A model arm for testing motor control theories on corrective movements during reaching

Based on a simple robotic toolkit, we have developed a robotic arm control system to be used as a humanoid benchmark for testing trajectory planning models and control hypotheses for both reaching and corrective movements during reaching. The developed system integrates visual sensory feedback of the end-effector of the arm allowing controlling its movement online. The system may operate in 1) open-loop configuration providing the servos at the joints of the robot one point of an initially planned trajectory every 20 ms; 2) correction closed loop configuration using visual feedback only for planning the corrective movement trajectory or 3) continuous closed loop configuration using initial conditions derived from visual feedback information and computing the next trajectory point at every time step. Although the planning and control of reaching movements has been extensively investigated, not much is know on the planning of corrective movements. The research tool we developed will be used to implement and test different trajectory planning and movement control models.