Robust and reliable localisation required for autonomous driving
Catherine Enright, Valeo Expert and Algorithm Team Manager at Valeo Vision Systems takes to the stage at AutoSens in Brussels this September to deliver a cutting-edge presentation on “Visual SLAM and localisation.” Valeo is the world’s leading provider of driving assistance systems to automotive manufacturers and Valeo Vision Systems (VVS), located in Ireland, is Valeo’s automated parking and computer vision research hub. Catherine’s current research focuses on Visual SLAM with an emphasis on the robust and reliable localisation required for autonomous driving. We caught up with Catherine to uncover more about her previous positions and her move into AI and ML.
You worked in telecommunications before moving to the area of Artificial Intelligence and Machine Learning, what was behind this move? What attracted you to AI and ML?
It is the potential of AI and Machine Learning that attracted me to the field in the first place. In the last 30+ years we have started amassing vast amounts of data in domains ranging from climate to medical and now more than ever in the social domain. The ability to learn from this data and then use this learning is so powerful.
Your PhD was on the application of probabilistic graphical models and efficient inference techniques, could you share the headline results of your research?
While learning models solely from data is incredibly powerful in my PhD I took the approach of using existing mathematical models as prior knowledge and then using the data to adapt the models. We often have prior knowledge which can be exploited to reason and learn more effectively with data, we should use this knowledge. My PhD provided a framework to use the knowledge encapsulated in the form of systems of differential equations. I applied my research to the medical domain. Using real-time data in a probabilistic framework I was able to adapt a drug delivery model to the individual patient in real-time. Along with the probabilistic framework, one of the main contributions was an adaptive-time particle filter which automatically adjusted time-steps to account for periods of rapid change and periods of slow change.
Have you drawn from your academic work in your commercial roles?
Absolutely. While my academic work was in a completely different domain, the techniques I worked on are directly applicable to computer vision and automotive. But even more valuable were the skills I acquired in problem analysis and the critical analysis of results. Having to stand up and present results to leading academics you learn quickly to question everything!
Your current research focuses on Visual SLAM with an emphasis on the robust and reliable localisation required for autonomous driving, which is also the focus of your presentation, what are your main challenges in this area?
The main challenge is developing a product that will work reliably, in all conditions. As a researcher and engineer I understand the limits physics imposes on sensors, not to mention the limits imposed by the available resources. But as a car owner I want cutting edge features that are available to use whenever I need them. So my challenge is to understand the constraints in which we work and engineer solutions to deliver functions that drivers will want to use day in day out.
Watch latest videos from AutoSens on YouTube
► Testing automotive camera modules – the difference between theory and reality
► Detectivity – Ranking Cameras for Machine Vision Tasks
► No Bit Lost – High Precision Data Recording in ADAS Development
What keeps you busy outside of work?
Family mainly but I also love the outdoors. Luckily Valeo is located on the West Coast of Ireland where we have both the sea and mountains. I am learning to sail. I can’t say I’ve mastered it but I do enjoy dinghy sailing. Racing a little dinghy in the wild winds on the Atlantic is great fun. Mountains are my other passion and Connemara provides wonderful quiet peaks you have to yourself. I am a member of the Galway Mountain Rescue team so between training and rescue operations I get to explore all areas of these mountains.
What is your favourite thing about the area of Ireland in which you live and work?
You can probably guess that from the last question – its the landscape. It is wild, varied and every changing. And in addition I get to live next to the vibrant city of Galway – there always seems to be a festival going on there especially in the summer which starts with Sea Fest, followed by the Arts Festival, then the Horse Racing Festival and Autumn is welcomed with the Oyster festival. At some stage in the year there is a Food festival, Film Festival … I really can’t keep up!
What are you looking forward to about presenting at AutoSens?
To be honest I am looking forward to hearing the other speakers as opposed to presenting myself. There is a great line up and its a wonderful opportunity to hear what other people are doing. And of course there is the venue – I have never been there but the pictures of Autoworld look so impressive, I can’t wait to walk around and get a sense of the place. It is such an amazing location for a conference.
Come and hear Catherine Enright discussing “Visual SLAM and localisation” at AutoSens in Brussels on Thursday 20 September. Book your tickets here >>
Read latest news and updates from AutoSens
- Image processing, active safety and deep learning complete a technically engaging agenda for AutoSens in Detroit for 2020 ~ 11 February 20
- IDTechEx Attends Autosens: Latest Trends in Automotive Lidars ~ 7 February 20
- From chip to city: the future of autonomous mobility ~ 6 February 20