Machine learning and data fusion algorithms for characterisation of muscle activity towards myoelectric prosthesis control

This exciting multi-disciplinary PhD project will develop machine learning algorithms to characterise temporal data of human muscle activity and extract temporal patterns that could be used to control robotic hand prostheses.
The successful applicant will do research in a multi-disciplinary and cross-School setting at the interface of machine learning, signal processing, and physiological human recordings. Applications of artificial intelligence to biomedical time series are expanding nowadays, and this PhD provides an excellent opportunity to be trained in this area.
We will devise machine learning methods for the analysis of temporal signals recorded via electromyography (EMG) from upper limbs. The recordings contain redundant information and patterns that we will seek to extract and exploit in novel strategies for the control of robotic prostheses. Of particular interest is the exploration of tensor decompositions as a framework to achieve parsimonious representations of patterns in EMG signals.
We have recently explored the benefits of tensor decompositions of EMG data for muscle synergy analysis, establishing a proof-of-concept for efficient and reliable extraction of meaningful muscle activity patterns associated with wrist movements. Our approach exploits the naturally multidimensional nature of EMG recordings for upper limb myoelectric control. It also provides a methodological framework to integrate information from additional sensors (such as inertial ones) to improve the extraction of patterns of the muscle activity. Therefore, we now seek to develop this approach further to better characterize the multifaceted nature of muscle activity.

To do so, we will build on recent developments by the supervisors and others in the areas of tensor decomposition, machine learning, and data fusion. We expect that this PhD will lead to developments that could eventually be transferred to the control of robotic prosthesis in the future.

Further Information: 

Early application is encouraged.
Informal queries to Dr Javier Escudero at


The University of Edinburgh is committed to equality of opportunity for all its staff and students, and promotes a culture of inclusivity. Please see details here:

Closing Date: 

Friday, January 15, 2021

Principal Supervisor: 

Assistant Supervisor: 


Minimum entry qualification - an Honours degree at 2:1 or above (or International equivalent) in a relevant science or engineering discipline, possibly supported by an MSc Degree. Further information on English language requirements for EU/Overseas applicants.

Enthusiastic and self-motivated candidates are sought with, at least, an Honours degree at 2:1 or above (or International equivalent) in electronic engineering, computer science, mathematics or cognate disciplines. An MSc qualification will be advantageous but it is not an essential requirement.
The candidate is expected to have very good programming and analytical skills.
Previous knowledge in areas related to signal processing (e.g., Fourier, time series analysis), and machine learning (e.g., matrix algebra, optimisation, regression algorithms, etc.) would be expected.


Applications are welcomed from self-funded students, or students who are applying for scholarships from the University of Edinburgh or elsewhere.

Further information and other funding options.

Informal Enquiries: