Autonomous Navigation Based on Pattern Matching Using 3D Sensors
Résumé
The significant impact of conventional agriculture on climate change is likely to be reduced through more sustainable practices, such as agroecology. These emerging practices require to adapt the machines used to perform agricultural tasks. Robotics then arises as a promising solution to alleviate these constraints by carrying autonomous work. These machines require precise guidance in the highly variable agricultural environments. It is achieved by path tracking algorithms using the Global Navigation Satellite System (GNSS), currently the standard in the industry. However, GNSS-based autonomous navigation is subject to signalloss and does not provide plant-referenced control for the vehicles. Researchers have developed new vision-based navigation strategies to ensure precise and adaptive tracking of crop rows. The sensitivity to environmental conditions (illumination, rain, dust) of cameras has motivated the development of lidar-based navigation systems. This paper presents an adaptation of a path detection algorithm, from 2D lidar tilted with the ground to 3D sensors. The impact on navigation of the previous and new algorithms are compared with a GNSS path tracking algorithm over the recorded RTK-GNSS path associated with a longitudinal strip cropping configuration of beans and bare soil. The performances are compared in regards of the mentioned algorithms interfaced with different sensors on the same robot with the same generic control law.
Fichier principal
Autonomous_navigation_based_on_pattern_matching_using_3D_sensors.pdf (25.6 Mo)
Télécharger le fichier
Origine | Fichiers produits par l'(les) auteur(s) |
---|---|
licence |