Track 10 Forklifts ±2cm Across 660ft | Marvelmind

0:07 Hello colleagues. Let's talk today about tracking of forklifts in a warehouse. This is a very, very typical task and we have an approach many times with this. So let's discuss in details how it could be done or how it is done. No, first of all, what's the task? In this particular case, every warehouse is around 200 meters by 20 meters, and the task is to precisely track 10 forklifts in real time in order to analyze their driving behavior, spending time, and also allocate the tasks to the drivers based on their current location. How is the system arranged? Each forklift has a mobile beacon Super-Beacon installed inside and is being tracked. There are 14 stationary beacons installed on the walls around 4 meters above the ground.

1:08 These 14 Super-Beacons are arranged in Inverse Architecture. Please check on our website about the terminology. They are divided into 11 submaps. This is a submap which is covering this area, and you can see this gray service zone for this submap. And there's another submap covering the same area. What is it for? It's basically for resiliency. So it means that the system will pick up automatically the best signal from both submaps. And if the signal and, let's say, measured location is not matching, then it will be using the performance or, let's say, the allocation from the best performing map.

2:08 It's also a resilience against the obstruction because if any of these beacons are obstructed, so the mobile beacon cannot see or hear the signal from any of these beacons, then it cannot be tracked. But at the same time, it may be seen by the beacons from the opposite direction. It's basically to increase the resiliency of your network. So there are several overlapping submaps. So this one and two are overlapping submaps, three and four, et cetera. In this area, there is no overlapping submap simply because the area was not narrow and it wasn't possible to install the beacons on this side because of the shape of the warehouse. In this area, there's an interesting case. The distance between, for example, this stationary beacon and this

3:06 farther end of the service zone is significantly more than 30 meters that we recommend. It's up to 40 meters, and in this case it could be even more. Why? Basically, when we did the network planning, they were born beacons originally. But during the rollout, the integrator made some mistakes, and we had to adjust the network planning on the fly. And effectively, they rearranged the beacons so that low frequency beacons, stationary beacons like in this case 19 kilohertz and 25 kilohertz beacons, they can tolerate their large distances more than the recommended 30 meters—in this case, up to 40 meters. And also, we deployed the forklifts using omni microphones, which are extending the range of confident tracking even further. So this is the forklift. This is

4:05 forklifts tracked. These are other forklifts, and these are the tracks of the forklifts. And these are the stationary beacons, which are facing down and basically placed in such a way that they would provide their best coverage, the best ultrasonic coverage to the mobile beacons. The mobile beacons are placed on the forklift, and the omni microphone is on top of their forklifts' roof so that the stationary beacons, which are high on the ceiling or high on the wall next to the ceiling, would be seeing or hearing their mobile beacons as optimally as possible. What data do you have? No, first of all, you are getting the raw data about the location. Second, there is a real-time player, which is

5:05 basically a special filter and special smoothing which allows occasional jumps or missing location data to be tolerated and smoothed. Additionally, all the data is recorded in a large CSV file, which allows post-processing. And using post-processing, it's possible to eliminate even further all the jumps and make the signal absolutely smooth and absolutely flawless, of course, when the jumps are not too long and when there is sufficient data to interpolate their measurements. It is used for this analysis of the data and is basically the core where the system provides. How did we deploy this system? We deployed the system

6:03 remotely. So we did the network planning based on the inputs from the customer—basically, the floor plan, the task, all the needs—and then we provided the data and the settings to each and every stationary and mobile beacon, and we shipped those beacons 5,000 miles away. Then the beacons were installed by their customer's integrator subcontractor, basically physically installed inside the building. And then we connected to the system remotely. We didn't fly there. There's no need. This is why it's possible to deploy the system all over the world, and we do this remotely from our location to your location over TeamViewer. So basically, we have full control over the system during the deployment. We can tune. We can adjust

7:03 we can provide the settings. We can optimize the handovers between the service zones and in the handover zone until the tracking is smooth, like in this case. And then we basically hand over the whole network to you, and you can use it as a complete map with the data being streamed to your location. So if your case is similar to this or you have something in mind, don't hesitate. Go to our website www.marvelmind.com or send us an email to info@marvelmind.com, and we'll be happy to help in your case as well. Thank you very much.