Submap Architecture & Beacon Placement | Marvelmind

0:01 Hello colleagues. Let's discuss how to build submaps. First, about their topic—it's pretty deep and pretty broad. As you see, there are quite many items. We'll start with already available materials because we addressed this topic already, so check those materials before you go deep into this presentation and into this video. Why some maps are needed at all—let's touch this terminology. Many new words like NIA architectures, submaps, several zones—let's discuss them. Some basic starting hints before going to submaps: how to place the beacons, how to place the beacons properly. We have three architectures: Non-Inverse Architecture, IA (Inverse Architecture), and Multi-Frequency IA (Inverse Architecture). What are the differences and why we focus in this presentation on Inverse Architecture? Basic Non-Inverse Architecture, submaps, and maps are redundancy.

1:00 What is it for? What? Inverse Architecture. Inverse Architecture is the most complex topic, so this is why there are so many slides and there are so many ways to build it and there are so many ways to make mistakes. So this is why we suggest that you start with Non-Inverse Architecture, and only after mastering it, you go to IA (Inverse Architecture). Then some special cases, of course. It's a subject for, let's say, deeper discussions in next presentations. And of course, the summary—just to conclude: why the submaps are needed? There are many ways, but probably the three most important are the following. Example: you want to cover a huge warehouse like this one, and their largest submap they can build is 2G, for example. Submap would require that from the mobile beacon to the stationary beacons, not more than 30 meters.

1:59 But if you have the warehouse which is, I don't know, 200 by 200 meters, what to do? Clearly you can cover this—no problem. But you need to build some maps. You cannot have a single map consisting of only two beacons, for example, at the corners and covered. No. The maximum size of the submap—let's say, not the maximum size of the submap, but the maximum distance from the mobile beacon to stationary beacons serving in the submap—30 meters. So that's the first one: range. Then non-line of sight. For example, even in this, you have shelves. Shelves clearly have no line of sight, so it means that you need to cover between the shelves. Shelves separate submaps. Walls, rooms, floors—so there are many ways when you have a static obstruction, and that creates non-line of sight. And non-line of sight is the key element.

2:58 That is driving everything—where to put the beacons, at how many stationary beacons you have. But then you have non-line of sight from the mobile beacon. For example, some obstacles are moving or people are moving. Oh, there is the chance that your mobile beacon on the robot, for example, will be obstructed by something else. That's okay. There are some solutions, and we will discuss those solutions. So once again, three major ones: range, abstraction from static objects, and abstraction from mobile objects. Um, we have already plenty of materials—some videos on how to build those submaps. Check them. You know, post this video, jump to those simpler videos, and have a glimpse of what we will be discussing. So check it: Operating Manual again. Browse through the Operating Manual before you go deeper or in this presentation. It will help with some understanding.

3:57 Some terminology. So this is the extract, for example. It's not very deeply discussed in the Operating Manual, but it helps to start. So the Placement Manual also—a very nice source because it addressed several ways and several key topics like submaps, zones, numerous architecture. It's much deeper in this presentation, but check it first there, then step by step. The refresh video, which is basically saying one simple thing: don't jump immediately to the complex maps. Before you build a less complex one—and the least complex one is Non-Inverse Architecture in 2G with just two stationary beacons. Build it, and then go to more complex maps and submaps step by step. Then terminology: so what is the map? Map is a...

4:55 ...set of submaps. Map may consist of only a single submap. For example, if you get a starter set from us and you want to deploy a lovely 2D Non-Inverse Architecture, you place two stationary beacons, and your map will consist only of two stationary beacons, one mobile beacon, and one Modem. So it will be one submap and one map at the same time. That's it. So map can be very, very simple, but those are trivial cases. You don't need such a detailed video. So our Operating Manual is good enough, and Placement Manual is good enough. This video is about more complex areas when you need multiple submaps, when you need a complex Inverse Architecture submaps, etc. What is Supermap now? Supermap is a future item. We don't have it yet, but some customers are already saying, "Oh, we need more than 250 beacons. We need more than 250 submaps." So yes, it will be covered in...

