基于高斯混合容积卡尔曼滤波的UUV自主导航定位算法  

Gaussian mixture cubature Kalman filter based autonomous navigation and localization algorithm for UUV

作　　者：王宏健[1] 李村[1] 么洪飞[1] 周佳加[1] 

机构地区：[1]哈尔滨工程大学自动化学院,哈尔滨150001

出　　处：《仪器仪表学报》2015年第2期254-261,共8页Chinese Journal of Scientific Instrument

基　　金：国家自然科学基金(E091002/50979017;E091002/51309067);教育部高等学校博士学科点专项科研基金(20092304110008);哈尔滨市科技创新人才(优秀学科带头人)研究专项资金(2012RFXXG083)资助项目

摘　　要：针对过程噪声为非理想高斯分布时无人水下航行器(UUV)自主导航定位存在噪声模型失配的问题,将高斯混合密度模型与容积卡尔曼滤波(CKF)相结合,设计了基于高斯混合容积卡尔曼滤波(GM-CKF)的UUV导航定位算法。建立了UUV运动模型及观测模型,利用CKF完成各高斯分量的预测更新,并将更新结果进行融合缩减与加权求和,从而实现UUV自主导航定位。通过与EKF、UKF和CKF算法仿真对比实验,验证了GM-CKF可以提高估计精度;通过UUV湖试试验,验证了基于GM-CKF的UUV自主导航定位精度和稳定性优于传统算法,其计算时间满足实时导航定位的要求。Aiming at the problem of mismatched noise model of autonomous navigation of unmanned underwater vehicle（ UUV） with non-ideal Gaussian distribution noise,the Gaussian mixture cubature Kalman filter（ GM-CKF） based navigation algorithm of UUV is designed through combining the Gaussian mixture density distribution model with CKF.The motion model and observation model of UUV are established;the Gaussian components are predicted and updated with CKF;the results are merged and weighted,and the autonomous navigation and localization of UUV is realized.The simulation comparison experiments with EKF,UKF and CKF algorithms were conducted,which prove that the GM-CKF algorithm could improve the estimation precision.The UUV lake trial experiment was also conducted,and the result indicates that the proposed GM-CKF algorithm can provide better accuracy and stability than conventional navigation algorithms,and the computation time satisfies the requirement of real time navigation and localization of UUV.

关 键 词：无人水下航行器 导航定位 高斯混合密度模型 容积卡尔曼滤波 

分 类 号：TH766[机械工程—仪器科学与技术]