基于不变误差定义的INS/USBL紧组合导航算法

Invariant error definition-based tightly integrated navigation algorithm for INS/USBL

针对现有惯性导航系统/超短基线定位系统(INS/USBL)组合导航算法误差定义不准确,引发空间不一致性及导航精度下降的问题,提出了一种基于不变误差定义的INS/USBL紧组合导航算法。首先,引入李群理论及左不变/右不变误差定义下的误差状态方程,保证了误差状态的空间一致性。然后,从噪声无偏性分析的角度阐述了松组合的缺陷。最后,利用超短基线测量的方位角及斜距信息,根据两种误差定义推导相应的紧组合量测方程。半实物仿真结果表明:相比于现有的INS/USBL松组合导航算法,所提出的基于右不变误差定义的紧组合导航算法有着更高的定位精度,2000 s后东向和北向定位精度分别提高14.65%和30.26%,天向定位精度提高73.91%;不变误差定义使得组合导航算法对初始对准精度的需求降低,即使初始姿态误差较大也能够对导航误差进行准确估计。

An inertial navigation system/ultra-short baseline(INS/USBL)tightly integrated navigation algorithm based on invariant error definition is proposed to address the problem of inaccurate error definition,which causes spatial inconsistency and navigation accuracy degradation.Firstly,the Lie group theory and the error state equation under the definition of left-invariant/right-invariant are introduced to ensure spatial consistency.Then,the shortcoming of the loose couple algorithm is described from the perspective of noise unbiasedness analysis.Finally,using the azimuth and slant distance information,the corresponding tight couple measurement equations are derived based on the two error definitions.The semi-physical simulation results show that compared with the existing INS/USBL loosely integrated navigation algorithms,the proposed tightly integrated navigation algorithm based on right-invariant error definition has higher positioning accuracy.After 2000 s,the eastward and northward positioning accuracy is improved by 14.65%and 30.26%,respectively,and the height positioning accuracy is improved by 73.91%.The invariant error definition reduces the requirement for initial alignment accuracy and enables accurate navigation error estimation under large initial attitude errors.