Testing a driverless car in Guangzhou’s Nansha New Area

This autonomous car, created by Pony.ai, can already drive on its own, but during our test drive someone is sitting in the driver's seat to talk us through the process, but he’s not touching a thing. It's extraordinary and terrifying in equal measure to be a passenger in something which feels so very new. These cars will be implemented in a Didi-like trial before the end of 2018, after two years of testing in both Silicon Valley and China.

Up until this test drive, I’d been vaguely aware of driverless cars, but those that I had seen were in early development, able to reach 5-10 kilometers an hour in real-life situations. However, as soon as we fasten our seatbelts, the Pony.ai car quickly speeds up to 50 kilometers per hour. Our co-driver explains that the vehicle has been pre-programmed to comply with local speed limits.

The car uses high-grade sensors to understand everything going on around it, from upcoming traf-fic lights and other vehicles on the road to pedestrians and unexpected obstructions. The car slams on the breaks on a couple of occasions as it approaches a slip road, assuming a car may make a quick appearance round the corner.

The company, Pony.ai, started in Silicon Valley of California, going on to establish the Pony.ai Insti-tute of Artificial Intelligence in Guangzhou’s Nansha New Area in May 2018. Christina Cao, the company’s business development director, says that opening in Nansha New Area, the city’s Free Trade Zone, was a huge help in the product’s quick development.

“Nansha New Area granted us a certain area in the city to deploy our driverless cars in real scenar-ios so more solid data could be collected. We’ve been working on the research and development of L4 driverless cars since the foundation of our company in 2016. L4 accommodates almost 80 to 90 percent of possible scenarios these cars will come up against.”

The company’s mission is to build the safest and most reliable technology for autonomous vehicles, and Vince Dai, from the corporate strategy department, explains that working in China has allowed their team to collect more data. “We have an advantage in data collection compared to our coun-terparts in other countries since China has more complicated scenarios when it comes to road tests. The more data we collect, the more advanced the system will be.”

The company’s first trial will be implemented before the end of 2018, when fifty autonomous cars will take part in a taxi trial. Similar in design to bike sharing apps, customers will be able to walk to predetermined pickup locations to take a ride in a driverless taxi. Over time, Cao explains, they will develop an app so that these taxis can drive anywhere a regular cab can, cutting driver costs.

When it comes to AI, you can’t get much more intelligent than a taxi that can drive itself. “Driverless cars represent the future of A.I. People in our industry hold the tenet that the driverless car is the mother of all A.I. projects. Being able to sense, predict, control and plan the development of driver-less car covers almost every aspect of the A.I. industry,” Dai noted.