Return to Session E3

Session E3: All-source Intelligent PNT Method

3D Ray Tracing Based BDS Multipath Characterization in Urban Canyons
Shaoqian Li, Rong Yang, Xingqun Zhan, School of Aeronautics and Astronautics, Shanghai Jiao Tong University, Shanghai

Peer Reviewed

Global navigation satellite system (GNSS) has been widely used in various positioning applications under multiple challenging scenarios. Vehicle positioning and navigation have become the largest civilian high precision application and attracted much attention in recent years. In an urban canyon, multipath is the main source of GNSS positioning error. The multipath signal arrives at the receiver through multiple propagation paths. The pseudorange errors caused by multipath delays lead to significant positioning errors. The current work mainly focuses on multipath identification, characterization, modeling, and mitigation [1] – [6]. Multipath characterization is the prerequisite of multipath error modelling and mitigation development, which is worthy of in-depth investigation.
The existing work on multipath characterization can be categorized into measurements and simulation based methods. The former requires dynamic or static observations in reality. References [7]-[9] analyzed multipath features, containing carrier phase errors, statistical models of static multipath delay, and frequency domain characteristics by real data processing. This method is accurate in certain specific regions but has difficulties in recognizing short-delay multipath. The simulation-based characterization requires a 3D map assistance. The multipath errors with 3D city model aiding are respectively evaluated in [10]-[12]. This method is a low cost virtual implantation and suitable for long-time or wide-range tests. In addition, the short-delay multipath can be exactly identified in this virtual framework.
Apparently, multipath characteristics are associated with the GNSS satellite geometry and the reflection surface distributions. As the latest developed GNSS system, BeiDou System (BDS) has some specific features, e.g., hybrid constellation deployment on geostationary orbit (GEO), inclined geosynchronous orbit (IGSO), and medium Earth orbit (MEO), and diversified signal modulations. Especially in the east Asia area, the BDS can provide a regional augmentation navigation service with more than 15 visible BDS satellites. The special orbital design of BDS leads to different multipath attributes, including multipath delay distribution, from other GNSS constellations. Such BDS multipath properties and behaviors attracted much attention.
For example, Chen X utilized raw data on BDS B1I signal to establish multipath statistical models and estimate multipath parameters [13]. Wang investigated the multipath code delay characteristics of BDS GEO satellites, using the method of combining observations of multipath [14]. Ye compared the multipath errors from BDS GEO, MEO, and IGSO satellites [15]. Zhang analyzed the BDS multipath repeat periods and proposed a BDS multipath mitigation algorithm in tracking loop design [16].
As mentioned above, these works mainly conducted based on the real observations. Combing the accurate 3D city model to generate and analyze the BDS multipath features have not been covered. In this paper, we applied the 3D ray tracing method to simulate BDS signal propagation paths. The open source toolbox developed by Ollander [12] has been used to carry out the high-density multipath simulation in urban canyons. The Lujiazui CBD in Shanghai is selected as the urban canyon representation in our study. The 3D city model is constructed and three BDS satellite orbits are generated according to the Rinex data file.
The simulation test starts from South Pudong street to Century Avenue with 5km total length. For each building surface, a mirror receiver relative to the surface is generated. If the path between this mirror receiver and the satellite passes through the building surface, the reflection path is considered to exist, and the reflection point is the intersection point. The multipath delay for a sample point is the arithmetic mean of each signal propagation path delay. Receiver position is calculated based on pseudoranges. All sample points are counted to build the distribution of multipath delays and positioning errors. The path generation rate is set to 1 Hz with consideration of computational burden and real implementation. A 12-hour simulation is carried out to calculate the multipath variations from each visible BDS satellites. We characterize multipath for BDS 5 GEO, 8 MEO, and 3 IGSO satellites in terms of code delays and carrier phase errors. In the meanwhile, the relationship between multipath delay distribution or positioning errors and the environmental effects is investigated. The results show that the GEO multipath errors exhibit a stable pattern, with also a smaller average multipath error as compared to MEO and IGSO multipath features. In addition, the multipath delays and positioning errors are highly correlated to environmental features, e.g., buildings height, density, and distributions.
Consequently, taking the advantage of reflection geometry, the distributions of multipath delays and positioning errors are computed. Furthermore, a 12 hour simulation is carried out as it helps in analyzing the fluctuation of the multipath properties due to the movements of satellites.



Return to Session E3