Multibillionaire, tech genius and superhero, Tony Stark, had one of the most impressive houses in the fictional Marvel Universe. Its AI system takes charge of the heating and cooling of the house, keeps intruders out and even runs Stark’s conglomerate for him.

We may still be a while away from being able to talk to our houses. But Singapore is already working towards AI-powered public housing estates that learn and adapt to residents’ behaviour. The ultimate vision is to have estates “‘trained’ to be self-sensing and self-thinking, and constantly learning to improve” using data from smart amenities, says Er. Dr. Johnny Wong, Group Director for Building & Research Institute of the Housing & Development Board (HDB).

HDB houses more than 80 per cent of Singaporeans, making the agency the nation’s “largest master planner and housing developer”, Wong says. He shares how HDB is using 3D models and predictive data to create more liveable spaces in Punggol Northshore, one of its housing districts.

Sun’s out, plants out

Tropical Singapore gets a lot of love from the sun, and this can be both a good and troubling thing when planning for housing estates. Increasingly, HDB uses 3D models and environmental simulation tools to understand how to create cool, comfortable neighbourhoods while making the most of the sunlight.

For instance, in Punggol town’s Northshore District, HDB studies factors such as wind flow, temperature, and shade when planning housing estate layouts. Urban planners can then decide how to orient buildings to increase ventilation, where they should introduce plants to cool overheated spaces, and which well-shaded areas can be turned into outdoor amenities like playgrounds and fitness corners, says Wong.

The agency also looks at how much sunlight each area gets. This helps planners identify the best locations for installing solar panels in the future, he adds.

What’s more, HDB has built a more advanced environmental modelling tool to make simulations more accurate. This new system, built together with the Singapore government’s research institute, examines how different environmental factors interact. For instance, it shows how heated surfaces influence wind flow.

Design driven by human behaviour

HDB will use behavioural science to plan its future housing estates. It is running a study to predict how demographics in residential towns are likely to evolve over time, and understand their needs and lifestyles, says Wong.

The project, dubbed the ‘New Urban Kampung’ research programme, alludes to the close-knit spirit in the village communities of Singapore’s past, but its strategy is forward-looking. It will use a combination of human traffic sensors, advanced modelling tools and traditional surveys to understand whether residents will be receptive to a new initiative, and what they look for in a neighbourhood.

The goal is to create estates that serve residents to the fullest potential. If people prioritise easy access to amenities and urban greenery, for instance, HDB will incorporate these elements into its design. Human traffic sensors can also identify under-utilised spaces, which planners can then redesign to encourage higher use, shares Wong.

Predictive maintenance

The Punggol Northshore housing estate will collect data on the usage patterns of common amenities, including lights and fans. This could help HDB to lower energy consumption, as well as to pre-empt potential maintenance issues over time, Wong says.

For example, smart lighting in common areas have sensors that react to human traffic. When there’s nobody around, the lights automatically dim, potentially reducing energy use by 60 per cent, Wong notes. Fans in the neighbourhood centres at Punggol Northshore also run at higher speeds only on particularly hot days to optimise energy use.

Even the watering of plants is powered by sensors. The irrigation system detects the amount of water in the soil to determine watering schedules. “This is a water-efficient and less labour-intensive approach to maintain greenery and landscapes,” says Wong.

All this data is stored in a central data hub, “akin to a ‘brain’ of the operations in our public housing estates”, Wong says. Through this hub and a network of estate sensors, HDB can monitor and analyse the performance of our public housing estate services in real time.

Over time, this information can be used to plan maintenance cycles, and even pre-empt breakdowns. “The ability to monitor the health of these services in real-time can greatly minimise any disruption of services and inconvenience to residents,” Wong explains.

Not all of us will need our houses to design superhero gear or run businesses for us. But we will need housing estates that learn how to cool overheated spaces, lower energy consumption, and pre-empt breakdowns. HDB is making use of 3D simulations, predictive data and behavioral science to inch Singapore’s homes closer to that reality.