# Building a Proactive Mine Security Program at FQM Kansanshi with Autonomous Drones - Event: NestGen ’26 - Published: October 2, 2026 - Length: 29:15 - Watch: https://nestgen.org/library/watch/0QuIgYnhVTU/ > 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 Mine security programs often start reactive, responding to incidents after they happen rather than catching them early. Autonomous drones are changing that equation, giving security teams continuous site visibility that turns patrols from reactive sweeps into proactive coverage. In this NestGen session, First Quantum Minerals and FS Systems will share how a proactive mine security program was built at FQM Kansanshi with autonomous drones, and what mining teams should think through when moving security operations from reactive to continuous. ## Transcript **[0:00]** Welcome to the mining track of Nestl 26. Great to have all of you with us. I am Unati from the customer success team here at flight pace and I will be moderating your session today. Before we dive in, let's chat where are you joining from. So in the chat just comment where you've currently our speakers are joining us from Cape Town and from the Kstanchi mine in Zambia where it's I think early afternoon, right? Uh okay, picture this. It's a copper mine. so big that the fence line runs farther than any patrol can walk in. In a shift, organized groups come in on motorbikes out of the bush, head straight for the tailings and the equipment yards and are gone before the ground can even get there. Fixed cameras can, you know, have the coverage where they're installed, but both of us know, all of us know that the threat usually comes from somewhere else when a mine is this big. That's exactly the problem that Kansanchi mine is set out **[1:01]** to solve. Kansanchi is one of the Africa's largest copper mines and a part of first quantum minerals. Their partner in doing this is FS system who fly the program remotely from their control room in Cape Town which is more than 2,000 km away. And to tell that story, we have two speakers, one from each side of the program. First, Robert Peacock, the drone operation managers at FS systems. Robert runs the operational side of FS systems autonomous drone deployments including remote operation, mission design, optimization, integration and reporting. He has overseen the Kanchansi program from its first talk. Welcome on board, Robert. And joining us today is also uh joining Miss Martin is Martin Smith from FQM Kanchi. Martin brings an extensive security and operational background having previously served in special forces before moving into the mining environment. He has been with **[2:04]** the Kansanchi mine for the past 10 years and currently serves as the superintendent of operations. His experience and long-standing knowledge of the Cananchi operation gives him a value valuable perspective on the security challenges across the site. Martin, welcome to NextGen. >> Thank you. >> Welcome. >> Before we start, uh for everyone who's attending, feel free to drop your question and answers in the box as they come up. We'll have time for them at the last of the session. This session is recorded and will be published after Nest. To whoever is listening, please, you know, make sure to capture your favorite moment and share it on LinkedIn and Twitter with # Nestchen. A few of the lucky people who share might have a surprise in store. With that, Robert, I am uh opening the stage up for you. **[3:04]** Thank you very much Nati for that great introduction and thanks for having us and for everyone joining at home. Thank you very much. Um I think if we can get the slideshow up so long and I can start just giving the basic introduction. So like said again I'm Robert Peacock with FS systems and quite simply we run the whole process of drones when it comes to installation um setting up the operations automation integrations and then after obviously support and piloting ours. Yeah. Thank you very much for that. I want to take you through our evolution of our drone operations. We're not going to get too deep and too technical. Um it is quite a secure site and I know that there's been a lot of other webinars explaining a lot of the stuff that we're going to go through as well. So we'll focus on our main journey. Um that is the story of our evolution where we start with a focused we've got obviously got our problem. We start with the focused solution. We've then improved that model and then now we're going to scale as the capability as becomes part of the wider security operation. Obviously the biggest thought that we've touched on throughout um the index next gen is the integration **[4:07]** getting everything to work together all the systems to speak and not just having another standalone gadget but something that adds into the overall ecosystem. Awesome. So if we can move into the problem uh the reason that we got into this in the first place and what we needed and the main problem being the complexity and the sheer scale of the mining site as well. The drone adds something different obviously like explained we have fixed cameras all along all around the site. Um but we add that something different that mobile aerial sensor that can proactively patrol the priority areas and move quickly when it's sent to any other kind of event that pops up. Um if I can just hand over to Martin just to give us a bit of a background and explain the main challenges that they posed on site um before we came with the solution of the drones. >> Um very basic. Thank you Robert. Uh basically what we had was uh we got a full-on security system uh in place uh but it's uh very much one-dimensional **[5:12]** uh on the ground uh people uh cameras on poles and uh although we have one helicopter on site we're only allowed to fly in daytime. So, uh we thought of extending this