5:59 ...the Supermap, but not there yet. So what is submap? Submap is a combination of two or more stationary beacons. And, oh, anyway, it will be discussed in much deeper details. So submap is a combination of beacons, a smaller map that will be a part of the larger map that we just discussed. In each submap, there's a table of distances. There will also be a separate slide about this. So, um, is a part? So table of distance is a part of the submap where the distances between each of the beacons inside the submap are listed—either self-measured or manually entered by you. Service zone is also an element of the submap, which is basically saying, "Okay, this submap will serve in this area." Because there's always an area where these beacons in this submap are...

6:57 ...trying to serve, but you manually define it. So it's very, very useful and very helpful when you build large maps consisting of multiple submaps. When you have only a single submap per map, it's a recommendation but not a must. Handover zone: it's very simple. You have two overlapping submaps, and the overlapping area between these submaps is a handover zone. When the mobile beacon is moving from service zone of one submap and then moving to this—during this time, it is served by two beacons, and then it's served by the second submap. So it's typical handover, like in cellular networks. Then IA (Inverse Architecture) versus Non-Inverse Architecture versus Multi-Frequency NIA (Non-Inverse Architecture). Their definitions could be many, but the basic distinguishing factor is: who is emitting ultrasound?

7:57 We are talking about ultrasound all the time. In terms of radio, they talk all the time, and remember that mobile beacons and stationary beacons don't talk to each other over radio. They talk to the Modem. The Modem is like a central controller, a spider controlling everyone. So they all talk to the Modem. So what? Once again, building submaps and building complex maps—it's not about the radio. Mostly not about the radio. It is mostly about ultrasound frequencies. So when I'm referring to frequencies, unless specifically addressed, it's ultrasound frequencies, not radio frequencies. So in Inverse Architecture, stationary beacons are emitting ultrasound. In Non-Inverse Architecture, mobile beacons are emitting ultrasound, and stationary beacons are listening. And in Multi-Frequency NIA (Non-Inverse Architecture), mobile beacons are emitting ultrasound but emitting ultrasound at the same time...

8:56 ...and on different frequencies. So we have eight different frequencies today available, and you can use up to eight ultrasound frequencies in IA (Inverse Architecture) and up to eight ultrasound frequencies in Multi-Frequency NIA (Non-Inverse Architecture). For regular Non-Inverse Architecture, it doesn't matter because your stationary beacons can be in any frequency because they are receiving. And for receiving, it doesn't matter because the Super-Beacons can receive any and all frequencies at the same time, but transmit only one frequency, which is native frequency for the beacon. Okay, ultrasound frequencies—we already touched. So there are eight frequencies today. That's great, because only one or two years ago we had five, which was fine for relatively smaller maps. But for large maps, the more frequencies you have, the more freedom you have. So this is why, when you are acquiring and you are not sure, get as many frequencies as you wish.

9:55 Because it will bring you the most flexibility for your maps. Okay, so back to a deeper definitions and explanations. What is the map? A map is a combination of submaps. For example, this is a map consisting of two submaps: one submap and another submap. Each submap consists of two beacons: Beacon 10 and Beacon 11. Beacon 11 and Beacon 12 is the second submap. Notice that Beacon 11 belongs to both submaps, Submap 1 and Submap 2. Today we support up to 150 submaps, and if each of the submaps can be up to roughly 1,000 square meters, so theoretically if you have a huge, huge, huge open hangar, it could be up to 250 square meters. So huge. One in practice the limitations are typically non-linear site limitations like

10:54 walls, shelves, some other things. So in reality, you can use those 250 submaps in much, much smaller area, or there could be some other limitations. Today we support up to 250 beacons in a map. Why? No, because each map is controlled only by one Modem. Remember, one map, one Modem. And when map is ready, in order not to move anything, in order not to, because usually you spend quite significant time trying to build the map, freeze it. So freezing map means that nothing will be changed in terms of mobile jobs of stationary beacons. None of their service zones, position of the stationary beacons—nothing changed. And you can freeze and save the map. It's very, very helpful because it

11:51 prevents accidental loss of this map, because it's a substantial amount of time you will build those maps. The future item is a super map. As mentioned, some customers are coming to us and say, "Wow, we love it. Can you cover even larger than 250?" Yes, we can. There's no dramatic change between map and super map. Simply, super map is a combination of maps, the same way as a map is a combination of submaps. So if each of the maps contain 250 beacons, by combining those maps into the super map, you can cover virtually unlimited space. Absolutely the same as cell networks. So if you are familiar with cell networks, think about our submaps as cells, maps is a cluster of cells, and a super map is a super cluster of those cells. So in this case you can cover, you know, the whole world. It doesn't matter.

