regulators, the achieved errors, the overall control torques, and the several contributions to the control action. The results show that the integral regulator achieves overall smaller tracking errors, even though high-curvature portions of the test trajectory are better tracked using the Sliding Mode Control algorithm.