capability um of the drones uh to integrate it with everything uh both the helicopter uh our cameras on site, the people in the bush and uh we added the the drone in the box idea. We discussed it with Robert extensively and we installed one and it actually evolved until where we are today where we cover most of the site at the moment. we we've got some phases that we are going through um automation and uh but at this stage I would say that uh we covered uh the gaps that we had uh 2 3 years ago uh we're closing up uh up quite uh quickly now. >> Brilliant. **[6:12]** So that was our our main kind of issue as Martin explained there and again we started small with it. We didn't just drop a bunch of docs and a bunch of cool gadgets for there and then we just let them kind of run with it. Um we started small. We got a simple installation. We have the dock. The first obviously dock we installed is a doc 2. Um where we had it installed close to the main security site um simply on a slab as obviously DJI recommends and quite a simple kind of setup for the initial proof of concept if you may. Um so centered around like I said the dock 2 and the aircraft through the FlytBase operations. Um the goal was not to automate the whole mine immediately but the first priorities like I mentioned is the reliable scheduled patrols, repeatable procedures, remote oversightes and coordination with the on-site team. Um again the main thing is we're not replacing ground patrols or replacing any of the systems they had in place. They already had great systems in place. We're just bringing the drones in to add to [clears throat] their forces. A force multiplier obviously makes them again like someone mentioned **[7:13]** in our previous webinar. It doesn't work unless you have a strong team behind it. You can have all these cool drones, have all these gadgets, but if you don't have the the set procedures, the organized personnel, the the real professionals in the background like Martin on site, um then it quite simply just becomes another cool gadget to play with and not into the whole ecosystem. So awesome. So now transitioning to the next one, once the operation was proven, the focus shifted from can this work because now we know it works to how do we make it work better and how do we integrate it better into the existing ecosystem. So if you can go yeah the placements and hours integration. So every new deployment um obviously benefits from the operational lessons of the previous deployment. Um as you can see as the dock placements evolve the straightforward ground level installations to a more purpose designated locations. Um our latest example being the one that you showed earlier is that elevated tower installation with surrounding **[8:13]** protective netting that we'll touch a bit more on as well. Um quite simply it doesn't mean or imply that every dock needs to be on a tower. Uh the main lesson is that the placement now considers the coverage, the environment, the protection and the infrastructure and the actual operational needs of the guys on site in the area that we work on. Yeah. So with saying that obviously our operating window grew um we initially started with about 5 hours a day of flights which we split between day and night operations. As we've grown, as an Martin mentioned, we started adding docks in. We now grew to about 10 hours a day. Um, with the main goal obviously to move on to the full 24 hours we're having drones constantly cover the site, constantly picking up and then giving that overall oversight to the ground teams just at a 247 level. I think if Martin would like to touch on any of that, what might have been the main kind of proving points and the factors that you guys felt on site that made you guys decide to okay, this is something worth investing a bit more in and moving to **[9:16]** probably I said 24 hours a day operations. What were some of the key kind of >> proof of concepts on your side? >> Yes. Uh Robert, if I can take you one step back just to explain to everybody. So what we had was um we had a Land Cruiser and we had a small DJI in the vehicle in a in a case and we go around and if something happened you rush there with a vehicle and get out of the vehicle, get the drone out, find a space first where you can launch it in the bush and then we fly the small drone, try and make coms with the guys on the ground, bring them in and uh the next step was to bring in the drone in a box. It's basically handsfree for us now. So there's we still have small drones in our vehicles that just for proddding, you know, if I can call it proddding. You just go out and you just go and have a look at the small corner in the bush or on the dumps. Uh but the the docs actually made it such a hands-free **[10:16]** um optimization that uh wherever you are we gave our drones names and uh I know drone one is southeast of me so many kilometers etc etc and uh basically we just say I want to see that side what is going on there and Robert and then jump in and show me and monitoring is not just in the vehicle we the The ultimate for us is where I can sit on my phone. I can sit on a laptop in my vehicle and uh and see what I want to see and uh move my people on the ground uh to make arrests. >> Awesome. Yeah. So, just to touch a bit on the integration here that we're moving to. Um obviously, we've proven our concept. We move forward from just 5 hours to 10 hours operationally per day with the main goal heading to 24 hours. And like I mentioned and touched on earlier, the most important site or the most important **[11:16]** layer is that integration layer um to have everything speaking to each other to have our visual or moss radar systems or any other information or technology that they use on site. That was my main kind of winning point I would say is speaking to the guys on site learning