Second Year Report: Integrated Control and Scheduling
2000-08-11
Karl-Erik Årzén
Klas Nilsson
Dept of Automatic Control Dept of Computer
Science
Lund University
Lund University
Project plan fulfillment
The project started 980501 and is intended as a 2+2 year
ARTES project at 80% activity level. The first half of the project will,
hence, terminate on 001031. The project plan from the original proposal
and the updated project plan from the project evaluation 9908 have been
followed without deviations and the licentiate theses for the two PhD students
have been successfully presented (Patrik Persson, 000412, and Anton Cervin,
000526).
Main achievements
The project has so far been very successful. This can for
example be seen in the recent ARTES evaluation by Andy Wellings where we
cite:
"I think this project borders on VERY GOOD.
There are good conference publications and journal publications are beginning
to appear. The collaboration with Lui Sha is most welcome and important"
The evaluation also recommends that funding is maintained
to allow completion of the PhDs.
The main achievements in the Feedback scheduling subproject
are the following. An in-depth state-of-the-art survey of the field has
been written. A first prototype of a simulator that allows co-simulation
of control algorithms and feedback scheduling strategies has been developed.
The scheduling of different parts of a control algorithm has been investigated
and a new iterative deadline-based priority assignment scheme has been
developed. The performance of different scheduling strategies under overload
conditions and its influence on control performance have been studied.
Experiments with feedback and feedforward-based scheduling strategies have
been performed. A special invited session for the CDC 2000 in Sydney (Conference
on Decision and Control, the most prominent annual major control conference),
on Integrated Control and Scheduling has been organized and accepted.
The main achievements in the Interactive Execution
Time Analysis subproject are the following. A declarative way
of implementing WCET has been developed based on reference attribute grammars.
To correctly schedule automatic memory management in hard real-time
systems, we have presented a live memory analysis that allows this information
to be automatically determined based on straight-forward programmer annotations.
The WCET analysis and the required annotations have also been extended
to cope with virtual method calls and dynamic class loading, which is needed
when developing flexible object-oriented real-time systems. Finally, for
such system development, a first prototype of an interactive development
environment (called Skånerost) utilizing the proposed techniques
has been developed. Although much of our research is not restricted to
any particular language, we use Java as our target language due to its
popularity, portability, safety, and built-in support for concurrency (to
which we add real-time support).
Embedded real-time control systems are of great importance
to Swedish industry. There is currently a trend towards on-line scheduling
in favor of static scheduling. There is also a trend towards using off-the-shelf
computing platforms also for real-time control applications. As the embedded
systems grow more complex due to customer demands, the need for using high-level
languages for time-critical systems increases. Real-time Java plays a key
role in this development. The interactive timing analysis in this project
is an important prerequisite for this. In related projects work on real-time
garbage collection and real-time Java run-time systems are performed, partly
in collaboration with ABB, Ericsson, and Sun Microsystems.
All together this means that the results that have and
will be developed in the project are very relevant for important parts
of Swedish industry. The co-design and simulation environment, and the
interactive WCET tool for obtaining the timing of specific control blocks,
clearly are candidates for industrial application.
Industrial statement
"I den del av industrin vi jobbar d.v.s. telekom,
verkstadsindustrin och medicinsk industri ser vi dagligen problemställningar
som berör hur man på bästa sätt behandlar inbyggda
system lösningar. Det arbete som bedrivs inom projektet avser att
föra fram den svenska industrin så att man använder nya
och effektivare sätt att tackl dessa problemställningar. Denna
forskning med dess lösningar är ett måste för att
svensk industri skall klara av den hårda konkurrensen som råder
på världsmarknaden samt att vi inom landet kan upprätthålla
en god kompetensnivå inom dessa områden.
SES är att bolag som jobbar inom konsultbranschen.
Detta innebär att vi ser många olika företags frågeställningar
inom inbyggda system, dock gör vi ingen egen produktutveckling. Då
det nu finns resultat och god substans i projektet, som kan visas upp för
svensk industri, så har vi en god möjlighet att knyta ihop och
fungera som en brygga mellan industrin och forskningen.
I den inledande fasen av projektet så har SES
deltagit som diskussionspartner. Detta har till stor del inneburit att
vi har försökt att ge en bild av vad som idag används av
industrin inom de områden som projektet berör. Vi klan referera
till hur industrin löser sina problem samt även peka på
varför industrin löser dem på detta sätt.
