Matjaž Colnarič

ARCHITECTURAL SUPPORT FOR PREDICTABILITY IN HARD REAL TIME SYSTEMS

Abstract There is evidence that, among all design domains of hard real time systems, architectural issues gained the lowest research interest. Universal architectures, which are generally applied as hardware bases for hard real time applications, are seldom behaving in a fully predictable wav. In the paper, several commonly used techniques which prevent temporal determinism of instruction execution are enumerated. An asymmetrical multiprocessor architecture for hard real time applications is presented, whose temporal behaviour is fully predictable. Some adequate features are discussed which are incorporated into the processors implementation.

Computational intelligence in control

Abstract This Milestone Report addresses first the position of the areas of computers, computational intelligence and communications within IFAC. Subsequently, it addresses the role of computational intelligence in control. It focuses on four topics within the Computational Intelligence area: neural network control, fuzzy control, reinforcement learning and brain machine interfaces. Within these topics the challenges and the relevant theoretical contributions are highlighted, as well as expected future directions are pointed out.

A reconfiguration pattern for distributed embedded systems

A reconfiguration pattern for UML-based projects of embedded (real-time) systems is defined. It enables to set up hardware/software configurations, and to specify conditions and methods for dynamic reconfiguration. The reconfiguration pattern was inspired by the reconfiguration management solution of the Specification PEARL methodology, which is based on the standard for Multiprocessor PEARL whose original idea it was to extend the language to enable the programming of distributed real-time applications in PEARL. In Specification PEARL, the possibility for abstract descriptions of hardware and software architectures and for defining mappings from software to hardware components has been enhanced in correspondence with the standard. Here, a UML pattern for reconfiguration management in distributed embedded applications based on concepts from Specification PEARL is presented. Its behavioural, structural and functional aspects are outlined. It addresses stereotype entities from the Specification PEARL language, which were joined in a UML profile, and outlines the related reconfiguration management mechanisms, which were carried over to the mentioned UML pattern. The proposed reconfiguration pattern is to facilitate the development of distributed embedded application in UML with consistent and temporally predictable reconfiguration support. It should also support and enhance the applications’ flexibility and portability.

Programming and Time Analysis of Hard Real-Time Applications

Abstract To assure temporal predictability in embeded hard real-time programing, layer-by-layer predictability of the system must be provided. Experimental hardware platform and coresponding operating system are being designed and built. Parallet to this, a programing environment for hard real-time systems is under construction, based on a simple structured programing language. A compiler with an integrated analyser for execution time analysis of tasks is used to get non-pesimistic run times.

Computerised controllers for safety critical medical applications

A novel design for a computer control system to be employed in safety critical applications, as found in medical environments, is presented. It features a low complexity, fault detecting hardware architecturally supporting a strictly cyclic operating mode, as known from programmable logic controllers, and a specification level, graphical programming technique based on the interconnection of application oriented standard software function modules. By design, there is no semantic gap between the programming and machine execution levels. Thus enabling the safety licensing of application software by an extremely simple but rigorous method, the software verification problem is satisfactorily solved for a large application area, where failures may cause hazards to, or even loss of, human safety and lives.

Some basic ideas for intelligent fault tolerant control systems design

Abstract Some basic ideas underlying the EU project IFATIS (intelligent fault tolerant control in integrated systems) are described. Faults in the plant (controlled system) as well as in the controllers (computer systems) are taken into account. Plant modules and controller modules are considered as resources. Needed application functions (partial processes) are to be allocated to such resources to run in a normal or degraded mode. In case of fault induced resource reduction, resource managers on different levels can take care of arbitration and allocation, thus providing for unrestricted continuation or graceful degradation. Some outlines concerning implementation and hardware/software architecture are presented.

Issues in the Implementation of a Fault-Tolerant Hardware Platform

Abstract In the paper, some discussion of the implementation of a fault-tolerant hardware platform is presented. The discussion is focused on the control applications with less severe integrity requirements (SIL1) and where the intelligent system reconfiguration is preferred (instead of the redundancy) in the case of failures. To build a dependable control system, all aspects of the implementation must be considered and integrated into the development from the beginning. A network of simple monitoring modules should be integrated into the system to detect and react to faults as soon as possible. To streamline the implementation, not only the hardware, but also the appropriate system software, must be constructed to hide the particularities of the low-level matters from the control application.

Dealing with exceptions in safety-related embedded systems

Abstract In embedded hard real-time systems, tasks must complete their executions within predefined time frames. A necessary pre-condition to achieve this requirement is predictability of their temporal behaviour. Here, the main focus is on handling exceptions in such systems. When handled in a classical way, they necessarily jeopardise the ultimate requirement, temporal predictability. Hence, it is argued that exceptions must be either prevented or avoided, as far as this is possible. For the remaining non-preventable and non-avoidable catastrophic exceptions, a technique in form of syntactic means is presented allowing to handle them in a well-structured and predictable way, and as painlessly as possible. The technique is based on recovery blocks with pre- and post-conditions. Finally, a method for the estimation of the resulting temporal behaviour (worst case execution time) is described.

Object Orientation in the Real-Time System Lifecycle

Abstract The design of the hardware, the software and the operating system for real-time applications is studied. To achieve the necessary properties and performance, it is proposed to co-design the constituents in a holistic way. An attempt is made to unify the design and development approaches. A modified spiral model of the design lifecycle is adopted where specifications are gradually refined, and prototypes generated to allow for verification and validation of both their functional and temporal properties. For the design in all domains, the object oriented approach was utilised.

A Real-Time Programming Language as a Means to Express Specifications

Abstract A possibility of using a real-time programming language to formulate hardware and software specifications is elaborated. Instead of using strict formal specifications, a system is described in a simple and straightforward manner using a terminology which is close to the application programmer and his or her way of thinking. The description is a mixture of clauses in syntactically correct formal notation and natural language inserts. The specifications are then stepwise refined until a program in the real-time programming language miniPEARL is obtained. Its syntax is based on the standardised language PEARL for Distributed Systems, of which certain constructs are renounced and to which some others are added to support the necessary features of the consistent hardware model to be employed in real-time control.