Radar imaging using linear prediction for data extrapolation

The authors describe an investigation wherein the resolution capability of range-Doppler imaging radars is extended well beyond the limitation imposed by the conventional FFT (fast Fourier transform) range-Doppler processing. The approach adopted uses linear prediction for data extrapolation before performing Fourier reconstruction. Preliminary results of rotating platform imaging based on experimental data acquired in a microwave anechoic chamber show that this technique shows promise for improving the resolution capability of range-Doppler imaging radars. With this approach, one can obtain either higher-resolution images for the same effective bandwidth of transmitted signals and total rotation angle in imaging, or equivalent-quality images from a smaller bandwidth and rotation angle.<>