Transforming Industry Through Data Analytics

Chapter 4. Smart Cities

Urban populations are rising. In 2014, 54 percent of the world’s population lived in cities. According to the United Nations, this number is projected to increase to 66 percent by 2050. In 1990, there were 10 megacities (defined as a city with a population of more than 10 million). This number almost tripled over the next 30 years; in 2014, there were 28 megacities. The UN estimates there will be 41 megacities by 2030.

Cities currently consume 75% of the world’s resources and generate 80% of the greenhouse gases. Given the projected continued urbanization of the planet, we can expect these numbers to increase. It’s clear that we need better methods for managing cities. The issues we see today will only be amplified as the population grows.

Throughout history, humans have applied technological advances to improve cities. Skyscrapers, mass transit, wastewater treatment, and traffic lights are all examples of the use of technology to solve an urban living problem. We are at a turning point in the evolution of the city for which the use of information and communication technology in general, and the IoT in particular, can be used in a practical way to address a wide variety of issues facing today’s cities—in some cases, dramatically transforming the city in the process.

Specifically, technology exists today or is on the horizon to improve the following areas:

• Energy efficiency

• Traffic management

• Parking

• Environmental monitoring

• Structural monitoring

• Waste management

• Safety

• Crowd control

Environmental Monitoring

Environmental monitoring involves making use of sensors installed on devices along with location-based sensors to track conditions related to the environment in order to respond directly to an issue or collect data to be analyzed.

Sensors that track water usage can identify both macro- and micro-issues. Total water usage by building can be tracked and analyzed for outliers, allowing the city management to focus on high-value improvement areas. Similarly, water flow sensors make identifying and finding leaks much faster, reducing, and in some cases eliminating, waste (sensor-enabled valves can automatically close when leaks are detected). In addition, sensors on the water meters automate the process for collecting the data for usage and billing.

In parks and public areas, ground sensors monitor the moisture content of the soil to be used to more efficiently and automatically manage the water-delivery systems (sprinklers, drip-feed, and so on). As the sensors grow in sophistication, other nutrients can also be supplied automatically, providing specific amounts of fertilizer and food to the plants.

In a similar vein, sensors can automatically sample the air, collecting data such as temperature, barometric pressure, carbon dioxide, carbon monoxide, light levels, and moisture content. The only limiting factor with respect to what we can monitor is the sophistication of the sensor itself. Moisture content is particularly interesting because it’s not what comes to mind when considering air quality, but it has interesting follow-on uses including identifying where to plant trees that ultimately reduce the overall temperature and lower cooling costs.

Sensor packages deployed throughout the city provide a wealth of data for analysis to aid in making data-driven decisions.

Energy Efficiency

We will discuss the transformation in the energy sector in more detail later on. As it pertains to cities, buildings and streetlights are the obvious targets to improve energy usage. Advances in technology allow for streetlights that use bulbs requiring much less energy and that emit dramatically less “light pollution.” These improvements can be extended further when coupled with sensors enabling the streetlight to adjust its brightness. In addition, as solar power and battery technologies advance, the streetlights themselves can be disconnected from the electric grid, or even net providers of electricity, particularly during the longer days of the summer months.

Finally, the sheer number of streetlights and their distribution throughout the city make them prime candidates to house the aforementioned sensor packages (the city of Chicago is already doing this). These sensors are not limited to just the environment: audio, and even video, can be added, which makes for some interesting capabilities.

Buildings represent a large portion of the resource consumption of cities. The American Council for an Energy Efficient Economy (ACEEE) estimates that even a modest improvement in only 35 percent of the commercial buildings in the United States alone would save up to $60 billion annually.

An integrated building energy management system monitors heating, cooling, and electricity consumption optimizing usage in real time while also identifying problems and issuing appropriate responses. Predictive maintenance capabilities like those discussed in the manufacturing section will reduce cost and improve overall equipment efficiency.

