ABSTRACT TESIS
      The use of wireless networks is having an ever greater impact on industry (systems controlling automation, industrial manufacturing, distributed control and supervision…), and wireless networks in particular have rapidly penetrated the area of the control networks that support  public utility services (gas and electricity supply, sanitation and water supply). These types of networks are usually formed of a central station where the information is processed, stored and controlled, different remote stations, and remote clients who are able to remotely monitor or even control the working of the system. 
      To connect remote stations to the central station, various heterogeneous communication infrastructures are used depending on location and circumstances, such as: xDSL, HFC, WiMAX, GPRS, UMTS and HSDP. To create an efficient, useful system, especially in times of emergency, users located in remote stations need to exchange information between each other or with the central station. In these cases, in the remote stations, the users that are monitoring the operation of the facility using their wireless devices can create a spontaneous mobile ad hoc network capable of exchanging information in the case that source and destination cannot achieve direct communication.
      The global objective of this thesis is to analyze the behavior of a MANET in an industrial environment, using the example of a network for the control and supervision of the sanitation network of a large city, and to study how the movement of the nodes affects their performance and the services provided.
      Although there have been many advances in the study of ad hoc networks in recent years, there still remain many issues that need to be studied and improved.
      There has been little research work evaluating ad hoc networks with formal models. In some cases, a routing protocol in ad hoc network has been modeled in detail, and the authors have evaluated how mobility affects the behavior of the protocol, but there has been little or no research work that incorporates the exact topology of the network. In other cases, the entire ad hoc network is not built node by node, since the size of the model is too big to obtain an exact numerical solution, and what has usually been done is that what happens in a particular node is modeled in detail, and the rest of the network is modified in a very simplified form. There are no formal models that take into account the entire network with knowledge of the entire topology and where the nodes are mobile.
      This has been a determining factor when setting the first objective of this work, the design of mathematical models that allow us to represent the behavior of ad hoc networks which can be generated in industrial environments and to evaluate their performance and whether the typical services can be offered and in which conditions.
      Regardless of the topology of the network, it is not possible to know the exact route to be chosen when requesting a communication between origin and destination, so it is impossible to know whether the movement of a node in the path will cause a break and loss of route, since we do not know whether the node that has moved belongs to the active route or not.
      Therefore, taking into account the characteristics of the monitoring systems mentioned, in this thesis a scenario (number of nodes, area size, speed of nodes, mobility, radio transmission range...) has been designed and then modeled using stochastic activity networks (SANs). This model represents the operation of a complete ad hoc network, in which the complete topology is known, which uses a reactive routing protocol and which will show whether the network is able to provide the typical services that the systems that govern these facilities should offer (images or video streaming and alerts). The model also takes into account the fact that the nodes are mobile. Regarding the methodology used in this case, we wish to highlight the incremental development of models and the progressive validation, contrasting the results of each stage.
      The second objective is to contrast and verify the results obtained, for which we created a testbed where we can take measurements to complement the results of the formal models.
      This testbed was set up to carry out experiments with a reactive routing protocol and multiple nodes to evaluate the behavior of the network during the transmission of video streaming. At the same time, changes to the protocol’s default parameters were made to adapt its performance to the application under evaluation, leading to significant improvement in the results. Note that the evaluation of the testbed was carried out in a controlled environment which allowed experiments to be repeated under the same conditions.
      In general, applications of the network under study are rarely taken into account and therefore the results can not be transferred to the operation of a real network. In this work, we have considered the environment and situation in which these networks operate. Therefore, knowing the temporal requirements of industrial networks of metropolitan scope and knowing how the routing protocol during transmission of the typical information used in these systems works, we have determined to what degree mobile ad hoc networks generated in these environments can fulfill these requirements.