Research On Autonomous Navigation And Integrity

With the progress of technology, various navigation technologies which are used in military field, have been applied in or are being popularized to civilian field. In civilian field, the safety of navigation system is of great importance. Integrity which is a measurement of the safety of navigation system, is a vital criterion for evaluating navigation system. Autonomous navigation can guide the moving objects without goals and information sources which are manually set. Inertial navigation is a kind of autonomous navigation, and integrated navigation which consists of inertial navigation is also a kind of autonomous navigation.In navigation system, measured information should be obtained for navigation solution. This dissertation mainly focuses on how to improve the integrity of navigation system by means of fault detection.Firstly, this dissertation analyzes inertial navigation system. In inertial navigation system, initial alignment should be carried out firstly, in order to determine initial attitude of the moving object. Initial alignment consists of coarse alignment and fine alignment. The fine alignment estimates the error of the coarse alignment by means of Kalman filtering. This dissertation proposes an improved fine alignment algorithm, which uses the five output values of two accelerometers and gyroscope as the observed values of Kalman filter. A mathematical model is set up to simulate the output values of accelerometer and gyroscope, and inertial navigation system platform is established to simulate system performance. Then traditional fine alignment algorithm and improved fine alignment algorithm are compared, and simulation results show that the improved fine alignment algorithm improves the accuracy of the azimuth error angle estimation.Secondly, GPS navigation system is simulated, and then the loosely coupled and the tightly coupled integrated navigation are respectively simulated based on GPS navigation system and the established inertial navigation system platform. Simulation results of various navigation systems are compared, and the established navigation system platforms provide a foundation for the research of fault detection.Thirdly, fault detection algorithms are researched. The optimal parity vector algorithm, which is the basis of multiple fault detection, is introduced. However optimal parity vector algorithm has a disadvantage which is that the faults of various observed values may make the residual error be small. In order to address the problem, this dissertation proposes an improved optimal parity vector algorithm. Simulation results show that the improved optimal parity vector algorithm can solve the problem effectively.The accuracy of the observed values may be different in practical applications, then the residual standardization method and the weighted parity vector algorithm are simulated, and the simulation results show that the weighted parity vector algorithm have an advantage in fault detection. In addition, we find that the parity vector algorithm can’t be used to detect the fault of observation equation, when the number of the values to be estimated is much more than the number of the observed values in Kalman filter, therefore, the algorithm of constructing new statistics should be applied to detect the faults. Simulation results show that the algorithm of constructing new statistics has good performance in fault detection.Finally, the method and feasibility of introducing fault detection to GPS and integrated navigation system are researched. Simulation results show that the proposed fault detection algorithms can effectively detect the faults in GPS and integrated navigation system.