多源GNSS数据与深度学习驱动的智能手机RTK定位误差预测

doi:10.11947/j.AGCS.2026.20250355

测绘学报 ›› 2026, Vol. 55 ›› Issue (5): 761-775.doi: 10.11947/j.AGCS.2026.20250355

• 北斗/GNSS多源传感器融合PNT • 下一篇

多源GNSS数据与深度学习驱动的智能手机RTK定位误差预测

施闯¹^,²^,³^,⁴(), 陈鑫鑫¹^,²^,³, 王家乐¹^,²^,³^,⁴^,⁵(), 夏鸣¹^,²^,³^,⁴

^1.北京航空航天大学空间与地球科学学院，北京　100191
^2.卫星导航与移动通信融合技术工业和信息化部重点实验室，北京　100191
^3.空地一体新航行系统技术全国重点实验室，北京　100191
^4.北京航空航天大学江西研究院，江西　南昌　330096
^5.香港理工大学土地测量与地理资讯学系，香港　999077

收稿日期:2025-09-04 修回日期:2026-04-30 出版日期:2026-06-23 发布日期:2026-06-23
通讯作者: 王家乐 E-mail:shichuang@buaa.edu.cn;wang_jiale@buaa.edu.cn
作者简介:施闯（1968—），男，博士，教授，中国科学院院士，研究方向为北斗高精度导航定位授时、通信导航融合。　E-mail：shichuang@buaa.edu.cn
基金资助:
北京市自然科学基金(3264040);国家重点研发计划(2024YFB3910103);中国博士后科学基金面上项目(2025M784291)

Multi-source GNSS data and deep learning-driven RTK positioning error prediction for smartphones

Chuang SHI¹^,²^,³^,⁴(), Xinxin CHEN¹^,²^,³, Jiale WANG¹^,²^,³^,⁴^,⁵(), Ming XIA¹^,²^,³^,⁴

^1.School of Space and Earth Sciences, Beihang University, Beijing 100191, China
^2.Key Laboratory of Satellite Navigation and Mobile Communication Fusion Technology, Ministry of Industry and Information Technology, Beijing 100191, China
^3.National Key Laboratory of CNS/ATM, Beijing 100191, China
^4.Jiangxi Research Institute of Beihang University, Nanchang 330096, China
^5.Department of Land Surveying and Geo-Information, The Hong Kong Polytechnic University, Hong Kong 999077, China

Received:2025-09-04 Revised:2026-04-30 Online:2026-06-23 Published:2026-06-23
Contact: Jiale WANG E-mail:shichuang@buaa.edu.cn;wang_jiale@buaa.edu.cn
About author:SHI Chuang (1968—), male, PhD, professor, academician of the Chinese Academy of Science, majors in high-precision BeiDou navigation positioning and timing, and navigation-communication integration.　E-mail: shichuang@buaa.edu.cn
Supported by:
Beijing Natural Science Foundation(3264040);The National Key Research and Development Program of China(2024YFB3910103);General Project of China Postdoctoral Science Foundation(2025M784291)

摘要/Abstract

摘要：

智能手机作为当前普及率最高的低成本GNSS终端，受限于内置线性极化天线的物理特性，其GNSS信号接收易受城市楼宇和树木等复杂环境遮挡干扰，导致观测值存在显著多路径效应、载波相位连续性差等问题，进而引发定位精度大幅退化。为解决这一难题，本文以安卓系统开放的多源观测数据为核心输入，包括GNSS原始测量信息、惯性传感器衍生的姿态偏航角与速度，以及定位解算过程中的伪距残差、位置精度因子等质量指标，在典型城市复杂场景中开展动态数据采集，构建面向智能手机RTK三维定位误差的预测与修正框架。针对定位误差的时空关联性与多特征耦合特性，本文提出融合通道-空间双注意力机制的卷积长短期记忆（CNN-LSTM）神经网络：通过卷积层提取多源特征的空间关联性，LSTM层捕捉误差序列的时间依赖关系，双注意力机制则分别强化关键卫星通道与核心观测特征的权重，实现对复杂环境下误差模式的精准建模。基于小米Mi 8与谷歌Pixel 6两款不同硬件配置的智能手机，在非对称遮挡过渡环境以及遮挡严重环境下的测试结果表明，Mi 8的定位精度分别提升了约49.3%和63.9%；Pixel 6则分别提升了37.5%和47.1%，验证了本文方法在不同硬件终端与复杂场景下的通用性与有效性，为智能手机高精度定位提供了轻量级算法支撑。

