printer friendly version

A smart start for South African cities

Gregg Sanders.

High crime rates are a reality of South African life. For that reason, a number of our cities have already created camera-based monitoring systems to deter criminals or identify and prosecute them if they transgress the law.

One example is Safe City Msunduzi, an entity of the Msunduzi Municipality, which monitors 169 CCTV cameras across Pietermaritzburg. Some are going even further, like the City of Cape Town, which is building on its camera-based system to deploy drones, gunshot sensors and a tech-heavy highway patrol unit that will automatically scan number plates.

Steppingstone to a smart city

While crime remains an unfortunate negative, camera-based security implementations could provide an unintended benefit. They provide a natural springboard for strategic smart city management systems, which also use cameras, but add sensors and analytics engines, ideally consolidated in a central command and control centre.

An IoT-based monitoring system could use smart data analytics to collect, correlate and flag relevant data. Machine learning can identify trends and outliers in critical infrastructure such as water and electricity supplies to automatically dispatch technicians to deal with outages, or even predict outages and prevent them in the first place.

We see more smart technologies in other areas such as transport, where cities use smart sensors in roads to improve traffic flow, and connected buses give commuters accurate travel times. Such systems could monitor and automate (with human intervention where necessary) a broad range of city processes and intelligently allocate resources for greater efficiency and better services to citizens.

During large events, a combination of camera-based monitoring and backend analytics could automatically re-allocate public transport from low-volume routes to deal with sudden peak demand outside sports stadiums, for example. Cameras could also detect accidents and send first responders to handle medical needs and manage traffic while clearing the scene to minimise congestion. The potential applications are myriad, but they tend to start with cameras, which are already in place in many instances.

Cameras on their own are not that exciting. They still require humans to watch screens and pick up events. Generally, dashboards and reports tend to be retrospective. City leaders want and need a view of events as they unfold. They need automatic alerts when thresholds are exceeded, and automated processes are initiated in response to events.

When you start using analytics to identify people and events, the potential benefits become really exciting. Appropriate camera technology connected to a smart backend can kickstart the process and help city authorities understand events and the patterns around them. AI systems learn continuously, and while humans are still needed to verify some events through direct observation, proper analytics can cut down the need for human monitoring of screens by about 70%. Those people can be reallocated to more important tasks, or tasks that really require human intervention.

Using AI, machine learning and process automation, important alarms are raised and key contacts alerted automatically. Mean time to respond (MTTR) to critical events can be massively reduced while also eliminating human error and ensuring process adherence. We have seen the following real-world results:

● Average MTTD (Mean Time to Detect) reduced from over 2 hours to less than 1 second.

● Average MTTR reduced from over 6 hours to less than 2 hours.

● Monitoring staff reductions from over 400 to under 170 (the staff were re-deployed to more critical areas of the business).

Connectivity, power and organisational silos

Connectivity is currently the smart city’s biggest potential Achilles heel. You may have hundreds of cameras, but without connectivity, they cannot feed into a real-time smart city system. Fixed-line systems with large groups of people monitoring screens are being replaced with wireless systems, where analytics capabilities are built into the cameras themselves. However, you have to make sure the connectivity is built and in place and, in a country affected by loadshedding, redundant power supply is also a necessity.

In a country plagued by load shedding, power is another obvious challenge. Battery backups with solar technology provide an obvious solution, so that each site becomes a self-contained unit with built-in connectivity, power and analytics. This provides a natural redundancy in case of failure in a particular unit.

Something we find all too often is that cities are not enjoying the full benefit of all their systems because they deploy disparate systems in silos. A smart system should ideally provide centralised command and control with feeds from all sensors in the city’s ecosystem. In South Africa, so many city departments are not linked to a backend that consolidates all city data. Consolidation makes cross-referencing possible and creates benefits that may even have been invisible.

For example, water and power should naturally interact, as pressure pumps that fill reservoirs are obviously power-dependent. Smart data analytics across both these functions could provide crucial alerts that flag or even prevent possible water shortages due to early detection or prevention of power outages. In an ideal world, one could even integrate all cities in a centralised national view and reallocate water from one city to another for example – but this is a long way off.

Working smart by building for the future

It is important to build with your end goal in mind, to install the appropriate underlying technology. Is the system just there to monitor? Do you want to send an alarm if there is an unusual event? Alternatively, even take the next step and automate certain functions like rerouting water supplies in the case of a leak. Existing systems may have limited camera capabilities in place, depending on your needs you might need to replace static cameras with pan, tilt and zoom capabilities linked to backend analytics that include license plate and facial recognition.

In terms of sustainability, strong ecosystem relationships may be just as important as redundant infrastructure with the necessary capabilities. One way to collaborate is to drive community involvement in smart city efforts. This not only educates the public around the possibilities, but also illustrates convenience, builds trust between a local government and its citizens and ultimately drives adoption. Educational programs for children at school level is a good way to encourage early acceptance of new technology.

Local and national governments do need to look at the regulatory environment to ensure it is conducive to public and business involvement in smart cities, but without relevant buy-in from the public, adoption could be slow. Public RFIs can also drive involvement from big business able to support smart city initiatives.

We need smart city solutions that meet South African requirements. Build on camera-based systems by deploying IoT-based monitoring systems for water, electricity, environmental and safety areas. Digitalise city processes, automate, understand the patterns and implement preventative maintenance instead of dealing with events after they occur.

Find out more at www.nec.xon.co.za

Share this article:

Further reading: