A NEURAL NETWORK FOR AEJ SITUATION ASSESSMENT AND RESOURCE MANAGEMENT

THE INABILITY OF THE CURRENT STATE-OF-THE-ART IN COMPUTING TO SUPPORT CONVENTIONAL ALGORITHMIC APPROACHES TO THE SITUATIONAL ASSESSMENT AND RESOURCE ALLOCATION PROBLEMS FACING THE APACHE ESCORT JAMMER NECESSITATES THE INTRODUCTION OF NOVEL SOLUTIONS, SUCH A NEURAL NETWORK TECHNIQUE, TO THIS PROBLEM. IN PHASE I, WE PROPOSE THE INVESTIGATION OF A DUAL NEURAL NETWORK ARCHITECTURE TO PERFORM THE TASKS OF SITUATIONAL ASSESSMENT AND RESOURCE MANAGEMENT: THE PROPOSED SITUATIONAL ASSESSMENT NEURAL NETWORK IS BASED UPON THE CONCEPT OF A FUZZY COGNITIVE MAP (FCM). BEING FEEDBACK GENERALIZATIONS OF SEARCH TREES, FCMS OVERCOME THE PROCESSING SPEED PROBLEMS OF CONVENTIONAL AI OR EXPERT SYSTEMS APPROACHES AND HAVE THE ADDITIONAL ADVANTAGE OF BEING ABLE TO NATURALLY COMBINE EXPERT KNOWLEDGE FROM MULTIPLE SOURCES. THE NETWORK IS DESIGNED TO INTERFACE DIRECTLY WITH THE AVIONICS AND EW EQUIPMENT AND PROVIDES INPUTS TO A NEURAL NETWORK CONTROLLER. THE NEURAL NETWORK CONTROLLER UTILIZES AN EXTENSION OF PROVEN CONCEPTS FROM OPTIMAL CONTROL THEORY. THE NETWORK MAKES USE OF INHIBITORY FEEDBACK WHICH PROVIDES COMPETITION AND COOPERATION AMONG THE AVIONICS AND EW RESOURCES IN DETERMINING THE OPTIMAL RESPONSE STRATEGY. THE NETWORK OUTPUTS PROVIDE FOR THE CONTROL OF BOTH TIME-RENEWABLE (E.G., POWER MANAGEMENT) AND EXPENDABLE (E.G., CHAFF/FLARE) ASSETS.