To simplify the structure of decentralized automation systems it is necessary in the future to transmit the information and the power supply for the bus nodes over one and the same line. The paper gives a summery of existing serial bus systems with integrated power supply. The possibilities to use such systems in the sensor/actuator area is discussed and the physical and technical limits are shown. The conclusion is, that it is advantageous to adapt existing bus system of the sensor/actuator area by changing their physical layer. This can be done by several kinds of modulation as known in communication engineering.

CAN systems are designed to be used in automotive and automation environments where it is likely to have a high degree of electromagnetic interfer- ence disturbing the data transported between the transmitter and receiver. CAN uses several error detection mechanisms to prevent receivers from accepting disturbed data. Assuming a two-state symmetric binary channel model for the physical transmission medium, this paper analyzes the probability for errors to be undetectable at receivers (residual error probability). The contributions from different error mechanisms to the residual error probability are identified and quantified.

DeviceNet is an open network developed by Allen-Bradley based on CAN that is designed to allow low cost industrial control devices to communicate with each other. DeviceNet is defined in terms of an abstract object model which presents the suite of communication services available and describes the externally visible behavior of a DeviceNet node. The DeviceNet Model is application independent. DeviceNet provides the communication services needed by various types of applications. Many of today's lower level industrial control devices must retain their low cost/low resource characteristics even when directly connected to a network. DeviceNet takes this into consideration by defining a specific instance of the Model for communications typically seen in a Master/Slave application. This is referred to as the Predefined Master/Slave Connection Set. This paper describes the DeviceNet Communication Model and presents the Predefined Master/Slave Connection Set.

The trend towards distributed intelligence in automation applications is gathering pace. Many different bus systems are now emerging, aimed at the industrial automation market place. The Controller Area Network (CAN) offers several important features which can be exploited to give an optimum solution to many industrial applications. This paper presents some of the work carried out with CAN-based servodrives over the past few years. With thousands of drives in the field, the advantages of CAN have been clearly demonstrated and the features of CAN have been exploited in full. The multi-master capability is used to implement electronic cam and gear functions over the bus while the broadcast facility is used to synchronise the sample periods of all drives down to the micro-second level. Within the ESPRIT project ASPIC1, a device profile has been defined which details an open standard for addressing drives on a CAN network. This profile defines two channels of information over the bus, an operational channel for high speed, real-time data, and a service channel for parameter data. In designing the CAN Drive Profile, drive profiles for other bus systems were also taken into consideration.

DC motor control is the most distributed application in automotive body systems. As the number of DC motors in cars increases, a multiplexed architecture becomes the only viable solution, reducing cost and weight, improving reliability and control efficiency, and significantly increasing passenger convenience. This paper demonstrates in a practical example how products from National Semiconductor help minimize the cost of these applications, using CAN as the state of the art, fault tolerant networking solution.

Many time critical applications, e.g. measurement devices, require a real-time clock with an accuracy in the order of microseconds. In a centralised system this is easy to implement with standard timer devices, but in a distributed system (like a number of sensor and actor nodes connected via the CAN-bus) this is more difficult as there is no global system tick. This problem can be solved by synchronising the local clocks of all nodes with a sufficient accuracy. The tight timing guaranties of a CAN-network offer a simple and cheap possibility to provide such a global clock without additional hardware. The real-time group of GMDs CREW Project has designed and implemented a clock synchronisation protocol on the CAN-bus that provides a global time base with an accuracy of about 20 microseconds. The protocol is simple and hardware-independent. It uses only a small amount of bandwidth (< 20 messages/second) and works with a single, arbitrary CAN-object. If necessary, e.g. in large scale networks, the protocol can be synchronised with an external time-base, like a GPS satellite receiver.

The acceptance to use the Controller Area Network (CAN) in industrial automation systems has been grown noticeable in the last time. One reason surely is the short reaction time which goes back to the bus access method (CSMA/CA). The disadvantage of this bus access method is, however, that a defined reaction time only may be guarantied for the object with the highest priority. The question to be answered was, wether it is possible to realize message delay times and object cycle times which are acceptable for typical industrial automation systems. This also concerns the message time equidistance. Does the prioritized bus access conflict with the demands on message time equidistance? The investigations based on conditions of the sensor/actuator area in industrial automation systems. That means a transmission distance of a few hundred meters and a baudrate of 500 kBit/s.