12:51 So, as I mentioned, super map is a future item. As soon as you need it, we'll be ready. But so far all our customers are fitting into 150 beacons and 250 submaps, because it's still large enough areas. If you want to know about the super map, super map will be controlled by Super-Beacon. So Super-Beacon is basically a software that is collecting the data from each of the maps. Each of the maps is controlled by Super-Modem, because it's a large one, and Super-Modem is streaming data to the super map and Super-Modem. And Super-Modem shows you as the end user the whole super map as adjusted map. So if you don't know that it consists of other maps, you probably wouldn't pay even attention. Because for you it would look like just a huge map of thousands and thousands of beacons and thousands and thousands of submaps.

13:50 So it's virtually the same then. Submap. What is the submap? The submaps, as mentioned, can be one per map, like in your basic starter set, or there may be up to 250 submaps per map. So this is the submap. Remember, submap is a combination of beacons so which are combined in order to serve some area. What area? Service zone area. Even if you don't define the service zone area specifically or explicitly, like in this case you click, click, click, click and say these beacons or this submap will serve this area. Even if you don't define this, there is an implicit limitation of distance, which is 30

14:49 meters by default. So it means that as soon as 30 meters, unless you increase it further manually, the beacons will stop listening to ultrasound. So it means that they will not be able even to determine the location of their mobile beacon further than 30 meters. It's possible to set manually to 50 meters, for example. So our recommendation, of course, is use up to 30 meters, but you can set 50 meters with manual settings so you can cover, or you can build manually a very large submap like 40 meters or 60 meters. Yeah, in some cases it is even possible. It's not recommended, but in some special cases when it's very low noise and when you have low frequency ultrasound like 19, 25 kilohertz, yeah, vertically it can, in practice as well, can cover the distance up to 50 meters.

15:47 But in general it's up to 30 meters. So this is why for simplicity we say each submap is up to 1,000 square meters, so 33 by 33 meters or something like this. Submaps can be 1D, 2D, 3D, and this is example of 2D so XY. But if you have only one mobile or one stationary beacon, one stationary beacon, you can build a 1D submap. That's also possible. For example, when you have an air corridor or near an aisle, you don't care about their location. You don't care about that location, or tunnel, for example. You are caring only about, you know, X location or X coordinate. That's it. That's a well-recommended way to build a 1D submap. 2D setup is the most typical one because, you know, people usually do either people tracking or

16:46 robot tracking or forklift or something like this, so it's 2D. But also 3D. 3D doesn't differ from 2D. Simply, instead of 2, they can still have 4 beacons. We will be discussing more about redundancy, but 4 is recommended. Because in order to have 3D tracking, you need at least three, but one of those can be easily blocked. So in order to have at least some redundancy, our starter sets consist of five beacons: four stationary beacons and one mobile beacon. And four stationary beacons provide you three plus one redundancy for 3D tracking. This is why we have four stationary beacons for 3D. And recommendations for 3D, now already mentioned, for 2D, for example, you can build fully overlapping submaps, and we discuss them a bit deeper. And normal setup for 3D tracking is three plus one redundancy.

