Artificial intelligence is set to take a key role in the design and engineering of new cars, dreaming up lighter, stronger and more complex structures than humans can envision.
Just as computing power exceeds the mathematical capability of the human mind, smart software capable of innovation and problem solving is set to push product development into new territory.
Hack Rod, a team of designers, engineers, geeks, Hollywood insiders and stunt drivers is working on a way to harness the power of artificial intelligence in tandem with powerful design software produced by Autodesk.
The team started by gathering data through a rudimentary chassis. Photo: Supplied Experimenting with connectivity surrounding the emerging "internet of things", the Hack Rod crew built a basic sports car, fitted it with dozens of race car-like sensors, and set about testing, racing and crashing the vehicle. They then fed millions of data points into a computer powered by NVidia processors capable of machine learning, and asked Autodesk's "Dreamcatcher" software to take that information and use it to design a better car.
Mike Geyer, Autodesk's director of evangelism and emerging technology, says the generative design software uses cloud computing systems to create the best geometric shapes for any given project. While some shapes are man-made, others have a natural, flowing look more akin to the biology of plant and bone structures.
The idea is that you can give a computer a set of parameters dictating the desired size, weight and materials of the finished product, and that it will use sheer mathematical power to draw up the most suitable structure.
"This stuff is pretty exciting," Geyer says.
"We said 'Ok, Dreamcatcher, tell us what a generative design chassis would look like'.
"I define the problem and the cloud computing system uses these cloud computing systems to come up with the geometry. You get these really wild solutions that we as humans wouldn't think of ? these spider web-like designs that might be significantly lighter or optimised for new materials.
"We're kind of unlocking a new era in automotive design that is going to yield incredible results."
The first Hack Rod prototype created using generative design resulted in a chassis that is 35 per cent lighter than the team's engineers first managed.
Built using a tubular frame similar to a bicycle or the roll cage of a race car, the design's intricacy was limited by existing manufacturing techniques.
AI-powered software then used the data to design a better chassis. Photo: Supplied - Cassidy-Rae Wilson The next Hack Rod will more closely follow the computer's ideal design by using 3D printing to realise Dreamcatcher's vision for a lightweight sports car chassis. Geyer says the team will then cloak it in carbon-fibre bodywork impregnated with electronic sensors that collect complex data surrounding the vehicle's performance, which will then be fed back into the computer's design process.
The implications for artificial intelligence in design are enormous.
"You could you have a fleet of vehicles that not only drive themselves but design themselves, or at least play a part in the design process," Geyer says.
Dreamcatcher's sports car vision, which uses organic shapes to design lighter, stronger engineering solutions. Photo: Supplied - Cassidy-Rae Wilson "You could play out this dystopian future where humans are out of the loop but we play such a huge part that it's more about using AI as a better tool, a better shovel.
"The involvement of AI allows people involved in the design and engineering process to work on higher order problems. As AI and generative design becomes more online the engineer and the designer will spend more time examining the problem question, understanding the constraints.
"It will be more like human-aided design than computer-aided design."
Source: www.bing.com