| Project/Area Number |
20K19893
|
|---|
| Research Category |
Grant-in-Aid for Early-Career Scientists
|
| Allocation Type | Multi-year Fund |
|---|
| Review Section |
Basic Section 61050:Intelligent robotics-related
|
|---|
| Research Institution | Kanazawa University |
|---|
Principal Investigator |
|
|---|
| Project Period (FY) |
2020-04-01 – 2023-03-31
|
| Project Status |
Completed (Fiscal Year 2022)
|
| Budget Amount *help |
¥4,290,000 (Direct Cost: ¥3,300,000、Indirect Cost: ¥990,000)
Fiscal Year 2021: ¥2,340,000 (Direct Cost: ¥1,800,000、Indirect Cost: ¥540,000)
Fiscal Year 2020: ¥1,950,000 (Direct Cost: ¥1,500,000、Indirect Cost: ¥450,000)
|
|---|
| Keywords | Graph SLAM / Autonomous Vehicles / Mapping Systems / LIDAR / LIDAR Mapping Systems / Largescal Map Generation / 2.5D Elevation Maps / SLAM in Multilevel Roads / Graph Slam / LIDAR Maps / 2D Intensity Maps / LIDAR SLAM / Mapping System / SLAM / Elevation Maps |
|---|
| Outline of Research at the Start |
1.Generating accurate road surface maps based on Graph Slam technology.
2.Enhancing the relative-position estimation of the map data in environments of long tunnels, dense trees and high buildings.
3.Combining map data to generate large-scale maps.
4.Increasing the global position accuracy of the maps.
|
| Outline of Final Research Achievements |
We are glad to report that the goal of the project has been achieved scientifically and practically with very impactful and impressive results. We analyzed the reasons of generating low quality maps using GNSS/INS-RTK systems and studied the influences on the localization accuracy during autonomous driving. A unique Graph-SLAM framework has been designed and implemented to operate in the image domain instead of the conventional methods in the 3D point cloud domain. The obtained results have outperformed accurate/expensive GNSS/INS-RTK systems to generate accurate maps in challenging environments such as long tunnels, high buildings, dense trees, underpasses, bridges and so on. Accordingly, the mapping module has become very robust to generate accurate maps for autonomous vehicles regardless of the complexity of the road structure in the XY plane as well as extend the map automatically to contain new areas with coherent representation of the road surface in the global coordinate system.
|
| Academic Significance and Societal Importance of the Research Achievements |
Precise Mapping of challenging environments is very important to commercialize autonomous vehicles and this research stands to safely conduct hand-free mapping with taking into account the driving scenarios, traffic flows, road topological representations and sensor configurations.
|
