Abstract:

 Considering the coupled and nonlinear relationship between the relative pose in the line of sight
equations, an iterative decoupling algorithm is proposed to determine relative pose. Application of the method
to aerial refueling is discussed. Based on the pseudo-range of satellite navigation, an observation matrix is
constructed by using line of sight vectors. The relative position are updated by the least square method, and
then the relative attitude is updated. By introducing an iterative error, convergence of the iterative algorithm
is proved. The trace of error variance matrix is used to analyze the effect of feature geometry configuration
on the precision. Simulation results show that the proposed algorithm can improve performance by 23% and
40% in speed and robustness than Gaussian least square differential correction (GLSDC), respectively. Steady
state errors of relative pose are 0.1 m and 0.2± lower respectively. The precisio can satisfy the relative pose
requirements of UAV aerial refueling during docking.

Key words:  aerial refueling, relative pose, line of sight, decoupled iterative