How drones and automated rovers monitor critical infrastructure | Tech News
Dan Patterson: Robots and AI don’t rule the world quite yet, but they’re coming close. AI is powering drones that help save companies time, energy, and money, and can help protect critical infrastructure. Danny, thanks a lot for your time today. I wonder if you could help us understand how rovers and drones help construction sites, farms, and can send back real-time day about what they’re monitoring.
Daniel Rodriguez: Yeah, definitely. So, essentially, when you look at it in two different worlds: Construction and Agriculture, construction has a big problem where you have to monitor the daily progress of a building as it’s being built up, either from the outside or from the inside. The problem with that is, it’s very hard to do it autonomously until now. Now, these rovers go inside the interior of the building, they can capture it with 360 cameras, the interior, every 4 X 4 X 4, and that way, all the engineers, the architects, the designers; they get to be able to see what’s going on back at home office every single day. They have a full log of that.
Now, when you combine that with VOS.AI, and you’re actually doing the machine learning and the crunching. Guess what? We can find variances, differences, issues in the construction; is there rust? Is there humidity? Are there leaks? Is something happening? That’s all very important. Now, if you think it’s just a house, not a big deal. When you think about high-rise, big deal. Multiple floors, office buildings.
When it comes to agriculture, when you think about illnesses, thousands upon thousands of acres are infected with different types of illnesses every single year. How do farmers keep up with it? The truth is, they don’t. Using aerial drones, you capture the imagery, you upload it into VOS.AI, VOS.AI crunches it for you and spits out, “We found illnesses here, here, and here.” That quickly allows the farmer to basically isolate that tree or that plant, and solve the problem immediately.
SEE: Drone policy (Tech Pro Research)
Patterson: That’s fascinating. So, Danny, I grew up in Iowa, a state that is driven by agriculture and farming. I know that it’s a very low-margin business, so how much savings could automation provide to farmers and farm owners?
Rodriguez: Well, it’s hard to always measure that. I also come from a farming community. My father, my family have been in farming for years. One of the biggest problems is: How do you make it cost-effective? It’s always a challenge, especially here in the United States. And so today, without drones, they actually use a helicopter with a pilot and they go once a quarter, once a month out there. That costs upwards of $200,000 just for one crop. How do you possibly do that with more? We proved that just by using an off-the-shelf, a $1,000 quad-copter and using machine learning. They could do it every week, and do it at a tenth of the price. That’s essentially how it helps them drive that down.
Patterson: That’s fascinating, and of course, with a helicopter, you have human surveillance, and feedback. Machine learning can do so much more. Can you help us understand some of the differences in the ways machine learning analyzes data, versus the way a human would in a helicopter?
Rodriguez: It’s actually pretty similar in some regard. I mean, a lot of these algorithms are designed after the human mind, and how it works, but essentially, just taking in data. You have to train it on thousands upon thousands of images to understand what to look for. Once you’ve classified the data, and you’ve trained on it, then it just becomes like a filter. Now, when you start feeding in terabytes and gigabytes of data, you’re just tagging every time you see the same thing over and over again. Essentially, it’s pattern recognition. So that’s how it does it.
Patterson: And I see that some of your tech is trend of the year, maybe trend of our lifetime, we’ll see. It’s built on the blockchain. Tell me how you use AI and the blockchain together.
Rodriguez: Yeah, definitely. I’ve been an early adopter of blockchain, back since 2011. I was always fascinated by the space, and when we started digging in, and looking to see how this could be leveraged for what I could do, specifically around machine learning and drones and whatnot, I immediately found this one thing that just kept standing out.
There are 500 million miners in the world today with these huge mining rigs, and I said, “Wow. That is a thousand times greater than Amazon, Google, Facebook, everybody combined. Their whole infrastructure. How could I leverage that and tap into a little bit of that?” That’s what started that, and seeded it. From that, we said, “Okay, let’s leverage the world of miners, and then let’s go ahead and make a decentralized ledger that stores the learning of the actual machine on the chain itself.” So, it differs from other blockchains in that regard.
SEE: Getting started with drone photography (TechRepublic)
Now, what that allows us to do, is that now we get to expose all this learning to everybody; at the same time, it’s private and secure.
Patterson: Privacy with security is always something that gets me about the blockchain. It might be the killer app. Danny, thanks a lot for your time today. I wonder if you could leave us with a forecast, a prediction. Look at, say, the next six, 12, 18 months in both blockchain and machine-earning technology, particularly, when it comes to drones. Where are we in the short-term inside that window?
Rodriguez: When it comes to drones, that market is still growing, and they’re finding more and more use-cases every single day. I would expect that to gradually grow. Now, the blockchain space is very hyperactive. There’s always something new going on, and there’s a lot of innovation. I can imagine in the next year, there’s going to be a very different landscape. You’ll see a lot more things going on, a lot more hyperactivity.
When our platform gets launched, you’ll see a lot of that go on as well. Now, in terms of AI, I think it’s basically a race right now, even though some people don’t think of that. That race is: Who is going to make the biggest, baddest AI that solves all the problems? And we should be able to see those things happening in the next three years.