A communication profile for real time systems which is based on the layer 7 specification CAN Application Layer (CAL) was developed in the ESPRIT pro- ject „ASPIC“. It supports the quick exchange of real time data as well as the utilisation of conventional device profiles. The concept involves the introduction of two communication channels with differing features: an operational channel for real time data and a service channel for parameter communication. Both are established between the cen- tral controller and the connected devices. The concept uses a subset of CAL. NMT and DBT are employed to perform network management and identifier dis- tribution whilst two CMS services are sufficient to accomplish the communica- tion channels. Device profiles interface to the communication profile. Here the communication objects are defined which enable standardised access to functions and fea- tures of the devices. The objects (e.g. parameter, modes or programs) can be addressed via an index in the service channel.

To wash aircraft’s a truck-based washing system was developed. The com- puter system is divided into two parts, An industrial robot control (IRC) and a bord computer. This bord computer is responsible for the diagnostic and con- trol of the system. A lot of signals must be analysed (about 250). To sample all these signals a CAN system is used. The advantage of this system is the distribution of connecting points and cabling. The safety of the whole system is higher than comparable systems with traditional wiring. The presentation describes the advantages, constraints, requirements and drawbacks for the implementation of CAN-system in the aircraft washing- system.

The CAN Protocol1 is currently implemented as on-chip peripherals integrated on microcontrollers and as stand-alone CAN chips. On-chip peripherals are available on several microcontroller architectures, including the MCS_ 51 and the MCS_ 96 microcontroller families. Likewise, there exists a variety of production-level stand-alone CAN chips such as the Philips PCA82C200 and the Intel 82527. The decision to use an integrated CAN peripheral or a stand-alone CAN chip should consider the tradeoffs between both alternatives. These tradeoffs include implementation cost, design flexibility, level of CPU burden and system reliability. This paper discusses these tradeoffs from both qualitative and quantitative perspectives. The goal of this paper is to identify the key issues that differentiate these two alternatives for various design and production goals.

CAN is known as a protocol for high performance and high reliable serial communication links between electronic control units in the field of automotive and industrial control applications. A new method of real-time support in the network allows to enhance the dynamic behaviour of the overall system as well as the diagnostic capability. With the microPD72005, NEC has developed a CAN controller, which allows to establish a global time system among the loosely coupled nodes of the network.

First the concept of the CAN-based in-vehicle network for the next generation of Mercedes-Benz trucks is introduced. This implies the connection of different systems, as e.g. engine- and braking-controller. A special characteristic of this information-network is the structure, where the components are combined in functional groups, which are connected via gateways, and organised in several hierarchical levels. Emphasis is put on concepts to the integration of external electric/electronic functions in the in-vehicle network. This will be shown on the interfaces to external systems such as bodyshell-electronics (e.g. concrete-mixer) and trailer-electronics (braking and non-braking applications), and on the data transmission via telephone ore satellite-communication for other services (e.g. vehicle pool management, emergency calls, remote-diagnosis and so on). The current position of this integration is illustrated by some examples. At the end of the discourse the question of standardisation is considered critically, and the consequences for the relationship in information exchange between car manufacturer, trailer and body manufacturer and suppliers for electronic systems are discussed.

A Controller Area Network (CAN) based communication protocol and control environment are presented for plant floor automation and control. This robust protocol allows for messaging between sensors, actuators, man-machine interfaces, controllers, and other plant floor control devices. This architecture provides a communications environment upon which a high-speed, real time centralized or distributed control platform can be created. In addition, the communication system provides the enabling technology for intelligent sensors, integration of advanced device diagnostics, and plant floor data-gathering, for communication to higher level networks. State of the art technology allows for the integration of the network interface into the smallest of plant floor devices. This system provides benefits in a variety of areas. Several real applications are discussed, including a description of an installation at a General Motors Engine Plant in Australia. A description of the benefits achieved in each installation is provided. In particular, machine uptime is improved through the integration of more intelligent devices such as programmable, re-configurable sensors with integrated diagnostics. Labor costs are reduced, due to the simplified interconnect system, and the easy addition to and modification of existing installations. Also, reliability is improved through diagnostics and reduced system complexity, and maintenance is improved through advanced tools for device and system configuration, monitoring, and troubleshooting. Reduced wiring costs can also be achieved, due to the bussing of devices.

