This paper presents a computationally efficient algorithm for attitude estimation of remote a sensing satellite. In this
study, gyro, magnetometer, sun sensor and star tracker are used in Extended Kalman Filter (EKF) structure for the
purpose of Attitude Determination (AD). However, utilizing all of the measurement data simultaneously in EKF
structure increases computational burden. Specifically, assuming n observation vectors, an inverse of a 3n×3n
matrix is required for gain calculation. In order to solve this problem, an efficient version of EKF, namely Murrell’s
version, is employed. This method utilizes measurements separately at each sampling time for gain computation.
Therefore, an inverse of a 3n×3n matrix is replaced by an inverse of a 3×3 matrix for each measurement vector.
Moreover, gyro drifts during the time can reduce the pointing accuracy. Therefore, a calibration algorithm is utilized for
estimation of the main gyro parameters.