The tech underlying these new self-driving vehicles is NVIDIA's DRIVE computing platform. It uses the latest in artificial intelligence concepts, including neural networks, which are basically computer science's way of modeling decisionmaking after the way the human brain works. The network is able to learn and improve as it goes by making new connections. Machine learning tasks being handled by neural networks include computer vision, which can use a combination of sensors and cameras to help a computer, in this case the car, figure out what's going on in an environment to navigate through it.
The companies are already demonstrating a version of DRIVE, the DRIVE PX 2, in the autonomous Audi Q7 shown above. The crossover is able to figure out its own path and can sense and drive on different types of surfaces, including pavement, grass, and dirt, plus it can navigate the cones of a simulated construction zone while reading dynamic detour signs.
In 2018, Audi will expand testing of its autonomous vehicles on California public roads. The manufacturer has been permitted to test cars in the state since 2014. In announcing the expansion, Audi makes a point to mention it intends to follow applicable laws to the letter, which seems like more than a veiled reference to the trouble Uber found itself in recently when it didn't quite meet testing requirements in San Francisco.
Audi has also promised Level 3 autonomy from the 2018 A8, which will feature a system called Traffic Jam Pilot to control the steering, throttle, and brakes at speeds below 35 mph. The company's previous autonomous work includes a self-driving RS7 track car named Bobby, which our own Jonathan Buckley got to race against in the Translogic episode below.