what technology that they are using and then thinking how can this work with the drone? How can we implement this? How can we use the drone to respond faster or work together with the existing system? Um so another example that we could have is a radar system that they have on site. Um a main thing was obviously we have this radar picking up targets but we can't necessarily put PTZs everywhere to see these targets. So the integration we came up with was just to dispatch the drone straight from that radar coordinates or that target coordinates attent. We obviously dispatch the drone straight to the area and that's the more deeper level integrations that we're now focusing on moving towards on site where they have these technologies. We focus on the integrations. We've kind of tested it on our demo facilities and now we're moving to that full implementation on site where it's going to be practically useful to the **[12:18]** >> Robert if I can just come in here. I just want to explain as well that we the key here is is location obviously GPS coordinates uh where are you or what do you want to see where where is this uh target and uh we use it in in not only the drones the drone is the one the tool that we tell to go and have a look uh the radar would give us an indication uh our range finding will give us a distance and that comes to a grid reference and the drones fly there and It it cuts time from some of the places on the mind. It'll take you about two hours to get to uh by road. Uh now what we do is it's a matter of uh we're doing the pre-check flights uh pre-flight checks and uh we fly the drone there and we have visibility on what we want to see where it is and that then the next step is to get the people on the ground in and uh do the need for that starting **[13:22]** impact. So proactive security is not only responding faster but is detecting earlier and maintaining awareness before the ground teams arrived such as Martin just touched on. Now um we can just touch on this one incident. I believe this is one of our very first successful intrusions that we detected. It was within a few few days of setting up the drones itself. You can see we got the guys. We picked them up early on the fence line. They came in and this was a great example of how the drone itself is used as a deterrent. um before the guys even get on site, before the ground team was even able to respond. So, we managed to detect these guys early as they were coming in. As the one guy kind of slipped through the fence, we responded with the drone. Um like I said, these were very early days, so it's the first time they might have even ever heard a drone, some of these guys. So, they didn't really know what was going on and we f focused on them. And as you can see, we kind of start pushing them out of the site. Um there's a couple good clips here. I think you'll see now the guy starts actually they've got slingshots. So, they'll start throwing things at the drones. you see the rocks going up and flying down. Um, it gives us a nice oversight. Again, they can't **[14:27]** get the drones due to the altitudes we we operate at, but then that gives us the further information to give the ground team the information to respond to. Now, they know what these guys have. They know where they're moving. They kind of know what kind of direction they're looking at. And that was the great part of to have that deterrence. As you see here, I believe you can play the video again. As we chased those three guys out, you see in the center, we moved back and we actually picked up a much larger group of guys that were starting to well get the ideas to start entering. Um, and the great thing was once they obviously saw the drone, they almost hold on the road there. Um, not sure what to do. And then that gives us time now again to send the coordinates to the ground team. The ground team then responds to that situation and makes the arrests or blocks them off wherever they need to be as well. Awesome. So supporting the ground team. So this is a nice example um where we have the drone itself picking up the guys and instead of moving in as a deterrence, we wanted to **[15:30]** make an arrest. Um so we kind of kept our distance, we let the guys enter. I followed them with high altitude. Follow them over the road. I think believe here you'll see. Yeah, you'll see as this bottom cruiser goes past, you're going to see a group of guys running over the road. Um, so now they're heading to the tails. There you see the guys starting to go. And if we can skip a bit forward now. So whilst we got this location, I waited for the ground team. Obviously, we show the ground team what's going on. I can move around. As the ground team gets in position, they tell me when they're ready. And then I swing the drone around and I almost push the intruders back into that ground team. Um, if you move forward, you start seeing the chase. Yeah, you see they're really busy chasing them through the bushes there. So, as the intruders are actually running away from the drone, we're kind of pushing them or guiding them into the grounding that's coming on the side over there. Um, which obviously led to a couple successful arrests, which was great. We can just play that last one. Yeah, the last clip is obviously just the final confirmation. Um, we had the second drone kind of stand, take off, go over to check, **[16:34]** and then again ensure that the ground team have the guy, give the overall overwatch as they're making the arrest, and then do that final sweep of patrols to see any other guys that might be in the area as well. I just want to mention here what we do do >> uh I just want to mention here what we do do is we we we plan uh in advance that uh when we do these operations we watch obviously watch your battery life your return to home and the next drone