A Review of OBD-II-Based Machine Learning Applications for Sustainable, Efficient, Secure, and Safe Vehicle Driving.

Area of Science:

• Automotive Engineering
• Computer Science
• Data Science

Background:

• The On-Board Diagnostics II (OBD-II) system collects real-time vehicle data via embedded sensors.
• Conventional OBD-II applications are limited compared to emerging machine learning (ML) capabilities.

Purpose of the Study:

• To investigate ML-based applications utilizing OBD-II sensor data.
• To enhance sustainability, operational efficiency, safety, and security in modern vehicles.

Main Methods:

• Examination of supervised, unsupervised, reinforcement learning (RL), deep learning (DL), and hybrid ML models.
• Analysis of ML applications for fuel optimization, emission control, driver behavior, anomaly detection, cybersecurity, road perception, and driving support.

Main Results:

• ML significantly surpasses conventional methods in leveraging OBD-II data for advanced automotive functionalities.
• Identified a diverse set of ML approaches applicable to various driving analytics tasks.

Conclusions:

• ML-driven OBD-II applications offer substantial improvements in vehicle performance and safety.
• Future research should address challenges and explore new directions in ML for automotive systems.