Need for Speed Tokyo Drift: Hotwheels Edition (A Drifting Control Case Study)

Providing safety guarantees for automotive system controllers operating is an important and difficult problem across a wide range of industries. However, most approaches to automotive controllers don’t have formal verification, ignore adverse conditions, or both. A particularly dangerous yet common adverse driving condition is the presence of rain or snow, resulting in a loss of traction with the ground. As an initial step in studying systems that operate after losing traction, we model and analyze the specific motion of drifting, the intentional loss of traction between the tires and the ground, and design controllers that choose a safe time to drift at a rate that reaches a desired degree of turn. Using Keymaera, a theorem prover for differential dynamic logic (dL), we provide a formal proof of the safety guarantees afforded by our controllers of not drifting off the road and facing the desired direction.