| | 41 | === Week 7 |
| | 42 | * Adapted a tutorial to send and receive OFDM signals via gnuradio and the GRC |
| | 43 | * Edited the GRC generated python code to script it so that I could have it automatically run trials given certain relevant parameters from the command line |
| | 44 | * Wrote up a draft of Experiment 1 about my data from the trials with the spectrum analyzer ^([#hn Experiment 1])^ |
| | 45 | * Discovered an error on the sandbox I was using, causing me to switch to using the grid as I was planning on doing anyways (on the grid, the radios transmit waves via the air, as opposed to sandboxes where the USRP radios are connected via a wire). |
| | 46 | |
| | 47 | === Week 8 |
| | 48 | * Continued tests to verify that there was an error on the sandbox I had been using (sandbox 1) |
| | 49 | * Learned how to work with different USRPs than the version I had been using previously (N210s as opposed to B210s) |
| | 50 | * Managed to get a connection between USRPs on the grid and download the data (and adapted the python code again so I could make it automatically run) |
| | 51 | * Wrote a few MATLab codes to analyze the data files, find the envelope of the wave to find the noise floor and the -3dB bandwidth, and calculate the total area of excess bandwidth contained in the wave ^([#hn Image 7])^ |
| | 52 | * Started plotting various sampling rates vs the excess area they had in the MATLab plots |
| | 53 | |
| | 54 | === Week 9 |
| | 55 | * Realized that the receiving bandwidth was too narrow for some of the sampling rates, causing me to set a standard for this experiment at 30 MHz |
| | 56 | * Finished scripting/automating the data collection process (including the naming of files to include metadata about the sampling rate and other factors at which the signal was sent) |
| | 57 | * Found a better way to account for noise by taking a sample of measurements with no signal transmission, finding the midpoint of each resulting graph, and averaging those midpoints to find the general noise level (as opposed to just assuming the edges of each graph necessarily represented the noise level). |
| | 58 | * Took 20 samples of each sampling rate to find a more accurate plot comparing sampling rate vs excess area with a 90% margin of confidence bar ^([#hn Image 8])^ |
| | 59 | |
| | 60 | === Week 10 |
| | 61 | * Wrote up Report 2 about the experiment from the previous week |
| | 62 | * Worked to get signals using the new spectrum analyzer to get data from the millimeter wave frequencies (with a center frequency of 27GHz as opposed to previous experiments which were centered at 5GHz) |
