Highly dexterous control of upper limb prostheses

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


The aim of the project is to develop and test control algorithms for active upper limb prostheses. For establishing a reliable interface with the user, we will investigate the potential of both invasive and non-invasive EMG sensors. The focus will be on the information transfer rate for the mapping from EMG signals into wrist/hand kinematics. The transfer rate will be analyzed first for the state of the art control strategies based on invasive and non-invasive EMG. This analysis will provide the
baseline foundation for the new solutions. Further, new control strategies will be developed, with the key addition with respect to the state of the art of including machine learning adaptation. This will allow the optimal co-adaptation of the user
with the upper-limb prosthetic device. In addition to novel machine learning approaches for unsupervised adaptation, the role of artificial sensory feedback in user's learning will also be explored.

The final aim is the development of a man-machine interface for wrist/hand prosthesis control that allows the proportional and simultaneous activation of multiple degrees of freedom, in a natural and intuitive way for the user and with adaptation to nonstationary conditions, such as changes in posture and fatigue. The project will be translational, covering all aspects from the algorithmic design, prototype development, and test, in realistic operational conditions in patients.


10 25 50

Studentship Projects

Project Reference Relationship Related To Start End Student Name
EP/N509486/1 01/10/2016 31/03/2022
1975944 Studentship EP/N509486/1 01/11/2017 30/04/2021 Martyna Stachaczyk
Description The most significant achievement that has been accomplished as a result of the work funded through the EPSRC grant so far is the discovery of the universality of the peripheral neural information ('neural code') in humans, i.e. proving the fact that at the (peripheral) nervous system level, humans control their movements according to the same paradigm.
This finding is crucial for a very effective clinical translation of dexterous control of neural interfaces, such as neural upper limb prosthetic devices, but is not limited to lab-and-clinics applications (i.e. it can be employed in any day-to-day neural interfaces that can be controlled by human movement, such as gaming devices and smart housing environments).
The initial findings leading to the publication regarding the major results discussed above (being submitted to Nature Biomedical Engineering, March 2020), are the research on multi class detection and tracking of movement (DOI: 10.1109/NER.2019.8717077) and adaptive neural and myoelectric signal processing (DOI: 10.1109/ICORR.2019.8779482).
Exploitation Route At the final stage of my doctoral research project I aim to create an open GitHub repository and upload all my algorithms for further development by interested individuals/research groups.
The findings provided by my PhD can be taken forward and applied in control systems of various neural interfaces ranging from clinics to BMI applications.
My research on peripheral neural correlates of volitional action was one of the first of its kind in the field of Human-Machine Interfaces, thus it can (and should) be followed up with further validation and exploration of the subject.
Sectors Communities and Social Services/Policy,Digital/Communication/Information Technologies (including Software),Electronics,Healthcare,Leisure Activities, including Sports, Recreation and Tourism

URL https://ieeexplore.ieee.org/abstract/document/8717077
Description The results and application of my findings is considered for the implementation by the CTRL-labs company which has recently been purchased by Facebook.
First Year Of Impact 2020
Sector Digital/Communication/Information Technologies (including Software),Healthcare
Impact Types Cultural,Societal