DriveC: Web Application for Classification of Driving Event

Author : Kevin Servat, Michael Cuadros, Pedro Castaneda

Abstract :This paper presents a web application designed to analyze and classify driving behaviors using data from gyroscope and accelerometer sensors embedded in smartphones. By harnessing real-time sensor data, the tool accurately calculates driving risk, enabling continuous and comprehensive driver behavior assessment. An advanced Long Short-Term Memory neural network model was implemented, chosen for its superior capability to capture temporal dependencies in sequential data and effectively identify complex driving patterns. The model achieved a notable accuracy of 86.36 percent, underscoring its reliability and strong potential for real-time deployment. This innovative approach provides a practical and precise method for driving risk assessment, with significant implications for enhancing safety in the insurance industry and road management systems.

Keywords :Driving rating, Insurance customization, LSTM.

Conference Name :International Conference on Science, Engineering & Technology (ICSET-24)

Conference Place Oruro, Bolivia

Conference Date 18th Nov 2024