UC Berkeley

Powered lower-extremity exoskeletons have traditionally used four to ten powered degrees of freedom to provide ambulation assistance for individuals with spinal cord injury. Systems with numerous high-impedance powered degrees of freedom commonly suffer from cumbersome walking dynamics and decreased utility due to added weight and increased control complexity. This work proposes a new approach to powered exoskeleton design that minimizes actuation and control complexity through embedding intelligence into the hardware. Two novel, minimally actuated exoskeleton systems (the Austin and the Ryan) are presented in this dissertation. Unlike conventional powered exoskeletons, the presented devices use a single motor for each exoskeleton leg in conjunction with a unique hip-knee coupling system to enable their users to walk, sit, and stand. The two types of joint coupling systems used are as follows.

The Austin Exoskeleton employs a bio-inspired mechanical joint coupling system designed to mimic the biarticular coupling of human leg muscles. This system allows a single actuator to power both hip and knee motions simultaneously. More specifically, when the mechanical hamstring and rectus femoris of the exoskeleton are activated, power from the hip actuator is transferred to the knee, generating synchronized hip-knee flexion and extension. The coupling mechanism is switched on and off at specific phases of the gait (and the sit-stand cycle) to generate the desired joint trajectories. The device has been proven to be successful in assisting a complete T12 paraplegic subject to walk, sit, and stand.

The Ryan Exoskeleton (also called the Passive Knee Exoskeleton) uses dynamic joint coupling. Dynamic joint coupling refers to a method of generating knee rotation through deliberate swinging of the hip joint. This minimalistic system is the first powered exoskeleton that weighs less than 20 pounds and has a compact form factor that more closely resembles a reciprocating gait orthosis than a conventional exoskeleton. The Passive Knee Exoskeleton has been validated by several SCI test pilots with injury levels ranging from T5 to T12. The lightweight, ambulation-centric assistive device have been tested to be able to comfortably reach an average ambulation speed of 0.27 m/s and have demonstrated high levels of maneuverability. The dynamic joint coupling paradigm has been proven to be effective especially for newly injured individuals who have not yet developed significant amounts of joint contracture or sustain high levels of spasticity.

Overall, this dissertation focuses on the design and operation of the Austin and Ryan Exoskeletons.