GM patents autonomous learner vehicle that doesn’t need a driving instructor
Technology would see learner’s inputs compared with what the autonomous vehicle thinks is correct.
American car giant General Motors has filed a patent for an autonomous vehicle that can teach you how to drive.
The move would allow learners to jump in a vehicle without an instructor and receive training on public roads.
In the filing, the description of the technology indicates that the autonomous vehicle will be constantly interpreting the world around it, making its own decisions about the correct inputs, then comparing this with what the driver does.
This includes looking at their steering and pedal inputs as well as external factors such as speed and position in the road compared with other vehicles.
An algorithm can then determine a score for each of the factors and feed this back to the driver as well as third parties. The filing appears to imply that the vehicle can take over one or more aspects of driving if the score drops below a certain threshold.
For example, if the car registers the driver is braking too late, it could help to slow the car, or move it back to the centre of the lane if they are drifting across.
It can also start out by giving new drivers control over some inputs and introduce more as their competency increases.
The benefits of such a system are said to be a reduced cost for those learning to drive, as well as increased availability of driver education, as the vehicles could be available 24/7.
Lower cost and improved availability and safety are obvious on paper, but it remains to be seen whether regulators buy into the idea of teaching people to drive with autonomous cars, while learners themselves might not feel comfortable alone in the car.
Regardless, autonomous cars are not currently legal to drive anywhere in the world outside of very specific testing circumstances, so autonomous learner vehicles will be some way off yet.