These systems implement advanced control strategies such as the following:

• Multispeed fans and demand control ventilation technologies for heating and cooling

• Smart windows that lighten and darken the windows to let in or keep out sunlight

• Integrated heating, ventilation, and air conditioning (HVAC) decisions with the smart windows to choose the most efficient use of the energy

• Sensors on all electricity-using devices, adjusting their use based on demand, in particular, reducing usage automatically during peak times. For example, lowering light levels and turning off nonmandatory displays.

In addition to these specific improvements, there is the larger general benefit of having the data to analyze to understand usage and make decisions based on that data. A next step is integrating the energy management systems of each building with all the others, thus enabling a city-wide view of both usage and response to demand.

Traffic Management

Traffic management in cities is on the cusp of a major transformation. Cities around the world are already implementing advanced technologies to improve transportation to, from, and within cities.

Sensor-enabled public transportation vehicles can provide real-time status on location and expected arrival times. They also can adjust to demand, changing their stops, routes and frequency as needed.

Video analytics is progressing to a point that new capabilities are coming online. Real-time license-plate reading, originally used to enforce stoplight and toll-road laws, is now commonplace. This kind of surveillance-based video analytics can be expanded to include many other traffic laws, especially speeding.

The same systems can identify congestion and post alerts to commuters, and also integrate with the traffic-light system to improve traffic flow. This can then be integrated with emergency vehicles to aid first responders in getting to their destination as quickly as possible.

No discussion of traffic management is complete without considering the impact of autonomous vehicles. Real-world tests of self-driving vehicles are underway with estimates that they’ll be commonplace ranging from 5 to 20 years from now.

The transformative nature of this technology cannot be overstated. We have already seen a sea change in private transportation via crowd sourcing with companies like Scoop (for carpooling) and Uber and Lyft (for taxi services). These companies use connected and mobile technologies to change how we travel to, from, and within the city; but it is easy to see self-driving, electric cars take this one step further, either individually-owned private cars, or fleets of cars owned by a new kind of transportation business. Fleets of self-driving cars, shuttles, and buses owned and run by the city government will replace the current public transportation system and potentially extend it much further.

The benefits are profound. All-electric, autonomous vehicles have zero emissions to reduce pollution, are safer, and reduce congestion because they require fewer vehicles overall and can coordinate their activities. Package delivery services will also change. There is an opportunity to dramatically improve delivery services, a real problem on crowded city streets, particularly in conjunction with delivery drones.

It is not unreasonable to envision a time when autonomous vehicles will be mandatory in the city or parts of the city, and the use of human-controlled vehicles will be banned.

Parking

Adding sensor technology to parking lots and structures and making this information available to mobile-connected drivers is more efficient and reduces traffic (no more driving in circles looking for a spot). We already see this in a primitive fashion at some parking structures where the number of open parking spots by floor is displayed at the entrance.

Here is another case for which autonomous vehicles will have a dramatic impact, both near- and far-reaching. Autonomous vehicles don’t need to park at their destination. If they’re part of a shared usage scenario, they don’t need to park at all (only a temporary pick-up and drop-off). If they’re privately owned, they don’t need to park nearby. They simply can drop off their owner and then be directed to a parking area in a less congested part of the city.

A far-reaching consequence is the change in land use and architecture. When autonomous vehicles are ubiquitous, city buildings will not need parking structures, instead they can use the space more productively, but will need to provide mechanisms for dropping off and picking up passengers on a much larger scale than is currently needed. Further, valuable land will not need to be used for parking structures. This space can be put to better use for other buildings, parks, or open space.

The same applies to airports, stadiums, or other event venues. Huge swaths of land are currently used to park cars that sit there for hours and days. This kind of parking, located near the destination, will be extinct in the future.

Safety

As cities grow in size and complexity, keeping citizens safe becomes more challenging. In recent times, there has been an explosion in the number of surveillance cameras installed worldwide. Beyond the obvious benefit of helping law enforcement solve crimes and apprehend criminals, many studies have shown that video surveillance acts as a deterrent reducing crime by its mere presence. Video analytics technology is advancing rapidly, and with new software algorithms and powerful processors, new capabilities are becoming available.

