MARB Testing:

Purpose: Determine effects of traffic on secrecy capacity.
Location: MARB, main floor

Procedure:

• A transmitter (ADALM-Pluto) was set up on the southern wall of the eastern doors on the ground floor of the MARB, approximately 93 inches off the ground
• An amplifier was attached to the transmitter (will check type when campus opens)
• 6 Receivers were set up on desks approximately 30 inches high at varying distances from the transmitter. See marbRxLocations_labeled.pdf for image (Note that 2 more were set up on stairs, but the data was unused)
• A laptop for each receiver was set up in a way that put the receiving antenna directly between the laptop camera and the transmitter, which allowed for clear identification of line-of-sigh obstructions in the recorded video
• Each laptop ran date_time_test.py (which generated a .avi video file and a .txt file with timestamps for each video frame) and MarbRx_Feb2020.m (which recorded the received signal, and timestamps, in a .mat file)
• Data was recorded two times: the first when the halls were mostly empty, the second during a class break when there was more traffic. Both scenarios lasted for four minutes.

Processing:

Most of the data processing was done in the 'DataProcessing' folder by the process.m script.
This file took in our receied data, pwelched & FFT shifted it, and then saved only the center 42 carriers (originally this step was done a little bit differently, but we adjusted everything to just take the center X carriers, and selected 42). The average noise power was scaled to 1, and then the reference signals were scaled so that the sum of the signal carriers was equal to one. This reference signal was then divided out of the data (in the frequency domain). Finally, the all the data sets were normalized by the overall max.

1. Pwelch, FFT Shift
2. Scale avg noise power to 1
3. Normalize reference signal so that carriers sum to 1
4. Divide the signal carriers by the reference carriers (frequency domain)
5. Repeat 1-4 for each frame
6. Repeat 1-5 for each channel
7. Normalize all channels by overall strongest carrier.

Data:

• Each .mat, .avi, and .txt file can be found on Box at https://byu.app.box.com/folder/103939543408

Other:

• Two other mini-experiments (Donut Capacity: Distance vs Capacity, and MarchControlledTests: Obstacle proximity vs capacity) were also done during this experiment, and are located in their respective folders with ReadMe.md files explaining them more thoroughly.

• Our initial results showed that for two individuals (Redd and Richmond, the closest two) the increased traffic levels increased the average capacity somehow, which seemed a bit weird. We decided to gather data again in June - 20 seconds with the area completely empty (thanks COVID!) and then 30 more seconds with about 5 of us walking around simulating traffic. This data is stored under "ValidationData_June2020" on Box.

• This data was used for paper XXXXXXXXXX that can be found at YYYYYYYY