Researchers at Stanford University used a neural network to create a self-driving system that allowed an autonomous car to learn to make high-speed turns without spinning out.
The researchers trained the neural network on data from more than 200,000 motion samples taken from test drives on a variety of surfaces.
The team then equipped a Volkswagen GTO with the algorithm and tested it on a race track. The car drove as fast as physically possible, monitoring its motion to adjust its steering and acceleration.
When driving around a turn at 50 kilometers per hour (around 30 m.p.h.), the vehicle had a low tracking error, deviating less than 50 centimeters from its desired turning path.
The researchers also found the neural network continued to work when the track was covered in snow or ice.
From New Scientist
View Full Article
Abstracts Copyright © 2019 SmithBucklin, Washington, DC, USA