A comparison of self-organizing neural networks for fast clustering of radar pulses

摘要: Abstract Four self-organizing neural networks are compared for automatic deinterleaving of radar pulse streams in electronic warfare systems. The the Fuzzy Adaptive Resonance Theory, Min–Max Clustering, Integrated and Self-Organizing Feature Mapping. Given need a clustering procedure that offers both accurate results computational efficiency, these four examined from three perspectives – quality, convergence time, complexity. quality time measured via computer simulation, using set pulses collected field. Estimation worst-case running each network allows assessment effect pattern presentation order is analyzed by presenting data not just random order, but also radar-like orders called burst interleaved.