AI • IOT • NETWORK EDGE

It’s happened to us all. It might be late at night; it might be out in the country. You’re sitting at a lonely red light minute after minute, with no cross traffic in sight. It seems like the signal will never go green! Isn’t there some way—you wonder, tapping your fingers on the steering wheel with impatience—to trigger the light to change in response to the actual conditions on the ground?

Intelligent traffic management (ITS) could provide a solution. And not only for the impatient country driver. Commuters, city planners, even passengers and pedestrians have an interest in the use of up-to-the-minute technologies, including AI, to improve conditions on the road. Joseph Harvey, ITS Market Sector Lead from video analytics company Intelligent Security Systems (ISS), talks to us about the challenges and opportunities of making roadways smarter, safer, and more efficient for us all (Video 1).

Talk to us about intelligent traffic management.

For a very long time, I’d say that ITS has been grounded in pushback to technology—wrongly and rightly, because there have been some great solutions that were already working in this space. But as AI and neural networking come into everyday consciousness more, we are seeing more and more technology feed into the ITS area.

Now end-user agencies—the customers for ITS technologies—can see what solutions like the one from ISS can offer from a safety standpoint and from a data standpoint because of neural networking and video intelligence. So we’re seeing things change rapidly. The excitement in ITS now is in how the interconnected realm is really going to work with everyday motorists. And companies like ISS can be at the forefront of that conversation.

What are some ways you can integrate AI into traffic management?

AI, neural networking, video intelligence—they are at the core of every single one of our vast portfolio of products. Devices are gathering data from everyday motorists—urban, arterial, out on the freeway—in a manner that is 95%, 96%, 97% accurate. It’s in real time and it’s automated, so operators only respond to specific needs, which is valuable for an end-user agency.

For example, one of the products that we have here at ISS is a pedestrian-safety device: a dynamic illumination for pedestrians within crosswalks. When you think about the more traditional measures, you’re asking a driver to react to, say, that static yellow sign in a school zone that indicates: This is a school crossing. A pedestrian may be in the crosswalk.

What we have done is leveraged AI and camera technology to dynamically illuminate at dusk or nighttime hours a pedestrian, a child, a mobility device that is within the crosswalk, and actually show where they are. This was a revolutionary thing in the industry for our company to do; no one had done anything like it before.

How does an AI solution for ITS improve on traditional means of gathering traffic data?

The traditional road measure at a signalized intersection is something like magnetic loops in the ground. Or if you look up you might see a number of devices. But with the development of cameras along with AI we’re really seeing the ability for the controller, the smarts, the brains behind those intersections to have greater understanding of the environment and be able to react in real time if there is an incident.

The original adoption of cameras was a little hit-or-miss because of their unreliability at times, especially during weather events. But with the advancements in cameras and then with this neural network, the system is able to understand the environment and make adjustments on the fly. Instead of getting only partial data and outliers, instead of making assessments on ones and zeros that an engineer might be looking at, you can go back and actually pull video feeds and really drive true meaning and understanding from the data that you’re looking at.

With the development of cameras along with #AI, we’re really seeing the ability to have a greater understanding of an environment and react in real time if there is an incident. @isscctv via @insightdottech

Near miss is a really big topic for us right now. If someone calls in and reports that a car has gone through a red light at 50 miles per hour, traditionally an engineer would have to go out to the field and take a look at what had happened on the cameras. Now, with AI and with video intelligence, all of that is constantly at the fingertips of operators, and they’re able to react much more quickly. Or they can look at the data sets long term to effect change on the roadways when they’re considering design or considering reshaping the roadway itself.

What about that scenario of the seemingly endless red light and no cross traffic?

I think there are 300,000 to 400,000 signalized intersections in the United States, and somewhere around half of those are still without detection or else have some sort of outdated detection method. But absolutely: There is the ability to add the visual aids, the algorithms, and the video intelligence to advise the traffic light that a vehicle or vehicles are waiting there, and basically to place a call into the signalized controller in order to effect the signal phase and timing—or SPAT. That is something that we are presently doing, and we’re seeing it in the marketplace.

