Applying a Modified Discrimination Model to Enhance Defense and Sensor Systems Security.

摘要

National defense security systems constantly face cyber attack, espionage, and hacking. The primary objective of this paper is to explore the application of a modified discrimination model to replace current encrypted user id and password authentication. This discrimination model will be based on minimum distance as compared to the traditional discrimination model which is based on maximum probability. The modified discrimination model treats user id and password as a multisensor information fusion technology problem. The model converts the input user id and password into a digital pattern feature vector. The model then processes the newly converted vector for distance between all known feature vectors stored in the secure knowledge database. The new pattern feature vector with the minimum distance generated by the modified discrimination model will be the authorized person. The new application is demonstrated using mathematical simulation, and is then verified by comparing its performance with the traditional multisensor correlation model.