Symbolic analysis of time-varying, dynamic phenomena

The authors present a method of analyzing dynamic waveforms using symbolic techniques. They successfully analyze waveforms with multiple degrees of randomness, integrate contextual knowledge into the analysis and interpretation process, and achieve a level of performance that can far exceed a human's. They tested the method on numerous waveforms generated by a waveform simulator. With few exceptions, the system found every moving pulse and tracked it through all its transitions, even when it became completely obscured by or embedded in other pulses. The rare anomalies occurred when more than one pulse became embedded in the same group of stopped pulses.<<ETX>>