Exact and approximate filtering in signal detection  An example
Abstract
The innovations approach in communication theory solves the signal detection problem in terms of a related nonlinear filter. Simulation data are presented which show that, at least for one example, using the appropriate linear filter instead results in a very small loss of detection power even though the variance of the linear filter is considerably greater than that of the nonlinear one. The results also illuminate the role of 'predictable processes' in communication theory.
 Publication:

IEEE Transactions on Information Theory
 Pub Date:
 November 1977
 Bibcode:
 1977ITIT...23..768D
 Keywords:

 Error Analysis;
 Linear Filters;
 Nonlinear Filters;
 Signal Detection;
 Signal Processing;
 Variance (Statistics);
 Approximation;
 Communication Theory;
 Computerized Simulation;
 Confidence Limits;
 Probability Theory;
 Communications and Radar