search

open-sdr / openwifi

open-source IEEE 802.11 WiFi baseband FPGA (chip) design: driver, software
GNU Affero General Public License v3.0
3.68k stars 629 forks source link

some questions about csi_fuzzer #163

Open xiaoxiaochuzm opened 2 years ago

xiaoxiaochuzm commented 2 years ago

Hello! I've been thinking about some problems in csi_fuzzer recently. I have the following two questions to ask you.

  1. What is the time rate at which the filter tap changes in csi_fuzzer? I guess it is consistent with the sampling rate of 40MHz, but I don't know if this is correct.I looked at the shell script.,it will sleep for 10ms after each ./csi_fuzzer.sh 0 $i 0 0 command, so the change rate seems to be 100Hz again.

  2. After enabling csi_fuzzer, what is the relationship between the change of CSI data obtained by MATLAB and the coefficient change rate of the filter? Does the tap coefficient change once and the 56 subcarriers of a data packet change once at the same time?

Look forward to your answer!

JiaoXianjun commented 2 years ago

The time when you issue "./csi_fuzzer.sh" and the value reaching fpga csi fuzzer, it takes effect immediately. Irrelevant to anything else. Please read the app note: https://github.com/open-sdr/openwifi/blob/master/doc/app_notes/csi_fuzzer.md

xiaoxiaochuzm commented 2 years ago

Thank you for your quick reply. After reading the document carefully, I found that the parameters in the csi_fuzzer.sh file are passed to csi_fuzzer_scan.sh through two for loops, and the tap coefficient is constantly changing. What is the change rate of the coefficient in fpga csi_fuzzer? In addition, I am concerned about the relationship between the CSI data obtained by MATLAB and the change of tap coefficient. For example, how many packets of CSI are affected by the coefficient change at a time? I think the information of 56 subcarriers of a packet is received at the same time. I read the timestamp and found that a packet is received in about 0.12 seconds. Obviously, the change rate of tap coefficient is greater than this value, so I don't understand the synchronization relationship between the change of coefficient and CSI data.