Lower limb exoskeletons are gaining prevalence to augment mobility of people with gait impairments. However, current exoskeletons using serial elastic actuators do not have the torque and bandwidth capabilities to assist in balance control in daily activities. Considering the critical role of mediolateral stability in balance control of individuals affected by neurological and age-related gait impairments, we propose an abduction-adduction exoskeleton for mediolateral assistance. The exoskeleton utilizes harmonic gear driven actuators with active torque feedback to ensure high bandwidth and admittance control capabilities, designed as an assistive device for enhanced balance. The exoskeleton is developed to be able to provide assistance up to a bandwidth of 5 Hz at the hip to individuals weighing up to 90 Kgs. Benchtop testing of the exoskeleton showed closed loop velocity bandwidth of more than 5 Hz while under loading with a user weighing 85 Kgs. Further, analysis of admittance controller using both static and dynamic testing showed that the developed controller can simulate stiffnesses up to 95 Nm/rad with an R 2 of over 99%. Pilot human testing individuals showed that the exoskeleton can modulate step width behavior of individuals while ensuring compliance with the users, facilitating seamless human robot interaction. The implication of the developed exoskeleton for balance control in elderly and neurologically impaired populations, and future avenues of research are discussed.

Abstract-Using wearable robotics to modulate step width in  normal walking for enhanced mediolateral balance has not been  demonstrated in the field. We designed a bilateral hip exoskeleton  with admittance control to power hip abduction and adduction to  modulate step width. Objective: As the first step to show its potential, the objective of  this study was to investigate how human’s step width reacted to  hip exoskeleton’s admittance control parameter changes during  walking. Methods: Ten non-disabled individuals walked on a treadmill at  a self-selected speed, while wearing our bilateral robotic hip  exoskeleton. We used two equilibrium positions to define the  direction of assistance. We studied the influence of multiple  stiffness values in the admittance control on the participants’ step  width, step length, and electromyographic (EMG) activity of the  gluteus medius. Results: Step width were significantly modulated by the change  of stiffness in exoskeleton control, while step length did not show  significant changes. When the stiffness changed from zero to our  studied stiffness values, the participants’ step width started to  modulate immediately. Within 4 consecutive heel strikes right  after a stiffness change, the step width showed a significant change.  Interestingly, EMG activity of the gluteus medius did not change  significantly regardless the applied stiffness and powered  direction. Conclusion: Tuning of stiffness in admittance control of a hip  exoskeleton, acting in mediolateral direction, can be a viable way for controlling step width in normal walking. Unvaried gluteus  medius activity indicates that the increase in step width were  mainly caused by the assistive torque applied by the exoskeleton. Significance: Our study results pave a new way for future  design and control of wearable robotics in enhancing mediolateral  walking balance for various rehabilitation applications.