Information
Coordinator:José Simó
Finantial institution:Spanish Ministry of Education, Science and Technology - DPI2005-09327-C02-01/02
Start: 01/01/2006
End: 31/12/2009
Home page: http://www.gii.upv.es/kertrol/
Partners: Institute of Systems Engineering and Control - UPV
Researchers: Gines Benet
Javier Coronel Parada
Patricia Balbastre Betoret
Pedro José Martínez Tárraga
Vicente Nicolau Gallego
Abstract
Embedded systems have a wide range of applicability into many sectors and their relevance is continuously growing. Many application fields involve running control tasks. It is well known that most control tasks have hard real time constraints. In an embedded systems environment, where the information is flowing among different processing units, it is necessary to be aware at any moment about the resources availability as well as the delivering time of the control action.

In this framework, the concept of a control kernel, similar to the OS kernel, arises. And this concept can be extended to many other control systems. It implies to determine the minimum control code we must run at any control task instantiation to ensure a safe operation of the controlled system, accepting the degrading of some performances and even the evolution to a safe shut-down.

The control kernel implementation implies to take into account the OS kernel. The current revolution providing miniaturisation, ubiquitous communications and digital convergence is leading to a major shift in Embedded Systems, which are moving to spread the processing resources, by selectively running changeable tasks and becoming strongly interconnected systems. From this point of view we can foresee new levels of complexity in the control that, therefore, will lead us to new levels of safety, comfort, and higher levels of productivity covering all areas from industry to people.

These characteristics require the joint consideration of the control design and implementation. The relevance of each control signal should be taken into account. This will determine the relevance of each control task and, thus, the admissible delays in order to establish the priorities and the time (resources) allocation: The goal is the control action delivering reduction or, in a more general viewpoint, to balance these delays in order to achieve the lowest performance degrading of the complete system, autonomously operating under resources limitation.

The modularity and adaptation of the system as well as the rapid application development features will be supported by services and middleware related to distributed systems, ubiquitous computing, mobility of code and real-time constraints.

The general goal of the coordinate project is to provide embedded systems with a high level of intelligence by means of the interaction with the environment and the communication among all components forming a solution based on embedded systems. This new level of intelligence must be compatible with fulfilling real-time restrictions of all control tasks.

The development of hardware platforms, operating system, and services necessaries to execute control systems defined following the concept of control kernel will be the matter of the first subproject. The development of the concept of control kernel itself will form the second subproject.

In order to test the results, the construction of a demonstrator platform is part of the project. This demonstrator is a biped robot that will be controlled using algorithms developed in the other subproject. In fact, the demonstrator will drive the integration of results and the coordination between the two subprojects.