We push forward the state-of-the-art in vehicle control at the limits of handling. We pursue expert-level performance in autonomous racing and drifting and then leverage these skills in critical situations, such as emergency obstacle avoidance. We envision a world where these tools allow each vehicle to support its driver, creating a safe and fun-to-drive experience for all.

Our team advances vehicle safety and performance capabilities by pursuing a cooperative approach to performance driving.

Our solutions leverage the latest tools in machine learning and optimal control to push state-of-the-art vehicle control at the limits of handling.

Our customers are engineers preparing the next generation of cars for the everyday driver and for motorsport competitors.

Expert-level racing and drifting are challenging tasks due to their highly dynamic and unstable nature. Despite this challenge, we need to drive the vehicle very near its absolute performance potential while adapting to changes in vehicle and environmental conditions.

Autonomous Drifting 
The main goal is to learn how we can advance technology to augment and amplify driving. We developed algorithms that can autonomously control the vehicle in critical situations where grip and traction reach their limit and are fully utilized. We want cars to have the skills of an expert driver and be able to correctly react to difficult situations like sliding on ice. When drifting in a vehicle, there are a lot of forces at play, and the driver must understand how every degree of steering angle can slow the car down. 

Handling a drift requires balancing all of these objectives to the very limits of the car's capabilities. We are looking at how to control the car in its entire spectrum of performance and build technologies that do not replace humans but rather make expert driving skills available to assist regular drivers in certain situations.

 

We are developing agents that can drive as well as the best racing drivers while also adapting to the subtle changes in total grip and cornering balance brought about by vehicle step-up changes, allowing vehicle engineers to optimize the performance of their cars efficiently using simulation tools.

 

 

 

Autonomous Tandem Drifting 
Tandem drifting, popularized in the US by Formula Drift, is a competition where two skilled drivers drift simultaneously. While the lead vehicle focuses on tracking an ideal path, or line, at a high angle, the chase vehicle has a much more complicated task. It must chase the lead car, maintaining close proximity as they drift through the course together, all while avoiding any collisions. This is a valuable problem to study for autonomy systems because it requires them to balance multiple conflicting objectives, such as avoiding collisions, staying on the road, and reacting to other vehicles in real time.

Leveraging the latest AI technology, Stanford Engineering and Toyota Research Institute demonstrated the world’s first-ever autonomous tandem drift in 2024. 

• One Model to Drift Them All: Physics-Informed Conditional Diffusion Model for Driving at the Limits
• Risk-Averse Model Predictive Control for Racing in Adverse Conditions
• Gliding on Simulated Ice: Effect of Low-μ Emulation on Drift Training
• Safe Stability Envelopes and Shared Control for Active Vehicle Safety
• An observer robustified control barrier function filter for vehicle control at the limits of handling
• Drifting with Unknown Tires: Learning Vehicle Models Online with Neural Networks and Model Predictive Control
• Learn Thy Enemy: Online, Task-Aware Opponent Modeling in Autonomous Racing