Development of an excavator-avoidance system for buried pipes

• Published in: Advanced robotics Vol. 35; no. 23; pp. 1468 - 1483
• Main Authors: Yajima, Ryosuke, Katsuma, Shinya, Suzuki, Makoto, Matsushita, Fumiya, Hamasaki, Shunsuke, Chun, Pang-jo, Nagatani, Keiji, Yamauchi, Genki, Hashimoto, Takeshi, Yamashita, Atsushi, Asama, Hajime, Ozawa, Kazumasa
• Format: Journal Article
• Language: English
• Published: Taylor & Francis 02.12.2021
• Subjects: Autonomous excavation; buried pipes; deep learning; ground-penetrating radar; underground utility database
• ISSN: 0169-1864; 1568-5535
• DOI: 10.1080/01691864.2021.2007167

Currently, accidents in which hydraulic excavators mistakenly damage buried pipes are becoming a serious challenge. Although prospecting is required to avoid damage to buried pipes, it is a difficult task for workers, and it increases the period required for construction. In this study, we developed an excavator-avoidance system for buried pipes to realize safe and efficient excavation that can prevent damage to buried pipes without prospecting. This system automatically detects buried pipes, generates an excavation path, and autonomously excavates by hydraulic excavators, including the underground utility database. We proposed a detection method using ground-penetrating radar and deep learning, as well as an excavation path generation method in which the autonomous hydraulic excavator is capable of avoiding buried pipes. In addition, we built a database to store information on underground utilities and buried pipes. Experiments were also conducted to validate this system. Accordingly, the position of the buried pipe can be automatically detected from the visualized images of the reflected waves of the ground-penetrating radar. Furthermore, based on the information provided in the database, the autonomous hydraulic excavator could perform excavations without collisions while avoiding the buried pipe.