Focus on research: Dr Eoin O’Connell, Confirm
Dr Eoin O’Connell is a lecturer at the Department of Electric & Computer Engineering at UL and a funded investigator with Confirm, the Science Foundation Ireland centre for smart manufacturing and Industry 4.0. In this interfview he talks about being a homebird and the next generation of wireless technology.
Unlike a lot of academics you’ve kept your career in Ireland. Was that a conscious decision?
The best way to describe it is I’m a homebird. My family are all Limerick city. I went to school in Limerick, I went to the College of Art, Commerce and Technology (COACT) which became became Limerick Institute of Technology. Then I went to work in Dublin doing nights and it was horrific, so I ended up getting a job in Johnson & Johnson in Limerick and I still didn’t like working nights, so I applied to UL and worked there at that same time – so I had to stay in Limerick. Then I married a Limerick woman. I can’t escape it.
Was there any particular area that brought you back to college?
I’m a technophile. I love finding out how stuff works. I hate looking at a black box and not knowing what’s happening inside. I ask a lot of my students if they have ever taken apart a machine and not been able to put it back together because that’s how I started out. And when I broke things I thought ‘now I need to know how to not break them’.
As an engineer I make things that talk to each other – like a database talking to a VR environment to create a visualisation tool. It’s not that my interests are diverse, it’s that they’re different verrsions of the same thing.
My PhD was in fibre optic sensors but it was actually an early example of the Internet of Things. It was a fibre optic sensor connected to a mobile phone network that would send a message to people that were managing a water vessel with a sensor, telling them when it was ecologically unsound. That was an early example of an environmental sensor that was active instead of people having to go out to the field and take a sample and bring it back to a lab.
Nowadays everybody has sensors and the problem is with so many out there that it becomes a scaling issue. When you have thousands of sensors, you need a big computer, when you have a big computer you need a big data centre, when you need a big data centre you need high bandwidth.
Do you see this evolution as a natural intertwining of what’s out there or a case of people looking for applications without knowing what they might be?
There’s a difference between knowledge and data. The biggest problem with sensors is you get loads of data. The celeverness comes from being able to turn that data into knowledge. As more and more things connect that means you get more and more data and that means you have to apply better algorithms, better ways of rationalising the data. That’s where the evolution has come in the way of AI and machine learning – what you and I would have called simple coding years ago is now ‘AI’ and ‘machine learning’.
People are talking about ChatGPT and you can get that AI engine to write code for you. All you have to do is tell the AI engine ‘I want you to create a visualisation tool of this data. So it’s democratising the technology. Say what you want, the technology should be able to do it for you. You have to move with the times, so that’s the evolution of it.
Another string you have to your bow is your experience of project management. We often hear about the lack of ‘soft skills’ in engineering and tech. Is that still a bit of a stereotype?
It’s interesting because in order to become a chartered engineer you have to have a demonstrable ability to communicate. There’s no point in doing research is you can’t commuicate.
There’s a difference between communication and dissemination – communicating to the masses and disseminating to your team. Engineers are usually very good at talking to engineers, which is why as engineers we need marketing people in our research centres to actually get our message out and show what we’re doing is of value.
There is a challenge for some engineers to get their point across in a language that allows everyone to move along with them. I can talk for hours on the minutae of something but you might not be interested in it, you have to keep it high level. I always find a good engineer is someone who can relate an issue or topic of interest to the person they’re talking to. You have ot know your audience.
Another problem with a lot of engineers is that presentatinos skills are a skillset in itself. To be able to get your message across in a way that is engaging is very hard. We’ve all been in classes with teachers who do ‘death by powerpoint’ but if you don’t get that level of engagement or conversation, you don’t get interest. If you don’t get interest yu’re not creating engineers, you’re creating automatons.
You’ve dealt with industry quite a bit. Do you see yourself as a bridge between it and acadmeia?
Very much. Academics are very good at doing the basics well. We can take a problem, discern what needs to be done and communicate it back in a reported structure. Industry can be about networking, getting access to the talent pipeline as well as the core research – these are ‘unspecified values’.
The biggest problem between bridging the gap is to get across the point that industry works in quarterly cycles and academics work in annual cycles. A four-week sprint for industry could be the equivalent of a 15-week semester. That disconnect between timelines is a big issue and you have to get that out in the open at the very beginning.
I’m doing a project with a company whose CEO is ringing us every week for updates and every week we have to have something to tell him and show explicitly what we’re doing. If that were an academic project is would be every month. The timelines are basically a multiplier for academia.
What areas are you excited about at the moment?
I do a lot of work in the area of wireless technology. Milimetre technology is something I see as a go-to space. At the moment 5G in Ireland is high speeds on relatively low frequencies. The 3.5-4.5GHz range. As we move up through the spectrum that will mean much higher speeds.
What happens when you go up the wireless spectrum is there are engineering trade-offs. What you gain in speed you lose in signal penetration. This is why in the past mobile phones were so big because the antenna had to be bigger to receive the signal.
As we moved up to 5G the antennas got smaller. Now we are talking about moving up the spectrum up to 300GHz and as we do that phones and sensors and other technology will becomes smaller and smaller to the extent that they’ll be called ‘sensing dust’.
You could have a situation where the nail on your little finger could hold hundreds of sensors. That’s an area that excites me. How to get data from those sensors to a database in real time or as close to as possible and have it fed into AI engines so they can make decisions – there will be so much more data and decisions to be made from it as we move up the radio spectrum and with that comes better compute power, less power consumed, which means everything will become environmentally friendly. But there will be questions about sustainability, scalability and security. There are lots of challenges as we move into the future.