2016 Fiscal Year Annual Research Report
Sensor Integration for Autonomous Vehicle Self-Localization in Urban City
| Project/Area Number |
16F16350
|
|---|
| Research Institution | The University of Tokyo |
|---|
Principal Investigator |
上條 俊介 東京大学, 大学院情報学環・学際情報学府, 准教授 (70334357)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
GU YANLEI 東京大学, 大学院情報学環・学際情報学府, 外国人特別研究員
|
|---|
| Project Period (FY) |
2016-11-07 – 2019-03-31
|
| Keywords | Autonomous driving / Vehicle localization / Sensor fusion |
|---|
| Outline of Annual Research Achievements |
Vehicle self-localization in urban canyon is a significant and challenging issue for Intelligent Transportation Systems. Integration of multiple sensors is one of possible solutions for the accurate localization.The most popular sensor is Velodyne, which is a Light Detection and Ranging (LiDAR) sensor. We firstly focused on LiDAR based matching with a pre-prepared map. The average error of the developed LiDAR based localization system is less than 20cm in general traffic situations.We also developed a low cost localization system, which integrates the 3D building map aided GNSS positioning technique, inertial sensors and vision sensors. The system can achieve sub-meter accuracy with respect to positioning error mean, and has 95% correct lane rate in the localization in the urban city.
|
| Current Status of Research Progress |
Current Status of Research Progress
2: Research has progressed on the whole more than it was originally planned.
Reason
1.In the past half year, we have developed the LiDAR based localization algorithm, and tested this proposed algorithm in urban city. The average localization error of the developed system is less than 20cm, which has proved the effectiveness of the LiDAR sensor for vehicle localization.
2.We also started the development of the integrated localization system in advance, in order to prepare for next phase of this research. The GNSS positioning technique, inertial sensors and monocular camera are integrated by the sensor fusion technique for vehicle self-localization. This system achieves sub-meter accuracy with respect to positioning error mean, and has 95% correct lane rate in the localization.
|
| Strategy for Future Research Activity |
The pre-developed localization system, including GNSS positioning technique, inertial sensors and monocular camera, has shown the effectiveness for localization. To reduce the positioning error, the system detects lane changing and keeping behavior from the camera when lane marking is visible. In the next phase, the visual odometry or other marking detection technique will be introduced into the system to improve the localization accuracy in the intersection area. Moreover, this system will be integrated with LiDAR based localization to achieve higher accuracy and better reliability.
In addition, we will develop the precise pedestrian positioning and navigation system using sensor fusion technique based on smart glasses or smartphone with camera, GNSS receiver and inertial sensor.
|
