Aubrey James Roche
Cognitive Approaches for Multimodal Sensor Data Perception
Aubrey is originally from Ireland and is currently studying a PhD under the Institute for Digital Technologies. Prior to his studies at Loughborough, Aubrey studied at several universities across Ireland.
Aubrey holds a DipEng in Process control & Instrumentation from Carlow Institute Of Technology (Ireland), a BEng in Mechatronics from Dublin City University (Ireland), a MSc in Bioengineering from Trinity College Dublin (Ireland) and an MSc in Mobile Internet from Loughborough University London.
Aubrey has accumulated over ten years of work experience with companies such as Irish Rail, Schering Plough Pharmaceuticals and Help International (NGO). Most recently, Aubrey has worked as a Consulting Forensic Engineer providing expert witness evidence to the Irish Courts. During this time, Aubrey completed numerous CPD programmes in accident investigation and project management
PhD research description
Driverless vehicle technology is currently being developed and tested around the world. Automakers, technology companies, and research companies are amongst the many stakeholders investing in this vehicular revolution.
The technology and sensors that make self-driving cars possible already impact the way we drive. The first steps towards hands-free driving have been taken by advancements like Parking Assist by VW and Parktronic by Mercedes-Benz. One area that appears to have been overlooked is networking autonomous vehicles. Networking of such vehicles could be seen as introducing another layer into an already complex system. Elements like security would of course be a concern as they are with so many emerging technologies. However, there are multiple benefits to balance those concerns, such as improved road user safety.
One key concern for road users, pedestrians and cyclists alike is safety. For cyclists and pedestrians, synchronised sensor data from autonomous vehicles would greatly assist in accident prevention. Research has shown that high collision numbers occur at signal-controlled junctions with obstructed visibility. Networking the output from sensors would enable autonomous vehicles to create a real-time virtual map of the road. Combining sensor outputs (LIDAR, Near Field Vision, Radar, GPS, Ultrasonic Rangefinders) from several cars would reduce the possibility of cyclists or pedestrians emerging undetected into the path of a vehicle.
PhD supervisors
Professor Ahmet Kondoz and Dr. Varuna De Silva
Awards, grants or scholarships received
Aubrey was awarded the Wolfram Studentship in October 2016.
Interests and activities
Aubrey is interested in engineering, mechanics, electronics and programming embedded controllers. In Ireland, he was involved in the setup of a social enterprise that focused largely on different applications for embedded controllers and their benefits for local communities. Aubrey also enjoys reading, kayaking and travelling.