Injuries to the neuromusculoskeletal systems often result in muscle weakness, abnormal coordination strategies, and gait impairments. Functional resistance training during walking—where a patient walks while a device increases loading on the leg—is an emerging approach to combat these symptoms. While simple passive devices (i.e., ankle weights and resistance bands) can be applied for this training, rehabilitation robots have more potential upside because they can be controlled to treat multiple gait abnormalities and can be monitored by clinicians. However, the cost of conventional robotic devices limits their use in the clinical or home setting. Hence, in this dissertation, we designed, developed, and tested passive and semi-passive wearable exoskeleton devices as a low-cost solution for providing controllable/configurable functional resistance training during walking.


We developed and tested two passive exoskeleton devices for providing resistance to walking and tested their effects on able-bodied participants and stroke survivors. First, we created a patented device that used a passive magnetic brake to provide a viscous (i.e., velocity-dependent) resistance to the knee. The resistive properties of the device could be placed under computer control (i.e., made semi-passive) to control resistance in real-time. Next, we created a passive exoskeleton that provided an elastic (i.e., position-dependent) resistance. While not controllable, this device was highly configurable. Meaning it could be used to provide resistance to joint flexion, extension, or to both (i.e., bidirectionally). Human subjects testing with these devices indicated they increased lower-extremity joint moments, powers, and muscle activation during training. Training also resulted in significant aftereffects—a potential indicator of therapeutic effectiveness—once the resistance was removed. A separate experiment indicated that individuals often kinematically slack (i.e., reduce joint excursions to minimize effort) when resistance is added to the limb. We also found that providing visual feedback of joint angles during training significantly increased muscle activation and kinematic aftereffects (i.e., reduced slacking).


With passive devices, the type of passive element used largely dictates the muscle groups, types of muscle contraction, joint actions, and the phases of gait when a device is able to apply resistance. To examine this issue, we compared the training effects of viscous and elastic devices that provided bidirectional resistance to the knee during gait. Additionally, we compared training with viscous resistances at the hip and knee joints. While the resistance type and targeted joint altered moments, powers, and muscle activation patterns, these methods did not differ in their ability to produce aftereffects, alter neural excitability, or induce fatigue in the leg muscles. While this may indicate that the resistance type does not have a large effect on functional resistance training during walking, it is possible that an extended training with these devices could produce a different result.


Lastly, we used musculoskeletal modeling in OpenSim to directly compare several strategies that have been used to provide functional resistance training to gait in the clinic or laboratory setting. We found that devices differed in their ability to alter gait parameters during walking. Hence, these findings could help clinicians when selecting a resistive strategy for their patients, or engineers when designing new devices or control schemes.



Date: Thursday, May 27, 2021

Time: 10:00 AM

Zoom: https://umich.zoom.us/meeting/register/tJIufumrrDgtHd3z5Jg3Y_BG4ZC70OPrjTjk (Zoom link requires prior registration)

Chair: Dr. Chandramouli Krishnan