USRP Software Radio

Showing results for 
Search instead for 
Did you mean: 

finding Channel Impulse Response using GNUradio and USRPs

I would like to share my project where I am calculating the Channel Impulse Response using the Auto-correlation Method in the Time domain. I have uploaded a video demonstrating the experiment, utilizing the USRP B210 with antennas fixed at a distance of 3 meters.

However, I am not entirely satisfied with the output. I observe a single peak that changes its position frequently, despite the transceivers being stationary. The experiment takes place in a large hall with several walls, potentially causing reflections.


Here are my observations:

It seems that my known input sequence (of length 1024) is repeating, and the vector bins do not occupy the same sequence over time during the correlation process with the receiving vector. I suspect that the first receiving vector does not align with the known vector for correlation, leading to a changing peak position over time.

When I disable the sequence repetition at the source ("No repeat"), I observe nothing.

I am considering an alternative approach where I obtain complex values at the output of the FFT in the "file sink" and manually perform correlation with the known complex data. While I attempted this, the data I obtained was unreadable.

I have attached a flow graph for reference.

I appreciate any insights or suggestions. Thank you in advance


0 Kudos
Message 1 of 1