粒子滤波在光纤陀螺四位置寻北中的应用研究  

Study on the application of particle filtering in FOG four-position north-seeking

作　　者：段苛苛[1] 李邓化[2] 

机构地区：[1]北京交通大学电子信息工程学院,北京100044 [2]北京信息科技大学自动化学院,北京100101

出　　处：《仪器仪表学报》2013年第8期1749-1755,共7页Chinese Journal of Scientific Instrument

基　　金：北京市自然科学基金项目(4122028);北京市高校人才强教计划(PHR201107218);北京市研究生联合培养基地项目(71C1311016)资助

摘　　要：为了减小噪声对光纤陀螺(FOG)寻北的影响,提高寻北精度,提出了将粒子滤波这种非线性滤波方法应用于光纤陀螺四位置寻北的方案。根据四位置寻北模型建立光纤陀螺寻北系统的非线性状态空间模型,将基于系统重采样和在重采样后引入马尔可夫链蒙特卡罗(Markov chain monte carlo,MCMC)移动的粒子滤波器分别应用于光纤陀螺寻北系统的非线性滤波。使用一个零偏稳定性为0.05°/h的闭环光纤陀螺进行实验,实验结果表明,这2种粒子滤波器均能够有效地提高光纤陀螺的寻北精度。基于系统重采样的粒子滤波器有随时间发散的趋势,这是由于在选定粒子数目较少的情况下,重采样导致粒子多样性丧失的结果;而在重采样后引入MCMC移动的粒子滤波器有很好的收敛性能。重复实验结果表明,使用基于MCMC移动的粒子滤波器可以达到更好的寻北性能。To reduce the impact of noises on fiber optic gyroscope north seeking and improve the north seeking accuracy, a processing scheme was proposed, which uses a nonlinear filtering method, i. e. particle filtering methoed in FOG four position north seeking. A FOG north seeking system model based on the four-position scheme was investiga- ted and the nonlinear state space model was established. The particle filters based on systematic resampling and based on MCMC moving algorithm were applied to the nonlinear filtering of the FOG north-seeking system, respectively. A closed loop FOG with the bias stability of 0.05 ~/h was used to perform experiment operation. The experiment results show that both the particle filters can improve the north seeking accuracy and north seeking speed effectively. The systematic resampling based particle filter has a divergence trend, which is caused by the loss of particle diversity under small number of particles;while the MCMC based particle filter has good convergence performance. The repeat experiment results prove that better north-seeking accuracy can be obtained using the MCMC based particle filter.

关 键 词：光纤陀螺 寻北 四位置方案 粒子滤波 重采样 马尔可夫链蒙特卡罗移动 

分 类 号：V249.3[航空宇航科学与技术—飞行器设计]