An electric vehicle with artificial intelligence (AI) sensors and computers is set to embark on a 1,200 kilometer (745 mile), three-month journey in Queensland, Australia.
The zero-emissions Renault ZOE will be used to map roads in the state, which is in the northeast of the country. Researchers from the Queensland University of Technology (QUT), which is based in Brisbane, will man the car.
"As researchers drive the car across Queensland, onboard sensors will build a virtual map to help refine AI-equipped vehicles to drive safely on our roads," Mark Bailey, Queensland's minister for Transport and Main Roads, said in a statement Wednesday.
Bailey added that while it was "early days", AI technology and smart road infrastructure had the potential to transform the way people travelled in Queensland and "reduce road trauma."
The project will look at how the vehicle and its AI system adapts to lane markings, traffic lights and street signage. Additionally, it will look to overcome GPS systems' limitations "in built-up areas and tunnels for vehicle positioning."
Michael Milford, from QUT's Australian Centre for Robotic Vision, said that as the vehicle was driven, AI would "watch and determine if it could perform the same as a human driver in all conditions."
While there is a great deal of excitement surrounding autonomous vehicles, there is still a large amount of work to be done.
Milford added that early testing had shown that something as inconsequential as a paint spill on a road could confuse a self-driving AI system.
"Past studies, along with initial experiments conducted by QUT, show how automated cars have difficulties on rural roads which can lack lane markings," he explained.
"A motorist on a rural road knows to stick on the left or imagine there is a line in the middle of the road. People will also cross the imaginary line to go around obstacles, it's quite difficult for an automated vehicle to do this."
Millard went on to explain that the "primary goal" of the project would be to look at how current advances in both machine learning and robotic vision enabled the research car to "see and make sense of everyday road signage and markings that we, as humans, take for granted."
