Computer scientists at the University of Illinois Urbana-Champaign developed of method to ensure end-to-end safety of autonomous vehicles (AVs) by guaranteeing the results of the machine learning algorithms used to inform them.
The strategy focuses on making sure the AV's perception system is reliable by quantifying the "known unknowns" using a perception contract.
The researchers found that safety can be ensured by keeping the AV within a specified range of uncertainty, provided the determination of that range is accurate.
From Wired
View Full Article
Abstracts Copyright © 2024 SmithBucklin, Washington, D.C., USA
No entries found