The Society of Automotive Engineers’ Heavy Duty Truck and Bus Division created a committee to form a standard for multiplexed control busses in their vehicles. The architecture of these vehicles are complex in nature, comprised of several modules (cab, trailer1, trailer2, etc.). Therefore they chose CAN as the basis of their standard, specifically version 2.0B, for the 29 bit identifiers. Several documents were created that built upon the Bosch specifications but details all seven layers of the ISO/OSI (Open System Interconnect) model. This paper will show how and why certain implementations were used. Using 29 bit identifiers allowed segmentation of the identifier into content, source, destination address, etc. The physical layer is also specified it is similar to ISO 11898. The need for very fast transmission rates, typically under 20 milliseconds, forced the application layer to pack similar data inside one CAN message. The process of decoding these messages is difficult because in some cases one data parameter may affect how other parameters are decoded. Due to this, several tools have been created to help facilitate decoding of messages.

The smart sensor concept offers the benefit of a digital error compensation of non-ideal sensor characteristics and temperature influences. For this purpose, a dedicated correction processor was implemented which uses individual calibration data organized as a two- dimensional characteristic diagram to calculate the true measurement value. Sensor values up to 14 bit are supported. The characteristic diagram can be set up in an end-of-line calibration process using a serial EPROM data interface. The calculated sensor value is transmitted via a digital, bus-capable sensor interface which offers the possibility of interference-free signal transmission and supports future multi-sensor structures. In the presented implementation, this task is performed by a CAN protocol controller with a reduced message memory, thus leading to a minimized chip area. In the design process, the CAN protocol's reference environment, consisting of a protocol controller model, a testbench, waveforms and a simulator kernel was used to verify the implementation of the CAN interface. By describing the other parts of the chip including test patterns in the same high level language, the whole chip could be verified on functional level before breaking down the design to the schematic entry of the selected IC-CAD. Testvectors automatically generated by the functional model ensured the correctness of this transformation. The chip named CC400 was fabricated in a 1.2 micrometer CMOS process with EPROM option.

The CAN (Controller Area Network) is one of today’s most widely accepted car networking systems. Various protocol implementations are available from different suppliers. Dedicated protocol controllers - Full-CAN controllers - are found as system bus interfaces connected to a main CPU or integrated into them. Yet in some applications, particularly in the low speed arena, these devices don’t meet the price target or offer the flexibility required by the system designer. This paper outlines the application interfaces available for the CAN protocol, gives an overview to National Semiconductors CAN chips and it demonstrates in a practical example how these products can help to minimise the cost of Full-CAN controller applications while increasing the flexibility of such systems.

"Based on CAL, the Standard CAN Application Layer of the CiA, two general approaches for implementation of open networks which provide interoperability and interchangeability of devices are presented. In the first approach the application processes have direct access to the services and protocols of CAL. Application depended specifications are provided and additional standardized or nonstandardized ""Application Profiles"". With the second approach (""CANopen"") a standard application is specified which provides the environment for the usage of standard devices in different applications."

Since the number of available automation components with CAN interfaces is increasing more and more, there is a strong demand on interoperability between these components in multi vendor systems. However, the CAN in Automation community is still using a wide range of manufacturer specific communication solutions. In order to achieve interoperability of control components, communication and device profiles are to be employed together with the CAN communication layers that form the basis for specific implementations. CANopen, a set of existing and emerging profiles based on CAN Application Layer (CAL) is presented. These profiles are open to manufacturers and users. The CAL based Communication Profile For Industrial Applications (CiA standard DS 301) allows the definition of a wide range of device profiles e. g. for decentralised I/O, drives, vision systems, encoders, etc. The communication profile which is presented in detail provides fast event driven or cyclic messages as well as asynchronous data transfer. Since several companies have already adopted CANopen, an overview is given on ongoing implementations.

The DeviceNet Physical Layer and Media include the following features: CAN technology, multiple baud rates, cable distance up to 500 meters, Thick and Thin drop line or trunk line, drop lines as long as 6 meters, isolated and non–isolated physical layers, branching on drop lines and protection from wiring errors. In addition to providing CAN based communications, DeviceNet also provides power. Because power and signal conductors both are contained in the cable, devices can draw power directly from the network without the need for separate device power supplies. DeviceNet provides a flexible approach to supplying power along the bus. Both single and multiple supply configurations are supported allowing high current capability. Power supplies can be placed anywhere along the trunk line resulting in greater flexibility for the system designer. The DeviceNet power bus is supplied by one or more nominal 24 volt power sources and can support up to 8 amps on any section of Thick cable trunk line or up to 3 amps on any section of Thin cable trunk line. This paper describes DeviceNet’s physical layer, media and power capabilities.

