Spinal motor neurons control of redundant robotic system

Supervisors

Dario Farina (Bioengineering)
Etienne Burdet (Bioengineering)

This project will develop non-invasive decoding of spinal motor neurons activity from high-density electromyographic (EMG) signals, implement and investigate voluntary control enabled by the extracted motor commands, and test these techniques for the control of a robotic system with multiple degrees-of-freedom (DOF). The first supervisor has pioneered the use of high-density EMG and dedicated signal processing to tackle the activity of individual motor neurons at spinal level in a non-invasive way. This opens up completely new possibilities for real-time, non-invasive control of prostheses, whose practical use by amputees suffers from the very restricted control capabilities of current systems. This project will first develop robust online deconvolution techniques to extract the neural spike trains of the output layers of the spinal cord from high-density EMG signals. It will then design and investigate control strategies to effectively map the decoded neural signals into intuitive commands for robotic devices, including the development of suitable sensory feedback and training protocols. Implementation of the control on a robotic device with multiple DOF will test the limits of the capabilities of the developed techniques for the control of complex movements with a redundant mechanism and for interactive robotic control adapting the mechanical impedance.

Student

Mario Bracklein