Design, program and manufacture of a smart below-knee prosthetic limb
Group Members
Rishabh Arora, Iason Avlonitis, Nathan Nile, Clinton Olateju, Konstantinos Pitsillos, Gregory TidanianSupervisors
Professor Liudi Jiang, Dr Andrew ChiperfieldSupporters
Dr Piotr Laszczak, IBMThis project involves the design, program and manufacture of a smart lower-limb prosthesis, with the intention of restoring functionality to people with below-knee (transtibial) limb loss. 51% of all amputations are below-knee (transtibial), and the solution to restore the functions of these amputated limbs is to use prostheses. The market is currently dominated by systems with non-motorised joints. These systems can only perform basic gait functions at lower walking velocities compared to those of people without limb loss.
The use of motorised joints was chosen as clinical trials show that they allow for preferred walking velocities similar to those of people without limb loss. The novel aspects of this project are an electromyography (EMG), sensor-based control system, a motorised gear system, deep learning via a neural network, and a mobile app. The control system involves EMG to move the foot and gives the user more control. A lightweight servo motor has been employed to reduce weight, with torque amplified by using a gear system. The IMU sensor can detect ground slope and adjust motor force accordingly. A neural network has been designed to sense and adjust walking speed in real time. A mobile app has been compiled to display key information and control specific movement to the prosthesis.
The performance was evaluated by walking on a treadmill with the sensors attached to the muscles intended to be used by a person with limb loss on their residual limb. The prosthesis responded to muscle signals and was held in place on a demonstration rig. Signal processing and material testing was also carried out.
The use of motorised joints was chosen as clinical trials show that they allow for preferred walking velocities similar to those of people without limb loss. The novel aspects of this project are an electromyography (EMG), sensor-based control system, a motorised gear system, deep learning via a neural network, and a mobile app. The control system involves EMG to move the foot and gives the user more control. A lightweight servo motor has been employed to reduce weight, with torque amplified by using a gear system. The IMU sensor can detect ground slope and adjust motor force accordingly. A neural network has been designed to sense and adjust walking speed in real time. A mobile app has been compiled to display key information and control specific movement to the prosthesis.
The performance was evaluated by walking on a treadmill with the sensors attached to the muscles intended to be used by a person with limb loss on their residual limb. The prosthesis responded to muscle signals and was held in place on a demonstration rig. Signal processing and material testing was also carried out.