17:44 So the system is automatically choosing which of the triangles to choose out of their four different triangles from four beacons and three combinations. So automatically choosing non-crossing line. In 2D, it's very important. You see there's a nose, and if you drive the mobile beacon too close to this, then the system cannot distinguish between this and this area. We'll discuss a bit deeper, but the main message is: don't cross the line connecting to stationary beacons in 2D. In 3D, there's no such limitation. But in 2D, it's not possible because the system is not able to distinguish simply by geometry whether the mobile beacon is placed in this area or in this area, because it measures the distances and distance will be the same. So don't cross it. In majority of cases

18:44 don't even come close, because the error will increase exponentially. Because there's always some mistakes with placement or in calculations, etc. So do not come too close to this line and never cross it. Service. We will discuss about this a bit deeper, but we recommend always have service zones when you have only a basic submap consisting of one or two beacons. Then service zone is recommended but not a must. But then if I have more than one submaps, like in this case, sorry zone is virtually a must. Why? Imagine that distance between this beacon and this beacon is just five meters, and between these beacons—oh sorry—between this beacon and this beacon is also five meters. Okay, but remember that each of this beacon can

19:42 Listen up to 30 meters or even more. And for example, you have here a mobile beacon—not here, for example, but in this area. This area is certainly better to be served by these beacons because they are closer, because they are tuned to this direction, etc. So they are better to be served by this. But from this area to this area, I don't know, maybe 10 meters or 15 meters. So this beacon can easily try to serve the same, or actually these two beacons will try to serve in this area. But that's not good because there will be two opinions of location from this submap or from this submap about this location and from this submap about the same location. Now first of all, there are two opinions. Second, it creates additional load. When opinions match, that's great. But they're always mismatching.

20:40 And the system has to choose. But now imagine we are discussing only two submaps, but you may have many small submaps. And when the distances are like five meters—the size of the submap is just five meters—then you may have like five, ten different submaps overlapping. They will be absolutely a mess, and the system will simply not work. So our recommendation is very, very basic: as soon as I have more than one submap, service zone is virtually a must. Notice that a submap is a combination of beacons. But service zone is basically a zone or area of responsibility of this submap. By submap, I mean beacons—two beacons, or three beacons, or four beacons—depending on how many beacons you have. A submap. So submap and service zone is not the same. I hope it's clear. Then, heights of the beacons. Remember, for 2D, you must.

21:39 Place. You must put the height of both stationary beacons and mobile beacons because a submap has—you know, all beacons are placed at some height against the floor. The floor is some kind of virtual thing inside the submap. In practice, it's a real floor. And you put the beacons, whatever three meters, five meters, or ten meters, above the floor. In some cases it's a virtual thing, but nevertheless. So you must put, enter manually, because the system is not able to calculate the height. The system is able to measure the distance between the beacons and build the table of distances automatically, but not the height. There's nothing to measure against because there's no knowledge about the floor. So the height of the beacons, stationary beacons, must be placed, put manually. And for 2D, also the height of the mobile beacon, because it's 2D the system must know what's the height.

22:38 So if your stationary beacons are on three meters and then your forklifts on 2.2 meters, you must enter the height of the mobile beacon. For 3D, it's not the case. For 3D, only the height of stationary beacons is a must. And for mobile beacon, of course, in 3D, that is also measured, self-building. That's a great advantage of Super-Beacons because you just place them physically inside your—whatever house, area. Like you put one beacon, another beacon, and you build a submap. So this map is self-built because until you freeze the submap, until you've done this, the system is trying to measure the distances between the beacons—from beacon 1 to beacon 2, or in this case from beacon 10 to beacon 11, and from beacon 11 to beacon 10. And the system is basically populating the table.

23:37 Consisting of 10 to 11, 11 to 10. And when the distance is, for example, five meters zero one centimeter and five meters zero three centimeters—okay, system says, okay, just two centimeters difference. So it assumes that measurement was done correctly. Shows that the table of distances is white. So that's okay. You can freeze the map. Map is ready. Or submap is ready. For what? For building another submap, for example, this one. So once again, as soon as you are fine with placement of the beacons, you are fine with the table of distances, you are fine with the service zone, you can freeze the submap. Oh, sorry, freeze the submap. If your map consists of only one submap, that you can freeze the map as well. If not, then you build another submap, another submap, another submap, another submap. And then all the submaps are built, frozen. Then you can freeze the map because the map is basically a.

