Vehicle Reidentification in a Connected Vehicle Environment using Machine Learning Algorithms

Deployment of connected vehicles will become possible for most American cities in the next 10 to 20 years. Connected vehicle (CV) applications (e.g., mobility, safety, environment) are constantly receiving vehicle data. The current ID protection mechanism assumes a vehicle’s ID changes every 5 minutes, so the topic of rematching vehicles is of interest in privacy protection and performance measure research. This paper explores the possibility of rematching connected vehicles’ IDs using popular machine learning techniques, including logistic regression (LR), linear discriminant analysis (LDA), quadratic discriminant analysis (QDA), linear and nonlinear support vector machine (SVM) and nearest neighbor algorithms. An experiment is conducted using a microscopic traffic simulation model through a software-in-the-loop technique. The best average mismatching rate is 14%. To assess potential factors’ effects on matching accuracy, a Poisson mixed regression model is analyzed under the Bayesian inference framework. Findings are: different matching algorithms vary in matching performance and the linear SVM, the QDA and the LDA have the best accuracy results; traffic volume and market penetration rate have little impact on matching results; location and number of vehicles to be matched are considered significant. The results make the performance measurement of future CV applications feasible and also suggest that more secure mechanisms are needed to protect the public.