Communications of the ACM
AI Helps Design an ltra-Aerodynamic Bike
A spin-off of the Swiss Federal Institute of Technology in Lausanne has developed software that rapidly calculates the most aerodynamic shape for a bicycle through the application of artificial intelligence to a set of user-defined specifications.
Credit: Neural Concept
Neural Concept, a spin-off of the Swiss Federal Institute of Technology in Lausanne (EPFL), has developed software that rapidly calculates the most aerodynamic shape for a bicycle via the application of artificial intelligence (AI) to a set of user-defined specifications.
The software was trained on a convolutional neural network to calculate the aerodynamic properties of various forms embodied by generic polygon meshes.
In the first layers, the program identifies a shape's contours, and then assigns them to an object and categorizes that object according to the expected outcome.
"Our program results in designs that are sometimes 5-20% more aerodynamic than conventional methods," says Neural Concept's Pierre Baque.
The software helped a team from IUT Annecy in France design a recumbent bike for an upcoming competition, with the researchers inputting the vehicle's maximum length and width and the space needed for the drivetrain and wheels.
From Swiss Federal Institute of Technology in Lausanne
View Full Article
Abstracts Copyright © 2018 Information Inc., Bethesda, Maryland, USA
No entries found