IA vs NIA: Architecture Selection Guide | Marvelmind

0:00 Hello. Today I will be talking about Inverse and Non-Inverse Architecture—the difference between IA, Inverse Architecture, and NIA, Non-Inverse Architecture. The difference is in the emitting beacon. So this is a very typical configuration for our starter set, which is the smallest configuration for 3D tracking. The configuration consists of four stationary beacons, one mobile beacon, and one modem. From this perspective, both IA, Inverse Architecture, and NIA, Non-Inverse Architecture, are the same. But there is a big difference. The difference between IA and NIA is who is emitting ultrasound. Let me, one more time, stress that in terms of radio, all beacons and the modem

0:58 are both transmitting and receiving. So when I will be talking about emitting or not emitting ultrasound, it's always referring to ultrasound. Also, another aspect before I am going to dive into ultrasound is that beacons don't talk to each other over radio. They all talk to the modem. So we have star architecture, at least at the moment. So back to ultrasound: in Non-Inverse Architecture, the mobile beacon is emitting ultrasound and the stationary beacons are receiving ultrasound. This is happening in the mode when it's already working. When the map of stationary beacons is built, at the same time when there is an initial stage, the map is being built. Then the modem is controlling each of these beacons and they can emit ultrasound. For example, this is emitting sound.

1:57 This is receiving ultrasound. And then the distance between them is measured and the modem is collecting a table of distances in order to build our self-built map of stationary beacons. So it means that even in Non-Inverse Architecture, the stationary beacons may emit ultrasound if you are using, for example, Beacon HW v4.9 or Super-Beacons. At the same time, you could use Mini-RX like this one. So Mini-RX cannot emit ultrasound. So in this case, for Non-Inverse Architecture, you would need to provide the distances between the stationary beacons manually or provide location of the stationary beacons manually because this is RX only—only receiving, not able to transmit anything. But it's smaller and, let's say, as versatile—let's say sensitive in terms of ultrasound—Super-

2:56 Beacons, but slightly less expensive. So back to Non-Inverse versus Inverse: so in Non-Inverse Architecture, this is emitting ultrasound and stationary beacons are receiving ultrasound. They all work at the same frequency, so usually 40 kilohertz. At the same time, in Inverse Architecture, which I would say is around 10 times more complex in terms of internal structure, stationary beacons are emitting ultrasound and mobile beacon is receiving. And from this comes their greatest benefit of Inverse Architecture because you can have many receiving beacons—many mobile beacons—without location update rate reduction. It's like in real GPS. In your GPS, satellites are emitting radio and your

3:54 receiver is receiving radio pulses from all those satellites. And we have millions of receivers receiving beacons. The same in Inverse Architecture: you may have many mobile beacons without location update rate reduction. In both cases, IA and NIA, through Marvelmind, we support 250 beacons. Beacons means combined stationary and mobile. In both cases, but the difference between Inverse Architecture and Non-Inverse Architecture, particularly, is in location update rate. So it means that if you take a Non-Inverse Architecture setup, at any given moment you're able to locate only one mobile beacon. So when you have one drone flying in the warehouse, that's great because I have only one drone and location update rate per drone and per

4:53 system is the same: 8 Hertz, 10 Hertz, 12 Hertz, around that value. But if you have 10 drones, then you would be able to locate—measure position of one drone, then another drone, then third drone, then tenth drone, and then the first one. So it means that if you have location update per system 10 Hertz, then per drone it will be only 1 Hertz. In Non-Inverse Architecture. At the same time, in Inverse Architecture, if I have a person and the person is being tracked, for example, with Mini-RX or with a badge or with a helmet, then we have many persons at the same time because these beacons will be emitting at the same time and this will be receiving and they can receive at the same time. But there is a great 'but,' and that 'but' is that the stationary beacons must transmit already frequencies—different frequencies, different ultrasonic frequencies. This

