June 3 — The innovative achievements of ZGC Smart City in the smart transportation sector once again garnered national media attention. CCTV News Channel conducted an in-depth report on the company's “Dual-Intelligence” urban smart transportation scenarios, praising the shift from “vehicles watching traffic lights” to “traffic lights watching vehicles”, which facilitates smooth traffic flow and has progressed from pilot programs to large-scale deployment.
In recent years, Beijing has been making all-out efforts to advance the “Dual-Intelligence” city construction, enabling traffic intersections to possess autonomous perception and real-time analysis capabilities. The Haidian District “Dual-Intelligence” urban smart transportation project, built and operated by ZGC Smart City, was launched as a pilot in 2025 and has now achieved comprehensive coverage at multiple district intersections this year.
Taking the pilot area around Haidian District's Dazhongsi Subway Station as an example, 13 intersections within a 1-square-kilometer range are undergoing a “smart revolution”. In the past, vehicles queued up waiting for red lights; today, the traffic lights have learned to “read the situation” — generating approximately 200 timing schemes every 50 seconds through the AI large model and automatically selecting the optimal one for deployment. In the Lenovo East Bridge area, the congestion index during morning peak hours dropped by 43%, and by 47% during evening peak hours.
From technology R&D to practical application, and with sustained focus from national, municipal, and district-level media, ZGC Smart City has consistently deepened its presence in the smart transportation field and has launched the “Dual-Intelligence” traffic intelligent agent. The system's core capability lies in: real-time monitoring of traffic flow through roadside binocular and trinocular cameras and other perception devices, fusing full-element data of motor vehicles, non-motor vehicles, and pedestrians, and relying on the AI signal control large model for dynamic timing optimization. Traffic lights are no longer mechanical countdowns; instead, they “think” in real time based on traffic flow about which direction to release and for how long. Meanwhile, the system can also automatically identify illegal U-turns, lane crossing, and other violations, and achieve en-route monitoring through vehicle historical trajectories, alleviating congestion from the source.
Sui Zongbin, COO and Senior Vice President of ZGC Smart City, explained that the company has explored a “pure vision + 3D spatial intelligence” technical route. Based on video fusion technology, it achieves low-cost digitalization of intersections, completing traffic perception and 3D calibration. It adopts a “full-volume perception, dynamic fusion” approach to refine real-time identification of full elements including motor vehicles, non-motor vehicles, and pedestrians. He noted: “This solution fully leverages existing traffic management equipment resources, achieving a fundamental shift from ‘vehicles watching traffic lights' to ‘traffic lights watching vehicles.'“
Next, the company will focus on advancing three aspects of smart transportation: first, expanding the scope of technology application and driving more intersections to connect to the AI signal control system; second, deeply integrating with vehicle-road collaboration and autonomous driving to provide real-time signal control support for unmanned driving; third, continuously iterating the AI large model to enhance perception and decision-making capabilities in complex scenarios. It will continuously build a smart transportation system that is perceptive, capable of thinking, and able to evolve, bringing the “Haidian practice” to more cities.