that is the the most accessible or the nearest drone will then be ready to fly and take over from the previous drone uh that we will automate later on as by looking at the u the overlapping areas where the these drones are flying. So one drone will basically uh in future hand over to the second drone to say listen I'm going home come in come and step in and fly and uh keep an eye on the on the guys uh with the same location it's easy to **[17:38]** do and uh I think we near uh this final solution for this Yeah, that's pretty much touched on what Martin said. Now, supporting the ground team. Um, that picture we have is just a a little example of a ground team or I think it might actually be one of the chopper teams just monitoring as we fly. Um, and obviously once we fly and as Martin says, we we coordinate with those guys. We coordinate between the drones. So, as that drone goes might have pick up the guys, he said low battery, we take that drone back. the helicopter will then come to respond or the next drone will then come to respond to go check out that area and carry on which is the great part of having that evolution moving forward and again the main part being involved with their already existing security operations infrastructure and procedures that they're following. The nice part of this is if I can uh just uh explain to you uh we have for instance a holding area where we have uh dog handlers and some other security personnel **[18:42]** uh ready made for any arrest or reaction that we need to have on onto a specific uh area. uh we identify weapons or what what force do we need and uh we plan the team accordingly put them in a helicopter and drop them off nearby and make the arrest. So uh the drone the capabilities the zoom capabilities helps us quite a lot uh to make those decisions to send the right people to the right place at the right time. >> Perfect. Um just to quickly touch on the new features that we've been using a lot um that FlytBase integrated with us. It's really been a great kind of boost on our operations. We have the coverage intelligence. Um obviously this moves from individual incidents to repeatable measurement. We're actually observing the areas we intend to cover. Whereas we know a flight path tells us where the aircraft is flying. The coverage intelligence now moves deeper and tells us what the camera is actually observing. Um are we hitting our assets? are we hitting the the coverage **[19:44]** the necessary coverage of the fence line that we want? Uh we getting the patterns that we're observing and from that we even pick up any blind spots. Um so that's the nice part. We kind of adjust our operations from these integrations where we can now see we fly, we measure the outcome, we might identify the gaps and then we refine and then we measure again. Um which is the great part of what what these kind of coverage intelligence reports offer us as well. Yeah. And then again once we've proven the value and measured all that effectiveness it gives us a much stronger basis to decide how to scale that's the important thing. The program is scaling through a much better placement longer operating windows that we described deeper integration being the main next step and the measurable coverage not simply just by adding more aircrafts. Um in the physical deployment obviously we kept it simple ground level. We've moved to a more purpose-built designed installations and we're obviously just going to keep evolving as we go. Um yeah just to save on time I think if Martin you want to touch anything on and before we start **[20:45]** getting into the questions. Um yeah thank you very much. >> Uh just to explain to you on on the same uh we use uh human intelligence uh uh information from from people around us, people from within and that information and the intelligence is fed into uh one central system that we use with the drones and the drones intelligence system. uh to plan our operations sometimes uh before there's any intrusions. We decide that we're going to do zone one or zone 14 or whatever because we know the the focus for this week is in that area and we we do our operations very specific that uh thus to save time and uh obviously manh hours uh to in a specific area problem area. >> Perfect. Uh thank you so much Robert and Martin that was really really insightful especially like seeing those videos where you can actually see the impact that this **[21:48]** technology that you know you've introduced onto the mine uh prove its impact. It's it's really amazing and what you makes like what you iterated again and again throughout the session makes so much sense that of course having this technology uh you know making sure that there are right processes behind it so that we can actually turn these detections into intelligent uh you know decisions that then maybe save us money uh keep our workers safe all of these things it's very crucial uh so thank you so much for sharing getting that insight. We have a lot of questions coming in. So, if you don't mind, let me just, uh, pull them up and we can go through them one by one. Perfect. So, uh, one of our attendees is asking how does the system navigate in a GPS denied environment? >> Um, so we've been quite lucky to not have that problem just yet on site. It is **[22:52]** quite a remote site. Um, and I'm not sure if Martin mentioned, but we operate on Starlink is our main kind of deployment for these drones. So, that is the big plus that unlocks a lot of doors to kind of have that reliable connection. Um, as far as GPS denied environments, we haven't had any issues. We've had some issues where obviously line of sights is an issue, interference from other equipment is a big issue. um the kind of again that where it comes with the placements of the docks establishing where we want it maintain or not maintaining but