Information fusion in underwater sonar simulation

This paper discusses information fusion methodologies, selection of one of these methodologies, and application of these fusion methodologies to underwater sonar simulation. Bayesian Inference and Dempster-Shafer are the two methods that have been studied in detail. In conclusion, the Dempster-Shafer approach was selected as the preferred method. Dempster-Shafer¿s main advantage is that it does not need conditional likelihoods. Also, Dempster-Shafer does not have computational complexity problems when multiple hypotheses and multiple conditional dependent events are examined. This method was applied to the multisensor information fusion problem in a simulation which includes a passive sonar, an active sonar, and a radar. The simulation is conducted on a geographical information system.