Research On Key Technologies For High Altitude And Long-flight-time UAV Information Fusion Autonomous Navigation

Unmanned aerial vehicles (UAVs) are widely used in both military and civilian domains for theadvantages of low cost, zero casualty and so on. An important aspect of UAVs is the research on highaltitude and long-flight-time UAV. Autonomous navigation system with high precision has been amajor constraint on the improvement in performance of high altitude and long-flight-time UAV.Combined with the pressing need for navigation system with high precison and reliability from UAV,autonomous control system, this dissertation studies the information fusion techniques and algorithmsfor high altitude and long-flight-time UAV. The objective is to form an effective theoretical base forthe research and application of autonomous navigation system for high altitude and long-flight-timeUAV.Based on the comprehensive analysis of the demands of high precison and reliability for highaltitude and long-flight-time UAV, high precison strapdown inertial navigation algorithms andinformation fusion autonomous navigation algorithms are porposed to improve the precision and thestability of autonomous navigation system by making full use of the drone airborne sensorinformation.The inertial/celestial intergrated localization algorithms assistanted by celestial navigation arestudied to reduce the GPS dependence. Multi-round intersection iterative astronomical localizationalgorithms with better numerical stability are deeply studied. The positioning covariance matrixprovided by the the method is used to propose inertial/celestial intergrated adaptive navigationalgorithms based on H∞filter, and effectively improved the positioning accuracy and robust ofintegrated navigation. The altitude angle, the platform error horn, and the longitude/latitude errormodels are all deduced. Then a new intergrated localization algorithm using the astronomical altitudeangle as measurement information is presented which is coupled with platform error horn. Sufficientsimulations have proved that the proposed algorithms are effective.Aerial reconnaissance, as an important function for high altitude and long-flight-time UAV, needshigh precision of attitude and position of navigation system. The INS/GPS/CNS intergratednavigation system is put forward to improve the precision and stability of the integrated navigationsystem. GPS information compensation algorithm based on extrapolating method is established. Theseparation filters is designed which is centralized and asynchronous with time update and measureupdate. As GPS is susceptible to be interfered in the lead to instability, in-depth analysis is done to themutual transformation of the inertial attitude and geography attitude. The transformation matrix of the navigation position errors is introduced, and coupling error model base on inertial attitude measuringby star sensors and geography attitude outputting by inertial navigation are established. The airborneinertial/celestial intergrated algorithm base on the geography position error is put forward, andenhances the stability of the combined system and autonomy.As celestial navigation system is limited by the altitude and weather and the GPS information isunstable, the SINS/GPS intergrated navigation system assisted by Air Data System is proposed. Usingthe attitude and the airspeed information provided by pressure sensors, the SINS/ADS measurementmodel is derived and the SINS/GPS/ADS navigation algorithms is designed. On the other hand, theSINS/GPS tightly coupled navigation algorithm is the effective methods to improve the SINS/GPScombination's stability.The non-linear measurement model of pseudo-range and pseudo-range rateare deduced. Due to the non-linear characteristics of measurement equations, a UKF filteringalgorithm is established. Simulations show that the algorithm can effectively guarantee precision ofthe system.Inertial navigation system is the required system for high altitude and long-flight-time UAV.Thorough analysis in the traditional strapdown inertial navigation algorithms is done.To improve theperformance of SINS, an improved strapdown inertial algorithm based on the fourth-order R-Kmethod is presented. Moreover, a trace-generation algorithm based on kinematics and attitudemaneuvers model is designed and to confirm the improvement of the strapdown inertial algorithm andsemi-physical simulation system.Finally, SINS/GPS/CNS/ADS navigation semi-physical simulation system is established. Thealgorithms presented in this paper have been proved by the semi-physical simulation system. At thesame time, considering the autonomous navigation requirements of one UAV, aSINS/GPS/ADS/HMR principle prototype is designed and tested by actual car-based experiment andairborne ground test. The test results show that the navigation system accomplished in the principleprototype has better precision and autonomy.