Vehicular Environment Identification Based on Channel State Information and Deep Learning.

Area of Science:

• Computer Science
• Electrical Engineering
• Artificial Intelligence

Background:

• Connected vehicles rely on environmental awareness for safe operation.
• Current methods often require dedicated sensors like cameras or radars.
• Vehicular wireless channel characteristics offer untapped potential for environmental sensing.

Purpose of the Study:

• To develop a novel vehicular environment identification approach using deep learning.
• To leverage Channel State Information (CSI) for environment classification.
• To enable environment recognition without additional sensors in connected vehicles.

Main Methods:

• Environment identification framed as a classification problem.
• Development of a new Convolutional Neural Network (CNN) architecture.
• Utilizing estimated CSI as input features for the CNN model.

Main Results:

• The proposed CNN model achieved high accuracy (96.48%) in environment recognition.
• The method reliably identifies the type of environment a vehicle is driving in.
• Performance evaluation demonstrated superior results compared to existing approaches.

Conclusions:

• Deep learning-based CSI analysis offers a viable sensor-free approach for vehicular environment identification.
• The novel CNN architecture is effective for classifying driving environments.
• This method has significant implications for enhancing the situational awareness of autonomous vehicles.