A complex mapping network for phase sensitive classification

The design of a network that incorporates the complex relationships present in the structure and learning algorithm, thereby enforcing the formation of a complex mapping of the problem space, is detailed. The network is applied to two phase-sensitive problems: interpretation of chaotic oscillator phase plane plots, and eddy current defect detection and characterization. In chaotic oscillator analysis, the network, in conjunction with the oscillator, demonstrates the ability to interpret small signal behavior. In eddy current impedance plane analysis, the network demonstrates a clear performance advantage over both real-valued multilayer feedforward networks (MFFNs) and human subjects, with overall classification accuracy improvements of 45% (to a 99% level) and 48%, respectively. This network structure and learning algorithm should provide similar results in other signal processing applications where time or phase considerations are critical for class discrimination.