Building models of real-time systems from application software

An approach to the modeling and verification of real-time systems is presented. The Specification PEARL description language is used for hardware/software modeling. It has its origins in the standard Multiprocessor PEARL language. Its usefulness has been enhanced for asymmetrical multiprocessor system modeling, and parameters for schedulability analysis have been added. It is meant to be an over-layer for PEARL programs and supports the corresponding program model. The detailed program models are represented internally by timed state transition automata. The model of the co-designed system is verified for feasibility with co-simulation. The resulting information may be used for changes in the current design to gain a feasible model of the system. After that, the program models may be enhanced to their full functionality. Schedulability analysis, which should provide the designer with more precise information, can then be done. The new information can be reused on the system model in order to discover bottlenecks or possible discrepancies with the estimated times. By utilising this methodology, the possibility of implementing a non-feasible system should be minimised.

Task automata: Schedulability, decidability and undecidability

A wide application of real time systems is safety critical systems. These systems should be highly reliable. Fault tolerance is one of the approaches to achieve reliability. In this paper, a fault tolerance model for real time systems is proposed. This model incorporates the concept of time stamped fault tolerance. This model is based on distributed computing along with feed forward artificial neural network methodology. The proposed technique is based on execution of design diverse variants on replicated hardware, and assigning weights to the results produced by variants. Thus the proposed method encompasses both the forward and backward recovery mechanism, but main focus is on forward recovery. The main essence of the proposed technique is the inclusion of time for decision mechanism

Schedulability analysis of software-intensive systems requires solution techniques that go beyond worst-case assumptions, fostering a cross-fertilization between the areas of real-time systems and performance engineering. We address probabilistic schedulability analysis of tasks in single-processor non-preemptive real-time systems. To this end, we consider periodic tasks with offsets, scheduled by Fixed Priority (FP) or Earliest Deadline First (EDF), and managed by the discarding policy or the rejection policy to control deadline misses. Each task has a Worst Case Execution Time (WCET) which can be a deterministic value or a random variable, notably characterized by a non-Markovian distribution possibly with bounded support. A continuous-time stochastic model of task-set is specified through Stochastic Time Petri Nets (STPNs) and solved by regenerative transient analysis based on stochastic state classes. The evaluation of performance measures on resource allocation and missed deadlines enables the analysis of design choices and the estimation of over-provisioned resources that are likely to be unused at run-time. Feasibility and effectiveness of the approach are validated on randomly generated task-sets for different processor utilizations.

Piezo actuators are widely use in positioning and control applications. Piezo actuators are nonlinear and possess time varying hysteresis nature. For controlling piezo actuators mostly inverse hysteresis model or adaptive inverse control model are utilize. To implement adaptive inverse control, adaptive model of piezo actuator is to be determined. There are various adaptive methods viz. Recursive Least Square (RLS), Kalman Filter (KF), Least Mean Square (LMS) and Normalized Least Mean Square (NLMS) by which the model can be determined. This paper discusses the system of piezo actuator and different methods of finding adaptive model for piezo actuator system. Further a comparative study of adaptive model identification is carried out and suitable model for this system is identified and discussed. A novel approach of identifying a suitable model is discussed in this paper. The experimentation is carried out using electro pneumatic positioner, which utilizes the piezo valve for positioning. LabVIEW 2014 is used along with data acquisition hardware for modeling of real time system. The results of experimentation shows that, RLS adaptive model is more suitable as compared to other adaptive algorithms.

