FLIGHT TEST DATA FILTERING AND SMOOTHING

REDUCTION OF AIRCRAFT TRACKING DATA FOR PRECISE TRAJECTORY DETERMINATION REQUIRES SMOOTH DATA. IF THE TRUE TRAJECTORY HAS ABRUPT CHANGES IN IT, CONVENTIONAL POLYNOMIAL LEAST SQUARES TECHNIQUES WILL BE INADEQUATE TO MEET PRECISION REQUIREMENTS. FURTHERMORE, MANY TEST MISSIONS COULD SIGNIFICANTLY BENEFIT FROM REAL-TIME SOLUTIONS INSTEAD OF CURRENT BATCH PROCESSING. THIS PROPOSAL PRESENTS AN ADAPTIVE OPTIMAL ESTIMATION TECHNIQUE THAT CAN MEET THESE REQUIREMENTS. CONCEPTS TO RELIABLY HANDLE THE RESPONSE/STABILITY TRADEOFF IN AN ADAPTIVE REAL-TIME FILTER HAVE BEEN FORMULATED. THE APPROACH INCLUDES ADAPTATION OF BOTH PROCESS AND MEASUREMENT NOISE, AND USE OF NON-LINEAR STATE PROPAGATION MODELS. EXTENDED STATE ESTIMATION IS DISCUSSED FOR PARTICULAR APPLICATIONS REQUIRING OTHER INFORMATION OR FOR CASES WHERE TRACKING SENSORS POSSESS SIGNIFICANT, DYNAMIC ERRORS. POST-FLIGHT OPTIMAL SMOOTHING APPLICATIONS ARE DISCUSSED.