Deconvolution and Compensation in the Measurement of Broadband Signals

This paper considers the recovery of broadband non-random signals distorted by the dynamics of the transmission system and corrupted by additive noise. The design and performance of deconvolution and compensation systems for a mean-square performance measure are discussed. After pointing out the deficiencies of deconvolution without compensation, the joint design of the two signal processing schemes under engineering constraints is presented and illustrated by some examples. Finally the results obtained when the signals are samples of stationary random processes with known statistics are briefly discussed.