Detection of ABS Events in Electronic Brake Systems using Machine Learning Algorithms

• Published in: Proceedings (IEEE International Engineering Management Conference) pp. 1 - 6
• Main Authors: Maties, George, Fosalau, Cristian
• Format: Conference Proceeding
• Language: English
• Published: IEEE 23.03.2023
• Subjects: ABS; automotive; Electric potential; electronic brake system; Feature extraction; hardware in the loop; Industries; Machine learning; Machine learning algorithms; Prediction algorithms; Safety
• ISSN: 2159-3604
• DOI: 10.1109/ATEE58038.2023.10108131

A novel approach for predicting Anti-lock Braking Systems (ABS) activation in electronic brake systems using Machine Learning (ML) algorithms is presented in this paper. ABS is a critical safety feature in modern vehicles, and early detection of function is essential for preventing accidents and ensuring the safety of passengers. Traditional approaches for detecting ABS failures rely on hand-engineered features and heuristics, which can be time-consuming and prone to errors. In contrast, our proposed method ML algorithms is devoted to automatically learn relevant features from raw sensor data, allowing more accurate and efficient prediction of ABS function. In the paper, we evaluate the performance of some ML models on a dataset of hardware in the loop (HIL) braking events, demonstrating its superiority over traditional approaches. Our results show that the use of LSTM-based models is a promising direction for improving the safety and reliability of electronic brake systems.