When Data Scientists Race Some Fast Cars

Elon Musk and his Tesla Model S started it all. The premium electric sedan was equipped with advanced and self-learning Autopilot system, which can take the chauffeuring work out of the driver. It was a complete breakthrough because if it works as good as it should; it is going to change driving forever. It is all because of Artificial Intellect (AI) and the way it works fueled by data sensors and powered by video analytics. And since we have, to be honest, it wasn’t started by Elon Musk but the automotive industry itself. It doesn’t matter if we talk about rain sensing wipers (Cadillac) or adaptive cruise control (Mitsubishi) or lane departure systems (Mercedes-Benz). Such systems work with predefined events with some variables in the input data and one particular, predefined outcome.

The rapid growth of AI enabled tech giants like Google to heavily experiment with fully autonomous vehicles. Their ambitious plan was preceded by DARPA Grand Challenge competitions, funded by the most prominent state defense research organization in the USA. The first one was held in 2004 and bunch of autonomous vehicles had to cross 150-mile/240 km distance in the Mojave Desert. All of them failed but the following year was way more fruitful – Stanford University won the first place with a time of 6:54 h. They were just 9 minutes faster than the second team, coming out of Carnegie Mellon University in Pennsylvania.

darpa challenge

Almost 10 years later, Audi followed up with a challenge of their own – to set fast lap at Hockenheim Racetrack with a driverless car. Their RS7 was not a record-breaking vehicle but it managed to reach speeds up to 150 mph / 240 km/h on a closed track.

Take a look at the video to feel what it was like!

The past few years were quite successful for autonomous vehicles development and it seems that it is time for driverless cars design. The cars competing in Roborace are technically based on the Formula E racers (F1 satellite electric series) but they do not have a driver cockpit. In fact, they look pretty close to unmanned plane drones but Roboracers have 4 wheels. The Roborace competition is going to be among identical driverless cars, equipped with cameras, radars, and a Nvidia Drive PX2 computing engine alongside the Formula E mechanics.

The main goal is not to race the cars but the Artificial Intellect driving the Roboracers. It is going to be really interesting since data scientists will compete for their skills in a pretty entertaining manner. Their algorithms will overtake each other at speeds up to 320 km/h. It is the first time when a bunch of AI machines are on the track.

The actual outcome of such effort is more precise collision preventing systems in actual cars, more sophisticated autopilot functions as well as more advanced data algorithms.

So, data scientists, beware! Someone might say, “My algorithm is faster than yours”.

Leave a Reply

Your email address will not be published. Required fields are marked *