Developing a Robotic Spasticity Simulator with a Robust Control Strategy to Mimic Various Joints

Abstract

Spasticity assessment has been criticized by its problem of inter-rater reliability. This lack of reliability compromises the assessment and evaluation of the spasticity grade. In order to address this issue, robots that mimic spastic human joint have been proposed by different studies to assist in training of therapists. In which these robots are designed specifically for certain joints and specific characteristics, such as fixed inertia and fixed limb length. Our goal is to provide a more versatile robotic device that can be easily adjusted for different joints and models to recreate feeling of different spastic human joints. A torque tracking control scheme based on time delay estimation was implemented. Time-delay estimation was implemented in the control scheme for dynamics cancellation of system (feedback linearization), and a Proportional-Integral-Derivative (PID) control is responsible for torque tracking. Providing a robust control method. Elbow and knee joint models and different spasticity levels were tested. Results have shown a satisfactory performance of proposed controller for elbow and knee joints to track the torque provided by the implemented joint model. We were able to successfully mimic spasticity levels from MAS 0 up to MAS 3 for elbow joint, and from MAS 0 to MAS 2 for knee joint.