关键词: 多源GNSS观测数据, 智能手机RTK, 卷积长短期记忆网络, 定位误差预测

Abstract:

As the most widely used low-cost GNSS terminals at present, smartphones are limited by the physical characteristics of their built-in linearly polarized antennas. Their GNSS signal reception is vulnerable to occlusion interference from complex environments such as urban buildings and trees, leading to significant issues in observations including prominent multipath effects and poor carrier phase continuity, which in turn cause a substantial degradation in positioning accuracy. To address this problem, this study takes multi-source observation data opened by the Android system as the core input, including GNSS raw measurement information, attitude yaw angle and velocity derived from inertial sensors, as well as quality indicators such as pseudorange residuals and position dilution of precision (PDOP) during positioning calculation. Dynamic data collection was conducted in typical complex urban scenarios, and a prediction and correction framework for the 3D real-time kinematic (RTK) positioning error of smartphones was constructed. Considering the spatiotemporal correlation and multi-feature coupling characteristics of positioning errors, this paper proposes a convolutional long short-term memory (CNN-LSTM) neural network integrated with a channel-spatial dual attention mechanism: the convolutional layer extracts the spatial correlation of multi-source features, the LSTM layer captures the temporal dependence of error sequences, and the dual attention mechanism enhances the weights of key satellite channels and core observation features respectively, thereby achieving accurate modeling of error patterns in complex environments. Based on tests conducted with two smartphones of different hardware configurations, the Xiaomi Mi 8 and Google Pixel 6, under asymmetric transition occlusion environments and severe occlusion environments, the results indicate that the positioning accuracy of the Mi 8 was improved by approximately 49.3% and 63.9%, respectively, while the Pixel 6 achieved improvements of 37.5% and 47.1%, respectively. These results verify the universality and effectiveness of the method across different hardware terminals and complex scenarios, providing a lightweight algorithmic support for high-precision positioning of smartphones.

Key words: multi-source GNSS observation data, smartphone RTK, convolutional long short-term memory (CNN-LSTM) network, positioning error prediction

中图分类号:

P228

施闯, 陈鑫鑫, 王家乐, 夏鸣. 多源GNSS数据与深度学习驱动的智能手机RTK定位误差预测[J]. 测绘学报, 2026, 55(5): 761-775.

Chuang SHI, Xinxin CHEN, Jiale WANG, Ming XIA. Multi-source GNSS data and deep learning-driven RTK positioning error prediction for smartphones[J]. Acta Geodaetica et Cartographica Sinica, 2026, 55(5): 761-775.

