Smart Beats Number Crunchers
Long gone are the days when car manufacturers offered "country" editions of their cars featuring front rails ("cow catchers") which led to serious injuries in the event of accidents involving other persons. Instead, it is more often the "inner values" that matter today – and these include performance in crash tests. There is scarcely a manufacturer who is not working towards the full 5-star rating. This top rating is only awarded if the protection of more vulnerable road users – i.e. pedestrians and cyclists – has also been taken into account during vehicle development, with the hood for example.
The traditional way of calculating the risk of injury to pedestrians or cyclists through impact with the hood is based on the finite element method (FEM) (and hardware testing), which simulates how loads, forces and boundary conditions affect the component and the person involved in the impact. However, this method is both complex and lengthy – not least because it calls for "number crunchers", i.e. high-performance data centers.
An alternative using a smart solution instead of massed computing power has been developed by EDAG: artificial intelligence based on a neural network. The white paper "The Smart Route to Five-Star Tests", which you can download here, reveals what the "AI Crash" tool can do and what advantages it offers. You will also find information on other areas in which the AI framework could be put to use.