24:36 Combination of submaps. These are the table of distances. What I refer to with Super-Beacons or Industrial Super-Beacons or any kind of Super-Beacons which are able to emit ultrasound and transmit ultrasound: the table of distances can be built automatically. So you don't need to enter the location of the beacons manually. You don't have to measure the distance between the beacons manually. Even more, if it's only possible, we strongly recommend letting the system build the table of distances automatically. Why? We discussed this in separate topics because there are real meters, the ultrasound meters, etc. So before calibration, it's not the same. So make sure that the table of distances is white when you freeze it. So in this case, it's not white, it's green.

25:35 Because it was populated manually. But in normal conditions, it would be white. So like 30 meters, one to 129 millimeters, and 30 meters, whatever, 120 millimeters. Okay, nine millimeters difference from this direction to this direction. That's okay. It would be shown white. If you have any cells in the table of distances red, the system clearly says, okay, my distance from beacon and to beacon—do not match. There's something wrong. What's wrong? Okay, there could be no line of sight. There could be some suboptimal position. There could be line of sight, but they are not turned correctly to each other, et cetera. So the system is basically saying the table of distances is not ready. Don't move forward. Don't freeze the submap. Don't do anything unless you fix it because, again, everything can stop. And there is a separate video about this. So.

26:33 Until you have a perfectly white table of distances, don't move forward. Don't try to track. There will be poor tracking and jumps and everything. Table of distances must be white. And then you freeze it. Service zones: we already discussed. So I will probably do it relatively quicker. So we recommend to have a service zone even in a single submap. Why? Because if you don't build a service zone in a single submap and place, for example, beacons two meters from each other—that's okay, or two or four meters, whatever—you will think, okay, I will have a small map, so I will have a high update rate. No, you will not have a higher update rate because, effectively, if you don't build a service zone defining, whatever, three by three meters or five by five meters—what is your expected submap size—the system doesn't know that.

27:31 And the system will build a 30-meter submap. It's not shown, but it's a 30-meter submap. If there is a map at 30 meters, so ultrasound will propagate 30 meters before the next update rate. So instead of having a very small three or four meter submap, you'll have a 30-meter submap. So you define the service zone even for a single submap. For multiple submaps, for more than one, it's nearly a must. We already discussed. So if you don't define this, then there will be multiple opinions about the location about a beacon from neighboring submaps, and you'll have a mess. When you have two, three, it works. But it may work again. It may work. Do not recommend. It may work. But most likely, the tracking will be poor because there will be different opinions, et cetera. But if you have, for example, a relatively small area and many submaps, there will be complete mess. So always build service zones when you.

28:30 Have more than one submap. Service zones can be 1D, like map 2D, and 3D. Okay, in 1D, it's basically limitation of distance. So using your service zone, limitation of distance is basically saying to the beacon: do not listen more than the set value. For example, 10 meters. So in this case, limitation distance—after 10 meters of time-of-flight, the system will simply not listen to it. And if the beacon is, let's say, the mobile beacon is 11 meters, it will basically go: go, go, go, go—9, still tracked. 9.5, still tracked. 10, still tracked. 10.5? That's it. It's already not seen. It's outside. It's just, you know, went outside. No, not visible. So be careful with this because if you set too little, then your mobile beacon is outside and it's not tracked. It's simply not there.

29:28 So this is why we always recommend—if you're stuck, press default, kill the map, rebuild the map from scratch with default settings. The default is 30 meters. Most likely your mobile beacon will be inside this. This is also, of course, try to make as small a submap as possible, server zone as possible. But if you stuck or lost your mobile beacon, delete your map, save it first for the future, but delete it and rebuild from scratch with 30 meters service zones. Implicit limitation of distance: implicit server zones. For 3D, a submap is like a layer. You define the 2D submap and the height above the floor and below the floor, so it's like a layer. It has the shape of 2D submap and height—just height above and height below.

