Question: Python parser
Not an issue, just a question.
On the Picoscenes website documentation. It states that the MATLAB parser does some pre-processing on the phase data.
Can we be sure this also happens with the Python parser?
I think it does but i'm not 100% sure.
Essentially I'm just wondering what kind of processing one would suggest applying on phase data. DWT? Hample? Some LP,BP filters? Additional unwrapping - but the parser already has it or no? I honestly don't know.
I feel the magnitude data is ready to go - as we can see a gesture around time index 1400 to 1600-, but I don't know what I can do on the phase data to have it work properly, it feels like the raw phase data is useless as shown in the figure below.
The generated data is acquired by using injector+monitor:
array_prepare_for_picoscenes 1 "5180 80 5210"
PicoScenes "-d debug -i 1 --mode logger --plot"
PicoScenes "-d debug -i 1 --mode injector --preset TX_CBW_80_HESU --repeat 1e5 --delay 5e3"
Using the following pseudocode:
1. stream = Picoscenes("somefile.csi")
2. first_frame = stream.raw[0]
3. csi_segement = first_frame.get("CSI")
4. N = len(stream.raw)
5. S = csi_segement.get("numTones")
6. Nsts = csi_segement.get("numTx")
7. Nrx = csi_segement.get("numRx")
8.
9. dimensions_stream = N, S, Nsts, Nrx
10. dimensions_frame_csi = S, Nsts, Nrx
11.
12. mag_stream = np.zeros(dimensions_stream)
13. phase_stream = np.zeros(dimensions_stream)
14. frame_idx = 0
15. for frame in stream.raw:
16. csi_segement = frame.get("CSI")
17. mag_data = csi_segement.get("Mag")
18. phase_data = csi_segement.get("Phase")
19. mag_stream[frame_idx] = mag_data.reshape(dimensions_frame_csi)
20. phase_stream[frame_idx] = phase_data.reshape(dimensions_frame_csi)
21. frame_idx +=1
22. heatmap(mag_stream[:,:,0,0]) #pseudo implementaiton
23. heatmap(phase_stream[:,:,0,0]) # ^
Any help would be greatly appreciated as my master thesis is running out of time, and I don't know how to continue. Finally I want to use the CSI Ratio model approach, but I think its beyond this issue tracker.