 Think Tech Away, civil engagement lives here. It is a noon hour on Thursday folks. Ted Rawson here, downtown Honolulu, our Think Tech Studios where we have our show where the drone leaves for every week or two weeks. We bring to our trusted audience information and views, facts and issues associated with the emerging world of unmanned air system, drones, remotely piloted aircraft, whatever you want to call them. One thing about this show that's really incredible is the people you get to meet that are already involved in the business and the way we can work together on bringing ideas together that work on the future of education, the future of problem solving and such like that associated with drones. Today we have two newcomers on the show, Dr. Zach Trimble, right down the street from UH Manaw, mechanical engineering department. Thanks for coming on Zach. Thank you Ted. I'm definitely nervous are you? Yeah of course I am nervous. You said first, first time right? First time, right. Thanks for having me on. So here across the country here, across the Pacific and across the country in St. Evine, Salem, Massachusetts is Bob Gellar. Bob you there? Yes, hello. You are? Thank you Ted. Okay and Bob is coming to us as an aerospace expert and a small unmanned air systems pilot, operator, designer and such and as a practitioner of the trade and what's most interesting I think is that both of our guests look at the problem of UAS from the inside out as perhaps more than from the outside in. And that's what I wanted to bring to our discussion. I think so much of the time we see our people we work with thinking about it from the outside and you open the box, take the thing out and you get amazed and amused by what you've got. But what we don't understand is what's inside, what makes this thing work. And so from a perspective of a practitioner such as Bob and from an educator such as Zach, we've got to somehow break through that and get the kids interested and get even the users interested in operating from a perspective of what's inside and how can we make that inside better. So Bob let me start with you. Give us your kind of background on where you come from and how you enter the UAS mix and how we might take this issue of understanding what the computational software and mathematics are that make these things work and how do we stand that tall in our educational outreach. And I realize that's a lot. So I'm just tossing it your way, sir. Okay. Thank you. I'm really glad to be here today. So just a little bit about my background, you know, loved aviation for a very young age. Gone to air shows. I always wanted to be a pilot. Got my private pilot license about ten years ago, fixed wing, single engine land and then I was also getting into the remote control aircraft as a hobby. And in conjunction with that, work in the aerospace industry and get in business for about 15 years, working with regulators and understanding kind of the safety practices that go along with manned aviation and kind of coming along in parallel and my free time was the unmanned stuff. So starting with fixed wing and then helicopters, multi rotors when those first came out and very early stages of that when they are still taking apart Nintendo remote, they get the accelerometers out that make the flight controllers. They've come a long way in the last five or six years and just learning more about, you know, the programming that goes into that, the TID tuning, we'll talk more about that later. But to touch on your other question about how to bring that to the forefront of the schools, I think there's a huge STEM opportunity with UAVs because it touches on everything. It touches on mechanical engineering. If you want to design the structures and the mounts that hold the components on the aircraft, it touches on electrical engineering. You have to understand circuits and control laws and soldering and dirt flow, voltages, watts, all those things. And it touches on your science. You have to know how to code and how to write software and different parameters through command lines and other functions. And it also touches on physics. You know, as you tune these things through the control laws and as you fly them and understanding the relationships and the trade-offs of, you know, torque versus top speed versus power, different propellers, different motors, the effect of center of gravity, you know, all those things can be learned by anyone, really, and especially students. And so getting those things and again, 3D printers as well have a huge role in UAVs because a lot of the stuff just doesn't exist. You can't go buy parts for these things. You have to make it. Thinking creatively on how to design a 3D model and then make that into a product that you can put on your aircraft. So getting those things into the schools, I think the cost has come down quite a bit. It used to be a pretty high barrier to entry from a cost standpoint. Now you can get, you know, entry-level setup for, you know, $100 and cheaper if you 3D print your own. So I think getting those into the hands of young students and letting them kind of learn and grow with that is the way to get them thinking about those parts as opposed to buying an off-the-shelf product that has all those things already done for you and all you have to do is charge the battery and take off. So starting at the point of level and understanding those trade-offs, I think it's important. So we have the inside outlook