Crutch gait pattern for robotic orthoses by the use of feature extraction

Abstract

A robotic orthosis has been developed for spinal cord injury patients which consists of an exoskeleton robotic suit with actuators at the hip and knee joints. In this article, a presentation is made on how to generate a gait pattern to control the robotic orthosis. First, the features needed to characterize a gait pattern are defined, and then a description is given of how to extract features from the normal gait patterns contained in experimental encoder data on the hip and knee joints of a person walking normally. Next, the physical meaning of these features in a gait pattern is discussed, and methods of varying the walking speed and step size through the adjustment of features are then considered. It is shown that the desired gait pattern can be obtained by connecting these features with spline interpolation. Finally, a crutch gait pattern is produced using these features, and this is applied to the exoskeleton robot ROBIN-P1 in order to verify the proposed method.