Control methods for a sports shoe testing robot.

摘要

There is a need to test sports shoes for their ability to control ankle pronation and supination. The Sports Shoe Testing Robot is being developed as a tool to standardize the testing. A prototype robot, using a Stewart platform type manipulator, positions the ground against a stationary shoe/foot fixture, which is fitted with a force/torque sensor at the ankle. The control of this robot is unique because of the non-anatomical design and the need for the shoe to influence the path of the robot. A proposed solution, called Biomechanical Simulation Control (BSC), uses a simulation of a human to control the robot. Ideally, BSC would occur in real time, but the speed of the robot is limited to that of the simulation. At this slow speed, the dynamic conditions are lost and the process is known as Quasi-static Simulation Control (QSC). To compensate for the lost dynamics, iterative methods of BSC, which allow the robot to run at full speed by separating it from the simulation, are needed. This thesis provides a general mathematical description of BSC and of two methods for dynamic compensation: the Single Step Method (SSM), which builds the corrected trajectory point-by-point, and the Dynamic Correction Method (DCM), which corrects the full trajectory on each iteration. A simulation of a simplified dynamic system is used to evaluate and compare the effectiveness of the two methods. It is concluded that the Single Step Method out performs the Dynamic Correction Method under the given conditions.