A groundbreaking achievement was announced by Toyota Research Institute (TRI) and Stanford Engineering: two cars were autonomously drifted in tandem, marking a world first in driving research.
Over the course of nearly seven years, the teams have worked together on research aimed at enhancing driving safety. These experiments focused on automating the motorsports maneuver known as “drifting,” which involves precisely controlling a vehicle’s direction after breaking traction by spinning the rear tires—an essential skill for recovering from a slide on snow or ice.
The addition of a second car drifting in tandem has enabled the teams to more accurately replicate dynamic conditions where cars need to swiftly react to other vehicles, pedestrians, and cyclists.
“Our researchers came together with one goal in mind – how to make driving safer,” said Avinash Balachandran, vice president of TRI’s Human Interactive Driving division. “Now, utilizing the latest tools in AI, we can drift two cars in tandem autonomously. It is the most complex maneuver in motorsports, and reaching this milestone with autonomy means we can control cars dynamically at the extremes. This has far-reaching implications for building advanced safety systems into future automobiles.”
“The physics of drifting are actually similar to what a car might experience on snow or ice,” said Chris Gerdes, professor of mechanical engineering and co-director of the Center for Automotive Research at Stanford (CARS). “What we have learned from this autonomous drifting project has already led to new techniques for controlling automated vehicles safely on ice.”
During an autonomous tandem drifting performance, two cars—a lead vehicle and a following vehicle—maneuver through a course at times in close proximity while operating at the brink of control. The team utilized advanced methods to develop the car’s AI, which included a tire model based on neural networks, enabling it to learn from experience, much like a skilled driver.
“The track conditions can change dramatically over a few minutes when the sun goes down,” said Gerdes. “The AI we developed for this project learns from every trip we have taken to the track to handle this variation.”
Over 40,000 people die in car accidents in the US, and approximately 1.35 million die worldwide each year. Many of these incidents occur due to a loss of vehicle control in sudden, unpredictable situations. Autonomy offers great potential to assist drivers in responding appropriately.
“When your car begins to skid or slide, you rely solely on your driving skills to avoid colliding with another vehicle, tree, or obstacle. An average driver struggles to manage these extreme circumstances, and a split second can mean the difference between life and death,” added Balachandran. “This new technology can kick in precisely in time to safeguard a driver and manage a loss of control, just as an expert drifter would.”
“Doing what has never been done before truly shows what is possible,” added Gerdes. “If we can do this, just imagine what we can do to make cars safer.”
Two modified GR Supras were put to the test at Thunderhill Raceway Park in Willows, California. The lead car’s algorithms were meticulously developed by TRI to ensure consistent and safe performance. Meanwhile, Stanford engineers focused on creating AI vehicle models and algorithms for the chase car, allowing it to dynamically adapt to the lead car’s motion and drift alongside without collision.
GReddy and Toyota Racing Development (TRD) made significant modifications to each car, including the suspension, engine, transmission, and safety systems, to match the specifications used in Formula Drift competitions. These carefully executed experiments were conducted with expert drivers in a controlled environment, providing invaluable data for the teams.
The vehicles are equipped with advanced computers and sensors that enable them to autonomously control their steering, throttle, and brakes while also monitoring their motion, including position, velocity, and rotation rate. Crucially, they are connected through a dedicated WiFi network, allowing real-time communication to exchange information such as relative positions and planned trajectories.
To achieve autonomous tandem drifting, the vehicles continuously plan their steering, throttle, brake commands, and the intended trajectory using a cutting-edge technique called Nonlinear Model Predictive Control (NMPC). Each vehicle begins with specific objectives, represented mathematically as rules or constraints that it must adhere to.
The lead vehicle’s objective is to sustain a drift along a desired path while complying with the laws of physics and hardware limits, such as maximum steering angle. The chase vehicle’s goal is to drift alongside the lead vehicle while proactively avoiding collisions.
Both vehicles continually solve an optimization problem multiple times per second, making decisions about steering, throttle, and brake commands that best align with their objectives and adapt to rapidly changing conditions. By utilizing AI to continuously train the neural network with data from previous tests, the vehicles make improvements with each trip to the track.
Journal reference:
- T. P. Weber and J. C. Gerdes. Modeling and Control for Dynamic Drifting Trajectories; IEEE Transactions on Intelligent Vehicles, 2024; DOI: 10.1109/TIV.2023.3340918
- T. Kobayashi, T. P. Weber, D. Mori and J. C. Gerdes. Trajectory Planning Using Tire Thermodynamics for Automated Drifting. IEEE Intelligent Vehicles Symposium (IV), 2024; DOI: 10.1109/IV55156.2024.10588753
- Nicholas Drake Broadbent, Trey Weber, Daiki Mori, J. Christian Gerdes. Neural Network Tire Force Modeling for Automated Drifting. arXiv, 2024; DOI: 10.48550/arXiv.2407.13760
- Jonathan Y. M. Goh, Michael Thompson, James Dallas & Avinash Balachandran. Beyond the stable handling limits: nonlinear model predictive control for highly transient autonomous drifting. Vehicle System Dynamics, 2024; DOI: 10.1080/00423114.2023.2297799