Skip to Main Content
Freeway management systems are becoming increasingly important for many ITS applications. To achieve better freeway management, effective traffic performance evaluation parameters are crucial. Most of the current evaluation parameters are from a macroscopic point of view, e.g., level-of-service based on traffic density. This paper proposes evaluation parameters in a microscopic way and investigates the relationship between macroscopic level of service (LOS) derived from roadway sensor readings and the microscopic quality of service (QOS) of traffic from the drivers' point of view. This is important for several ITS applications that need to monitor roadway network traffic conditions, e.g., incident detection. It has been found that macroscopic LOS does not represent the microscopic traffic flow quality very well. A new criterion for determining microscopic traffic flow quality is proposed using average speed values and speed coefficient of variation (CV) of micro-trips. A Gaussian mixture model and expectation maximization (EM) algorithm are used to find the best partitioning thresholds. Experimental results verify the effectiveness of the proposed quality of service measure.