GovTech expert shares challenges and considerations of building a smart city

Mr Christopher Tong, engineering project manager at GovTech, discusses the challenges and complexities of scaling smart city solutions.

Dense clusters of buildings, sprawling infrastructure and hordes of people do not a smart city make. To become truly ‘smart’, cities need to sense and respond to the needs of its denizens, and that ambition begins with data.

Like the nerves of the body that send sensory feedback to the brain, networked sensors powered by the Internet of Things (IoT) allow the remote collection of data, which can then be analysed by city planners for insights.

“The basic idea is to install sensors throughout the city which collect data that can be used for different applications,” said Mr Christopher Tong, engineering project manager at the Government Technology Agency of Singapore (GovTech), during a brown bag seminar held at GovTech’s Mapletree Business City headquarters.

‘Smart’ must also be scalable

Given an empty plot of land, or ‘greenfield’ territory, a city planner or architect can design and create anything he or she wishes. However, in reality, smart city projects are generally ‘brownfield’ developments, where new capabilities have to be overlaid onto existing structures.

Mr Tong thus highlighted that smart city planners should adopt the approach of ‘retrofit, not replace’. The smart lamp post, a project that GovTech’s Sensors and IoT (SIoT) team works on, is a case in point. Rather than replace ‘dumb’ lamp posts across the island with new ones that come with a variety of sensors built-in, why not just mount those sensors onto the old lamp posts? Not only is this more cost-efficient, but also allows for smart city solutions to scale more easily, Mr Tong said.

“For us in the SIoT team, we’re always playing on the two themes of: what we can do in a given environment, and how we can improve things economically,” he added. “That’s the key difference between applications and operations—applications are all the wonderful use cases we can think of… but operations are about engineering and maintenance. When applications and operations complement each other, we have a better chance of actually scaling smart nation solutions out.”

Of options and trade-offs

When it comes to the practical implementation of a nationwide sensor network, smart city planners need to contend with the issue of powering each sensor almost indefinitely if a constant stream of real-time data is required. Here, Mr Tong noted that there is a trade-off between power consumption and computation.

Devices that need to perform computationally intensive functions typically need “to draw power from the mains”, he said, adding that such power sources may not be easily available or accessible. Solar energy might be an option, although frequent cloud cover in Singapore may result in intermittency in power generation. In the long term, the ideal setup would be to use a combination of low power devices, improve energy storage, and perhaps raise the efficiency of energy harvesting, Mr Tong explained.

Because power consumption corresponds to the intensity of on-device computation, an alternative power-saving strategy would be to offload data to the cloud for analysis. However, Mr Tong noted that for heavy bandwidth applications like people-counting in videos, “you might not even be able to send all that data to the cloud”. Some processing at the edge may be necessary so that only the numerical data is sent to the cloud. Therefore, smart city planners will need to carefully evaluate each use case to determine what hardware, software and network specifications are most suitable.

“So when I think about the projects that we’re doing [at GovTech], these are some of the considerations that we have to look into. There’s no easy answer for any of them, but we’re figuring things out every day,” he concluded.