CAN network protocol originally had been invented for automotive applications. Because of its characteristics, its robustness in conjunction with its excellent performance/price ratio CAN very soon was adopted for industrial control applications widely. The high sales quantities of more than 9 mio units until 1994 and the still more promising forecast illustrate the importance of CAN as a protocol in fieldbus and sensor/actuator bus applications. The following paper discusses the historical evolution of CAN until today, giving the reasons for the big success of CAN. In a second paragraph the technical characteristics of today’s CAN systems are analyzed and explained. Paragraph 3 dares a prognosis how and where CAN systems might go in the future.

DeviceNet connections can be configured with a multitude of behaviors based upon the transport class types and data production triggers desired by the end communicating applications. The DeviceNet communication model provides a variety of connection object transport class types and data production triggers. Connection objects can be configured to produce only or consume only. They can be configured to both produce and consume and to do so in a fashion that is either synchronous or asynchronous with the end application to which they are linked. DeviceNet connections are capable of producing their data based upon a connection timer (cyclically), upon a change-of-state of the I/O or at the application’s discretion. With this flexibility an end device may set up a myriad of communication links that are useful in architectures including Master/Slave, Peer- to-Peer and Distributed Control. This paper will present the communications links that can be established using DeviceNet connection objects and give examples as to how these communications links can be applied in the various network architectures.

This paper investigates the use of new alternative transport media for CAN, with specific focus on narrow band systems. Designing a physical layer for these kind of media implies using modulation techniques on a carrier wave. In an introductory study we give an overview of different digital modulation schemes and compare their performance. We identify the additional requirements imposed by CAN on the choice of modulation scheme and discuss some possible complications. As an application of the and results following from this study, we present a concrete design of a new narrow band physical layer for CAN, using the mains wiring as transport medium.

M3S stands for Multiple Master Multiple Slave and is a system concept designed to improve access to assistive technical devices by disabled people. It is a proposed standard architecture for general-purpose integrated and modular systems, which specification is available as an open standard. It is based on an industry-standard digital communication bus, the Controller Area Network (CAN), and includes additional signal lines to increase system safety and integrity. M3S provides a standard interface between input devices and end-effectors, allowing devices from different manufacturers to be linked in the same system. During revalidation phases this facilitates the process of evaluating different input devices and making the decision what input devices are optimal for a specific user. Furthermore the M3S architecture enables the user to operate more end- effectors, from categories like mobility, manipulation, environmental control and communication, using a single input-device.

The integrated transmission of information and power is a very important point of interest and a demand of many industrial bus system users. First results of such bus systems especially for simple binary sensors are known. The main task of our research activities is the integrated power transmission (IPT) in CAN systems even for analogue sensors with features of signal conditioning and data pre- processing. The paper presents a solution of such a bus node. Beside the conventional data processing a power management is necessary. In relation with the abilities of the node two methods of integrated transmission of information and power will be explained. The advantages and disadvantages for several application fields will be shown. Finally the results of the investigation of dynamic behaviour, error behaviour, and power capacity will be summarized.

The ever-increasing complexity of today’s electronic products creates a continuing demand for more sophisticated diagnostic systems, whilst the need for a simple user interface remains unchanged. A popular approach to system diagnostics involves the use of diagnostic fault trees. These can provide the product designer with a simple mechanism for the testing of operational parameters and a clear graphical display of the paths within the diagnostic sequence. This paper discusses the mechanisms that have been implemented to allow the use of fault- tree diagnostics on CAN bus systems. Also discussed are the additional capabilities that the diagnostic system can provide on the CAN network, and other ways of implementing diagnostic systems on existing software platforms.

The Controller Area Network (CAN) is a communication bus for message transaction in small-scale distributed environments. Continuity of service and bounded and known message delivery latency are requirements of a number of applications, which are not perfectly fulfilled by existing networks, CAN included. One key issue with this regard, is that networks are subject to failures, namely partitions. However, most of non-critical applications can live with temporary glitches in network operation, provided these temporary partitions are time-bounded. We call these periods of inaccessibility. Should one call for hard real-time behaviour, a worst-case figure for these non-negligible periods should be derived and added to the worst-case transmission delay expected in the absence of faults. This paper does an exhaustive study of CAN inaccessibility characteristics, presenting figures for intervals in CAN operation when the network does not provide service, allthough not being failed.

