The ACT Government, needs your help to improve autonomous vehicles. In collaboration with Seeing Machines, known for their monitoring & intervention technologies & services for the automotive, mining, transport & aviation industries, a new research project will the test how quickly a driver can take back control of an automated vehicle. At the Sutton Road Driver Training Centre in Canberra, phase 1 of the program, known as CAN Drive, focuses on driver behaviour and leverages the latest monitoring technology for face and eyes.
Project Manager Nicolas Riquelme features in the video that shows off the in-car technology, added to a Tesla Model S. Being on a dedicated track, they can challenge drivers in unique ways safely. At low speed, the CAN Drive program tracks the user’s face, watches their eye movements and even monitors brain activity while being presented with challenges. These may be mobile alerts or obstacles that would represent a time the driver needs to respond and take back control. By understanding the time requirement for an average driver, we start to have a benchmark for what autonomous auto makers should be aiming for in the temporary phase of semi-autonomous driving.
As good as Tesla’s Autopilot is, we know from international events, there are fringe cases that the technology still has issues. One of the biggest issues is wildlife, like kangaroos that don’t move like other vehicles, or humans on foot or on bikes and scooters. Kangroos are random, they sometimes jump towards headlights and often are found in groups. This is an incredibly difficult thing to program for and Kangaroos are just one example of wildlife that could mean humans need to resume driving, at least for now.
In a few years (maybe sooner) level 5 autonomy will be here, the point where automakers are comfortable that their hardware/software combination can accommodate all the randomness found in the world.
If you’re interested in participating, then you need to email firstname.lastname@example.org.