Man-machine interfacing based on ultrasound wearable technology for controlling upper limb prostheses

Lead Research Organisation: Imperial College London
Department Name: Dept of Bioengineering

Abstract

With up to half of active myocontrol prosthesis users reportedly abandoning their prosthetic device it is important to use resources to consider and research new methods of control that may offer a better user experience and encourage fewer users to stop using their prostheses. This requires more reliable, accurate and stable control systems, which this project will explore.

The aim of the research project is to develop novel methods for the control of upper-limb prostheses using ultrasound of the remnant muscles as the control signal. The aim is for the ultrasound technique to improve upon tried and tested methods using surface EMG to decode neural signals from the spinal cord used to control the musculoskeletal system. These electrophysical signals are limited by their poor spatial resolution, their shallow penetration and their dependence on the interface (e.g. their degradation in quality due to sweating). The project will therefore focus on similar techniques to decode spinal cord activity using the mechanical signal from ultrasound. Ultrafast ultrasound, developed by one of the supervisors of the project, will be used as it works on the required time scales at the required resolution and depth. With this new method, the aim is to be able to see activities of deeper muscles that sEMG cannot and to, with the higher resolution, decode the firing of individual motor neurons which has previously not been possible. Initial parts of the PhD project involve fusing techniques from the research groups of the two main supervisors, then running initial feasibility testing with laboratory sized ultrasound equipment. After this, the system will be downscaled for use in a wearable prosthetic device.

The research carried out in this project clearly falls within the EPSRC's research area of assistive technology, rehabilitation and musculoskeletal biomechanics. In particular, the project aims to vastly improve the capabilities of non-invasive techniques to extract musculoskeletal information, preventing the need for invasive surgeries and direct brain interfaces that may be suggested alternatives for control signals. This supports ambition H5 of the EPSRC, which aims to promote minimally invasive yet effective systems that are able to restore function.

Publications

10 25 50

Studentship Projects

Project Reference Relationship Related To Start End Student Name
EP/S023283/1 01/04/2019 30/09/2027
2295839 Studentship EP/S023283/1 01/10/2019 30/09/2023 Emma Lubel