When Didi, China’s erstwhile Uber rival, bought out the last of its main competitors and put to an end four years of a bitter market fight, there was little time for triumph before the inevitable question came: with those battles behind them, where was there left for them to go?
In 2015, having beaten out Kuaiche, Shunfeng Che, Daijia and other Chinese ride-hailing services, Didi acquired a new name, Didi Chuxing (or “Didi Travel”), and it started to reevaluate its purpose and position. Simply hailing rides was no longer enough, and so it set for itself the goal of becoming the “world’s largest one-stop mobile transportation platform.”
Cheng Wei, CEO of Didi, has said that urban transit hasn’t fundamentally changed in the hundred years since the arrival of automobiles. True, technology has improved the speed and performance of cars, with some now being able to easily surpass 300 kph. But in Didi’s recently released “Second National Keypoint City Transportation Report,” the average car speed in large cities in China is a mere 30 kph, meaning that the driver behind the wheel of a car with a W12 engine could hardly beat out a four wheel horsecart from a century ago. And despite a flurry of road construction over the last several decades, China has become the home to some of the worst traffic jams in the world, with several record-breaking cases of traffic jams that took days to clear. So what has gone wrong?
At PingWest’s SYNC 2017 conference in Silicon Valley this July, we sat down to speak with Dr. Gong Fengmin, Didi’s VP of Info Security Strategy and the head of its research lab. And while he believes gasoline engines may have managed to achieve speed, they don’t offer any solution for the problems of everyday traffic. But if Didi can leverage mobile networks, big data, and AI, it may at last be able to address modern car culture’s own worst problems.
Sharing as Smart Allocation
Sharing resources is the simplest way of solving bottlenecks, and with the emergence of mobile networking that has become easier than ever before, in part by widening the span of who can share what with whom. Shared transportation is what Didi has been doing all along, and at present it is already one of the most significant transportation service platforms in the world, with some 400 million users and more than 20 million rides per day. The upshot of this is that private cars sit idle less of the time, and more people ride in fewer cars.
But China’s regulatory policies have been keeping the ride-hailing industry from progressing further, the main hurdle being restrictions on which cars and drivers are allowed to operate. For instance, Beijing’s municipal government has instituted rules that limit Didi’s drivers in the city to those with local residency status and to cars with Beijing license plates. The result is that many otherwise qualified drivers and cars are excluded from working, creating inefficiencies for all sides.
“The current car ownership model, from either the perspective of energy or resource usage rates, just isn’t desirable, and it can’t continue,” Gong said. In Beijing, many can’t afford a car (or can’t buy one because of the quotas on license plates), while those who do own cars face problems with overcrowded parking and traffic jammed streets, perversely encouraging them to leave the cars they went to such expense to buy sitting at home. There is a gross mismatch between the use rates of cars themselves, along with parking spaces and roads, and people’s actual needs for getting around. Companies like Didi were born because of that imbalance of resources.
Apart from Didi, there are a number of international companies looking for a breakthrough in either production or market strategies to change the current car ownership model. Some, including electric car manufacturers, are examining time-shared car rentals and partnerships with ride-hailing services and providing customized vehicles for those purposes. Such cars would be owned and maintained by companies, not individuals, and aside from when they went in for maintenance would not usually require parking spaces, improving efficiency. “The shared model is more efficient, and in the future the car ownership model is certain to see some change. The trend is already taking shape,” Gong said.
Crowdsourcing for Smart Transit
Of course, sharing resources alone can’t entirely solve the problems of urban transit. Many city managers are already aware that they can use big data and machine learning to tweak traffic regulation and allocate transportation resources, but harnessing new technologies is a challenge, and so city governments have been partnering with companies. Gong notes that Didi is already working with Chinese city governments to connect with their transit information systems. For Didi, becoming a “one-stop mobile transportation platform” also requires a connection with public transit information.
There is precedent for this as well in Santa Clara, CA, which has already adopted rules for self-driving cars. Accident rates are low, at least in part because the local transportation department opened up its traffic signalling system to car companies so that self-driving cars could access them and make better judgements. In return, the companies have provided the government with big data processing and analysis.
Gong pointed out that the mutual feedback that transportation companies and governments have been providing one another is an aid for developing smart transportation further. “We want to solve problems through technology, but that can’t be done if companies and government don’t cooperate,” adding, “Many transportation departments are already using Didi’s open information platform to check local realtime traffic conditions.” He hopes Didi will be able to forge more such partnerships. “We can become information brokers for transportation, and provide even more help to cities.”
Gong says Didi is already using big data and machine learning for optimizing its services. To go even further, Didi’s system could draw in more layers of data on users’ digital lives, such as predicting their trips based on where they are and what they’re doing. “I believe that Didi’s platform has a lot of potential, and we have the capability to do a lot of things, but we want to serve consumers as well as possible, and we have to embrace an open ecosystem for partnerships.”
For those on the outside of the transportation industry looking in, the most salient AI application has been self-driving cars. The value of autonomous driving would be not only in making driving safer, but in the information that cars could exchange with transit systems, giving a boost to urban transportation networks with coordination.
This is what Didi’s lab in the US is working on right now, Gong says. The lab is especially focused on information security, which has become more important than ever with IoT devices and equipment, including cars. Didi is not the leader in self-driving technology, but it is nevertheless willing to put itself behind researching the kinds of information security protections that self-driving tech will require, because it is looking at an even larger picture.
Exactly what role Didi will be able to play in reshaping transportation and trying to resolve the excesses and inefficiencies of modern car culture remains to be seen. But its experiments in self-driving technology and smart urban transit are worth following, the company has the whole of China to use as its lab, and it clearly has no shortage of vision.