# From a Gunshot Alert to a Drone Overhead in Minutes Across Montevideo, Uruguay

- Event: NestGen ’26
- Published: October 2, 2026
- Length: 25:27
- Watch: https://nestgen.org/library/watch/75oh_2viWbU/

> 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.

## About this session

When a gunshot is detected in a dense city, the first few minutes decide how much responders know before they arrive. Getting eyes on the scene fast, and sharing that live view with dispatch and the officers on their way, is hard to do with ground units alone.

In this NestGen session, Uruguay's Ministry of the Interior and Martín Fablet y Cía. will share how gunshot detection alerts are being connected to drone response across Montevideo so a drone can be over the scene within minutes, and what public safety agencies should think through when linking automated alerts to drone operations.

## Transcript

**[0:01]** Okay. Well, the idea is to share a bit about the experience we are having in Montevideo, Uruguay, of using drones as a first response. In this case, what we are going to see are examples of alerts and shooting. Uh, but the first video we're going to see is... well, on the next slide, the first video we're going to see is an operation that has nothing to do with a gunshot alert, but it's an event that happened a week immediately after the equipment was installed. Nivelitas, if you can change the slide. Okay. What we're going to see here is a video of a person falling into the river, the Río de la Plata, on the shores of Montevideo. Through a telephone report from a person, a witness who saw this situation, called

**[1:06]** the CCU, the command center. The command center immediately dispatches a drone. An aerial view is taken and a joint response is made with the Naval Prefecture, who are in charge of beach safety and national coast security , and in parallel, the state emergency medical services of the government are called in. All because of this person who had fallen into the sea and a person had witnessed this event. Nivelitas, I ask you if you can post the video of what happened. And we'll keep you updated. Well, that's our promenade, the promenade of Montevideo, of the capital. What happens there, what is focused on there, is this, a person who had fallen, had tripped on the stones and had fallen into the sea. There we see the thermal

**[2:09]** imaging camera for the detection of this person. A man is trying to help, a man who was passing by, uh, is trying to help this person who had fallen into the sea. All of this is live, meaning the drone was already there seconds after the 911 call. And the coordination with the prefecture, who are the rescuers. Uh, they were going to the place already equipped with their wetsuits, because the water in the river is very cold in winter. And that, that can produce this, hypothermia. So, the rescue has to be very quick. And we see that everything is being broadcast live and in real time, all in an exact coordination [clears throat] between command, prefecture and ambulance. There they remove the person from the river, from the water.

**[3:10]** This, the prefectural agents. We can see that we also have police officers on the scene, that the police have already arrived at the scene, in addition to the prefecture. We see how they are helping this woman who had fallen, not noticeably, I don't know if it's clear, but she has an injured leg, because there are rocks all over that area. They also have to help the person who was the Good Samaritan who went to lend a hand to this poor woman. Everything coordinated and everything live, that is, everything transmitted live to the command center. This video we are watching is public, public from the Ministry of the Interior here in Uruguay, that is, it can be found on the pages. It hung up on us or it hung up on me. I

**[4:17]** 'm not getting a connection. Yes, perfect. Okay, perfect, perfect, perfect. Well, what we saw there is a video of an event and how the system was activated through a 911 phone call. What we want to explain here, or try to convey as a concept, is that an incident is not just a drone flight. What we're going to use as an example here is that we have a digital alert with a geolocation, and we need to transform that into a flight, a drone flight. On the next slide,

**[5:21]** okay, we have the event, its coordinates, and we need to put an aircraft into action; we need to put a drone into action. They cannot communicate on their own. So this is where the platform comes in, where Flight TV comes in, which is our orchestrator of this type of thing. That alert could be a shot spotter, a gunshot detection, or it could also be a call to, of course, 911. We transform a digital event into action. Yes, we move on to the next slide, and here we see the power of FlytBase, where it functions as an orchestrator and gives us autonomy and orchestration of procedures, allowing us to input certain alarms or certain digital events, and that will allow us to put equipment, drone equipment, into motion . In this case they are DJI Dock 3 with their

