There has been much debate on autonomous or ‘self-driving’ cars and how they will impact society. One of the concerns is over how well the computer controlling them will cope with changing conditions on the roads. A team at the University of Liverpool believe they have a solution – a VR training system for autonomous vehicles.
The ALEAD system was devised by a team from Future Coders, CGA and the University of Liverpool, who received a grant of £1 million (GBP) from the The Centre for Connected and Autonomous Vehicles (CCAV) to create the technology.
The system uses information gathered from Liverpool’s road, including vehicle density, hazards, road layout and junctions. It then uses machine learning to predict typical driving scenarios to give self-driving car systems a save environment to learn in.
The technology makes it both cheaper and safer to train self-driving cars, and can be used to simulate a variety of events or extreme weather conditions, such as snow, fog, debris in the road or vehicles moving erratically or unpredictably.
Jon Wetherall, Managing Director of CGA Simulation, explains how Space Ribbon, an intergalactic racing car game, helped inspire them: “Our team at CGA has many years’ experience working in the creative and innovative world of games development. The console and virtual reality games we’ve created, like Space Ribbon, helped us visualise how we could take emerging technologies and apply them to autonomous vehicle technologies” He added: “We hope that by combining our love of gaming, with our skills in simulation and futuristic technologies, we’ll help make autonomous vehicles a reality on Liverpool’s roads.”
Max Zadow, CEO of Future Coders said: “We are very proud to be working as partners on the exciting and groundbreaking ALEAD project, as it highlights the cutting edge tech work that’s coming out of Liverpool now.”
The team say they will be exploring commercial applications for the technology once the project is complete. VRFocus will brig you further news on ALEAD and other VR projects as it becomes available.