ROCOIDS · Robust control of infinite dimensional systems
6РП — Действия „Мария Кюри“
- Период
- 2005-05-15 → 2007-05-14
- Финансиране от ЕС
- 80 000 €
- Участници
- 1
- Схема
- IRG
Линиите свързват координатора с партньорите. За проекти отпреди 2014 г. CORDIS не винаги дава точни координати. Тези точки са на ниво град или държава.
Накратко на български
Методите за управление на сложни системи с времеви закъснения се прилагат при регулирането на потоци от данни в компютърни мрежи или аеродинамични потоци при самолети. Това помага за намаляване на грешките и смущенията, което подобрява стабилността и работата на тези системи.
Кратко обяснение, генерирано от езиков модел по текста на CORDIS. Оригиналът е по-долу.
Резултати накратко
Final Activity Report Summary - ROCOIDS (Robust Control of Infinite Dimensional Systems)
Feedback control is an essential part of almost all engineering systems. Its main purpose is to reduce the effect of the uncertainty in the form of disturbance signals and modelling errors. This project dealt with robust control of a class of physical processes where the dynamical model was an infinite dimensional system. These types of mathematical models appeared when there was time delay in the process or in cases where the system was spatially distributed. There were three major lines of work within the framework of the ROCOIDS project: 1. the development of new robust controller design techniques; 2. applications on time delay systems, in particular the design of congestion control schemes in computer communication networks, e.g. data flow controllers, and the development of active queue management algorithms; 3. the design of aerodynamic flow controllers, with the specific problem of interest being the suppression of cavity flow oscillations in aircrafts and other aerodynamics' applications. The newly developed robust controllers displayed improvements in comparison to existing methods. For example, in the active queue management of Transmission control protocol (TCP) flows, tracking of a target queue led to reduction in delay jitter as compared to standard techniques, such as random early detection and proportional plus integral control scheme which were proposed by others. The results on active queue management were demonstrated using ns-2 simulations. Collaborations with a group of researchers at the Ohio State University (OSU) led to novel controllers for cavity flow suppressions, which were successfully tested on an experimental facility at OSU.
Текст от CORDIS, на английски · Данни: CORDIS, © Европейски съюз
Цел на проекта
Infinite dimensional system models appear in the feedback control of physical processes that have spatially distributed dynamics, and/or contain time delays. We develop new robust control techniques for such systems, and apply our methods to solve certain specific problems in computer networks and in aerodynamic flow control. Fast and reliable data transport is an essential goal in communication networks. One of the difficulties in this area is the fundamental performance limitation characterized by the pro duct of the return trip time (delay) and the data flow rate (bandwidth). This means that in order to move to higher speeds in data transfer we need to pay attention to time delay. If possible we need to reduce the delay, and the delay jitter.A fluid approximation of data flow in computer networks results in an infinite dimensional continuous time system. We investigate new robust controllers for this type of systems, in particular, for the case when the time delay is uncertain and time varying, and the available capacity of a congested link varies with uncertain cross-traffic. Fluid flow control is another application area involving infinite dimensional mathematical models. In this case we are dealing with spatially distributed systems, where the underlying dynamical behaviour is characterized by the Navier-Stokes equations. These equations are known to be very difficult to solve in real-time to do any type of model-based control. Therefore, typical approach is to obtain a reduced order model through simulations, and experiments. In aerodynamic flow control community Proper Orthogonal Decomposition (POD) is gaining popularity as a model reduction technique. The reduced order models obtained from POD are non-linear. We investigate new controller designs for POD-based models of aerodynamic flow processes.
Оригинален текст от CORDIS (на английски).
Участници
- BILKENT UNIVERSITY · ANKARAКоординаторНиво градТурция
Връзки
Данни: CORDIS, © Европейски съюз