Optimization of probe vehicle deployment for traffic status estimation

Traffic congestion in urban areas is posing many challenges, and traffic flow model provides accurate traffic status estimation and prediction can be beneficial for congestion management. With the limitation of infrastructure, probe data from individual vehicles is an attractive alternative to inductive loop detectors as a mean to collect traffic data for traffic flow modelling. This paper investigates the optimal deployment strategy of probe vehicles. Data assimilation technique, Newtonian relaxation method, is used to incorporate probe data into macroscopic traffic flow model, and synthetic traffic is used to study the optimization problem. The tradeoff between the quality of traffic density estimation and operation cost of probing are investigating using multi-objective genetic algorithm. The results indicates that it is possible to decrease probe data for congested traffic with negligible degradation on the quality of traffic status estimation.