CogX 2018 – Using AI to Turbo Charge African Economies and Reduce Environmental Damage | CogX
So I’m going to talk to you today about the industrial use of drones. One of the things that’s happened in the field that I’m in is that Amazon has created the impression that everyone’s mind that drones are pretty much only about delivery. In my experience of building these things, the bottom line is that’s one of the least productive uses of a drone. Partly because the batteries don’t last that long and they can’t go very far. But I’m going to walk you through other reasons why this machinery is all about gathering data so that you can manage problems much more effectively. The system that my company has created, we’re called H Robotics and this is the physical aspect of a system we call high-sight. And it’s designed only for companies. So this is not a retail product. It is designed for huge corporations that are trying to solve problems in an ongoing way. In our system, I don’t know if it’s coming up on the slides, in our system it comes in three parts. This is a stable platform in the sky. That sounds very simple. But interestingly, so many of the drones in the market today are just not able to handle the simplest of challenges like wind, like rain, like literally weather interferes with the flying process for most of them. And so we built something that’s an incredibly stable platform in the sky to which you can attach pretty much any kind of sensor, camera or device that the market comes up with. We put it inside this case. That is not because it’s just convenient to put all those cameras and sensors and the flying devices, the iPad that you run it on. It’s because this is how you transmit the data from the thing in the sky to where you want it. In other words, we’ve created a broadcasting system that works on site. It’s easier than not totally dependent on any external source. And the third part of our system is what we call the flight cloud. Flight cloud is a very interesting innovation because again, in the world of drones, because most people think of them as an individual thing that you fly one off and it’s not really about data gathering, you’re not too focused on what are you going to do with the data that’s collected. What we found is that it makes sense to make sure that every single flight is automatically tracked and logged and goes up into the cloud in what we call the flight cloud. So for us, this is a three-part system designed to help companies deal with the problems that they face. The current business model in this world is literally when you need aerial data, you call an external company and they usually send a couple of guys who bring a machine that personally I would describe as a toy and most of the drones have been developed from retail basis and they will come and fly manually around your site and it costs you about 5,000 pounds for a day and then you have to pay for the data because it doesn’t belong to you. It belongs to them. This business model I think is not going to work in a world where you can actually own the system yourself. And we’ve designed it very specifically so that anybody can fly this. If we were permitted to legally, I would let all of you have a go at flying it but they won’t let us do this here. But I promise you, it would take about 30 seconds for you to be able to fly it. That’s partly because we’ve built it around a system where all you have to do is point and click it where you want it to go and create pre-established flight paths. Now why is that important? That’s critical because if you have a bunch of pilots who are flying it manually, none of the data can be compared. So you can’t see slow moving problems over time. You can’t adjust the behaviors of your company to accommodate the problem that you’re trying to fix. So think of it this way, when you talk about drones in the industrial world, mainly people are talking about companies that supply human beings that manually fly these aircraft. Everybody says we’re going to have these things dropping packages in suburban neighborhoods, right? We’re all going to have delivery in the backyard. When we tested these things for thousands of hours, let me tell you, there’s no way they’re going to be dropping packages in the backyard. And that is because every dog and every child is going to leap higher into the air than you have ever seen at leap before in an effort to grab it and pull it down from the sky. And everybody’s going to end up in the emergency room. The idea that they’re going to be used for this kind of general delivery, I think, is really a notion that’s gotten a foothold, but it’s just incorrect. And I don’t think it’s where any of the civil aviation authorities really want this industry to go. So we just don’t see it. What we see as a world where most of the drones that you know about, that you hear about, were designed to be toys for retail use. And what I’m going to try and explain to you today is what it looks like when you start talking about an industrial perspective. Mining. And this was the area that we started with, working with mines in Africa. It’s harder to find a more challenging environment to operate in than an African mine. But we wanted to be sure that what we have built will work in the worst possible conditions. And handled by people who are not going to treat the equipment delicately at the end of the day on a Friday, really important because most drones, when you buy them, if they