Toyota Taught Self-Driving Cars To Drift In Tandem

Researchers from Toyota and Stanford University put two AI-controlled racing cars into synchronized drift. The goal is to make self-driving cars safer.

One of the leaders of the Center for Automotive Research at Stanford, Professor Chris Gerdes, explained that the physics of drift is similar to the behavior of cars on snow or ice. Balachandran adds that the technology developed can intervene to prevent loss of control of the vehicle. The system analyzes an array of parameters 50 times per second to generate control impulses for steering, gas and brakes.

The modified Toyota Supra cars use AI, which is further trained with each drive. TRI developed control algorithms for the lead car, and the Stanford School of Engineering developed AI models for the pursuit car to ensure it follows the leader without collisions. The vehicles communicate via a Wi-Fi network that was configured by specialists from GReddy and Toyota Racing Development.

According to Balachandran, drifting two cars in tandem is an “important milestone.” He believes this has “far-reaching implications for the development of advanced safety systems” in future passenger cars.