针对卫星在轨道机动时平均轨道根数的高精度实时估计问题,提出了一种采用UKF滤波算法在轨实时估计平均轨道根数的方法.根据轨道根数形式的摄动运动方程分别推导了在轨道机动控制力、地球非球形摄动和大气阻力摄动作用下平均轨道根数的变化率,由此建立了状态方程和量测方程并对测量噪声的特性进行了蒙特卡洛法分析;提出了基于球形单边采样的平方根UKF滤波的平均轨道根数估计算法,并给出了相应的计算流程.仿真结果表明:所提算法在卫星进行轨道机动时有效,偏心率和轨道倾角等轨道参数滤波误差分别提高到4.0×10-6和1.2×10-5,并具有较好鲁棒性和在轨实时性.
An onboard real-time estimation algorithm of the mean orbital elements is presented for satellite missions with maneuvers, in which the unscented Kalman filter is employed. The motion of the satellite was analyzed using classical orbital elements and the rates of change of the mean orbital elements were formulated by the effects of the control forces for maneuvers, geopotential of zonal harmonics and atmospheric drag, respectively. The state equa- tions and measurement equations were established using mean orbital elements as state variables and osculating or- bital elements as measurements. The distribution of measurement noise was analyzed by using the Monte Carlo meth- od. The filter procedures of the spherical simplex square root form of UKF were summarized. The simulation results prove that the proposed algorithm is effective, with good robustness and onboard timeliness, and has improved the eccentricity and orbit inclination angle error to 4.0×10-6 and 1.2×10-5 respectively.