Didi zhangbo: facing three major changes in the industry 5g accelerates the commercialization of automatic driving

0
129

Sina Technology News on the afternoon of September 22, Sina 5g open day, jointly sponsored by Sina Technology and Sina 5g, opened recently. Zhang Bo, co-founder of didi travel, CTO and CEO of didi auto driving, elaborated his views on 5g and other new technologies enabling future travel in his speech at the future travel sub forum. He said that new technologies such as 5g, AI and big data will promote the transformation of the travel industry in the next 20 years from the three levels of infrastructure, transportation and shared travel.
At the infrastructure level, Zhang Bo said that didi travel, with the help of AI and big data of online car hailing, has reduced traffic congestion rate and improved road use efficiency and traffic rate through intelligent traffic lights, dynamic road assignment and more scientific road network design.
He said that didi currently works in more than 100 cities. Through real-time control of traffic light duration, Didi can extend the red light time at the entrance of congested road sections and the green light time at the exit, and form a green wave band of regional traffic with adjacent streets to improve the traffic efficiency of the whole region. In the infrastructure layer, intelligent map has become the core technology of Didi. With the help of precise positioning and recommendation of boarding point technology, 70% of orders no longer need to be communicated by drivers and passengers in advance.
In the middle of the transport layer, there will be two important changes in the next 10 years. First, energy will change from gasoline to electricity; second, it will be intelligent. Didi is working with car companies and hardware suppliers to promote the transformation of vehicles to intelligent and new energy sources.
Zhang Bo said that didi has been developing automatic driving in a low-key way for four years and now has more than 300 engineers and scientists in China and the United States. At present, Didi has formed a series of complete solutions and technical modules such as automatic driving perception, prediction, planning, control, and high-precision map, as well as the ability of automatic driving landing service. In June this year, Didi autopilot has provided a manned demonstration application to the public in Shanghai.
In the shared travel layer, he thinks sharing will become a trend. “All kinds of transportation resources should meet the needs of you when I’m not using them, and I’m using them when I’m not using them. Each car has three seats and one trunk. The ideal state is that one car will be involved in the didi network, and the three seats can be used immediately. The efficiency of vehicle use will be significantly improved. There should be no idle vehicles on the road. This is the scenario we imagine in the future. We hope to promote traffic reform and make everyone’s travel safer and more efficient. ”
In Zhang Bo’s opinion, Didi’s huge amount of big data and the world’s leading shared travel network are the advantages of didi in these three levels of reform.
For example, Didi’s vehicle will upload location, speed and direction every few seconds, add 106tb of daily trajectory data every day, process 4875tb data every day, route planning data is 40 billion times a day, and positioning times are 15 billion times, which is a big map trading engine. Using these big data of online car Hailing operation and road network safety data, combined with didi mixed order dispatching mode, we can determine which areas of automatic driving technology can be implemented and have commercial value, and finally realize the rapid landing and commercial operation of automatic driving.
For 5g, Zhang Bo said that it will bring higher bandwidth, lower delay and greater access density, which is an important infrastructure for the commercialization of automatic driving in the future. After the arrival of automatic driving, real-time communication between vehicles is needed, and communication between vehicles and road infrastructure is also needed to realize vehicle road coordination. 5g will accelerate the construction of communication between vehicles and vehicles, between vehicles and roads, and will make automatic driving land on more intelligent roads and vehicles, thus accelerating the commercialization of automatic driving.
Welcome to follow Sina 5g open Day, together with Wu Hequan, academician of Chinese Academy of engineering; LV Tingjie, executive director of International Telecommunication Association (ITS) and professor of Beijing University of Posts and telecommunications; Zhao Zizhong, President of New Media Research Institute of Communication University of China; Meng Pu, chairman of Qualcomm China; Qin Lihong, co-founder and president of Weilai; Han Xu, founder and CEO of Wenyuan Zhixing; Li Di, CEO of Xiaobing company Top executives talk about 5g new opportunities. (Zhang Jun)
The following is the full text of the speech:
In the past 20 years, the Internet and mobile Internet have basically solved the problem of information flow. Didi hopes to solve the flow of people and objects in the physical world. The construction and development of “new infrastructure” such as 5g, big data and artificial intelligence will promote the transformation of the whole travel industry in the next 20 years from the three levels of infrastructure, transportation and shared travel.
In the infrastructure layer, Didi travel reduces the traffic congestion rate and improves the road use efficiency and traffic rate with the help of AI, big data of online car hailing, intelligent traffic lights, dynamic road assignment and more scientific road network design. Didi currently operates in more than 100 cities. Through real-time control of traffic light duration, Didi can extend the red light time at the entrance of the congested road section and the green light time at the exit, and form a green wave band of regional traffic with adjacent streets, so as to improve the traffic efficiency of the whole region and reduce the traffic congestion rate of urban roads.
In the infrastructure layer, intelligent map has become the core technology of Didi. The map runs through the whole travel link. When passengers open didi app, the first step is to accurately locate the passenger’s location, and then recommend the most suitable boarding place for passengers. At this time, you will see that there is a small green dot on the home page. You can recommend the most suitable place for you to get on the bus according to your positioning. With the help of precise positioning and recommendation of boarding point technology, 70% of orders no longer need to be communicated by drivers and passengers in advance.
In the middle of the vehicle layer, there will be two important changes in the next 10 years. First, energy will be changed from gasoline to electricity. Second, it will be intelligent. Didi has invested more in automatic driving. We have been developing automatic driving very low-key for four years. Now there are more than 300 engineers and scientists in China and the United States. Didi is working with car companies and hardware suppliers to promote the transformation of vehicles to intelligent and new energy sources. At present, Didi has formed a series of complete solutions and technical modules such as automatic driving perception, prediction, planning, control, and high-precision map, as well as the ability of automatic driving landing service. On June 27, Didi automatic driving has provided a manned demonstration application to the public in Shanghai.

