ANRO is currently under a Phase II SBIR contract to evaluate an Ultra-Wideband (UWB) linear radar array and the processing of co- and cross-polarization returns for detecting and discriminating buried mines. Experiments indicate that both metallic and plastic ines located several inches below the suface are reliably detected using a novel real-time cancellation scheme; however, it appears that target discrimination is not solved, practically, using co- and cross-polarization discrimination alone. The proposed approach will consider information about target size, shape and material that is inherent in the resonant region of co- and cross- polarized backscattered fields independently, as well as both together. To accomplish this, it is shown that reducing the nominal center frequency of the present array from 2.5 GHz to 1 GHz provides reflected signals rich in mine discrimination information. ANRO proposes to sequentially sample the received waveforms such that the microwave responses with bandwidths exceeding 500 MHz are reduced to low frequency video data. These returns can be processed, digitally, in near real-time. Since the istantaneous amplitudes and phases of backscattered pulses will be different for objects with different sizes, shapes and material compositions, a proven neural net approach is proposed to provide the required discrimination of targets.
Keywords: Uwb Reflectometer Buried Mine Detection Discrimination Neural Networks