um what's the word I'm looking for having a look or breaking down what we get out of that dock and then adjusting it um based on the RF signals and stuff like that. So the main thing obviously we've used the one tough site we had at our initial deployments is obviously um the 4G dongle integration as well that DJI offers. So we put a 4G dongle into the dock and the drone and that again just switches over to the cell network. Um but yeah that is the main thing. Luckily we haven't had to deal with any major GPS denied environments yet. **[23:54]** >> So uh also like Dhiraj from our team has uh you know answer to this as well. So these are open mines. So drones do not enter the tunnels which is why the GPS issues uh are you know not like negligible. Perfect. We have another customer asking that they have been looking at drone security for their mine, but their biggest concern is coverage at the night, right? So, have you like figured out a couple of methods or sensors that make this possible? >> Yeah. So, mainly our operations especially initially when we started um taking place at night. Like I said, they've got a great security system in place um that gives them visibility during the day for most of the areas. At night, the main thing is having that thermal payload. Um and I wouldn't say there's any specific sensors and things to add on that we found, but to have that thermal payload, we operate great. Um you can see the see the intruders coming in from quite a good spot away. **[24:54]** Um obviously the doc 3's come with a bit better thermal camera, which has been great. Um but the main thing is actually just operational experience, I would say. Um it's having an operator that's flown these kinds of mission. He knows what to look for. Obviously the heat signatures are a bit different at night for what you're looking distinguishing between an animal and a person. Um for that I would say the main thing that kind of allows it further than the technology already does um would just be operator experience. >> Perfect. Uh also for everyone who's asking for Robert and Martin's email addresses, we will be adding those to uh the chatbot here. So feel free to connect with them after the session as well. Uh we have another customer who's asking us we're taking the last question. uh everyone who's so the question is basically with the mine being such a harsh environment with you know such processes going on how do you ensure that the drone **[26:00]** operations take place safely [snorts] and uh yeah >> yeah so um I think that would be the answer with most general drone operations is just have your maintenance down your regular maintenance having your schedules making sure that everything is operational Not just when the checklist says or every two weeks, but every day you check the systems, you're making sure everything is looking good. You check the cameras. Um, as far as the operational side of what we've been experiencing, like I said, we started with our DOC 2. Uh, the DOT 2 is quite a big jump from the doc 3. As you guys know, the DOC 2 is a bit more open. It's a bit more sensible to dust. Obviously, not as great operating in the wind. But in saying that, I've never had a problem even with the DOT 2. Um, it's held itself very well. Well, we got operated in a very very dusty area the one dock 2 is placed and although the dust does collect a bit sometimes over a few months um it's never been an issue for the dock itself. Just give it a wipe and these systems have proven that they are quite reliable in these operational um areas. **[27:01]** >> Perfect. I know I said >> if I can just if I if I can just add to to Robert there what we also do is um when a drone is dormant >> uh we either put it in the container or and either a mobile container or just a physical container and uh we just take it off. So uh the most dust that we accumulate on these drones is when it's flying. Uh but while it's static it's it's quite well protected. Yeah, we've done great. They've got, like I said, we mentioned that trailer you saw in the images before. We pushed that trailer, the dockers into the trailer completely sealed off and Martin and them even took that extra initiative to have customuilt dock covers as well for site, which has been great. It's been a great help, especially with the dock to use again. They just roll that cover over, patch it close, and then again that extra bit of environmental protection always helps a bit. >> Perfect. Uh, Robert, last 30 seconds. As I said it was the last question but one of our attendees is asking how long will it take for someone to develop that kind of **[28:03]** an experience. Uh and yeah >> um that's quite a difficult question to answer. It very much depends on the operator and the circumstances and the site that he's dealing with. Um obviously I had to learn quite quickly to pick up these kind of differences just due to the amount of animals on site versus the intruders coming in. um it takes a while but it it obviously as you go and the more you deal with these situations the more you kind of build that um information and that knowledge. So I would say depending on how demanding your operations is for an operator to get used to those kind of heat signatures and stuff. I'd give it a about a two months timeline of dealing with that stuff consistently before you kind of get your eye in for what you're looking for. >> Perfect. Thank you so much Robert. Thank you so much Martin. It is rare to hear the operator and the mind tell the same story side by side and it's very clear that the program is doing better with placement, longer hours, deeper integration, not just a lot of uh additional drones. Thank you so much for joining us at Nestchen **[29:06]** and uh we'll see you at the next sessions. >> Thank you much and thanks everyone for joining me. >> Bye-bye.