30:27 That's the 3D of that. Now, service zones and limitations of distance are very, very similar. If you don't define the service zone, there will be limitation of distance. But if you define the service zone, the service zone prevails and there will be no limitation of distance anymore. But otherwise, for example, what would be limitation of distance of listening to ultrasound for this beacon? It will be the maximum size of the submap plus some additional margin for safety. Already mentioned: service zones are important for many reasons. First of all is handover zones—this, you know, avoiding minority reports and multiple opinions. But another element is location update rate. Of course, typically you want to have as large an update rate as possible, but the update rate depends on time of flight to ultrasound. But time of flight to ultrasound depends on the size of this

31:26 map. It depends also on some other parameters—I mean the update rate like radio profile, etc.—but the largest contribution is from the size of the submap. So this is why the smaller submap you have, the higher update rate you will have. So this is why, if update rate is important for you, try to reduce the size of the submap by defining the service zone. Service zone today, you know, can be clicked—click, click, click. Like click, click, click, click, click, click. Up to eight points. So if for whatever reasons you would need more, let us know, maybe we will increase it. But today, service zone may have eight points or eight corners. Handover zone—it's very, very simple. Handover zone is a zone of overlapping of two submaps in this

32:24 zone. Let's say in this area, only first submap is serving. In this area, the second submap is serving. But in this area, two neighboring submaps are serving. And this is why there are soft handovers and hard handovers. Soft handovers are of course better because in this area both are serving. So when it's entering there, there are several types of handovers. There's first type, second type, and third type. First type is when it enters the handover zone—it already can be tracked by former submap and future submap. So first type of handover is when it immediately jumps to the new submap.

33:24 Second is: it goes, it's ghost. It's already served by both submaps, but it still reports location from the former submap, from the old submap, until it reaches the end of its serving zone. And then here it immediately jumps to the coordinates reported by the second submap. Of course, the whole network planning and placement is done in order to avoid any jumps. And typically the jump is not visible. But if the submap is not built perfectly or misaligned or something, there will be a small jump or not small, depending on how bad the error is, when it enters the service zone or handover zone, and when it leaves. So it depends. And handover type 3

34:23 is averaging. So when it's close, it's mostly on this submap, and then in the middle it will be 50-50, and at the end it's mostly this submap. And when it exits, it's of course on the other submap. So it helps to make these potential jumps smoothly. But when you build it, sometimes it's easy and nicer to have knowledge of this jump. Hard handover is when the mobile beacon is entering the service zone, but for some reason it's not handed over to the second submap, and the system must initiate, you know, hard handover procedure. The hard handover procedure is basically trying to find them about leaking from scratch. With soft handover there is no delay—it just goes, goes, goes, goes and goes through like this.

35:22 So it will go like this. So this will be the soft handover. In hard handover, it will go here, it will be lost for whatever reason, and then it will be lost for some location updates—maybe two, three, five, even ten location updates. So if you have eight times per second, it may be lost for one or two seconds before it will be found already here and then it will go good, like normal tracking. Hard handover, of course, must be avoided. But hard handover is better than being completely lost. Hard handover is a functionality, but this functionality is the kind of last resort when the beacons are lost.

36:21 How to place the beacons? Before we go to how to build submaps and multiple submaps, let's briefly touch how to place the beacons. And there's only one rule governing everything: line of sight, line of sight, line of sight. How to place the beacons? The beacons must be placed so that there will be least probability of non-line of sight from the mobile beacon to the stationary beacons. So this is why typically, not always, but in 95 percent of cases you place the stationary beacons high on the ceiling or high on the floor, and the mobile beacon is placed on the top of vehicle, robot, or drone. So this is why now the sensor is on the top because there will be least probability of obstruction. Of course, people can hide under the desk, table, or behind, I don't know,

37:20 a forklift. But those chances are smaller than if, for example, you put your mobile beacon on your chest or in your pocket. It's almost, almost guaranteed that you will be most of the time nonline of sight. So this is why stationary beacons on the top, mobile beacons also on the top. Stationary beacons are looking downwards, mobile beacon is looking upwards. Looking means this: because remember and check the operating manual, each of these transducers emits ultrasound in roughly 90 degrees beam. So this is why, for four, for four, you have 360 degrees horizontal coverage and vertically there's 180 degrees coverage. The microphone, which is slightly offset, has automatically 360 degrees horizontally and almost 180 degrees

38:19 vertically. It's difficult to say because, you know, to the horizon it still listens but with a weak signal. So the sensitivity drops at the horizon. So this is why we are discussing: your stationary beacons must be typically more than 20 to 30 degrees above the horizon. If your stationary beacon is very far, for example here, and it's almost at the horizon, then the sensitivity may drop and all nasty things may start happening. It's no subject for this particular presentation. But remember, the stationary beacon becomes high enough so it means it will not be next to the horizon. But for example, if you put it below the horizon for the mobile beacon, like like this, below the horizon, the microphone will be simply not able to even hear because it will not see it—it's below. You know, it will create the shadow.

