Methods of linear system characterization through response cataloging.

Various methods of linear system characterization are presented. Each method requires cataloging the system response to a set of input stimuli. The continuum orthonormal basis set response method, in principle applicable to all linear systems, has as special cases point-spread function and frequency response superposition integral characterizations. The piecewise isoplanatic approximation, sampling theorem approach, and discrete basis set response methods are three schemes whereby restrictive assumptions are placed on the system and/or allowable input class. In each case, the number of system responses required for system definition is reduced to a countable number. Contrasting properties and applicability of the three schemes are discussed. Last, shift-invariant and spreadless systems are briefly reviewed in the context of requiring only one input-output relation for complete system characterization.