A Design Of Low-cost Vehicle Tightly Coupled Navigation System Of Dual-mode GNSS/SINS

With the development of automatic driving technology and BeiDou Navigation Satellite System(BDS),the vehicle navigation system with low cost,high precision and high stability has always been a research focus of scholars and engineers.Compared to the traditional loosely coupled system,a tightly coupled navigation system based on Global Navigation Satellite System(GNSS)and Inertial Measurement Unit(IMU)of Micro-Electro-Mechanical System(MEMS)technology is small,precise and stable even under the circumstance of less than 4 satellites available.Due to the above mentioned considerations,the research in this thesis focuses on the design of a low-cost vehicle tightly coupled navigation system of dual-mode GNSS/SINS.Firstly,the basic technical principles,algorithms,advantages and disadvantages of GNSS and Strapdown Inertial Navigation System(SINS)navigation are discussed.An error model is then proposed to meet the demand of parameter estimation and error propagation.The Kalman Filter and its state parameter compensation are also introduced.Based on the error model of GNSS and SINS system,the discrete Extended Kalman Filter for tightly coupled system is constructed.Besides,models of the vehicle kinematics constraint,dual-antenna GNSS and other vehicle sensors are discussed to help localization.In order to solve the problem of time synchronization,a double rate Kalman filter is also proposed and verified by experiments with the consideration of the different output frequencies between GNSS and SINS,as well as the adaptive algorithm to deal with the time-varying noise.According to the vehicle trajectory producer,simulations of loosely coupled system and tightly coupled system are made to compare the two kind of algorithm.In addition,data fusion schemes with different sensor combinations are suggested in this section.The algorithm is transplanted to Microsoft Foundation Classes(MFC)by C Language and then completed by the function of anomaly detection and reconstruction.With the analysis of GNSS and SINS sensors' output,some of the sensors are chosen for the design and manufacture of the circuit board which refers to the minimum system of microprocessor.A satellite capture test of the snow environment is carried out in this thesis as well as a low speed test and a car test.The result shows that the tightly coupled algorithm is able to achieve the precision of vehicle localization in the actual environment.The tightly coupled system is more stable than one with loosely coupled algorithm.