39:18 will be in the shadow. For example, here's the shadow for this microphone. So this is not a shadow, but this would be the shadow. Clear? So basically, line of sight—line inside line of sight. And you place a beacon so that you would cover the same area with the least number of beacons and the least chance of the abstraction. And typically, least chance of the abstraction, and the least number of decants—place them high on the wall, not too high, not too low, more than 20 degrees to the horizon to the mobile from the mobile beacon—and with the least chance of the abstraction. And remember as well, if possible, place the beacon so that you can build the smallest submap and the smallest service zone, so you'll have the highest update rate. Other hints already mentioned: remember to put the height

40:18 to the stationary beacons. Those are typical mistakes—avoid them for my build, for mobile beacons, for 2D and 1D. Also, the height is a must. Otherwise it will be, you know, not measured location. It will be measured, but it will be offset. So you don't need that error. You see, sometimes when you install the beacon in the corner you don't want, or you don't need the sensor emitting upward. For example, what's the point? There's no nothing to be served. So disable their transducer facing upward if it's not needed there. Or another transducer is basically facing to the wall neighboring—you don't need it. Typically it doesn't hurt, but it, you know, additional power, additional something. But sometimes it hurts because it emits. There could be some special standing

41:17 waves or something. And if it's too close, then it may distract the pulse. The ultrasound powers—because, for example, imagine for the large distance you have. We typically recommend you enable 50 pulses. What does mean 50 pounds? Each pulse, and for our let's say typical 31 kilohertz frequency, each pulse is around one centimeter. So the length of the pulse is 50 centimeters. If your wall is like 10 centimeters and you emit to the wall, and your pulse is started running—of course, the reflected signal will come later, but the reflected signal may go in 180 degrees phase, which means that it will try to kill of our main wave. Of course, the first part it will not be able because it's already delayed. It goes this way and this way. But their last part it may effectively

42:15 kill. So instead of 50 good pulses, you'll have only 10 or 15 good pulses, and the rest will be probably killed by your too-close reflection. How to combat: easily, don't place too close to the walls. Close means closer than, let's say, 30 centimeters. Don't place, or don't enable their transducers emitting to the wall. There's no point—it will not help you, but it may harm you. So and then rotate the beacons. For example, we recommend okay installing on the wall, but in reality, don't install on the wall. Install on the wall and slightly rotate to the center. It's better because you know you want the best coverage. The best coverage means like this. Remember that each of their transducers gives you around 90 degrees. So sometimes you need only one transducer like this one. Only five

43:14 example: you have a corner. Corner is 90 degrees. So if you if you enable in the corner right in the middle, so you disable all the transducers and all the transducers from this, and only transducer number four, which is the center one, which is pointing to the center of your service zone. And from this it's perfect. It's perfect for several directions. Now, first of all, there will be no this additional cancellation because you installed in the corner. Second, remember our system is very precise. So people typically want to achieve even highest, highest precision. And highest, highest precision is achieved when you have a single emitting transducer, not five. Five are great, you know, to provide the courage, but they are slightly offset from each other—like one, two centimeters. And it contributes. It's very difficult to say, but it's not a dot anymore. So it's kind of a thick dot, and it brings some uncertainty.

