UGRacing’s Post

This weekend Joseph Agrane and I took part in the first European Defense Tech hackathon in Munich. There were lots of great mentors, fellow hackers, and project suggestions, as well as an awesome location for the work to take place. We worked on an idea using a network of microphones to detect UAVs, completing a simulation of our solution and almost a full prototype. Huge congratulations to Benjamin Wolba, Jonatan H. Luther-Bergquist, and the team from Entrepreneur First for organising such a smooth event that pushes for solutions to such important problems!

From the classroom to the track, simulation is pushing the boundaries of EV innovation. Watch the latest episode of #DrivenBySim to see how Porsche Motorsport and UGRacing are working to engineer high-performance electric vehicles.

Over the last few weeks, the UGRacing Driverless team have completed some successful testing sessions at Knockhill Racing Circuit Ltd in Fife. Testing is always a great opportunity to get the team motivated - for all of our new recruits, this is the first time they'll have seen their code move the car! Additionally, it's an important time for knowledge transfer - the younger team members must be familiar with the car and know how to debug it. We're looking forward to demonstrating our progress at the IMechE Formula Student competition in July and seeing how far all of the other #driverless entries have come! Thanks to Stuart Gray for letting us use the facilities at Knockhill, Richard Atkinson for your help transporting the car, and Jolie Bonner for supervising us at each session. #formulastudent #ai #autonomous

Delighted that after 7 months and 40 pages of hard work, I was able to submit my dissertation today! ⭐️ Writing my dissertation, 'Evaluation of Object Detection for Autonomous Race Cars from Camera Data Using Deep Learning Methods', has been both a rewarding and challenging experience. Having had no prior knowledge of computer vision, I've really enjoyed delving into the statistics behind Convolutional Neural Networks and how they are implemented. It's been interesting to see a real-life application of the work I've done in courses throughout university, whilst also learning a lot of new concepts. The second part of my project looked at comparing CNNs to YOLO, and exploring the reasons why YOLO is preferred in the autonomous driving industry. This involved developing a YOLO V8 model which detected and classified traffic cones used at the IMechE FSAI competition, and evaluating its performance. It has been great to complete my dissertation in a topic I have heard about all through my past two years on the UGRacing Driverless Team. Being able to marry the academics I have been learning at university along with a personal interest has been extremely fulfilling, and I now feel I have a better understanding of the software used on the car, particularly for perception. I'd like to extend a huge thank you to my supervisor, Dr Xiaochen Yang, for her ongoing support in the project. I also wouldn't have been able to complete this project without the unwavering help of the Driverless Team at UGRacing, who not only sparked my interest in autonomous technology but helped to point me in the right direction throughout my project. I'd finally like to extend thanks to my friends and family, who without their support throughout my time at university as a whole, I would not have been writing this dissertation. One step closer to graduation! 🎓