When experts in automatic control and those in real-time solutions work independently, the result may be expensive solutions. Or even total misunderstanding. Researchers in automatic control (Lund) and embedded systems (Linköping) have worked in collaboration for many years to develop computer tools that bridge the gap.
Nowadays, hundreds – even thousands – of computers are interconnected in networks, in embedded systems, in cars, aircraft and manufacturing industries. These dynamic systems are controlled with the aid of methods developed in automatic control. The engine, car, aeroplane or factory is caused to act as desired with the aid of feedback, in which the results from control signals are returned to the control system, which issues new control signals in an unceasing loop.
The science of real-time systems is another field that is highly relevant for embedded systems. This branch of computer science studies time-critical computer systems, in which the calculations must be performed so rapidly that control signals can be issued in real time.
These two disciplines, automatic control and real-time systems, have a common origin in the 1960s, but have developed in different directions. They have become separate fields, each with its own concepts and methods.
“Few researchers and engineers have expertise in both branches, which sometimes leads to expensive solutions and may lead to complete misunderstanding”, says Anton Cervin, senior lecturer in the Department of Automatic Control, Lund University.
For many years, however, scientists at Lund University and Linköping University have worked together to bridge the gap between the two disciplines.
“A doctoral student working with Professor Petru Eles and Professor Zebo Peng at LiU contacted me. He was working with real-time systems and was interested in automatic control. So we set up collaboration, and when the ELLIIT programme eventually started, we already had projects under way”, says Anton Cervin.
The collaboration within ELLIIT has led to several computer tools that bridge the two scientific fields. These tools have been welcomed with open arms by industry. The most widely used is TrueTime, a simulation tool available free of charge and based on the standard software in the field, Matlab.
“TrueTime is an engineering tool for rapid prototyping in which all or parts of a system can be built and simulated in the computer before a physical prototype is built. ABB has used TrueTime to simulate a new wireless communication protocol for industrial automation, just to take one example”, says Anton Cervin.
The researchers initially kept track of how many people downloaded the tool.
“We’ve stopped doing that, but we know that we have thousands of users in academia and industry all over the world.”
Several tools – Jitterbug, JitterMargin and JitterTime – have been developed for research purposes. These carry out deep theoretical analyses of system reliability, robustness and performance.
“Our current research is a collaboration with General Motors, using Jitterbug and JitterMargin for the co-design of communication and control systems in a new vehicle”, says Anton Cervin.
The tools are particularly significant when the developers need to determine how powerful, and thus how expensive, the required components are.
“The tools give information about how different subsystems that share common resources can collaborate without disturbing each other. Such bottlenecks are difficult to discover when automatic control specialists and real-time specialists work independently. We provide a common language that everyone agrees on, giving an overall understanding of the complete project”, says Anton Cervin.
Five senior researchers and two doctoral students are working in the field. Anton Cervin, senior lecturer, Martina Maggio, senior lecturer, and Karl-Erik Årzén, professor, all from the Department of Automatic Control, Lund University, and Professor Petru Eles and Professor Zebo Peng, from the Department of Computer and Information Science, Linköping University.