Synthesis of perturbations for gait rehabilitation robots

Perturbations are an important measure to improve gait training and to investigate the the learning success of stroke patients during rehabilitation. This paper describes the development and implementation of algorithms for the synthesis of perturbations during gait training on a robotic walking simulator. The perturbation algorithm performs an automatic on-line adaptation of the robot motion override. Hence, it is independent from the type of motion interpolation which the robot motion is based on. Perturbations we have implemented so far include stumbling, i.e. hitting an obstacle during swing phase, and slipping, i.e. sliding on an obstacle during stance phase. An additional catch-up control algorithm ensures that both feet will regain synchronous motion again after the event. The algorithms have been implemented in the control software of the HapticWalker, a programmable footplate based human walking simulator for gait rehabilitation, which we have developed. So far a number of healthy subjects have successfully tested the algorithms on the simulator. Clinical trials will follow.