China News Service, Guangzhou, November 10 (Reporter Wang Jian) The 2026 China International Urban Rail Transit Exhibition was held in Guangzhou on the 10th. Focusing on AI applications, signaling systems and the open source Hongmeng digital base, a number of smart rail transit "black technologies" were unveiled at the exhibition, attracting exhibitors at home and abroad to learn about it.
In recent years, the urban rail industry has shifted from "building lines and expanding scale" to "refined operations, reducing costs and increasing efficiency". Artificial intelligence is moving from display scenarios to daily dispatching and operation and maintenance. At the exhibition, Chinese companies focused on the self-developed "Knowledge and Action Big Model" to embed AI into line network dispatching, passenger flow control and equipment operation and maintenance: image detection agents assist in identifying vehicle appearance faults, emergency dispatching agents can generate operation chart adjustment plans in emergencies within 3 minutes, and platform door failure prediction agents can provide early warning and locate the cause.
The reporter learned at the exhibition that multiple related AI products have covered various professional scenarios such as vehicles, signals, communications, electromechanical and so on.
Changes on the operation and maintenance side are more straightforward. Traditional subway equipment is repaired separately by profession, and fault detection and dispatching rely on experience; the "Jiayiwei" operation and maintenance intelligent system opens up comprehensive monitoring, platform doors, communications, signaling and other systems to promote the shift from "planned maintenance" to "status maintenance". The accuracy of intelligent dispatching is no less than 99%, and maintenance efficiency can be improved by about 30%. At present, this type of product has been used in subways in more than ten cities including Guangzhou, Shenzhen, Chengdu, and Chongqing.
Ensuring driving safety is still the bottom line for technology implementation. The exhibition site showed the train signal control system self-developed by Chinese enterprises, covering the four core links of monitoring, interlocking, protection, and automatic driving, and adapting to the mainstream technology routes of CBTC and FAO. Among them, the computer interlocking system meets the highest safety level and has been implemented on many lines at home and abroad. The simulated driving platform can restore a variety of fault scenarios and is used for emergency training of drivers and signal personnel.
It is worth mentioning that in the past, cameras, broadcasts, turnstiles, and escalators in stations belonged to different manufacturers, with different protocols and fragmented data. However, the "Transportation Jiahong" operating system unveiled at the exhibition connects heterogeneous equipment to the same platform through a unified object model: the equipment is plug-and-play and can be linked across systems with one click. Station staff can check the status nearby and quickly troubleshoot. At present, dozens of self-developed and ecological hardware have passed the compatibility certification and have been pilot verified in smart stations in cities such as Shenzhen.
When AI and large models no longer stay in concept demonstrations, but enter the dispatching console, vehicle control room and rail monitoring links of urban rail transit, the competitiveness of rail transit is also changing from "running fast" to "managing smartly, repairing proactively, and the base being autonomous". (over)
