Research On The Key Technologies And Application Of Multi-GNSS Precise Differential Positioning With Tight Integration Of Mixed Frequencies

With the development of GNSS,the satellite navigation and positioning has entered an era of multi-system co-existence.The precise positioning technique has become an important support for the information construction of emerging industries(e.g.,the smart city and intelligent transportation).However,multi-system integrated data processing is still inadequate,for instance,the insufficient use of observation data,the inconsistent models of data processing,and the poor positioning performance in complex environments such as urban canyon.Aiming at the issues of deep-fusion of multi-GNSS information,this thesis is focused on the four parts:(1)The technique of tight combination for both same and different frequency of CDMA systems.(2)CMDA/FDMA systems inter-system model with consideration of interfrequency bias and inter-system bias.(3)On-line determining the time-varying characteristic parameters of DISB for Network RTK users.(4)Inter-system model for Network RTK.Based on the innovation or improvement of some algorithms and models,the positioning performance in complex environments such as urban canyon will be improved.The main work and contributions are listed as follows:1.The tightly combined model between the inter-system observations with mixed frequencies is proposed.Conventional DISB processing models for overlapping frequencies and short baselines are difficult to effectively compatible with current mixing processing of multi-system and multifrequency signals.For this problem,the thesis intends to research the unified function model of inter-system bias and break through the fast estimation of carrier DISB,so that the model can be applicable to mixing processing of multi-system and multi-frequency signals and tightlycombined high-precision positioning can be implemented.The specific ideas are as follows: the rank-defect problem among DISB,receiver clock,ambiguity,ionosphere delay and other parameters will be studied,so that a full-rank single-difference un-combined observation model can be founded.(1)For same frequencies with the homogeneous receiver types,the DISB is closed to zero,and for the heterogeneous receiver types,the DISB are generally very stable.For nonoverlapping frequencies,the DISB are also stable,the amplitudes relative to the mean value are all within 0.1 cycles for a short time,and the STD are all near 0.01 cycles.(2)Through the real-time estimation of DISB parameters in multi-epoch mode,the strength of positioning model can be enhanced.Experimental results show that compared with the conventional intra-system differential model,the tightly combined model can effectively improve the positioning accuracy and reliability,especially for the obstructed environments with a small number of satellites available.When the number of visible satellites is no more than seven,the positioning accuracy can be improved by over 20%.2.The tight combined model between different signal transmission mechanisms(CDMA/FDMA)is proposed.(1)After analyses the relationship between GLONASS code IFB and channel for double differencing,we proposed a code IFB correction model and an inter-system code-differencing model between GLONASS and GPS.Based on the experimental results,we can conclude that the code IFB have no linear relationship with channel numbers and should be estimated for each GLONASS satellite.Moreover,although the code IFB for the same or different receiver pairs are different,from the general point of view,the code IFB are all stable over whole time and the standard deviations of code IFB for most satellites are less than 0.2m,and therefore the average code IFB estimates can be considered as the final results for each baseline to generate correct table to calibrate GLONASS inter frequency code bias.We can also find that even if the receiver types and the firmware versions are same,the code DISB between GPS+GLONASS are nonzero,but the code DISB are very stable as a whole for the four baselines and the standard deviations of code DISB for four baselines are all less than 0.45 m.Regardless of the random terms caused by observation noises,it can be considered that the code DISB for GPS+GLONASS is stable over a few days,with this feature,it is expected to increase the redundant observation information for DD code relative positioning.(2)A single difference-based uniform inter-system model suitable for all of the CDMA and FDMA frequency combinations is proposed in this study.We first utilize a particle filter approach to extract the IFPB rate.Afterward,the multi-ISB parameters can be estimated by the proposed model simultaneously.Four short baselines containing the signals of GLONASS,GPS,BDS-3,BDS-2,Galileo,and QZSS are selected to validate the model.The quantitative results indicate that the phase IFB rate can be extracted precisely by particle filter within 10 epochs.The phase DISB valuations are sufficiently stable on a daily scale and have standard deviation(STD)near 0.01 cycles.Furthermore,imposing the stable DISB on the multi-GNSS intersystem model can effectively improve the single-frequency real-time kinematic(RTK)positioning accuracy and reliability for severely obstructed situations with only a small number of satellites observed.3.An On-line phase DISB determining method for Network RTK users is proposed.In Network RTK scenarios,when the number of observed satellites for each GNSS system is small,it is difficult to estimate F-ISB by the traditional method to form inter-system model.In this study,we proposed a state optimal estimation-based F-ISPB estimate approach to exact F-ISPB for multi-GNSS inter-system positioning for the Network RTK users.Four different FISPB handling schemes,which are step-by-step PF,step-by-step PSO,multidimensional PF and multidimensional PSO were proposed,and two baselines observation data were adopted to validate the optimal method under open and simulated complex environments.Besides,positioning results after calculate F-ISPB have been achieved.The results show that compared with step-by-step PF,the estimate results of PSO were much worse than that of step-by-step PF,especially for long baseline,which is due to the wrong F-ISPB may bring about maximum Ratiofor a long time during initial times,and the PF have a persistent variance for each particle than PSO,thus step-by-step PF method is recommend than step-by-step PSO.Compared with the one-dimensional situation,under the multi-dimensional situation,the Ratiois related to multi-F-ISPB,thus the estimated results may be better than one dimensional case.However,the two-dimensional results also shown that all of the F-ISPB still cannot be extracted under complex environments by multidimensional PSO.Furthermore,compared with step-by-step PF,the multi-dimensional PF method cost too much to get the right value.For step-by-step particle filter,each epoch only needs to calculate 200 times,but the calculation times of multi-dimensional particle filter increase exponentially.For example,in the two-dimensional case,each epoch needs 40000 times,while in the three-dimensional case,each epoch needs to calculate 8000000 times,which is unbearable for any receiver.More importantly,the step-by-step PF can obtain the right F-ISPB with about 100 epochs no matter what scenarios.And the positioning results show significant benefits can be attained under complex environments by inter-system model which obtained F-ISPB by step-by-step PF compared to normal intra-system situation.Hence,making a comprehensive survey of advantages and disadvantages of the four methods,for the Network users,a step-by-step particle filter-based F-ISB estimate method is recommended for multi-dimensional inter-system model.4.A network RTK tight combination positioning model considering the DISB between reference stations is established.Aiming at the problem that the data center only calculates the intra-system atmospheric correction in the network RTK Positioning model,the inter-system atmospheric error is studied,and the network RTK Positioning Software Earth Net 3.0 is developed.In order to achieve uniform and high-precision positioning in the region,the code and phase DISB between reference stations are corrected in advance,the tight combination solution of baseline between reference stations is realized,and the atmospheric error interpolation method between network RTK systems is constructed.In particular,aiming at the problems of few available interpolation models and atmospheric interpolation accuracy of conventional triangular modeling unit.Using Delaunay triangulation,a regional atmospheric error modeling method with sub optimal network elements is proposed.According to the user location,the main network element is used to realize the automatic search of sub optimal network elements,which breaks through the limitation of the number of base stations of conventional triangulation units;The elevation factor H is introduced to realize the transformation of tropospheric error from plane(x,y)to three-dimensional(x,y,H)modeling,and weaken the influence of inter station elevation difference and satellite altitude angle change.The problem of inaccurate estimation of atmospheric error caused by ionospheric activity and elevation difference is effectively solved.On this basis,a unified compact combination positioning processing model in the observation domain is constructed to improve the observation redundancy and optimize the observation structure,so as to improve the positioning performance of Network RTK clients in complex environments.