In recent years, the product life cycle has decreased and demands for new products have emerged due to competition, modern industrial needs and rapidly changing technology. This has necessitated changes in design, development and manufacturing processes so as to improve quality and efficiency as well as reducing costs. Computer Aided Design (CAD) helps to meet this challenge in the design evaluation and final product design stages. This paper presents the development of an interactive software for the optimal design of a motor intended for variable speed applications. The use of finite element analysis methods is proposed as an indispensable part of the CAD system for electrical machine design. An illustration of the method is given for the design of a switched reluctance motor.

Programmable logic controllers (PLCs) are commonly programmed using relay ladder logic. There are many disadvantages to this approach. The state transition diagram affords a systematic means of providing precise and unambiguous specification while at the same time yielding a good system overview for ease of troubleshooting and maintenance. Computer aided logic design techniques help to automate the process of relay logic design for the state transition diagram. A software package which runs on an IBM PC/compatible machine has been developed to automatically transform the state diagram specification into relay ladder logic and the associated (PLC dependent) source code.

The aim of this research is to develop the system for generating the robot assembly task program by using the data which are automatically designed when the design specification is given, such as the input data of NC machine is made by using the data designed in CAD/CAM system. In this paper, the fundamental research for this purpose is described. When the internal pressure of a pressure container is given, shown is the method by which the robot assembly task program is generated using the attributes of the through bolts, which are designed by calculating the allowable stress of its material.

A computer aided engineering (CAE) approach is presented to study, investigate and verify the design primitives of a multiprocessing system for speedy engineering process. The multiprocessing system considered here consists of N processing elements, communicating with one another through a shared memory. An appropriate analytical model to represent the mechanism of the multiprocessing system is proposed to realize the CAE package. The proposed model meets with the requirement of being time efficient to run on a computer, which is necessary for a CAE system. The package is interactive, accepts various system design parameters and produces results in quantitative and graphical form. Typical results obtained from CAE package are the average system throughput, average processing element utilization, and average backlog etc.

This paper presents CAD system named DPACS, which has functions not only CAD but also real time contro] and test. For the CAD, it has functions of control system analysis, design, simulation, identifcation and data handling and management. A real time control program based on CAD data is produced automatically and the digital control is possible by using A/D and D/A converter even if an operator doesn't know the programing technique or any programing language at al]. The details of CAD system is explained by the example of designing a controller for the inverted pendulum system.

Multivariable linear, time invariant control system design invariably involves loop shaping. When it is necessary to shape several time and frequency responses simultaneously and keep gains within prescribed bounds, classical design techniques become overpowered. Fortunately, at this point the designer can call upon semi-infinite optimization (see [Pol.3]) which is a natural tool for solving control system design problems that require squeezing time and frequency responses into preassigned tubes, as well as for taking plant uncertainty into account. Semi-infinite optimization can also be used to carry out trade-offs when all the design requirements cannot be satisfied at the same time. The use of semi-infinite optimization in control system design is most successful when carried out in an interactive computing environment.

To date, applications of artificial intelligence in control system analysis and design are primarily associated with the design process. These applications take the form of knowledge bases incorporating expertise on a design method, such as multivariable linear controller design, or on a field such as identification. My experience has demonstrated that, while such expert systems are useful, perhaps a greater benefit will come from applications in the maintenance of technical databases, as are found in real-time data acquisition systems, and of modeling and design databases, which represent the status of a computer-aided design process for a human user. This reflects the observation that computers are best at maintaining relations about large sets of objects, whereas humans are best at maintaining knowledge of depth, as occurs when a design option involving a sequence of steps is explored. This paper will discuss some of these issues, and will provide some examples which illustrate the potential of artificial intelligence.

This paper proposes a new method of inductive learning that incorporates an element of inexact reasoning. The first part of the paper is devoted to the critical analysis of existing methods of inexact reasoning (Certainty Factors, Bayesian Inference, and Dempster-Shafer belief functions). In the second part we propose and describe a new method for inexact reasoning and its application in the domain of knowledge acquisition.

