Investigations of Adaptive Detection Techniques

摘要

This report is concerned with the determination of the characteristics of adaptive signal detection and waveform estimation systems. The signals of interest are pulse like, recur sporadically and are masked by noise. The objective of the adaptive system is to detect and describe these signals without having detailed knowledge of their waveshapes, strength, signal-to-noise ratio or statistics of occurrence. A system is developed which determines the coefficients of an orthonormal expansion of its input waveform, decides at what instants a signal is most likely to be present and builds up a description of the signal waveform. The best orthonormal expansion to use in such a system is determined by the statistics of the set of signals with which the system is to operate. An adaptive system, using a Laguerre expansion was analyzed in more detail. Such a system was constructed and evaluated using analogue computer and digital logic techniques. Results of the simulation show that the system operates in accordance with theory and should be useful in defining the capabilities and limitations of such systems. The performance of this system is illustrated by means of plots and graphs. (Author)