# Why We Need Ground-Based Digital Visual Observers

- Event: Xpress ’23
- Published: June 26, 2023
- Length: 14:44
- Watch: https://nestgen.org/library/watch/6UiASmBWRpE/

> The transcript of a recorded NestGen session, for reading or for use with an AI tool. It comes from the video's
> captions; automatic captions can mishear names and terms.

## Transcript

**[0:00]** [Music] [Applause] [Music] welcome back to NestGen Express we're here at boot 3909 at exponential 2023 here in wonderful Denver Colorado my hometown welcome you all we're here with 8 000 of our closest friends and we are ready to talk to uh you know what I'm actually really excited about this one we're going to talk to Dan O'Shea Iris automation director of global sales and customer success welcome to the show thanks for having me you are so welcome all right so we're gonna do this we have a couple questions for you we're gonna obviously talk about Cassie G clearly but we also want to talk about some other things maybe get a little deep get a little personal we never know let's do it let's do it all right so so first question so why because I use visual observers right most of my students are visual observers I'm like look sure why is using visual observers only a short-term fix and why would a tech-based solution be required yeah I mean ultimately it comes down to

**[1:02]** two things right safety and scalability right and so the safety part of it is eating hopefully apparent because the studies that have been put out there about individual observers I mean humans just aren't that good like okay so where we as humans are good at spotting that there's an aircraft out there but we're pretty bad at determining whether or not it actually poses a relative risk and our field of view is pretty narrow we get about a two degree field of view and then outside of that we're getting into peripheral vision that has that's reliant on movement so but the human visual Observer not looking at the right place at the right time it becomes difficult to identify that threat and then it becomes really hard to determine whether that threat is really a threat how far away is it how high is it and whether or and will that pose an actual risk to the operation so a game an AI based system like cassia Jeep or any of the other detect and avoid systems that are exhibiting here at the show are going to be able to give more of a Precision um give you more Precision about where that threat is and how much of the threat it

**[2:06]** represents to your actual operation so safety is Paramount with any drone operation but especially BV loss operations and so being able to identify that threat early being able to know exactly where it is and in the case of cassia G what it is we'll talk about that later I think but um is absolutely critical the other part of this is just scalability um the human factor right so we talk with lots and lots of police departments right um there is a massive Crunch and a massive shortage of officers here in the United States and so if you have you know police departments looking to scale BV loss operations and they need to have officers on rooftops well those are officers that now can't do their regular jobs and similarly if you're doing like a 50-mile uh mile pipeline inspection do you want to have humans out there every step of the way at a certain point it becomes do you have the human resources to actually accomplish that and do you have the cot like the budget to be able to pay those right so in those

**[3:06]** two instances that's why we need these digital Solutions and like Cassie G to be able to you know present that safer argument that's safer operation and to be able to provide scalability and I've been a visual Observer I can tell you it's fairly boring yeah and knowing now that like everything right here is like amazing everything out here I'm a T-Rex yeah there you go that's good to know that's good now that's good it's good okay so so what is what if we were to give cassia g a super power what would it be I I gotta know can it detect UFOs okay that's whoa that's what it could whoa bye okay so all right a little background on Iris automation right like so when we were founded um we started with onboard detecting the voices right and so the way that our system fundamentally works is we do classification Ventures yes so we'll spot a moving object and we'll drop it into a box bird playing drone helicopter right so in the

**[4:08]** initial days virus we had another classifier did it UFO oh for the stuff that was we just didn't know what the heck it was so um so yes we'll detect UFOs we'll probably classify it as like a burner helicopter or something but yeah we'll do it um but no is it because you're ran out of fingers maybe okay so um but that that honestly is the superpower of a system like cassia um any of the iterations but cassia G2 um and what it comes down to is that you know we'll look at a moving object and we'll be able to positively identify what it is in addition to where it is so you know when we talk about like Aerospace taking foot some people care a lot about birds so if you're at a vertaport for instance you don't want to take off into a flutter flock of birds but wait and so you have a class flyer to identify that if you don't care about birds at all turn off that classifier so now we're able to provide some functionality and be able to identify these threats and provide that information in real time to the user so that this remote

**[5:13]** piling command potentially across the country can say oh that aircraft that's actually 747 flying at 7000 feet I don't care about it right or that's a sparrow that just overflow my my drone Nest okay um well it's gone now I don't care about it or oh yeah that's uh Bell 212 that's like circling my operational area I care a lot about that and that's really what the differentiator of our system is that we're able to make those positive identification that makes sense yeah yeah that makes a lot of sense actually because every time every every single time I remember just being on radio and it's like you're hearing oh blah blah no Factor blah blah no factor and like a I'm questioning the no factor a little bit but I'm also that's not helpful it's not helpful information I don't need that in my ear I don't I don't really care I appreciate you doing your job but I would rather have to software just do that for me right and that's the point I know okay so tell me a story can you tell me a story where cassia G helped prevent actually a potentially

**[6:15]** dangerous situation so I got stories about the onboard system the ground-based system is relatively new but I got other stories about the ground basis so let's talk about that one so let's so so we recently did a large a really really impactful approval with the city of Pearland Police Department and so why this was so important was because uh cassia is named as part of this waiver as being an alternative means it applies to the 91-113 B see an avoid requirement okay that's a mouthful so what does that mean it means that we are an acceptable substitution to individual observers um and we get named at that which is awesome and so when we were doing the initial testing um in Pearland Texas we set up a system on top of a water tower um permanently installed it there and uh the police officer that we were working with Brenda Khan um he was like he called up all the local PDS around him Houston called up their helicopter company called the local airports and it's like guys I want you to beat this come in at 100 feet coming at 200 feet 300 feet 400 feet I want you to try to

