Research On Multi-information Fusion And Adaptive Fault-tolerant Navigation For Near Space Hypersonic Vehicle

The near space hypersonic vehicles(NSHV) get the attention of the military powers with its great strategic application value. The NSHV has a speed of more than 5 Mach and maximum flight heights of 30 to 70 kilometers with the characteristics of global reach, penetration ability, and high combat effectiveness. In order to ensure NSHV flight safety in rapidly remote maneuvering and the precision of strikes over the horizon, adaptive fault-tolerant autonomous navigation technology with high reliability and precision becomes key technology in the field of NSHV navigation.For the purpose of improving the precision and reliability of NSHV navigation, this thesis developed the technology of multi-information fusion autonomous navigation for NSHV. Based on analyses of characteristics and demands for NSHV flight, the thesis proposed a series of information fusion schemes and algorithms of multi-information autonomous navigation. The research contents are as following.Aiming at reducing the impact of non-Gaussian noise disturbance on astronomical attitude measuring, the robust adaptive astronomical attitude measuring algorithm of inertial/celestial integrated system was studied. The influence of near space hypersonic flight environment will cause the performance loss of celestial observation. To enhance the stability of the inertial/celestial navigation, astronomy attitude error covariance matrix online calculation model was presented, and then robust adaptive filtering algorithm for inertial/celestial attitude determination was proposed based on generalized maximum likelihood estimation and navigation star geometrical configuration.To improve the reliability of navigation positioning in near space hypersonic flight, Inertial/celestial tightly-coupled integrated positioning system which use starlight angle as observations can be utilized as a redundant positioning information source for NHSV. This thesis proposed a new dynamic filtering gain adjustment method based on P-value quality map for inertial/celestial tightly-coupled system. This method can evaluate the quality of observations online when navigation stars changes dynamically in the flight, and the difference between elevating angle observations of stars will be used to adjust filtering gains. The proposed method improved the adaptability to environment change and disturbance, which can enhance the autonomy and adaptability of navigation system.The complicated attitude maneuvering of NSHV will motivate the measuring offset errors in integrated navigation systems. In this thesis, measuring offset errors compensation method of NSHV integrated navigation system was also researched. For the integrated navigation system that uses position and speed as observations, an iteration compensation method was proposed to improve navigation accuracy under the attitude continuous maneuvering condition. For the integrated navigation system that uses pseudo range and pseudo rangerate as observations, the linear/non-linear model was built to resolve the problem that measuring offset errors cannot be compensated directly, and then the non-linear filtering compensation method for inertial/celestial is studied to improve error states estimation performance in the maneuvering flight.In order to enhance attitude measuring performance for inertial/GNSS integrated system in near space hypersonic flight, double-layer structure and filtering approach for multi-information fusion system were further studied. The double-layer information fusion scheme of embedded inertial/ satellite integrated system assisted by star sensor was proposed. According to the characteristics of the open-loop and closed-loop hybrid, double-layer information fusion algorithms were deduced under the condition of synchronous and asynchronous measurement respectively. The proposed scheme and method provide an effective approach of attitude performance upgrade for NSHV navigation system.Taking into account the fault-tolerant performance deficiencies of traditional federal filtering algorithm in the case of slowly varying and rapid fault existing in NSHV, adaptive fault-tolerant techniques of multi-information fusion navigation were studied. A kind of adaptive fault-tolerant algorithm based on sequential probability mapping for integrated navigation was proposed. The statistic model of local estimation states was built by fault dynamic mapping. Then quality of local estimation states under soft fault was calculated online by sequential probability ratio test and finally used in adaptive tuning of global fusion. The simulation indicated that the proposed algorithm can significantly improve the fault-tolerant performance when soft faults happen, which result in a higher accuracy of integrated navigation.The digital and semi-physical simulation systems of NSHV multi-information fusion and fault-tolerant autonomous navigation have been built to verify proposed schemes and algorithms. Based on the constructed simulation platform, integrated navigation with inertial, celestial and GNSS in dynamic flight were carried out. Through digital and semi-physical simulation test, the reliability and validity of the system were tested, which laid a good foundation for further engineering test and development of actual system. The researches of this thesis are helpful to improve the autonomy, reliability and adaptability of NSHV multi-information fusion navigation, which have both important theoretical meaning and engineering value for NSHV applications.