that you just expressed and then we know we have the outside outlook that all of our peers and folks who work with drones normally come from. The concept of a drone is really that of motion management and position measurement and somehow putting those two together in some kind of a software orientation that achieves the right positioning in some efficient way. So Zach, how do we take what Bob just said and think about packaging that certainly at a graduate level, undergraduate level, and even down to the elementary school level? How do we take an IMU and hold it in your hand and have somebody understand what that's all about? Not feeling any pressure? Well, certainly we need the outside in look. We need people to be able to look at drones and imagine what I can do with them. Imagine what new data I can get. Imagine what new mission I can do. We need people to do that to drive a desire and end use. But as you just mentioned, it's an issue of do I know where I am? Do I know where I'm going? Am I under control? I'm doing the drone. Yeah, I'm being the drone. Am I going where I am supposed to be going? And we need people to see that from the inside. And as you said, do I understand that IMU in my hand? Do I understand its limitations? Do I understand the GPS signal? And do I understand all of that? Yes, it's giving me a position in a location, a pose estimate. But there's an error in that. And we need people to imagine those new uses of drones, as I said, from the outside in. But understand that I need to advance the insides to get to those new ones and get excited about advancing those insides. And I think we're being relatively successful that at the University of Hawaii. We've got a fairly large contingent of students running in what's known as a VIP program. And I know that you're familiar with that. But for the audience, it's known as vertically integrated projects. And what happens is we get together as was mentioned, we get together mechanical engineers, electrical engineers, all working together at various ranges, from freshman up through doctoral students, on these insides to say, OK, how do I get it, the drone, it being the drone, to have better location to be able to do the next mission better? So we can imagine the next. So we're actually putting a screen on there and saying, hey, can we see the stuff that falls through the screen and fix it? Is there better ways to do motion management? Better ways to do the math of projecting the right trajectory and that kind of thing? Better ways to look at fault propagation? If we can find a way to make that what is the takeaway out of these projects and out of the education system, that is the place where the future lies right there, getting all that. I mean, as an educator in mechanical engineering, I can't agree with you more. That's why I say, this program, we get to get people who think I like together a lot. They don't generally pick people on a show who don't agree. Bob, what do you think about that from your experience in industry and your perception of what operates successfully and what doesn't operate successfully from an operator perspective? And I'm really thinking here of fault accommodation, symptom identification, and what to do about it. As we all know, we're all three of us airplane pilots. And you have 30 seconds from the time, the first indication of something to fix it. As long as you have a lot of altitude and a lot of ideas, you're probably OK. In drones, you have five seconds. And you may not have the proper knowledge of what the fault symptom is and what the indication is and what the consequence is. Therefore, you don't know what to do about it. So Bob, from your operating experience, how do we tie that all together? Yeah, so a couple of different ways. One way is to reactively to through the telemetry data that you're being sent from the aircraft while it's in the airport, and you're going to get alerts. So depending on the flight control system that you have in the ground control system, so the ground control system being the laptop, if you're using a laptop to control the aircraft, some systems are quite sophisticated. You might just only have a handheld remote, but maybe you're getting audible or visual cues from the aircraft through flashing lights or buzzing cones. But however it comes to you, you're getting some feedback that there's an issue, whether that be you've lost GPS, the battery's getting too low, other things like a multi-rotor, if you start losing travelers or motors, things like that, usually that departs controlled flight, and there's not a lot you can do. But things like batteries getting too low or GPS failure reacting to those things at the pilot. So if you have, are you getting too low, you land it. A lot of times there's things you can do ahead of time, and your configurations that set fail-safe. So when those things happen, they're dealt with automatically. So when the battery gets too low, it turns around and comes back and lands. If you choose, you can switch into manual mode and do that yourself. Whether the industry's really heading is trying to get more proactive. And so by that, I mean looking at flight log. So in performing statistical analysis, I've been trying to forward-look and forward-project on anomalies. So as you see, vibrations start to increase. Maybe you've got a motor that's about to go bad or a bearing that needs maintenance. Or as you see, current draw starts to increase. Maybe you have some