Title: Neural Driven Motion Synthesis for Assistive Technology
Supervisor: Dario Farina
Host institution: Imperial College London Project summary (abstract style):
Unintuitive and inefficient control schemes of electronic lower limb Prosthetic and Orthotic (P&O) devices limit the number of cases where they are a worthwhile intervention. Furthermore, commercial control algorithms are challenged by uneven terrain and turning during walking. Muscle signals contain information about the intended upcoming motion, which is required for responsive control strategies. Predicting the motion directly from these signals is prone to producing unfeasible or unsafe movement. Instead, an abstract representation of the underlying intent can be first estimated from muscle signals, from which stable motion is synthesised in a separate step. This project investigates data-driven, hierarchical control schemes targeting lower limb P&O devices.
Project summary (biography style):
My project investigates the requirements for synthesising walking motion. I’m quantifying the available information from muscle signals for lower limb device control, and I’m looking for hierarchical systems, intent representations and control strategies that increase the reliability and versatility of the generated movement.
Tabletop gaming, hiking, doodling