During the driving process, the navigation software will issue real-time commands and prompts: "Turn left 100 meters ahead", "There is congestion one kilometer ahead, and it is expected to take 5 minutes to pass"... Many people may wonder why the navigation software is so "smart"? Is it relying on cameras on both sides of the road or induction coils buried underground? The project leader and researcher at the Institute of Geographic Resources, Chinese Academy of Sciences, Lu Feng, told reporters that traditional traffic information acquisition relies on devices such as induction coils deployed under road sections or video probes deployed above roads. These devices have fixed locations, high costs, and cannot fully reflect the real-time status of urban transportation

Therefore, the project team must take a different approach.
The project team has set their sights on taxis, which are equipped with satellite navigation positioning systems that can periodically transmit location and status information to the monitoring and dispatch center. These taxis, which roam the streets every day, have become an important source of urban traffic information collection.
However, taxis are not a 'once and for all' solution. The scale and range of taxis are limited after all. Especially with the increasing popularity of self driving travel, timely access to intercity and even national transportation information has become a "must-have" for people's travel.
The project team decided to establish a new "traffic state induction coil", which covers a wide range of tracks, including long-distance buses, trucks, mobile map client tracks, Internet sites and social network text messages, etc.

According to national requirements, long-distance buses, tourist charter buses, dangerous goods transport vehicles, and general cargo trucks across the country must be mandatory to install driving monitoring equipment to monitor the entire driving process, prevent driver fatigue driving, and reduce traffic accidents. These instruments will upload the vehicle's driving trajectory in real time with extremely high accuracy, so we have a mobile information source covering the whole country, "said Lu Feng.
Moreover, the project team also keenly observed that in the Internet era, a large number of users use mobile map services online every day, and these high-frequency satellite positioning tracks undoubtedly greatly expand the mobile information source team based on cars. In addition, there is a large amount of traffic information online, including official and self posted road conditions, road signs, and other information by netizens on social networks. Then, can we make full use of these high-frequency trajectories of Internet maps and a large number of texts on social networks to give full play to the power of the masses and obtain real-time traffic information?

In view of these technical bottlenecks, the project team has broken through such technical bottlenecks as large-scale high-low frequency moving target track map matching and traffic state analysis, traffic information extraction contained in Internet and social media texts, real-time integration of multi-source traffic information, built a traffic information ecosystem for service provision and user feedback, established a nationwide traffic information service system, and achieved a national traffic information map. The project team successfully upgraded the public from traffic information users to traffic information collectors. These pieces of information provide the public with the operational status of the urban road network, effectively guiding their travel and avoiding congested roads.

"When you open the navigation software, your current road information will be uploaded automatically, that is to say, everyone is both a beneficiary and a provider of traffic information, which is also an embodiment of the Internet sharing economy," said Lu Feng.

It is reported that nowadays, the system processes over 2 million taxis and vans, over 8 million mobile phone map trajectories, over 10000 fixed sensors, and over 10000 traffic event texts every day. This national traffic information map has built a traffic information ecosystem that couples service provision with user feedback.

The efficiency of updating navigation maps has been greatly improved
In navigation, we often worry that map updates won't keep up with changes in roads or surrounding commercial areas. Due to the intensive construction of urban areas and frequent changes in commercial outlets, current applications such as mobile maps generally suffer from insufficient timeliness.
In the past, updating navigation maps relied on methods such as collecting vehicles. Unlike ordinary vehicles, the collection vehicle is equipped with a white rack that is about half a meter high on top. The camera on the rack can capture the surrounding scenery from all angles and transmit it to the computer. Then, the data operator performs post-processing on the captured video, generates or updates navigation maps, stores them in the system, and the data transmitted back by the collection vehicle becomes visible road information here.

This method of updating is very labor-intensive, often requiring a navigation map company to maintain a data collection and update team of hundreds of people, with huge costs, "said Lu Feng.
The project team believes that mobile Internet provides a good information sharing environment for businesses and people, and various highly timely road network maps and business service information emerge in endlessly. These non professional maps from volunteers can provide faster and better services for people if they are organically integrated with professional navigation map data.
In response to this technological bottleneck, the project team conducted in-depth research and first proposed an interest point and road network integration model based on geometric patterns, establishing a mapping relationship between the two and breaking through the problem of semantic consistency detection. Then, a heuristic probability relaxation matching model was proposed, achieving a perfect integration of volunteer sources and commercial navigation road networks with a matching accuracy of better than 95%, effectively ensuring the richness of information in travel navigation and map applications.