In the shared travel layer, I think sharing will become a trend. All kinds of traffic resources should meet the needs of you when I am not using them, and I will use them when you are not using them. Each car has three seats and one trunk. The ideal state is that one car will be involved in the didi network. The three seats can be used immediately, and the vehicle efficiency will be significantly improved. There should be no idle vehicles on the road. This is the scenario we imagine in the future. We hope to promote traffic reform and make everyone’s travel safer and more efficient.
Didi has a huge amount of big data and the world’s leading shared travel network. These are the foundation and advantageous resources for didi to constantly update the iterative technology.
Based on this huge network, Didi now has very valuable traffic big data. Our vehicles will upload location, speed and direction every few seconds. We add 106tb daily trajectory data every day and process 4875tb data every day. The daily path planning data is 40 billion times and the positioning times are 15 billion times. This is a great map trading engine. Didi will upload the location, speed and direction every second. If we put together the tracks of didi vehicles in Beijing, we will cover Beijing 400 times a day. If there is any change of road conditions on the road, where there are traffic accidents and congestion, the backstage will know for the first time, and inform the whole network of car Hailing to update the path planning. Hundreds of thousands of passengers and drivers report and share the latest traffic information every day.
Using these big data of online car Hailing operation, road network security data, combined with Didi’s mixed order distribution mode, we can determine where the automatic driving technology can be implemented and has commercial value, and finally realize the rapid landing and commercial operation of automatic driving.
We will refine the relevant information every day to improve the travel efficiency of drivers and passengers. At the same time, the orange video camera installed on Didi’s car Hailing has more abundant data, and these rich real-time map ecological data are also the unique advantages of Didi. At present, big data is playing a role in didi map and path planning, research and development of automatic driving, improving the accuracy of online car Hailing estimated arrival time, and improving travel safety.
Based on big data of online car hailing and advanced artificial intelligence technology, Didi eta (estimated arrival time) has been continuously optimized in the past few years. For example, when passengers leave, Didi app estimates that the journey will take about 20 minutes, but whether it is 19 minutes or 21 minutes, we will continue to optimize, and the time error continues to decline. In 2015, the estimated arrival time has an average error of 20%, which has been reduced to about 10% in 2019, which is in the leading position in the industry.
In addition, Didi’s rich big data has brought many advantages to Didi’s development of map business. Didi map originated in 2015, when didi was a start-up company just three years ago. Why did didi make a map? I find that Didi’s upper travel network and map have obvious positive feedback. When our upper travel network becomes more prosperous, we will have more real-time and fresh map data. When our map data becomes more and more abundant, we will be more able to make better maps, which will make our upper travel network experience better and more efficient.
5g technology will bring higher bandwidth, lower delay and greater access density. It is an important infrastructure for the commercialization of automatic driving in the future.
After the arrival of automatic driving, real-time communication between vehicles is needed, and the infrastructure of vehicles and roads also needs communication to realize vehicle road coordination. I think the arrival of 5g will speed up the construction of communication between vehicles and vehicles, between vehicles and roads, and will enable automatic driving to land on smarter roads and vehicles, thus accelerating the commercialization of automatic driving.