Multipath exploitation for radar target classification

Inclusion of multipath ghost target images, alongside direct target images, is shown to improve performance in radar target classification. When ultra wideband waveforms are employed there is sufficient range resolution that the components of the received signal relating to indirect paths appear as unique peaks. The additional returns result in a point spread function with peaks at locations other than that of the target - the ghosts. A method for localizing the ghosts is provided, and is then used to include the ghosts in target classification testing. A gain of up to 17% in classification performance is shown when the ghost targets are included. This study makes use of real radar data.