Mapping out the road ahead for China’s autonomous vehicles

6 min read
(Image credit: BigStock/Akarat Phasura)

Shanghai taxi driver Yuan Wei isn’t concerned about being replaced by autonomous vehicles (AVs). “No one will be able to afford an unmanned car,” he says. “It must be expensive. Maybe RMB 1 million (around $150,000) or RMB 2 million.” Yuan may be right about the price of a driverless car—at least given today’s technology—but the threat to his livelihood may be closer than it appears.

The middle-aged cabbie was driving around Anting Town, the hotbed for automobile innovation that lies 40 kilometers northwest of downtown Shanghai. In Anting, charging stations for electric vehicles line the streets and electric taxis seem to outnumber their gas-guzzling counterparts. As Yuan pulled over to pick me up, a sign overhead drew my attention: “Intelligent Connected Vehicle Test Road,” it read. Unbeknownst to Yuan, he had stumbled upon one of two government-approved testing areas for self-driving cars in the city.

“There are driverless cars here?” he asked later. “I haven’t seen any.”

China has set ambitious goals for AVs. By 2020, half of all new cars on the country’s roads are expected to be autonomous or semi-autonomous. For now, they’re still restricted to driving on designated roads, but that will change. The number of these vehicles is expected to reach 8.6 million by 2035.

Self-driving cars and several related industries are crucial to China’s long-term plan to upgrade its economy by shifting away from traditional manufacturing. But autonomous vehicles are especially important. Their success is underpinned by the country’s artificial intelligence (AI) prowess, for which the national government has set formidable goals. The State Council, China’s cabinet, wants to be a world leader in AI by 2030, making the country’s self-driving development even more pressing.

A number of forces are driving China to take the AV wheel faster than other countries, but widescale adoption will be challenging. Legacy vehicles and self-driving cars will share the roads for some time, and traffic infrastructure will have to be drastically rethought. For better or worse, the transportation experience and the shape of our cities will be significantly shaped by the model and pace of AV adoption.

A call to arms

To keep up with the United States, China laid out national guidelines for testing self-driving cars in April last year. City governments have followed suit. Beijing, Chongqing, Shenzhen, and Guangzhou, in addition to Shanghai, have opened their roads to AVs.

Tech giants and startups are taking advantage of the government’s favorable regulatory environment. Baidu, Alibaba, and Tencent are all developing platforms for self-driving vehicles while partnering with vehicle manufacturers.

Shanghai-based electric vehicle startup Nio is developing a self-driving car dubbed Eve, expected to be released in 2020, while Byton, which plans to open a vehicle factory in the eastern Chinese city of Nanjing this May, is developing its autonomous K-Byte model for launch in the same year.

In the parlance of AVs, most cars currently on the road are considered Level 0 vehicles—wholly dependent on their drivers. At the other end of the spectrum, Level 5 systems are entirely independent of human intervention in all situations.

Most Chinese autonomous driving companies are currently focused on Level 4 autonomy—fully independent within certain conditions. Level 3’s conditional automation systems are often seen as too dangerous for public use, as drivers are slow to take back control of the vehicle if a problem arises; companies like Google are opting to skip this level entirely. However, some companies, including Beijing-based Autobrain, are looking to take Level 3 vehicles to market by 2020.

Mobility services

At around 120 cars per 1,000 people, China’s rates of vehicle ownership pale in comparison to Europe and the US, says Bill Russo, ex-Chrysler executive and founder of consultancy Automobility. According to World Health Organization data, there were 830 vehicles per 1,000 people in the US in 2015, six times higher than in China. In the European Union, that number totaled just over 500.

“China … is starting at a different place and is perhaps willing to experiment in different ways, not just because the government wants it to, because the market is different, and because people don’t have deeply rooted [car ownership] habits,” Russo said at an industry event in Shanghai.

