Neuroprosthetics is at the intersection of neuroscience, biomedical engineering, and physics. A biocompatible neuroprosthesis contains artificial neurons exhibiting biophysically plausible dynamics. Hybrid systems analysis could be used to prototype such artificial neurons. Biohybrid systems are composed of artificial and living neurons coupled via real-time computing and dynamic clamp. Model neurons must be thoroughly tested before coupled with a living cell. We use bifurcation theory to identify hazardous regimes of activity that may compromise biocompatibility and to identify control strategies for regimes of activity desirable for functional behavior. We construct real-time artificial neurons for the analysis of hybrid systems and demonstrate a mechanism through which an artificial neuron could maintain duty cycle independent of variations in period.

In this dissertation, we implement efficient lock-based synchronization by a novel, high performance, simple and scalable hardware technique and associated software for a target shared-memory multiprocessor System-on-a-Chip (SoC). The custom hardware part of our solution is provided in the form of an intellectual property (IP) hardware unit which we call the SoC Lock Cache (SoCLC). SoCLC provides effective lock hand-off by reducing on-chip memory traffic and improving performance in terms of lock latency, lock delay and bandwidth consumption. The proposed solution is independent from the memory hierarchy, cache protocol and the processor architectures used in the SoC, which enables easily applicable implementations of the SoCLC (e.g., as a reconfigurable or partially/fully custom logic), and which distinguishes SoCLC from previous approaches. Furthermore, the SoCLC mechanism has been extended to support priority inheritance with an immediate priority ceiling protocol (IPCP) implemented in hardware, which enhances the hard real-time performance of the system. Our experimental results in a four-processor SoC indicate that SoCLC can achieve up to 37% overall speedup over spin-lock and up to 48% overall speedup over MCS for a microbenchmark with false sharing. The priority inheritance implemented as part of the SoCLC hardware, on the other hand, achieves 1.43X speedup in overall execution time of a robot application when compared to the priority inheritance implementation under the Atalanta real-time operating system. Furthermore, it has been shown that with the IPCP mechanism integrated into the SoCLC, all of the tasks of the robot application could meet their deadlines (e.g., a high priority task with 250us worst case response time could complete its execution in 93us with SoCLC, however the same task missed its deadline by completing its execution in 283us without SoCLC). Therefore, with IPCP support, our solution can provide better real-time guarantees for real-time systems. To automate SoCLC design, we have also developed an SoCLC-generator tool, PARLAK, that generates user specified configurations of a custom SoCLC. We used PARLAK to generate SoCLCs from a version for two processors with 32 lock variables occupying 2,520 gates up to a version for fourteen processors with 256 lock variables occupying 78,240 gates.

One of the key challenges in service-oriented systems engineering is the prediction and assurance of non-functional properties, such as the reliability and the availability of composite interorganizational services. Such systems are often characterized by a variety of inherent uncertainties, which must be addressed in the modeling and the analysis approach. The different relevant types of uncertainties can be categorized into (1) epistemic uncertainties due to incomplete knowledge and (2) randomization as explicitly used in protocols or as a result of physical processes. In this report, we study a probabilistic timed model which allows us to quantitatively reason about nonfunctional properties for a restricted class of service-oriented real-time systems using formal methods. To properly motivate the choice for the used approach, we devise a requirements catalogue for the modeling and the analysis of probabilistic real-time systems with uncertainties and provide evidence that the uncertainties of type (1) and (2) in the targeted systems have a major impact on the used models and require distinguished analysis approaches. The formal model we use in this report are Interval Probabilistic Timed Automata (IPTA). Based on the outlined requirements, we give evidence that this model provides both enough expressiveness for a realistic and modular specifiation of the targeted class of systems, and suitable formal methods for analyzing properties, such as safety and reliability properties in a quantitative manner. As technical means for the quantitative analysis, we build on probabilistic model checking, specifically on probabilistic time-bounded reachability analysis and computation of expected reachability rewards and costs. To carry out the quantitative analysis using probabilistic model checking, we developed an extension of the Prism tool for modeling and analyzing IPTA. Our extension of Prism introduces a means for modeling probabilistic uncertainty in the form of probability intervals, as required for IPTA. For analyzing IPTA, our Prism extension moreover adds support for probabilistic reachability checking and computation of expected rewards and costs. We discuss the performance of our extended version of Prism and compare the interval-based IPTA approach to models with fixed probabilities. / Eine der wichtigsten Herausforderungen in der Entwicklung von Service-orientierten Systemen ist die Vorhersage und die Zusicherung von nicht-funktionalen Eigenschaften, wie Ausfallsicherheit und Verfügbarkeit von zusammengesetzten, interorganisationellen Diensten. Diese Systeme sind oft charakterisiert durch eine Vielzahl von inhärenten Unsicherheiten, welche sowohl in der Modellierung als auch in der Analyse eine Rolle spielen. Die verschiedenen relevanten Arten von Unsicherheiten können eingeteilt werden in (1) epistemische Unsicherheiten aufgrund von unvollständigem Wissen und (2) Zufall als Mittel in Protokollen oder als Resultat von physikalischen Prozessen. In diesem Bericht wird ein probabilistisches, Zeit-behaftetes Modell untersucht, welches es ermöglicht quantitative Aussagen über nicht-funktionale Eigenschaften von einer eingeschränkten Klasse von Service-orientierten Echtzeitsystemen mittels formaler Methoden zu treffen. Zur Motivation und Einordnung wird ein Anforderungskatalog für probabilistische Echtzeitsysteme mit Unsicherheiten erstellt und gezeigt, dass die Unsicherheiten vom Typ (1) und (2) in den untersuchten Systemen einen Ein uss auf die Wahl der Modellierungs- und der Analysemethode haben. Als formales Modell werden Interval Probabilistic Timed Automata (IPTA) benutzt. Basierend auf den erarbeiteten Anforderungen wird gezeigt, dass dieses Modell sowohl ausreichende Ausdrucksstärke für eine realistische und modulare Spezifikation als auch geeignete formale Methoden zur Bestimmung von quantitativen Sicherheits- und Zuverlässlichkeitseigenschaften bietet. Als technisches Mittel für die quantitative Analyse wird probabilistisches Model Checking, speziell probabilistische Zeit-beschränkte Erreichbarkeitsanalyse und Bestimmung von Erwartungswerten für Kosten und Vergütungen eingesetzt. Um die quantitative Analyse mittels probabilistischem Model Checking durchzuführen, wird eine Erweiterung des Prism-Werkzeugs zur Modellierung und Analyse von IPTA eingeführt. Die präsentierte Erweiterung von Prism ermöglicht die Modellierung von probabilistischen Unsicherheiten mittelsWahrscheinlichkeitsintervallen, wie sie für IPTA benötigt werden. Zur Verifikation wird probabilistische Erreichbarkeitsanalyse und die Berechnung von Erwartungswerten durch das Werkzeug unterstützt. Es wird die Performanz der Prism-Erweiterung untersucht und der Intervall-basierte IPTA-Ansatz mit Modellen mit festen Wahrscheinlichkeitswerten verglichen.

