Baidu: still a while before self-driving cars operate on roads

A self-driving robotaxi developed by Chinese search provider Baidu and Chinese carmaker FAW Hongqi participates in a test ride in Changsha, Hunan Province

Search engine giant Baidu has said that it will be a while before autonomous vehicles finally get working on the roads.
Last year, the Chinese firm registered nearly 140,000 kilometers in Beijing in self-driving distances, representing about 91 percent of total distances traveled by the eight licensed transportation companies in China.
Baidu has been running a platform – Apollo – which lets its third-party developers access its autonomous tech.
In a talk yesterday at the University of Macau, Zhang Yaqin, president of Baidu, recalled that the first milestone took place in 2009, when Google started to invest resources and capital into autonomous vehicles.
In China, Baidu started its serious research about five or six years ago.
“Another major milestone was in 2017 when Baidu opened Apollo to the public and made it an open source platform. Most car companies are using it,” said the scientist.
“Why is there such a big undertaking and so much interest? The first and foremost reason is safety, […] and secondly, safety,” said Zhang.
In China, car accidents kill more than 200,000 people each year, 95% of which are caused by human error.
For Zhang, economic value also remains a significant factor as to why market players are investing in the project.
Research co-conducted by the firm foresees that the automotive industry will amount to $1.6 trillion in the next 10 years.
“There’s an additional $1.6 trillion opportunity from the ecosystem of driving [including] system, hardware, technology. New revenue opportunities can be created from these,” he said.
Still, “it’s going to be a while” before these cars get onto real road lanes, said Zhang, citing the joint research of Baidu and McKinsey & Company, which shows that in 10 years, about 15% of new cars will be level 4 and above, while 50% of automobiles will be level 3 and above.
Zhang was referring to a system defined by the Society of Automobile Engineers, which details six levels of automation starting from Level 0, indicating no automation, to Level 5, indicating full automation.
Meanwhile, contrary to the comments made by Elon Musk on the use of lidar – a device that acts as an ‘eye’ for autonomous vehicles – who said that “anyone relying on lidar is doomed,” Zhang thinks the opposite.
Lidar is a technology that uses safe, invisible laser beam pulses to detect objects both in motion and at rest.
“Lidar provides additional data which serves your advantage, so why not? We were able to build cars […] with lidar and it’s more safe and has more capabilities,” the scientist explained.
“Lidar is more expensive but we believe as an approach that it will be cheaper. […] With all this data, it gives us more things to compute,” he added. LV