VISION-BASED DECISION SYSTEMS

A PIPELINE STRATEGY FOR MAN-MACHINE DECISION MODULES HAS BEEN DEVELOPED. IT HAS HIGHER OVERALL DECISION PERFORMANCE THAN THE SINGLE DECISION MODULE STRATEGY, AND IS MORE COMPUTATIONALLY EFFICIENT THAN THE MULTIPLE MODULE APPROACH. IT PROVIDES A FRAMEWORK FOR CONTROLLED CONSTRUCTION OF ORTHOGONAL DECISION MODULES, AND DIFFERS FROM THE MULTIPLE MODULE APPROACH BECAUSE EACH MODULE ONLY OPERATES ON DATA THAT REMAIN UNDECIDED IN THE PIPELINE. IT ENABLES A CLEAR APPROACH TO MAN-MACHINE DECISION CONFIGURATIONS IN WHICH THE HUMAN OPERATOR FOCUSES ON ONLY THE DIFFICULT PROBLEMS. BECAUSE THE MAHCINE MODULES FOUND WITHIN THIS CONFIGURATION PRODUCE NARROWED DECISION CHOICES, THOSE CHOICES CAN BE AUTOMATICALLY ROUTED TO A HUMAN SPECIALIST TRAINED TO RESOLVE THIS DECISION AMBIGUITY. UNDER THIS ORGANIZATION, THE HUMAN'S PERFORMANCE IS FURTHER EXTENDED BECAUSE SHE/HE SEES FEWER PROBLEMS, AND THOSE PROBLEMS CAN BE DISPLAYED MOST EFFICIENTLY SINCE THE DECISION TASK IS MORE FOCUSED.