HD maps steal limelight in auto driving
China a strong player in sector as it leads efforts in data collection
Just imagine: You are alone relaxing in the back of your car, which is driving itself on a highway in China, and you’re reading a book or watching a movie or doing whatever else the average human driver cannot do.
By now, you might be asking some questions: How can this car drive smoothly and avoid crashes without the help of natural intelligence?
Besides the self-driving car boasting sensor technologies such as a camera, radar, lidar and laser sensor that can help it “see” its surroundings by shooting out laser beams which then capture images, it boasts another crucial technology – a high-definition (HD) map.
Amid all the hype about autonomous driving, HD map technology is in particular attracting more and more attention in the sector.
Different from navigation maps, which are used to help people drive more efficiently, HD maps are specifically designed for self-driving vehicles with extremely high precision at centimeter-level accuracy, as such vehicles need to pinpoint their positions in a much more precise way.
Compared with sensors, the demand for HD maps was a bit shy at the beginning of the development of autonomous driving technology, Yu Jing, manager of the market development department at Beijing-based technology firm Navinfo Co, told the Global Times in a recent exclusive interview.
However, sensors have their own unique limitations. “Sensors can only ‘see’ within 100-200 meters, based on which, the car will make decisions immediately, but, that will not bring good user experience,” Yu said, adding that sensors cannot guarantee a 100 percent accuracy rate due to factors such as weather and light.
“But if the car is embedded with the precisely computed digital map, enabling it to observe from several kilometers away, it can have a comprehensive outlook and predict possible events.”
When it comes to driving, humans still have the edge in real-time decision-making capabilities such as stopping the vehicle in the right place or choosing a faster lane according to the driver’s own preference, something a robot would find difficult to execute, Yu said.
“Then here comes the map to strengthen autonomous driving to help make the right decision at the right time,” she noted.
Driving information, including stop signs, exit ramps, traffic lights and lane markings, is all featured in the HD map, since relevant road data are collected beforehand, which is a monumental task.
Two main methods are usually used when collecting HD map data: The first is sending a fleet of cars decked with state-of-the-art technology onto the roads, which then collects data, and the second is drawing on data scooped up by scanning systems that mobility service providers agree to install into their vehicles, including taxies and private cars operated by Didi, a popular ride-hailing platform in China, according to Yu.
“A fleet of cars is able to collect high-quality and high-precision HD map data, but those cars cannot satisfy the need to track real-time traffic updates due to their limited numbers, that’s why we’re trying to usher in the latter model,” she noted.
So far, Shenzhen-listed Navinfo has covered 50 percent of domestic highways in its mission to collect comprehensive HD map data, and it aims to have covered 80 percent of highways as well as expanded its urban road sample coverage by the end of this year.
Future business
“In China, ground-mapping belongs to a top-secret industry with a high threshold, which makes it difficult for foreign map providers to enter the sector,” said Zhu Jiangming, chairman of electric vehicle company Leapmotor, a start-up based in East China’s Zhejiang Province.
For example, Here, a map company started by Finnish multinational Nokia, which was then acquired by German automakers BMW, Daimler and Audi, is developing its business in the European and US markets, but if it wants to enter the Chinese market, it needs to partner with domestic players, and under current laws, it cannot obtain a license by itself.
In May 2017, Navinfo announced the establishment of a joint venture with Here in China to broaden its business within autonomous driving and HD mapping.
Given the strategic significance of HD mapping and its cutting-edge technological capacities, various players involved in the self-driving sector have already made early layouts, Zhu told the Global Times on Monday.
Apart from map providers, internet tech firms, traditional automakers as well as sensor-making companies have all expressed their ambitions to take a leading role in the niche market, Zhu noted, adding that the future trend will be strong players joining hands across different territories.
Chinese tech giant Tencent Holdings became the second-biggest shareholder in Navinfo in 2014 with a stake of nearly 10 percent, and that year also marked the beginning of the niche market gaining capital favors.
“Leapmotor, which is also making foray into autonomous driving, will choose a suitable map provider according to a comprehensive evaluation of its research ability and user experience, and then publicize the result when appropriate,” Zhu said.
In terms of technology standards, compared with the widely used navigation maps currently circulating in the domestic market – the development of which once lagged a decade behind such technologies in developed economies such as Japan and Europe – the HD map sector in China is now easily keeping abreast with that in developed economies, Yu noted.
“The current phase of HD mapping is still focused on cash-oriented research and development, but it’s potential to become a future business is very obvious,” she added, noting that services will definitely become an essential part of a safe and appropriate autonomous driving solution.