Research On The Key Technology Of Small Satellite Multi-Sensor Autonomous Navigation

Modern small satellites are attracting more and more attention all over the world because of the low-cost, ability of new technologies rapid adopting, and smart flexibility. In this dissertation, the technical characteristics of LEO (low earth orbit) small satellite for the study, the autonomous navigation technology and the integrated navigation technology is carried out, which mainly consist of the following parts:Firstly, in view of the superiority of geomagnetic navigation method, a scheme of using magnetometer/radar altimeter as the small satellite's basic autonomous orbit determination sensor is bring forward based on the research of the conventional methods of geomagnetic navigation, so a filtering algorithm in the earth-centered inertial Cartesian coordinate system of this scheme is established. In addition, for enhance the geomagnetic navigation plan's accuracy and reliability effectively, the star sensor and infrared horizon sensor based autonomous navigation method and the ultraviolet sensor based autonomous navigation method was studied. Taken these two methods as the effective supplement of the geomagnetic navigation plan, a corresponding small satellite multi-sensor autonomous navigation scheme which use federated filtering structure is set up. The autonomous navigation system's accuracy and reliability are improved.Secondly, because of the typical nonlinear characteristic of small satellite orbit determination system, the implementation method of unscented Kalman filter (UKF) which based on the symmetrical distributed sampling strategy is analyzed, becaused of the low efficiency of this algorithm, the improved square root unscented Kalman filter (SRUKF) algorithm which uses the minimal skew sampling strategy is put forward. On the basis of it, a federated SRUKF algorithm which combines the federated Kalman filter with the square root unscented Kalman filter is proposed subsequently, it can increase the accuracy of the multi-sensor navigation system. Furthermore, aimed to the problem of the local filters have different intervals, a relevant multi-information fusion filtering algorithm is presented. So the integration of multi-sensor information is ensured. The study mentioned above has important theoretical significance and engineering value to the practical application of the small satellite navigation system.Thirdly, for the purpose of enhance the reliability and the fault-tolerant performance of the integrated system, the fault detection algorithms of federated SRUKF is studied. For the limitations of state chi-square detection method, the improved double state chi-square detection algorithm is put forward, furthermore, the improved residual chi-square detection method which used fuzzy threshold is discussed. The fault-tolerant scheme and algorithms of small satellite autonomous and integrated system are designed on the basis of it. The double state chi-square detection method is used in the master filter to avoid the gradual fault and the improved residual chi-square detection method is used in the local filters to avoid the abrupt fault, so the gradual faults and abrupt faults are both detected and isolated.Finally, for verify the proposed schemes and algorithms in this paper effectively; the orbit determination simulation platform for small satellite is constructed. This platform is composed of the professional satellite simulation software STK (satellite tool kit) and MATLAB. The small satellite fault-tolerant multi-sensor integrated navigation system has been simulated on the platform and gained good effect; it laid a solid foundation for the practical application of algorithms.