An Approach to Dependability Modelling of Real Time Systems

Real-time software is becoming more prevalent in avionics systems such as unmanned aerial vehicles and spacecrafts. More often, these systems operate with limited human interaction or an increased latency in human response. Therefore, these systems require autonomous fault detection and a repair mechanism to respond to off nominal conditions that could arise. In order to prevent a hazard, these systems require assurance that they will operate in a safe manner while successfully completing a mission. Assurance, as a software engineering activity involves the reduction of risk, and inversely, an increase in confidence in a software system. In the context of fault management, this becomes a question of increasing confidence that a software system can detect and respond to the offnominal conditions occur. This activity is not always straight forward as off-nominal conditions in complex real-time systems do not always manifest as singular events, but rather as a combination of symptoms. Furthermore, for fault management assurance to be useful it needs to be rigorous. Therefore, the approach needs to be supported by a standard, repeatable framework. The foundation for this framework is in the ability to model the fault management system integrated in the hardware and software avionics real-time system. Models can assist in assuring both functional and quality attribute requirements such as reliability. The Architecture Analysis and Design Language (AADL) is a Society of Automotive Engineers (SAE) standard notation (AS5506/1) for the modeling and analysis of real-time systems. AADL has been extended to model fault management behavior through the AADL Error Annex, also an SAE standard. At its core, the AADL Error Annex provides rules for how software and hardware components are allowed to propagate an error across the avionics real-time system modeled in AADL. Because a standard notation is being used, there is intrinsic confidence to the approach. This paper will explore how the assurance of software fault management can be applied in practice. First, it is important to briefly review AADL and the AADL Error Annex as well as common terminology related to software fault management. Next, the application of software fault management assurance will be discussed. The first step in this application is to understand the very basic (atomic) hardware and software faults that occur at the component level. This information can be collected in the form of a failure modes and effects analysis (FMEA). An analysis is required in order to evaluate the ability of the software fault management system to, at the very least, detect the symptoms. This analysis answers two fundamental questions. First, to what degree are faults visible to the avionics software? Second, to what degree does the avionics software handle these faults? This paper presents an example to demonstrate the assurance of software fault management. The software fault management assurance framework proposed in this paper utilizes a standard notation, coupled with a formal mechanism for modeling faults. This provides a rigorous foundation for analysis on the avionics software real-time system in a repeatable manner. Therefore a model based assurance activity of a software fault management system can successfully be applied.

The paper deals with problems of high level modeling of real-time systems in system level languages. Main advantages of transaction level modeling in SystemC are emphasized. The universal kernel for functionality, communication and timing issues modeling is presented. The problem of multithreading is exemplified on sorting procedure implemented in multi-core microprocessor system. Advantages of system level approach to modeling and design of real-time systems are outlined.

An important development trend of modern real-time systems and embedded systems is Mixed criticality real-time Systems, Which is integrating functions with different level of importance into a sharing hardware platform. The biggest research focus of mixed criticality real-time Systems is scheduling problem. At present, the definition and connotation of Mixed criticality real-time system are generally accepted, but there is no unified task model. In this study, according to the research requirements of Mixed criticality real-time system, we first give a task model of mixed criticality real-time system, and define the schedulability of the system. Then it introduces two scheduling strategies based on time and priority, analyzes their advantages and disadvantages, and leads to the classic algorithm EDF in dynamic priority strategy. Finally, the paper focuses on the EDF-VD scheduling algorithm. Based on the task model of Mixed criticality real-time system, an example is given to analyze the EDF-VD algorithm in the Mixed criticality real-time system.KeywordsMixed criticality real-time systemsScheduling algorithm and analysisEDF-VD (Earliest Deadline First-Virtual Deadlines) algorithm

As redes de Petri temporais são uma das extensões das redes de Petri convencionais para modelagem e análise de sistemas a eventos discretos temporizados. A cada transição da rede é associado um intervalo de tempo, o qual delimita o instante de disparo da transição. A análise das redes de Petri temporais tem sido feita por métodos enumerativos que relacionam todos os estados alcançáveis da rede. Neste artigo, usando recursos da álgebra intervalar e utilizando o conceito de classe de estados, é apresentada uma abordagem para resolver problemas relacionados à caracterização de intervalos de tempo de disparo em redes de Petri temporais. A partir de uma expressão algébrica, desenvolvida para cálculo de intervalos de tempo entre duas classes de estados consecutivas, obtém-se uma equação intervalar que possibilita o cálculo de intervalos de tempo entre duas classes de estados quaisquer. Assim, é possível calcular o intervalo de tempo para ocorrência de uma seqüência de eventos. Utilizando um método de redução da rede, a equação intervalar assume a forma matricial e, com isso, amplia a classe das redes de Petri temporais à qual a equação intervalar é aplicável. Por fim, uma aplicação explorando as potencialidades da equação intervalar no cálculo de medidas características de um sistema a eventos discretos temporizados é apresentada.

