search

wodls929 / BLE_AOA_Positioning

MUSIC을 이용한 AOA 기반 Positioning
3 stars 1 forks source link

how to understand this? #1

Open upupsky opened 2 years ago

upupsky commented 2 years ago

according to you code "calculate_new.py " file, line 47~49. d = phase_diff wavelength
e = 2math.pi*distance f = d/e I have some confuse. the phase_diff comes from the 250khz or 500khz IQ samples, but you used in 2.4ghz? how to understand this? thks.

wodls929 commented 2 years ago

In BLE structure, carrier frequency is 2.4GHz and CTE frequency is 250kHz. So I/Q sample frequency will be around 2.4GHz. Then I just consider I/Q sample frequency as 2.4GHz(Channel hopping was ignored in calculation). Am I missing something?

upupsky commented 2 years ago

but it seems can not "ignored" . the distance 12cm is matched with 2.4g. first of all, I can not understand the phase of I/Q . if the I/Q of the f=250kHz, the lamda = C/f, the lamda is not 12cm.

upupsky commented 2 years ago

can we discuss with real time chat software, such as WeChat, or others.

wodls929 commented 2 years ago

Sorry for late response. I was searching for related material.

  1. I agree that we should consider frequency hopping effect for more accurate calculation
  2. I/Q frequency is carrier frequency + CTE frequency. So, 2.4GHz + 250kHz = around 2.4GHz. This is what I want to express
upupsky commented 2 years ago

thank you. but I have some IQ data comes from HW, can you calculate the phase diff for me, sorry, I am still confused.

wodls929 commented 2 years ago

Is your data raw data? or was filtered? (filter like discarding data when switching slot)

upupsky commented 2 years ago

what HW are you used? I used nrf52833,not TI hw.

wodls929 commented 2 years ago

I used TI Boostxl-AOA.
According to BLE CTE structure, we should sort out useful I/Q data from raw I/Q data. So what I want to know was your problem is whether from data filtering. Did you filter your data earned in switching slot and reference slot already?

upupsky commented 2 years ago

maybe the I/Q is different from nrf52833. the raw data from one anntena , I = 3 -5 4 -4 3 -4 4 -3 , Q= -1 1 -1 1 -1 2 -3 2 . there is no data filter. can you calculate the angle ?

upupsky commented 2 years ago

can you leave an email. I send some materials, then we discuss. the theory https://infocenter.nordicsemi.com/pdf/nwp_036.pdf I used.

wodls929 commented 2 years ago

okay. My email is wodls929@gmail.com

upupsky commented 2 years ago

email is sent, thank you.