No abstract

-Symmetric ankle propulsion is the cornerstone of efficient human walking. The ankle plantar flexors provide the majority of the mechanical work for the step-to-step transition and much of this work is delivered via elastic recoil from the Achilles' tendon -making it highly efficient. Even though the plantar flexors play a central role in propulsion, body-weight support and swing initiation during walking, very few assistive devices have focused on aiding ankle plantarflexion. Our goal was to develop a portable ankle exoskeleton taking inspiration from the passive elastic mechanisms at play in the human triceps surae-Achilles' tendon complex during walking. The challenge was to use parallel springs to provide ankle joint mechanical assistance during stance phase but allow free ankle rotation during swing phase. To do this we developed a novel 'smartclutch' that can engage and disengage a parallel spring based only on ankle kinematic state. The system is purely passivecontaining no motors, electronics or external power supply. This 'energy-neutral' ankle exoskeleton could be used to restore symmetry and reduce metabolic energy expenditure of walking in populations with weak ankle plantar flexors (e.g. stroke, spinal cord injury, normal aging).Keywords-human walking, ankle exoskeleton, plantar flexors, elastic energy storage and return, passive dynamics, 'energyneutral', metabolic cost I. BACKGROUNDCoordinated ankle propulsion is a critical factor for efficient human walking. The ankle plantar flexors contribute the majority of the mechanical work done on the center-of-mass during push-off (Phase 4, 50-60% Stride) (Fig. 1). The efficiency of human gait is particularly impacted by the timing of the push-off and collision impulsive ground reaction forces [1]. For example, following stroke, propulsive impulses delivered by the ankle plantar flexors are often highly asymmetric. Push-off asymmetry results in slow walking speeds and increased metabolic energy consumption [3][4][5]. Due to the vital role of the ankle plantar flexors in shaping the normal mechanics and energetics of walking, it is essential to investigate ways in which we can improve gait impairments by focusing on aiding ankle joint push-off.Portable wearable robotic devices hold considerable promise for restoring ankle function in populations with musculoskeletal or neurological impairments. In general, current devices fall into two distinct categories (1) fullypowered [6-9] and (2) purely passive [10]. Fully-powered devices employ motors under high gain force control that can mimic the normal torque output of the lower-limb joints over the full gait cycle. Some major downsides to this approach are that powerful motors are heavy, require bulky gears and mounting frames, and rely on finite power sources that must be donned by the user. The consequence of this added mass is a marked decrease in walking economy (i.e. no metabolic savings) during assisted locomotion with portable powered devices.