Where AI = Auto Innovation – Capgemini

This post was originally published on this site.

The hype around generative AI has been hard to miss but, behind the headlines, teams of tech specialists are working to test its capabilities and harnessing its power to make a positive impact in many different industries. 

One of those is the automotive industry, where Gen AI tools are enabling new ways of designing electric vehicles (EVs), in which artificial intelligence and human creativity work together to produce innovative vehicles of the future.   
Capgemini colleagues have been playing their part in this too. Capgemini’s Applied Innovation Exchange (AIE) in Mumbai, India created a GenEV Design Challenge for the students of MIT Art, Design and Technology University in Pune. The challenge was an opportunity for the students to showcase their talent and creativity, while upskilling them in using the cutting-edge Gen AI tools. 

Collaboration for innovation 

“With the GenEV Design Challenge, we wanted to bring together academia, technological innovation and industry knowledge and experience,” says Anshul Pandey, head of AIE in Mumbai. “We wanted to see how those three areas could collaborate and innovate together on the design for a product that would be fit for the purposes of the automotive industry.” 

For the project, Anshul worked closely with Vinayak Yannam, ecosystem lead at AIE in Mumbai. 

“Pune is a big automotive hub; there are lots of universities in our backyard as well as one of the top automotive design schools. We wanted to see if we could use Gen AI in the research and development and design processes of the automotive world. How would you use it to design an EV for the Indian market?” 

Vinayak says that one of the initial challenges was to change academic mindsets about the design process.  
 
“Design schools are often still very traditional,” he says. “Students hand-draw initial designs and then use AutoCAD design software to refine the design. We challenged them to rethink that and reconsider what the future of design will look like, as we know that some major auto manufacturers are already starting to use Gen AI to help design their vehicles.” 

Meeting industry standards 

Led by Vinayak, teams of students at the design school were coached and mentored for four months, gradually evolving their designs to meet industry standards and specifications.  

“We wanted to make the challenge as realistic as possible, just like it would be in a real job in the automotive industry,” says Vinayak. “We asked them to design an EV for a mid-tier segment – a family car, rather than something high-end like a Rolls-Royce. While their imaginations could run wild, the design needed to work for the market.  

“They also needed to create a presentation that was worthy of showcasing to our clients. It had to be very polished and communicate their designs clearly. It needed to demonstrate practical considerations – for example, how could you replace hard plastic parts such as the dashboard with a more sustainable material such as bamboo.”  

Anshul says the teams’ early Gen AI designs were all very different, but the mentoring process highlighted how human input could help refine them. 

“We can’t be completely dependent on technology. It can help us to come up with new and innovative ideas, but human intervention is still needed. In future, the technology will mature, but right now we still need human input to ensure we develop designs with all the right specification for manufacturing,” she said. 

The quantum test 

As part of the project the designs also needed to be tested to see if they met industry standards. This was also an opportunity to bring the students up to speed with technology, including quantum computing.   

“Testing means running simulations for things like topology optimization,” says Vinayak. “It’s about optimizing the structure of the car and testing the tensile strength of the materials it is built with. In industry, these simulations might take weeks to run. By using quantum computing algorithms, they can run much more quickly.” 

“You can also simulate things like crash testing, without needing to build a real car, and crash it to know what needs improving. By translating the car design into what we call a mesh structure – a virtual representation of the car – and running simulations using quantum algorithms, you can do this in minutes.” 

Winning designs 

From all the students’ designs, three teams were selected as finalists with one, Watt-up, chosen as the winner. “In terms of design, ultimately they combined innovative design and were the closest to the industry specifications we asked for,” says Vinayak. “They really understood that an automotive manufacturer in the world today works under a lot of restrictions, and they stuck to that, demonstrating a real understanding of the market.” 

For Anshul, the project showed how different skills and technologies need to combine to deliver design success. “Innovation, technology, design and quantum. These all needed to be integrated for the vehicle design to be a success.”  

The project also highlights how bringing together partners across academia, technology, and industry can accelerate innovation. In fact, Capgemini has partnerships with more than 30 leading institutes across India, helping develop relevant skills in technologies including Gen AI and quantum. 

As Anshul notes: “Collaborating with students like we have on the GenEV challenge allows us to be part of a real cultural shift in innovation – and that shift is happening right now.”