An inverse kinematics algorithm for upper-limb joint reconstruction during robot-aided motor therapy

The paper presents an analytical procedure, based on inverse kinematics with an Augmented Jacobian matrix, for reconstructing human arm joint motion during robot-aided motor therapy with an end-effector machine. The presented method can be exploited for investigating arm motion in the joint space and, accordingly, determining objective indicators of patient kinematic performance in order to plan and adapt the therapy to his/her capabilities. The proposed technique exploits hand pose data, provided by the robot, and a constraint on the elbow trajectories to solve redundancy. In order to compute elbow trajectories, two additional magneto-inertial sensors located on the upper arm are used. Theoretical formulation and experimental validation of the method are presented in the paper; finally, a comparative analysis with an optoelectronic motion capture system is carried out to assess the level of accuracy and reliability of the proposed algorithm.