Sign In

Communications of the ACM

ACM TechNews

Accident Prevention Software for Autonomous Vehicles


View as: Print Mobile App Share: Send by email Share on reddit Share on StumbleUpon Share on Hacker News Share on Tweeter Share on Facebook
The software tracks other vehicles, with an eye toward avoiding accidents.

Researchers at Germany's Technical University of Munich (TUM) say they have created accident-preventing software for autonomous vehicles that anticipates variations of traffic scenarios.

Credit: LV4260/iStock

Researchers at Germany's Technical University of Munich (TUM) say they have created accident-preventing software for autonomous vehicles that anticipates variations of traffic scenarios.

The software analyzes and predicts events while driving, recording sensor data and re-assessing the traffic situation each millisecond, calculating all possible movements for every traffic participant (provided they comply with traffic regulations).

The system projects three to six seconds ahead and from this ascertains movement options for the vehicle, while simultaneously calculating potential emergency maneuvers.

TUM's Matthias Althoff said testing the software in simulations proved "that the safety module does not lead to any loss of performance in terms of driving behavior, the predictive calculations are correct, accidents are prevented, and in emergency situations the vehicle is demonstrably brought to a safe stop."

From Technical University of Munich (Germany)
View Full Article

 

Abstracts Copyright © 2020 SmithBucklin, Washington, DC, USA


 

No entries found