A star tracker should be well calibrated before it is equipped in order to achieve high accuracy. There exists, however, the coupling problem between the internal and external parameters for most commonly used laboratory calibration methods, which affect the star tracker’s performance. We theoretically analyze the major aspects of the coupling mechanism based on the star tracker laboratory calibration model, which means the coupling between the principal point and the installation angle. The concept of equivalent principal point error, which illustrates the effectiveness of the calibration even with poor decoupling accuracy between the principal point and the installation angle, is introduced. Simulation and bench experiments are conducted to verify the laboratory calibration method and its coupling mechanism. The decoupling accuracy can be improved with more samples during calibration. In addition, the equivalent principal point error converges quickly and hardly affects the attitude of the star tracker, which is verified by both theory and experiment. The comprehensive calibration accuracy can still reach a high level even with poor decoupling accuracy.

In order to solve the problem that the calculation result of mathematical model and the measured data of the whole machine deviate greatly due to the conditional simplification in the modeling process and engine component difference, an engine model correction method based on Quantum-behaved Particle Swarm Optimization(QPSO) is proposed. And the engine model calculation results are compared with the measured data of the whole machine performance. The results show that using the QPSO to modify the engine model can significantly improve the accuracy of the model. Before the correction, comparing the performance calculation result of the engine model with the measured data, the maximum error reaches 4.84%. After the correction, the accuracy of the model is greatly improved, and the maximum error is only 0.966%. The correction effect is good.