Payload Estimation for Hydraulic Excavators using a Depth Camera
Hiroshi Yoshida, Tatsuya Yoshimoto · 2021
We propose a payload estimation system for hydraulic excavators. The proposed system achieves the following advantages: (1) it can be externally mounted on existing excavators, (2) it is relatively inexpensive, and (3) it achieves high accuracy. To realize these advantages, we install a depth camera externally on the inside of an excavator’s arm and use a method to determine the depth of the soil in the bucket from a bird’s eye view. Moreover, we utilize kernel ridge regression, which is a machine learning technique, to achieve highly accurate estimation. Furthermore, we enhance the method for small- scale training datasets by combining it with geometrical volume calculation.