针对粒子滤波的粒子退化和匮乏等问题,在粒子滤波和网格近似的基础上,提出了一种采用分层近似策略的粒子滤波改进算法。改进算法利用高斯分布对后验概率密度进行近似,并在连续分布的后验概率密度上进行分层近似,从而获得更具效率的粒子,提高了粒子滤波的精度。光纤陀螺的寻北精度主要取决于光纤陀螺自身的性能以及所采用的寻北方案,将采用分层近似策略的改进粒子滤波算法应用于光纤陀螺的寻北方案中,能够有效地解决光纤陀螺寻北中的非线性状态估计问题,提高光纤陀螺的寻北精度。
Aiming at the problems of particle degeneracy and deficiency,an improved particle filtering algorithm with stratified approximation strategy is put forward based on particle filtering and grid approximation.In the improved algorithm,posterior probability density is approximated with Gaussian distribution.A stratified approximation is conducted on the posterior probability density with continuous distribution.Hence,particles with higher efficiency are obtained and the precision of particle filtering is improved.The north-seeking precision of fiber optic gyroscope is mainly determined by the performance of the gyroscope itself and the adopted north-seeking scheme.The application of this proposed particle filter algorithm with stratified approximation in north-seeking scheme of fiber optic gyroscope can effectively solve its nonlinear state estimation problem and improve north-seeking precision.