Anders P. Ravn

Specifying and verifying requirements of real-time systems

An approach to specification of requirements and verification of design for real-time systems is presented. A system is defined by a conventional mathematical model for a dynamic system where application specific states denote functions of real time. Specifications are formulas in duration calculus, a real-time interval logic, where predicates define durations of states. Requirements define safety and functionality constraints on the system or a component. A top-level design is given by a control law: a predicate that defines an automation controlling the transition between phases of operation. Each phase maintains certain relations among the system states; this is analogous to the control functions known from conventional control theory. The top-level design is decomposed into an architecture for a distributed system with specifications for sensor, actuator, and program components. Programs control the distributed computation through synchronous events. Sensors and actuators relate events with system states. Verification is a deduction showing that a design implies requirements. >

Device Monitors

A driver is the part of an I/O system used for processing of an I/O request for a specific channel. The interaction of the CPU with a channel is described through the monitor concept of Hoare and Brinch Hansen. The implementation of monitors using hardware interrupt facilities is described. The resulting device monitor is compared pared with the device processes of Wirths Modula. The concept is illustrated through an extension to Concurrent Pascal with examples drawn from the PDP11 system. Problems of missing interrupts and power failure are also discussed.

Hybrid Control of a Robot - A Case Study

An experiment with a distributed architecture to support a hybrid controller for a robot is described. For a desired trajectory, the controller plans a schedule for switching between a fixed set of control functions. Initial results indicate that the proposed architecture is better at achieving a desired trajectory than conventional control algorithms. The experiment also illustrates a division of concerns between software engineering and control engineering. Development of controlling realtime state machines and their mapping to processors and network is the task of software engineering, while the control engineer must identify plant phases, switching conditions and relevant control laws. These define algorithms for a planner and for the control functions. The format of schedules produced by the planner and the algorithms constitute the interface to software developers.

Design of embedded, real-time systems: developing a method for practical software engineering

The methodological issues and practical problems in development and industrial use of a theory-based design method for embedded, real-time systems are discussed. The method has been used for several years in a number of smaller industries that develop both electronics and software for a professional market. The design is expressed in a notation for communicating sequential processes, while data types and operations are expressed in a notation built on mathematical set theory. The authors present an order in which to use the notations, a technique for deriving states and operations, and a method to provide systematic checks of a design with respect to system requirements.<<ETX>>

Pointer variables in concurrent Pascal

Concurrent Pascal and similar languages are inefficient when used to program systems where processes exchange large messages. The inefficiencies, caused by copying of data, can be removed by introducing pointer variables, and this paper shows how this is done without allowing unrestricted sharing of variables. Pointer variables also give means to solve problems of dynamic resource allocation and dynamic addressing. Examples with performance figures for an actual implementation are given.

Synchronization under a commercial operating system

Mutual exclusion and general synchronization of concurrent processes (‘tasks’) are well‐known principles for constructing reliable real‐time systems. This paper shows how to implement these principles under a typical commercial operating system which provides incomplete synchronization operations. The problem of synchronizing erroneous tasks is discussed briefly.