calibrate the CSI phase

[yujianyuanhaha]

greetings,

I have been able to collect CSI and the Amplitude is pretty stable over a short time(not for long time), while for the phase (even the offset phase) is in chaos. Have anyone met the issue and got a solution? @DanielAW @schmittner @matthiasseemoo @zeroby0 @mzakharo

There have been some study calibrate the intel 5300 CSI-Based Fingerprinting for Indoor Localization: A Deep Learning Approach and ArtherOS firmware π-Splicer: Precise CSI Measurement With COTS Wi-Fi Devices.

I mentioned it cause I note the amplitude is not stable over the time issue, while some signal processing way, especially delay-profile or STFT require both the amplitude and phase information.

Best.

Jet

[wanghe97]

Hi,

Have you been able to find a solution yet? I have the same issue.

Thanks, Edana

[evileleven]

There are several ways to eliminate the random phase offset, such as: linear transformation; CSI tario; CSI multiplication; and find reference path, which are the methods that I know right now. There are papers about the above methods.