This week, a new video was released from Stanford Engineering and Toyota Research that shows two autonomous vehicles drifting side-by-side in close proximity.
The Stanford Engineering researchers, in collaboration with the Toyota Research Institute, achieved the world’s first fully autonomous tandem drift, with the goal of advancing the potential of AI to improve safety.
The event took place on the skid pan of Thunderhill Raceway, CA, a motorsports complex located 7 miles west of Willows, California, United States, in the Sacramento Valley.
Both cars, lead and chase, were loaded with autonomous software that had access to vehicle controls to perform the drift. While we see multiple shots of people in the vehicle, when in autonomous mode, these are simply passengers, hands by their sides and the car is doing the work.
The drift gets initiated, the slide controlled by constantly adjusting the steering angle and throttle input using only AI.
The lead vehicle was tasked with path planning and driving to its line without human input. The chase car was required to follow closely, while avoiding the lead car.
The result is a pretty stunning achievement, a showcase of what’s possible when engineers lend their minds to the possibility of AI.
Trey Weber of Stanford Engineering has a PhD in Autonomy and says they used a neural network vehicle model and instead of hand tuning this, they decided to learn through data (sound familiar?).
Chris Gerdes, Professor of Mechanical Engineering, Stanford University goes on to say they’ve gone from being skeptical about using AI, to now understanding that they couldn’t do this without it.
“We think what we have done here can be scaled up to tackle larger problems like automated driving in urban scenarios”
Drifting is a real spectacle with the noise, tyre smoke and unusual vehicle movements (sliding sideways), but it’s hoped that by learning how to control vehicles in these conditions, that controlling vehicles in safety situations will be possible.
As we know from developments by other automakers, AI and particularly computer vision combined with neural nets is helping vehicles to navigate the world around it autonomously.
The only question I have, is how fast can Toyota take these learnings and apply them to new production vehicles.
