TRI / Stanford Engineering Autonomous Tandem Drift
A Toyota Research Institute researcher describes how TRI and Stanford Engineering developed autonomous tandem drifting and what it may teach engineers about vehicle control and road safety.
In this talk
- 1:29
The challenge of the first autonomous run
The days before the experiment were the hardest of the researcher’s career. Inside the car, loud engines, tire noise, vibration, and smoke made the run an intense experience.
- 1:49
Two cars drifting together
The team demonstrated two vehicles drifting in tandem without drivers controlling them. The trailing car responded to the lead car in real time and adjusted to avoid a collision.
- 2:04
Learning vehicle behavior from data
The researchers used a neural network model trained from data instead of relying on a hand-built physics model with manually tuned settings. They found AI helped the control system improve with experience, and came to view it as essential for this project.
- 2:35
Building confidence through track testing
After years of development on a test track, the team judged the technology ready to run two cars in an autonomous tandem drift. The maneuver demands continual adjustments to steering, throttle, and brakes, since a small error can make a car spin.
- 3:00
Applying drift control to safer driving
The team sees similarities between controlling a drift and managing a car on ice or snow. They hope techniques refined in tandem drifting could eventually help drivers in difficult conditions and make roads safer.
Briefs summarise talks and interviews; timestamps open the video at that moment. Claims in a talk are the speaker's own.