Temporal and Spatial Interference Mitigation Strategies to Improve Radar Data Quality

The microwave band is well suited to wireless applications, including radar, communications, and electronic warfare. While radar operations currently have priority in a portion of the microwave band, wireless companies are lobbying to change that; such a change would force current operators into a smaller total bandwidth. Interference would occur, and has already occurred at the former National Weather Radar Testbed Phased Array Radar. The research in this dissertation was motivated by this interference --- it occurred even without a change to radar's primacy in the microwave band. If microwave operations had to squeeze into a smaller overall bandwidth, such interference, whether originating from other radars or some other source, would only become more common. The radio frequency interference (RFI) present at the National Weather Radar Testbed Phased Array Radar altered the statistical properties at certain locations, causing targets to be erroneously detected. While harmless enough in clear air, it could affect National Weather Service decisions if it occurred during a weather event. The initial experiments, covered in Chapter 2, used data comprised of a single channel of in-phase and quadrature (IQ) data, reflecting the resources available to the National Weather Service's weather radar surveillance network. A new algorithm, the Interference Spike Detection Algorithm, was developed with these restrictions in mind. This new algorithm outperforms several interference detection algorithms developed by industry. Tests on this data examined algorithm performance quantitatively, using real and simulated weather data and radio frequency interference. Additionally, machine learning classification algorithms were employed for the first time to the RFI classification problem and it was found that, given enough resources, machine learning had the potential to perform even better than the other temporal algorithms. Subsequent experiments, covered in Chapter 3, used spatial data from phased arrays and looked at methods of interference mitigation that leveraged this spatial data. Specifically, adaptive beamforming techniques could be used to mitigate interference and improve data quality. A variety of adaptive digital beamforming techniques were evaluated in terms of their performance at interference mitigation for a communications task. Additionally, weather radar data contaminated with ground clutter was collected from the sidelobe canceller channels of the former National Weather Radar Testbed Phased Array Radar and, using the reasoning that ground clutter is simply interference from the ground, adaptive digital beamforming was successfully employed to mitigate the impact of ground clutter and restore the data to reflect the statistics of the underlying weather data. Tests on digital equalization, covered in Chapter 4, used data from a prototype receiver for Horus, a digital phased array radar under development at the University of Oklahoma. The data suffered from significant channel mismatch, which can severely negatively impact the performance of phased arrays. Equalization, implemented both via older digital filter design methods and, for the first time, via newer machine learning regression methods, was able to improve channel matching. When used before adaptive digital beamforming, it was found that digital equalization always improved system performance