Annual Review in Automatic Programming最新文献

This paper describes a real-time architecture based on a deadline driven dataflow machine and the hardware implementation of the architecture in the form of resource adequate Simple-processors (S-processors). The machine-level language and a high-level visual language are described, which enable implementation-detail-free programs to be written. The implementation of a real-time system on a resource adequate S-processor architecture requires no effort to be spent on schedulability analysis for temporal correctness.

Adaptive and dynamic behaviour is seen as one of the key characteristics of next generation hard real-time systems. Whilst fixed priority pre-emptive scheduling is becoming a de facto standard in real-time system implementation, it remains inflexible in its purest form. One approach to countering this criticism is to allow optional components, not guaranteed offline, to be executed at run-time. Optional components may be guaranteed a minimum set of resources at run-time, with competing resource requests scheduled according to the best-effort approach. This allows increased dynamic behaviour and improves the utility of the system. In this paper, we discuss this integration of fixed priority and best-effort scheduling.

A hierarchical approach to correctness verification of real-time software specifications is presented. The verification is distributed into successive steps that correspond to the design phases. The three languages: Rule Charts, LACTATRE (graphical specification) and Communicating Real Time State Machines are used for specification of real-time software within corresponding abstraction levels. The correctness is defined as a coincidence of a system specified in a phase w.r.t. requirements established ing the previous phase. This correctness concept leads to an application of the relative correctness methods (developed in former works) for the verification. The approach is examined in Preliminary and Detailed Design phases for the verification of several types of properties: structure, functions and time constraints.

In this paper, we present two different methods to make proofs on the GRAF-CET language. The first one is based on Transition Systems and the second one uses Polynomial Dynamical Systems. Theoretical and pratical aspects are presented.

Fuzzy identification means to find a set of fuzzy if-then rules with well defined attributes, that can describe the given $\text{I}\text{O}\text{-}\text{behaviour}$ of a system. In the identification algorithm proposed here the subject of learning are the rule conclusions i.e. the membership functions of output attributes in form of singletons. For fixed input membership functions learning is shown to be a least squares optimization problem linear in the unknown parameters. Examples show applications of the algorithm to the linguistic formulation of a PI control strategy and to identification of a nonlinear time-discrete dynamic system.

In this paper we discuss issues in real-time image processing, including applications, approaches and hardware. In particular, we discuss the failure of existing programming languages to support these considerations and present requirements for any language that can support real-time image processing.

With the advancements in multimedia and networking technologies, many distributed multimedia applications have been developed recently. These new applications rely on the underlying network support to transmit the multimedia real-time data between sites. Multimedia Transport Protocol (MMTP) is an experimental protocol designed for transmission of various multimedia data. MMTP uses multiple priority queues to support different levels of service requirements, and it discards packets from the transmission queue to reduce the network loading, and to ensure packets transmitted will meet the real-time constraint required by the data streams.

This paper presents the actual work in real-time planning as search [1] [2]. Based in this work we tried to solve the path planning in numerical state space. We found that precision, performance, and time were very linked. In real-time problem solving, the agent can fall in traps made of forbidden zones and to go out it, have to spend too much computing time. To solve this problem we propose a multilayer inference based in subgoals computation. An architecture based in two agents, one for low level task with the maximum precision and other for subgoals computation is proposed here.

This paper describes a learning algorithm to be applied in process control techniques, covering several topics of control applications such as system identification, or process supervision in a simple and useful way which make the method reliable to be applied on industrial process control. Knowledge acquired by means of a proposed learning algorithm is stored into a DAM or FAM (deterministic or fuzzy associative memory) for finally be applied on model parameter mapping. With such mappings, it is possible to know if the process is being disturbed and the intensity of such excitation inputs when no variation exist in system parameters or it is also possible to know the variation of system parameters when all input forces (control force and disturbances) are known.

Embedded systems tend to be event driven. The BED kernel supports an event-action programming model for an embedded systems development environment called Testbed. The kernel features deterministic scheduling and message transfer services.