Time-Frequency Processing of Wideband Radar Echo: from Fixed Resolution to Multiresolution and Superresolution

The electromagnetic energy backscattered from an unknown target can provide information useful for classifying and identifying the target This is commonly accomplished by interpreting the radar echo in either the time or the frequency domain. For example, the natural resonances of a target are manifested in the frequency domain as sharp, discrete events and can be attributed to the unique global features of the target. Similarly, scattering centers are manifested in the time domain as distinct time-pulses and can be related to the local features on the target For target characteristics which are not immediately apparent in either the time or the frequency domain, the joint time-frequency representation of the radar echo can sometimes provide more insight into echo interpretation. The usefulness of the time-frequency representation of signals has long been recognized in the signal processing arenas. In the electromagnetic scattering community, the joint time-frequency analysis was first introduced by Moghaddar and Walton 1.2 to explain the measurement data from an open-ended waveguide cavity. In this paper, we present our work 3–9 in the generation and interpretation of time-frequency representations of wideband backscattered signals from targets, as well as our efforts to improve resolution in the time-frequency plane using wavelets and superresolution techniques. Similar efforts along these lines have also been reported by other researchers recently10–12.