In a groundbreaking collaboration, Toyota Research Institute (TRI) and Stanford Engineering have achieved a remarkable milestone in autonomous driving: the world’s first fully autonomous tandem drift sequence. This achievement, while seemingly a stunt for motorsport enthusiasts, has profound implications for the future of driving safety.
For nearly seven years, TRI and Stanford Engineering have been at the forefront of research aimed at making driving safer. Their latest success involves automating a complex motorsport maneuver known as "drifting," where a driver expertly controls a vehicle after breaking traction by spinning the rear tires. This skill is crucial for managing vehicle control in challenging conditions like ice or snow. Now, by adding a second car to the drift sequence, the teams have simulated dynamic driving scenarios where vehicles must respond quickly to other cars, 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. This has far-reaching implications for building advanced safety systems into future automobiles."
A Deep Dive into the Physics of Drifting
Drifting isn't just about showing off; it's about mastering control at the vehicle's limits. The physics involved in drifting are similar to what a car might experience on snow or ice. "What we have learned from this autonomous drifting project has already led to new techniques for controlling automated vehicles safely on ice," explained Chris Gerdes, professor of mechanical engineering and co-director of the Center for Automotive Research at Stanford (CARS).
In an autonomous tandem drifting sequence, two vehicles—a lead car and a chase car—navigate a course while operating at the edge of control. This requires sophisticated AI systems and a deep understanding of vehicle dynamics. The team used a neural network tire model, which allowed the AI to learn from experience, much like an expert driver. The AI developed for this project adapts to changing track conditions, ensuring the vehicles can handle variations in real-time.
Implications for Driving Safety
Car crashes claim over 40,000 lives annually in the U.S. and about 1.35 million worldwide. Many of these incidents are due to loss of vehicle control in sudden, dynamic situations. Autonomous driving technology holds the promise of assisting drivers in these critical moments, potentially saving countless lives.
"When your car begins to skid or slide, you rely solely on your driving skills to avoid collisions," added Balachandran. "An average driver struggles in these extreme circumstances, and a split second can mean the difference between life and death. 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."
The Technical Marvel of Autonomous Tandem Drifting
Experiments were conducted at Thunderhill Raceway Park in California, using two modified Toyota GR Supra. TRI developed the algorithms for the lead car, while Stanford focused on the chase car. Each car's suspension, engine, transmission, and safety systems were modified to Formula Drift specifications. Equipped with advanced computers and sensors, the cars could control their steering, throttle, and brakes, while constantly communicating via a dedicated WiFi network to exchange crucial data on positions and planned trajectories.
The core of this technology lies in Nonlinear Model Predictive Control (NMPC), which allows the vehicles to plan and re-plan their steering, throttle, and brake commands up to 50 times per second. This ensures that the cars can respond to rapidly changing conditions and maintain control even at the limits of their performance.
Leveraging AI for Real-World Applications
At the heart of this autonomous driving technology is a sophisticated AI system. The AI learns from every trip to the track, continually improving its performance. By using a physics-informed neural network, the team blended data-driven methods with physical models to achieve greater accuracy and robustness.
This approach is crucial for safety-critical applications. The AI models need to predict real-world behavior accurately to ensure reliable performance. To build confidence in these models, the team collected extensive data, validated their training processes, and developed strategies to handle different conditions, such as varying track temperatures.
A Safer Future on the Horizon
The successful demonstration of autonomous tandem drifting is a significant step forward in automotive safety. While your future car may not drift its way to the supermarket, the underlying technology developed in this project could revolutionize how vehicles respond to dangerous situations.
"If we can do this, just imagine what we can do to make cars safer," Gerdes concluded.
This pioneering work by TRI and Stanford Engineering exemplifies the potential of AI and advanced vehicle dynamics to create a safer driving experience. As we look to the future, these innovations may well be the key to reducing accidents and saving lives on the road.
