There has been a sharp rise in the growth of wearable electronics in recent years, reaching over $186 billion globally by 2030 [1]. However, many long-term healthcare applications, e.g., epileptic seizure monitoring, microsleep detection, etc., have not been deployed due to limited energy and on-device computational power. In particular, low-power microcontrollers on smartwatches and wireless earbuds consume 1-100 mW while most flexible batteries only provide <5 mWh/cm2 and wearable bioenergy harvesters can only generate <1 mW/cm2 power [2]. Thus, there is an imminent need for a solution that can provide ultra-efficient computation while being universally deployable on various wearable/implantable platforms.
Analogue computing, such as in-memory and in-ADC (analogue-to-digital converters) computing is an emerging approach that tackles the fundamental bottleneck of all von Neumann computers, i.e., the need to move data back and forth between the computing and memory units, aka the “memory wall”. This bottleneck leads to an inherent inefficiency in power consumption and latency as the computing unit and memory bus always need to wake up from sleep mode for all the calculations. By blurring the boundary between computing and storage units, it is possible to achieve significant gains in computational efficiency. A similar effect could be observed in extremely energy-efficient human brains where memory and processing are tightly coupled with each other.
Aims: In this project, we will explore non-von Neumann approaches to overcome the “memory wall” by exploiting analogue properties of memory and ADCs to provide local computations without the need to wake the main CPU. In particular, the project will focus on the following objectives.
(1) Explore the methods to produce basic logical Boolean operations on memory or ADCs by exploiting their electrical properties and biosignals’ characteristics.
(2) Devise the algorithms to mitigate intermittent power loss on battery-less devices to maintain the correctness and progress of the computation.
(3) Develop a software-to-hardware abstraction layer and/or a compiler to provide advanced computations to software developers based on the basic logical Boolean operations.
(4) Simulate and evaluate the developed solutions on important computation tasks such as biosignal filtering and transformation, machine learning and neural network operations.
(5) Fabricate and deploy a prototyped system on practical use cases, such as epileptic seizure monitoring to study the efficiency and usability of the proposed solution.
[1] Grand View Research. Wearable Technology Market Size, 2023 - 2030.
https://tinyurl.com/4t965e3w.
[2] Yin, Lu, and Joseph Wang. "Wearable energy systems: what are the limits and limitations?." National Science Review (2023).
Contact for information on the project: Dr Nhat (Nick) Pham (phamn@cardiff.ac.uk)
Academic criteria: A 2:1 Honours undergraduate degree or a master's degree, in computing or a related subject. Applicants with appropriate professional experience are also considered. Degree-level mathematics (or equivalent) is required for research in some project areas.
Applicants for whom English is not their first language must demonstrate proficiency by obtaining an IELTS score of at least 6.5 overall, with a minimum of 6.0 in each skills component.
This project is accepting applications all year round, for self-funded candidates
Please contact the supervisors of the project prior to submitting your application to discuss and develop an individual research proposal that builds on the information provided in this advert. Once you have developed the proposal with support from the supervisors, please submit your application following the instructions provided below
Please submit your application via Computer Science and Informatics - Study - Cardiff University
In order to be considered candidates must submit the following information:
- Supporting statement
- CV
- In the ‘Research Proposal’ section of the application enter the name of the project you are applying to and upload your Individual research proposal, as mentioned above in BOLD
- In the funding field of your application, please provide details of your funding source.
- Qualification certificates and Transcripts
- References x 2
- Proof of English language (if applicable)
If you have any additional questions or need more information, please contact: COMSC-PGR@cardiff.ac.uk
Continue with Facebook