Need to track at least one object that would be located “outside” where the anchors would be.
Also need data, maybe fast.
Looking at the MDEK1000 kit as a possible starter for this and have some questions
With 3 anchors would this work if the 3 anchors were about 5 feet apart and the tag or tags were ouside or around the anchors? Issues?
if we had a location where the tags would never be found 180 degrees away… Could we make the system work with 2 anchors (again about 5 to maybe 10 feet between the anchors)
dilution of accuracy… if we had 3 anchors (or 2 described above) or maybe 4 in a rectangular config about 5feet square… If the TAG is say 40 feet away from the nearest anchor how accurate would we still be?
Data… Can we do data with the MDEK1000 dev boards? I see BLE as a way to talk but can they also talk over USB? What is the data rate we can sustain/burst?
I am leaning to get a dev kit but want to be sure before I spend the green.
The systems needs at least 3 anchors to calculate a position, ideally.
I would say that the geometric dilution of precision may be limited if you space your anchors only 5 feets. (1.5 meters)
If you increase the spacing, then you will be able to calculate a position out of the anchors area fairly accurately. A usual case is 15-20m between anchors, but I’ve observed good tag accuracy for a 3m*5m rectangle for example.
Regarding the speed, MDEK supports 10Hz position update rate.
I believe you will be using PANS which is the RTLS firmware delivered with the MDEK1001. In this case the BLE is already use for configuration and cannot be use for user data transfer. We will release an update of PANS in the coming month adding 34 bytes of data between nodes and a gateway (cloud).
The DW1000 supports data transfer as any radio, but if you wish to use it more specifically you will have to develop your own RTLS stack and it is a thorough job.
On your 3rd question:
If I set 3 anchors in a square pattern 1.5 m across and calculate the DoP 12 m away (about the spacing you ask for) I get a horizontal accuracy estimate of 50 cm.
If I add a 4th corner to the square that drops to 36 cm.
This is all assuming things are all at about the same height.
Also this is purely a mathematical model based on the geometry of the antennas, the real world has an annoying habit of not always matching idealized mathematical models. Consider this the best you could hope for rather than an easy to achieve accuracy.
For comparison with anchors at 3 corners of a 1.5 m square the same tool gives me an estimated horizontal accuracy of 4 cm for the centre of the square.
Interesting. Ive done some more measuments of where and how far the anchors would be and we are looking at a possible triangle of
2.5m x 2.7m x 1.2 m
We probably are looking needing about a 60cm accuracy. Maybe even larger
10 times a second is probably good. we are running other parts of the system at 5Hz.
That would be really cool to be able to send. We dont need to be fast and 34 bytes in a message is plenty (for now because in the future we may want more data but thats a ways off)
Right now we dont want to create our own RTLS stack. That is why I was hoping that the DW1001/MDEK would be able to do what we want.
Now I see that in the demo that there is location reporting on the android app over BLE. So for now that is working? and if I wanted to talk BLE to an anchor or tag to get the location of the tag then that is working?
(Idealy what I want to do is get the location of the tag via one of the anchors… and then later be able to send a message to the tag… but the first goal is to track the tag so I can show it on a screen)
You can query location from a tag over BLE. The anchors are not aware of tags location, you need to use a listener (there are some thread dealing with this in the forum, take a look).
MDEK objective is really to enable quick RTLS deployment and explore the technology, so I’d say it will give you a good idea of what can be done with UWB and also allow you to realise a first proof of concept for your project.