Home System Design AI: the Key to Smart Cities

AI: the Key to Smart Cities

by Benchmark

Smart cities are grabbing many of the headlines in the technology sector at present, and interest is growing. Whether councils, Government or commercial entities, a growing number of organisations understand how ‘big data’ can be used to introduce efficiencies across a wide range of management, safety and security needs. In an ever-increasing drive for efficiency, modern solutions must be able to offer smart services.

When considering the potential on offer from smart cities, it is very easy to distance such projects from the solutions on offer in the security market. Our focus has always been on the prevention of crime, the mitigation of risks and the management of incidents when they do occur. As vital as these tasks might be, the reality is they don’t encapsulate the ethos of a smart city. However, that doesn’t mean the design and implementation of smart cities is not relevant to the security and safety sectors.

Just as security and safety only represents one small part of a smart city operation, so to do other technologies. Traffic monitoring and management are an important part of a smart city, but as with security, the technology does not offer a complete solution.

Energy management is also an essential part of any smart city plan, as it transportation, communications, critical infrastructure, building management and a host of other technologies.

The main difference between these other technology sectors and security is a simple one: the other sectors are actively working to force their way into the smart city market, whereas security is dragging its heels a little.

While it is easy to dismiss smart city planning because security is typically implemented on a site-by-site basis, the reality is a smart city is nothing more than a collection of smart buildings, smart spaces and smart streets. All employ a variety of systems, generally working in different ways, but with one ingredient which makes them truly smart.

The smart ingredient

Intelligent cities, streets, buildings and spaces all rely on one thing which delivers the smart benefits: data. It is the sharing of data which implements the intelligence, allowing systems to take actions as and when required. If incidents occur, data relating to the status of spaces, streets or buildings is used to help control other parts of the city’s essential infrastructure.

The collection of accurate data, in real-time, is essential to any decision-making process. If that data is inaccurate, untimely, or based upon suppositions, then any smart environment will inevitably fail to meet expectations. Indeed, situations could be made worse if unreliable data, or insufficient data, is used. As such, smart cities will be data hungry.

However, it isn’t a case of simply collecting as much data as possible. The requirement is to collect useful data, and to mine it well, identifying issues before they become problems, spotting trends and instigating actions proactively rather than reactively. The right data, used in the right way, is what underpins a smart city.

The value for all stakeholders – city officials, law enforcement, Government departments, councils, residents and businesses and organisations – comes from having accurate and credible data which can be actioned. It is too easy to think that all data is vital. For many applications, certain data will be important, but when this is transferred to a smart city scenario, it cannot be actioned.

There is an IT industry term: GIGO. This stands for ‘Garbage In, Garbage Out’, and it is very much true in smart city applications. Integrators looking to branch into the delivery of smart site implementations need to understand that ‘Garbage’ does not solely refer to inaccurate or outdated data. It can also include data which, while accurate and credible, simply cannot be actioned effectively for other tasks.

If you consider security solutions, they are effectively data collection and analysis systems. Because of the nature of security and risk mitigation, the systems are designed to capture accurate real-time data, and often to verify that data. The data collected is usually relevant to site and building status, and therefore can be exploited to achieve many intelligent building goals.

This fact makes security a more relevant inclusion in the big data used to support smart city implementations than many other systems. It is therefore critical that integrators look closely at how their systems can add value for users at the highest levels.

Significant data

Considering the various technologies involved in a complete security solution, it becomes possible to understand the vast range of data that is being captured in real time.

Access control systems capture and process data about who enters (and leaves) a building, space or even street if automated vehicle access control is deployed. It is also possible to track people or vehicles around a defined area, knowing who is where at any given time. This enables definite occupancy data to be captured, along with easily searchable records of who is on and off site for specific spaces.

Access control records can contain much more information that a user’s identity and access permissions. Supplementary information can include qualifications such as first aider, H&S trained worker, fire marshall, first responder, etc.. They can also link specific individuals with assets, tools or other items to enhance management decisions if or when an incident occurs.

Access control data can be used to manage intelligent tasks based upon specific individuals being in an area, or non-specific individuals from certain access groups, or persons with defined qualifications. For example, all trained first aiders on a campus could be found by location, allowing a faster response to medical emergencies, regardless of which business or organisation they are attached to.

Status reports can be based upon occupancy levels, expected personnel or vehicular flows, or controlling processes dependent upon whether buildings, streets or spaces are densely occupied or unoccupied.

Because access data is often used for muster reporting, payroll and security, the data has to be accurate and delivered reliably in real-time. As such, it adds an important data element to larger-scale smart projects.

Data security issues can be addressed by only the important information being viewable, such as vital qualifications or training levels.

Even within the security industry, the potential for exploiting the data captured by intruder detection systems is not always appreciated. Data that identifies users is predominantly based upon setting and unsetting of the system, or the implementation of configuration changes. However, if you consider the status information gathered by alarm systems, it does offer a number of benefits when creating an intelligent solution.

For example, buildings (or premises in shared buildings) can be identified as armed or disarmed, and departments with or without motion-based activity can be identified in real-time.

Whether using external or internal detection, the power of intruder-based systems is the ability to create double-knock scenarios, timer-based actions and advanced logic rules. The latest intruder alarm systems include a plethora of additional features and functions, many of which can be applied to create smart events.

In recent times, advances in processing power and the increased use of GPUs to manage off-loaded data processing has seen the potential for capturing and exploiting data from video surveillance grow in an exponential way.

Video surveillance data offers a number of benefits as it allows visually detectable information to be exploited without the need for human intervention. Today’s video surveillance solutions are proactive thanks to the increased use of IVA and AI technologies.

Video surveillance offers data which can be processed for a very wide range of status reports. These range from simple events such as motion-based triggers, through a whole range of intelligent analytics triggers, through to site status reports based upon occupancy, traffic flow, entry and exit times, condition reporting, etc..

As the use of GPUs increases, so object recognition based upon deep learning algorithms becomes a more used element. The recognition can differentiate specific objects, even though they might of a similar size, shape, and display the same behavioural characteristics. Adding another layer of possibilities, the development of deep learning technologies ensures the potential on offer from video surveillance in smart cities goes far beyond security.

In essence, captured visual information can be processed as a part of a smart solution. Data can be collected about moving or stationary objects, appearing or disappearing people or items, speed and direction of flow, size and shape of objects (or object classification), dwell times, behavioural trends, anomalies and exceptions, etc..

In summary

It is important integrators do not think in terms of a smart city controlling every single function via an integrated solution. Smart cities will have legacy systems that perform certain management tasks, and these will continue to exist.

Additional smart functions will exploit the data being captured. For integrators, the key to additional business lies in ensuring users appreciate the power and flexibility of modern solutions.

Related Articles