Abstract: In the autonomous optical navigation system based on line-of-sight vector observation, the nonlinearity of measurement model is an importance factor which influences the navigation accuracy. By quantitatively calculating the degree of nonlinearity, the selection of measurement model for line-of-sight observation based optical navigation system was studied. First, navigation measurement equations and measurement noise statistics of two line-of-sight measurement models were proposed based on the perspective projection colinearity equations. Next, the effects of measurement model nonlinearity on the navigation performance were thoroughly analyzed within the framework of the extended Kalman filter. Then, the measures of nonlinearity of two measurement models using the differential geometry curvature nonlinearity measure theory were compared. Finally, the simulation of attitude determination system from vector measurement was performed to confirm the proposed theory. The results show that the curvature nonlinearity measure magnitude and the filter performance are consistent and the optical navigation system based on the unit vector measurement model has higher precision and faster convergence properties than that based on the focal plane model.

Key words: Opticalnavigation, Measurementmodel, Curvaturemeasure, Nonlinearity, Deep-spaceexploration