Spatial-Temporal Federated Transfer Learning with multi-sensor data fusion for cooperative positioning

被引:27
|
作者
Zhou, Xiaokang [1 ,8 ]
Yang, Qiuyue [2 ]
Liu, Qiang [3 ]
Liang, Wei [2 ]
Wang, Kevin [4 ,6 ]
Liu, Zhi [5 ]
Ma, Jianhua
Jin, Qun [7 ]
机构
[1] Shiga Univ, Fac Data Sci, Hikone 5228522, Japan
[2] Hunan Univ Technol & Business, Changsha Social Lab Artificial Intelligence, Changsha 410205, Peoples R China
[3] Hunan Univ Technol & Business, Comp Sci Inst, Changsha 410205, Peoples R China
[4] Univ Auckland, Dept Elect Comp & Software Engn, Auckland 1010, New Zealand
[5] Univ Electrocommun, Dept Comp & Network Engn, Tokyo, Japan
[6] Hosei Univ, Fac Comp & Informat Sci, Chiyoda Ku, Tokyo 1028160, Japan
[7] Waseda Univ, Fac Human Sci, Tokorozawa 3591192, Japan
[8] RIKEN, RIKEN Ctr Adv Intelligence Project, Tokyo 1030027, Japan
基金
日本学术振兴会;
关键词
Federated Transfer Learning; Siamese Network; Spatial-Temporal Clustering; Data Fusion; Cooperative Positioning; NEURAL-NETWORK; SIAMESE;
D O I
10.1016/j.inffus.2023.102182
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
With the development of advanced embedded and communication systems, location information has become crucial factor in supporting context -aware or location -aware intelligent services. Among these services, modern Intelligent Transportation System (ITS) has the strictest requirements for real-time, accurate, and private location data. In this study, a Spatial -Temporal Federated Transfer Learning (ST-FTL) framework is designed and introduced to achieve more precise cooperative positioning with multi -sensor data fusion while protecting the location data privacy in urban ITS. Specifically, a three -layer FTL architecture is constructed to enhance the prediction accuracy on GPS positioning errors for vehicles in different regions especially when facing missing local data in some specific scenarios (e.g., urban canyons), in which Transfer Learning (TL) is incorporated to optimize the initialization of global model with faster convergence but less communication cost in Federated Learning (FL). A multi -attribute based spatial -temporal clustering algorithm is developed to facilitate the finding of more appropriate source domains similar to the target domain in a density -based scheme, while convolutional -gated unit is newly designed to further filter out the useless features and add new features based on the convolution operations and gating mechanism, resulting in more effective global model initialization and weight aggregation from cross -region selection. A multi -sensor data fusion model is built in local to improve the prediction accuracy on positioning errors, in which an improved time -aware asymmetric attention mechanism is involved to selectively adjust the weight importance of the incorporated GPS and Inertial Measurement Unit (IMU) data in a more targeted fusion process. Additionally, an improved Siamese network structure, which breaks the traditional dual network structure and only adopts one single network structure, is leveraged to realize the lightweight data augmentation based on the better utilization of historical similar data from local vehicles themselves, when suffering from insufficient training data. Experiments and evaluations based on two different public datasets demonstrate the outstanding performance of our proposed model and method in achieving superior prediction accuracy and faster convergence for cooperative positioning compared with other state-of-the-art positioning methods in urban ITS.
引用
收藏
页数:14
相关论文
共 50 条
  • [1] Federated Filter for Multi-Sensor Data Fusion of Dynamic Positioning Ship
    Shi, Xiaocheng
    Sun, Xingyan
    Fu, Mingyu
    Chen, Youzhen
    Xie, Wenbo
    2012 IEEE INTERNATIONAL CONFERENCE ON AUTOMATION AND LOGISTICS (ICAL), 2012, : 13 - 18
  • [2] Cooperative vehicle positioning with multi-sensor data fusion and vehicular communications
    Md. Anowar Hossain
    Ibrahim Elshafiey
    Abdulhameed Al-Sanie
    Wireless Networks, 2019, 25 : 1403 - 1413
  • [3] Cooperative vehicle positioning with multi-sensor data fusion and vehicular communications
    Hossain, Md. Anowar
    Elshafiey, Ibrahim
    Al-Sanie, Abdulhameed
    WIRELESS NETWORKS, 2019, 25 (03) : 1403 - 1413
  • [4] Indoor multi-sensor fusion positioning based on federated filtering
    Li, HuiXia
    Ao, LongHui
    Guo, Hang
    Yan, XiaoYi
    MEASUREMENT, 2020, 154
  • [5] Support degree and adaptive weighted spatial-temporal fusion algorithm of multi-sensor
    Liu, Yuanze
    Zhang, Jiawei
    Li, Mingbao
    2010 CHINESE CONTROL AND DECISION CONFERENCE, VOLS 1-5, 2010, : 4129 - 4133
  • [6] Multi-sensor Data Fusion Diagnosis Method Based on Interpretable Spatial-temporal Graph Convolutional Network
    Wen, Kairu
    Chen, Zhuyun
    Huang, Ruyi
    Li, Dongpeng
    Li, Weihua
    Jixie Gongcheng Xuebao/Journal of Mechanical Engineering, 2024, 60 (12): : 158 - 167
  • [7] Integrity for Multi-Sensor Cooperative Positioning
    Xiong, Jun
    Cheong, Joon Wayn
    Xiong, Zhi
    Dempster, Andrew G.
    Tian, Shiwei
    Wang, Rong
    IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS, 2021, 22 (02) : 792 - 807
  • [8] Federated IMM-UKF Algorithm for Multi-sensor Data Fusion
    Xu, Tianlai
    MATERIALS PROCESSING AND MANUFACTURING III, PTS 1-4, 2013, 753-755 : 2117 - 2120
  • [9] UAV Positioning Based on Multi-Sensor Fusion
    Peng, Jing
    Zhang, Ping
    Zheng, Lanxiang
    Tan, Jia
    IEEE ACCESS, 2020, 8 : 34455 - 34467
  • [10] Multi-sensor data fusion via federated dual-band filter
    Fong, Li-Wei
    PROCEEDINGS OF 2008 IEEE INTERNATIONAL CONFERENCE ON NETWORKING, SENSING AND CONTROL, VOLS 1 AND 2, 2008, : 1059 - 1064