基于DVL信息辅助下的捷联惯导初始对准可观测性分析

Observability Analysis of Strapdown Inertial Navigation Initial Alignment Based on DVL Information

在利用DVL与SINS进行初始对准时,一般将系统误差及相关器件的误差作为状态变量进行滤波估计,状态估计的精度及速度决定着系统初始对准的速度及精度,而状态估计的精度及速度又与系统的可观测性有很大关系,当被估计的状态不可观测或者可观测性较低时,将状态估计值反馈给系统会造成系统的精度下降甚至发散,因此需要对系统状态变量进行可观测性及可观测度的分析,针对以上问题,论文构建了基于DVL信息辅助下SINS初始对准的系统模型,利于基于系统可观测矩阵奇异值分解(SVD)的方法分别对模型在静止和转动状态下进行可观测性分析,结果表明转动状态下能激励惯性器件的各误差参数,使基于DVL信息辅助下SINS初始对准的系统模型可观测。

When using DVL and SINS for initial alignment,the system error and related device error are generally estimated as state variables.The accuracy and speed of state estimation determine the speed and accuracy of the initial alignment of the system,and the accuracy and speed of state estimation.It is also related to the observability of the system.When the estimated state is unobservable or the observability is low,feeding back the state estimation value to the system will cause the accuracy of the system to drop or even diverge.Therefore,the system state variables need to be performed.Based on the analysis of observability and observability,this paper constructs a system model based on DVL information-assisted SINS initial alignment,which is beneficial to the system based on the singular value decomposition(SVD)method of the observable matrix.The observability analysis is carried out in the rotating state.The results show that the error parameters of the inertial device can be excited in the rotating state,so that the system model based on the initial alignment of the SINS assisted by the DVL information can be observed.