Ensuring the Safety of Autonomous Vehicles Through Proven Guarantees

Sunday - January 21 2024, 11:14 UTC - 10 months ago

Autonomous vehicles, such as driverless cars and planes, are becoming more common but raise safety concerns. Current testing methods may not uncover all potential flaws, prompting researchers like Sayan Mitra to develop a more reliable method. Mitra's team focuses on ensuring the reliability of a vehicle's perception system, which provides information to the control module for decision making. By quantifying uncertainties and creating mathematically proven guarantees, Mitra's method can ensure the safety of autonomous vehicles.

Driverless cars and planes are no longer the stuff of the future. In the city of San Francisco alone, two taxi companies have collectively logged 8 million miles of autonomous driving through August 2023. And more than 850,000 autonomous aerial vehicles, or drones, are registered in the United States — not counting those owned by the military.

But there are legitimate concerns about safety. For example, in a 10-month period that ended in May 2022, the National Highway Traffic Safety Administration reported nearly 400 crashes involving automobiles using some form of autonomous control. Six people died as a result of these accidents, and five were seriously injured.

Traditional testing methods for addressing safety concerns in autonomous vehicles can be unreliable and may not uncover all potential flaws. Sayan Mitra, a computer scientist at the University of Illinois, Urbana-Champaign, has developed a method for proving the safety of lane-tracking capabilities for cars and landing systems for autonomous aircraft. This method is now being used to assist in landing drones on aircraft carriers, and Boeing plans to test it on an experimental aircraft this year.

Mitra and his team focus on ensuring the reliability of the vehicle's perception system, which feeds raw data from cameras and other sensors to machine learning algorithms based on neural networks. These algorithms re-create the environment outside the vehicle, and this information is then sent to the control module for decision making. While the control module relies on well-established technology, the perception system introduces uncertainty into its re-creation of the vehicle's surroundings.

The key to Mitra's strategy is quantifying the uncertainties involved, known as the "error band" or the "known unknowns." By determining the amount of error introduced by the perception system, Mitra's team can create mathematically proven guarantees that safety requirements are satisfied.