Return to Session E3

Session E3: All-source Intelligent PNT Method

Navigation Domain Multipath Characterization using GNSS Direct Position Estimation in Urban Canyon Environment
Jihong Huang, Rui Sun, Rong Yang, Xingqun Zhan, School of Aeronautics and Astronautics, Shanghai Jiao Tong University; Wantong Chen, Tianjin Key Lab for Advanced Signal Processing, Civil Aviation University of China

With the development of Global Navigation Satellite System (GNSS), the all-weather high precision navigation becomes reality. However, the GNSS is likely to suffer performance degradation due to the presence of multipath fading in the challenging environments, e.g., urban canyon, indoor, boulevard and tunnel [1]. Multiple Non-Line of Sight (NLOS) signals reflected by the nearby constructions, arrive at the GNSS receiver antenna, would significantly distort the direct Line of Sight (LOS) signals. In some extreme cases, only NLOS signals exist, and the LOS signal can be completely blocked. Such signal interruption or interference may cause code and carrier tracking errors, which leads to the cycle slips and loss of lock, and eventually degrades the GNSS positioning accuracy and robustness.
Multipath interference is one of the main factors affecting the GNSS performance and has attracted much attentions. Due to the multipath fading with random temporal and spatial futures, the context-aware capability of multipath mitigation algorithms and proper understanding of multipath characteristics in different environments are of great importance [2].
A great effort focuses on the characterization, detection, and mitigation of the multipath fading. Bellaed et al. [3] separated the LOS signal from the composite multipath signals under static and low dynamic condition based on the correlation map. Xie et al. [4] used long coherent time in correlation map to characterize the distribution of multipath in urban canyon. Strode et al. [5] combined three-frequency carrier-to-noise ratio (C/N0) measurements to detect the multipath. Wang et al. [6] utilized support vector machine (SVM) to classify the multipath scenarios using a vector including mean, standard deviation, blockage coefficient, geometric dilution of precision (GDOP) expansion ratios and strength fluctuations. Yang et al. [1] refer the ionospheric scintillation indexes to characterize the fast-varying multipath scenarios such as the frequency selective fading in the urban canyon.
However, these methods have some specific limits. The correlation map method with long coherent integration time is vulnerable to be utilized in dynamic circumstance. The C/N0, satellite elevations and GDOP parameters are over-optimistic to delicately characterize the multipath feature. And it is difficult to maintain lock under severe multipath fading to calculate the indexes such as S4. It can be seen that, the traditional multipath characterization methods utilize the pseudorange measurements of multiple satellites via acquisition and tracking is essentially based on the "pseudorange domain" processing strategy. The independent tracking architecture in each channel leads to a separated signal processing of different satellites, and the inner-channel correction is basically ignored. Therefore, in the traditional methods, the degraded satellite signals influenced by multipath fading are difficult to be acquired and cannot be accurately characterized.
To solve the problems, an open loop GNSS direct position estimation (DPE) algorithm based on collective detection [7] is utilized in this paper to characterize the multipath interference in navigation domain. DPE is an algorithm calculates the position directly from the signal correlators. The theoretically optimal estimation of navigation parameters in DPE has been proved in [8] and [9]. Gao et al. utilized the NLOS signal into DPE to improve the receiver robustness [10]. Cheong et al. focused on the collect detection DPE in GNSS receivers [11]. The collective detection theory directly projects the correlation energy of each satellite signal into the navigation domain and DPE can be performed based on the maximum likelihood criterion, regardless of the signal is successfully acquired, tracked or not. Therefore, compared with traditional methods, the GNSS DPE algorithm can significantly improve the positioning performance under multipath fading which also be helpful for multipath characterization.
In this paper, the DPE multipath characterization method is proposed based on the "correlation domain" and "navigation domain". In the correlation domain, the LOS signal correlation peak is distinct due to the combination of multi-satellite energies. After mapping the LOS correlation peak back to each satellite, the NLOS correlation peaks can be re-constructed separately. In the navigation domain, the multipath characterization obtained in correlation domain can be intuitively linked to the final positioning results. The common feature can be directly projected to the distance and the real geographical environment can be analyzed for further study.
The GPS L1 live data collected from the road test near Lujiazui CBD area in Shanghai is used in our work. The results show that the DPE utilized all visible satellites to complete the maximum likelihood estimation. The field test shows that the disturbed satellite signals can be effectively involved in the positioning solution, which improves the positioning performance under the severe multipath influence. Based on the field test, the multipath characteristic in the urban environment is given. The correlation LOS and NLOS correlation peaks are separated and the multipath distribution is identified. C/N0, code delay and Doppler measurements are obtained in correlation domain. The results show that for a single satellite, the correlator peak with the largest magnitude does not always correspond to LOS. However, the LOS peak energies are improved after utilizing DPE algorithm. In addition, the navigation domain results confirm the 3D projection of position and velocity errors. The NLOS peaks are direction-dependent with the satellite geometry, and the multipath Doppler errors are related to the receiver’s velocity vector.
Reference:
[1] Yang R, Zhan X, Huang J. Characterization and mitigation of multipath fading on multi-frequency GNSS signals in urban environment[J]. Aerospace Systems, 2021, 4(1): 19-27.
[2] Chen X. Statistical multipath model comparative analysis of different GNSS orbits in static urban canyon environment[J]. Advances in Space Research, 2018, 62(5): 1034-1048.
[3] Bellad V, Petovello M G. Indoor multipath characterization and separation using distortions in GPS receiver correlation peaks[C]//Proceedings of the 26th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2013). 2013: 3564-3576.
[4] Xie P, Petovello M G. Measuring GNSS multipath distributions in urban canyon environments[J]. IEEE Transactions on Instrumentation and Measurement, 2014, 64(2): 366-377.
[5] Strode P R R, Groves P D. GNSS multipath detection using three-frequency signal-to-noise measurements[J]. GPS solutions, 2016, 20(3): 399-412.
[6] Wang Y, Liu P, Liu Q, et al. Urban environment recognition based on the GNSS signal characteristics[J]. NAVIGATION, Journal of the Institute of Navigation, 2019, 66(1): 211-225.
[7] Chen W, Wang Z, Liu Q. The new method for GPS direction position estimation based on collection[J]. Aerospace Control, 2019(3):6.
[8] Closas P, Fernández-Prades C, Fernández-Rubio JA. Maximum likelihood estimation of position in GNSS. IEEE Signal Processing Letters. 2007 Apr 16;14(5):359-62.
[9] Dampf J, Frankl K, Pany T. Optimal Particle Filter Weight for Bayesian Direct Position Estimation in a GNSS Receiver [J]. Sensors, 2018, 18(8), 2736.
[10] Ng Y, Gao GX. Direct position estimation utilizing non-line-of-sight (NLOS) GPS signals. In Proceedings of the 29th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2016) 2016 Sep 16 (pp. 1279-1284).
[11] Li L, Cheong JW, Wu J, Dempster AG. Improvement to multi-resolution collective detection in GNSS receivers. The Journal of Navigation. 2014 Mar;67(2):277-93.



Return to Session E3