Hi all,
I am trying to measure the distance between two BladeRFs by comparing the phase of a pseudorandom (PN) code in GNU Radio Companion. Radio 1 streams the code (+/- 1s, so BPSK) on one frequency, Radio 2 just retransmits it on a different frequency. Radio 1 receives that delayed code. An FIR filter (taps are time reversed PN code) is used to perform the correlation. The signal that Radio 1 is streaming is added to the signal that Radio 1 receives and the combined signal goes into the FIR filter, which ideally produces two peaks, with the distance between them representing the propagation time plus the processing time at Radio 2 of the signal, which can then be used to find distance.
I understand there are random delays, mainly between the radios and software (and within the software) that we will need to account for, but first we need to actually be able to measure the overall delay.
Right now I am just trying to transmit from Radio 1 to Radio 2, and see a peak in the correlation output in GRC connected to Radio 2, but that is not happening. I don't know if it is a problem with my method or the radios. I have not been able to recreate the behavior in simulation. A distinct peak emerges in simulation even with noise 15 dB stronger than the signal. Time offsets simply cause the peak to scroll (I realize this will need to be compensated for) and it seems to handle relatively frequency offsets. I've tried putting taps in the channel model block but it seems to have little effect. The antennas are only a few feet apart, indoors.
With multipath, I'd expect to just see more peaks or a wider peak, although maybe since I'm sampling at 40MHz (7.5 meters range resolution), the reflections are within a single sample and destroy the correlation completely. Would that make sense?
Any ideas about how we can troubleshoot this more effectively? Or better model the channel?
Thanks,
Jonathan