In this paper we discuss how to generate the command value for the optimal regulator in an automotive antiskid system. First, the behavior of the vehicle at braking is expressed as a mathematical model with the formulation by physical consideration and identification of the hydraulic system by statistical methods. An optimal regulator with additional integral is applied to the automotive antiskid control in order to make the each wheel speed follow any command value. However, the desired command value to stop the vehicle efficiently and stably is dependent on ambiguous road surface conditions. Thus, how to determine the desired command value under the moment-to-moment conditions is most important. A method for inferring the conditions is developed using fuzzy logic, with three fuzzy variables expressing the conditions adequately. On the basis of the inference, the ideal command values are generated. Outstanding control performance and good adaptability are obtained in vehicle experiments. Consequently, the Expert Antiskid System, employing modern control theory and fuzzy logic, can stop a vehicle efficiently and stably under any condition.

This paper describes a frame-based knowledge representation system for the design of systems involving tranfer of mechanical motion. The system has a graphical interface to produce its output in graphical form. The structure of the frame representing any mechanical object and its location relative to other objects in space is given. The system performs first-hand mechanical checks before proceeding in the design. The sytem has been built in an open-ended way so that it can be used to design even larger mechanical systems without major changes.

Dynamic information, a specific feature of industrial applications on process control restricts the use of expert system in this field. Nevertheless, it allows the improvement of the quality of the knowledge representation, the reduction of possible combinations when the solution is being searched, and as the events occur at different times, it also allows, at any moment, the limitation of the amount of data to be processed, which is particularly interesting when a crisis happens. This paper describes an expert system for process control that copes with dynamic information and reasons intelligently on any new events.

This paper describes the different problems in process control at the level of the interface between the operator and the supervisory devices. We emphasize graphic man-machine communication and evaluation of graphic presentation of information, which seem very determinant for us. The aim is to use the results of this evaluation to fill out knowledge bases of SYNOP, an expert system, based on a first-order logic inference engine, for the ergonomic conception of industrial control views.

A rule-based, digital, closed-loop controller that incorporates 'fuzzy' logic has been designed and implemented for the control of power on the 5 MWt MIT Research Reactor under both steady-state and transient conditions. Based on this experience and having designed several other controllers for the same purpose, a comparison is made of the rule-based and analytic approaches. Differences in the division of labor between plant engineers and control specialists, the type of knowledge required and its acquisition, the use of performance criteria, and controller testing are discussed. The design, implementation, and calibration of rule-based controllers are reviewed with specific examples taken from the completed work at the MIT Research Reactor. An evaluation is then made of the possible role of rule-based technology in process control. It is noted that there are no comprehensive guidelines for the design of rule-based controllers and that such systems are quite difficult to calibrate. The advantage of rule-based systems is that they are generally more robust than their analy-tic counterparts. The rule-based and analytic technologies should therefore be used to complement each other with rule-based systems being employed both as backups to analytic controllers and as a means of improving the man-machine interface by providing human operators with the rationale for automatic control actions.

Designers and developers set out to create something, a hardware product, a computer program, an automated manufacturing application, or a piece of documentation. Their premise is that when developed and debugged, this creation will be good .... and that is as it should be. That premise fails to ask the question, "Under what conditions will this creation fail to perform as I have created it to do?" Too often there is an assumption that debug equates to testing. They are not the same. A well designed test should determine under what reasonable conditions the product will fail to perform as intended.

A print quality evaluation method has been developed and applied to inspection of wire-dot matrix printers. Newly developed algorithms adopt human characteristics such as edge emphasis, logarithmic sensitivity in optic nerves, and progressive weighted total evaluation. Experimental results show that the capability of the print quality evaluation system is on the same level as expert operators and that reproducibility is three times higher.

Distributed computer processing has led to the need for data communications over reliable links. This paper discusses the use of fiber optics in Manufacturing Automation Protocol (MAP) factory networks -- its potential and limitations. A fiber optic MAP system is described which provides completely compatible fiber optic interconnections for all 10 Mb/s broadband MAP equipment. The system uses innovative techniques to support broadband MAP as well as other broadband services such as video.

MAP, a communications protocol for factory automation proposed by General Motors [1], has been accepted by users throughout the world and is rapidly becoming a user standard. In fact, it is now a LAN standard for factory automation. MAP is intended to interconnect different devices, such as computers and programmable devices, made by different manufacturers, enabling them to exchange information. It is based on the OSI intercomputer com-munications protocol standard under development by the ISO. With progress and standardization, MAP is being investigated for application to process control fields other than factory automation [2]. The transmission response time of the network system and centralized management of data exchanged with various devices for distributed control are import-ant in the case of a real-time process control with programmable controllers, computers, and instruments connected to a LAN system. MAP/EPA and MINI MAP aim at reduced overhead in protocol processing and enhanced transmission response. If applied to real-time process control, a protocol based on point-to-point and request-response transactions limits throughput and transmission response. This paper describes an advanced MAP LAN system applied to real-time process control by adding a new data transmission control that performs multicasting communication voluntarily and periodically in the priority order of data to be exchanged.