MCS-5 is a new, decentralized CAN-based automation system used mainly in ship automation. The system can be divided into two layers: monitoring and control, and process. This paper describes the function and the automation devices for each layer. The main points of the paper are the communication principles, protocol gateways and the use of PLCs within the system. The protocol layers and the communication services are explained. The paper also gives an overview of the PLC languages used. The trend to decentralized automation systems significantly influenced the development of the new MTU Monitoring and Control System, MCS-5. The new quality of this automation system is data processing within the decentralized automation devices and data exchange between them over field bus systems.

As today’s systems are becoming more and more complex, simulation is often the only viable way to verify the functionality of a system, or to estimate its performance. Especially in time and money critical sections it is important to gain information about a designed system before any expensive hardware is to be implemented. Using an object oriented simulation framework eases the solving of this problem. Different CAN components were developed separately and are now available as a CAN part library. Complete and heterogeneous systems can now be simulated and evaluated by taking parts from the library and connecting them using the standardized interface from the simulation framework. Configuration of the simulation is supported by a simple to use description language. The gained results are presented in the last section. We use the SAE scenario to show how the simulation results can be used to dimension a CAN network.

"We will describe a system, that uses CAN as a medium to connect different control devices on board of vehicles of the public transport system. Among the connected devices, tehre are main board computer, graphical display panels for teh information of the driver and a wireless data communication to a local traffic control center. Possible future extensions are the integration of a high speed infrared communication or measuring the position with a GPS receiver. We will give a technical overview on this system. Especially the implementation of an open system approach, based on the results of the CiA working group CAN in mobile applications will be described. There will be also a discussion of the advantages using an open approach, and possible reasons why not using an open approach. The authors are involved in industrial projects concerning these topics. They are also active members of the CiA working group: ""CAN in mobile applications""."

A distributed real-time computer system consists of several processing nodes interconnected by commlunication channels. In a safety critical application, the real-time system should maintain timely dependable services despite component failures or enfironmental changes such as transient overloads. In this paper the imprecise computation technique is integrated with fault tolerance schemes and adopted as the Adaptable Management System (AMS) which adjusts the operating strategy of the real-time system in response to changing application environments and the internal fault patterns. The timing response performance of the AMS for a Controller Area Network (CAN) based distributed real-time system is evaluated as to whether the timing constraints is satisfied for normal, overload, and degraded operational modes. The simulation study employs the Society of Automotive Engineers (SAE) real-time control system benchmark as the workload model. The simulation results also show that the quality of service of the AMS can be improved by varying the optional message in the workload.

CAN is well-suited to distributed real-time control of electric passenger vehicles. This paper describes the use of CAN to provide a tightly-coupled dual drive system, with real-time control distributed over the two drives. In order to achieve this reliably, the application layer has been extended. In order to maintain the integrity of periodic variables a Promptness feature using producer-consumer semantics is available as well as time-out mechanisms on inter-application events. Furthermore, a Global Data philosophy has been adapted where all data shared between the applications are tagged with a unique ID. This has many benefits as well as making it easy to monitor the overall system state via a Bus Analyser. The above approach can be used in and extended to a wide variety of performance-critical advanced vehicle applications.

The design of a CAN based system can be a complex task. Several issues need to be addressed: choosing the right modules, verifying that the chosen modules really are able to communicate, assignment of bus identifiers, verifying that the real time requirements of the system are met, selecting correct bus parameters, and so on. Kingdom Founder is a graphical design tool which aids the system designer in this process. Fetched from a database or defined on the fly, the modules and the CAN bus are drawn on a worksheet. The information structure of the system, as well as several module and system parameters are then defined. The system designer can then use the program to check for errors or potential problem sources, produce documentation and generate system startup information and code skeletons for the modules.

In this paper, we describe the hardware and software architecture of an autonomous mobile robot. ROMAIN, INstrumented Autonomous Mobile RObot, is an experimental platform dedicated to a wide range of applications. Its main characteristics are flexibility, high speed velocity, energetic and decisional autonomy and real time processing. This mobile robot owns an open structure based around a Control Area Network for data transmission between microcontroller cards and a real time operating system. After the description of this structure we propose an application consisting in the tracking of an other robot in an unknown environment with ultrasonic sensors.

