The Land Transport Authority (LTA) will progressively roll out an AI-powered video analytics system using existing cameras on public buses, with deployment beginning in the fourth quarter of 2026.
The technology is designed to detect motorists who encroach into bus lanes or park illegally, while also identifying road infrastructure defects and non-compliant road works.
The system is expected to be expanded progressively across all bus routes by 2030, potentially giving Singapore a much wider network of eyes on its roads without requiring a completely new fleet of roadside cameras.
AI could turn everyday buses into mobile road monitors
The concept is relatively simple: instead of relying only on dedicated inspection vehicles, enforcement officers and fixed cameras, LTA can make use of cameras that are already travelling through Singapore every day.
According to LTA’s earlier innovation brief, the proposed system can analyse footage, identify potential violations or defects, capture supporting images, record the location and alert LTA officers when action is required.
Among the traffic offences considered in the system are:
- Vehicles travelling in bus lanes during restricted hours
- Illegal parking
- Stopping in yellow boxes
- Failure to give way to buses at bus priority boxes
The original LTA proposal also called for vehicles involved in detected offences to be classified and for registration numbers to be recorded.
But the current rollout announced by LTA specifically confirms detection of bus-lane encroachment and illegal parking, together with road defects and non-compliant road works.
The AI has already been tested
This is not simply a future concept.
LTA said trials of the system have been completed, with an average accuracy of 90 per cent across the scenarios tested.
The authority says the technology will strengthen monitoring across the bus network and allow problems to be addressed more quickly, whether that means enforcement against motorists or rectification of road and infrastructure issues.
The move follows LTA’s earlier effort to find technology that could make better use of bus-mounted cameras.
In 2024, LTA said conventional enforcement and inspection methods could be resource-intensive and labour-intensive. Its innovation call sought AI and video-analytics solutions capable of automatically detecting a range of road-related problems.
It is not just about catching motorists
One of the more significant aspects of the project is that the same technology is intended to improve road maintenance.
LTA’s proposal included detecting problems involving roads, bridges, tunnels, underpasses, road works and bus-stop infrastructure. These could include damaged structures, surface defects, poorly reinstated road surfaces and problems with road-work safety measures.
LTA has already been using AI in road-maintenance work. In January 2026, the authority said vehicle-mounted cameras were being used to identify visible defects such as cracks, potholes and ponding, with AI achieving accuracy of more than 90 per cent in that system.
The bus-camera initiative therefore represents a broader push to use vehicles already moving around Singapore as an additional source of road-condition data.
Will the cameras identify people’s faces?
This is where the distinction between road monitoring and general surveillance becomes important.
In a 2024 parliamentary reply, Transport Minister Chee Hong Tat said LTA planned to use the existing road-facing bus cameras to detect bus-lane encroachment, illegal parking and road or infrastructure defects.
He also stated that LTA had no plans to use the footage to identify faces and link them to Singapore’s National Registration Identity Card database. Government agencies’ use of data remains subject to prevailing laws and security guidelines.
In other words, the stated purpose of the AI system is focused on vehicles, traffic violations and infrastructure, rather than creating a facial-recognition system linked to residents’ identity records.
Why bus-lane enforcement matters
For motorists, driving into a bus lane during restricted hours may appear to be a relatively minor traffic offence.
For bus passengers, however, the consequences can be more serious.
Senior Minister of State for Transport and National Development Sun Xueling said reducing bus-lane encroachments could help prevent situations where bus captains have to brake suddenly to avoid collisions.
That matters because sudden braking can create safety risks for passengers who are standing or moving inside a bus.
The AI system could therefore serve a purpose beyond enforcement: keeping bus lanes clear may help buses move more predictably while reducing situations that force drivers into sudden evasive manoeuvres.
Singapore is also upgrading bus safety in other ways
The AI camera rollout is part of a wider push to improve public-bus safety.
LTA said it will progressively deploy a fully automated safety audio-announcement feature across the bus fleet from the end of 2027, alongside an upgrade to the Bus Fleet Management System.
The feature has already been piloted on 11 bus services since August 2025, with announcements reminding passengers to hold onto hand grips and grab poles before buses move off.
Meanwhile, Singapore has been progressively equipping buses with technologies including 360-degree collision-warning systems, anti-fatigue technology and camera-mirror systems. Existing buses will be retrofitted where feasible.
The bigger picture: Singapore’s roads are becoming increasingly data-driven
The latest AI bus-camera rollout reflects a larger transformation in how Singapore manages its transport network.
Instead of depending exclusively on fixed cameras, roadside officers or periodic inspections, authorities are increasingly combining AI, video analytics, connected systems and mobile camera platforms to monitor what is happening across the road network.
The potential advantage is scale.
A bus travelling its normal route can simultaneously transport passengers and collect information about road conditions and traffic behaviour.
If the technology performs reliably, thousands of daily bus journeys could provide LTA with an additional stream of information about problems that might otherwise take longer to detect.
And that is what makes the development particularly significant.
The next time a bus passes through a road, its cameras may not simply be recording what happens around the vehicle — AI could be analysing it too.
What happens next?
The AI-powered bus-camera system is scheduled to begin progressive deployment from Q4 2026, with expansion to all bus routes targeted by 2030.
For motorists, the message is straightforward: bus lanes are becoming harder to ignore.
For commuters, the potential benefit could be even broader — faster detection of road defects, safer bus operations and quicker responses to problems across Singapore’s streets.
And with cameras already moving through virtually every part of the city, the question may no longer be whether Singapore can see more of its roads — but what else AI will eventually be able to detect.
Leave a Reply