Study On The Key Technologies For Autonomous Star Sensors

Star sensor is a high precise attitude measurement instrument for satellites in the fieldof modern spaceflight. Autonomous star sensor measures attitudes very fast in any positionof space without any apriori information. Study on its key technologies has significancesfor laying the technical foundation for development of autonomous star sensors with highstability, update rate, and performance.In chapter one, the issue studied in this dissertation, its significance and applicationare put forward. The state of art of star sensors at home and abroad is introduced.In chapter two, the concepts of the associated coordination systems and star catalogsused in star sensors are introduced. The constructions and the working principles ofautonomous star sensors are discussed.In chapter three, according to the requirements for autonomous navigation,relationships, determination and examples of optical system performances such as the fieldof view, the focal length, and the entrance pupil diameter and so on, are analyzed and givenout in detail.In chapter four, star image extraction algorithms are studied. Cross-link list structureand region-growing algorithm are introduced. A fast star image extraction algorithm basedon cross-link list and pixel clustering is presented and realized in order to save the time toextract star images.In chapter five, optimal selection algorithm for guide stars and star patternidentification algorithm are studied through analyzing distribution characteristics of starsover the celestial sphere. The optimal selection algorithm based on the Multi-scale ImagePlane Segmentation (MIPS) is proposed in order to have them uniformly distribute. Inorder to reduce the wrong identification rate as a result of false stars and star imageextraction errors, Pentagram star pattern identification algorithm is proposed andprogrammed through investigation and improvement of conventional Pyramid algorithm.The performances and the validities of MIPS and Pentagram algorithms are evaluated bymeans of processing simulated star maps. In chapter six, a method to simulate the imaging of star fields is introduced utilizingthe extension interface of available optical design program. The simulation results aregiven out, and their correctness is tested. The simulated star maps are suitable to study onthe star image extraction algorithm, optimal selection algorithm for guide stars and starpattern identification algorithm.In chapter seven, the verifying experiment of the key technologies relative toautonomous star sensors is implemented. The experimental methods, setups and results areintroduced and reported respectively. The key technologies researched in this dissertationincluding the method to determine the performances of optical system, star imageextraction algorithm based on cross link list and pixel clustering, MIPS optimal selectionalgorithm for guide stars, Pentagram star pattern identification algorithm are verified.Finally, the main work and the innovative aspects in this dissertation are summarized,and the further research is prospected.