5:51 is why they have different resonant frequencies like 19 kilohertz, 25 kilohertz, 21 kilohertz, 23 kilohertz, 45 kilohertz, and more. So currently we have up to 5 frequencies, and you cannot have in the same submap the same frequencies because the mobile beacon wouldn't be able to distinguish the ultrasonic signal coming from which of the stationary beacons. Their great benefit of Inverse Architecture is that you may have many mobile beacons without update rate reduction. That's great. But at the same time, the complexity of the system increased significantly—again, 10 times more complex than Non-Inverse Architecture—because the mobile beacon must be able to receive at the same time different ultrasonic frequencies. They are pretty close to each other, filter them, detect the signal, and still be able, in terms of external noise and

6:50 in terms of interference, to calculate its own position. Because in this case, the mobile beacon is calculating its own position, not the modem. So it meant that Inverse Architecture system is more scalable. And again, update rate reduction is slow or not at all when you increase the number of mobile beacons. What is the drawback of Inverse Architecture? Since the mobile beacon is receiving ultrasound, you cannot place the mobile beacon on something noisy like a drone because the drone is emitting ultrasound, sound, in a very wide, nearly white-type noise which is going from audible to ultrasound range as well. And it would simply be blocking the reception of ultrasound from stationary beacons. So it means that instead of 30 meters, the

7:49 maximum distance would be reduced to 20, to 10. In case of very noisy drone, it could be reduced up to 5 meters, for example. So we do not recommend to use Inverse Architecture in case of drone or any kind of noisy mobile object. And we always recommend to use it when you have multiple mobile objects like people taking, for example, or multiple robots or multiple forklifts. Many means more than 5, 10, and those kinds of things. And update rate is important. So if you are okay with update rate 1 per 10 seconds, then Non-Inverse Architecture is good enough as well. But if you need, for example, 10 Hertz update rate per mobile beacon, then in this case Inverse Architecture is your solution. So once again, let me repeat there. The summary of differences and benefits and

8:48 present costs between Non-Inverse and Inverse: we support both systems. Both systems support 250 beacons combined. Inverse Architecture is mainly used when we have multiple mobile objects like people and when like update rate is important for you. Non-Inverse Architecture is used when you have noisy mobile objects like drones. In this case, Non-Inverse Architecture is your choice. Then let's talk about a bit about submaps. We will be having a separate video on this, but in terms of submaps and in terms of this architecture building, both architectures are the same. So if you want to have 2D tracking of a robot or of a person, you can employ Non-Inverse Architecture or Inverse Architecture. But you still build the map the same. For example, for 2D tracking, this is one room, another one, third room, and so on. In each room, you need to have at least two stationary beacons anyway.

9:47 Configuration, and then of course, them about beacon that would be moving from one to another. The difference in terms of submaps is complexity. Because when you build Non-Inverse Architecture, each of these beacons is working the same frequency. I will default to three frequencies: 31 kilohertz. But if you have Inverse Architecture, you cannot have the same frequency. So it means that the network planning is a bit more complex. We will be describing it more and teaching a bit more, but the logic is the same. So for example, you have one room. You have 19 and 25 in this room. You must also have different frequencies, but you cannot use the same 19:25. You must use 21 and 37, for example, because otherwise when they will handover between submap for five and submap for... the system would, you know, wouldn't be able to do the handover because the frequencies are the same. So the neighboring submaps cannot have the same ultrasonic frequencies in this.

10:47 Case. This already can do the same frequencies, but there is another complexity we'll talk about this different in different video in a bit more details. But remember that you can build submaps and larger maps with up to 250 submaps in both Inverse Architecture and Non-Inverse Architecture. So from this perspective, like capacity and capabilities, overall systems are the same. But there's a fundamental difference in terms of ultrasonic working. And so you can build as large and as complex maps, consisting of up to 250 submaps in, from both Inverse Architecture and Non-Inverse Architecture. As usual, if you have additional questions, please ask or send us a mail to Marvelmind. We will be happy to help. Thank you.