Autonomous Delivery Robot Navigation Guide | Marvelmind

0:00 In this video, we will be showing what the robot or what the system sees when the robot drives autonomously and delivers goods from point A to point B in the warehouse. So the robot is being filled in with buckets, and this is the robot on the map. This is exactly what the robot actually sees. We have the controlling system, so their indoor navigation system is covering several aisles, but there are what we will be using only this aisle and this corridor, and then brings to this place. But the system is larger, so the robot, if it actually could, would drive to any of these points covered by the map. So this is one submap, another submap, third, fourth, and so on. Some maps have many submaps, and here you can

1:00 see the service zones. And this is the service zone for this submap. It's basically telling the system where each of the submaps is, what area each of the submaps is covering. So this submap is covering this area, and this submap is covering this area, and the overlapping is the handoff resolved. So this is important when you do the network planning and when you do the tracking, how well the system is working and how well the handoffs are performed. This plot is player. It's embedded inside the Dashboard, and it's showing the statistics about the particular task. In this case, the drive of the robot total driving time, total time driving time, holdup time, inactivity, charging time.

1:58 Time. So you will see in real time the performance of this in a particular task. You can see performance for any task. So you can upload their log file and analyze post facto, or you can see the same in real time while it's performing. Let's check more. So as you can see, the robot received the command, just one button click to go from point A to point B. Robot knew in advance that this will be the path because this path was uploaded by the dispatcher, and basically the task of the robot is to bring from this particular area to this particular area. That's it, as simple as that.

2:57 It could do many other areas, so there may be point A to point B, point C to point D, and so forth. We covered all this territory, so it means that the robot could effectively drive from this dot to like this, to this, from this to this, and back—any to any. So that's the whole idea. And as you can see, the statistics are already changing. Please notice how the robot is controlled. So the robot sees not only location but also the direction. We are using our own feature, which is paired beacons feature. So there are two mobile beacons: one is here and the other one is here. And you can see the track from both, and by knowing the position of mobile beacon 20 and mobile beacon 21, the robot is able to calculate not only its

3:56 current XY position but also the direction, and the robot is using it in order to drive it directly in the right direction and right place on the aisle. As you can see, we didn't provide well coverage, so there was probably some sort of obstruction in this area. So there the coverage in this area should have been better. So we don't know the details, but most probably there was no direct line of sight between this beacon and in this area. This arc in this, you know, lack of coverage brought this jump. And there's also some strange behavior in the last meter of driving.

4:55 And also, you can see already the statistics based on this. So out of the total driving time, four minutes, the movement was three minutes and about one minute was inactivity—basically waiting time while loading. So this drive back, and the same story. So one button was clicked, but in this case the button corresponding to this point, point A, and the robot was driving back with the same load. And this notice now there will be a nice accidental test actually. A worker engineer from the

5:53 factory decided to test whether their obstacle detection and avoidance works on the robot, and it worked successfully. Please notice that the person jumps in front of the robot, or but stays in the place. Detects there, the person stops. Then when the path is again available, the robot drives and continues driving. OK, so that was exactly the time the robot successfully detected it, stopped with, you know, well in advance, no danger for the person. When the person went off the roof of the path, and the robot continued.

6:39 As you see, there's a hold of time now. It's becoming less and less visible because it was just some fraction, and most of the time the robot was driving. So these statistics are showing the utilization of your mobile assets. In this case, the robot—as you can see, the robot successfully delivered the goods in this case from point B to point A and ready to be offloaded. Thank you very much. One more time, check our website and ask additional questions if you have any at info@marvelmind.com. Thank you very much.