随着我国月球探测工程的开展,为弥补地面测控的局限性,月球卫星的自主导航技术已成为一项亟待解决的关键技术。本文提出了一种基于天文和陆标观测的月球卫星自主导航新方法,该方法利用星上姿态敏感器和有效载荷相机获得天体观测信息和陆标观测信息,通过非线性Unscented滤波方法进行信息融合,实现了二者的优势互补,在提高精度的同时也提高了可靠性。同时针对传统导航姿态估计方法中欧拉角存在奇点,四元数及其误差滤波更新后不能同时满足约束条件的问题,采用了先进的四元数-广义Rodrigues参数姿态估计方法,该方法用四元数描述姿态,用广义Rodrigues参数描述姿态误差,自适应跟踪误差变化,尤其适用于初始姿态误差大且误差特性不确定的姿态估计问题。仿真结果表明该方法可提供包括位置、速度和姿态在内的全部导航参数,且具有较高的估计精度,是一种非常适于月球卫星的高精度自主导航方法。
With the development of our lunar exploration project,the capability of autonomous navigation is in great demand for lunar satellites for limitations of radiometric tracking. A new autonomous navigation method based on celestial and landmark measurements is proposed in this paper. An advanced unscented Kalman filter and corresponding information fusion technique are used to get a better navigation performance and higher reliability. Furthermore,because updated quaternion hardly obeys a unit norm constraint,and Euler angles always exist singular,an advanced Quaternion-generalized Rodrigues attitude estimation approach is proposed. It guarantees both quaternion and quaternion-error within the norm constraint,and provides the adaptive ability to compensate for system errors. Results verify that all navigation information,including position,velocity and attitude are given with a high accuracy. This method is a promising and attractive scheme for autonomous navigation of lunar satellites.