During the development of the DeviceNet Specification, it became apparent that it would be necessary to provide a mechanism for common user friendly configuration of all DeviceNet products. It would have to be robust enough to cover a wide range of products, from the very simple, to the very complex. It would be required to be used by arbitrary configuration tools and be independent of the implementation of these tools. The configuration mechanisms would also have to be easy to implement in order to facilitate its adoption by product developers. It would also have to be cost effective to implement so as not to add excessive cost to inexpensive devices. This paper describes how the Electronic Data Sheet achieves these goals.

DeviceNet has emerged as one of the low level fieldbusses optimised for industrial control. It uses the robust and powerful CAN (Controller Area Network) technology as the backbone. Interoperability between various DeviceNet devices, advanced failure prevention and fault diagnosis, and lower implementation costs are some of the immediate advantages of DeviceNet. Interoperability brings the issue of compatibility and conformance among the DeviceNet devices from various vendors. This paper describes the issues involved in implementing a DeviceNet system both from developers' and end-users' point-of-views. It investigates the 'plug-and-play' and interoperability of DeviceNet devices. A study for realising a fully automated compliance test for DeviceNet is done.

The development of distributed architectures that rely on communication protocols for data exchanges between remote ECUs produced by different equipment manufacturers, sets new problems for car manufacturers who have the responsibility of the whole system. At each communication level, the car manufacturers have to be sure that the protocol implementations are compliant with the chosen standard. The CAN standard is a Data Link Layer protocol generally implemented in ASICs. So, there is a need for tools verifying during the design process that all the components produced, today and in the future, by a lot of different silicon suppliers, are and will be compliant with the CAN standard.

The Controller Area Network (CAN) is a serial bus with high speed, high reliability, and low cost for distributed real time control applications in both the car and the industrial environments. In an industrial environment, while the CAN is used by manufacturing sections to control the systems, the management department can use Ethernet (or Token Ring) LAN in its building. This implies that the CAN should communicate with a LAN. A solution is to use the internetworking devices to connect a CAN and a LAN. Bridges are high performance devices that are used to interconnect two similar or dissimilar LANs. The aim of this study is to design and implement a bridge which is capable of connecting a CAN and an Ethernet LAN. In the following, a brief overview about the CAN, the Ethernet, and bridges is presented. The required processes are detailed and the implementation of the bridge is explained in section 2 and 3, respectively. In section 4, the modelling environment for simulation is summarized. Finally, the results obtained from simulation are presented in section 5, and the conclusion is given in section 6.

Honeywell has been a long-term participant in the industrial control marketplace. Participation in this market has kept Honeywell in constant communication with customers concerning issues and needs in the marketplace. Over the past five to seven years, several key trends have driven requirements for change in control system architectures. One of these trends is increased levels of automation and decreased levels of available human resources. A second trend is decreasing product and technology life-cycles which have made fixed automation systems cost prohibitive. These trends have driven automation users to require systems that can be modified, adapted, or upgraded as required to sustain a long-term competitive position. Control systems must advance to provide information to assist users in the maintenance of complex automation systems. These market trends provided the force to encourage Honeywell to develop the Smart Distributed System. The Smart Distributed System is a device level network which is designed as a control platform for use in a wide variety of industrial machine control applications. Honeywell has now expanded the scope of the Smart Distributed System to include Personal Computer (PC) hardware and software to execute real-time control and provide Operator Interface software. The PC Control platform provides a mechanism that is well suited to take full advantage of information available from the networked Input/Output (I/O) system.

Since CANopen communication and device profiles are available the influence of this CAN communication standard is increasing thoughout the CAN in automation arena. Various CANopen-based components as well as OEM protocol implemenations are available. This contribution concentrates on experience drawn from software implementations on different devices and set up of CANopen networks. The software structure of network master and slave devices is presented in detail together with realisation hints. Achieved performance in terms of code execution time and consumption of hardware resources are described. The configuration of CaNopen devices and set-up of complete networks by the user forms the second section. Tools are presented which allow the access to object dictionaries of devices via electronic data sheet and parameter up- and download. Graphical user interfaces allow the layout of CANopen networks, support the distribution of identifiers and the cooperation of bus nodes. Finally, an application example in form of a multi-vendor system with several controllers, decentralized Input/Output devices and multiple drives is given.

