The Relationship Between Stiffness and Walking
Open Access
- Author:
- Schmitthenner, Dave
- Graduate Program:
- Mechanical Engineering
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- December 14, 2020
- Committee Members:
- Anne Elizabeth Martin, Dissertation Advisor/Co-Advisor
Anne Elizabeth Martin, Committee Chair/Co-Chair
Bo Cheng, Committee Member
Daniel Humberto Cortes Correales, Committee Member
Joseph Paul Cusumano, Outside Member
Andrew Michael Geronimo, Special Member
Daniel Connell Haworth, Program Head/Chair - Keywords:
- Exoskeletons
Biomechanics
Nonlinear Control
System Identification
Gait - Abstract:
- Leg stiffness is an important factor in human walking. Understanding the ways in which stiffness impacts walking could lead to improvements in gait rehabilitation, exoskeleton design, and exoskeleton control. Three studies were conducted to better understand the effects of stiffness on walking. The first study investigated how changing stiffness of the foot itself affects walking. To do this, a rigid foot plate was attached to the foot of 16 subjects in order to increase the stiffness of the foot. This is because many exoskeletons and orthoses use a rigid foot plate, but this may affect the natural mechanics of the foot. The resulting spatiotemporal parameters, kinematics, and kinetics were compared with normal walking. It was found that the foot plate did not have an effect on spatiotemporal parameters, but did decrease the ankle range of motion and decreased the intersegmental motion of the different parts of the foot. It was also found that a foot plate that ends before the toes has less of an effect than a foot plate that extends to the end of the toes. Therefore, exoskeletons and orthoses should implement a jointed toe segment or a shorter foot plate in order to increase the transparency of the device. For the second study, an ankle exoskeleton was developed with the requirement that it was capable of a versatile range of control modes, and was robust. Versatility is important for an exoskeleton because the type of rehabilitation a person may require can vary due to a number of factors. The exoskeleton must also be robust in uncertain environments because humans undergoing rehabilitation may act unpredictably. Therefore, a pneumatically powered exoskeleton was designed and controlled using sliding mode control. This exoskeleton is capable of position, force, and impedance control. Four subjects were used to test the performance of these controllers while following position and force trajectories. The tests showed that regardless of the behavior of the wearer, the exoskeleton acted appropriately and under an acceptable degree of error. This demonstrates that sliding mode control can be used to make exoskeletons capable of the versatility required for rehabilitation. Lastly, in the third study, the control of leg stiffness while walking was investigated. This was done by using human data from 12 subjects to extract system models and leg stiffness controllers for walking using three different methods and informed by the spring-loaded inverted pendulum model. The first method used ordinary least squares to linearize the system about a Poincaré section; the second linearized the spring-loaded inverted pendulum model and combined that with the results from the first method. The third method used the sparse identification of nonlinear dynamics procedure to identify linear combinations of nonlinear functions that describe the system. These methods were evaluated based on how well they predicted experimental data, and based on how well they controlled a spring-loaded inverted pendulum model walker in simulation. Results showed that all three methods resulted in similarly accurate models and controllers. Therefore, using human data in conjunction with data driven identification techniques may be used to reveal the mechanics of gait or to design robotic controllers for walking.