Abstract Networked Control Systems is the term used to refer to a special kind of control systems, characterized by employing a shared communication medium to transfer information between controller and controlled plant. This means that in difference from a conventional discrete control, in this kind of systems the same communication medium is used to perform several control loops and even for other tasks not related with the control. By using a non exclusive link to close the feedback loop a set of drawbacks arises, having a negative influence in the evolution of the interest variable. Basically, there are two forms of facing this problem. In one side it is possible to act to improve the performance of the shared communication medium, increasing the available bandwidth and so, reducing its impact in the control objectives. From other point of view, it can be assumed that the communication medium exists and imposes limitations, and trying to reduce its influence by means of designing appropriated control structures. This second approach, closer to the knowledge area devoted to automatic control, is the one that has been used in this thesis. The first one of the drawbacks characteristics of this kind of systems is that the fact of sharing the link reduces the available bandwidth, further when greater is the number of devices contending for using the link. This limitation on the sampling and control frequency, can determine the lower bound of the sampling period used in a conventional discrete control loop and so, it will become a limit if the performance that can be achieved for the interest variable. As the communication medium is not always available, it can be not possible to transmit all the information by time unit that should be desirable, because the capacity of the link must be distributed between all the devices contending for its utilization. In a conventional control system this limitation would compel to increase the sampling period until it can be guaranteed that it is possible to transmit all the information, leaving available the link enough time for the other devices, which in most cases will suppose to be in the worst case. Nevertheless, it is also possible to design a control structure adapted to this situation, trying to make the best use of the available bandwidth and, even it can not achieve the performance of a control system without limitations, it provide an improvement respect to the worst case. The second drawback imposed by the shared communication medium is the presence of random access delays due to the stochastic nature of the information traffic in the link. They are not transport delays which cause the information to arrive more or less late, but without loss. In this case the delay appears in the access to the shared resource and, depending on the relationship between the delay and the frequency of the transmitted discrete signal, it is possible that some samples do not get to be sent. It is not about information that arrives late and in an irregular way, but information that is lost while waiting for the link grant. Logically this loss of information will have a negative influence in the control performance that can be reduced with an appropriated design of the controller, taking into account that part of the information it is going to be lost and acts consequently. The third drawback that has been considered is the lack of synchronization between the time bases governing the tasks performed by devices placed in both ends of the shared link. In a conventional discrete control it is assumed that the sampling of the interest variable, the calculation of the control actions and the application of them over the controlled process coincide in time. Because there is no way of communicate both ends permanently it is not easy to guarantee that the events happening in them are perfectly synchronized. In normal conditions this lack of synchronization would not have a great influence in the system performance but, when it is combined with the random access delay it causes loss and reusing of transmitted information that can be significant for the control objectives. In addition if the lack of synchronization is caused by the shared medium access delay, it will cause irregular sampling and actuation. Again, this drawback can be foreseen in the design of the controller to reduce its influence in the interest variable. The work is approached with a fundamental objective: the development of control structures adapted to the presence of the shared communication and to the drawbacks caused by it. To achieve this objective and following a classic procedure in control theory, the problem scenario is described, characterizing the impact of the link with a collection of parameters. Afterwards the control loop is modeled, including the communication medium, and this model is employed to design the control structure trying to solve, al least partially, the problems caused by the shared link. With this aim, the work has been structured in two parts. In the first one it is considered the combination of limited bandwidth with the presence of random access delays, assuming a perfect synchronization between senders and receivers of the transmitted information. In the second part it is considered the combination of the lack of synchronization with the access delay and the irregularities caused by it, assuming that there are no limitations for the sampling and control frequency. In both cases, a model of the loop through the shared medium is developed and control structures adapted to the considered drawbacks have been proposed. The control structures have been evaluated by using a simulation model of this kind of systems, developed intentionally as a part of the work. This doctoral thesis has been approached as the beginning of an investigation line but, of course, it is not the end of the same. There are still a lot of topics to consider and many problems to solve among which, as an example, it can be mentioned the extension of the proposed control structures to the case that considers the presence of the three drawbacks described, the improvement of the proposed solutions with the addition of a local feedback through a link of exclusive use and the implementation over a real system. Those topics, which are proposed as future work, are just a sample of the many open problems to continue the investigation line devoted to Networked Control Systems.