The multiplexing systems currently being implemented are essentially information sharing systems. In the body system in particular, it is possible to reduce the cost of implementing features by growing one of the micros into a 16 bit device and slimming down the others. The bus will carry commands from the 16 bit master device as well as information. Many of the tasks in the body system do not need to run often. The 16 bit device can make use of large (economical) flash memory for reprogrammability, but in order to benefit from reprogrammability, the whole bus needs to run under a communications manager that will guarantee the real time performance of the bus when program updates occur. This type of development will further divorce electronics hardware from software, and will require organisational changes in companies to implement it effectively.

The paper gives an overview over the object-oriented methodologies including related workbenches, which were developed especially for modelling of distributed reactive systems. The applicability of this tools will be discussed with a concrete motion control application based on CAN . Also some aspects of hardware/software co-design of such kind of systems regarding to the quality and efficiency of the development process will be part of the presentation.

Machine control software are in general highly embedded in target systems. Distributed processors and standard CAN-buses have provided suitable tools to design machine diagnosis and fault tolerance based on user requirements or safety aspects. In this paper some general and on target machine designed diagnosing elements and monitoring functions are explained. The surveyd target machine is just developed for loading explosive into holes in mining applications. Because of highly dangerous work environment machine fault detection and diagnostic maintenance are of crucial importance. Both continuous operation and rapid responses are peculiar to the developed machine. On the basis of real time maintenance the error sources are divided in critical requiring immediate user operations and in slowly developing faults caused mainly by component wearing. The distributed computer architecture provides means to implement fault tolerating systems and especially CAN contollers facilitate to embed diagnosis messages both in measuring data and in special safety functions. The distributed control makes it possible to diagnose the functions of the machine more effectively, but, however, on the other hand the distributed control is a new and a serious risk factor. Without sufficient checking intelligent subsystems may behave unexpectedly: erroneous interpretation of system state or fault sensor can trigger the actuator on at wrong time with dangerous consequences.

This paper describes the introduction of new concepts into a tin can manufacturing line and its re-engineering with the aim of doubling the production speed. The design of the hardware and control system played an important part in matching the dynamics of the machinery to the production processes, and also in reducing costs and improving reliability.

This paper gives a survey on the historical background of CAN. Additionally, the paper handles different CAN controller implementations and discusses the use of CAN in automotive systems and in automation, including initiatives for open systems in both areas. Although originally developed for in-vehicle purposes, automation area started to apply CAN as soon as the first components became available. Since the introduction of the first CAN chips the sale of CAN components has been remarkably higher in automation than in automotive sector. However, the roles are supposed to change soon due to the growing use of CAN in automotive electronics.

Electronic control units (ECUs) linked together within an automotive network are generally supplied by different companies, and mostly include different micro-controllers and software architectures. Moreover, applications provided by different suppliers have to co-exist in a single processor in the future. Today's individual development processes for distributed, communicating ECUs hinder the integration of automotive systems and increase the overall costs. In order to achieve a significant reduction of these costs, services and protocols for communication, network management, and operating system must be standardised. The OSEK/VDX group worked out a respective specification in coperation with several car manufactuerers and suppliers. The specification will permit a cost-effective system integration and support the portation of system functions between different electronic control units. This paper gives a brief summary of the OSEK/VDX results and an outlook on the continuation of the project.

"""Objectmodul"" is not just another layer 7 for CAN networks. It is a universal SW Interface best fitting in distributed real-time systems and decouples the application from the different ways of transporting data. So the application can be developed independent of the layers (CAL, Devicenet, SDS, CAN Kingdom, ...) of the network (CAN, RS232, Dual port Memory, ...). The existing ""Objectmodul"" is implemented as a standalone SW module, but it is also possible to integrate it as an extension of an operating system. By using the ""Objectmodul"", the applications on various CAN nodes do not need any knowledge about the topology. This means, no knowledge is needed about the I/O is local or distributed by the network. The application just reads or writes objects and the ""Objectmodul"" controls the handling of the data. So the application program is completely independent of the way the data is transferred, the type of communication partner or data layout."