导出引用管理器 EndNote|Reference Manager|ProCite|BibTeX|RefWorks

链接本文: http://xb.chinasmp.com/CN/10.11947/j.AGCS.2026.20250355

http://xb.chinasmp.com/CN/Y2026/V55/I5/761

图/表 16

表1

图1

图2

表2

图3

图4

图5

图6

图7

图8

表3

图9

图10

图11

表4

图12

参考文献 33

[1]	毕京学, 甄杰, 郭英. Android手机GPS和A-GPS定位精度分析[J]. 测绘通报, 2016(7): 10-13.
	BI Jingxue, ZHEN Jie, GUO Ying. Android phone GPS and A-GPS positioning accuracy analysis[J]. Bulletin of Surveying and Mapping, 2016(7): 10-13.
[2]	王颖喆, 陶贤露, 朱锋, 等. 利用智能手机实现GNSS原始观测值的高精度差分定位[J]. 武汉大学学报(信息科学版), 2021, 46(12): 1941-1950.
	WANG Yingzhe, TAO Xianlu, ZHU Feng, et al. High-precision differential positioning with GNSS raw observations from smartphones[J]. Geomatics and Information Science of Wuhan University, 2021, 46(12): 1941-1950.
[3]	袁良雄, 王浩, 申志恒. 基于扩展天线的智能手机GNSS RTK定位性能研究[J]. 全球定位系统, 2023, 48(3): 77-84.
	YUAN Liangxiong, WANG Hao, SHEN Zhiheng. Research on the GNSS RTK positioning performance of smartphones based on extended antenna[J]. Global Positioning System, 2023, 48(3): 77-84.
[4]	潘宇明, 丁乐乐, 王震. 安卓手机GNSS信号增强对定位精度的影响分析[J]. 测绘通报, 2022(S2): 43-47.
	PAN Yuming, DING Lele, WANG Zhen. Analysis of the influence of GNSS signal enhancement on positioning accuracy of Android phones[J]. Bulletin of Surveying and Mapping, 2022(S2): 43-47.
[5]	高成发, 陈波, 刘永胜. Android智能手机GNSS高精度实时动态定位[J]. 测绘学报, 2021, 50(1): 18-26. DOI: . doi: 10.11947/j.AGCS.2020.20200107
	GAO Chengfa, CHEN Bo, LIU Yongsheng. Android smartphone GNSS high-precision real-time kinematic positioning[J]. Acta Geodaetica et Cartographica Sinica, 2021, 50(1): 18-26. DOI: . doi: 10.11947/j.AGCS.2020.20200107
[6]	TAO X, LIU W, WANG Y, et al. Smartphone RTK positioning with multi-frequency and multi-constellation raw observations: GPS L1/L5, Galileo E1/E5a, BDS B1I/B1C/B2a[J]. Journal of Geodesy, 2023, 97(5): 43.
[7]	李仕辉, 王虎, 马宏阳, 等. 城市环境手机GNSS定位随机模型自适应构建方法[J]. 测绘科学, 2024, 49(3): 8-18.
	LI Shihui, WANG Hu, MA Hongyang, et al. An adaptive stochastic model construction method for smartphone GNSS positioning in urban environments[J]. Science of Surveying and Mapping, 2024, 49(3): 8-18.
[8]	DENG J, WANG H, WEI S, et al. A stochastic model based on optimal satellite subset selection strategy for smartphone pseudorange relative positioning[J]. Sensors, 2024, 24(8): 2598.
[9]	WANG J L, XIA M, ZHANG D, et al. Urban GNSS positioning for consumer electronics: 3D mapping and advanced signal processing[J]. IEEE Transactions on Consumer Electronics, 2025, 71(2): 7059-7072.
[10]	王式太, 张博宇, 殷敏. 基于动态阈值的移动终端GNSS定位阴影匹配算法[J]. 大地测量与地球动力学, 2021, 41(11): 1101-1105.
	WANG Shitai, ZHANG Boyu, YIN Min. A shadow matching algorithm for GNSS positioning of mobile terminals based on dynamic threshold[J]. Journal of Geodesy and Geodynamics, 2021, 41(11): 1101-1105.
[11]	KIM D, JANG M, LEE K, et al. Application of shadow matching technique to improve smartphone-based global navigation satellite system positioning accuracy[J]. Sensors and Materials, 2022, 34: 383.
[12]	GROVES P D. Shadow matching: a new GNSS positioning technique for urban canyons[J]. Journal of Navigation, 2011, 64(3): 417-430.
[13]	冯家昊, 任晓东, 张小红, 等. 一种智能手机GNSS精密定位随机模型构建方法及定位性能分析[J]. 导航定位与授时, 2024, 11(5): 136-144.
	FENG Jiahao, REN Xiaodong, ZHANG Xiaohong, et al. A method for constructing a stochastic model for smartphone GNSS precise positioning and its performance analysis[J]. Navigation, Positioning and Timing, 2024, 11(5): 136-144.
[14]	HUANG X, LI W, DAI Z, et al. Improving smartphone GNSS positioning in challenging urban environments using GA-BPNN[J]. GPS Solutions, 2024, 29(1): 3.
[15]	LIU W, SHI X, ZHU F, et al. Quality analysis of multi-GNSS raw observations and a velocity-aided positioning approach based on smartphones[J]. Advances in Space Research, 2019, 63(8): 2358-2377.
[16]	冷宏宇, 秘金钟, 徐彦田, 等. 智能手机终端RTK定位性能分析[J]. 测绘科学, 2020, 45(12): 15-21.
