Characterizing Optimal Performance of a Passive Elastic Ankle Exoskeleton during Human Locomotion

Perhaps because humans are already so well-tuned for locomotion, no autonomous, wearable device intended to assist walking or running has succeeded in reducing metabolic energy consumption for healthy individuals during typical walking conditions. The ankle provides the majority of positive mechanical work during walking and much of this work is delivered via elastic recoil from the Achilles' tendon, which may serve as an energy savings mechanism. The 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 exoskeleton was designed to be as transparent to the user as possible having minimal interference with gait kinematics and lightweight enough to minimize the metabolic penalty of adding mass to the user. The exoskeleton provides plantarflexion torque during stance, and does not interfere with toe clearance during swing. To do this a lightweight custom composite frame and two clutches that can engage and disengage a parallel spring based only on ankle kinematic state was developed. The primary system is purely passive containing no motors, electronics or external power supply. A secondary clutch has an additional low power, servomotor to control the timing of engagement of the clutch, which still passively provides assistance but is more versatile and can handle dynamically changing gait (e.g. increases in speed, asymmetry due to impairment).;To test the validity of our exoskeleton design a variety of studies of individuals walking in many conditions with and without the exoskeleton was performed, on a split belt instrumented treadmill. Kinetic, kinematic, electromyography, and oxygen consumption and carbon dioxide expiration were recorded during all trials. Initial testing of the exoskeleton suggests the utility of the clutch, to act in series with the parallel spring. Results indicate the clutched exoskeleton design addresses all three of our design criteria: (1) it does not hinder natural gait kinematics, (2) it is lightweight enough so that added mass has minimal effect on net metabolic energy consumption, and (3) it can produce significant plantarflexor torque assistance during stance, but does not resist toe clearance during swing.;As human adaption studies to a passive exoskeleton have never been performed, user adaptation to the exoskeleton was studied. Regression fits to indirect calorimetry data indicate that users began to decrease their metabolic energy use below normal walking after ~18.5 min of training with intermediate exoskeleton parallel spring stiffness. These data also suggest that users could decrease their metabolic energy consumption even more than the values we reported with additional training.;An in depth view of the neuromechanics and energetics of walking with a passive elastic exoskeleton was performed after adaptation. Study participants (n=9) reduced their net metabolic power by 7% below normal walking with an intermediate spring stiffness after 28 minutes of walking in the exoskeleton. Kinetic analysis indicates the exoskeleton offloads plantar flexor muscle forces during stance and assists in plantarflexion, which might be key to reducing metabolic energy expenditure. Electromyography (EMG) data indicate that reductions in plantarflexior muscle activation (e.g. soleus) plays a role in decreased net metabolic power, but increases in dorsiflexor (e.g. tibialis anterior) EMG activity at high parallel spring stiffness confounds this reduction and ultimately leads to an increase in net metabolic power for the stiffest exoskeleton parallel springs. Future studies with direct muscle level measurements will be necessary to identify the exact mechanism of energy savings.