Embedded Systems UCC
Nominated Award: Best Application of AI to achieve Social Good
Embedded.Systems@UCC is an award winning group pursuing research in a variety of domains which better the humanity. Our motto is Engineering Inspired by Life. Since its inception in 2005, the group was reward ed by more than 50 awards and distinctions for its high quality work. In fact it is the only group which won 3 times in the IEEE/IBM smarter planet challenge with projects such as improving healthcare access for disadvantaged communities, building toys which can interact with children regardless of their disability or even catch Santa, building smart beehives using the latest advances in tech. AI is an enabler for these applications and not even cloud is the limit.
The Edge is where we thrive. Recently the team prides itself in some major advances towards building a universal stethoscope, some sort of Star Trek tech, which can be used to listen to the heart or even brain and can detect congenital heart disease or brain seizures. State of the art machine learning is at the core of these algorithms and the inference models are optimised for tiny devices while still keeping very high accuracy of detection. Lower cost and ease of use would make such a device more used in providing improved care in disadvantaged communities and not only.
Recently the group was also the recipient of a major award in urban planning and design for a project looking at air quality for cyclists, by using low cost sensors. Irish weather is also an inspiration and after building a wind directed orchestra of miniaturised wind turbines (that can compete with any AI-driven music synthesis) we also created some (Irish) weather modulated protocols for communications using AI at the Edge which also help us understanding better the environment by providing high quality data at a frac tion of power consumption cost. Securing the Internet of Things using hardware and AI is also increasingly important and the group was awarded recently a major 1st prize in an international competition. While all these projects are very different or exotic, they have one common feature and that is humans empowered by AI.
We are proud of our people and also proud to be part of some amazing centres including CRT-AI, INSIGHT and Marei.
Reason for Nomination:
As described in the profile, the group prides itself in some outstanding projects which use AI. Here is a list of some of the recent award winning projects in the group; security.ai: using a scalable ultra low power inference for detecting anomalies in the power supply of IoT devices using ARIM A/CNN/FPGA. By using our methodology we are able to detect anomalies in computation, anomalies in inference and anomalies in networking and communications. This project won 1st prize in the Elektor/Maxim AI and Electronics International Competition, 2021.
chd.ai: using XG-Boost for detecting congenital heart disease. The algorithm and the (open) paper won the best engineering paper at UCC, 2021. Parts of the work are also published in the leading conference in the area of biomedical electronics (IEEE EMBC). Currently the inference models are being implemented on the EDGE hardware towards a Universal Stethoscope.
neurobell.ai: using AI (namely convolutional neural networks), a number of sonification algorithms for an improve d seizure detection and their implementation on resource constrained platforms. Best paper award in IEEE ISSC conference, state of the art implementations on low power microcontrollers
comms.ai: using AI at the EDGE together with Lora communication protocols to drastically reduce power consumption for communications while keeping or improving the quality of data. I EEE ISSC 2021 publication.
iBEES: using AI in the beehive, analyse bees behaviour and predict weather(rain or no rain) with astonishing accuracy. Published in a top journal in Electronics in Agriculture. Research lead also to the formation of a successful start-up
u-play and Unique: intelligent interfaces using AI for interacting with toys for children with disabilities (using EoG, EMG, IMU signals). Eventually we made toys to eve n dance Riverdance. 2nd prize IEEE/IBM smarter planet challenge, top demos in conferences and outreach.
The breath of application, recognition through international awards already won and the ultimate goal of achieving social good make the reason for this nomination for such prestigious awards. We hope that you also find this team unique.
Additional Information:
Group website: sit es.google.com/site/embedded0101
Publications website: https://scholar.google.com/citations?hl=en&user=_6jD76IAAAAJ
Youtube channel:
https://www.youtube.com/watch?v=LFGj_rw40oQ&list=PL4NKSX7HxQjWy5Tm9GOIOGbaK3uKHVKLp&index=1
https://www.youtube.com/watch?v=b0YBn1mEMJM&list=PL4NKSX7HxQjWy5Tm9GOIOGbaK3uKHVKLp&index=2
https://www.youtube.com/watch?v=X8fdvc8NSd0&list=PL4NKSX7HxQjWy5Tm9GOIOGbaK3uKHVKLp&index=3
https://www.youtube.com/watch?v=-PiB8ongzOY&list=PL4NKSX7HxQjWy5Tm9GOIOGbaK3uKHVKLp&index=4
https://www.youtube.com/watch?v=SUsKNU-_Skw&list=PL4NKSX7HxQjWy5Tm9GOIOGbaK3uKHVKLp&index=11
https://www.youtube.com/watch?v=u5tztrGPqBg&list=PL4NKSX7HxQjWy5Tm9GOIOGbaK3uKHVKLp&index=7