Stellar inertial attitude determination for LEO spacecraft

This paper describes the normal mode attitude determination algorithm for the low earth orbit (LEO) spacecraft using stellar inertial sensors. Included in the algorithm is a 6-state extended Kalman filter (EKF) which corrects the spacecraft attitude error as well as gyro bias error using celestial observations from star trackers. The detailed derivation of the EKF using a quaternion formulation is given in the paper. Time-domain simulations are presented to show the predicted filter performance which is a function of star density (or number of stars seen), star catalog error, star tracker noise, and gyro noises. Covariance analysis indicates that precision attitude with accuracy better than 12 arc seconds (3 sigma) can be met with the proposed attitude determination and control system hardware components.