FLEXCON - Flexible Embedded Control Systems 
FLEXCON is an SSF/IT research programme between January
2003 - December 2005. The budget is 10 MSEK.
Program Summary
Control and automation systems constitute an important subclass of
embedded real-time systems. Control systems have traditionally been relatively
static systems. However, technology advances and market demands are rapidly
changing the situation. The increased connectivity implied by Internet and
mobile device technology will have a major impact on control system architectures.
Products are often based on commercial-off-the-shelf (COTS) components.
The rapid development of component-based technologies and languages like
Java and C# increases portability and safety, and makes heterogeneous distributed
control-system platforms possible. The evolution from static systems towards
dynamic systems makes flexibility a key design attribute for future systems.
The key challenge of FLEXCON is how to provide flexibility and reliability
in embedded control systems implemented with COTS component-based computing
and communications technology. Research will be performed on design and
implementation techniques that support dynamic run-time flexibility with
respect to, e.g., changes in workload and resource utilization patterns.
The use of control-theoretical approaches for modeling, analysis, and design
of embedded systems is a promising approach to control uncertainty and to
provide flexibility, which will be investigated within FLEXCON. Other focal
points are quality-of-service (dos) issues in control systems, and testing-based
verification and monitoring of flexible embedded control systems. The main
application area is adaptive industrial automation systems. An industrial
robotics-based demonstrator will serve as the carrier of the project results.
Workpackages:
- WP1: Flexibility in real-time embedded control systems using COTS
platforms, languages and components
The rapid development of COTS component-based computing and communications
platforms lacking stringent timing guarantees makes static system designs
based on worst-case assumptions increasingly conservative. Research is needed
on design and implementation techniques that allow dynamic run-time flexibility
with respect to, e.g., changes in workload and resource utilization patterns.
In addition it is necessary to improve the understanding of how this dynamic
flexibility may be combined with more traditional real-time system approaches
based on static design approaches. For example, how should event-driven execution
be combined with pre-scheduled time-driven execution in embedded control systems?
- WP2/3: Control-based and Quality-of-Service approaches in embedded
control systems:
Using control-based approaches for modeling, analysis, and design of embedded
computer and communications systems is currently receiving increased attention
from the real-time systems community, as a promising foundation for controlling
the uncertainty in large and complex real-time systems. Areas of growing interest
include feedback architectures for adaptive real-time computing, theory for
performance guarantees under uncertainty, integrated resource scheduling
and feedback control, control-theoretical models of dynamic real-time systems,
application of control theory for controlling timing behavior, and optimal,
robust, or adaptive feedback control in real-time systems. The use of a control-based
approach has the potential to increase flexibility, while preserving dependability
and efficiency. For example, control techniques can be used to compensate
for shortcomings and imperfections in the implementation platforms.Control
approaches to resource allocation are especially interesting for distributed
control systems. For example, a feedback scheduler can distribute the computing
and communications resources in such a way that the global control performance,
or Quality-of-Control (QoC), is maximized. QoC is also an alternative approach
to increasing dependability, e.g., through dynamic reconfiguration of resources
in critical situations or for graceful degradation. The hypothesis for this
work package is that flexibility in complex distributed feedback control
systems can be achieved by developing a new and innovative concept for automatic
negotiation of resources between control tasks based on new and appropriate
measures of control quality. The overall idea is similar to the quality-of-service
(QoS) concept developed for multi-media applications, but the means
for estimating the achieved level of control, as well as timing requirements
and resource structures are very different here. Hence, very few of the specific
mechanisms and methods developed for QoS are applicable in QoC.
- WP4: Testing-based verification and monitoring of flexible embedded
control systems:
Testing-based verification of flexible real-time control systems is inherently
hard. Besides testing in the value domain it is also necessary to test real-time
systems in the time domain. Furthermore, the flexibility offered in event-driven
real-time systems require substantially more test cases for complete test
coverage than a corresponding time-triggered system. From an industry perspective,
research in the testing area is interesting, as there are reports stating
that as much as 50% of the resources spent in software development projects
can be linked to testing activities. The requirement for third-party testing
of COTS components also poses a challenge. Event monitoring in such systems
can support continued testing as well as performance evaluation efforts. In
safety-critical control applications, the need for rigid verification is
high.
- WP5: Robotics and Automation Demonstrator:
In addition to publication of scientific results, we will confront those
results with industrial aspects in an experimental setup. Apart from adding
a focus to all of the workpackages, the demonstrator makes results more tangible
and applicable for industry and applied research. Because of their needs for
flexibility and safety in combination with needs for efficient real-time solutions
and programming on several levels, robots comprise the most challenging case
for flexible control. The focus here is not on robotics as such; the aim
is to experimentally verify mechanisms and systems solutions for flexible
embedded systems. Special emphasis will be put on dynamic deployment of control
components, including testing and performance evaluation. This must be supported
via host simulation, virtual models, and system access over the Internet.
