GNURadio Reference Signal Test Setup 2 sudo stations

GNURadio Reference Signal Test

Setup • 2 sudo stations • • Pi 4 Hack. RF Amplifier Antenna • Both controlled manually (1 by hardware, 1 by remote desktop) • Recording 10 e 6 samples centered at 162 MHz sampling at 2 MSPS • 5 seconds of data • 10 data sets (pairs) were recorded

GNU Radio Recording Hack. RF Receiver Specify preamp voltage bias Saves samples as. bin file Function Parameters Cosine function with frequency of center_freq - channel_freq Limits number of sampled recorded • Python script run on the pi with all parameters as function input variables – can be called from terminal and runs till completion • The files saved are centered at channel_freq

Recorded Data Signal Presence 2 nd Pi 1 st Pi DC Spike at recording center frequency

Offline Processing • Step 1: Filtering • Use low pass hamming filter • Cutoff frequency at 12 k. Hz • Transition Width at 2 k. Hz Before Filter After Filter

Offline Processing – Matlab (Data Set 3) -114999 sample delay = 57 ms -71 sample delay = 35. 5 us -386323 sample delay = 193 ms The filtered IQ data was put through the three most robust cross correlation/covariance methods

Offline Processing – Matlab (Data Set 1 vs 3) This is a false positive from two sets of data that did not overlap in time. Measured 5187 sample delay (2. 6 ms)

Offline Processing – Time Domain Analysis • The information in the signal is frequency modulated • Signal has uniform amplitude • Cross correlation on the FM signal is not consistent Pure cross correlation on the direct IQ data will not yield reliable time delay results

Offline Processing • Step 2: Demodulation WBFM Receiver Block can be used without decimation to generate real samples at the sample rate

GNU Radio Offline Processing Function Parameters Saves data as a. wav file • Python script run offline with all parameters as function input variables – can be called from terminal and runs till completion • The files saved are centered at channel_freq

Offline Processing – Matlab (Data Set 1) 21887 sample delay = 10. 9435 ms

Offline Processing – Matlab (Data Set 2) 76891 sample delay = 38. 4455 ms

Offline Processing – Matlab (Data Set 3) 80170 sample delay = 40. 085 ms

Offline Processing – Matlab (Data Set 4) 43421 sample delay = 21. 7105 ms

Offline Processing – Matlab (Remaining Data) Set 5: 39180 Delay (19. 5900 ms) Set 8: 56437 Delay (28. 2185 ms) Set 6: 105247 Delay (52. 6235 ms) Set 7: 65355 Delay (32. 6775 ms) Set 9: 112859 Delay (56. 4295 ms) Set 10: 50060 Delay (25. 0300 ms)

Summary • Time delays were found using a reference signal • These time delays were due to latency in the recording process • In order to properly analyze the ref signal data, filtering and demodulation are necessary • These processes do not sacrifice time accuracy or sample data • Repeatable results were found across the 10 data sets