How To Build a Holistic Smart City Architecture

Smart cities leverage the power of IoT to improve service delivery and optimize city infrastructure. Here, drivers don’t need to search for parking spots – they receive automatic notifications; waste containers ‘tell’ truck drivers when it’s time to collect waste; there are no traffic jams - smart traffic lights automatically adjust signal timings based on real-time traffic analysis, etc.

These examples show what happens ‘on stage’, but what is behind the curtains? In this article, we'll answer this question, reveal the components that make smart cities function and present our own concept of a smart city architecture.

A Contextual Approach to a Smart City Architecture

In designing a smart city architecture, it is important to think about the context. It means we must have a clear vision of the environment it will be integrated into and the people it will interact with. With this in mind, we’ve worked out a three-dimensional architecture concept, consisting of:

• IoT-based smart city platform
• Service management solution
• Citizen portal

Well-rounded and comprehensive, it will help cities to deal with large-scale streams of data, automate service requests, and create a community of engaged city residents. And now let’s dwell on the details.

IoT-based Smart City Platform

The Network of Smart Things

As any IoT system, a smart city uses smart things equipped with sensors and actuators. Sensors collect data and pass it over to the cloud for processing. For example, connected streetlight’s sensors measure illuminance, sensors installed in waste containers measure the amount of waste, road surface sensors measure the average traffic speed, etc.

Actuators make things act. They receive and execute commands from control applications and user apps, e.g. alter traffic lights, switch on and off the lights, etc.

Field Gateway

The data gathered by sensors cannot pass directly to the cloud - it goes through field gateways. They preprocess and filter the data before passing it to the cloud. They also transmit commands from control apps to things’ actuators. Additionally, with local intelligence, field gateways can provide time-critical responses even when cloud connectivity is lost.

Cloud Gateway

The cloud gateway addresses security issues and ensures safe data transmission between field gateways and the cloud. It protects data through encryption, prevents unauthorized access and interception, as well as provides compatibility between different protocols.

Streaming Data Processor

In smart cities, data is generated continuously by thousands of sensors, which send data records simultaneously. A streaming data processor makes it able to act on this data directly upon receiving it. It passes the data over to control applications and loads it to a data lake.

Data Lake

A data lake is a data repository used to store data in its raw format. When data is needed for meaningful insights and its value is identified, it is extracted from a data lake, structured, and loaded to the big data warehouse.

Big Data Warehouse

If you think of data as of some amount of water, a data lake is a large storage pool, while a big data warehouse is more like the number of packaged water bottles with cleansed water. The big data warehouse contains structured data and contextual information about connected things, e.g. where and when they were installed, as well as the commands sent to things’ actuators by control applications. The big data warehouse makes the reuse of sensor data possible allowing different smart city services, like smart traffic or smart environment, to access and use the same data sets. 

Data Analytics

Data analysts use data analytics solutions to examine sensor data sets in order to draw meaningful insights and communicate results to data analysts. To make the results of analysis easier to perceive, they are visualized with the help of data visualization tools. Data analytics and visualization tools are often integrated into dashboard applications that display data on a single screen and can be updated in real time as new information becomes available.

Data analysts use data analytics tools to monitor traffic performance, reduce accidents, identify potential crime scenes, etc. For example, using data analytics tools for monitoring traffic over a period of time, it is possible to reveal patterns of traffic distribution across the city and make significant progress in relieving congestions, emissions and noise.

Machine Learning

Machine learning uses advanced computational techniques to adapt the “behavior” of smart things to the needs of citizens. Machine learning algorithms are applied to reveal hidden correlations and, based on these correlations, build predictive models.

These models determine how connected things will react to certain conditions. Models are tested for accuracy and, if the output action differs from the expected one, they are revised and tested again until they function as intended. Then they are used by control applications, which send commands based on the models to things’ actuators.

For example, ML can be applied to a traffic management solution. Monitoring traffic over time, ML algorithms create patterns of traffic distribution across the city. Based on these patterns, cities can detect the areas with the heaviest traffic load and take steps to relieve them without human intervention, e.g. adjust traffic lights timings, reroute part of the traffic, etc.

Control Applications

Control applications send commands to things’ actuators. For example, actuators of a streetlight can receive a command to brighten the lights when a movement is detected.

Control applications can be rule-based and machine learning-based. Rule-based control applications use rules programmed manually. This way, variables in the rule are substituted with incoming data records. If they meet the conditions defined in the rule, the output action is triggered.

For example, a city uses a traffic management solution to identify and relieve congestions. For that, traffic management platform users define a speed threshold which signals that there is a congestion, say, if the average traffic speed drops to 9 mph, the rule identifies a congestion and takes an output action – alters traffic lights.

Machine learning-based control applications use the models created by applying machine learning algorithms.

User Applications

With user applications, citizens can send commands to control applications and receive notifications and alerts. For example, user apps can receive notifications when a parking spot has vacated. A driver can also use the app to view the map of parking spots in the area, register their car for the spot and pay for parking, as well as extend parking time.

Service Management Solution

The service management solution helps to provide timely support for citizens and address their service requests. The service management solution gathers citizens’ requests from a variety of different channels, including emails, online communities, social media, web chats, etc. It processes requests, automatically creates cases and assigns them to agents.

For example, if a citizen notices that a traffic light is out of order, they can take a photo, enter additional information, mark the location on the map and report a case, using a mobile application. Service management solution receives a query, processes it, creates a case and automatically notifies a field worker.

Citizen Portal

The main purpose of smart cities is not mere automation, but making citizens’ life better. That is why a smart city infrastructure is not complete without a citizen portal. A citizen portal creates a common space for city administration, employees, field workers and citizens. A successful citizen portal should address the needs of city dwellers. That is why it is important to examine citizens’ needs first. As a rule, citizens want to:

• Get their questions answered in time.
• Be able to see the status of their requests and activity.
• Quickly access required information.
• Leave feedback and recommendations.
• Report malfunctions and file cases.

For example, the city of Chicago has created a citizen portal that allows citizens to report smart things’ malfunctions, access knowledge bases to find information on how connected things work, view the data on where connected things are installed and which of them are out, etc. Moreover, citizens can view a live map of road closures, broken streetlights, potholes and much more.

In Conclusion

Here is a quick summary. Starting an IoT-based smart city development, you need a basic smart city platform, consisting of:

• The network of smart things for gathering data.
• Field gateways for facilitating data gathering and compression.
• Cloud gateway for ensuring secure data transmission.
• Streaming data processor for aggregating numerous data streams and distributing them to a data lake and control applications.
• Data lake for storing data the value of which is yet to be defined.
• Date warehouse for storing cleansed and structured data.
• Data analytics tools for analyzing and visualizing data collected by sensors.
• Machine learning for automating city services based on long-term data analysis and finding ways to improve the performance of control applications.
• Control applications for sending commands to the things’ actuators.
• User applications for connecting smart things and citizens.

Besides, to ensure you identify problems before they pose any threat to citizens’ wellbeing and handle citizens’ requests in a timely manner, you need a service management solution. And finally, to engage citizens and create a convenient space for communicating, a citizen portal is vital.

Written by Alex Grizhnevich, process automation and IoT consultant at ScienceSoft.