Payload Estimation in Excavators Using a Machine Learning Based Parameter Identification Method

Abstract

This paper documents a part aspect of a broader work, where the goal is to develop a self-sufficient method for continuous and dynamic payload estimation for hydraulic excavators. Self-sufficiency here implies that the required unknowns are either measurable or can be identified using simple sensors. Results from field tests have showed that the approach for identification in its basic form is easy to implement which comes at the cost of diminished estimation accuracy, especially concerning the moment of inertia and friction behavior. Specifically, this paper covers the work done regarding application of machine learning methods to improve the accuracy and reliability of the identification approach, thereby consequently improving the accuracy of the payload estimation.