Abstract
The main objective of the present thesis is the design of a new methodology for modelling the levels of service of motorways in real time, based on the information provided by sensors currently located on the road. It is intended that this modelling methodology can be used by traffic management systems, to obtain a more accurate characterization of the levels of service. This characterization will improve traffic management, due to the fact that the determination of the levels of service in real time will help decision-makers in the activation of traffic management measures.
This research has been based on the fundamental hypothesis that the six levels of service defined in the Highway Capacity Manual (HCM2000), are not suitable for dynamic traffic management, the two main reasons being that: a) the methodology is completely static and it is oriented towards the design of new road networks and b) there are no significant differences between traffic conditions of certain levels of service.
The research development consists of two phases. The first one is a data reduction phase. This reduction has been based on the analysis of the quality of the data provided by monitoring stations, based on loops. The result of this reduction is the creation of a set of filters to ensure data quality.
In the second phase, the proposal and analysis of different methods is carried out to obtain the level of service. These methods are:
1. Obtaining the level of service based on dynamic HCM defined on Highway Capacity Manual. This method uses the intensity values and percentage of heavy good vehicles from the DCS.
2. Practical-analytical methodology for obtaining the level of service based on dynamic HCM. This method considers the definition of free speed based on a geometric average speed. This definition considers those intervals where the intensity values of the measured section are below 400 veh/h.
3. Practical-analytical methodology for obtaining the level of service based on the speed-occupancy matrix. This new proposal is based on one of the three basic curves of Traffic Engineering. This curve establishes the relation between the average speed and the percentage of occupancy time, obtained from traffic data capture stations or DCS.
Once the different methods have been defined and analyzed, a comparative study has been made. Also, a contrast test has been carried out to determine the goodness of the analysis. In addition, through an analysis based on cluster theory, the working hypothesis of method 3 has been confirmed. It has been possible to conclude that for dynamic traffic management in real time, the set of levels of service suitable for traffic management are the four levels of service, because it presents a clear transition between them. 
Abstract