Research group leader: ​Professor Martin Fabian  
All researchers in the Automation research group are listed below.

About the Automation research area

Within the Automation research area, a wide range of research is conducted within industrial automation. The focus is mainly on optimization, supervision and control of production systems. Typical examples are robot cells, conveyor systems, and complex machines.
Automation originates from the Greek word ‘automatus’, which means self-acting. A self-acting manner is achieved by a control function, which autonomously controls a system in order that a desired behavior is obtained. On lower levels, control actions are often determined by continuous-time sensor signals, while higher-level control functions are based on discrete events and logical signals.
The research within the Automation area is mainly concerned with discrete event and logical systems, where complexity and information integration are two main challenges. Sustainable production and the relation between product, production and automation design are also important research areas.
The Automation research area, which started in 2000 and now includes around 20 people, is also a part of the Wingquist/Vinnex Excellence Centre and the Production Area of Advance at Chalmers. Extensive collaboration with the industry includes research projects with Volvo Cars, Volvo Technology, ABB, TetraPak, FlexLink, and General Motors R&D.
The Automation research focuses on four areas; Discrete Event Systems, Sequences of Operations, Integrated Product, Production and Automation Design, and Sustainable production.

​Discrete event systems

What are “discrete event systems”?
Many phenomena in our society and daily lives can be modeled as discrete event systems (DESs), where an instantaneous event occurs when a system changes from one discrete state to another. Examples of discrete states are available/occupied resources, number of elements in buffers, and high/low values of discrete sensor signals.
Although individual subsystems are often quite simple, the total integrated DES may become huge in terms of number of states. Examples of such systems are integrated locking and alarm functions in modern vehicles, and robot cells in flexible manufacturing systems, where the number of state combinations may be billions and even more.
The Chalmers involvement
The Chalmers research is focused on modeling, analysis and optimization of such complex DESs. Approaches from computer science and applied mathematics are used and further developed, including formal methods, model checking, AI planning and operation research.
The complexity is dealt with by exploiting the modular structure of the problem, applying abstractions and efficient data structures such as binary decision diagrams. A special model structure, extended finite automata (EFAs) has recently been introduced, where ordinary automata are augmented by discrete variables. This model class is flexible and useful from an industrial perspective, where for instance control programs implemented in programmable logical controllers (PLCs) are naturally modeled by EFAs.
Control functions can be verified but also synthesized to fulfill desired specifications. More recently, techniques to diagnose failures in DESs have also been developed based on EFAs. By extending discrete event models with the time that is spent in different states, optimal routes can also be determined. Different approaches to perform such optimization are now evaluated and further developed. Most of the results on DESs are implemented in a software tool called Supremica.
Ongoing projects:
  • Compositional synthesis of supervisors for modular discrete event systems (VR)
  • Optimal control and diagnosis of discrete event systems (VR)

Sequences of Operations

A specific type of DESs is operations, and sequences of operations. Operations, such as moving or processing a part, are either active, completed, or not yet started. Preconditions between different operations, for instance that one operation must be completed before another one can be started, result in sequences of operations including alternative and parallel behavior.
A sequence of operations (SOP) can act as a recipe of how to manufacture a product, as a protocol that determines how different nodes communicate with each other, or as an activity list that informs the staff at an emergency department how to organize and prioritize among incoming patients.
Graphical language makes sense of operation sequences
To model and visualize these tasks or operation sequences, a graphical language for hierarchical operations and SOPs has been introduced based on self-contained operations.
This means that operations can be grouped and viewed from different angles, for instance in a production system from a part flow, machine/device, or operator perspective. These multiple views increase the interoperability between different engineering disciplines.
The suggested specification language has been implemented in a software tool called Sequence Planner, where SOPs can be modeled, viewed and optimized. This is done based on formal models for the given operations and their relations to each other. Control functions in terms of PLC code can also be generated automatically from the given SOP models.
Ongoing projects:
  • VINN Excellence Centre for Virtual Product Realization (VINNOVA)
  • COOP - Configuration and optimization of operation sequences (VINNOVA)
  • AVM – Adaptive vehicle make (DARPA)