Facial recognition is probably the most widely talked about safety-related technology and is beginning to see practical applications in real-world use, particular in airports. But the technology is more far-reaching than identification. Video analytics will be able to identify suspect behavior. For example, analysis shows it takes a certain amount of time to find your car in the parking lot, and there is a specific pattern to how people move from the entrance to their car. Video analytics can identify when someone does not match this profile and can alert security in real time.

Many cities are investing in extensive networks of surveillance cameras. Some of the challenges facing human operators is finding someone or something (e.g., a car) that matches a description and then having found them tracking them through the various camera feeds. This needs to be done both historically, after the fact, as part of an investigation, and in real time, as events unfold.

Video analytics is advancing to the point where the human operators will be able to input a description—for example, male, about six feet tall, wearing a green shirt—and in real time find all people matching that description. After an identification is made, the system will be able to track that person through the various surveillance camera feeds, also in real time.

After the fact, the human operators could institute a search with the same description, specifying the time and geolocation parameters and the system would present them with all people found matching that description. The system can then “follow” that person forward through time.

These examples of identifying suspect behavior, finding someone matching a description and tracking something through the system are easier than facial recognition in some ways and much more difficult in others. The sheer amount of data that must be transferred across the network and processed is mind-boggling and simply wasn’t possible prior to recent advances in network and processor technology.

These systems are not limited to government agencies. Private surveillance systems can identify suspect behavior and alert security, but they also can integrate directly with police and issue an alert based on the suspected behavior. An example would be detection of a firearm. Another is detecting a fire and notifying the fire department.

These are very dramatic uses of video analytics, but more mundane uses are potentially more meaningful. Video surveillance systems at intersections can be used to identify people in a crosswalk and delay the changing of the traffic light to green. The same systems can identify lost children in stores, airports, train stations, or any public area.

Earlier, we discussed sensor packages installed on streetlights. Audio sensors as part of that package can be used to detect a gunshot, triangulate its location, and direct police to the area. This is already being done using custom hardware. What makes it interesting is when the audio sensors are everywhere. It’s not difficult to envision a time when someone, anywhere in the city, can call for help (literally, “Help, police!”) and have the call detected, routed, and responded to automatically.

Crowd Control

Continuing with video analytics, a basic capability in the future is counting the number of people entering and leaving a building or area. This can be used to automatically detect violations of the fire code and alert the appropriate personnel (management or security). It can also be used to help the fire department understand how many people are still in the building, both from the count and by switching the system to an emergency mode that is specifically designed to interact with emergency services personnel.

There is another aspect to consider. It will be possible to automatically identify people entering a building and determining if certain individuals are authorized to be there. This can be done in a very sophisticated way through the use of facial recognition. A simpler method is to interact with their smartphones. In this way, the smartphone replaces the badges most companies use.

The Internet of Everything

We call it the Internet of Things, referring to sensor-enabled devices connected to the internet, but it’s really the Internet of Everything. Every “thing” in the world, literally everything, can be connected to the internet and interact with the environment. By way of closing out this section, let’s consider a few miscellaneous things that are not obvious candidates for the IoT.

First, trash cans: sensors installed on waste receptacles detect when the container is full and communicate that information to a central station. This data is used to make more efficient use of the trucks and personnel that collect the trash, skipping receptacles that aren’t full and prioritizing those that are, optimizing the route in the process. It might seem far-fetched but it’s already being done.

Secondly, consider the structural monitoring of bridges. Sensors installed on bridges collect data on the load and the environment (wind, especially) identifying issues before they become a catastrophe. The data is also analyzed and used to better understand how the structure reacts to different combinations of load and weather, ultimately helping to build better bridges. The same can be done for buildings, particularly skyscrapers, especially those in earthquake zones.

Finally, let’s address a bane of modern life: pot holes. Street maintenance is expensive and ripe for optimization. If the city’s maintenance services can identify cracks in the street, they can respond before the cracks become expensive-to-repair holes. Cameras on the underside of vehicles can use video analytics to identify cracks and transmit the data back to a central station. This is actually a much easier video analytics task than those already discussed. But, what vehicles? Post office delivery vans are one choice. They’re government owned and they go everywhere, but ultimately, every single autonomous vehicle can be used for this purpose, given that the hardware needed is already installed.