Abstract:
High measurement accuracy is the reason for the importance of the star sensor as a key instrument in spacecraft navigation systems. However, the existing methods cannot b...Show MoreMetadata
Abstract:
High measurement accuracy is the reason for the importance of the star sensor as a key instrument in spacecraft navigation systems. However, the existing methods cannot be solved or solved inaccurately when considering the error factor of focal plane tilt and rotation, making it difficult to further improve the accuracy of star sensor measurements. To address this problem, we proposed an on-orbit calibration method for star sensors based on the virtual plane. The optimization algorithm is used to find the coordinates of its projection in the virtual plane using star point coordinates according to the idea of finding the best. Utilizing two sets of coordinates to calibrate the influence factors, including principal point, focal length, focal plane tilt and rotation, and distortion, and determine the attitude angle of the star sensor from the projected coordinates. First, a nonlinear equation system among the star point coordinates, the projection coordinates of the star point on the virtual plane, and the error factor was established based on geometric relationships. Then, particle swarm optimization was adopted to find star point projection coordinates for the calibration and attitude angle determination of the star sensor. To address the problem of solving systems of nonlinear equations and the shortcomings of standard particle swarm optimization, we proposed a method based on space resection for solving systems of nonlinear equations and improved particle swarm optimization. Finally, the simulation results demonstrated that the proposed method requires less data and is highly accurate. The validity of the method is verified by night sky observation.
Published in: IEEE Transactions on Instrumentation and Measurement ( Volume: 73)