Potholes are often a costly and frustrating problem for cities to manage. TechRepublic's Karen Roby talked with the CIO of Memphis about a new AI program that is making the issue much easier to solve.
The city of Memphis turned to technology to fix a pesky and costly problem that so many cities face—potholes! TechRepublic's Karen Roby talked with the CIO of Memphis, Mike Rodriquez, about how AI is making a difference and saving the city time and money. The following is an edited transcript of their conversation.
Karen Roby: So your office turned to AI and machine learning to address the problem with potholes. Tell us about this and how big of a problem were you facing?
Mike Rodriquez: Memphis is a little unique in that they've kind of underfunded the efforts to fill potholes or pave streets since like 2010. What that meant was, we had a cycle going that a road would be paved every 75 years, which is somewhat insane since the, I guess the practice is on average around 25 years. We found ourselves with a lot of potholes, and we've got about 324 square miles or about 6,800 lane miles, which is enough to drive all the way to California and back more than one time, so our problem is somewhat big.
We were trying to solve that problem, and I remember the day the mayor asked me, "Mike, is there an innovative way we can approach this issue? What can we do to try and solve this problem?" It wasn't a week later I was in a conference, and I was sitting at a table with a number of Google representatives, and I started asking them what they thought they might be able to do to help us challenge this issue and try to solve the problem. That's how this first thought came about with, let's see what we can do together to try and solve this problem.
Karen Roby: Tell us about what you did specifically, how you set up the program, and what the program actually looks like.
Mike Rodriquez: The first thing we did is, we tried to think differently about how to solve it. As you know, if you look at a pothole issue from the driver's perspective, you're really not even concerned about it until it's big enough for it to impact your drive. The way the city of Memphis does it is, most of the citizens who complain than give us attention to where those potholes are, and then we send a crew out to try and fix those potholes, and then they discover more potholes. That's how we were addressing the problem.
What we decided to do is inventory all the assets we had and try and approach it differently. One of the assets we had was a video that was recorded off the front of our MATA buses. MATA buses travel most of the routes that bring us concerns and issues with potholes. By taking that video, applying machine learning to it, we were able to teach the computer basically how to identify potholes, and then mark those potholes for us to start repairing those potholes.
It gave us a lot more insight on the conditions of our streets, but it also gave us the ability to get ahead of the demand. We didn't have to wait any longer for citizens to make complaints, we started getting visibility immediately on those potholes. And if you go a little bit further with machine learning, we're considering the opportunity of overlaying weather patterns on what makes potholes go from six inches to 12 inches.
SEE: How smart tech is transforming the transportation industry (TechRepublic Premium)
Memphis has all four seasons. It rains quite a bit during the winter. We have a lot of freezes and thaws and freezes and thaws, which takes a small hole and turns it into a large pothole. At the same time, in the winter it's a hard time to try and pave potholes because you'd have to have the right temperature and it cant be raining, which you see a lot at that time. By doing this, we can start looking at patterns and seeing what weather patterns come through and take a three-inch hole, turn it into a six-inch hole and turn it into a 12-inch hole, and we can start paving those three-inch holes before they start impacting the citizens.
Karen Roby: In terms of your reaction, what did you think when you started seeing the data when it started coming in? Were you surprised by it or just relieved? I'm sure the mayor was.
Mike Rodriquez: We were excited to see that this opened the door on how we address that problem. No longer are we asking our citizens to tell us when they're having a problem. We understand what's going to be a problem for them and we try and get ahead of that. As we've improved the video feeds off the front of the MATA buses, our abilities have actually grown in identifying other issues. So yes, we started with potholes, but we're also looking at a lot of other things in terms of street conditions so that we can better manage that process.
Karen Roby: So much of what we're talking about now with smart cities and with the video technology now and with sensors and IoT devices, all of these things, cities are getting smarter and able to detect things ahead of time. This is just one more step; it seems in terms of being a smarter city.
Mike Rodriquez: It is, and it's leveraging something I think we all have. We have video that's literally being recorded in community centers, and libraries, and schools and it's just there for when there's an issue. Using this type of technology, you're able to educate your insights, and you can make decisions based on things that are occurring at the moment. We like to say our situational awareness is growing. We're going to know when you have a problem, and we're going to get ahead of it.
If the video from the MATA bus happens to catch a field that's growing and not being attended to, we can train the computer to identify that, alert us and then we can send someone to go cut the grass before the citizen calls us and says this is out of control. It's changing the way we look at what we do and how we do it.
Karen Roby: How important are partnerships in IT when it comes to a city the size of Memphis?
Mike Rodriquez: The partnerships internally have greatly changed the way the other departments see the value of what we can do with technology. I don't know that it's the dynamic change between how they used to use technology, but if you think about it, in the old world they would gather all of the complaints, and then they would go about addressing those. Now we have the ability to alert them when they're in the area where there's an issue so they can start attacking that issue and it's making that partnership get tighter within the business.
Then if you look at externally, us leveraging Google and SpringML, who are the partners that we're working with, it allows us to be a lot more innovative on a government side than normally we would be because you have to build an infrastructure, first of all, to try and do some of these things. We're leveraging the Google cloud and the expertise the SpringML is bringing to the table, and we're moving quicker on innovative solutions for the city.
- Machine learning: A cheat sheet (TechRepublic)
- Telemedicine, AI, and deep learning are revolutionizing healthcare (TechRepublic download)
- Managing AI and ML in the enterprise 2019: Tech leaders expect more difficulty than previous IT projects (TechRepublic Premium)
- What is AI? Everything you need to know about Artificial Intelligence (ZDNet)
- 6 ways to delete yourself from the internet (CNET)
- Artificial Intelligence: More must-read coverage (TechRepublic on Flipboard)