What question did this study set out to answer?

The research aims to evaluate driving behavior characteristics of articulated equipment in underground mines using XGBoost.

March 13, 2026Open Access

Evaluation of Articulated Equipment Driving Behaviors in Narrow Confined Roadways of Underground Mines Based on XGBoost

Key Points

The research aims to evaluate driving behavior characteristics of articulated equipment in underground mines using XGBoost.
Developed a simulator using Unity3D to replicate articulated equipment driving in confined roadways.
Established an evaluation index system for driving behavior characteristics of articulated equipment.
Utilized XGBoost for performance evaluation of driving behaviors with 2,660 sample data points.
Applied SHAP for analyzing key influencing factors of the driving behavior model.
Achieved a coefficient of determination (R²) of 0.975, indicating high prediction accuracy.
Identified spatial safety distances and articulated angles as the dominant factors influencing driving behavior.
Observed significant feature interactions that provide insights for safety management and risk control.

Abstract

● Unity3D was used to develop a simulator of articulated equipment (AE) in underground mines. ● The evaluation index system for the driving behavior (DB) characteristics of AE was established. ● Extreme Gradient Boosting (XGBoost) was employed to evaluate the driving behavior of AE in underground mines. ● The SHAP method was used to analyze the prediction contribution mechanism of the model and identify the main influencing factors. ● The evaluation model accurately predicts DB of AE, with interpretable results. Articulated equipment (AE) is one of the main types of equipment for underground mining. Accurately assessing miners’ operational behavior characteristics (MOBC) using AE is critical for risk prevention, safety control, and operational efficiency improvement in underground mining. This study develops a MOBC model by adopting the extreme gradient boosting (XGBoost) to evaluate the driving patterns and characteristics of AE in the underground roadways. To this end, this study uses the Unity3D engine to develop an AE driving simulator, simulates the driving process of AE in the confined roadway space, constructs an evaluation index system, and establishes a database by combining on-site underground experiments. The database contains 2,660 samples with multiple operational features and one target parameter reflecting steering behavior. Automatic optimization is performed using intelligent optimization algorithms, and the optimized model achieves high prediction accuracy with a coefficient of determination (R²) of 0.975 and a root mean square error (RMSE) of 61.75. Shapley Additive exPlanations (SHAP) is further adopted for model interpretability analysis. The results quantify the relative importance of key input features and reveal that specific spatial safety distances and the articulated angle are the most dominant factors. Significant feature interactions are also identified, which provides important insights into AE driving behavior and supports equipment safety management and risk control in underground mining.

Read Full Paperexternally

Demander à l'IA

Bookmark

View Full Paper

Cite This Study

Gao et al. (Sun,) studied this question.

synapsesocial.com/papers/69b3ab3c02a1e69014ccbf62 https://doi.org/https://doi.org/10.1016/j.rineng.2026.109947

Demander à l'IA

Bookmark

View Full Paper