APPLICATION OF NEURAL NETWORKS TO COMMAND AND CONTROL

EVOLVING NEURAL NETWORK TECHNOLOGY HAS THE POTENTIAL TO PROVIDE THE RAPID INFORMATION REDUCTION AND ASSESSMENT THAT WILL BE REQUIRED FOR NEW GENERATION COMMAND AND CONTROL SYSTEMS. IN COMPARISON WITH OTHER TYPES OF CLASSIFICATION AND DECISION ALGORITHMS, NEURAL NETWORKS ARE ROBUST AND DEPEND ONLY WEAKLY ON ASSUMPTIONS MADE ABOUT INFORMATION OR NOISE DISTRIBUTIONS. AS A RESULT, THERE IS LESS DIFFICULTY ADAPTING NEURAL NETWORK TECHNOLOGY TO A VARIETY OF INFORMATION SOURCES. IN ADDITION, THIS TYPE OF PROCESSING PROVIDES THE POTENTIAL FOR ADAPTABILITY IN THE FACE OF NEW OR CHANGING THREAT SIGNATURES. RECENT RESULTS IN NEURAL NETWORK RESEARCH HAVE SHOWN THE DESIRABILITY OF FUSING DATA FROM SUCH DIFFERENT SENSORS AS ACOUSTIC AND VISUAL INPUT TO PERFORM SPEECH RECOGNITION. THE ABILITY TO ENCODE FUZZY RELATIONS AND DECISION RULES IN NEURAL NETWORKS CAN OVERCOME THE SPEED LIMITATIONS OF SERIAL DECISION-MAKING AIDS SUCH AS RULE-BASED EXPERT SYSTEMS. THE PRIMARY OBJECTIVE OF THE PHASE I EFFORT IS TO CONVERT THE STATE-OF-THE-ART WORK THAT HAS BEEN COMPLETED IN NEURAL NETWORKS TO SOFTWARE, WHICH CAN BE USED FOR INFORMATION PROCESSING AND DECISION MAKING IN A COMMAND AND CONTROL OPERATING ENVIRONMENT, SUCH AS THE PLANNING AND EXECUTION OF AN AIRBORNE ASSAULT. IN ADDITION, ORINCON WILL PROVIDE AN APPROACH FOR DEVELOPING AND TESTING A PROTOTYPE APPLICATION IN THE FIELD.