break, they break. They’re done. They can’t be repaired. They can’t be fixed. We decided to build something profoundly different where it’s actually modular. So the arms actually snap on and off. And all of the equipment snaps on and off. So if you manage to break it, which is pretty darn hard, then you just replace the broken part and you keep going. In fact, we decided to ensure them for an entire year for the clients because it’s just so difficult to break them. Most of the business models are assuming that the way you make the money is you make a thing and you hope and you pray the client breaks it as quickly as possible and they come by another one. We think that business model just isn’t going to work in a modern world economy. And what you want is like you have a hammer and like you have a screwdriver. Now you’re going to have a robotic tool that gives you your aerial data. Now everybody goes mines. What the heck are you going to do with a drone at a mine? Well the answer is mines have a whole bunch of problems that we’re now able to solve. And I think these are the kinds of problems that are going to help Africa diminish its exposure to environmental risk issues, diminish its exposure to theft and to violence on some of its sites. Anyway, let me walk you through mining. One of the biggest problems is every mine in the world has what you call a tailings dam. What that means is a pool of usually very toxic fluid. And that toxic fluid is used to blast through the rock and the stone when you hope it’s water but it’s not. So this huge pool and they’re supposed to know what is the volume in that pool of water, pool of toxic fluid. Well you would think the easy thing is you put a boat on it and then you drop a stone on our device and you find out. No. Number one, the toxic fluid eats through the base of the boat so now you have a sinking boat onto the toxic pool. And in some of the African sites that we’ve worked on they have this small detail, crocodiles on the toxic lake that literally started trashing the boats of the mines that we were working with. So they said is there any way you can help us to know what is the volume of fluid in the tailings dam. So what we did is fitted our drone with a sonar device that hangs underneath and we drag it over the surface. Now we have an exact volumetric analysis of the tailings dam. Further, we can tell you the shape of the basin underneath. This matters because all these basins are lined and if the liners come unstuck, now we have environmental damage on an epic scale. If you don’t know the volume of fluid in your tailings dam all the local communities are at risk of overflow. So suddenly we’re able to use this for a purpose that was not what this client originally envisaged. Originally they said well listen, people are always trying to steal from mines in the middle of the night and we need something to help deal with this problem. It said fine, we can fit these with very bright lights. You can use it as an anti-thaf device but fixing the tailings dam problem is a much more profound use. They had another problem as well which is really interesting. Volumetric analysis of another kind. How do you value a mine? How does a mine know what it’s worth? You would think it’s because they pick up all the pieces of gold or diamonds and they count it out. No. They send someone out with a measuring tape once a quarter and the guy with the measuring tape looks at the size of the pile of dirt they’ve pulled out of the mine and calculates because there’s always a proportional relationship with the gold or the diamonds or whatever it is they’re mining. So he said you know we can fly this right over the top and give you a very precise volumetric analysis. He said this is incredible. It totally changes our understanding of what we’re accomplishing here on this site. So in our system the way we built it was to try to give these big institutions the opportunity to solve a lot of different problems using one machine. And you do that by basically changing what are the sensors, what are the cameras, what are the devices that you attach to it. It’s modular in its structure. Again, this is not a retail product. It is all about helping business to business environment improve their margins, manage the environment better, have less issues with the local community as well. The one delivery mechanism we do see and we’ve encouraged our clients to do is if you own fleets of these tools, if you have local communities nearby then why not you? Use them when you’re not using them to do things like deliver medicines. That makes sense in that context where you’re already out in the middle of nowhere. But if they’re not, they’re still very limited on the whole delivery front. Another use, oil and gas. Currently in places like Africa, in Nigeria, but frankly even here in the United Kingdom in the North Sea. The way they manage platforms, these huge oil and gas offshore platforms is sometimes they actually have to hang human beings on ropes over the side to inspect what is happening to the surface of the side of the platform. In some of these rope configurations, a human being can last literally 10 minutes. If the guy is on there for 10 minutes and two seconds, he’s dead. Literally these rope configurations are incredibly dangerous. Not all of them, but there are two kinds they use and one of them are called groin ropes where the weight of the human being is literally resting and if you’re in it too