electrical issue or other issues that are presenting themselves before the failure, so that you can do that maintenance proactively before you have the problem. Because as you mentioned, you don't have a lot of time when the failure occurs, especially the more catastrophic one. So trying to get ahead of that is really the way that the industry is trying to trend and drive what's through software and data analytics. That's interesting to know. And that also brings up a question. If some element of ministry or some element of academia comes up with a bright idea, you mentioned forecasting and potential failure, if there was a brilliant solution that showed up somewhere, how would we, as the larger industry, capture that and find a way to employ it widely, as opposed to have it controlled by some proprietary means by which you have to buy it? Because this whole industry is based on sharing a lot of information, a lot of capability. Do we have any kind of a commercial advisory group that takes out those kind of bright ideas, Bob, and does something to standardize them? They're in the process of being informed. So there's lots of different groups that are trying to shape some of that through either legislation or just through guidelines and guidance. There's a couple of those. There's also quite a bit of open source in the industry. PIXHawk is one example through Archipilot, so there's conduits to kind of get things out there that in a way that'll be broadly shared in a kind of commercial fashion. So that would be one conduit. And then I guess kind of integrating it into software platforms that are existing. So there's some web-based tools that are used for logbook analysis and data log analysis that are trying to do some of these things, and that would be the place to kind of inject and infuse some of the students' work potentially in a broader sharing of that knowledge. So it sounds like, Zach, we need some students involved here on the websites all the time, hitting different websites for different information and pulling this insight together. It's all over the place. Yeah, I would agree with that. And I think it is the student-built websites that actually are an interesting repository for this type of information. There are many universities, probably all of them, that have some form of a UAV program going on, and nearly all of those UAV teams maintain some sort of a website. And so the hard part is, is how do I check them all? So much information. Yeah, it's an information overload to an extent. But yeah, the student teams themselves building up their websites and sort of monitoring what's out there. Now, there's a challenge, right? If I come up with something very novel, do I really want to just go open source on it? Bob, close your ears. Right. And that is a challenge. And we do have students that struggle with that. This idea of, do I really want to do this while I'm at school, or I have this great idea, and should I wait another year and then try to... When I'm out of school, I can make some money on this thing. That's correct. Instead of working and having to do it as part of my schoolwork. Yeah. We do struggle with that a bit, not a ton, to some. Well, most interesting. Well, let's get back to where this all fits into education when we get back from our first and only one minute break today. My name is Mark Shklav. I am the host of Think Tech Hawaii's Law Across the Sea. Law Across the Sea is on Think Tech Hawaii every other Monday at 11 a.m. Please join me where my guests talk about law topics and ideas and music in Hawaii Ana all across the sea from Hawaii and back again. Aloha. Hi, I'm Bill Sharp, host of Asian Review here on Think Tech Hawaii. Join me every Monday afternoon from 5 to 5.30 Hawaii Standard Time for an insightful discussion of Contemporary Asian Affairs. There's so much to discuss and the guests that we have are very, very well informed. Just think we have the upcoming negotiation between President Trump and Kim Jong-un. The possibility of Xi Jinping, the leader of China remaining in power forever. We'll see you then. It is still the noon hour on Thursday, folks. Ted Rawson here and Zach Tribble, downtown Honolulu. Zach, thanks for coming on the show. First time from UH Manoa and Far Across the Sea over in Salem, Massachusetts. We have Bob Gettler waiting to come back on. There you are, Bob. So we were talking before the break about education and such and how critical that is to all of this and there's so many elements to that. We could never exhaust all the time we have and still be talking about education. But about problem solving. We were talking about the information, there's so much information everywhere in terms of different methods that have been put together for the technical elements of drones and such. Bob, you from an industry perspective, how do you make sure when you generate a solution to a problem, you've taken advantage of whatever's out there? How do you package up a solution when somebody comes to you with an operational or a design problem? If you just take us through that, we really need to import that into our educational process. One of the things that I try to do is stay up to speed on the different trade journals and news feeds, websites, follow a lot of different companies on LinkedIn just to kind of keep my fingers on the pulse of the new products that are coming out. Also attend trade shows at Xfinancial this past spring. Interest to see, talk to the