|
Organizers and Guest Editors: Christer Norström and Hans Hansson, Mälardalen University, Västerås, SWEDEN Guest Editorial: The never ending demand on higher productivity in our factories leads to requirements on a total integration from sensors to decision making. This vertical integration includes asset optimisation on all levels, and calls for standards that allows new production sites to be designed and built by equipment from various vendors, including providers of network technology, compatible controllers, devices, machines, and machinery, and thus the vertical integration promotes industrial system growth without being dependent on vendor propriety technology. One interesting observation we can do as a community is that the automotive sector has developed considerably with respect to networking in the last 15 years and are sharing more and more technology and solutions with industrial automation. Owing to the size of the automotive sector, the cost pressure on the suppliers of shared technology has increased dramatically, and has thus made solutions cheaper also in the industrial domain. Common architectural elements can be identified between industrial automation and automotive systems. For example, a truck shows similarities with a larger production system, as it has a comparable hierarchy of networks, all the way from the low-level sensor networks to the interfaces to business systems. In both these domains the networks are the integration platform of these highly integrated systems, and thereby have, besides being standardised, very high requirements on reliability and availability as they are the core of the integrated system. The trend is also that the requirements on reliability and availability are continuously increasing to achieve higher quality and productivity. Further, the networks must be able to handle different kinds of non-functional requirements such as quality of service requirements, real-time requirements, cost effective implementations as well as different types of network traffic including critical control loops, business data, monitoring and network management. Today there are many networking solutions, but there are still many issues remaining to be solved, and further improvements and optimizations of solutions are in high demand.. This “Special section on Factory Communication Systems” is devoted to advancements in industrial network technology and aims at presenting a sample of current research in the area. Common for the papers in this issue is that all has focus on guaranteeing the timing behaviour, which is an utmost important factor for achieving systems with high reliability. The four papers in this special section cover Ethernet networks, reliability in wireless industrial LANs, Quality of service, and real-time server based communication. The papers are introduced through an overview of the issues they address. In modern process control systems, Ethernet is becoming more and more common on all levels in a factory. The main problem of using Ethernet at the field level is the non-determinism of the Ethernet MAC protocol, that is, the MAC protocol cannot guarantee an upper bound on the channel access time. Thus, standard Ethernet is not suitable for hard real-time systems. However, the standard Ethernet protocol can be successfully used in soft real-time applications and by introducing higher level protocols also in hard real-time applications. The paper “Improving the Real-Time Behavior of Ethernet Networks Using Traffic Smooting” by L. Lo Bello, G. A. Kaczynski and O. Mirabella considers industrial applications with soft-real requirements. The paper presents fuzzy traffic smoothing, a technique to perform adaptive traffic smoothing over Ethernet networks at the field level to enable statistical bound on packet delivery time. The paper “FTT-Ethernet: A Flexible Real-Time Communication Protocol that Supports Dynamic QoS Management on Ethernet-based Systems” by P. Pedreiras, P. Gai, L. Almeida, and G. Buttazzo presents the FTT-Ethernet protocol to support hard-real-time operation in a flexible way, over both shared or switched Ethernet. The FTT-Ethernet protocol employs an efficient master/multi-slave transmission control technique and combines on-line scheduling with on-line admission control, to guarantee continued real-time operation under dynamic communication requirements, together with data structures and mechanisms that are tailored to support dynamic QoS management. Wireless LANs is an attractive networking technology for industrial applications. The main objectives for not using Wireless LANs in industrial applications are the error-prone behaviour of wireless channels which may cause system unavailability or system failures. Most techniques for increasing the probability of transmitting a message before a deadline use automatic repeat request (ARQ) protocols. The paper “Redundancy Concepts to Increase Transmission Reliability in Wireless Industrial LANs” by A. Willig presents three