A Driver Rating System Based on Driving Pattern

Recent technological progress is reshaping the execution of daily tasks particularly within the transportation sector, where tools for data collection and analysis have the potential to enhance safety and convenience for travellers. However, despite these advancements, road accidents due to reckless driving continue to occur and effective control methods remain limited. This paper presents a system designed to collect and analyse driver behaviour data from a vehicle's On-Board Diagnostics (OBD-II) port to improve road safety by encouraging responsible driving. Real-time data such as RPM, vehicle speed, throttle position, and engine load, are continuously recorded and processed to assess driving habits. Using test drives with different driving styles, a model has been trained that identifies driving patterns and implemented a rating system to evaluate driver's behaviour over time. In this study, a random forest classifier was applied, achieving an accuracy of 91% in identifying risky driving patterns. The system effectively translates complex driving behaviours into straightforward driver rating, offering actionable insights for safer road practices. This approach demonstrates that machine learning applied to vehicle data can be a valuable tool in reducing accidents and promoting responsible driving.