long when they let you go, the pressure on your heart gives way and you’re dead. So we’re like, why would you put human beings in such a dangerous position? Why would you have people on the ground in Africa, frankly in the North Sea, in this situation? This is where we can use this kind of tool to get very close to assess down to a few centimeters of granularity and now immediately understand what is the situation that you’re dealing with without ever having to put a human life at risk. Pipelines, it’s another thing in Africa, sub-Saharan pipelines. Some of the biggest oil and gas companies in the world are losing 100 days a year to sabotage a theft on the pipelines. Trucks come up to the pipeline, they literally hack into the pipeline and use a garden hose basically to take out, siphon out as much as they can get and then they run away and which you end up with is huge environmental damage around these locations. If you have a system that lets you see what’s happening to do some detection and deterrence to identify when a truck is approaching a pipeline in the middle of nowhere, to be able to send something up that has voice projection as we see you, we know who you are, please stop packing into the pipeline. At least we could do something about diminishing the amount of environmental damage that happens in these environments. We came up with a system where we can put these in self-charging solar paneled boxes so that they can deliver without a human being have to be close by. I think this is a really important point by the way. Again, because the business models we’re all familiar with say you have to have a human being flying one of these things, you have a notion in your mind there has to be a human involved. We think that more and more, there are not going to be any people. Why would you put a human being on an offshore decommissioned oil or gas platform, an incredibly dangerous place to be? Recently there was some guy who went on to one of the platforms in the North Sea to check it out in the decommissioning process and just touched the handrail on the whole side of the complete platform literally sheared off. These are not places that we want human beings. The other thing is Africa is leapfrogging. It is skipping the stages of development that we’ve all been through in the industrialized world. Their telecoms networks are all mobile phone-based. We interface with those networks instead of having to rely on all the traditional telecoms. They’re leaping into artificial intelligence. They’re leaping into data management much faster than we have in the West. So when we work with the local people employed by these companies in places like Tanzania, what we found is the guys are thrilled that someone is giving them a robotic tool that they can deploy, not some external person who’s flown in from London or New York, not some external person who doesn’t work for the company, but the people at the front end who know the company’s problem is better than anybody else. Suddenly, they’re empowered with something that they can operate, that they can manage. That’s been quite exciting for us to work with the local employees. It’s very interesting how prejudice people are and say, well, they’ll never be able to operate it because it’s a complicated technical machinery. Our actual experience is they were fantastic in identifying new uses that never occurred to us. We’ll say, what are your drones used for? What they’re going to be used for? The whole point is it’s a platform. It’s up to the person using it to decide how are they going to deploy the hardware and software mix on that platform. These guys know much better than any of us what’s required to solve the problems that they face every day. Also when you’re talking about drones and data, the way you see what you’re looking at matters. Because most of the drones have been developed as toys, they’re not really focused on how do you see the data. In our system, we work very hard to make it incredibly easy for the user that they should always be within just a few clicks away from being able to see it in the format they want, which may be raw data, that may be orthomisex, that may be photogrammetry, and may be volumetric analysis. Every day the market’s coming up with some new system for delivering a new way to analyze the data. We’ve actually been using more and more virtual reality as an example so that you can viscerally walk through an asset that’s on the other side of the world and understand what has just happened in that place, compare what it was like before to what it’s going to be like going forward. In this world of technology, everything is moving so ridiculously fast. So one area people always ask us about is agribusiness, agriculture. Guess what? Satellites are going to do that. There are lots of technologies that compete with this, but bottom line is what we’ve built is basically a hammer that’s robotic. It’s literally a tool that gets used by companies, their own staff on the ground. It’s not meant to be glamorous. We’re in a world where people talk about drones, the way they talk about cars. I like cars and they can flate a go-cart and afford an a fiat and a Ferrari. Well in drones they’re not all the same. We make a kind of land rover. That’s our story and we’re working closely with Africa to do this to improve the quality of the environment they’re operating in. I think now we’re going to go to another room for questions. So thank you very much.