It limits congestion; it limits the environmental impact of a car sitting and waiting two, three minutes at some of these intersections. But it also then gathers all of that data: What time of day is it? What type of vehicles are there? Are there pedestrians? All of these data points are really helping engineers continue the ITS conversation.

And as more and more devices get connected and as more and more end user agencies are able to take the inputs of an alerted event and interconnect everything that they have at their fingertips, they’re making the road safer; they’re having a real impact at time of event. This is honestly the reason a lot of us are in this industry, because we can see that change, and right now we are seeing it happen very quickly.

What are the challenges to implementing these solutions?

What ISS is looking to do is leverage the infrastructure an end-user agency might have in the field already—say, the cameras—and build on top of that. That’s what scalability or flexibility means to us. We are able to take the camera input and just leverage our video intelligence and our neural network to provide whatever outcome the customer may be looking for: Is it for pedestrian safety? Is it just incident detection?

Ultimately, the funding comes from our tax dollars, and we need to make sure that we make the best use of those dollars as spent by an agency. Ultimately, the traveling public is who is funding what is going out onto the road space.

Can you give any real-world examples of the ISS technology in action?

I would say the biggest-scale project we’ve done was in Mexico City. We implemented our operating system, SecurOS^®, within their entire transit agency. So we were the interconnected network for somewhere north of 65,000 cameras and other physical devices.

And we were able to then leverage that neural network to open up the possibilities. If you think about the amount of personnel they would have needed in order to review that number of cameras or to look at them live—allowing our system to be the point of the spear for them, while everything else lives behind it, that was transformational for Mexico City as a smart city.

So that’s the feather in our hat. That was a very large project for us, but it allows you to understand what the scale of the need is in some of these very large cities—in the world or here in the US—when reviewing just inbound video into their system.

But in some places we’re just doing flow estimation: speed, volume, gap, occupancy. This is the data that engineers hold dear and need constantly in order to make decisions about reshaping where we’re driving for the future.

We also have a few examples with tolling agencies, where we provide our LPR solution to do license plate capturing. If you’ve ever driven in a heavily populated area and been on a toll road, there are generally a number of cameras and devices up on the gantries that are already in use by that agency. So we’re able to be a part of the totality of what that governing body may be doing.

Does making sure personal data is protected factor into your solutions?

It absolutely does. From the privacy standpoint the industry has taken a branched approach: There’s what we are actually capturing out on the roadway where we can blur faces, blur license plates; but then we also have the intelligence within the cameras to help us remove any of that personal data.

ISS had its grounding in what physical security meant for real-world applications, and we have continued to build on that. We work with each end-user agency individually on what its standards are. And we also make sure that, for any advancement we bring to market, there is a parallel path concerned with security and privacy. Because trust and understanding from our users is paramount to our success, so it has to be a part of what we do.

Are you partnering with other technology companies, such as Intel, to make this happen?

From a performance standpoint, a company like Intel is almost invaluable for companies like ISS. More and more of our end-user agencies are asking us to include different data points and just to push the capabilities of the physical hardware and software.

Intel partners with us to react to and solve the problems we see in the real world. But the understanding and forethought from a company like that of what the marketplace is going to need—that has a huge impact on us generally, and ultimately the industry is relying on Intel and companies like it to continue to push that forward for us.

At ISS, we are trying to solve the real-time, actual problems and events that you are seeing out on the roadway. But we are also looking to solve problems that we don’t even understand yet. And we believe we are up for that challenge.

Related Content

To learn more about intelligent transportation systems, listen to Intelligent Traffic Management: Keeping Your AI on the Road and read Video Intelligence Illuminates Path to Pedestrian Safety. For the latest innovations from ISS, follow them on X at @isscctv and on LinkedIn.

This article was edited by Erin Noble, copy editor.