MAXIMUM ENTROPY ALGORITHM FOR TRACK FUSION

THE PROBLEM ADDRESSED IS THE FUSION OF MULTISENSOR TRACKS, INCLUDING ATTRIBUTES, OBTAINED IN AN AIRBORNE MULTITHREAT ENVIRONMENT ON A SINGLE PLATFORM. FUSION IS REQUIRED TO ENHANCE SENSOR SUITE CAPABILITY, SIMPLIFY CLUTTERED DISPLAYS AND REDUCE OPERATOR WORK LOAD. SENSOR TRACKS AND THEIR ATTRIBUTES ARE THE GIVEN INPUTS. THIS EFFORT COVERS THE PRELIMINARY DEVELOPMENT OF A TRACK FUSION ALGORITHM BASED ON THE MAXIMUM ENTROPY METHOD (MEM) FORMALISM. RELATIVE ENTROPY IS USED COMBINING BOTH A PRIORI INFORMATION FROM PREVIOUS TIME STEPS AND NEW SENSOR DATA AT THE CURRENT TIME STEP. THE ALGORITHM CONCEPT IS PRESENTED WITH EMPHASIS ON THE ASSOCIATION PROBLEM AND IS BRIEFLY COMPARED TO AN ALGORITHM BY BOWMAN AND GROSS (1985).