Estimating Ground Conditions for Efficient Operation of Robotic Lawn Mowers
-
- Zushida Kazuki
- Department of Mechanical Science and Technology Graduate School of Science and Technology, Gunma University
-
- Zhang Haohao
- Department of Mechanical Science and Technology Graduate School of Science and Technology, Gunma University
-
- Kojima Masaki
- Research and Development Division, Nippon Mobility Inc.
-
- Motegi Kazuhiro
- Department of Mechanical Science and Technology Graduate School of Science and Technology, Gunma University
-
- Shiraishi Yoichi
- Department of Mechanical Science and Technology Graduate School of Science and Technology, Gunma University
-
- Shimamura Hideaki
- Life Creation Center, Honda R&D Co., Ltd.
-
- Yamada So
- Life Creation Center, Honda R&D Co., Ltd.
Bibliographic Information
- Other Title
-
- ロボット芝刈り機における効率的な作業を行うための地面状態の推定
Abstract
<p>The objective of this research is to estimate the length of the grass and the condition of the ground when a robotic lawn mower runs using the Random Forest and Neural Network, and to change the running control according to the ground conditions for efficient lawn mowing. The main robotic lawn mowers currently on the market cannot recognize the length of the grass or the condition of the ground other than the lawn, such as bare soil. Therefore, the robotic lawn mower always rotates its blade at the maximum speed while running randomly over the specified area. In order to mow the lawn efficiently, we propose to estimate the length of the grass and the ground condition using sensors and machine learning, and to control the mower appropriately. In this study, we experimentally show that the system achieves a correct answer rate of more than 90% in estimating the ground conditions.</p>
Journal
-
- Journal of The Japan Institute of Electronics Packaging
-
Journal of The Japan Institute of Electronics Packaging 25 (3), 240-249, 2022-05-01
The Japan Institute of Electronics Packaging
- Tweet
Details 詳細情報について
-
- CRID
- 1390573407666398336
-
- ISSN
- 1884121X
- 13439677
-
- Text Lang
- ja
-
- Data Source
-
- JaLC
- Crossref
-
- Abstract License Flag
- Disallowed