Global self-localization and navigation using 3D-LiDAR for headland turning in vineyards
When using the real-time kinematic global navigation satellite system (RTK-GNSS) in orchards or vineyards, autonomous navigation can be disrupted due to the inability to obtain ambiguity-fixed solutions. In this study, we employed light detection and ranging in a three-dimensional space (3D-LiDAR) f...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-12-01
|
| Series: | Smart Agricultural Technology |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2772375525005271 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | When using the real-time kinematic global navigation satellite system (RTK-GNSS) in orchards or vineyards, autonomous navigation can be disrupted due to the inability to obtain ambiguity-fixed solutions. In this study, we employed light detection and ranging in a three-dimensional space (3D-LiDAR) for global self-localization only during headland turning in a vineyard without RTK-GNSS in real time. The RTK-GNSS was only used in advance to create a map of structural poles. The 3D-LiDAR information was used to detect and track poles at the ends of rows of grapevines in headlands. The global self-localization was conducted based on the distance between the poles and the vehicle, combined with pole positions obtained in advance via RTK-GNSS. Experimental tests were conducted through growing seasons in a commercial vineyard by two different vehicles at two different sites. The results showed a 0.101 m root mean square error (RMSE) between the estimated position and the ground truth acquired through the RTK-GNSS. We discussed the RMSE by calculating horizontal dilution of precision (HDOP), and it was found that the accuracy was improved to 0.091 m on average under the condition that the HDOP was less than 20. An algorithm for robust localization, missing pole detection, was suggested and verified to improve accuracy by 14 %. Additionally, the travel results revealed a 0.075 m lateral RMSE between the trajectory navigated by the RTK-GNSS and that by the 3D-LiDAR. These findings suggest that using 3D-LiDAR as a navigation sensor can serve as a viable alternative to RTK-GNSS for the turning in headlands of autonomous vehicles in vineyards. |
|---|---|
| ISSN: | 2772-3755 |