The physical robots and the necessary industrial experience for such a demonstrator
are already available in the group, and results will be made available to
other related projects. Hence, supporting this project is a golden opportunity
to improve applicability of real-time systems research results.
Partners:
University Partners
Industrial Partners
- ABB Automation Products Technology, Malmö and Västerås
- ABB Robotics, Västerås
- Enea Systems
Funded PhD Students:
Industrial Advisory Group:
- Bertil Emmertz, ABB Robotics
- Magnus Larsson, ABB Corporate Research
- Tomas Pauly, ABB Automation Technology Products
- Thomas Vesterlund, Enea Embedded Systems
- Hans Hansson, SAVE
Academic Research Connections:
Publications:
Presentations:
FLEXCON Publications:
- Resource-Constrained Embedded Control
Systems:, Karl-Erik Årzén, Anton Cervin, Dan Henriksson,
CERTS2003 - Co-design in Embedded
Real-time Systems Workshop, Porto, Portugal 2003
- Trends in software and control, Sanz, R.; Årzén, K.-E. IEEE
Control Systems Magazine, 23:3, pp. 12--15, 2003
Related Background Material:
- Integrated
Control and Scheduling , Anton Cervin, PhD thesis ISRN LUTFD2/TFRT--1065--SE,
April 2003 .Department of Automatic Control, Lund Institute of Technology,
Sweden
- How
Does Control Timing Affect Performance?, Anton Cervin, Dan Henriksson,
Bo Lincoln, Johan Eker, Karl-Erik Årzén, IEEE Control
Systems Magazine, 23:3, pp. 16--30, 2003
- Feedback-Feedforward
Scheduling of Control Tasks, Anton Cervin, Johan Eker, Bo Bernhardsson,
Karl-Erik Årzén, Real-Time Systems, 23:1,
2002
- The
Control Server: A Computational Model for Real-Time Control Tasks, Anton
Cervin, Johan Eker, In Proceedings of the 15th Euromicro Conference on
Real-Time Systems, Porto, Portugal, 2003
- Jitterbug:
A Tool for Analysis of Real-Time Control Performance, Bo Lincoln, Anton
Cervin, In Proceedings of the 41st IEEE Conference on Decision and Control,
Las Vegas, NV, 2002
- TrueTime:
Simulation of Control Loops Under Shared Computer Resources, Dan Henriksson,
Anton Cervin, Karl-Erik Årzén, In Proceedings of the 15th
IFAC World Congress on Automatic Control, Barcelona, Spain, 2002
- On
Dynamic Real-Time Scheduling of Model Predictive Controllers, Dan Henriksson,
Anton Cervin, Johan Åkesson, Karl-Erik Årzén, In Proceedings
of the 41st IEEE Conference on Decision and Control, Las Vegas,
NV, 2002
- Feedback
Scheduling of Model Predictive Controllers, Dan Henriksson, Anton Cervin,
Johan Åkesson, Karl-Erik Årzén, In Proceedings of the
8th IEEE Real-Time and Embedded Technology and Applications Symposium,
San Jose, CA, 2002
- An
Introduction to Control and Scheduling Co-Design, Karl-Erik Årzén,
Anton Cervin, Johan Eker, Lui Sha, In Proceedings of the 39th IEEE Conference
on Decision and Control, Sydney, Australia, 2000
- Integrated
Control and Scheduling, Karl-Erik Årzén, Bo Bernhardsson,
Johan Eker, Anton Cervin, Klas Nilsson, Patrik Persson, Lui Sha, Technical
Report ISRN LUTFD2/TFRT--7586--SE, August 1999, Department of Automatic
Control, Lund Institute of Technology, Lund, Sweden
- A Feedback
Scheduler for Real-Time Control Tasks, Johan Eker, Per Hagander, Karl-Erik
Årzén, Control Engineering Practice, 8:12,
pp. 1369--1378, 2000
- Flexible
Embedded Control Systems. Design and Implementation, Johan Eker, PhD
thesis ISRN LUTFD2/TFRT--1055--SE, December 1999 .Department of Automatic
Control, Lund Institute of Technology, Sweden
FLEXCON Meetings:
- FLEXCON Kick-off meeting, Lund, 7-8 October 2002
- FLEXCON Meeting, Västerås, 17 August 2003
- FLEXCON Industrial Advisory Group Meeting, Strängnäs, 20
August, 2003