Lane-Level Vehicular Localization Utilizing Smartphones
- 20 March 2017
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-5
Abstract
Lane-level vehicular localization has been regarded as a critical technology component for future vehicle navigation services. Current lane-level vehicular localization systems require dedicated devices, making the systems difficult to popularize. Moreover, most systems depend heavily on continuous accurate GPS data, which may be interfered under specific environments. Efficient lane-level map building is another problem to deal with. In this paper, we propose an integrated system with the capability of map building and lane- level localization using smartphones. This system employs a crowdsourcing-based approach to collect information from multiple sensors (including GPS, orientation sensor and acceleration sensor), such as lane changes and turns. Based on the information, a lane localization schemes is designed using the tool of machine learning. The experimental results show that the proposed system achieves high accuracy of map building and lane-level localization.Keywords
This publication has 10 references indexed in Scilit:
- Map-aided localization with lateral perception2014
- Reliable Vehicle Pose Estimation Using Vision and a Single-Track ModelIEEE Transactions on Intelligent Transportation Systems, 2014
- Fusion of sensor signals for navigation on an unmanned land vehicle prototype2014
- A lane boundary detection method based on high dynamic range image2012
- Terrain-based road vehicle localization on multi-lane highways2009
- Towards a Driver Model: Preliminary Study of Lane Change Behavior2008
- Using Support Vector Machines for Lane-Change DetectionProceedings of the Human Factors and Ergonomics Society Annual Meeting, 2005
- Lane detection using color-based segmentation2005
- A new multi-lanes detection using multi-camera for robust vehicle location2005
- Mining GPS Traces for Map RefinementData Mining and Knowledge Discovery, 2004