一种基于NGA-QPF的航天器姿态估计方法

针对航天器姿态确定中的非线性非高斯的滤波问题，提出一种基于遗传算法的粒子滤波的航天器姿态估计方法。该方法将姿态四元数作为采样粒子进行粒子滤波，并将小生境遗传算法（NGA）引入粒子滤波算法中，以改善粒子滤波的性能；用遗传算法单独进行陀螺偏差估计，以减少粒子滤波的状态维数。该姿态估计方法保持了四元数的归一化性质，通过引入小生境遗传算法解决了重采样阶段的粒子退化问题，并且由于单独估计陀螺偏差避免了粒子滤波状态的扩展。该方法能够在较少粒子的情况下实现高效率高精度的定姿，仿真结果说明了方法的有效性。

A Spacecraft Attitude Estimation Method Based on NGA-QPF

A spacecraft attitude estimation method of particle filters based on the genetic algorithm is proposed to solve nonlinear non-Gaussian filtering problems in attitude determination. In this method, the attitude quaternion is used as sampling particles for the particle filter and the niche genetic algorithm (NGA) is introduced into the particle filter algorithm in order to improve its performance. At the same time, gyro bias is estimated by the genetic algorithm separately to reduce the state dimension of the particle filter. This method not only maintains the properties of the normalized quaternion but also solves the particle degradation problem in the resampling stage by introducing the NGA. And as gyro bias is estimated separately, the expansion of the state dimension is avoided. The method can achieve attitude determination with high efficiency and precision for the case of relatively few particles. Simulation results show the validity of the method.