Driving behavior characterization and traffic emission analysis considering the vehicle trajectory

Area of Science:

• Environmental Science
• Transportation Engineering
• Data Science

Background:

• Growing need for resource-saving, environmentally friendly societies necessitates sustainable urban transportation.
• Motor vehicle exhaust is a primary source of urban pollution.
• Understanding the link between driving behavior and emissions is crucial for reduction strategies.

Purpose of the Study:

• To analyze the relationship between driving behavior and traffic emissions.
• To develop methods for constraining driver behavior to reduce pollutant emissions.
• To utilize data-driven insights for promoting low-carbon urban transport.

Main Methods:

• Preprocessing GPS data using Navicat (integration, screening, sorting).
• Cleaning speed data with box-and-line plots and linear interpolation in SPSS.
• Applying Principal Component Analysis (PCA) for indicator dimensionality reduction.
• Clustering driver behavior using K-MEANS and K-MEDOIDS algorithms.
• Analyzing driving states via symbolic approximation aggregation.
• Utilizing the MOVES traffic emission model and decision trees for analysis and prediction.

Main Results:

• Identified distinct driver behavior clusters based on speed, acceleration, and other indicators.
• Established a quantifiable relationship between specific driving states/modes and traffic emissions.
• Developed a decision tree model for predicting driver modes and estimating associated emissions.

Conclusions:

• Driver behavior significantly impacts traffic emissions.
• Data-driven clustering and emission modeling can identify emission-reducing driving strategies.
• The findings support the development of targeted interventions for sustainable urban transportation.