AV testing in China

LinkedIn +

The world’s largest automotive market – China – has also become a global research and development hub, with the scale and scope to solve some of the industry’s biggest technical challenges first. With foundations put in place for autonomous vehicle testing, China looks set to be one of the technology’s earliest mass-market adopters, and an important proving ground for systems that will be used all over the world.

Recent McKinsey research suggests timescales will be short, predicting two-thirds of the 13.8 trillion kilometers (8.6 trillion miles) traveled by all light-duty vehicles in China by 2040 will be driven by autonomous systems, and that 40% of all new vehicle sales will be AV-capable by the same point in time. Despite short-term disruption from the Covid-19 pandemic, McKinsey believes research and development efforts in China will “double down”, and it will continue to spearhead development and roll-out.

Domestic policy plays a big role, according to Pony.ai. Established four years ago, the company is undertaking on-road trials of robotaxi and truck fleets, and says governments – both national and regional – are acutely aware of the significance of autonomous driving in developing notably better transportation systems. It expects increasingly favorable policies for the foreseeable future, designed to lure stakeholders to conduct research and development there.

Zhang Bo, chief technical officer of DiDi Chuxing and CEO of its autonomous driving subsidiary, agrees. “Central and local governments are very supportive of autonomous vehicle development. China has a relatively well-formatted regulatory framework for autonomous vehicle compliance and safety, especially in Shanghai [where DiDi has a permit for on-road testing],” he says.

“The country also has a favorable environment for large-scale autonomous vehicle applications in the long run, as the nation seeks to build a comprehensive digital infrastructure network based on artificial intelligence, 5G and Internet of Things technologies, and is already thriving in advances in developing a Cooperative Vehicle Infrastructure System [CVIS].”

Regulatory changes and infrastructure developments happen in China incredibly quickly, according to Alexander Kraus, senior vice president, mobility, and global head of automotive at TÜV SÜD and chairman of the International Alliance for Mobility Testing and Standardization (IAMTS).

The alliance is aimed at developing and establishing global best practices for autonomous vehicle testing and validation methods, and developing a commonly accepted framework of regulations, test scenarios, validation and certification methods, and terminology. Kraus says that China’s early development of standards and pre-standards provides very clear guidelines for stakeholders to work toward this.

“What makes things easier in China, compared with many other regions, is that the regulation and standards development has the same velocity as the technology development,” Kraus explains. “That is really a hurdle we see at the moment in Europe and also the USA – the tech companies located in those regions invest huge amounts of money into disruptive technology development, but they cannot deploy it into scalable business models because the regulations are not there.

“Government-level support also extends to investment in test facilities, enabling them to focus on adding the required assets and infrastructure for testing,” Kraus adds. “These are massive investments – you need to build roads, artificial urban environments, crossroads that are interconnected, and so on.

“There are some top-notch test facilities in China, which are not run as a business model and totally concentrate on the technology development. If you look at Europe or the USA, this is always a business model, which limits the investments because they want to earn money.”

Technology giant Baidu was one of the first to utilize that support; its Apollo autonomous driving platform is undergoing widespread trials across China and the USA, in partnership with both domestic and international OEMs. It says advances rely on a ‘trinity’ model where technology companies, governments and OEMs work together, adding that its home market is very supportive. Growing digital road infrastructure is reducing the technical difficulties as well as cost of deployment, while the country’s test procedures are rigorous and standardized, enabling closed test sites and guidelines for carrying passengers.

China might also have a need for this technology first. McKinsey says autonomous vehicles are a potential solution to alleviate the pressure on its increasingly overstretched road network. With 30 million light-duty vehicles sold in 2018, the market in China is 70% larger than the USA, and Beijing residents’ 1.3-hour average commuting time is three times greater than that of their American counterparts.

