5G Millimeter Wave Leakage

About This Project

The goal of this Project is to determine how well a given millimeter wave frequency (i.e. waveforms sent at frequencies above 20 MGhz, as 5G will need to be) can stay within its allotted bandwidth. This will be accomplished by measuring the amount of "leakage" - the amount of extra energy that is present due to the transmission of a wave - at various frequencies between about 25-28GHz.

Week 1

• Learned how to use some initial Linux commands and access nodes (i.e. remote computers) in Orbit Lab
• Discovered how to image nodes to load a previous software setup
• Began first tutorial on sending and viewing analog waveforms with Software Defined Radios ^{(Image 1)}

Week 2

• Sent analog waves of various shapes (Square, Sinusoidal, and Sawtooth), gains, sampling rates, wave frequencies, and center frequencies
• Tested hardware capabilities of a node of wave frequency vs sampling rate holding other factors constant
• Learned what WinSCP is and how to use it to get files from Orbit Lab nodes onto a personal laptop
• Figured out how to use WinSCP to transfer supplied code to a node
• Studied and adapted MATLAB code to plot waveforms in the time and frequency domains ^{(Image 2)}
• Began researching OFDM (Orthogonal Frequency Division Multiplexing)

Week 3

• Began sending digital waveforms with modulations (so they could actually carry information) using GNUradio and the benchmark_tx and benchmark_rx programs
• Tested out the differences between various modulations and sampling rates using the programs rx_ascii_art_dft to see the waveforms and benchmark_rx to note the quality of the transmission ^{(Image 3)}
• Found the limits of the b210 USRPs (the physical radios) at certain modulations in terms of the maximum sampling rate they could handle before underwriting (found the limits of the hardware).

*Began learning Bash to write my own scripts.

Week 4

• Wrote a bash program to simultaneously run the transmitter and receiver
• Authored another program to cycle through various modulations and sampling rates, parse the output for the final line with the results, and write that into a file
• Researched more about benchmark_tx and benchmark_rx to better understand the output
• Plotted the results in MATLab for each modulation, comparing the sampling rate of the transmission vs the number of successfully received packets of data ^{(Image 4)}
• Varied other factors such as tx gain and the packet size (a total of 1Mb of data was transmitted in a number of packets, with a smaller packet size resulting in a larger number of packets, which creates more granular data)
• Began connecting to a Spectrum Analyzer

Week 5

• Figured out how to see a signal on the spectrum analyzer by increasing the gain
• Researched and practiced effectively using the Spectrum Analyzer and how to save and move files onto a personal computer
• Conducted 49 tests to see the effects of varying gain and sampling rate on different modulations
• Began analyzing the spectrum analyzer's output in MATLab to find the total leakage in each waveform and plotting it vs the sampling rate for each modulation at various gains ^{(Image 5)} (defining the intended bandwidth to be within the 3dB bandwdith of the waveform from the peak of the signal to the peak of the noise floor)

Week 6

• Attempted to use an alternative method to determine what constitutes leakage vs intended bandwidth to be a function of the sampling rate at which the signal was sent
• Learned how to access the gnuradio companion (GRC) and began working on tutorials to send and recieve OFDM signals.
• Wrote a few smaller programs to test out various predefined blocks ^{(Image 6)}