With low rates of car ownership and high demand for mobility, China has become the biggest ride-hailing market in the world. According to market research company Statista, the number of people in China using mobility services increased by 16% in 2018, reaching nearly 260 million. The number is expected to rise to more than 290 million this year.

In 2017, Didi Chuxing, China’s largest ride-hailing firm, facilitated more than 7 billion rides, according to the company, compared to Uber’s global total of 4 billion. Last year, the Chinese giant operated around 30 million rides a day.

While Didi dominates the market, other players are seeing increasing growth. Dida Chuxing became the second largest ride-hailing platform in China in October, jumping to 10 million daily active users. Other players include Meituan and Banma.

Just as the maritime industry served as a catalyst for public adoption of radios in the early 1900s, ride-sharing networks will act as a forerunner in AV adoption, using the data they collect to improve their self-driving abilities, while paving the way for more widespread adoption.

Still, nobody envisions a full-scale shift to driverless ride-hailing services anytime soon. Level 4 self-driving cars may be capable of functioning autonomously, but certain weather conditions pose a significant challenge to these vehicles’ self-driving capabilities. Consequently, mobility services would be well-equipped to handle such limitations by deploying AVs when road conditions are right, but continuing to use human-driven cars when their driverless counterparts can’t operate.

Those same limitations would make private ownership of self-driving cars less practical, at least in the initial stages of development.

A move away from sharing

On the tip of the eastern peninsula of Shanghai’s vast Pudong District, hugging the Yellow Sea, lies the almost perfectly circular Dishui Lake. Although the name translates to “Water Droplet Lake,” when seen from above, this man-made lake would seem to have been created with a hole punch.

Concentric roads surrounding the lake spread out like ripples. Similar to Anting, the area has been designated for autonomous vehicle testing. To a casual visitor, however, the sparse lanes feature more street sweepers than vehicles of the future. It looks much like any other developing part of Shanghai.

Throughout the 20th century, urban development around the world has been inexorably shaped by cars as highways feed sprawling low-density suburbs.

“The car has dictated major city developments in China,” Vicky Chan, founder at Avoid Obvious Architects, told TechNode. “It’s quite dramatic. Many cities are home to big company headquarters, which are meant to be appreciated from the highway.”

Private ownership of AVs, which is expected to become viable in a decade, could exacerbate urban sprawl. Imagine a scenario in which you could sleep or work in your car en route to the office. You could choose to live further away from work. But if everyone decided to do so, traffic congestion would rise commensurately.

Some have argued that in the next 10 to 20 years, AVs could become even cheaper to own and maintain than legacy vehicles, with their simpler electric engines and lighter bodies. The cost of manufacturing could also be reduced by removing driving interfaces—the steering wheel, the dashboard, and foot pedals.

The increased affordability and convenience of owning a car could make driver’s licenses an anachronism in an automated world, further increasing the number of vehicles on the roads.

“It can’t be that everyone is alone in their own capsule in their own vehicle,” said Tom Kirschbaum, founder at European public transport technology solutions provider Door2door.

“It’s clear that by only making vehicles autonomous you won’t win in terms of effects on congestion,” he said.

Research shows that the knee-jerk reaction of building more roads to alleviate congestion doesn’t work. A study by the US National Academies of Sciences, Engineering, and Medicine found that every 1% increase in a highway’s capacity results in a  0.7% traffic increase in one to two years and a 0.3%-1.1% rise after five years.

The question, then, is whether networks of autonomous vehicles can provide a feasible alternative to the appeal of privately owned cars. According to Henry Liu, vice president of Didi, the future of transportation is not hailing a vehicle, but instead a seat in a car, making driverless fleets a smaller version of public transportation.

To Kirschbaum, blurring these lines is vital: “The gold standard in mobility would be a scenario where the user has a very seamless way of using a variety of transportation modes.”

The challenge lies in pushing cities to factor in technological innovation when making long-term policy decisions. “They try to figure out the impact for the next 10, 15, or 20 years. At times when things are changing so fast, and technology is changing so fast, this is a big clash,” he says.