	LENG Hongyu, MI Jinzhong, XU Yantian, et al. Performance analysis of smartphone RTK positioning[J]. Science of Surveying and Mapping, 2020, 45(12): 15-21.
[17]	WANG Jiale, ZHEN Fu, HU Yong, et al. Instantaneous sub-meter level precise point positioning of low-cost smartphones[J]. NAVIGATION: Journal of the Institute of Navigation, 2023, 70(4): 597.
[18]	ZHANG X, TAO X, ZHU F, et al. Quality assessment of GNSS observations from an Android N smartphone and positioning performance analysis using time-differenced filtering approach[J]. GPS Solutions, 2018, 22(3): 70.
[19]	HÅKANSSON M. Characterization of GNSS observations from a Nexus 9 Android tablet[J]. GPS Solutions, 2018, 23(1): 21.
[20]	王家乐, 郑福, 施闯, 等. 智能手机GNSS数据质量分析及大气增强PPP精度评估[C]//第十三届中国卫星导航年会论文集. 北京: 中国卫星导航系统管理办公室, 2022: 1-9.
	WANG Jiale, ZHENG Fu, SHI Chuang, et al. Analysis of smartphone GNSS data quality and accuracy assessment of atmospheric-augmented PPP[C]//Proceedings of the 13th China Satellite Navigation Conference. Beijing: China Satellite Navigation Office, 2022: 1-9.
[21]	VAZQUEZ-ONTIVEROS J R, MARTINEZ-FELIX C A, MELGAREJO-MORALES A, et al. Assessing the quality of raw GNSS observations and 3D positioning performance using the Xiaomi Mi 8 dual-frequency smartphone in Northwest Mexico[J]. Earth Science Informatics, 2024, 17(1): 21-35.
[22]	赵兴旺, 陶安迪, 陈健, 等. 智能手机GNSS数据质量分析与随机模型建立[J]. 大地测量与地球动力学, 2024, 44(7): 661-666.
	ZHAO Xingwang, TAO Andi, CHEN Jian, et al. Analysis of smartphone GNSS data quality and stochastic model establishment[J]. Journal of Geodesy and Geodynamics, 2024, 44(7): 661-666.
[23]	徐彦田, 刘巍峰, 李玉星, 等. BDS/GPS/GAL智能手机RTK动态定位算法[J]. 无线电工程, 2023, 53(5): 1061-1067.
	XU Yantian, LIU Weifeng, LI Yuxing, et al. BDS/GPS/GAL smartphone RTK dynamic positioning algorithm[J]. Radio Engineering, 2023, 53(5): 1061-1067.
[24]	姚子扬, 尚俊娜, 孙建强, 等. 一种多源融合室内外无缝定位方法[J]. 传感技术学报, 2022, 35(1): 72-83.
	YAO Ziyang, SHANG Junna, SUN Jianqiang, et al. A multi-source fusion seamless indoor and outdoor positioning method[J]. Journal of Transducer Technology, 2022, 35(1): 72-83.
[25]	WANG J L, XIA M, XIE Y, et al. Multimodal sensor fusion-based lightweight pedestrian wearable system for continuous health monitoring and location tracking[J]. IEEE Sensors Journal, 2025, 26(10): 14558-14575.
[26]	PAZIEWSKI J, FORTUNATO M, MAZZONI A, et al. An analysis of multi-GNSS observations tracked by recent Android smartphones and smartphone-only relative positioning results[J]. Measurement. 2021, 175: 109162.
[27]	LI Xingxing, ZHANG Xiaohong, NIU Xiaoji, et al. Progress and achievements of multi-sensor fusion navigation in China during 2019—2023[J]. Journal of Geodesy and Geoinformation Science, 2023, 6(3): 102-114.
[28]	JADERBERG M, SIMONYAN K, ZISSERMAN A. Spatial transformer networks[J]. Advances in Neural Information Processing Systems, 2015, 28.
[29]	ZHU X, CHENG D, ZHANG Z, et al. An empirical study of spatial attention mechanisms in deep networks[C]//Proceedings of 2019 IEEE/CVF International Conference on Computer Vision. Seoul: IEEE, 2019: 6688-6697.
[30]	WANG Jiale, SHI Chuang, ZHENG Fu, et al. Multi-frequency smartphone positioning performance evaluation: insights into A-GNSS PPP-B2b services and beyond[J]. Satellite Navigation, 2024, 5(1): 25.
[31]	YAN D, LI T, JIANG H, et al. PS-VINS: a visual-inertial SLAM system with pedestrian gait and structural constraints using smartphone sensors[J]. IEEE Sensors Journal, 2024, 24(5): 6777-6791.
[32]	WANG Jiale, SHI Chuang, XIA Ming, et al. Seamless indoor-outdoor foot-mounted inertial pedestrian positioning system enhanced by smartphone PPP/3-D Map/Barometer[J]. IEEE Internet of Things Journal, 2024, 11(7): 13051-13069.
[33]	LI Guangcai, GENG Jianghui. Characteristics of raw multi-GNSS measurement error from Google Android smart devices[J]. GPS solutions, 2019, 23(3): 90.

