How dConstruct Robotics and EM Services are reimagining routine HDB inspections with autonomous AI-powered robots that navigate corridors, ride lifts, and detect defects — so property officers can focus on higher-value work.
dConstruct Robotics is an award-winning AI and robotics company based in Singapore, focused on building smarter robots for a better world. We deliver end-to-end solutions for inspection, surveillance, and data-driven operations across facilities management, construction, smart city infrastructure, and urban logistics.
When we joined the BEAMP programme, we were looking for an opportunity to put our technology to the test in a real-world environment — and we found exactly that in a collaboration with EM Services.
A Challenge Hiding in Plain Sight
EM Services is one of Singapore's leading township management companies, dedicated to making communities better places to live, work, learn, and play.
Their property officers routinely inspect the common corridors and staircases of HDB blocks, checking for defects and hazards — a task that is essential but repetitive, physically demanding, and labour-intensive. With hundreds of blocks under management, the time and manpower required presented a clear opportunity for automation.
The challenge was straightforward but far from simple: could a robot reliably navigate the narrow corridors and multiple floors of an HDB block, detect real defects, and do so without disrupting residents?
Building a Solution That Moves Between Floors
Working closely with EM Services, we deployed our d.ASH-ER robots to perform autonomous corridor inspections in a live HDB environment. The robots generate accurate 3D models of HDB environments using SLAM technology, enabling reliable navigation and spatial awareness.
Critically, we demonstrated successful robot-to-lift PLC integration, allowing the robots to operate seamlessly across multiple storeys — a capability essential for any practical deployment in high-rise residential buildings. With a minimum of six hours of continuous battery life, the robots can inspect entire blocks in a single run.
On the detection side, the system performs near real-time video analytics to identify and classify building defects and hazards. A remote tele-operations capability allows operators at a command centre to intervene when needed, and all operations are managed through a centralised robotics management platform compliant with TR93 connectivity requirements.
Earning Trust on the Ground
One of the less obvious but equally important aspects of this project was public acceptance. Deploying robots in a residential setting requires trust.
Together with EM Services, we introduced early public education efforts — posting announcement notices in shared areas such as the void deck and lift lobbies — to reassure residents about the robot's presence and inspection operations.
The response was positive, and the approach demonstrated that thoughtful communication can smooth the path for new technology in everyday environments.
What’s Next
The proof-of-concept confirmed that autonomous robots can take on the repetitive, physically demanding corridor inspections currently handled by property officers — improving manpower productivity and freeing teams to focus on tasks that require judgement and human interaction.
With the technology proven in practice, EM Services is now exploring how such solutions can be deployed across similar inspection scenarios.
On our end, we plan to continue enhancing the solution based on what we learned during this project — improving navigation, perception, and system performance, and expanding into areas such as delivery, towing, and broader inspection applications.
We will also continue refining our AI models and robotics management platform to support more reliable operations in increasingly complex environments.
The BEAMP Difference
Participating in BEAMP gave us the opportunity to validate our technology where it matters most — in the real world. Through our collaboration with EM Services, we were able to deploy in an operational HDB setting, gaining invaluable insights into real-world constraints, user requirements, and system performance.
These learnings have directly contributed to the refinement of our robotics platform, from navigation and perception to system integration. The programme also deepened our understanding of industry needs and enabled closer collaboration with key stakeholders.
For a company like ours, that combination of real-world testing, structured support, and industry access is exactly what it takes to build scalable solutions that are ready for deployment.
Lim Chinn Hwa
CEO, dConstruct Robotics