A discussion is presented of PNSOFT, a user-friendly, menu-driven software package for Petri-net modeling and analysis of real-time systems. It aids in the design, management and analysis of Petri-net models. PNSOFT runs in an automated programming environment that provides both a graphical and a matrix interface for the creation and modification of Petri-net models in different environments. The proposed software package is written in the C programming language and is efficient with respect to both time and space. It can easily be upgraded to perform a variety of other functions. The data structures created by PNSOFT's editing facilities can be used by other application programs. Various functions and programs used for analytical operations can be directly accessed by the user, using a menu-driven interface, at different levels. >

To formally verify the real-time system using model checking technique, two necessary elements are required, the real-time system model and its expected real-time properties. In order to specify these real-time properties, an appropriate extended logic system, to cope with quantitative of time constraints is needed. We select one of the most popular real-time logic systems, called Metric Temporal Logic (MTL) to specify our mentioned real-time properties. However, the real-time properties defined in MTL are still not relevant to SPIN model checker. In this paper, we propose a novel scheme to translate the basic MTL formulas into Promela and LTL formulas as to be relevantly used in SPIN, in terms of Promela’s inline definitions and macro commands. The punctuality of the formulas is preserved during the translation. Practically, the basic timing constraints eventually and always operators of MTL are focused, and the translation of MTL formulas of the real-time safety property patterns is demonstrated.

There is a growing interest in adopting object technologies for the development of real-time control systems. Several commercial tools, currently available, provide object-oriented modeling and design support for real-time control systems. While these products provide many useful facilities, such as visualization tools and automatic code generation, they are all weak in addressing the central characteristic of real-time control systems design, i.e., providing support for a designer to reason about timeliness properties. We believe an approach that integrates the advancements in both object modeling and design methods and real-time scheduling theory is the key to successful use of object technology for real-time software. Surprisingly several past approaches to integrate the two either restrict the object models, or do not allow sophisticated schedulability analysis techniques. This study shows how schedulability analysis can be integrated with UML for Real-Time (UML-RT) to deal with timing properties in real time control systems. More specifically, we develop the schedulability and feasibility analysis modeling for the external messages that may suffer release jitter due to being dispatched by a tick driven scheduler in real-time control system and we also develop the scheduliablity modeling for sporadic activities, where messages arrive sporadically then execute periodically for some bounded time. This method can be used to cope with timing constraints in realistic and complex real-time control systems. Using this method, a designer can quickly evaluate the impact of various implementation decisions on schedulability. In conjunction with automatic code-generation, we believe that this will greatly streamline the design and development of real-time control systems software.

UML 2.0 is a multi-paradigm language which copes well with modeling of reactive and real-time systems by offering many visual and flexible notations for description of their various aspects including their time constraints. However, their development remains a complex task and error-prone even if object-oriented concepts are used. Consequently, one is interested in validation techniques which allow analyzing the behaviors of a design already in early phases. Nevertheless, the consistency and compatibility checking of designed artifacts for a reactive system is a difficult task; because of both the imprecise dynamic semantics of UML diagrams and the expressiveness gap between them. In this setting, this article proposes a methodology for semantics and time consistency checking between behavioral models and scenario-based specifications depicted, respectively, by state and sequence diagrams.

In the paper, we present a methodology developed in order to verify probabilistic timed properties related to dependability of real-time systems. The methodology is made of three essential steps. The first one is a UML profile called DAMRTS (dependability analysis models for real-time systems) designed using GME tool. The aim is to model a real-time system with qualitative and quantitative information related to its quality of service. In this profile, UML statecharts are used to represent the system behavior. An extension is introduced with probabilities, real-time requirements and nondeterministic choices. The second one proposes a translation from extended UML statecharts to probabilistic timed automata. In this step, synchronization by events is used in probabilistic timed automata to describe the concurrency in UML statecharts. The last one is about the probabilistic model checking. This requires specification of dependability properties with a suitable temporal logic.