Zac Vawter

Targeted Muscle Reinnervation

Thought Control

• 96 electrodes in a cylindrical grid in the socket, approx. 25 mm apart
• measurements of signal intensity and location taken on healthy limb and applied to amputated limb
• so the same way he would control a biological leg, he can control a bionic leg
• gait styles (ambulation modes) each have different joint stiffness and power:
• level ground
• up/down ramp (10 degree slope)
• up/down stairs w/ reciprocal gait (leg over leg)

Mechanical Stresses

Weight, Response, and other issues

• multiple motors for articulated knee and ankle makes the leg heavy
• less muscle to control leg + heavy leg = problems
• 4.7 kg is weight now
• notice the leg is very stripped down
• all excess material has been removed
• leg reacts/ can react faster than an actual human limb
• misinterpretation of signals
• difficulty getting signals: all electrodes must remain in constant, close contact with stump without becoming painful or moving very much
• leg is noisy, and most amputees don't want huge amounts of attention on their missing limb
• expense

• peak torque at the ankle joint of approximately 1.6 Nm/kg in a very small amount of time (±0.2 s for a walking rate of 1 step/s)
• average 0.35 J/kg of mechanical energy per step
• generated power at push-off has peak of approximately 3.5 to 4.5 W/kg
• Considering a 75 kg subject
• max torque output approximately 120 Nm
• power output between 250 and 350 W
• these are average values
• Vawter's data is protected by patient privacy
• joints can be stiff (no elasticity), compliant (elasticity used to store energy), pneumatic (uses opposing nonlinear forces), and/or electric (a type of compliant with multiple motors)
• physics and calculus far above my level are involved in figuring out how to make this work

• used for upper limb amputations to control bionics
• reattaches nerves that would go to the missing limb back to remaining muscle
• similar procedure used in lower limb amputations to prevent nueroma formation (this is what Vawter received)
• allows electronic signals used by nerves to be detected on the outside of the skin
• picked up by sensors in the socket of the leg
• interpreted by computers in the leg based on previous data
• allows for immediate response to thought, so the fake leg reacts like a real one

Programming and Processing Data

• pattern recognition based on other patients without TMR
• during first gait tests, transitions between gaits were controlled remotely by researchers
• mechanical sensors now partially control transition
• every attempted motion has a distinct electrical signal
• 90% accuracy with just ankle and knee flexion/extension
• 92% accuracy also including leg rotation
• average accuracy in non-TMR amputees: 91.0% and 86.8% respectively
• about 40% reduction in error rate
• ambulation error rate for only mechanical sensors: 12.9%
• with EMG from reinnervated muscles: 1.8%
• No buttons or hip movements to control transitions
• many mistakes are not noticeable or not critical errors, such as mistaking an uphill ramp for stairs
• each gait has different stiffness and torque
• large errors common when transitioning to stair climbing drastically reduced with EMG data
• control also possible during non-weight-bearing activity

Zac Vawter

• lost leg above the knee in motorcycle accident in 2009
• first user of bionic leg
• received surgery during amputation including nerve transfers to prevent neuroma formation