A comparison of three non-linear filters
Automatica (Journal of IFAC)
Evaluation of likelihood functions for Gaussian signals
IEEE Transactions on Information Theory
A general likelihood-ratio formula for random signals in Gaussian noise
IEEE Transactions on Information Theory
| Hi-index | 0.07 |
A pseudo Bayes approach to likelihood ratio determination for discrete time random processes imbedded in Gaussian noise is presented. The resulting computationally feasible algorithm for the likelihood ratio is expressed as a function of the one-step prediction conditional mean estimate of the message, suggesting an estimator-correlator type structure for the optimum digital detector. The use of smoothed or iterated conditional mean estimates in the pseudo Bayes detection scheme is also considered. The likelihood ratio formulas for continuous random signals in Gaussian noise are derived from the discrete results as a limiting case. Examples indicate the efficacy of the method.