Hong Kong’s new traffic management system is set to mark the dawn of citywide AI-empowered real-time traffic data collection and flow management.

Eleven sets of surveillance cameras across nine streets in Kwun Tong district will kickstart the project’s pilot phase, police said.

The Smart Traffic Management System, having secured HK$6.8 million ($874,840) in funding from the Innovation and Technology Bureau, will go into operation at year-end, following final refinements, said officials who are optimistic about its long-term prospects of extension to other areas.

READ MORE: Hong Kong AI research institute to open as soon as next year

“This is a notable attempt by Hong Kong to keep up its smart city transformation efforts. Technically speaking, this type of AI-enabled monitoring technology has been extensively tested and proven. With the right implementation, achieving a sufficient level of accuracy and operational stability can often take just a matter of a few months,” said lawmaker Gary Zhang Xinyu, who is also a seasoned engineer.

The initial version of the system was introduced by the police force in 2023 as the result of a project in collaboration with the Hong Kong University of Science and Technology, aiming to make traffic flow management more efficient and reduce manpower. A team of 50 officers in 2020 has since been downsized to six.

Seeking to ride on the wave of new technology, the force floated the idea of upgrading the old system by replacing manual data collection and analyses with artificial intelligence-enabled operations last April.

“Apart from detecting traffic accidents, the system will also help adjust traffic lights as necessary. Data collected will be used to enhance the understanding of local traffic conditions, ultimately contributing to the optimization of existing traffic management strategies,” said Chief Inspector Yu Shing-lam of the Kwun Tong police district.

ALSO READ: Chan: Hong Kong well-equipped to develop AI

Ng Ka-fai, a senior research engineer from the university’s Department of Civil and Environmental Engineering, explained the system’s operating mechanism at Wednesday’s launch ceremony, highlighting its capacity for real-time data collection, analytic working, as well as accident alerts.

Ng cited examples where congestion can be quickly identified when real-time footage shows causes such as temporary construction or parking on both sides of the road, adding that the system will indicate the intensity of traffic flow using a straightforward color scheme and suggest adjustments based on AI-enabled simulations.

“Previously, for many accident or traffic violation black spots, the police mostly relied on public reports or patrols. Even with closed-circuit television systems in place to monitor traffic, reviewing the massive amount of footage was time-consuming,” said Zhang.

“Deploying the system can automate all these procedures, making the whole process much more efficient.”

READ MORE: Beating the snags in AI quest

Addressing potential concerns regarding privacy violations, Yu said that the system is designed only for traffic data handling, with no facial recognition capacity, and footage will be deleted in a 28-day circle.

Kwun Tong has made several attempts to explore the AI era. Two sets of AI-enabled drowning detection devices were installed in public pools respectively in 2023 and 2024 within the district, and an additional HK$700,000 will be dedicated to their upgrading, the Leisure and Cultural Services Department announced earlier last month. 

Contact the writer at wanqing@chinadailyhk.com