**[7:16]** beat this system and it couldn't do it which was awesome and uh so like it does what it needs to do for sure um with the armboard system actually we we have a another partner here that that has our system integrating these are drone and they were telling a story like how they were given a demo for a customer It's Just Aircraft took off and then immediately descended they were like Western oh tried to take off again immediately to send it and what had happened was and then all of a sudden try to take off again immediately descended helicopter bust right over there so literally uh our system was picking up the helicopter initiated the avoidance maneuver no no no yeah exactly right exactly and they had no idea because it was flying so well wasn't broadcasting on adsb coming in over the Treetops yeah that is it's pretty neat that is pretty cool that's all of superpower yeah yeah a little Annoying if you're like come on man come on man you're like I'm so glad you didn't find that helicopter helicopter pilot and buying the gear right yeah because that was a demo for a customer yeah yeah amazing well that's really cool perfect that's it we did not plan that no no no why how would you plan that no you could okay so so

**[8:20]** this one may be a little bit obvious but like so how does the gaseous G system help pilots and ground-based operators comply with current regulations you had that big mouthful of uh well exactly so I mean the one thing that we're trying to do every step of the process is we want buy-in from the regulator and especially you know we're a North American company so we're you know working mostly with the FAA but also Transport Canada um discussions with ayasa and the whole point is that we want them to be comfortable with what we're trying to do and what our customers trying to do right and so you know we we've built out our system to asdm Performance standards we've made sure that we've done um extensive real world testing over 17 000 worldwide encounters to prove out our data sets millions of Encounters in simulation and then we did third party validation testing with Virginia Tech approval our test methodology to be able to say yes this is real and so we're doing all of that and of course lending lots of regulatory support to get our customers on the finish line but the real big thing

**[9:23]** that we're trying to do with our customers is get them engaged with the FAA early and the whole idea is that if you want to make sure yes you want to make sure that you're flying within the Google set as as much as possible but as soon as you start getting ambitious and start going for like I want to fly 50 miles Beyond visual in the site you want the FAA on your side and more importantly the FAA wants to be on your side early they want to understand that from their position yes they are charged with protecting the gnats right they want this airspace to be the safest in the world and every regulator feels the same way about their own individual airspace but the FDA also realizes they have the Monumental task of creating new regulation and they need to understand what's out there and so on the one hand building that relationship early with the FAA allows them to understand the safety and efficacy of the operation so that they know when they give their stamp of approval that they're doing so with a safe operation with Safe Systems that have been validated and tested

**[10:23]** are going to ensure that there's no any no damage or loss of life to um you know people or systems right um but they also want to be able to be aware of these systems so that when they're grafting those rule sets in the future they have you know examples of that they get that they can point to to say this was a safe operation and they want to be able to get the data from those operations to back up so they want quantitative and qualitative information to back up you know to get your rule make it let's talk about impacts and now about sure so so what kind of impact do you see a crown based digital visual Observer having on the overall aviation industry in the coming years and how do you see the technology evolving whoa okay this is this is good I see it's huge right like so the FAA in particular right is making strides towards you know shielded airspace and trying to make Beauty

**[11:25]** loss operations easier while still staying safe so a system like ours a daa system is able to the whole idea is to get Beyond shielding right like so if you're stuck within that 50 foot you know for some and for some folks like doing Inspection Services 50 feet that's probably more than enough um but for folks that need to get higher up at least arguments um certain types of Inspection Services mapping and whatnot then you need to be flying higher and then we get back into the question scalability right like how can you justify putting like 10 Vos along a 15 mile pipeline you just can't right like the cost just doesn't work out that way and so and then we get back to the safety argument as well so ultimately as these you know as the industry develops as drones continue to proliferate then these then systems like ours are are going to make sure that you know we maintain those scalabilities in terms of the evolution of the of these systems over time I think what we're going to see to start more like sensor fusions and incorporations of multiple different

**[12:26]** types of Technology um into local localized well in our case nodes right like so you know we're closely watching the development of a roller ID and when that rolls out in September opt to put a remote ID sensor inside our node right like to give because we already have Vision system we have adsb and then remote ID that just makes it even more comprehensive right um and then of course working with folks like FlytBase in flight 11 through the FlytNow system now all of a sudden we could talk about multiple sensor integration so now a system like ours an electro Optical system has its you know UPS there there are absolute upsides in terms of like classification Precision um and low level detections right but there are downsides to electrical Optical systems right relative range and of course you know weather conditions of one none so now if you're able to start fusing things like a radar into that as well into a singular system now you've created a safer operation with more flexibility right like so this is kind of where I see things going I see like a continued industry push for collaboration that's going to result in multi-layered detections

**[13:30]** um and of course you know just from us as a company perspective we're constantly doing updates to our system we do at least three software updates a year so then and we're always tweaking performance building new classifiers rolling out new features and functionality and of course every other company at this show today is going to be saying the same thing in some way shape or formed everybody's trying to a group and and we're all on the same page in terms of collaborating to get the regulator comfortable and so I think we're going to continue to see those sensor fusions those joint collaborations to get the Drone industry really up and moving we're starting seeing scaled operations globally so there's my mic drop um Sensor Fusion yeah Man Dan this has been amazing thank you so much thank you for your insights yo having a little fun with us here uh you know it's it's day two towards the end of day two exponential we're just having a good time here on Nest fan Express this is just kind of the place to be we're just filing the drain but highly no no we're not spiraling no look I may be done but but besides

**[14:37]** that's all you that's just for you uh but thanks so much for joining us we will catch you on the next episode NestGen Express thank you