Li Bo, head of the Intelligent Software Center and autonomous driving technology application at the Geely Research Institute, believes China offers some unique early use cases. Its megacities often feature elevated urban roads with no pedestrians or traffic lights, which are good environments for autonomous vehicles, while heavy traffic could also create a market for low-speed (sub-6km/h [3.7mph]) assistance systems. Scenarios and vehicle types are also different, he says.

“In China, there are a lot of big trucks, which are very difficult to detect with the camera and radar sensor,” he explains. “We do a lot of testing for these big trucks as well as the small vehicles. The operational situation is different from other countries, and driver behavior is different. For example, sometimes [Chinese] drivers have a very big [burst of]acceleration, or a very tight cut-in. This kind of driving behavior we also need to handle.”

For Daimler, China offers the chance to test systems on more complex road networks. One of the first major OEMs to get an autonomous driving license in 2018, it’s validating how sensors and cameras cope with higher densities of pedestrians and bicycles, small two- and three-wheeled vehicles and signs featuring Chinese characters. Highways, meanwhile, feature parallel lines to indicate safe following distances – these look like pedestrian crossings in other markets, and systems have to understand these differences, says Daimler.

Aside from traffic and population size, China also offers huge variation in climate conditions. During the first 500 days since Pony.ai began testing in Guangzhou at the end of 2018, its robotaxi fleet encountered 200 days of rain, with damp, heatwave conditions and high salinity posing big challenges for system stability. It’s a totally different operating environment from Beijing, where another of the company’s test fleets is at work in a dry, sunny and windy climate.

What’s less variable is end users’ willingness to accept the autonomous vehicle technology itself. According to a recent Bosch survey spanning six countries, 74% of Chinese respondents were in favor of accelerated introduction of autonomous driving, compared with 33% in Germany and 31% in the USA. Li Bo says consumers have embraced EVs and are as open to new features as they are driven to contribute to their development.

“For the last 30 or 40 years, the Chinese government has been very focused on education,” he comments. “The population is not growing so fast, but the number of good engineers is still growing because of the good education environment. The result is, there are a lot of technology startups in China, and something like 30 or 40 companies focusing on autonomous driving areas. They are working very hard to provide good solutions and products to the market.”

This also isn’t happening in silos. Kraus notes that strategic policies and government investment have helped establish hubs similar to California’s Silicon Valley, in turn driving advances that mean on-road testing is far more common. Although it’s attracting international investment, he believes that China will continue to spearhead development, and stakeholders are being clever with the data they collect.

“Test vehicles create
a huge amount of data,” he continues. “In China, they are collecting data in a very structured, highly standardized way, and that will bring China to the forefront in data use for AV development, which will accelerate development greatly.”

McKinsey research is similarly optimistic, predicting robotaxi operating costs in China will cross over with manually driven fleets by the year 2026, with mass commercialization to begin in 2027. The Pony.ai team has stepped up its collaboration with Toyota to enable easier integration of its technology into new vehicles and co-development of mobility services such as the PonyPilot robotaxi product. Baidu, meanwhile, is focused on creating a commercial proposition for its Apollo autonomous driving platform, also predicting widespread applications in robotaxis despite recent pandemic-related changes in customer habits.

“China will most likely become the world’s largest autonomous driving market,” says Zhenyu Li, corporate vice president and general manager of the Baidu Intelligent Driving Group. “Having been through the pandemic and supported the front line, we realize that automation and intelligence are the best solutions for humans to respond to large-scale emergencies.”

Focused on solving one of the industry’s biggest challenges, in a market with a genuine need for increased automation, China might turn out to be as influential for the autonomous vehicle sector as it has been for electrification. Kraus believes there’s a lot that others could learn from the experiences gained there.

“Looking at what’s coming up in China, I honestly believe that they will be the first ones to deploy mass standards and regulations that are really usable and practical. I think everybody is looking currently to China and what’s going on there, and this is something that will be of benefit to AV development worldwide,” he concludes.

Share this story:

About Author

Comments are closed.