| 摘要: |
| 针对车辆在GNSS拒止条件下的定位需求，提出了一种基于多源异构特征信息联合匹配的自主定位方法，并从数据库建立和匹配方法2个方面分别进行了分析研究。提出了基于众包数据的特征信息库建立方法，通过地磁、WLAN、蓝牙、蜂窝网等信息对众包数据进行分段与聚类，利用提出的DBSCAN算法进行地图重构；此外，还提出了基于HMM模型的异构特征联合匹配定位方法，可将多种传感器提供的观测量以特征方式进行联合匹配。试验结果表明，该方法可以很好地解决GNSS拒止条件下的车辆自主定位问题。 |
| 关键词: 众包 自主定位 HMM 异构融合 |
| DOI： |
|
| 基金项目:中国科学院光电研究院创新项目（Y80B07A1BY） |
|
| Vehicle Autonomous Positioning Method Based on Heterogeneous Feature Matching |
| LI Wen,WEI Dong-yan,LU Yi,WANG Xiang-dong,YUAN Hong |
| (Academy of Opto-Electronics, Chinese Academy of Sciences, Beijing 100094, China; University of Chinese Academy of Sciences, Beijing 100049, China;Beijing Institute of Space Launch Technology, Beijing 100076, China) |
| Abstract: |
| Aiming at the positioning requirements of vehicles in GNSS denied conditions, an autonomous positioning method based on the jointly matching of multi-source heterogeneous feature information is proposed and analysed in two important aspects of database establishment and feature matching. The paper raises a method for building a feature information database based on crowdsourcing data, by which the crowdsourcing data is segmented and clustered through geomagnetic, WLAN, Bluetooth, cellular and other information, and further clustered using DBSCAN algorithm for map reconstruction. In addition, the paper proposes a heterogeneous feature joint matching localization method based on HMM model, which can jointly match the characteristic of observations provided by various sensors. The experimental analysis results show that the proposed method can solve the problem of vehicle autonomous positioning in GNSS denied conditions. |
| Key words: Crowdsourcing Autonomous positioning HMM Heterogeneous feature fusion |
