Scientists create ‘world’s most perfect car’ after using AI to pore over motor designs to make fuel-saving speed demon
A GROUP of scientists are on their way to creating the world’s most perfect car - with help from artificial intelligence.
By poring over the data of 8,000 3D car designs and their aerodynamics, the tech experts are many steps closer to designing a more efficient and eco-friendly motor.
Engineers from the have revealed their DrivAerNet++ software - which is set to revolutionise the process of designing the cars of tomorrow.
The dataset, which features over 8,000 unique car designs based on the most popular car types, such as saloons, estates and coupes, began back in 2014 - with 3D models provided by BMW and Audi.
For these designs, researchers were able to simulate how the car cuts through airflow to gain crucial aerodynamic information.
Mohamed Elrefaie, a mechanical engineering graduate student at MIT, explains: “This dataset lays the foundation for the next generation of AI applications in engineering, promoting efficient design processes, cutting R&D costs, and driving advancements toward a more sustainable automotive future.”
Read more Motors News
DrivAerNet++, which is the largest open-source dataset of its kind, acts as a vast library of realistic car designs and their aerodynamic properties, and allows for AI models to be trained quickly and efficiently.
These models, which, along with the dataset, was presented at the NeurIPS conference in December 2024, can then generate new, potentially improved designs - drastically reducing the time it takes to develop better cars.
Faez Ahmed, assistant professor of mechanical engineering at MIT, said: “Often when designing a car, the forward process is so expensive that manufacturers can only tweak a car a little bit from one version to the next.
“But if you have larger datasets where you know the performance of each design, now you can train machine-learning models to iterate fast so you are more likely to get a better design.”
Most read in Motors
The dataset also allows for the inverse process, which means taking a specific car design and assessing its aerodynamic performance.
This can then predict its fuel efficiency or electric range - without the need for costly physical prototypes.
Ahmed added: “What this dataset allows you to do is train generative AI models to do things in seconds rather than hours.
“These models can help lower fuel consumption for internal combustion vehicles and increase the range of electric cars — ultimately paving the way for more sustainable, environmentally friendly vehicles.”
This comes as incredible footage has been unearthed that shows the world's first autonomous car in action, and it's not one of Elon Musk's Teslas.
Decades before self-driving cars were unleashed to the public a 1986 vehicle was moving without a human at the wheel.
A German man brought driverless vehicles onto European roads - decades before Musks machines.
Ernst Dickmanns, a German scientist, developed a vehicle that was able to navigate roads on its own.
A group of German engineers came together in the 1980s to transform a Mercedes van, bought by Ernst.
