Acoustic Detector of Road Vehicles Based on Sound Intensity.

Area of Science:

• Acoustic Sensing
• Traffic Monitoring Systems
• Signal Processing

Background:

• Traditional traffic monitoring often relies on inductive loops or cameras, which can be costly and intrusive.
• There is a need for low-cost, passive sensing solutions for comprehensive traffic analysis.

Purpose of the Study:

• To develop and validate an acoustic-based method for detecting and counting road vehicles.
• To assess the performance of the acoustic method under real-world traffic conditions, including dense traffic scenarios.

Main Methods:

• Utilizing an acoustic sensor positioned roadside to measure sound intensity in parallel and perpendicular directions.
• Implementing an algorithm for acoustic event detection, sound source localization, and vehicle movement/direction determination.
• Testing the algorithm on a continuous 24-hour recording under diverse traffic conditions.

Main Results:

• Achieved an overall detection accuracy with recall, precision, and F-score of 0.95.
• Demonstrated robust performance even in dense traffic, with worst-case results of recall 0.9, precision 0.93, and F-score 0.91.
• Validated the method's effectiveness through real-world testing over a 24-hour period.

Conclusions:

• The proposed acoustic sensing method offers a viable, accurate, and cost-effective solution for road vehicle detection and counting.
• The system's low complexity and passive sensor design make it suitable for integration into smart city traffic monitoring networks.
• The high detection accuracy supports its application in intelligent transportation systems and traffic management.