Annual Review in Automatic Programming最新文献

This paper presents an architecture to build distributed systems for the process control. The agents share information by the communication tools developed in Unix. In this way it is possible to link expert systems with other processes for the sensor input/output or interface with the world. At the same time modification for a new CLIPS structure are proposed in order to build Blackboards.

Processes with (partly) unknown or complex dynamic behaviours need complex control schemes. Neural networks offer interesting perspectives both for identification and control of these processes, because neural networks can approximate any (non-linear) continuous function. Especially adaptive control using neural networks offers good possibilities, due to the possibility to learn online. The NECTAR-project (Neural and Expert ConTrol of AircRaft) aims at studying both neural network and expert system structures for highly demanding control environments. In this publication the general structure of the project will be sketched, after which the neural network part of the project will be discussed in more detail. The final architecture will be implemented on-board of a laboratory aircraft to demonstrate the possibilities. Some preliminary results are reported.

This paper describes the configuration of real-time tasks using a PEARL-based programmable controller (“PEARL-PLC”). It will be shown that tasks can be consistently configurated with function-blocks for logic-, sequence- and feedback-control. Furthermore, the synchronization of cooperating tasks, controlling interdependent sub-processes, can also be configurated using communication function-blocks based on Multicomputer-PEARL. Moreover, using the explicit task-scheduling offered by PEARL, the needed processor-time can be reduced to a conceptual minimum. Details of the configuration of a sequence-control task and a fuzzy-control task, the last one as a representative of a feedback-control task, are presented. Finally, trends of future developments of the “PEARL-PLC”, e.g. the implementation of fault-tolerance, are pointed out.

The ability to supervise and control highly non-linear and time variant bioprocess is of considerable importance to the biotechnological industries which are continually striving to obtain improved productivity and to reduce process variability. The proposed Intelligent Supervisory System consists of several modules, but in this contribution most attention was given to the fault detection module. Four pattern recognition techniques (Artificial Neural Networks, Principal Component Analysis, Multi-way Principal Component Analysis and Autoassociative Neural Networks) were applied to an industrial fed-batch process. It is shown that a deviation from nominal behaviour of the process can be detected even early on in the fermentation run. Data from industrial penicillin G fermenters is used to demonstrate the procedures.

In this paper, we describe a development methodology for building real-time knowledge-based systems (RTKBS) using the REAKT toolkit. In RAM (REAKT Application Methodology), we attempt to combine elements from three separate streams of ideas: Knowledge based systems (KBS), real-time systems (RT), and object-orientation (OO). The concept of application frameworks for building and reusing real-time knowledge-based systems is proposed. The methodology is illustrated with examples from an alarm handling application at an oil refinery.

Expert systems have now reached maturity and an ever growing number of applications make use of this technology. Nevertheless, specific requirements of real time applications have significantly slowed down the use of these techniques in this area: this is first due to incompatibilities between time and memory requirements of AI algorithms and real time specificities, but also to a lack of adaptation to the growing complexity of these systems.

In this paper a real time architecture for Knowledge Based Systems: REAKT (Real Time Knowledge Tool) which aims at combining the expressiveness and power of Knowledge Based Systems with the efficiency and predictability of real time systems is presented.

Real-time database management systems have become a hot topic in the research and development community of late (Fort94a,Grah93,WCPP93). In addition there has been a movement in the standards community to examine and develop extensions to existing and proposed query languages to support real-time (Fish94,Fort94,Fort94a,FS94,Gord94).

This paper examines the state of research into real-time database management systems in the areas of database structuring, transaction structuring, transaction processing, concurrency control, recovery and real-time transaction scheduling. We then extend the findings and trends of this work into the high level specification of data definition language, data manipulation language and data control language extensions for the standard SQL2 and emerging SQL3 database query languages.

MAP provides communication support for manufacturing applications and factory automation. MMS and Network Management are some applications that depend on MAP for the communication support. These applications use MAP to communicate with manufacturing devices at the cell level. In this paper, we describe the architecture of LITMAP Stack implemented at LIT and discuss its ability to support soft real-time industrial applications. This discussion is based on the scheduling ability of this architecture and the performance of LITMAP Stack.

The lack of a common framework for the uniform application of the solution techniques when developing intelligent process control system, have given rise to the HINT project (Heterogeneous INTegration architecture for intelligent control systems) whose main objectives are the design and realization of a supporting software architecture and the development of a methodology that allows the integration and cooperation of different control techniques (mostly, Artificial Intelligence-related) within a consistent computational environment. The aim is to deal with the real problems encountered in in-place industrial process control.

A real time quality control system for flat products, applied to steel strip manufacturing, is introduced. The system is able to perform this task when the product runs at its maximum speed (15 m/s). The outlines of the overall system are explained, paying special attention to the neural classifier. A new model of neural network has been developed in order to implement this subsystem. Some benchmark results are resembled in order to ease the understanding of the model.