粒子滤波在卫星姿态确定中的应用

摘要

为了验证以星敏感器和速率陀螺作为卫星测量元件时，粒子滤波算法在姿态确定中的有效性．采用修正的罗德里格参数作为姿态参数建立了有陀螺和无陀螺两种模式下的系统状态方程和测量方程，并利用粒子滤波（PF）算法进行了姿态估计．和扩展的卡尔曼滤波（EKF）算法进行比较，仿真结果表明：PF算法在小初始估计误差下能够收敛，且具有和EKF相当的精度．大初始估计误差时，EKF算法不能收敛而PF算法仍能收敛．最后验证了PF算法在无陀螺模式下进行姿态确定的有效性．%In order to investigate the efficiency of attitude determination system using particle filter. The system state equation and measure equation of a satellite with star-sensors and gyros were established. And particle filter was used to estimate the attitude of satellite. The simulation results showed that the system can convergent not only for the small initial estimation error, but also for the big initial estimation error. But EKF can convergent only for the small initial estimation error. Finally, the effectiveness of PF applied to the gyro-free mode was also verified.