数据源/类名	参数/接口名称	描述	单位
GnssClock类	BiasUncertaintyNanos	描述FullBiasNanos（手机内部时钟与GPST的总偏差）和BiasNanos（GNSS芯片时钟与GPST的亚纳秒级偏差）的不确定度	ns
GnssMeasurement类	ReceivedSvTimeUncertaintyNanos	接收到的卫星的发射时刻不确定度	ns
	AccumulatedDeltaRangeUncertaintyMeters	载波相位不确定度	m
	Cn0DbHz	载噪比	dB-Hz
惯导传感器^[27]	姿态偏航角	描述设备运动方向与正北方向的夹角	°
惯导传感器	平面运动速度	设备在北方向与东方向的运动速度	m/s
后处理	伪距残差	观测伪距与根据定位结果计算的理论伪距之间的差值，反映观测值的拟合程度	m
后处理	PDOP	描述卫星几何构型对定位精度的影响程度	—
后处理	历元间位置差分	相邻两个历元之间计算出的三维位置变化量	m

设备名称	天线	系统	GNSS芯片	星座	频率	载波相位
Mi 8	线极化	安卓12	博通BCM4775	G/R/C/E/J	L1，L5	支持
Pixel 6	线极化	安卓13	博通BCM4776	G/R/C/E/J	L1，L5	支持

设备	指标	模型误差			RTK误差
设备	指标	北向	东向	天向	北向	东向	天向
Mi 8	RMS	4.560 4	6.742 3	5.438 2	10.490 0	11.906 2	11.012 5
	AVE	2.742 5	1.065 1	－3.270 4	－8.680 5	－8.823 9	－9.829 4
	STD	3.665 0	6.696 7	4.370 4	5.924 1	8.040 4	4.994 7
	MAX	10.700 5	12.200 4	12.760 0	16.160 6	21.701 1	20.776 4
Pixel 6	RMS	1.460 1	7.807 1	2.162 7	5.423 1	4.487 8	11.139 5
	AVE	0.275 6	－0.436 4	－1.288 3	3.317 2	1.163 3	5.839 5
	STD	1.442 3	7.840 6	1.747 3	4.315 5	4.359 9	9.541 9
	MAX	4.960 7	19.192 4	5.187 4	11.559 0	11.098 5	19.844 0

