A Deep Learning Framework for Driving Behavior Identification on In-Vehicle CAN-BUS Sensor Data

Area of Science:

• Computer Science
• Artificial Intelligence
• Automotive Engineering

Background:

• Driver identification is crucial for applications like auto-theft systems.
• Controller Area Network-BUS (CAN-BUS) data offers a rich source for analyzing driving behaviors.
• Traditional methods struggle to capture the complex temporal dynamics inherent in driving data.

Purpose of the Study:

• To develop an advanced deep learning framework for accurate and automated driver behavior identification.
• To overcome limitations of traditional methods in modeling temporal features from CAN-BUS data.
• To create a robust system for recognizing individual driving patterns.

Main Methods:

• An end-to-end deep learning framework fusing Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs).
• Integration of an attention mechanism to focus on salient temporal features within the time series CAN-BUS data.
• Automatic feature learning and correlation analysis across multi-sensor data for comprehensive behavior representation.

Main Results:

• The proposed framework successfully models complex temporal features from CAN-BUS sensor data.
• It automatically extracts relevant driving behavior characteristics without requiring manual feature engineering.
• Demonstrated superior performance in real-world driving behavior identification tasks compared to existing state-of-the-art methods.

Conclusions:

• The developed deep learning approach provides an effective solution for driver behavior identification using CAN-BUS data.
• The fusion of CNNs, RNNs, and attention mechanisms enhances the ability to model intricate temporal patterns.
• This framework offers a promising advancement for intelligent transportation systems and vehicle security.