Vehicle Classification Using the Discrete Fourier Transform with Traffic Inductive Sensors

Area of Science:

• Traffic Engineering
• Signal Processing
• Sensor Technology

Background:

• Inductive Loop Detectors (ILDs) are standard sensors in traffic management.
• Current vehicle classification methods often require data from multiple sensor loops.
• Vehicle speed can complicate accurate classification using inductive signatures.

Purpose of the Study:

• To develop a novel vehicle classification method using inductive signatures.
• To identify spectral features from inductive signatures that are invariant to vehicle speed.
• To demonstrate the feasibility of single-loop sensor-based classification.

Main Methods:

• Extraction of spectral features from inductive signatures using Fourier Transform (FT).
• Analysis of feature invariance to vehicle speed.
• Development of a classification algorithm based on speed-independent features.
• Validation using real-world inductive signatures from a hardware prototype.

Main Results:

• Identification of specific spectral features from FT of inductive signatures that do not vary with vehicle speed.
• Demonstration that vehicle classification is achievable using a single inductive loop sensor.
• Successful evaluation of the proposed method with real-world traffic data.

Conclusions:

• A novel, speed-independent vehicle classification method based on single-loop inductive signatures is proposed.
• The method offers a potential simplification for traffic management systems.
• The findings validate the use of specific spectral features for robust vehicle classification.