Driverless cars will need cities covered in sensors, DiDi VP says
Chinese ride-hailing giant Didi Chuxing wants to become one of the front-runners in developing self-driving cars, the company's chief scientist for smart transportation initiatives said on Tuesday.
Didi has been working to develop autonomous vehicle technologies for three years, and has teams based in the United States and China, Henry Liu told CNBC during a fireside chat at the East Tech West conference held in the Nansha district of Guangzhou, China.
"We already have autonomous vehicles being equipped with our sensors and we have licenses in both Mountain View, California, as well as in Beijing, China," he said. "We'll be one of the front-runners in terms of the autonomous vehicle technology development."
Automakers and internet companies around the world are investing millions of dollars and rolling out long-term plans for self-driving vehicles. Many analysts believe the widespread adoption of these vehicles will potentially start to pick up in 2021 or 2022.
"We can also predict in terms of what's going to happen in the next 15 to 30 minutes, in terms of traffic flows."
-Henry Liu, chief scientist for smart transportation, Didi Chuxing
For its part, Didi is developing autonomous vehicles on two fronts, said Liu.
First, by installing sensors in vehicles that can sense the environment on the road, detect objects, plan travel routes and ultimately, control the cars. The second front is what he described as "cooperative vehicle-highway systems" that rely more on the environment — that means having sensors installed on roads, buildings, lamp posts and the surrounding areas to provide relevant information to self-driving cars.
"The main difference is that we not only have the vehicle sensing capability, we're also going to have a roadside sensing capability, so we will be able to provide the autonomous vehicles with environment information, from the infrastructure side," he said.
But such a development will require the presence of very high-speed mobile internet connection readily available, Liu added. China is developing that technology very aggressively and hasoutspent the United States since 2015.
Didi's advantage: 550 million users
One of Didi's major advantages when it comes to developing self-driving, smart cars is that it has a massive transportation network, according to Liu.
Didi has about 550 million users taking an average of 30 million rides every day across more than 400 cities. That allows the Chinese firm to collect plenty of data about its users, from their travel habits to traffic conditions in various cities. Generally, artificial intelligence systems require large volumes of so-called training data to learn patterns and behaviors.
"We collect a hundred terabytes of vehicle trajectory data per day," Liu said, adding that Didi processes nearly five times as much information daily to better estimate travel routes, prices and demand for vehicles at any given time. The data also helps cities plan their traffic networks better to avoid congestion.
Earlier this year, Didi launched a so-called "Smart Transportation Brain" service with Chinese traffic management authorities. Using vast amounts of data from Didi, local governments and other businesses, the service uses artificial intelligence to recommend improvements to existing transport systems that can reduce travel times for commuters.
"We can also predict in terms of what's going to happen in the next 15 to 30 minutes, in terms of traffic flows," Liu added.
Didi remains one of China's most valuable start-ups, backed by major names including Apple, Alibaba and SoftBank, and it has a valuation of $56 billion, according to CB Insights. Two years ago, it acquired Uber's China business to establish its dominating position in the Chinese ride-hailing market.
WATCH: Didi Chuxing's chief scientist for smart transportation talks about self-driving cars
We use the data to help passengers and cities, says DiDi VP