Research On High-Accuracy Positioning Technology For In-Vehicle Integrated Navigation System

This dissertation studies on the high-accuracy positioning technology for In-Vehicle Integrated Navigation System, funded by National High Technology Research and Development Program (863 Program):"key technologies research and development on new generation of intelligent traffic control system"with NO.2006AA11Z228 and 985 Project of Jilin University:"research and development on intelligent in-vehicle information systems". The GPS/DR integrated location system and the technology of map matching are utilized to decrease the accumulative error of DR while GPS signal is not reliable among the buildings and tunnels in cities. The positioning accuracy and reliability of the In-Vehicle Navigation System can be obviously improved.The dissertation designs GPS/DR federated filter mechanization. The filter is composed of two sub-filters: a Kalman filter and an Unscented Kalman filter (UKF), and they are respectively used to process the information that comes from the linear GPS sub-system and the nonlinear DR sub-system. Meanwhile, the status messages of the sub-systems are collected and the reliability values are achieved through the fuzzy comprehensive judge module. Then, the values are transformed toβ₁,β₂ through the optimized algorithm for dynamic information-sharing coefficient. The main filter outputs the optimal estimate of state and covariance through integrating the two sub-filters. And it can be used to modify or initialize the state of the DR system. Meanwhile, a module of map matching is designed, it also can be used to update the state of the DR while the value of reliability is accepted.This dissertation comprises of 5 chapters and the contents are as follows: Chapter 1: First, present the background, purpose and significance of the research. Then, the development of domestic and international In-Vehicle Navigation System is discussed. At last, some main In-Vehicle Location technologies, the main contents of this dissertation are presented. Chapter 2: The GPS error and the GPRMC format in the integrated location system are discussed firstly. The real-time DR parameters are modified based on the principle analysis and error estimate. The subjection function is presented based on the real-time status messages of sensors of the GPS and DR subsystems and the reliability values of the subsystems are achieved through the improved fuzzy comprehensive judge algorithm.Chapter 3: Firstly, the definition of data fusion, multi-sensor fusion technology and several representative methods are introduced. Especially, some representative Kalman Filter methods and their characteristic are analyzed in detail. This dissertation builds an improved GPS/DR federated filter mechanization based on the NR mode through the comparison of common federated filter mechanizations. This filter is composed of two sub-filters: a Kalman filter and an Unscented Kalman filter (UKF), and they are respectively used to process the information that comes from the linear GPS sub-system and the nonlinear DR sub-system. Meanwhile, the status messages of the sub-systems are collected and the reliability values are achieved through the fuzzy comprehensive judge module. Then, the values are transformed toβ₁,β₂ through the optimized algorithm for dynamic information-sharing coefficient. It can keep the system with optimal state and fault tolerance ability. It can decrease the accumulative error of DR to modify or initialize the sub-system with the optimal estimate of previous position and covariance. The simulation testing of Matlab shows that the algorithm can improve the positioning accuracy and keep a high reliability when the error of GPS is too big for a long time.Chapter 4: The conditions, effect factors and principles of map matching are introduced, and then the main flow of map matching is presented. In this dissertation, the algorithm is utilized to check the road-links in the error range. Corresponding algorithms of map matching are used to pick out the link under different conditions. An improved dead reckoning and a projection algorithm are respectively used to confirm the position of the vehicle in the digital Cyrus-Beck map. And the position with high reliability will be used to initialize the DR system. This dissertation has designed a map matching module based on Visual Basic 6.0, the test proves that the algorithm can meet the demand of real-time positioning for In-Vehicle Navigation System.Chapter 5: The main contents of the thesis are reviewed, and then the innovation and the existing problems are pointed out. Finally, the problems that need to further study are proposed.