Millions of laps against yourself: Di Grassi wants AI twins for Formula E drivers

Lucas di Grassi wants Formula E to embrace artificial intelligence - and has an unusual idea for how it could help drivers. A virtual twin could mirror a driver's style, techniques and preferences, complete millions of simulated laps and then become a training partner. But the 2016/17 Formula E title winner insists AI should not replace humans. Instead, he wants it to make the championship even more driver-centred.

Lucas di Grassi may have ended his active Formula E career, but the Brazilian is still thinking about the championship's technological future. During a media round in London attended by e-Formula.news, the 2016/17 title winner outlined how he believes artificial intelligence could play a role in electric motorsport.

The subject is not entirely new territory for di Grassi. Alongside his racing career, he also spent time as CEO of Roborace, the autonomous electric racing project in which software and artificial intelligence played a central role.

"We need to start working right now on how to implement AI into Formula E in many different areas," says di Grassi. One of his most concrete ideas is a virtual twin of the driver. "One of the things AI could do, for example, is create a virtual twin of a driver to help the driver evolve and train better."

The concept would involve creating a digital model that reproduces a driver's techniques, preferences and individual characteristics. Di Grassi uses artificial intelligence as the broad term, but does not specify how such a system would actually be built. Machine learning - a subset of AI in which models identify patterns from data - could in principle provide one technical basis, but di Grassi does not say that this would be the method used.

That digital driver model could then complete millions of virtual laps before the real driver gets into the simulator and trains with it. "Instead of training against another driver or learning from another person, AI can actually create a model in which the driver trains against himself - against himself driving for millions of laps before he jumps on the sim, with his own tendencies, techniques and preferences."

"AI is not going to replace the driver"

Di Grassi is adamant that this is not about removing the human from the race car. "We need to be human-centred. Motorsport is about the best humans driving cars. AI doesn't mean it's going to replace the driver. On the contrary, AI will help."

He believes artificial intelligence could also reduce costs and accelerate development. "It can make the championship cheaper and better. We can evolve faster and find better solutions," says di Grassi. In his vision, AI would be used to drive Formula E forward rather than diminish the role of the people competing in it.

From racing lines to track design: Di Grassi sees wider AI potential

The virtual driver twin is only one part of di Grassi's vision. He sees numerous other areas in which artificial intelligence could potentially be used within Formula E.

"AI could control parts of the car, help design different racing lines or even help design different corners," he explains. Di Grassi also points to simulation and the cross-referencing of large amounts of data as areas with potential.

"There are a lot of things AI could do to make motorsport safer and cheaper, and Formula E more efficient and faster," says di Grassi. But his focus remains firmly on the human element: "It should make it more driver-relevant, not less - more driver-centred."

Di Grassi believes Formula E is particularly well placed to experiment with such technologies. "I really think Formula E is primed for this AI revolution. We need to bring it inside."

Whether any of his ideas will ultimately be adopted by Formula E, the FIA or individual teams remains open. Di Grassi presented them as his own vision rather than as part of an already confirmed programme. But the virtual driver twin is one concept he would personally like to pursue: "A virtual twin is just one thing that I would like to develop as quickly as possible, because I think it's going to help the teams a lot."