Large Scale Network (Initialy Tag follow up over 2 network PANID change)

Wireless Connectivity Ultra-Wideband
, ,
1

Hello,
After reading the documentation I found that the extend of a network where we may localize a TAG is a rectangular of 30Anchor*60m/2lines by 60m.

I was interested by a longer/bigger surface localisation system.
I planned to have the TAG connected to a computer.
As there is a limit to the expand of a PANID network, the tag needs to change network.

So I did imagine to have the tag changing it’s PANID from one region to the other. But I think it is not a good Idea because the dwm_panid_set writes also in the nonvolatile memory. Is there any way to change on the fly the PANID without writing to the nonvolatile memory?

Is there an other way to extend the localisation of a tag to multiple PANID networks ?

May be in the next release, it will be possible to change on the fly the PANID?
Thanks.

2

Hi DGE,

A few points :

  • The system can scale further than a rectangular of 30 anchors, you just need to make sure there is enough distance between anchors so they are not all clustered together.

  • The tag cannot change network easily, the system was not design to support this case. I guess you best chance there will be to write the non-volatile memory to change the panid.

  • A the moment, nothing is decided with regards to an additional release.

Thank you,
Yves

3

I do understand for the change of network.
What I did understood is the fact that with one initiator we should have only 30 Anchors.
What I do understand from your answer is that we can extend the network further. (Let’s say with a grid like in Figure 12 of the system overview).
I imagine that when I do this I need more initiators in the grid to synchronize the network on the right of the figure 12.
What is still unclear is how the border of the two regions builds.
It is said : “Fully scalable network: Anchor seats and TWR slots can be reused is sufficient space between nodes”

Does this means that we need to have a free space between the two zones so that anchor on the left don’t hear anchor on the right?

If so this means that in between the tag localisation is not very good/redundant.

Is it possible to have a handover between the two zones?

Could you please explain.
Thanks in advance.

4

Hi DGE,

if I understood correctly then your goal is to have a large installation where you think you might need more than 30 anchors.

PANS is fully scalable but a certain rules of deployment need to be applied. The 30 anchors limit is not per Initiator. A network can have one Initiator and tens or hundreds of anchors.

A very simplified explanation of the 30 anchors limit can be: each anchor requires a certain air-time reserved to advertise some network information. There are 30 of those air-time in the PANS design. If you put 40 anchors in one area where their signal would influence each other, then only 30 would get the air-time allocated. The rest need to wait until some air-time would be released. If the 40 anchors are spread across the area of interest, then the system will work and the resources can be reused.

As the tag passes across the network, the handover mechanism will function in the way which would allow the tag to locate and transmit the uplink/downlink data seamlessly.

I hope this would help.

Cheers,
TDK

5

OK now I do better understand and this does answer my initial concern.

Taking the "Figure 10:Scaling the DRTLS showing anchor’s seat numbers ", and with your explanation, I do now better understand.

Could you tell me if the following assertions are correct:

Taking as correct that the transmission circles in the figure show the exact limit of communication between two. anchors.
The Initiator A only sees slots occupied by A,B,F,E. A sees over slots as unoccupied.
The Anchor B only sees slots occupied by A,B,C,E,F,G . B sees over slots as unoccupied.
The anchor F only sees slots occupied by A,B,C, E, F, G, I, J, K. F sees over slots as unoccupied.
The Anchor J only sees slots occupied by E,F,G, I,J,K, M,N,O. J sees over slots as unoccupied,
Depending on the algorithm, boot time, … I imagine that N could have taken the slot of A,B,C because from its point of view it sees the slot as free, and all its neighbours (I,J,K … also), so it can be allowed to emit on this slot…

This explains why the network can be extended to infinite number of Anchors (save other limitations :slight_smile:).

Thanks in advance to your.

6

Hi DGE,

yes, you got it right. A new joining node can take any seat which is free from its point of view and also which is free for its neighbors. When the environment changes (which can happen when e.g. doors or obstacles are removed), the collision might happen. In that case it is detected and a collision resolution algorithm would solve that. The system should get into stable state after that.

I hope this helps.

Cheers,
TDK