Smart Cities, Challenging Issues

Smart cities are coming. Not everything will be connected, and not everything will be connected at once. Still, governments around the globe are beginning to tap into a world of connected devices and sensors for reasons ranging from cheaper lighting to less traffic, lower crime, and improved air quality.

Smart cities encompass all manner of usage models and equipment — parking meters, traffic lights, power and water meters, mobile telephone networks, apps on every resident user’s handheld phone, including cameras and microphones. The engineer’s challenge is integrating all of these devices using common communication links, ubiquitous GPS technology, algorithms that make sense of the data collected, and central repositories for relevant data.

“You need a horizontal platform,” said Mike Muller, chief technology officer at ARM. “Without that horizontal platform you don’t get innovation. You need the ability to let someone come in from the outside and go, ‘Here’s a brand-new thing that you can do by shuffling all of the pieces you had before.’ In a vertical application, people will have a view of how those pieces come together and what those pieces are. But you don’t enable the innovation from the outside without making it horizontal.”

Motivations behind smart cities vary greatly from one city to the next, and from one region to the next.

“In American cities it’s quite clear that IoT is a way of reducing operating costs and taxpayer burden,” said Muller. “In a lot of Eastern deployments, it may be, ‘We want to do something about emissions controls.’ The whole motivation is very different and you may end up with completely different solutions. The driver there may be air quality rather than an energy-saving issue, which would be the classic value proposition for home control in the West.”

There are enough studies underway or completed to show that collecting, mining and utilizing data can lead to all sorts of improvements. “Research by (UCLA urban planning research professor) Donald Shoup has found that up to 30% of the traffic in metropolitan areas is due to drivers circling business districts in order to find a near-by parking space,” Darrell West, founding director of the Center for Technology Innovation at the Brookings Institute, writes in a just-released policy paper.

So how do traffic cameras and entire systems for alleviating the typical Bay Area or D.C. denizen’s 80 minutes a day in traffic account for the human urge to spend an extra 20 minutes in traffic rather than park and walk an extra 5 minutes? How do cities keep people safe, so that people can get to school, to work and live together peacefully?

“The trend is to build platforms that decentralize intelligence at the edge,” said Remi El-Ouazzane, CEO of Movidius, which developed a compute fabric for what it calls Vision Processor Units (VPUs). Those VPUs capture images, process that data, and then using machine learning make assumptions about what’s occurring right next to the sensors, not necessarily back in a data center or the cloud. Intel announced it would acquire Movidius in September of this year.

A Movidius device is inside a new generation of cameras from Hikvision of China, which is capable of counting cars in a given traffic flow, identifying accurately 96% of the time the exact make and model of the car in the image captured. This lends itself to changing the timing of traffic lights in a given sector of a city, as well as working in concert with law enforcement and security agencies to spot a criminal.

But the on-board processing power and software learning in the camera system can go further, anticipating a crime might be about to occur.

“This allows cameras to prevent things from happening. You can simulate or modelize the behavior that leads to a theft or a crime. Our cameras might see that someone is placing a phone call and is driving and texting, for instance,” said El-Ouazzane. “We can understand in real time whether or not a package was left unattended. The goal is less toward reacting, and more toward preventing an event. The deployment of artificial intelligence in cameras that will be used in cities can actually make the cities safer.”

Safety vs. privacy
That assumes the same kinds of cameras will be used everywhere, of course. But a major sticking point for rolling out certain features in ultra-connected environment such as a smart city is privacy.

In fact, privacy concerns are driving demand for lensless optical sensors. Rather than producing images with the kind of visual acuity of a camera, they generate data and rough images.

“This is a new kind of optic that is extremely flat and easier to use,” said Patrick Gill, principal scientist for Rambus Labs. “So instead of generating an image with a photo-sensitive array and a lens, which is what you find in a camera, this replaces the lens with a refraction grading. You can still see people moving and tell whether they’re sitting, walking, avoiding certain regions, and identify a change in the traffic flow or pick up moving car headlights. But it also allows you to replace something that’s analog with a binary diffraction grading. It’s easy to manufacture—you can do this as large as 2 microns—and it’s a very good detector of motion. You also can have a wide space between sensors because these can have a 140-degree field of view, or even wider in some cases.”

Gill said the first use of this technology was expected to be in eye tracking in virtual reality and augmented reality. While there is still a lot of interest from those markets, the negative reaction to having cameras everywhere suddenly raised interest in this technology in places that no one had seriously considered in the past.

