Digital predistortion of wideband radar waveforms

As demands for increased radio frequency system performance have grown over time, sufficiently mitigating the distortion introduced by high power amplifiers has proven to be a major challenge to overcome. Digital predistortion is one technique commonly used to reduce the distortion products seen at the output of amplifiers, to great effect. While some of the most widely used predistortion algorithms to date may seem most applicable to waveforms with properties most commonly seen in wireless communication systems, it will be seen that with some specialization, these methods can be quite applicable to linear frequency-modulated pulse radar waveforms. In this thesis, a frequency-domain approach to parameter estimation is presented, and a simple window-based approach to predistortion of wideband linear frequency-modulated pulses is also explored. To test the performance of these predistortion algorithms, a modular radar signal processing FPGA testbench architecture is discussed. Finally, a software defined radio was used to measure the performance of the algorithms discussed in this thesis, and a significant improvement in the peak range sidelobe level was observed. These results enable the integration of digital predistortion into modern radar systems with minimal computational overhead to correct radar pulse envelope distortion