This paper discusses the design and performance requirements for the interconnection of wireless nodes to a standard CAN bus environment. It is focused on the network architecture as a single centralised cell consisting of a central node and wireless terminals. The central node provides data communication between the CAN and the wireless network. It maintains a self-learning algorithm for the frame filtering and forwarding requirements. The Medium Access Control Protocol used for the wireless network is specifically designed for supporting prioritised frame transmission in the wireless channel. The prioritisation is accomplished by using different sequence numbers which are added to each individual frame based on the CAN identifier. The performance of the system is analysed and presented in terms of probability, delay and channel utilisation under various traffic conditions.

A chip set for highly dynamic servodrives has been developed. Ist first application is a very compact, intelligent, CAN-controlled brushless AC servodrive. It can, however, be used for DC, brushless DC and induction servodrives as well. Main tasks of the chips are the calculation of rotor position, torque and current control as well as pulse-width modulation. The chip set not only enables the construction of economic and intelligent servodrives with high dynamics but, in addition, it considerably improves the reliability of the drives by drastic reduction of the component count.

As in other branches a central automation concept is the state of the art in industrial fluidic systems. When one wishes an open communication with CAN the problem arises to choose between several major Higher Layer Standards, such as CAL, CANKingdom, DeviceNet or SDS. In order to make open communication more efficient the idea is to have a Higher Layer independent API for fluidic applications (profile). Due to several Higher Layer Standards for CAN it is advantageous to divide such an API into a Higher Layer determined part and an application determined part. A proposal will be given in the presentation. We will show, that an open communication which fits the needs of fluidic systems causes a change of its automation structure. It is determined by decentralized devices with processing capabilities (e.g. device configuration and self diagnosis) which can fit to a special application by IEC 1131 compliant configuration facilities.

In recent years systems of CAN-connected automotive control units have been increasing rapidly both in complexity and size. Therefore not only functional but also performance aspects must be considered during development. For the system to be feasible, all communication timing constraints for all functions have to be proven. This feasibility has to be assured in early design stages already. This paper presents simulation results of a bus model as a basis for performance analysis of automotive CAN systems. The single server queue with non-preemtive priorities was chosen as a simple model of the bus. Each priority class represents a message of a certain identifier. Bus arbitration is implicitly modelled by the non-preemtive priorities of the service discipline. The simulation model considers not only cyclic messages (messages with constant interarrival times) but also sporadic traffic sources. Those sources emit only when their corresponding function is active. The obtained simulation results give the mean waiting time for the different priorities and can be used to tell if the system with the planned communication load can work acceptably for all realized functions. If necessary the design can be changed in an early stage to improve network performance.

MiniCAN is a concept that combines CAN compatibility with simplified functionality. It is optimised to attach basic components directly to a CAN bus. Those components are e.g. switches, lamps or sensors and actuators that do not require more than 6 bits of data. It is a master-slave system that uses an inframe response within the CAN data field to run a simplified protocol. Special concern was taken for security, low cost and ease of use. It may be used both for low and high speed communication, compliant to CAN specification 2.0, parts A and B.

Designers of soft real-time systems often ignore worst-case behavior analysis when designing their systems. We have developed a discrete-event model of Controller Area Network (CAN) to assess performance under both worst-case and normal conditions. The analysis revealed that many network events and attributes can lead to message serialization, which causes large network delays typical of worst-case behavior. This transient network behavior has serious implications on the application performance. In this paper, we present these results and show how the actual performance of the system varies over time between the normal and worst-case scenarios.

This paper deals with the analysis of real time systems subject to the time delay in the communication networks. Different approaches to handle the effect of communication delays are discussed. To improve the robustness of the control system, a fuzzy controller is used. The use of the fuzy technique is motivated by the fact that time delays give rise to phase lag, which often degenerate system stability and performances. The case of the Controller Area Network bus is particularly discussed.

"""In this paper we study the wireless communication extension for the Controller Area Network (CAN) protocol to suit industrial applications. Two different network topologies and medium access methods have been considered. The remote frame and the prioritised frame medium access control (MAC) methods are proposed for the centralised and distributed wireless CAN based network. The performance of these protocols is evaluated by simulating the protocols in the wireless CAN network. The “SAE Benchmark"""" is used as the workload to illustrate the industrial applications of CAN based system. This paper discusses the applicability of the proposed wireless MAC protocols and confirms its usefulness for real-time communication base on the benchmark."""