**[6:25]** 4T matrix. So, this is basically how it works from the trigger or from the alert to the response. We have the event, the alarm is received, the connection between the weapon and FlytBase is generated, which will give us a logical response that we will pre-program. We're going to send a drone to the location, we're going to go to that location, we're going to carry out a mission. All this so that the command operators, the police pilots, have eyes on the scene, all through the platform. This system can operate automatically or it can also be manually confirmed by an operator. It works perfectly either way . That's a basic, brief summary of how it would work. In the next slide we see how until a little while ago we were talking about one team and one alert,

**[7:27]** but what happens when the city scales up, as happens in Montevideo, and there are many, I don't know, for security reasons we are not going to say the number, but there are many teams. How does FlytBase help us to answer, to choose which team should go? It's not always the team that's closest that will necessarily respond. There are going to be many things we have to take into account: the type of incident, the distance, the viability of the equipment, the airspace in which they will fly, the weather, connectivity, and the battery capacity of each piece of equipment. Okay, so if we continue to the next slide we'll show... There it is. What this FlytBase allows us to do, and this is where it absolutely shows the potential of the system, is that with many devices, many drones and many alarm inputs

**[8:28]** we will have control of the fleet, the mission, the live video of all these devices, the telemetry that will allow us to have that availability of the devices. What is the purpose of all this? So that the operator or the pilot on the other side doesn't see the drones as separate pieces, but quite the opposite. What is available is a network of equipment, a network of drones, where everyone will work together. It's like a ballet, to put it that way. That's why I said that the drone that responds isn't always the one that's closest. It will depend on many things, specifically the platform is the one that will give us this information and allow us to work in that way. We continue on the next slide.

**[9:34]** So, what we're going to have here, and what we need, is information. What will this give us? It's not just a drone flight, it's information. What's happening there? A gunshot was detected , uh, what's going on there ? To enable zooming in on the area that interests us. So, what can't we see with thermal cameras, uh, what can we see with a normal camera? We use thermal imaging cameras. As we saw in the video, the thermal imaging cameras are activated. Uh, where exactly is it? Beyond having geolocation, with the live camera, we can see exactly where it is. And what is all this for? To give us some context on how to respond to this digital event that started this whole process. All to give you information. Information for the pilot, information for the police. Uh, what for ? In the next slide,

**[10:34]** we'll see. Here. So, we had that event, that shot, that ShotSpotter detection. The drone delivered that information to the police. The police will physically arrive at the location with much more information and will go with much greater security. In other words, it's also about safety for the people themselves, the police officers themselves . That's why what we always say is information. And all handled through FlytBase. That's why this isn't just about flying a drone, it's about many things. Find, locate this, control a fleet of equipment, have information, video, alarm income. On the next slide, and Y is basically the potential of FlytBase.

**[11:34]** One platform, many triggers. Gunshot detection, 911 calls, motion detection cameras, perimeter sensors, alarm detection, artificial intelligence, CAD operators, etc. Everything is entered into Flytibase, which will give us a logical answer based on the equipment and hardware we have. Which drone is going, which drone is enabled, which drone can be used. That's basically what FlytBase and its power are doing in this DFR project. It is one of the three important pillars. In the next slide, what I can do, or rather summarize, is the lesson or the advice that we can humbly give when one thinks about putting together a program like this for drone response, is to take into account

**[12:37]** the connectivity and the power available where the docks are going to be installed, that everything is stable, that everything is continuous, that there are no of these kinds of problems. The coverage area that that dock, that drone, will have. The weather issue is somewhat uncontrollable, but weather forecasts can be made. The airspace, which is not too close to no-fly zones, that is, airport runway ends or things or places where flying is not allowed. Redundancy, something we have also learned and that all this has taught us, is trying to ensure that the coverage areas can be set up. And not to see the drone, as we were saying a while ago, not to see the drone as the only piece, but also to be in redundant areas, where the areas covered by each piece of equipment can overlap . Having something very super important: flight procedures. How to proceed, having clear procedures in case of an alert, what steps to follow. I check the weather, I

