Subject tailored gait pattern planning for robotic gait rehabilitation

Abstract

Robotic is gaining its popularity in gait rehabilitation. Gait pattern planning is an important as to ensure that the gait pattern provided by the robot onto the patient is natural. It is known that cadence, stride length, and walking speed are the key factors of gait pattern. Moreover, a systematical methodology for gait pattern planning in terms of these factors is necessary. The present work proposes a methodology to generate waveforms of lower limb joint angles during walking, given a set of cadence and stride length. The walking motion has been captured with a motion capture system by using passive markers. The waveforms are then decomposed into Fourier coefficients. The decomposition reduces the amount of data for analysis by representing the entire waveform with eleven Fourier coefficients. A function for Fourier coefficients in terms of cadence and stride length are established with a multiple linear regression. The final constructed waveforms with the Fourier coefficients computed with the function are presented together with the original experimental waveforms. The final waveforms closely match the experimental waveforms.