STAR PATTERN RECOGNITION FOR REAL TIME ATTITUDE DETERMINATION.
- Additional Document Info
- View All
In this paper, an attitude determination approach for autonomous, on-board, near-real-time determination of spacecraft attitude with sub-ten-arc-second precision is discussed. The method employs a systematic pattern recognition procedure for identifying stars sensed by, for example, charged coupled device (CCD) electrooptical star sensors. An extended Kalman filter is used to predict spacecraft attitude at each data gathering epoch. A parallel processing division of the computations and logic associated with data acquisition/editing, pattern recognition/attitude determination and optimal prediction functions is proposed. Three intermittently communicating parallel processes are proposed which appear to optimize the rate and precision of attitude estimation, subject to the contraints of on-board computation. Numerical experiments are summarized which support the validity and practicality of the proposed star pattern recognition/attitude estimation strategy.
author list (cited authors)
Junkins, J. L., White, C. C., & Turner, J. D.