IDENTIFICATION TECHNIQUES FOR LARGE SPACE SYSTEMS

THIS PROPOSAL SEEKS TO DEVELOP IDENTIFICATION ALGORITHMS FOR FLEXIBLE STRUCTURES IN SPACE WHERE PERFORMANCE SPECIFICATIONS ARE IN TERMS OF RMS VALUES OF DEFLECTIONS AT MULTIPLE POINTS ON THE STRUCTURE (DERIVED FROM SHAPE OR POINTING CONTROL REQUIREMENTS). HENCE, COVARIANCE CONTROL OBJECTIVES ARE OF INTEREST IN SUCH PROBLEMS. THIS IDENTIFICATION PROPOSAL WILL DEVELOP ALGORITHMS TO PRODUCE MODELS OF ANY SPECIFIED ORDER R(Q) FROM EXPERIMENTAL DATA TAKEN FROM THE STRUCTURE. THE MODEL WILL MATCH THE FIRST Q DERIVATIVES OF BOTH THE IMPULSE RESPONSE MATRIX AND THE AUTOCORRELATION MATRIX IN THE VICINITY OF T=O, WHERE R(Q)<N(Y)Q, AND N(Y) IS THE NUMBER OF SENSORS FROM WHICH THE EXPERIMENTAL DATA IS TAKEN. THE IDENTIFICATION ALGORITHM WILL BE COMPATIBLE WITH CONTROL OBJECTIVES BECAUSE OF ITS CAPABILITY TO MATCH THE CONVARIANCE MATRIX THUS MAKING THESE DEPENDENT PROBLEMS (MODELING AND CONTROL) COMPATIBLE. THIS IS A REQUIRED RESULT BEFORE COMMERCIAL APPLICATIONS OF FLEXIBLE STRUCTURE CONTROL IN SPACE ARE FEASIBLE.