# CS434_Project README

Supporting Material

1. Code
  1.1 GNU Radio
    1.11 am_generator.grc
    1.12 treshold.grc
  1.2 Python
    1.21 power_superposition.py
    1.22 window_size.py
    1.23 test_powerlevel.py
  
1. Code
    1.1 GNU Radio
        ***NOTE: For all GNU Radio files, any time there is a file sink or a file source, update 
        the path to the directory that you are working from
    
        1.11 am_generator.grc
              This generates the AM signal. Can swap between the cosine wave (no noise) and 
              the audio input by plugging and unplugging the source on the left hand side.
        1.12 treshold.grc
              This sends values to the files so that the python program can determine the 
              averages needed to calculate the thresholds.
        1.13 backscatter.grc
              This performs the actual backscattering, sending a stream of random bits, 
              each repeated with reference to the size of a bit period. These bits are used 
              to multiply with the signal that is added to the carrier (AM signal) and then 
              are added together to emulate the reflecting of the signal by the two transmitting 
              devices.
        
    1.2 Python
        1.21 power_superposition.py
              This file implements the decoding algorithm discussed in the Power Superposition 
              Model. It calculates the averages of each threshold band based on the values from 
              threshold.grc. It then loops over the values of backscatter.grc, creating an 
              average of every window. It compares this average to the thresholds and determines 
              the bits that each device is sending in that bit window.
        1.22 window_size.py
              This file implements the decoding algorithm discussed in the Time Offset Model. It 
              also calculates the averages of each threshold band, but it ignores the middle bands 
              because it only cares about thresholds 1 and 3 (see paper). It then takes the values 
              sent from device A and the values sent from device B, offsets, and adds them. It 
              then decodes the averages across half of a bit period and determines the value of 
              the bit based on the threshold, and determines the device sent based on which window 
              it is in. 
        1.23 test_powerlevel.py
              This file is used to evaluate the performances of each model across a spectrum of 
              power levels. It modifies a file called config that contains the power levels used
              by the GRC files and reruns the GRC files and either power_superposition.py or 
              window_size.py. It takes the bit error rate for each power level and stores them in 
              an excel file, which we use to generate graphs. To change the number of iterations 
              (how many power levels between 0 and 1), change the value of the variable 'numiter'. 
              To change whether it tests the Power Superposition Model or the Time Offset Model, 
              change the import that is commented at the top (noted in the file).
              
              Recommended: Go into window_size.py and power_superposition.py and comment out 
              the print statements that print the decoded values if you are testing over a 
              large number of iterations.