Deep learning-based health monitoring electronic system with multi-wavelength optoelectronic sensor

Loughborough’s opto-physiological interaction research has led to a new generation of multi-wavelength opto-electronic sensor technology (mOEPS, www.lboro.ac.uk/carelight). mOEPS offers a flexible, biomedical monitoring engineering research platform to meet growing demands, from clinicians to individual, which outperforms all current worn smart devices for vital sign monitoring. However, the present state of opto-physiological monitoring is lacking in sophisticated, dynamic solutions to deal with complexity of individual health monitoring.

Hence this PhD aims to research how an electronics system effectively can monitor critical signs captured by the mOEPS system, integrating the smart approach of Deep Learning (DL) to cope with diversity of measurement environments.
The objective of consolidating a DL function into the present diversity of the mOPES electronic system (Smart-mOPES) is to deliver an ultra-lightweight, wearable and enhanced prototype, capable of real-time vital sign monitoring and providing better performances both at rest and during physical activity for the healthcare, sport and fitness industries.

Excellence in Science and Technology
The Smart-mOPES will be one of the outcomes derived from solid fundamental research of opto-physiological interaction against the existing lambert-Beer law based and motion artefact photoplethysmography (PPG). LU opto-physiological modelling based on physiological monitoring work is recognised as being at the forefront of worldwide new generation opto-electronic patch sensor research in the Atlas of Science. Together with a DL architecture (DoI: 10.1109/MIE.2018.2824843), fundamental research on DL will be undertaken to provide optimal solutions towards viable innovative products. The Smart-mOPES will possess miniaturised, system integrated, IoT and ultra-lightweight features with a reliable and enhanced performance to deliver one-stop solutions in critical signs monitoring. The developed Smart-mOPES prototype will be tested using a standardised physiological testing procedure to demonstrate its resistance to motion artefacts and its reliability through a designated sport physiological protocol. The Smart-mOPES prototype offers both clinicians and scientists the ability to gain a more in-depth knowledge of physiological processes within the body, as its design makes it easy to collect data continuously over long periods of time due to its ultra-light weight, comfortable, unobtrusive nature which does not impede or restrict movement.

Due to so many variations of individual physiological characters, it is unlikely to be possible to build up a specific filtering procedure for every person. To obtain better health status information, an underlying higher quality of vital physiological parameters seems critical to fulfil the requirements of healthcare and sport assessment. Hence, we are looking for an engineering architecture to establish a DL network or embedded DL algorithm capable of extracting physiological data from as many types of people as possible; this would mean that it would be smarter and more accurate than current versions. It would adapt its way of extracting data to the individual considering the contextual information and would be, overall, more efficient and noise resistant consequently.

Novelties
The project will consolidate the enabling mOPES sensing tech into a DL electronic architecture of a user prototype, leading to healthcare and sport physiological monitoring applications. The novelties of the project will be created with its unique characteristics of Smart-mOPES as follows:

1. Motion-resistant physiological monitoring through a sound fundamental of the opto-physiological interaction.
2. An intelligent physiological monitoring electronics system with a self-calibrating DL architecture, with its associated network or embedded algorithm, to be consolidated with our existing mOPES
3. Heterogeneous design to provide enhanced functionalities of Smart-mOPES and a better performance during the measurements indoors and outdoors using DL algorithms.
4. Physiological testing will be undertaken, not only to validate the heterogonous design of the Smart-mOPES, but also to gain further insight into physiological processes within the body both at rest and during physical activity through continuous, long duration, minimally intrusive measurement.

This multidisciplinary research cuts across photonics, DL algorithms, healthcare and sport monitoring applications, to lead a new research paradigm for real-time physiological monitoring and assessment.