Integrated product, production and automation design

Growing demands on product quality, cost reduction, and a short product life cycle requires concurrent development of new products and manufacturing systems. A key issue is then to have an integrated framework from early product and process design to the final control and operation of a manufacturing plant.
Such a framework has recently been developed, where formal relations between product properties and process operations have been identified. From these formal relations, including resource allocation based on product demands and resource capabilities, operation sequences are generated and optimized.
Challenges: New products on the line and system restarts
Introduction of new products in existing manufacturing plants is one of the more critical challenges, especially in the vehicle industry. Inherently, this brings a performance reduction caused by a down-time and a ramp-up-time. One of the most promising and crucial strategies to reduce this performance loss is to develop and verify the production system offline in a virtual environment.
In collaboration with the vehicle industry and system suppliers, a framework for virtual commissioning has been developed. In this framework, software such as Process Simulate, Delmia Automation and RobotStudio are integrated with Chalmers tools for automatic generation of control functions, e.g. to avoid collisions between interacting robots and other moving devices.
A new methodology for the restart of production systems has also been developed. The purpose is to aid maintenance personnel in resuming normal production after the occurrence and correction of an error. The restart method focuses on the generation of restart states that guarantee correct behavior based on given specifications for the controlled system.
The mass customization challenge
Mass customization plays an important role in modern manufacturing, but benefits of reduced costs come at a price; not all customizable product features are compatible with each other. Configuration Problem describes this situation of choosing only compatible features or determining valid configurations. Due to the high complexity of modern configuration problems, efficient methods and tools are currently developed and used by the industry.
Ongoing projects:
  • FLEXA - Advanced Flexible Automation Cell (EU-FP7)
  • Know4Car - An Internet-based Collaborative Platform for Managing Manufacturing Knowledge (EU-FP7)
  • VINN Excellence Centre for Virtual Product Realization (VINNOVA)
  • Virtual Commissioning of Manufacturing Stations Including PLC Logics (FFI)
  • LISA - Line information system architecture (FFI)
  • Virtual system optimization (TetraPak)

Sustainable production

Sustainable production is achieved by considering environmental, human and economical sustainability.
To meet the environmental challenges, novel methods for energy optimization have recently been developed for systems with interacting and moving equipment. For robot cells, where a number of robots are acting in the same area, energy optimal schedules are obtained at the same time as desired efficiency in terms of cycle time is fulfilled. Similarly, conveyor systems are optimized concerning energy but also lifetime of individual components.
Current work also involves energy reduction of AGV systems. So far energy reduction has focused mainly on individual machines and processes. The novelty of this work is that systems of resources such as robots, conveyors and AGVs as well as multiple products are considered in the optimization.
Humans and robots working together
Human sustainability is also investigated by introducing flexibility, so that for instance robots and humans can work together and replace each other depending on the desired level of automation. From a system point of view, robots and humans are resources with different capabilities that need to be considered when a production system is designed. A framework is currently developed based on operation sequences, where robots and humans can work together and also replace each other in a flexible way. The goal is to achieve an efficient production system, which is robust to failures and varying production rates.
Ongoing projects:
  • Human and automation optimization (Sustainable production initiative/VINNOVA)
  • COOP - Configuration and optimization of operation sequences (VINNOVA)
  • Reduced energy consumption for conveyor systems and paint shops (GM)

Research projects

This project addresses the problem of efficiently calculating a logically correct supervisor for reactive systems. The work is oriented towards design of computer algorithms in a formal mathematical way, mathematically proving their correctness... 
DARPA Adaptive Vehicle Make (AVM) is a portfolio of programs that address revolutionary approaches to the design, verification and manufacturing of complex defense systems and vehicles. The portfolio consists of three primary programs: META, Instant Foundry Adaptive through Bits (iFAB) and Fast Adaptable Next-Generation Ground Vehicle (FANG)...
The FLEXA project is set up to meet one common main objective which is to create the tools, methods and technologies needed to define, prepare and validate an automated flexible cell. This cell should be able to manufacture a generic process chain allowing for safe human interaction, and to deliver quality assured parts for the European aerospace industry...
Current digital manufacturing ICT platforms have provided a series of tools, including CAx, PDM, and PLM systems, to support engineers in a series of collaborative activities, allowing them to communicate as well as to design and validate the manufacturing processes...

The purpose of this project is to create an energy efficient, time and cost efficient preparation procedure of production systems for complex products. The project will also contribute to energy efficient and flexible production systems with a reduced number of discarded parts...

Future sustainable competitive production systems need to be productive and flexible, as well as environmentally friendly and safe for the personnel. To improve a production system and reach these objectives, production management must have access to coherent information regarding products and processes throughout the entire plant...
The project puts focus on the generation of virtual models based on mechanical layout and partly automatic kinematics, concentrating on efficient handling of information. An overall aim is to increase the parallel preparation of the mechanical design and the automated solution...
This project aims to reduce the human workload to avoid physical and mental stress, and at the same time minimize the energy consumption for machines and systems. This can be balanced by the use of optimization methods...
Turning raw parts to finished products consumes a lot of electrical energy in a manufacturing plant.  A large amount of this energy is consumed by processing machines and robots to perform some operations on parts, and by automated handling mechanisms to transport parts in the plant...


Published: Wed 05 Sep 2012. Modified: Tue 21 Jun 2016