Predictive modeling of loader's working resistance measurement based on multi-sourced parameter data

Abstract

Accurate measurement of loader's working resistance is crucial for autonomous intelligence and energy-saving optimization. This study highlights the limitations of strain sensors used in traditional measurements, such as difficult installation, difficult calibration, and short service life, and proposes a predictive modeling method for the soft measurement of loader's working resistance. First, the characterization parameters of the soft measurement model for predicting the working resistance are determined through theoretical and model simulations. Subsequently, the maximum information coefficient is used for feature selection. Second, a prediction model is developed using the random forest regression algorithm optimized through a grid search. Finally, the data under sand and loose-soil conditions are used to validate the prediction and generalization performances of the prediction model. The results show that the soft measurement model proposed in this study can effectively and accurately characterize the loader's working resistance and has significance in theoretical research and engineering applications.