基于协方差变换的INS/DVL全球组合导航算法

All-earth INS/DVL integrated navigation algorithm based on covariance transformation

横坐标系下设计的惯性/多普勒测速仪(INS/DVL)组合导航滤波器可以满足无人潜航器(AUV)极区导航的需要,但在AUV进入、离开极区时导航坐标系的切换会引起滤波器结构变化,使系统的误差状态参数估计值产生震荡,影响导航精度。针对此问题,提出了基于协方差变换的INS/DVL全球组合导航算法,设计了横地理坐标系下的INS/DVL组合导航滤波器,并推导建立了地理坐标系与横地理坐标系下滤波器系统误差状态及其协方差矩阵的转换关系。AUV进入、离开极区过程中,导航坐标系切换时同时完成导航参数、滤波器参数的转换。通过船载实验与半实物仿真实验对所提算法的有效性进行了验证,结果表明:在导航坐标系切换过程中,协方差变换算法能够有效改善滤波器估计值不稳定问题,相比于无协方差变换,导航参数震荡误差减小50%以上,提高了组合导航系统稳定性。

The inertial navigation system/Doppler velocity log(INS/DVL)integrated navigation filter designed in the transverse coordinate system can meet the needs of autonomous underwater vehicle(AUV)navigation in the polar region,but the switching of the navigation coordinate system when the AUV enters and leaves the polar region will change the filter structure,which makes the estimated values of the system error state parameters oscillate and affects the navigation accuracy.To solve the problem,an all-earth INS/DVL integrated navigation algorithm based on covariance transformation is proposed,the INS/DVL integrated navigation filter is designed in the transverse geographic coordinate system,and the transformation relationship of filter system error state and its covariance matrix is derived between geographic coordinate system and transverse geographic coordinate system.The conversion of navigation parameters and filter parameters is completed simultaneously during the navigation coordinate system switching.By ship experiment and hardware-in-the-loop simulation experiment,the proposed algorithm validity is verified.The result shows that:during the navigation coordinate system switching,the covariance transformation algorithm can effectively improve the instability of the filter estimates,and the navigation parameter oscillation error is reduced by more than 50%compared with non-covariance transformation,which improves the stability of the integrated navigation system.