Intelligent trajectory conversion and inverse dynamic control of a 3-DOF neuro-rehabilitation platform

Tracking a desired trajectory in joint space has been favored in several robot manipulators and end-effector control scheme due to the simplicity and high sampling rate offered by the joint space scheme. This usually require the trajectory conversion process, of the desired position, velocity, and acceleration, from Cartesian space to joint space using conventional inverse kinematics solutions which have been known to have several limitations and which often pose a big challenge, computationally, and even prohibitive, to achieve, for some robot designs. In this study, an intelligent approach to the inverse kinematics problem using adaptive neuro-fuzzy inference system (ANFIS) is proposed for control of a 3-DOF end-effector based neuro-rehabilitation platform. The joint positions, velocities, and accelerations are achieved/predicted by means of the ANFIS networks which is trained with data obtained from the forward kinematics, velocity Jacobian and the differential of the velocity kinematics equations. Simulation studies have shown that the proposed intelligent techniques has simplified both the trajectory conversion process and the control framework while tracking is achieve to a high degree of accuracy.