AI-Powered Early Brake Anomaly Detection With Explainable Predictive Intelligence

This study proposes a secure and efficient machine learning-based framework for predicting brake failures in heavy commercial vehicles. In modern transportation systems, the Air Pressure System (APS) of heavy vehicles is continuously monitored using IoT-based sensors, which generate large volumes of operational data. Manually detecting brake faults from such large and highly imbalanced datasets is both time-consuming and inefficient. To address these challenges, the proposed approach utilizes K-Nearest Neighbour (KNN) imputation to handle missing data and Synthetic Minority Oversampling Technique (SMOTE) to manage class imbalance. Various machine learning algorithms, including Logistic Regression, Decision Tree, Support Vector Machine, Gradient Boosting, and Random Forest, are implemented and evaluated using stratified cross-validation techniques. Experimental results indicate that the Random Forest classifier achieves superior performance in terms of accuracy, precision, recall, F1-score, and ROC-AUC. To improve interpretability and build trust in the prediction process, Explainable Artificial Intelligence (XAI) techniques such as SHAP and LIME are incorporated, enabling clear understanding of model decisions. Additionally, feature selection methods are applied to reduce computational complexity while maintaining high prediction accuracy. The proposed framework enhances the reliability of brake fault detection, minimizes maintenance costs, and supports predictive maintenance strategies in heavy transport systems.