Research On Seamless Positioning Theory And Method For Multi-source Information Fusion

With the continuous improvement of autonomous driving and robotics technologies,the demands for high-precision seamless positioning and navigation in complex environments are increasing.Multi-source information fusion has been a popular research hotspot for seamless positioning.The integrated navigation technology based on Global Navigation Satellite System(GNSS)and Inertial Navigation System(INS)has widely used for navigation and positioning services in outdoor environments.However,high-precision positioning services in complex outdoor environments,indoor-outdoor transition areas and complex indoor environments are still limited by the environment.Additionally,it is unable to provide positioning and navigation services with high-precision,high-continuity,and high-stability to users.Ultra-Wide Band(UWB)technology has the characteristics of low power spectral density,narrow pulse,low power consumption,high time accuracy,fast transmission rate,anti-interference,and penetration ability,etc.The centimeter-level ranging results in line-of-sight environments are suitable for applications in complex indoor scenes or areas with weak GNSS signals to provide available ranging information.This thesis focuses on the theories and methods related to the application of seamless indoor-outdoor positioning systems in complex environments.Multi-source sensor information fusion using GNSS,INS and UWB sensors and plane map information is utilized to achieve seamless indoor-outdoor high-precision positioning.However,UWB signals are susceptible to non-line of sight(NLOS)errors,multipath effects,and outlier errors in complex indoor environments.In addition,there are still difficulties in how to achieve high accuracy evaluation of indoor positioning technology.In indoor-outdoor transition areas or in complex outdoor environments,GNSS has a severely reduced number of available satellites,poor geometric configuration,and large multi-path effect errors.Moreover,if the INS does not have the constraints of other sensors providing available positioning results for a long time,the system will have the problem of error accumulation.This thesis researches on indoor positioning reference datum acquisition,UWB error analysis and calibration,UWB error reduction method,UWB/INS/Map integration model,GNSS/UWB/INS/Map seamless positioning and multi-source information fusion filtering method.The main research work and results are as follows.(1)UWB data quality analysis and autonomous integrity fault monitoring algorithms are proposed to analyze the performance of UWB indoor dynamic positioning system qualitatively and quantitatively.First,the distance standard time deviation in the UWB line-of-sight environment is calibrated.Then,two reference datum acquisition methods based on precision control points and automatic tracking total station are described.The time delay of the UWB data acquisition system can be calibrated by using the obtained high-precision reference datum and the improved dynamic time normalization model.The impacts of the number of UWB base stations,data quality,and base station distribution on positioning results are analyzed.One conclusion is drawn that improving UWB data quality is a key factor in improving the accuracy of UWB indoor positioning system.Therefore,the UWB autonomous integrity monitoring algorithm is constructed to perform global monitoring and fault location on UWB original observation data,eliminate abnormal data,and ensure UWB data quality.The algorithm reduces the Root Mean Square(RMS)of the easting positioning error from 0.170 m to 0.038 m,and the RMS of the northing from 0.348 m to 0.034 m.The maximum error in the east direction decreased from 3.418 m to 0.123 m,and the maximum error in the north direction decreased from 8.603 m to 0.106 m.The experimental results show that the UWB data quality analysis can effectively analyze the UWB data performance,assist the system to deploy the base station,and the UWB autonomous integrity fault monitoring algorithm can effectively improve the UWB data quality and system positioning accuracy.(2)To weaken the influence of UWB NLOS errors and other outlier errors,a UWB/INS tightly coupled high-precision indoor positioning method based on map constraints is proposed to realize high-precision indoor positioning.First,a mathematical model of the UWB/INS tightly coupled integration is constructed.Then,considering that UWB signals are easily affected by complex indoor structures,the indoor plane map is used to assist the NLOS identification of UWB signals,map information and INS mechanization prediction information are used,and map line segment matching algorithm is used to identify UWB NLOS signals.The weights of observations affected by NLOS errors are adjusted according to the status of identification,and UWB NLOS errors are identified,mitigated and eliminated.Finally,the effects of errors such as unmonitored faults,unidentified NLOS errors and other multi-path errors,outliers,etc.in UWB observations are processed using filtering to weaken the effects of colored noise.The feasibility of the proposed method in complex indoor environment is verified by experiments.The RMS of the position error of the method in the three directions of north,east and plane is 0.19 m,0.19 m and 0.27 m,the average value of position error in the three directions of north,east and plane is0.11 m,0.14 m and 0.21 m,and the maximum error in the north,east,and plane directions is0.56 m,1.05 m and 1.10 m,respectively.Experimental results show that the method can effectively improve the accuracy of the integrated system and achieve high-precision indoor positioning.(3)To achieve high-precision indoor-outdoor seamless positioning,multi-GNSS-tightly coupled(multi-GNSS-TC)RTK/INS/UWB/Map seamless positioning model and method are proposed.First,a multi-GNSS-Tightly Coupled(TC)unified observation model is established,which is tightly coupled with INS to form a multi-GNSS-TC RTK/INS tight coupled system to solve the problem of outdoor complex environment positioning.Then,in view of the characteristics that both GNSS and UWB observations exist in the transition area between indoor-outdoor,and both available observations are small and seriously affected by the environment,the multi-GNSS-TC RTK/INS tight coupled model and UWB/INS tightly coupled indoor positioning system based on map constraints are fused to establish a seamless positioning model with multi-GNSS-TC RTK/INS/UWB/Map integration system.The model improves the utilization of available observation information in complex indoor-outdoor seamless environments,improves the ambiguity fixation efficiency,fixation success rate,and ensures the high precision,high continuity,and high stability of the integrated system.(4)Aiming at the data fusion method of UWB/INS/Map indoor positioning system,multi-GNSS-TC RTK/INS outdoor navigation system and multi-GNSS-TC RTK/INS/UWB/Map seamless positioning system,an Improved Adaptive Robust Extended Kalman Filter(IAREKF)algorithm for multi-source information fusion is proposed.To prevent the noise parameters that are easily generated by the information fusion of various sensors and the probability distribution from being inconsistent with the filtering system,resulting in transitional convergence,transitional divergence,and collapse,the IAREKF algorithm is used.The algorithm can adaptively adjust the observation value noise using the innovation sequence,resist the influence of abnormal observation information,and set upper and lower limits for the noise level of the zero bias and scale factor parameters to effectively prevent the integrated system from appearing abnormal or collapsed,and improve the positioning accuracy and positioning performance of the whole multi-source data fusion system.(5)The hardware and software platform system equipped with GNSS,INS,and UWB sensors is designed independently.The platform is used to verify and evaluate the performance of the proposed seamless positioning system for multi-GNSS-TC RTK/INS/UWB/Map multi-source information fusion.The positioning accuracy and reliability of the system in outdoor complex environments,indoor-outdoor transition areas and indoor complex environments are analyzed.The RMS of the positioning errors of the system in the north,east and plane directions is 0.18 m,0.17 m,and 0.24 m,the average errors in the north,east and plane directions is 0.09 m,0.10 m and 0.16 m,and the maximum errors in the north,east and plane direction is 0.80 m,1.06 m and 1.13 m,respectively.The success rate of ambiguity fixation is89.4 %.The experimental results show that the system significantly improves the stability and accuracy of the parameter calculation of the combined positioning system and can be used as an effective solution to achieve seamless indoor-outdoor high-precision positioning.The paper has 79 figures,13 tables,and 174 references.