we want to use nRF52833 as BLE5.1 AOA/DOA locator

[Athos.Zhang]

we want to use nRF52833 as BLE5.1 AOA/DOA locator,here are some questions about to how to use it:

1. How many antennas are needed for direction finding/positioning? Is 6 enough? What are the maximum and minimum numbers of antennas supported?
2. What are the requirements for isolation, polarization, and cross-polarization isolation between antennas?
3. Do the polarizations of all antennas need to be consistent? Or should half of the antennas horizontally polarized and the other half vertically polarized?
4. Are there any requirements for the lobe width and radiation pattern on a single antenna unit?
5. What is the fundamental theory of the direction finding/positioning algorithm? Is there a ready-made algorithm that we can use?
6. Should the phase delay of each RF path be consistent? Are the amplitudes required to be consistent? What about phase error and amplitude error index ?
7. Can all GPIO ports be configured to switch antenna paths?

[Dmitry]

Hi,

Athos.Zhang said:

How many antennas are needed for direction finding/positioning? Is 6 enough? What are the maximum and minimum numbers of antennas supported?

Depends on your case. Two antennas can cover a 90-degree sector, if you need 360-degree coverage, you need at least four antennas. More antennas can improve accuracy with a good algorithm.

Maximum controlled GPIO lines is 8 - with SP4T switches, you can easy connect 16 antennas.

Athos.Zhang said:

What are the requirements for isolation, polarization, and cross-polarization isolation between antennas?

IMHO absolute polarization doesn't matter, as tags may have any orientation - two sets of differently polarized antennas would improve distance but increases complexity a lot. Of course, each two antennas involving in one measurement should have equal polarization.

Athos.Zhang said:

What is the fundamental theory of the direction finding/positioning algorithm? Is there a ready-made algorithm that we can use?

http://www.diva-portal.org/smash/get/diva2:1317231/FULLTEXT01.pdf gives a good overview of principles with the real evaluation.

Athos.Zhang said:

Should the phase delay of each RF path be consistent? Are the amplitudes required to be consistent? What about phase error and amplitude error index ?

Not necesary, but you should account phase difference in your algorithm. With equal antennas, amplitudes should be near equal (amplitude difference can give you a hint about mixing with reflected signal).

Athos.Zhang said:

Can all GPIO ports be configured to switch antenna paths?

Any GPIO that is not marked as "low drive, low frequency I/O only" in product specification.

[Athos.Zhang]

Hi Sir：

two more question:

1. If we design 4 pieces of  antenna element, For the antenna position, which arrangement is better? Antenna is on the one side of PCB?like this:

Or on the tow side of PCB? like this:

2. i can't open this link:http://www.diva-portal.org/smash/get/diva2:1317231/FULLTEXT01.pdf 

   can you help to share the PDF file in another way ?

[Dmitry]

Hi, 

here is another link.

Concerning first question, I'm not an expert in antenna design, maybe somebody with better experience could answer. Just notice that I can't see any difference in two pictures..