Baidu's autonomous vehicles set for public test

A driverless car, developed by internet giant Baidu Inc, on display at the Light of the Internet Exposition in Wuzhen, Zhejiang province, on Tuesday. The expo, part of the Third World Internet Conference, opened on Nov 15, 2016. (Photo/China Daliy)

Baidu's autonomous vehicles set for public test

On a 5-kilometer road in Wuzhen, an ancient river township in East China's Zhejiang Province, the future of the auto industry is in sight.

On Tuesday, several cars without drivers, guided by a global auto positioning system, were moving on Ziye Road at speeds of up to 50 kilometers per hour. They were able to obey traffic lights, halt when sensing barriers or pedestrians and make turns smoothly in accordance with routes pre-set by passengers.

Those autonomous vehicles developed jointly by Baidu Inc, carmaker BAIC and Chery Automobile are more than just a show for the 20 or so reporters who turned up to see the road test on Tuesday. They will be open for public trial services, and about 100 people in Wuzhen will be invited to try out the services later this week.

Engineers from Baidu told the Global Times on Tuesday that they tested the cars on Ziye Road during the past month, running up more than 20,000 kilometers, to assure the safety of the trial services.

A reporter from Beijing felt "amazed" when she tried out the technology on Tuesday. She called the test drive "smooth, stable and impressive."

But it still needs improvement.

"The parking is very sudden," she said.

The public trial in Wuzhen, to some extent, could push forward the process of bringing driverless cars from lab to reality, according to a press release the company sent to the Global Times on Tuesday.

In mid-December 2015, Baidu announced plans to get its driverless cars ready for a rollout by 2018 and achieve mass production by 2020.

"Baidu indeed invested a lot in building new technologies for the auto industry. And it is leading the trend together with foreign peers such as Google Inc," Ma Jihua, an independent industry analyst from Beijing, told the Global Times.

Starting with the exploration of opportunities in autonomous cars from 2013, Baidu completed a 30-kilometer test drive of its two prototype driverless cars in December 2015. And in October this year, Baidu also completed a self-driving car test in California.

Google, which set up its autonomous car project in 2009, has reportedly tested its driverless cars for over 2 million miles, but the US tech powerhouse has not given a clear timetable for commercial application yet.

Premium electric car manufacturer Tesla appears to be more aggressive.

Tesla's newly released autonomous car system Autopilot 2.0 promised Level 5 full autonomy, while Baidu's Autonomous Driving Unit is developing autonomous driving technologies at Level 4 automation.

Level 4 system means the car is driverless but may not be able to drive everywhere. A Level 5 automated driving system is expected to be more autonomous than Level 4.

The US carmaker claimed that its electric vehicle could drive autonomously within two years from Los Angeles to New York, according to media reports.

But the mass market is not ready for the application of autonomous cars, partly due to the lack of mature technologies, said Ma.

"Unlike smartphones or TVs, cars are closely related to the safety of human beings. So companies should not rush to capitalize on autonomous technologies," Ma noted.

According to a guideline released by the Society of Automotive Engineers of China in October, partially self-driving vehicles are expected to take a market share of 50 percent by 2020, highly self-driving vehicles will take 15 percent of the market by 2025 and totally self-driving vehicles will take 10 percent of the market by 2030.

Baidu's trial came one day ahead of the World Internet Conference (WIC), which has been held in Wuzhen annually since 2014. During this year's WIC, executives from Baidu and Didi Chuxing are scheduled to discuss the challenges that smart driving will face, among other issues.

The widespread application of self-driving cars requires the solid development of car-connected mobility, the Internet of Things as well as "discipline" on the part of pedestrians and other vehicles, Yan Xiaojia, executive director of the Beijing-based Chinese Investment Data Institute, told the Global Times on Tuesday.

"It's difficult for the computers of self-driving vehicles to judge complicated road conditions when pedestrians and other vehicles such as bicycles ignore red lights," he explained.