Inter-Disciplinary Applications

The fundamental research on theory and methodology is continuously stimulated and challenged by close interaction with cross-disciplinary applications projects:

Process Control

Controlling a process industry plant is a challenging task, since the plants are very complex, being multivariable and nonlinear with lots of human interaction and disturbances. Traditionally, these plants are controlled using a decentralized approach, consisting of hundreds or thousands of PID controllers in various  configurations. Some of these decentralized approaches have been replaced by centralized approaches in the last decade. Model predictive control is one of the most popular approaches. We believe that both the decentralized and the centralized approaches can be developed further, and we have several research projects where we explore benefits of both approaches.

Feedback-Based Resource Scheduling in Embedded Computing

Applying control-based approaches to software and computer systems has a very promising future. Feedback is today used successfully to control physical processes which are only partially known. In the future feedback will also be used to manage uncertainty and achieve performance in computer software systems in several ways: To build reliable systems that can handle errors dynamically, to control the software development process, to control performance in software systems, and to dynamically schedule the computing resources in time-critical embedded systems.

Automotive Systems

In many automotive applications, control systems have reached a critical level of complexity. The reason is a desire to include new functionalities and the introduction of new actuators to provide comfort, safety, and fuel economy and to meet environmental demands. Consequently, the interaction between and the coordination of different controllers is becoming a matter of great importance. The requirements for modularity and safety and the use of components from different manufacturers will most likely lead to a shifted balance between centralised and decentralised control architectures and, eventually, to significant changes in the design of automobiles.


Robotics is a multidisciplinary topic which spans across several different applications and research areas. For this reason, we cooperate with both academic colleagues and industrial partners in our Robotics Lab. Industrial robots are truly nonlinear multivariable dynamic systems. The achievable performance thus not only depends on the mechanics of the manipulator, but strongly depends on the control system (where savings in cost of mechanics and drives, in general, implies a harder control problem). The issues of complexity are particularly urgent in the study of interaction between several robots as well as interaction with humans.


Forced by the drive to use noninvasive methods for diagnosis and therapy, systems theory is expected to have an important impact on theory, methodology, and clinical application in several branches of future medicine. This is a particularly challenging and interesting application area, because of complex interaction with the built-in distributed feedback loops of the human body. Contacts with several departments in the medical faculty have been established over a long time, and the activity has recently been growing.