Analysis of Geometric Effects on Tightly-Integrated INS/Vision for Lunar Descent Navigation

The dilution of precision(DOP) used in GPS as performance measure can be applied to the vision navigation to analyze the errors introduced by the geometry between the vision sensor and the matched feature points. In this paper, we derive the DOP of the vision navigation for 4 matched feature points analytically and verify the DOP with numerical simulation. The result shows that the vision DOP is dependent on the distance to the feature points and distribution width of the points. We also analyze the influence of the vision geometry on the performance of the tightly integrated INS/vision with the star tracker through covariance analysis for lunar descent trajectory. The simulation results show that the accurate estimation of the attitude error of the integrated navigation system decouples the position DOP and the attitude DOP of vision navigation, results in smaller position error than the error calculated from the vision DOP.