modifications to an ARQ protocol. As one of these modifications, a specific transmit diversity scheme, called antenna redundancy, is introduced. The other modifications are error-correcting codes and the transmission of multiple copies of the same packet. These modifications are thoroughly investigated and compared under different error conditions. Network control systems based on the Profibus token bus protocol are common. Typically, the token passing protocol is subject to random network delays due to uncertainties in token circulation, but has an upper and lower bound of what constitutes and acceptable network delay. The upper and lower bound depends on the performance parameter target rotation time. The paper “QoS-based Remote Control of Networked Control Systems Profibus Token Passing Protocol“ by K. C. Lee, S. Lee M. H. Lee presents an algorithm for selection of target rotation time using a genetic algorithm to ensure quality of service of control information. The proposed algorithm was evaluated on a network control system (Profibus-FMS network) for a feedback control system for a servo motor. Share-driven scheduling provides temporal isolation between applications, and is thus an interesting approach for modular system design. In a networking context this approach provides flexibility for assigning communication bandwidth to different applications during development, and facilitates dynamic addition and removal of system components during run-time. In the final paper of this special issue, “Real-Time Server-Based Communication with CAN”, T. Nolte, M. Nolin, and H. Hansson apply share driven scheduling to the scheduling of communication in the Controller Area Network (CAN). Using servers to implement the share-driven scheduling, they show how fairness and bandwidth isolation between predictable as well as unpredictable streams of messages on the network can be provided.. We would like to thank all authors of papers submitted for consideration for publication for their contribution. We would also like to thank all reviewers for their careful review work which contributed significantly to the quality of the presented work. Guest Editors: Christer Norström is a professor in software and systems engineering. He received a Ph.D from Royal Institute of Technology (KTH), Stockholm in 1997, became Docent at KTH in 2001, and a professor at Mälardalen University 2002. Christer's research interests are design of real-time systems, reliability and safety methods, software engineering, and architectures for real-time systems. His research is almost always carried out in close collaboration with industry. Currently his main industrial partners are ABB Robotics and Volvo CE. He is very interested in technology transfer from academia to industry and he has manifested that through several successful transfers to the automotive industry. He is currently working as the Dean at Mälardalen University. Previously, he was working as a manager for future technology at ABB Automation Technology Products/ Robotics. He has also worked as a consultant, in particular for the automotive industry. Christer has given numerous courses on real-time system for industry both in Sweden and in Europe. Christer was previously department chairman at the Department of Computer Engineering, Mälardalen University. In year 2001 he was awarded best teacher at Mälardalen University. Hans Hansson is a professor in Computer Engineering, specialising in real-time systems, at Mälardalen University since 1997. He is the director of research at the Dept. of Computer Science and Electronics, heads the MRTC and its Real-Time Systems Lab. He is the director of the national graduate school SAVE-IT and co-ordinates the national research initiative SAVE. He received a MSc (Engineering Physics), a Licentiate degree (Computer Science), a BA (Business Administration), and a Doctor of Technology degree (Computer Science) from Uppsala University, Sweden, in 1981, 1984, 1984 and 1992, respectively. He was appointed “docent” in Computer Systems at Uppsala University 1998. Hans was previously the director of the national research programme ARTES, visiting professor and department chairman at the Department of Computer Systems, Uppsala University, and researcher and scientific advisor at the Swedish Institute of Computer Science in Stockholm, Sweden. Hans Hansson's current research interests include real-time system design, scheduling theory, distributed real-time systems, and real-time communications networks. An important current focus is on component based design of safety-critical real-time systems.
Section Contents:
Title:
FTT-Ethernet: A Flexible Real-Time Communication Protocol that
Supports Dynamic QoS Management on Ethernet-based Systems
Title:
Exploiting Redundancy Concepts to Increase Transmission Reliability in
Wireless Industrial LANs
Title:
QoS-based Remote Control of Networked
Control Systems via Profibus Token Passing Protocol
|
IEEE Transactions on Industrial Informatics