Smart street lamps will be lighting up Geylang and one-north business park in Buona Vista early next year, now that ST Engineering has won the tender for the Lamppost-as-a-Platform project. The local conglomerate won the $7.5 million tender to fit lamp posts here with sensors and cameras that can collect a wide range of citizen surveillance data.

ST Engineering edged out five other contenders - including local telco M1 and multinational tech firms Accenture and NEC - to win the tender, The Straits Times has learnt.

With the surveillance information, government agencies can increase their situational awareness, detect potential problems and respond quickly to incidents, such as unruly crowds, train breakdowns and traffic congestion.

The details of this tender were first reported by The Sunday Times on 8 April. Prime Minister Lee Hsien Loong also mentioned the plans in his National Day Rally speech in August last year.

A spokesman for GovTech, which called for the tender and is the central agency behind public sector tech transformation, said ST Engineering was picked as it has overseas partners and the in-house capabilities to provide what is necessary.

The new camera systems can, for instance, analyse faces - down to race, gender and age - for catching speeding e-scooter riders. Artificial intelligence-based video analytic systems - which allow for facial matching against a database - will enable real-time facial analysis.

These include end-to-end solutions for sensors and inter-operable video analytics. Mr. Ravinder Singh, President of Electronics at ST Engineering, said, "We will bring our experience from our global deployments of smart street lighting, water and energy management systems to work with GovTech."

More than 100,000 lamp posts nationwide may eventually be 'smartened up' after the trial in Geylang and one-north. Singapore has a network of security cameras, but this plan takes it up several notches because of the use of more sophisticated technologies that can merge data sets and draw detailed analyses. The present cameras can capture data such as licence plates but the data belongs to separate agencies which may not actively share surveillance data and analysis.

Artificial intelligence-based video analytic systems - which allow for facial matching against a database - will enable real-time facial analysis. The sophisticated system can also detect, classify and record objects such as backpacks, personal mobility devices, bicycles, and car models and licence plate numbers.

For instance, if a personal mobility device or bicycle is travelling at more than 15km/h on footpaths, which is illegal, the data will be captured and an alert sent to the relevant agency. The speed limit will be reduced to 10km/h next year.

According to tender specifications, real-time kinematic technology equipment will also be mounted on the lamp posts to provide line-of-sight connection to driverless cars, to determine their precise locations to help them navigate and avoid collisions. Other sensors to be installed include those measuring temperature and rainfall changes.