Abstract: This paper deals with a design approach of a gait training machine based on a quantitative gait analysis. The proposed training machine is composed of a body weight support device and a cable-driven parallel robot. This paper is focused on the cable-driven robot, which controls the pose of the lower limb through an orthosis placed on the patient’s leg. The cable robot reproduces a normal gait movement through the motion of the orthosis. A motion capture system is used to perform the quantitative analysis of a normal gait, which will be used as an input to the inverse dynamic model of the cable robot. By means of an optimization algorithm, the optimal design parameters, which minimize the tensions in the cables, are determined. Two constraints are considered, i.e., a non-negative tension in the cables at all times, and a free cable/end-effector collision. Once the optimal solution is computed, a power analysis is carried out in order to size the robot actuators. The proposed approach can be easily extended for the design study of a similar type of cable robots.

Key words: body weight support system, cable-driven robots, design optimization, gait rehabilitation