摘要
传统的小干扰失准角模型只适合于小失准角情况下的初始对准,对于处于大失准角下的舰船或飞机的对准必须寻求不做任何线性假设的非线性模型和非线性滤波方法。针对以上问题,建立了基于四元数的姿态误差方程,给出了基于复杂噪声模型的UKF算法,在该算法的基础上假设量测方程为线性,得出简化的UKF算法,避免了重采样、多次求解量测预测方程、计算量测预测方差等一系列繁杂过程。基于以上理论建立了适合简化UKF算法的非线性滤波模型,在大失准角、小失准角下与常规Kalman和EKF算法做对比仿真,结果表明,在小失准角下三种方法效果相当,但在大失准角下简化UKF和EKF显示出了处理非线性模型的优势,对准速度和精度都好于常规Kalman算法。由于EKF线性化造成的高阶截断误差使得对准精度略低于简化UKF。
Conventional small perturbation is only effective to initial alignment of small misalignment angle system rather than large misalignment angle system,therefore nonlinear model and nonlinear filter are introduced to solve the problem.Equations of attitude error based on quaternion and the UKF algorithm based on complicated noise model are established.Simplified UKF is introduced when measurement equation is supposed to be linear.It avoids the resample,numerous solving the measurement equations and variances.Nonlinear model suitable for simplified UKF is established,and simulations are made compared with UKF and EKF under the conditions of small and large misalignment angle.The result shows that three methods have similar effects under small misalignment angle but UKF and EKF had more advantage in dealing with nonlinear model than conventional Kalman on alignment time and accuracy.EKF is inferior to UKF because of high-order truncation errors for linearization.
出处
《中国惯性技术学报》
EI
CSCD
北大核心
2011年第5期537-542,共6页
Journal of Chinese Inertial Technology
基金
国家自然科学基金(60904088
60874092)
陕西省电子信息系统综合集成重点实验室基金资助(201101Y19)
东南大学微惯性仪表与先进导航技术教育部重点实验室(B类)开放基金资助项目(201008)
基本科研业务费重大引导基金(3222001102)
关键词
初始对准
非线性滤波
简化UKF
EKF
initial alignment
nonlinear filter
simplified unscented Kalman filter
extended Kalman filter
