The Effects of Geometric Scattered Signal Waveform Modeling on Target Identification Performance


SELVER M. A., SEÇMEN M., ZORAL E. Y.

11th European Conference on Antennas and Propagation (EUCAP), Paris, France, 19 - 24 March 2017, pp.1087-1090 identifier identifier

  • Publication Type: Conference Paper / Full Text
  • Volume:
  • Doi Number: 10.23919/eucap.2017.7928831
  • City: Paris
  • Country: France
  • Page Numbers: pp.1087-1090
  • Keywords: scattered signals, resonance region, target identification, geometric-modeling, feature extraction, CLASSIFICATION
  • Dokuz Eylül University Affiliated: Yes

Abstract

Target identification from scattered signals using time domain techniques depend significantly on the waveform. Recently, a novel feature set is proposed which encounter structural properties of the waveform and collects local extrema points to model the scattered signal via triangularization. Then, using this piecewise model, it extracts several morphological features and employs them for target identification through classification. This study expands that approach by modeling the scattered signal with other geometric shapes and accordingly, by enriching the feature set. Such an approach requires careful representation of the waveform model since more than one morphology is considered to represent sub-waves of the waveform. The effects of the proposed approach are observed by applications on spherical targets having different size and material type.