Intelligent loading and unloading stages of loader based on data-driven control

Abstract

Dynamically changing state variables increase the complexity of the loader during the Loading and Unloading Stages (LUSs). These variables include the height of the truck, the position of the stockpile, and the attitude of the working device. This paper proposes a novel data-driven control strategy to automate the complex, dynamic LUSs of loaders. Specifically, data-driven control strategies automatically lift, flip, and lower the bucket. These strategies ensure that the task requirements of different operation stages are correctly completed. Experimental validation revealed a control deviation in the attitude of the bucket. Based on the analysis, it was clarified that the lag of the control signal was the main factor causing the deviation. Subsequently, this paper proposes a control strategy based on Machine Learning (ML) that maintains error within 0.4% when the engine speed is relatively stable. This paper provides a clear, intelligent operation realization scheme for earthmoving machines.