Dynamic Adjustment of Physical Exercises Based on Performance Using the Proficio Robotic Arm

Physical therapy is an essential element in a comprehensive rehabilitation plan. Robot-assisted solutions have recently become more common to support the patient throughout the healing process. In this paper, we examine two techniques for assessing the movements a person performs when supported by a robotic arm. We propose DyAd, a Dynamic Adjustment system that recommends adjusted exercise configurations based on a person's performance. DyAd works with the Proficio arm, a commercial robotic arm designed for rehabilitation. The arm provides DyAd with kinematic data about the person's movement. DyAd has a graphical interface to support the user with a 3D visualization of the exercise trajectory to be followed. When a person tries to make a movement following the designed trajectory, DyAd compares this movement, the measured trajectory, to the designed trajectory. It evaluates their similarity using Dynamic Time Warping. DyAd also measures the smoothness of the movement using a modified version of the spectral arc length method. We designed DyAd to propose adjustments to the configuration of the exercise for specific users after each exercise trial. Finally, we validate the functionality of DyAd by presenting the changes in the difficulty level for users based on their performance.