different vendors and look at their product. To try to have a catalog and a roll deck in my mind of what's out there. Also it's a very good relationship with the different vendors and suppliers that they kind of give me ideas on things they're working on. We collaborate on a few things if there's something that's not currently in their catalog but maybe it's something that they could develop to fill a niche. And as much as possible is because of the speed that everything moves in this industry. We try to use integrating off-the-shelf product just because the product lifecycle and the development cycles to develop something brand new and do an MPI are often too long. And so knowing what's out there and what you can use in a somewhat creative way to meet the needs of your application is very important. The combination of that and then also just kind of collaborating with my peers in different industries to understand what they know. So that's kind of the first step. If we can't find anything that currently exists then we start thinking about what can we quickly wrap the prototype or make if necessary. Okay, that's interesting. So you basically think of object-based integration. You have a bunch of objects that you can integrate if somebody provided the API or some other integration ROS or something format for you to follow. Let me throw one at you just to hear how you think about it. Say you have a need to get a drone or a UAS that operates outdoors in clear air and clear signals, has to operate also in an urban canyon with a lot of potential signal reflection or signal blocking and has to operate undercover somehow that is under a roof or in a tunnel. So you have these transitions from the quality of the GPS for example is going to be quite varied across that range. How would you think through putting together a solution to that particular problem? So for that particular problem, and this is a new thought for me, this particular application I have had to deal with this, but what I would start with is for the canyon side is probably some kind of mesh network. Either that or look at what kind of data do we need to send back? What bandwidth do we need? And are we okay with just having the command and control or C2 which doesn't take a lot of data? Is that's going to help the system network that we need to build or take advantage of? So if it's low data rates, maybe you can get away with SACCOM if it's or lower frequencies which tend to have better object penetration in longer ranges. If you need high data rates, you got to go to the higher frequencies that you're going to need to probably have some additional radius placed strategically. And you'd want to do some analysis looking at spraying and doing some measurements and some testing to see where the optimal place is at, if you're essentially building a cellular network in a way or a Wi-Fi network with these repeater stations along your route. Once you get to the tower or the tunnel where you're going to be GPS denied, you almost need to have a second navigation system on the aircraft to switch to when you get to that environment because you're not going to have clear view of enough satellites that have accurate positioning. So that's where you're probably going to need to move to either like an obstacle flow, using camera systems or some kind of localized-based positioning system with some nodes in the tunnel and making that transition from it almost had two separate flight paths where the aircraft would fly to route through the canyon on GPS. And then once you get to the tunnel, it would switch over to a new flight plan with the GPS denied navigation system. So what Bob's outlined by going through that thinking process is something really, I think, instructive and useful to us. It makes me think that in the buildup of instruction and educational themes here, we have what we just thought of maybe as a positioning aspect. That is the position feedback, the GPS. You have to know your position regardless of what the mission is. And so the different environmental conditions determine what systems you're going to have, what electronics and what software to do that positioning work. We have position. So position management or position measurements, one thing, position management through motion understanding, motion modeling and motion compensation is another category. The transitioning from one of these modes to the other, the failure modes at a company and what the indication is and the response and what kind of training we go through in simulations. You can think of four or five major categories into which the technology is starting to collect here, buckets of technology. Is that how you guys look at it from an educational construction perspective? Yeah, I think so. I'm known for having a couple of different sayings that I repeat many, many times to students. One of them, the first one being, define what success looks like. That's great. Just hold that for thought. Define what success looks like. How do students react to that? Can they define success very well? It's a big challenge at the first. It is very much a challenge at the first. Once we've had them for a year, they're getting much better at that. But being able to just, in very simple terms, state to somebody, if I have to prove to you that my system works, what are the steps I have to take? What does success look like? What does this machine, this drone, this item look like when I'm done? And more importantly, I