Building an Urban Traffic Brain

Customer Background

Shenzhen has a land size of less than 2,000 square kilometers, and its city-wide road length is only slightly more than 600 km. There are about 530 vehicles per kilometer on average, the highest vehicle density in China. As a result, the contradiction between people, vehicles, and roads is increasingly prominent.

Against this backdrop, the Shenzhen Traffic Police Bureau has been boldly exploring new feasible ways and making innovations to improve interaction between people, vehicles, and roads. By doing so, the Bureau has long maintained a leading position in the industry.

Challenges

From the perspective of the traffic management department, the disadvantages of the existing video system are as follows: First, the collected transportation data are isolated from each other, resulting in data silos. Therefore, road condition prediction and violation warning are difficult. Second, traffic lights at intersections cannot be adjusted in a timely manner based on road congestion. Traffic data collection has little value in improving traffic conditions.

Solution

Huawei provides an integrated solution based on the Atlas AI server and ISV software to meet the real-world requirements of Shenzhen Traffic Police. Based on the Atlas hardware, Huawei helps Shenzhen Traffic Police build a video cloud in the city, which is physically dispersed and logically centralized. It supports flexible scheduling, video resource pooling, and video management and analytics capabilities. In addition, an efficient dedicated video network is established to ensure block-free video transmission and real-time data collection.

Benefits

With cameras deployed at intersections, the traffic management department can collect transportation information, and use intelligent analysis methods to improve the ability of traffic state awareness. The police department can extract valuable information and help to make better-informed decisions, improving the city's transportation management capabilities. At the same time, the system successfully leverages the value of intersection and vehicle data, and adjusts the traffic lights according to the congestion condition. As a result, the traffic rate of Shenzhen city road increases by 9%, and the congestion time in peak hours decreases by 15%.