SINGAPORE: The Housing and Development Board (HDB) has rolled out the use of artificial intelligence (AI) to enhance worker safety at its construction sites.
This AI system will focus on two scenarios which are common causes of worksite accidents, based on data from the Ministry of Manpower, HDB said in a media briefing on Thursday (Sep 12).
The system will monitor workers who come within one metre of a non-barricaded edge with a drop of more than two metres and those who are under the path of heavy loads lifted by tower cranes.
Currently, construction work sites rely on manual supervision by site supervisors and Workplace, Safety and Health Officers (WSHO) to ensure compliance with safety standards. According to HDB, this is a “resource intensive endeavour”, requiring multiple WSHOs and supervisors.
Connected to CCTVs in the worksite, the AI system will help identify high-risk zones and monitor it in real time. When workers enter these high-risk zones, the AI will send a Telegram alert and a photograph of the situation to site safety supervisors, prompting them to take action.
“A centralised and automated surveillance system will enable safety personnel to simultaneously monitor different work sites round-the-clock, thus allowing greater coverage with less manpower,” said HDB in its press release.
It added: “The automated system will reduce the risk of inconsistent safety standards across different supervising personnel, as well as human errors that could occur due to fatigue.”
Trials have been carried out at the Clementi Peaks Build-to-Order (BTO) site since September, and are expected to end by the fourth quarter of 2020.
If the trials are successful, HDB said it could potentially scale up the tool to cover other areas to enhance worksite safety, such as workers standing in the way of vehicle pathways, site traffic management and workers entering confined spaces without permission.
When fully developed in the longer term, the tool could also be shared with other developers and agencies to improve construction worksite safety in Singapore on a larger scale, it added.
HDB collaborated with the National University of Singapore’s (NUS) School of Design and Environment on the system supported by AI Singapore (under its 100 Experiments (100E) Programme).
3D CONCRETE PRINTING
HDB said the AI system is one example of how it has stepped up its investment in technology to uncover newer, better ways of working, amid the pressures and challenges in manpower and resources faced by the built environment sector.
One other way it is doing so is exploring the use of 3D concrete printing technology to expand its design capabilities.
The process removes the need for moulds or formworks, allowing objects with “intricate detail or geometric forms that would be near impossible to create with traditional methods”, said HDB.
This will give construction architects and designers “greater flexibility and free play in their designs”, said deputy director of building design & precast system Heru Santoso Soedarsono.
“And since the printing process is highly automated, this reduces their reliance on manual labour,” he added.
While traditional methods could take up to several months to build a room-sized component, it would take just 13 hours – not including steel reinforcement bars – to complete with 3D printing.
There are plans use 3D printing for smaller components used in precinct designs, such as landscape furniture or architectural designs in common areas for some projects in Tengah and Bidadari.
It is also looking at the feasibility of using drones to inspect the facades of HDB blocks for cracks, spalling concrete and even corrosion, said HDB.
Currently, town councils use manual methods such as suspended gondolas, scaffolds, boom lifts or rope access which requires inspectors to work at height.
While manual inspections for one block usually takes a few days, drone inspections can reduce the time taken to a day or less.
AI paired with the drones will also allow HDB to tag the detected defects to a digital model of the building, and the system will deliver a report with the severity of the defects and recommended solutions.
In anticipation of residents’ concerns, HDB said residents will be notified in advance, the area below flights paths will be cordoned off for safety and that “any images of residents captured during (the) scanning process would be masked out, before the captured data is used for analysis”.
The drones were first piloted at Yuhua and Sembawang from July to October 2018.