Effects of Elastic Ankle Exoskeleton Stiffness and Walking Speed on Human Locomotor Performance from Whole-Body Energetics to Individual Muscle Neuromechanic

Robotic assistive devices show promise for improving gait outcomes in both healthy and impaired populations. For individuals with spinal cord injuries or post-stroke, commercial devices are now available as assistive aids and therapeutic tools for rehabilitation (e.g. EksoGT, lndego). For healthy individuals, clever assistive devices have been shown capable of enhancing performance. However, the complex mechanisms by which the assistive devices can lead to positive (or negative) neuromechanical outcomes are not yet fully understood. A human-centric approach to designing wearable devices is crucial for optimizing efficacy of wearable robotic systems. The goal of this work was to (1) determine beneficial exoskeleton assistive strategies for providing ankle assistive torque and (2) identify why certain assistance strategies are or are not effective with a core focus on the physiological response of the user.;In Chapter 1, I constructed and validated a laboratory-based robotic exoskeleton emulator capable of applying plantarflexion torques to the human ankle during locomotion. We developed a series of bio-inspired control algorithms based on an ankle stiffness ('exo-tendon'), a neuromuscular-based model, and adaptive-proportional electromyography. In unimpaired human walking, we have shown that the exoskeleton emulator provided high-fidelity impedance (stiffness) control with minimum torque error and low net power generation. This hardware and control system allows us to rapidly and systematically test a variety of exoskeleton torque assistance strategies.;In Chapter 2, we used this system to evaluate the coupled effects of speed and ankle exoskeleton rotational stiffness on human locomotor performance. In taking a human-centric approach, we evaluated the effect of exoskeleton assistance on the user from whole-body metabolic cost, joint dynamics, neural activation, down to individual muscle length and velocity dynamics. First, we hypothesized that we could find an optimal exoskeleton stiffness at each of the three speeds (1.25, 1.5, and 1.75 m s-1 ) where metabolic demand of walking would be minimized. Our results showed that the optimal stiffness of 50 Nm rad-1 reduced metabolic demand by 4.2 and 4.7% at 1.25 and 1.75 m s-1 respectively. However, no stiffness condition provided metabolic improvement at 1.5 m s-1. Evaluation of joint dynamics and muscle activation demonstrated decreased biological ankle moment and decreased soleus activation concomitant with increased exoskeleton stiffness and torque. Similar trends in joint and activation dynamics were seen across all three speeds.;In Chapter 3, I used B-mode ultrasound imaging to look "under the skin" at soleus muscle fascicles during human walking with the exoskeleton assistance. We hypothesized that exoskeleton rotational stiffness would disrupt fascicle dynamics, resulting in longer fascicles and suboptimal metabolic improvements. Results showed that the stiffest assistance, compared to no assistance, resulted in a significant 11.8% increase in fascicle length at peak dorsiflexion. At the metabolically optimal stiffness (50 N m rad-1), the fascicle length was slightly shorter (-0.6%). We found a significant positive linear relationship between metabolic demand and fascicle length. Increased fascicle lengths in walking appear to result in increased muscle economy (force per activation) in early stance, but decreased economy in late stance trends with increased fascicle velocity.;In Chapter 4, I performed a preliminary study to demonstrate the application of elastic-ankle exoskeletons to alter the biological structure in older adults and restore performance. Because older adults have lower Achilles tendon stiffness, we hypothesized that by adding parallel stiffness via an elastic-ankle exoskeleton, we would improve the muscle economy and whole-body efficiency of the older user. Data showed that exoskeleton assistance on older adults results in increased fascicle lengths, reduced muscle activation, and improved walking efficiency up to 5%.