44:14 Difficult, again, to prove it mathematically. But if you have only single transducer, single emitter, it certainly helps because it's much more smaller dot. So you'll have higher accuracy. It also helps. And that helps also because of power. Instead of driving five transducers, you will drive only one transducer. You know, it significantly saves on power. And sometimes even the range, because you know the same power embedded, or you know, put into five different transducers and one transducer—it helps. Typically doesn't much because transducers are anyway saturated by their power. But sometimes it helps. So remember, the modem can be placed anywhere as soon as it provides the coverage to each of their stationary and mobile beacon. So we do not discuss much the modems. There is a separate study, radio

45:12 presentation about this. So when we are referring to frequencies, or when we are referring to anything, the modem can be anywhere. It may have no line of sight. It doesn't matter. But of course, it must have radio connectivity to each of their beacon—mobile and stationary. Particularly difficult is about because for station you can guarantee, nearly guarantee, you know, because you're tested, you have errors aside. The mobile beacons are moving. Antenna conditions are different. All the conditions are different. So for mobile, they can make it. It's more difficult. And even more why, these future submaps are available. For example, you have a high building with several floors. It's very, very difficult with one modem to cover several floors. It's possible, but some special antennas needs to be arranged in order to provide because radio signal, they have a very weak up to 10 dBm.

46:11 Radio. It's very difficult to provide the coverage of radio through the floors. And hint number two: this special video about this already, so check those videos. I will not touch too much. But in some cases—again, I will not go—some special cases when the beacons are too wide and too close. For example, this would be already. This is a different subject, but this will be too wide map because the mobile beacon is too close to the line of this. And when it's too narrow—when the mobile beacon is too far, for example if these two would build a map under, and the mobile beacon is here—it's too wide. Two why it means when the ratio is, I don't know, one to ten. Typical ratio, recommended ratio is, you know, one to one, one to one-third, one to three—something like this. That's

47:09 good. But one to ten or ten to one is already too much. Besides that, there is a special configuration about the precise. Check videos as well. I will not touch it. But pay attention—it's about starting, starting copter from the floor. If you place the beacons on the floor, there will be difficulty with that before it flies away. Check it more carefully there. So this is why, for example, there, the drone must not fly away but flies inside their building. The easiest way to achieve the precise is basically place their stationary beacons on the ceiling and never fly above the ceiling. Then you basically don't have any issues with precisely that at all.

48:08 Um, about their 2D and not crossing—I already touched it, but let's look at this. Why the problem is with not crossing: so this is a typical 2D submap. It doesn't matter—inverse architecture, non-inverse architecture. Does matter the same, the same way. So this is the real location, and you see this is a nose showing that we are serving in this area. But if you come closer, closer, closer, closer, closer, and even cross, then the system is not able to distinguish. You may be here, but the system will still believe that you are here. So, for example, if you move like this, so in practice, the system will show a mirror because the system will not be able to distinguish. It will be still able to listen to this area, but it will not be able to distinguish.

49:05 Whether you are here or here, basically because of basic geometry based on two stations, because it's impossible geometrically. But our recommendation: not even come too close to this. Why? Already mentioned: too wide, too narrow. As soon as you come too close, how the system works—it's trilateration. In this case, bilateration. It measures the distance, and the distance from this area and this area is almost the same. So the system may start confusing. And if you do not calculate or measure this very, very precisely—for example, you made whatever two centimeters error—so when you are too close, this will look like you're already on this line. So you will have all kinds of nasty things. And particularly, it's problematic when you have a high—for example, you place two. So this beacon and this beacon are this, and you know, over there. So these are two these

50:05 beacons, and they are placed, or whatever, three to five meters above the ground. And when you come too close to this and try to track, it will be again this too white and too narrow error. And effectively, you'll be very, very, very close to this line. Because in relative terms, you need to measure long distances, but you are very close to this line, and there will be high chance of the error, which will manifest itself. Then it will go, go, go, and then suddenly it will jump. For example, on this. Because it will be—you know, system will believe impossible. It's possible you're still not crossing the line, but because of margin, because of errors, it may put you on this. The rule is very simple: do not come too close to the line connecting two beacons. How much? You know, there is no clear-cut rule. It depends how well you build the map, but

51:04 overall, not closer than one tenth of the distance between the stationary beacons. Again, it's not a real rule; it's just, you know, a hint and advice. Let's stop with this first part because it looks like the presentation will be even longer. So check the second part and probably the third one. Thank you very much.