This paper deals with the study of LLC services for field buses in the case of FIP proposal (FIP : Factory Instrumentation Protocol). Various solutions are proposed. They are compared from the point of view of the error detection and of efficiency. The final solution is presented.

We show an automatic task or plan generator which is able of finding the solution of a problem without knowing all the information about its current state. Therefore, for achieving the proposed goal, modifying actions over the real world, which will be carried out through an industrial robot, are not enough. Also searching information actions are essential. This is accomplished by sensorial devices. The chosen knowledge representation is a new production system, that is expanded by the parallel planning method and is modified by the addition of an optimizing level composed by heuristics or evaluation functions. The plan generator allows to solve problems that have interacting subgoals without resorting to backtracking, reordering partial plans and patches.

In this paper, we give a comprehensive review of a number of practical problems associated with the use of static priority scheduling. We first present a new approach to stabilize the rate-monotonic algorithm in the presence of transient processor overloads. We also present a new class of algorithms to handle aperiodic tasks which improve the response times to aperiodic tasks while guaranteeing the deadlines of periodic tasks. We then study the problem of integrated processor and data I/O scheduling. Finally we review the problem of scheduling of messages over a bus with insufficient priority levels but with multiple buffers.

The reliability of plant operation is greatly improved by utilizing selected data for display in real-time on the screens of 12 color CRTs in the Central Control Room and voice information. The load on operators is reduced by automated operation of plant equipment and an operation guide. Information from the power station is transmitted to an off-site location via an optical fiber data way which makes it possible to display in real-time the same CRT screens displayed in the Central Control Room.

The Magnetic Supported Intelligent Hand (MSIH) is a robot hand for precision automated assembly. This paper describes the conception of the MSIH and explains its structure, control system, functions and experimental system. The MSIH has a unique structure incorporating a magnetic-bearing wrist mechanism, in which a floating body, holding a work piece, is supported by controlled magnetic forces without any mechanical contacts. The MSIH has three major functions: 1) the advanced Remote Center Compliance function; 2) the precision-actuating function of all three translations and two rotations of a work piece; 3) the sensing function of forces, moments and displacements. The MSIH can achieve precise insertion operations in various levels and conditions automatically by selecting or combining these functions.

The authors have developed a new moving data storage system where the data carrier (DAC) is mounted on a moving object. This DAC has a memory where data can be read and written using electromagnetic waves. This paper describes the configuration, the functions, the operating principles, and the performance of this moving data storage system. This paper also deals with the technology for miniaturization, low power consumption, and life extention of the DAC.

The paper describes a system in which a sensory-equipped, two-degrees-of-freedom gripper and its controller form an autonomous unit. The aim was to create a system capable of local decision-making, working within the hierarchically organized robotic cell. The microprocessor-controlled gripper system makes decisions based on the sensory information and on partial knowledge of both the task and the status of the cell. Specific reactions of the gripper to sensory inputs can be consistently repeated in a whole range of different tasks. Moreover, the system processes the majority of the sensory signals, leaving to the cell supervisor only the control of the flow of the main task. Elementary reactions are determined by responses of the servo loops, while other, more sophisticated reactions may be acquired by learning. It is shown how the capabilities of a relatively simple multiple-loop analogue controller benefit from this feature.

The CIM's view of manufacturing as a system, working towards common goals through the sharing of common information, should guide any design, analysis or retrofit efforts related to production systems. The line/cell information model presented here, which is based on the concept of an irreducible set of interacting information objects, reflects such a view. It allows access to information from higher levels of system hierarchy. CAD and CAM files supply the model with the context necessary for meaningful line autonomy and real-time reporting. Evolving knowledge about the line functionality is supported by an abstraction/inheritance class mechanism. The model is suitable not only for integration of hi-tech production technology, but also for use in traditional assembly lines (as shown in the implementation example).

Designing FMS control is a difficult task which necessitates a hierarchical system, a clear structuration of the functions and a test of the programs obtained in a simulated but realistic environment. This paper presents a methodology for specifying and emulating the control and the process behaviour on a distributed structure.