设备	指标	模型误差			RTK误差
设备	指标	北向	东向	天向	北向	东向	天向
Mi 8	RMS	5.912 5	3.404 0	7.050 7	10.531 8	4.433 0	24.622 4
	AVE	－2.034 9	－2.229 1	2.426 6	1.014 2	－3.367 4	21.002 2
	STD	5.589 2	2.590 2	6.665 1	10.554 4	2.902 7	12.939 6
	MAX	24.639 9	11.563 4	29.383 3	23.696 6	13.401 8	46.614 2
Pixel 6	RMS	9.164 4	6.814 8	10.749 7	9.812 3	7.094 4	27.085 3
	AVE	－3.633 2	4.771 9	4.538 0	6.469 1	1.766 3	23.347 4
	STD	8.467 2	4.896 3	9.807 1	7.424 9	6.914 9	13.817 6
	MAX	29.094 4	18.462 4	33.214 6	22.139 8	17.663 8	66.107 3

多源GNSS数据与深度学习驱动的智能手机RTK定位误差预测

Multi-source GNSS data and deep learning-driven RTK positioning error prediction for smartphones

RichHTML

PDF

可视化

摘要/Abstract

引用本文

使用本文

图/表 16

参考文献 33

相关文章 15

编辑推荐

Metrics

本文评价

[1]	李博峰, 章浙涛. 非模型化误差实时处理：探测、补偿及控制[J]. 测绘学报, 2026, 55(5): 776-786.
[2]	耿江辉, 王锋. 多源传感器“盲插即用”的组件化PNT融合框架、原理及试验验证[J]. 测绘学报, 2026, 55(5): 787-797.
[3]	朱锋, 廖元木, 周瑞, 张小红. 多传感器融合导航中的连续时间状态估计方法[J]. 测绘学报, 2026, 55(5): 798-808.
[4]	姚凌寒, 徐天河, 丛阳滋, 张震, 邢建平. 基于IESRKF的LiDAR/UWB/INS多源传感器紧耦合模型[J]. 测绘学报, 2026, 55(5): 809-825.
[5]	冯腾飞. 华北水储量变化与非构造形变的GRACE多源数据处理及信号分析[J]. 测绘学报, 2026, 55(5): 943-943.
[6]	李贤炮. 联合GNSS和GRACE/GFO数据反演陆地水储量变化方法及应用[J]. 测绘学报, 2026, 55(5): 944-944.
[7]	洪菊. GNSS精密单点定位增强技术研究[J]. 测绘学报, 2026, 55(5): 945-945.
[8]	周洁. 基于多源异构数据的城市新老城区地震应急避难场所空间适宜性比较研究[J]. 测绘学报, 2026, 55(5): 949-949.
[9]	祝程程, 李真, 郭金运, 周茂盛, 李婉秋. 基于限定区域的纬度做差法解算宽刈幅测高数据交叉点位置[J]. 测绘学报, 2026, 55(4): 673-683.
[10]	杨小磊, 肖云, 杨丽清, 洪晓东, 郭恩泽, 王翰. 时变重力场反演的改进基线法[J]. 测绘学报, 2026, 55(4): 708-720.
[11]	李洋, 黄海军, 刘素兰, 吴晓辉, 刘琦, 庞琪沛, 吴云龙. 鄱阳湖流域未来百年陆地水储量异常预测及干旱风险评估：基于CMIP6驱动的Transformer-GRU模型[J]. 测绘学报, 2026, 55(4): 721-738.
[12]	杨先圣. 高可用可见光智能定位技术研究[J]. 测绘学报, 2026, 55(4): 753-753.
[13]	耿继军. 基于智能手机多传感器融合的行人室内外定位研究[J]. 测绘学报, 2026, 55(4): 754-754.
[14]	马宏阳. 对流层模型增强的PPP-AR理论方法及在机载导航中的应用研究[J]. 测绘学报, 2026, 55(4): 756-756.
[15]	刘航. 联合地基GNSS和星载GNSS-R的低纬电离层闪烁监测及预报方法研究[J]. 测绘学报, 2026, 55(4): 758-758.