Use of importance sampling to investigate signal detection systems by simulated method
Abstract
The use of importance sampling for investigating signal detection systems by simulation employing only a posteriori information about the response of the system to a given input is examined. In the importance sampling method, the initial input is replaced in the simulation by a modified input, after which the system response is corrected by the appropriate weight coefficient. The input is modified such that the detection probability for a given accuracy is estimated according to the smallest possible number of experiments. The case is considered in which the set of modified inputs includes the probability density of which differs from that of the original input only in the values of the parameters f sub mod (x alpha) = f(x alpha sub mod). The number of experiments required to achieve the required accuracy in estimating the probability in direct modeling and in modeling employing modified inputs is calculated. It is found that the importance sampling method makes it possible to estimate small false alarm probabilities with an acceptable number of trials.
 Publication:

USSR Rept Electron Elec Eng JPRS UEE
 Pub Date:
 January 1985
 Bibcode:
 1985RpEEE.......41K
 Keywords:

 Digital Systems;
 Dynamic Response;
 Radio Signals;
 Sampling;
 Signal Detection;
 Estimating;
 Probability Density Functions;
 Probability Theory;
 Communications and Radar