On September 24, 2015, a Geely management team of about 40 people visited Baidu. At that time, we had only been developing autonomous driving technology for two years. After listening to our introduction, Geely said that although autonomous driving technology is good, it is not what car manufacturers need most. If our technology can significantly improve the safety of cars, Geely will definitely be willing to pay for it. This incident left a deep impression on me and made me start to pay attention to traffic safety issues.

Every day, there are reports of traffic accidents causing casualties. According to data released by the International Highway Safety Association, about 1.35 million people die in traffic accidents every year, which means that on average one person dies in a traffic accident every 24 seconds around the world; in addition, 20 million to 50 million people are injured in traffic accidents every year.

In China, driving under the influence of alcohol is a serious threat to traffic safety. In 2020, courts across the country concluded 289,000 cases of dangerous driving crimes such as driving under the influence of alcohol, ranking first in criminal cases, far exceeding the second-ranked theft. Moreover, the number of dangerous driving crimes and crimes involving dangerous driving under the influence of alcohol has increased year by year, and the figures are shocking.

"One drop of alcohol from the driver brings tears to the family", this sentence is almost known to everyone. Traffic accidents bring great grief to every injured person and the family of the deceased. At the same time, the economic losses caused by traffic accidents are also staggering. According to the official website of the International Highway Safety Association, road traffic accidents may cause countries to lose 2%-8% of their GDP.

The impact of transportation on climate is another major issue of global concern. Greenhouse gas emissions from transportation account for 16% of the world's total emissions of 51 billion tons. In China, the transportation sector is also an important focus for emission reduction.

In the past few years, I started by focusing on how autonomous driving can solve safety problems, and gradually went deeper and deeper, and gradually expanded the scope of my attention - focusing on the problem of dangerous driving by human drivers, as well as autonomous driving, smart transportation, smart cities and other issues. At the same time, as a member of the National Committee of the 中文 People's Political Consultative Conference, I have submitted proposals related to autonomous driving and smart transportation every year since 2016.

At the 2021 National People's Congress, I submitted a proposal titled "Accelerating the popularization of commercial autonomous driving and intelligent transportation, making travel greener and more convenient for the people, and achieving the carbon peak target." I believe that vigorously developing low-carbon transportation, using new technologies such as artificial intelligence and 5G, and accelerating the popularization of commercial autonomous driving and intelligent transportation can effectively alleviate traffic congestion, make travel greener and more convenient for the people, and achieve the coordinated unity of carbon peak and high-quality economic development.

In fact, as I pay more and more attention to traffic issues, the dimension of my view on autonomous driving has also changed significantly. Baidu's development direction has also gradually extended from autonomous driving to vehicle-road collaboration, intelligent transportation, and digital city operations. In this process, there are many practices and many lessons. However, I am increasingly certain of the following things:

An intelligent transportation system can greatly reduce the probability of traffic accidents. About 94% of traffic accidents are caused by human factors, including dangerous driving behaviors such as drunk driving mentioned above. I believe that as time goes by, people will increasingly realize that autonomous driving is safer than human driving. As driverless driving becomes more mature, "smart cars", "smart roads", vehicle-road collaboration, and intelligent traffic management systems can all reduce the probability of traffic accidents, making travel safer for cyclists, motorcyclists, and pedestrians.

An intelligent transportation system should be able to effectively solve the congestion problem. The "purchase restriction" and "traffic restriction" policies that are common in big cities should be cancelled. We have calculated that intelligent transportation based on vehicle-road collaboration will be able to improve traffic efficiency by 15%-30%. Within 5 years, China's first-tier cities may no longer need "purchase restriction" and "traffic restriction"; within 10 years, the congestion problem is expected to be solved by improving traffic efficiency.

An intelligent transportation system can improve traffic efficiency, save energy, reduce the level of air pollution, help achieve carbon peak and carbon neutrality, and benefit everyone who has travel needs. At present, traffic problems such as congestion, low parking efficiency, and unreasonable travel structure are all exacerbating urban carbon emissions and making environmental pollution more serious. By building high-grade intelligent roads, vigorously developing vehicle-road collaboration and autonomous driving, and building a safe, convenient, efficient, and low-carbon intelligent travel service system, we can make our cities greener.

An intelligent transportation system needs a dedicated operator. The operator's responsibility is to continuously optimize the system and continuously improve transportation efficiency and reliability. This is just like telecom operators optimizing mobile phone communication efficiency by continuously upgrading the network. From a more macro perspective, intelligent transportation is a microcosm of future digital city operations. The intelligent transportation operator model will bring disruptive innovation and far-reaching impact to transportation, cities, and even the economy and society.

Huangpu District of Guangzhou City adopted Baidu Apollo's ACE intelligent traffic solution and implemented a dynamic green wave control strategy on 6 main roads in the urban area. The average travel time on each road was reduced by 25%, and the average number of red light stops was reduced from 3-4 times to 0-1 times. Among them, the core arterial road Kaitai Avenue achieved green lights all the way through 12 intersections from east to west. By building "smart roads", Huangpu District of Guangzhou City allows drivers to drive with green lights all the way, no longer relying on "good luck".

In the past few years of being involved in autonomous driving and intelligent transportation, I have visited many cities in China and felt the changes in transportation. From the optimization of traffic light control to the construction of urban transportation AI engine; from buses to subways, from roads to ports, all industries and fields of transportation are conducting intelligent attempts, and local efficiency improvements are everywhere. At the same time, the speed of economic growth is also very fast. The once smooth road sections quickly become congested, so that some city managers use the degree of traffic congestion to prove the level of local economic development.

In the next 10 to 40 years, we will have bigger and bigger cities, more and more cars, and more and more people. This makes traffic problems more urgent and more serious than ever before. We need to find the best solutions for traffic safety, congestion relief, and low carbon.

Traffic affects everyone every day. It is both an economic issue and a livelihood issue. Decades of development of artificial intelligence technology, coupled with the top-level design and overall coordination of a capable government, can finally make traffic truly intelligent. Every time I think of this possibility, I am excited.

Of course, I know clearly that the current transportation infrastructure is still far from perfect, and digitization has not yet been completed in most places: traffic lights are still designed for people, not machines; camera data cannot be processed in real time; and traffic rules do not take into account driverless driving. I also know that the construction of transportation infrastructure and the formulation of traffic laws and regulations may involve many departments, and implementation is not so easy.

However, having witnessed the practice of intelligent transportation in certain areas of some cities over the past two years and seen how technology has brought real changes to people's travel, I am full of confidence and expectations for the future.