**[13:42]** check this flight, uh etcetera, etcetera. What condition is the pilot in? Tired? Not tired? It's practically like flying an airplane; there are certain safety procedures that must be followed, and the pilots are well aware of them . And of course another pillar is the people, the pilot, the person, the officer. In this case, luckily the pilots are police officers, trained, all licensed, and what's great about having police pilots is that sixth sense that the police have that a normal human being doesn't have; they see images and they notice certain things that we don't know. That comes from the expertise of being out on the street. And now, if you'd like, we can move on to the next slide. And here we have a summary of basically how the system is set up. We have the

**[14:43]** incident. We have the sensors to detect these incidents; they can be shot spotters, cameras as we mentioned, or phone calls. And FlytBase in the middle as a pillar, as we were talking about today, as a pillar that will integrate, decide, with the pilot of course also, not automatically, but operate and give us information, give us video. All so that the police response is in accordance with the situation caused by that event. In other words, as the La says, what FlytBase does is connect the incident with the response. Always thinking about one, about a platform that has helped us a lot, in terms of the number of teams. And here, of course, to conclude the talk a bit, so that something is clear, the goal is not to replace the police officer or the human being. That, for now, is

**[15:44]** impossible. In other words, the human being and the answer are needed. We saw it in the video; a human being is needed to go and rescue the person. In other words, that 's impossible. The idea here is to give the knowledge, to give the information to the officer who is in the field for his own safety and to know what to expect when he arrives at the scene of the event or situation that he was notified of. That's basically it.

**[17:48]** Well, uh, that will depend on the area that you want to cover. Well, I can't give information on how many drones are installed in Montevideo because it's a matter of public safety, and I can't say that. But it will depend on the coverage area of ​​ each drone. That's why I was saying that to set up the geopark, one has to know the maximum area that drone will be able to cover, taking into account that there will be procedures. I mean, a drone can fly 9 km, 10 km, but it has to be clear that when one is working in the geopark, afterwards that drone will have to stay, looking at videos, that is, it will be in the operation for X minutes, it's not just arriving, looking, and returning. So, that needs to be calculated carefully depending on the locations and the situation.

**[19:06]** No, not at the moment, there haven't been any complaints, I understand, of course I'm not the one who's going to find out about that, but no, not at the moment, there's no problem and the community hasn't been bothered at all. Remember that Montevideo has, sorry, remember that Montevideo has more than 10,000 cameras installed on the streets, that is, before drones there are the cameras that are installed, that is, there are also quite a few cameras for privacy on the streets. Well, the system responds to gunshot detection, yes, it's responding to gunshot

**[20:07]** detection and it's also used as a patrol system. As far as I know, the police haven't connected any other alarm inputs to it yet, but the platform allows it, so no, no, no, there aren't any major issues with that. But for now, it's responding to both gunshot detection and patrols or a 911 call. Please repeat the question.

**[21:22]** Well. For now, shooting a drone at over 120m high is not easy. I understand it's not easy, I've never tried this one and since the platform hasn't happened yet, right? But we understand that it is very difficult to shoot a drone at a height of 120 m. This should be a sniper. And the risk of the drone falling is practically zero; it would have to be a human error, a pilot error, because once it reaches a certain point, its battery automatically sends it back to its nest. In other words, the only way would be if a human prevented it from returning and the drone crashed, but that's not what pilots are trained for, and that's what the procedure we were talking about is for. Okay.

**[22:48]** Okay. Okay. Yes, the system is the one that the Ministry of the Interior has had for some time, it is the ShotSpotter gunshot detection system, that is one of the systems they are using and everything is communicated through the platform, it is the logical response that the platform generates. They enter the platform and from the platform sends the order to the drone. The dispatch can be either autonomous or manual. It can be used both ways. I understand that it's currently being used manually, meaning there's a pilot always following the command or always giving the OK to Fly B's suggestion.

**[23:54]** Well, I don't have that information, but I can give you an idea. [clears throat] The drone currently takes 15 seconds from the time the alert enters through the platform , between 10 and 15 seconds they are in the air and considering that it can travel at about 50 km per hour, it is very fast in the event. For example, the person in the video we saw arrived almost immediately at the scene, faster than a patrol car and , of course, avoiding any kind of accident from a possible car crash by trying to hurry to arrive. The drone travels very quickly to the area; it takes minutes to arrive.

**[25:14]** Thank you very much, and I hope you all have a good day too.