| Due to the acceleration of urbanization and the massive increase of modern urban infrastructure,the demand for indoor high-precision and seamless positioning and other location services is increasing.With the continuous development of GNSS technology,the problem of high-precision and seamless positioning under the complex shielding environment of urban buildings and dense buildings still cannot be solved.By introducing a large bandwidth,high rate of indoor UWB technology can further make up the above shortcomings,but there is still a benchmark portfolio architecture is not unified,dual system observation difficult integration,system synchronization problems in time,to solve above problems in GNSS/UWB indoor and outdoor seamless positioning technology research,combining with the characteristics of two kinds of sensor’s own positioning,Detailed theoretical research and experimental verification of the combined positioning model are carried out,and finally the seamless connection from outdoor to transition zone to indoor positioning is realized.The main work of this paper is as follows:(1)combination of GNSS/UWB positioning in the transition area affected by multipath effect observation environment lead to GNSS satellite signal quality serious decline,the satellite multipath space-time modeling study for the above problems,based on empirical mode decomposition method to establish the single station multipath error correction model,to carry out the transition area between multiple sites within the scope of multipath correlation analysis,An error fitting model of multi-parameter points in small area based on spatio-temporal correlation is proposed.The experimental results show that the multi-path error fitting model with multi-parameter points and small area can reduce the multi-path error in the observation range by 35%,effectively solving the influence of the multi-path error.(2)Aiming at the problem that the UWB ranging results are interfered by multiple factors in the transition region and indoor complex environment,which leads to the decrease of accuracy,the UWB ranging error study is carried out,and the UWB ranging error correction model based on Gaussian kernel RBF neural network is proposed.By using neural network to train the observation data,the problem of irregular distribution of ranging error in transition region and indoor complex environment is solved.Experimental results show that 90.21% of UWB ranging point accuracy is within 0.1m after correction based on the range error model,and the problem of UWB ranging precision degradation in the above environment is effectively solved.(3)Carry out UWB network positioning experiment in transition area and indoor complex environment,and finally determine the base station layout scheme and calculation method most consistent with the current environment through the positioning results of different base stations in the above environment;A UWB positioning solution model based on gradient descent was proposed,and the above schemes were verified by Kalman filter,least square method and gradient descent algorithm.Experimental positioning results show that the average point error of UWB positioning algorithm based on gradient descent algorithm is 0.221 m.Compared with kalman filter and least square algorithm,the positioning accuracy is improved by 28.17% and 58.51% respectively.(4)Carry out the combined positioning theory research in the outdoor,transition zone and indoor environment,establish the tight combined double difference positioning model,solve the problem of difficult fusion of multiple systems and multiple observation values in the process of seamless positioning,and improve the accuracy of seamless positioning;For seamless positioning in the process of conventional extended kalman filtering in modeling parameters in the process of the shortcomings and observation on gross error will lead to the filter accuracy retrogressive phenomenon research,set up the indoor and outdoor GNSS/UWB positioning adaptive extended kalman filtering model,by adaptive EKF filtering and fading factor correct weight problem between each system as a whole,Improve the accuracy of location results.The feasibility of the combined positioning model is verified by simulation experiments,and the above two positioning models are tested by static and dynamic field experiments.The experimental results show that the average error of the overall positioning results of the adaptive extended Kalman combined positioning model is 0.212 m and the maximum positioning error is 0.462 m in the above experimental positioning environment.Compared with GNSS system,UWB system improved by 87.38%,49.64%,and conventional extended Kalman filter combined positioning model improved by 31.16%.Effectively realize the goal of indoor and outdoor high precision seamless positioning.(5)After obtaining the combination results,a constraint model based on carrier azimuth was proposed combining with the structural characteristics of urban architectural complexes and map road information in the transition area and indoor environment,and the high-precision seamless positioning process was realized by making full use of the constraint conditions.Experiments verify that the average absolute errors of azimuth values with map constraints are reduced by 67.95%,78.74%,78.11% and 66.3%respectively compared with the original heading Angle errors.The maximum absolute errors of azimuth and true azimuth with map constraints are 88.43%,68.90%,93.23% and66.88% lower than those of original and true azimuth,respectively.The experimental results show that the constraint model improves the accuracy of azimuth estimation and provides a guarantee for the precision of combined positioning. |
PDF Full Text Request