Multi-source Information Integrated Positioning Research Under The Environment Of Urban Canyon

With the rapid development of the intelligent terminal and communication network, the Location-based Service(LBS) applications consisting of tourism resource,tourism activities, photos and Location tagging, buildings and landmarks recognition,social networks, games and sports, advertising and Augmented Reality(AR), are of great commercial value. The major occasions of using these applications is in "urban canyon", where are high density buildings and short distance between buildings. The traditional GNSS receiving the signals will be affected by the building block,reflection, diffraction and visual satellite distribution of strip in this scenario and it will causes positioning error up to hundreds of meters or can’t fix a position. Accurate location information is the key to realize the LBS application. Therefore, this paper will carry out this work in two areas: Tracking and Positioning algorithm research based on the simulation of GPS data and Integrated localization algorithm based on Intelligent terminal multi-source information in urban canyon. The main work is as follows:The research of the Nonlinear filtering applied in positioning and the carrier tracking. Analysis and simulation location algorithm based on Extended Kalman filtering(EKF) and Unscented Kalman Filter(UKF). The problems of model and noise statistical properties is not accurate and larger positioning error or even divergence problems for EKF and UKF. Introduces Particle filter(PF) to improve the positioning accuracy. Experiment results show that PF than EKF, UKF in the x, y, z direction of the position error average decreased by 0.69 m, 1.34 m, 0.72 m, the position error variance decreased by 10.23 m, 14.38 m, 8.40 m, respectively. Phase curve of the traditional carrier loop as periodic function. When the phase error is beyond the scope phase, which will prone to 180 °phase ambiguity. Besides, non-gaussian noise may introduce to carrier tracking in the actual application. Introducing PF in the GPS carrier tracking loop instead of carrier phase. Experiments show that the carriertracking loop based on PF, compared with the traditional carrier loop and carrier loop based on EKF, The mean phase increased from31.8 10-? to30.5 10-?, phase variance from38.4 10-? to30.5 10-?, which has better anti-noise performance.Multi-Source information integrated positioning algorithm research based on intelligent terminal in Urban Canyon. In urban canyon scenario, due to the structure of the shade to Cross the street dilution of Precision(CDOP) influence caused by crossing accuracy degradation, making positioning error in cross-street of convention GPS. Shadow Matching(SM) algorithm can improve the cross-street positioning accuracy.Which based on satellites visibility predicted by building boundary of 3D city model with elevation information of satellites and satellites visibility by GNSS receiver for matching to location. However, SM cannot improve the positioning precision of along-street. SM/KF and SM/PF algorithm can improve the precision of cross-street, but is not obvious to improve the precision of along-street in city ditch scene which building outline does not change significantly. Therefore, based on the distribution characteristics of PF position precision positioning algorithm. Proposing the SM/PF adaptive weighted integrated localization algorithm. This algorithm according to the SM positioning lingering time under motion characteristics and the statistical properties of the template score to calculate adaptive weighted factor.Adaptiving weighted SM and PF positioning algorithm to get the final localization results. Experimental results show that the proposed algorithm, compared with SM,PF, SM/KF, SM/PF. Increased average error from 5.04 m to 2.80 m in along-street.Increased average error from 4.87 m to 1.30 m in cross-street. Increased the root-mean-square error from 6.82 m to 3.47 m. At present, it is difficult to get detailed3 D city model, which make SM algorithm cannot widely used in all city scenes.Therefore, putting GPS/INS integrated positioning algorithm based on intelligent terminal sensors. GPS/INS can realize continuous and accurate positioning in the dense urban canyon compared with the traditional GPS positioning algorithm.Experimental results show that GPS/INS based on the intelligent terminal multi-source sensors compared to conventional GPS.Increased average error from4.76 m to 2.31 m and the root mean square error from 19.25 m to 2.52 m in along-street. Increased average error from 3.78 m to 1.49 m and the root mean square error from 66.83 m to 5.70 m in cross-street.