The aim of gait rehabilitation is to achieve independent ambulation after neurological injuries, such as stroke. Various technologies, including treadmill-based robots and wearable exoskeletons have been employed to address the challenges of regaining walking function. However, research on motor learning has suggested the need for a paradigm-shift towards personalized therapy, by which the robotic action is tailored and adapted to the patient’s needs. 

In detail, most treadmill-based rehabilitation robots assist the user to move according to a predefined trajectory, which is played back over time. These trajectories are usually based on recorded gait data from healthy people. However, the major drawback of pre-recorded trajectories is the lack of flexibility, especially in case of rigid mobilization exercises.  

In this project, to improve the personalization of gait rehabilitation exercises, you will first review available strategies to generate gait trajectories for locomotion, model physiological gait, and finally develop a synthetic trajectory-generation algorithm that creates parameterized gait profiles based on physiological measures. Finally, the algorithm is implemented (MATLAB/SIMULINK) in a treadmill-based gait rehabilitation robot (Lokomat, Hocoma AG, Switzerland). Further improvements can be introduced with online adaptation of these trajectories in terms of foot clearance, step length, step duration, walking speed, etc. according to the current status of the patient.