Self driving racing cars to battle it out in Formula E

Quote in 2 Mins
Retrieve a quote »
Fireworks, popping champagne corks and bikini models draped over bonnets won't be appreciated by the latest super self-driving race cars. Can this type of racing ever be exciting when it is just between competing software programs?

The self-driving racing car big breakthrough happened in California in 2015 when a self driving Audi TTS, nicknamed "Shelley", raced against an amateur touring class champion. Shelley won by 0.4 of a secondi. The autonomous car was developed by the Dynamic Design Lab at Stanford University, US, and the Volkswagen Electronics Research Lab. Analysis of the race results showed the amateur champion was quicker on some turns, and researchers conceded if the race had started at a different point on the track, the champion might have wonv.


But the astonishing fact is that a self-driving car is nearly comparable to a racing driver. The researchers said the car still lacked the human capacity to guess how far to push the limits and also overall driving smoothness. For example, the autonomous car can calculate the quickest way around a track, but then rigidly sticks to that plan, whereas a person can swing wider in one turn, knowing this will set him up in a great spot for the next turnv. The Stanford team has raced cars at 149mph (239kmh) and also programmed an autonomous DeLorean to do donuts and driftiii.

Stanford University researchers say the main point of pushing an autonomous car to extremes is to develop computer software that is better able to avoid accidents and prevent injuriesii. The researchers have gained their knowledge by placing electrodes on the brains of racing car drivers, which shows the most complicated reflexive actions involve minimal cognitive brain power - that is, they are intuitive. This provides a huge challenge for researchers, who will need to make autonomous cars less reliant on strict algorithms, and better able to perform intuitivelyi.

Racing software can also be programmed to mimic a personality type. Difficult decisions, such as overtaking competitors, will, in the future, ultimately be a battle between different types of computer softwareii.

Impatient Formula 1 fans won't have to wait long to see these cars thrash it out on the track. Formula E and investment company Kinetik have launched a driverless electric Roborace car to compete in its hour-long driver-free races at the 2016 Formula E electric car racing championship (the date hasn’t yet been announced)iii. Ten teams, including a crowd-sourced community team, will run two driverless cars each and all the cars will be mechanically identicaliii.

Human drivers still have an intuitive advantage.

If you'd like to see how an autonomous car sees things on the road, take a look at this video developed by the Computer Vision and Robotics Group at Cambridge University. It shows how the car's view of the road is segmented into a form that's readable: other cars appear as purple shapes, road markings are orange and pedestrians are olive green.

Experts predict, however, that this probably doesn't signal the end of motor racing, as people will still want to see champions performing at their imperfect limits, with all the thrills, rivalries and uncertainty that goes along with being a fallible humaniv.

iPrigg, M. 2015, 'Robots take the chequered flag: watch the self driving racing car that can beat a human driver',, 19 February, viewed 23 December 2015,

iiGerdes, C. 2012, 'The future race car - 150mph, and no driver', TEDx, May, viewed 23 December 2015,

iiiDavies, A. 2015, 'Racing self-driving cars will make roads safer for everyone', Wired, 12 February, viewed 23 December 2015,

ivPhilip, S. 2015, 'Back to the Future day: could self-driving cars kill motorsport?', BBC Top Gear, 21 October, viewed 23 December 2015,