SES har under den inledande fasen framförallt
deltagit som diskussionspartner men ser nu att projektet har nått
en stabilitet. Med stabilitet menar vi från SES att projektmedlemmarna
har bygget upp en kunskapsbank och en färdighet som kommer att kunna
användas på ett mycket konstruktivt sätt. Vi från
SES avser att deltaga under den efterföljande tiden in te bara som
diskussionspart utan även knyta ihop intressanta "verkliga" problemställningar
som vi vet finns i industrin. SES hoppas verkligen att projektet kommer
att erhålla vidrae finansiering, då vi under den gångna
tiden har sett att det har genererats goda resultat, publikationer samt
engagerade arbetsinsatser."
Malmö, 2000-08-16
Jan J Enejder
Manager Embedded Systems
Sigma Exallon Systems, SES
Publications
Theses
P. Persson: Predicting Time and Memory Demands of
Object-Oriented Programs. Licentiate thesis, Department of Computer
Science, Lund Institute of Technology, April 2000. |
A. Cervin: Towards the Integration of Control and
Real-Time Scheduling Design. Licentiate thesis, Department of Automatic
Control, Lund Institute of Technology, May 2000. |
Journal Articles
K.-E. Årzén and A. Cervin: A Simple Event-Based
PID Controller. Accepted for Control Engineering Practice, 2000. |
J. Eker, P. Hagander, and K.-E. Årzén: A
Feedback Scheduler for Real-time Controller Tasks. To appear in
Control
Engineering Practice, 2000. |
K.-E. Årzén, A. Cervin, J. Eker, B. Bernhardsson,
L. Sha: Towards the Integration of Control and Real-Time Scheduling
Design. In journal submission, 2000. |
Conference Publications
P. Persson and G. Hedin: Interactive Execution Time
Predictions using Reference Attributed Grammars. In Proceedings
of WAGA '99, Second Workshop on Attribute Grammars and their Applications,
Amsterdam, The Netherlands, March 1999. |
P. Persson: Live Memory Analysis for Garbage Collection
in Embedded Systems. In Proceedings of the ACM SIGPLAN 1999 Workshop
on Languages, Compilers, and Tools for Embedded Systems, Atlanta, GA,
May 1999. |
A. Cervin: Improved Scheduling of Control Tasks.
In Proceedings of the 11th Euromicro Conference on Real-Time Systems,
York, England, June 1999. |
P. Persson: Using Aliasing Information to Predict
Real-Time Garbage Collection in Embedded Systems. Position paper presented
at the Intercontinental Workshop on Aliasing in Object-Oriented Systems,
Lisbon, Portugal, June 1999. |
K.-E. Årzén: A Simple Event-Based PID
Controller. In Proceedings of the IFAC World Congress, Beijing,
China, July 1999. |
J. Eker and A. Cervin: A Matlab Toolbox for Real-Time
and Control Systems Co-Design. In Proceedings of the 6th International
Conference on Real-Time Computing Systems and Applications, Hong Kong,
China, December 1999. |
P. Persson and G. Hedin: An Interactive Environment
for Real-Time Software Development. In Proceedings of the 33rd International
Conference on Technology of Object-Oriented Languages and Systems,
St. Malo, France, June 2000. |
K.-E. Årzén, A. Cervin, J. Eker, and L.
Sha: An Introduction to Control and Real-Time Scheduling Co-Design.
39th
IEEE Conference on Decision and Control, Sydney, Australia, December
2000. |
A. Cervin and J. Eker: Feedback Scheduling of Control
Tasks. 39th IEEE Conference on Decision and Control, Sydney,
Australia, December 2000. |
Technical Reports
K.-E. Årzén, B. Bernhardsson, J. Eker, A.
Cervin, K. Nilsson, P. Persson, and L. Sha: Integrated Control and Scheduling.
Internal report TFRT-7582, Department of Automatic Control, Lund Institute
of Technology, August 1999. |
A. Cervin: The Real-Time Control Systems Simulator---Reference
Manual. Internal Report TFRT-7592, Department of Automatic Control,
Lund Institute of Technology, April 2000. |
P. Persson, A. Cervin, and J. Eker: Execution-Time
Properties of a Hybrid Controller. Internal report TFRT-7591, Department
of Automatic Control, Lund Institute of Technology, April 2000. |
Karl-Erik Arzen and Klas
Nilsson
Last modified: 2000-08-09
|