Research On Robust Train Positioning Method For Satellite Signal Constrained Observation Environment

Train positioning based on the integration of Global Navigation Satellite System(GNSS)and Inertial Navigation System(INS)can effectively overcome the shortcoming of error accumulation by INS,and provide accurate,continuous and reliable positioning information for train control system.Based on specific system state-space model and the Gaussian assumption,the Kalman filter algorithm is able to calibrate and constrain the cumulative INS error using the observation information of navigation satellites.However,during practical train operation,the characteristics of positioning observations would be time-varying.Thus,the conventional Kalman filter based on fixed parameter assumption cannot be adaptive to the dynamic changes and will suffer from positioning performance degradation or even the filtering divergence.In addition,the quality of satellite signal observation is susceptible to the train operating environment.Obstruction or signal interference along the railway will constrain the satellite signal observation condition,which may affect the performance of GNSS/INS integrated positioning.Therefore,how to cope with the constraint in satellite observation conditions and improve the robustness of train positioning has become a key factor to the implementation of Bei Dou Navigation Satellite System(BDS)in railway applications like train control systems.In this thesis,the robust train positioning method is studied according to the practical demand under the constrained GNSS observation environment.A GNSS/UWB/INS tightly coupled train positioning framework is proposed,and a robust Unscented Kalman Filter(UKF)algorithm with the adaptive capability to the observation characteristics is designed.The performance of the proposed train positioning framework and method are validated under different types of GNSS-constrained conditions.The contribution of this thesis can be summarized as follows:(1)The error sources and corresponding processing methods of satellite positioning are analyzed.Based on the establishment of error models for GNSS and INS systems,a GNSS/INS tightly coupled train positioning framework is designed.(2)Based on the robust estimation theory,a robust UKF estimation algorithm is proposed.An innovation-driven adaptive filtering strategy is designed to improve the robustness against different characteristics of the measurements.The performance of the proposed adaptive robust UKF algorithm is verified and analyzed using real scene data.(3)Considering two typical GNSS-constrained conditions including complete signal blockage and signal interference,a UWB/INS coupled positioning method is designed by using the Ultra Wide Band(UWB)technology,and a GNSS signal interference detection and exclusion algorithm is proposed.The performance enhancement capability of the proposed schemes is illustrated based on field data and simulation experiments.According to the research and validation results,the parameter adaptive robust estimation method proposed in this thesis can achieve expected positioning performance under different observing conditions.The utilization of UWB/INS coupled positioning mode and the involvement of the interference detection and exclusion mechanism can provide an effective protection to the positioning performance against typical GNSSconstrained conditions,which will greatly support the application of Bei Dou Navigation Satellite System in new generation railway train control systems.There are 78 pictures,14 tables and 55 references.