“This gives you more detail than a motion detector, but it also combines an element of privacy so you’re not going to see something end up on YouTube that you don’t want to be made public,” he said. The modules are in the $1 range, as well, which is far less than a camera, and the price tag is expected to drop even further as volumes increase. “So one deployment might be on a multi-function chip where you’re measuring temperature and humidity, but where it’s too expensive to integrate a lens. What’s noteworthy is that all camera modules do not survive a solder re-flow. The lens would come out as a puddle. But you can make a diffraction grading with a high-temperature polymer. And there are few things that can do this using low power.”

Power matters
Power is one of the constant themes in all of these implementations. Utility bills are line items in government spending, and savings through deployment of smart technology can be in the millions or even tens of millions of dollars per year. So not only do technologies in smart cities need to do something valuable, they also need to do it using less energy.

Movidius’ El-Quazzane pointed to two critical factors in its development—a power envelope that was less than 1 watt, and the development of proprietary machine vision algorithms and software ahead of the hardware so that the device could accomplish what the software was asking of it. “Machine vision does not give you any choice. You have to think of the entire system together, from the output of the algorithm down to the implementation of the gates.”

Keeping that max of 1 watt power consumption was critical, because of where the company saw these cameras going — pole mounted camera systems that could be run solely from a solar panel, or in the very limited power budget of a drone.

Nor is this just about vision. SST, a Newark, Calif.-based startup, has developed a technology called ShotSpotter that listens for and pinpoints gunshots. The technology is in use in more than 90 cities. Company CEO Ralph Clark said that the power envelope of the systems his company deploys is critical, and more component makers need to step up.

“My feeling is the solar and battery combination is going to have to get a lot better,” said Clark. “We have it working in Kruger Park in South Africa, where ShotSpotter is deployed to stop rhinoceros poachers, but that gets an awful lot of sunlight.” The normal mode is for every roof-mounted sensor system to be direct-wired.

The ShotSpotter compute platform today is a series of standard components, Clark said, a processor, memory, storage, an analog-to-digital converter connecting to the microphone capturing audio, as well as devices for mobile network connectivity and GPS location.

“We designed our own firmware running on this. It processes when we find three or more sensors detecting the same boom or bang sound, then it ignores the ambient noise, puts a timestamp on it, and sends that captured audio to the cloud with the metadata about location.

“Our SLA [ service level agreement ] is for that location information to be within 25 meters of the incident. It’s often about 10 meters, though,” he said. Police and security officers have this information on their smart phones within 30 to 45 seconds of a trigger being pulled.

And here is where the not so logical but ever-important human behavior part comes in. “We want law enforcement aware any time a gun is fired, which in most places is a felony–80% to 90% of all shootings never get reported. Sadly, that means the gun violence is normalized in that community. This is a way to denormalize that,” he said. “We think every shooting is worth investigating.”

The Brookings Institute’s West talks about the need for all city dwellers to understand the importance of water and air quality for Smart Cities, for sustainability to always be a part of the picture. Traffic congestion and mobility is a big factor in that. The world’s largest parking meter and transport ticket provider, Parkeon (formerly Schlumberger’s parking division), is deploying more and more electronic systems to bring that about.

In 2014, 54% of the world’s population lived in urban areas, and this figure will be nearly 70% by 2050, according to Parkeon. City centers will require increasingly greater technology investments to keep mobility under control, and congestion in check.

Only a few years ago, the average parking meter was a purely mechanical affair, with a meter for time and a bin to hold the coins. Today is a complex $10,000 system with embedded computing and rugged touch screen displays. Now, a person can park and use an app on their phone to alert them when time is up and add more money to the meter. Cities invest, and benefit, when they capture many more of the payments electronically.

But the reach of these technologies is deeper than one might imagine at first glance. By creating platforms, standards evolve that allow other devices and services to be connected.

“Once you have standards playing out in a smart city, you start seeing how a smart city plays with smart power into industrial,” said ARM’s Muller. “In some ways, consumer has the least buying power because you’ve got a relatively small number of pieces of domestic kit and you’re trying to link them together. This plays out like the remote control wars, where the company wants you to buy the video player that is the same brand as your television because that way you have one remote that makes both of them work. And the number of players and dollars is limited. But in a smart city, you are spending millions of dollars from clearly differentiated suppliers to build a system that has to all work together. That’s where a lot of the standardization will come.”

—Ed Sperling contributed to this report.

Related Stories
Smarter Cities
Technology catches on quickly as way of reducing costs, but big challenges remain.
GridComm: Smarter Grids
Singapore startup utilizes powerline infrastructure to reduce energy bills, overcoming longstanding challenges with this approach.
Smart Grids, Smart Cities (2015)
Today’s smart grids will power the smart cities of tomorrow, but how secure will they be?