Cumulative probability distributions such as occur in radar detection problems are approximated by a new version of the saddlepoint method of evaluating the inverse Laplace transform of the moment-generating function. When the number of radar pulses integrated is large, the approximation of lowest order yields good accuracy in the tails of the distributions, yet requires much less computation than standard recursive methods. Greater accuracy can be achieved upon summing the residual series by converting it to a continued fraction. The method is applied to evaluating the error-function integral and the Mth-order Q function, and to approximating the inverse of the chi-squared distribution. Cumulative distributions of discrete random variables, needed for determining error probabilities in optical communication receivers that involve counting photoelectrons, can be approximated by a simple modification of the method, which is here applied to the Laguerre distribution.
Approximate Evaluation of Detection Probabilities in Radar and Optical Communications
IEEE Transactions on Aerospace and Electronic Systems ; AES-14 , 4 ; 630-640
1978-07-01
2236774 byte
Aufsatz (Zeitschrift)
Elektronische Ressource
Englisch
Direct Evaluation of Radar Detection Probabilities
IEEE | 1987
|Radar Detection Probabilities and Their Calculation
Online Contents | 1995
|Radar detection probabilities and their calculation
IEEE | 1995
|