somehow have to prove to you that it solves your problem. And so I have to think about what are the tests I have to go through to do that? And then it very much defines what are the things I need to do. So even for the subject that Bob just described, that is the position identification, you could even define a success parameter for that. And a response to challenges or threats, electrical storms, for example, or heavy rain or whatever it might be that degrades signals. So you could think of a measurement of success and you could think of the challenges under which the thing has to survive. And that would structure the thought, wouldn't it, in terms of what the elements of the technical solution are? That's correct, yeah. That's exactly right. And by knowing what it has to look like at the end, it's much easier for me to find what I have to work on now. And then the other keyword is, what's your most critical module? And I think that's what you've just identified from what Bob was talking about is the most critical module. As a previous engineer, your mind just did it. And that most critical module was, okay, you gave me a mission and flying was pretty easy and communicating was typical, but localizing was hardest. And so I've defined the thing that I gotta focus on and I've gotta get that taken care of first because unless that doesn't work, I might as well not do the rest of it. And so that's the next thing that we talk about is once you've defined what success looks like, dig out of that the thing that everything hinges on, the most critical module. The reason I thought of that example is because we're getting into UAS integration with manned airspace. We have the FAA IPP program, which we're actually part of, to Alaska. That's dealing with 400 feet and below. And we have the NASA SOI or SIO program, which is dealing with 500 feet and above. So it's two different programs looking at integrating with manned aircraft. That means detect and avoid, it means remain well clear, it has all the proper terms, but that's gonna require precision information about location that works bar none. It has to work at all times or you haven't got to say system. And we think in terms of drones as positioning within two or three feet, if you wanna get up near a building or wanna get near a tree to look at what the bird's nest egg looks like or something like that. And these are really precise positioning issues in complex environments. And yet in our airplane experience, we have nothing like that. I mean, 100 feet goes by in the altimeter, you don't even notice it. And so we think of 100 feet is a positioning precision. Well, we're talking about a couple of feet in the world of drones. So this issue of integration is gonna define or drive a lot of these needs for thinking about these parametric issues, measures of success. And then the testing required to get to that point and the proof that you're done. Well, you've also hit on another thing that's important here. And that is the idea that oftentimes I can't get it done with one technology and that's really where the systems thinkings come in again. I can't get it all done with GPS. I have to coming in to be next to this building. I can't have enough GPS accuracy to be next to that building. So I still need GPS to get to where I need to be, but then suddenly I need to switch to a LiDAR or radar or something to give me that very fine precision. So again, now suddenly I have to start mixing systems together and we really need students that are interested in. That's interesting. And you talked about the school, the educational environment where you have a chance to speak about these things up front. You have a semester orientation so things aren't done till the end. Bob in the commercial world, you have to go through these discussions pretty quickly and get to a solution plan and a critical path that's gonna work for you in order to survive and make money. So do you have that same orientation that Zach has in terms of a wide open conversation about the various issues before you hit a actual execution path? Yeah, so I think what Zach touched on is an important part of the thought process of thinking about your kind of critical path and in your hinging point or your most difficult part, make sure you solve that. Because the last thing you wanna do is do all the easy stuff and then get to the hard part and not have a good solution. And either the other thing to think about too is if you're doing things that are gonna require a waiver so you get a back off of that all of the review cycle time for the FAA. So not only do you have to move fast in general but now a lot of your time is consumed with the regulatory agencies approving your solution. So part of that measure of success then that Zach talked about is having something that the regulators can understand and can find compliance with. So that's another whole dimension here. And we're about out of time. We didn't get to your favorite subject, Bob, of PID loop tuning and we'll have to hit that on a follow on conversation but I wanted to thank you very much for your insight from a commercial perspective on this emerging game of optimizing micro technology into successful drone operations. So Bob, once again, thanks for coming on from over in Salem, Massachusetts. And Zach, thanks for coming on down the hallway from Manoa to be part of this. Have a long drive. Have a long drive. Thank you, Ted. And we'll see you all next week.