New tools to obtain a high profit of automatic vehicle identification datasensors and models

Resumen

The continuous growth of the population and the demand needs of the latter to carry out activities that imply their mobility, is implying that the administrations, supported by scientific and technical organizations, seek new and innovative solutions that allow it to be managed in a more efficient and sustainable way. In this regard, the Smart concept applied to transportation infrastructures and systems can become one of them, understanding this as a set of digital technologies and tools that allow managers and users to make decisions in a way that is more in line with their needs and circumstances that take place in the environment, and that can condition the performance of a certain action. In everything that has to do with transportation network and traffic management systems, this intelligent or smart technology, manifested through traffic data collection sensors and advanced applications and tools that allow them to be exploited, is increasingly useful for try to manage traffic problems, present and future, and help in decision-making to mitigate the possible consequences that, in time, may affect the functionality and operation of the transportation network. However, implementing these advanced technologies entails an economic cost that, at times, cannot be addressed by any administration or organization involved in activities related to the engineering and management of transportation infrastructures and systems. In addition, it is necessary to take into account that, at a higher cost, in principle, the better the performance of said technology. The latter has to do with the type of sensor to be located in the infrastructure for the collection of data that must later be managed by an analytical tool that allows them to be exploited and obtain a benefit from their analysis. The aim of this Doctoral Thesis is framed in all this context, which consists of addressing one of the most advanced and complex technologies for monitoring and analyzing traffic flows that take place on the road network. This technology consists mainly in the automatic recognition of vehicles through the scanning of their license plates, with subsequent use of the data collected for the analysis and estimation of demand flows, with object to achieve a full diagnosis of the network. In all this, the achieved objectives and contributions with the completion of this Thesis, such that they allow addressing everything indicated above, have been: A review of the technical-theoretical aspects of the monitoring system based on the automatic recognition of vehicles through the scanning of their license plates, as well as the advantages and limitations of its implementation with the observation of real applications. In addition, a deep review of the modelling and prediction techniques of traffic flows is carried out, which is necessary to quantify a priori to know where to locate the sensors through a location model. Address the conceptual and technological design of a low-cost sensor that allows the automatic identification of vehicles, through license plate scanning, and that solves the cost and efficiency limitations that limit the applicability of this technique in road networks of several geographical scales. Adapt the modelling of the road network when the data comes from automatic vehicle identification sensors. Develop a new model to locate automatic vehicle identification sensors to minimize the impact (on the estimation of traffic flows) that uncertainty has in the knowledge of the routes used by the users of a road network. Incorporate as a variable or condition in the network sensor location problems, the precise position that each sensor must adopt in the corresponding link. The formulation of a mathematical model for estimating dynamic traffic flows on routes using data obtained with automatic vehicle identification sensors, which will allow the definition of both the Origin-Destination matrix and the flows in the links that make up the road network. Carrying out a field test that shows the applicability, the plausibility and the economy of the models proposed in the previous objectives.