一种适用于星敏感器的大角速度下星跟踪算法

Star tracking algorithm under large angle velocity for star sensor

传统的星敏感器预报方法是利用当前四元数和陀螺输出角速度来预测下帧姿态四元数,再通过预测的四元数计算视场内恒星的理想星像坐标,最后在这些理想坐标为中心的范围内来获取实际的星像坐标。但是,由于陀螺漂移的存在,随着时间的增加,预测的理想坐标与实际坐标偏差越来越大,获取实际星像坐标扫描的区域越来越大,星敏感器的数据更新率也随之降低。为此,本文提出了一种适用于星敏感器的大角速度下星跟踪算法。首先根据连续两帧的姿态偏差计算星敏感器的角速度,利用角速度来计算星敏感器当前时刻与最近上一帧的姿态偏差。然后以该姿态偏差来预报当前时刻的星敏感器姿态输出,这样不仅提高了数据更新率,而且满足飞行器在大角速度机动时单纯采用星敏感器进行姿态测控的要求。最后,采用2010年外场试验实际观星数据对该算法进行了仿真验证。

The conventional predictive method of star sensor uses the current quatemion and gyro angular velocity output to predict the attitude quaternion of the next frame, and the ideal locations of stars are calculated with predictive quaternion. The true locations of stars are obtained within the threshold scan windows of predicted locations of stars. The error between ideal locations of stars and true locations of stars would become larger due to the gyro drift, resulting in the decreased update rate of star sensor. So this paper puts forward a star tracking algorithm for star sensor with large angle velocity. In this algorithm, the angle velocity is calculated with continuous quaternion. Then the ideal quaternion of the next fi＇ame is calculated with current quaternion and angle calculated velocity. The ideal locations of stars in the star image of next flame ate calculated with the calculated quaternion. The true locations of stars in the star image of next frame are obtained within the threshold scan windows of ideal locations of stars in the next frame. At last, these algorithms are verified by simulation using night sky experiment data in 2010.