INTERACTING MULTIPLE MODELS ADAPTIVE ATTITUDE ESTIMATION FOR A STEREO-IMAGERY SATELLITE WITH UNKNOWN STAR SENSOR NOISE LEVEL

In this paper, we will design an interacting multiple models adaptive attitude estimator (IMMAE) for a stereo-imagery satellite in which the satellite requires to have a relatively large angle maneuvers around pitch and roll axes simultaneously. This method improves the accuracy of the attitude estimation by identifying the star sensor noise level changes at the end of maneuvers while imaging takes place. The structure of the proposed IMMAE method contains several extended Kalman filters with different star sensor noise levels that the probability of each level will be updated based on the model conditional likelihood functions. This algorithm will be designed once by using only kinematic model of the satellite which will be entitled 6-state IMMAE method and once by using kinematic and dynamic models of the satellite which will be entitled 12-state IMMAE method. In this method, high fidelity gyro and star sensor models are used to evaluate the design in realistic situation associated with gyro errors and star sensor noise level uncertainties. The effectiveness of the proposed IMMAE method will be examined and compared with previous proposed methods through numerical simulations.