Embedded systems have a wide range of applications in many sectors and their importance is continually growing. One field of application includes control tasks. 
The heterogeneity of current systems consist of multiple components from different networked features, suggests the development of distributed control systems in which the various control functions are implemented at different levels.
Considering the presence of distributed systems with communication networks, coupled with limited computing power, implies the need to consider the performance of control in unconventional conditions, such as a local control to ensure safety, alternative controls to the availability resources, activation of different modes of operation to ensure acceptable performance degradation in the presence of delays, losses, actions or excessive computation time. 
The control kernel, comparable to the kernel of an operating system, is defined as the minimum code to be executed in a control application for safe operation, although it may produce a strong performance degradation or even evolve into a state insurance disconnection. Control kernel allows modularity and adaptation of the system and the capacity for rapid development of control applications with support services (middleware) necessary to provide support control algorithms for distributed systems, ubiquitous computing, code mobility and real-time constraints. 
The objective of this thesis is the creation of a distributed control strategy, with embedded items, using the control kernel, which uses high-performance digital controllers in systems with limited computing power. Furthermore, the control system must address the aforementioned problems of data loss, delays or excessive computation time, achieving acceptable performance degradation. 
The thesis proposes a methodology, both to obtain a metric that allows comparison of different modes of operation for any loss of data, and for the choice of parameters of the controllers to be implemented in each node of the distributed system control. This methodology has been tested by simulation and analysis tool, created for this purpose, in processes with different dynamics. In addition, we evaluate the distributed control system on several real processes, which has permitted the validation of the simulation tool as well as supplies collected to data loss in communications.