Optimal and adaptive mismatch filtering for stretch processing

Abstract

Traditional stretch processing performs match filtering using a Fourier transform, which is essentially the matched filter bank for an LFM waveform mixed with an LFM reference. Here we consider how optimal least-squares based mismatched filtering could be used in place of the Fourier transform. This notion is then taken a step further with the formulation of an adaptive transformation. These new stretch processing filter structures are demonstrated in simulation and experimentally using open-air measurements, with the enhancements they enable applicable to a wide variety of legacy systems.