<p>When exploring a verification approach for Komatsu Forest's control system regarding their forest machines (Valmet), the context of soft real-time systems is illuminated. Because of the nature of such context, the verification process is based on empirical corroboration of requirements fulfillment rather than being a formal proving process.</p><p>After analysis of the literature with respect to the software testing field, two paradigms have been defined in order to highlight important concepts for soft real-time systems. The paradigms are based on an abstract stimuli/response model, which conceptualize a system with inputs and output. Since the system is perceived as a black box, its internal details are hidden and thus focus is placed on a more abstract level.</p><p>The first paradigm, the “input data paradigm”, is concerned about what data to input to the system. The second paradigm, the “input data mechanism paradigm” is concerned about how the data is sent, i.e. the actual input mechanism is focused. By specifying different dimensions associated with each paradigm, it is possible to define their unique characteristics. The advantage of this kind of theoretical construction is that each paradigm creates an unique sub-field with its own problems and techniques.</p><p>The problems defined for this thesis is primarily focused on the input data mechanism paradigm, where devised dimensions are applied. New verification techniques are deduced and analyzed based on general software testing principles. Based on the constructed theory, a test system architecture for the control system is developed. Finally, an implementation is constructed based on the architecture and a practical scenario. Its automation capability is then assessed.</p><p>The practical context for the thesis is a new simulator under development. It is based upon LabVIEW and PXI technology and handles over 200 I/O. Real machine components are connected to the environment, together with artificial components that simulate the engine, hydraulic systems and a forest. Additionally, physical control sticks and buttons are connected to the simulator to enable user testing of the machine being simulated.</p><p>The results associated with the thesis is first of all that usable verification techniques were deduced. Generally speaking, some of these techniques are scalable and are possible to apply for an entire system, while other techniques may be appropriate for selected subsets that needs extra attention. Secondly, an architecture for an automated test system based on a selection of techniques has been constructed for the control system.</p><p>Last but not least, as a result of this, an implementation of a general test system has been possible and successful. The implemented test system is based on both C# and LabVIEW. What remains regarding the implementation is primarily to extend the system to include the full scope of features described in the architecture and to enable result analysis.</p> / <p>Då verifikationstekniker för Komatu Forests styrsystem utreds angående Valmet skogsmaskiner, hamnar det mjuka realtidssystemkontextet i fokus. Ett sådant kontext antyder en process där empirisk styrkning av kravuppfyllande står i centrum framför formella bevisföringsprocesser.</p><p>Efter en genomgång och analys av litteratur för mjukvarutestområdet har två paradigmer definierats med avsikten att belysa viktiga concept för mjuka realtidssystem. Paradigmerna är baserade på en abstrakt stimuli/responsmodell, som beskriver ett system med in- och utdata. Eftersom detta system betraktas som en svart låda är inre detaljer gömda, vilket medför att fokus hamnar på ett mer abstrakt plan.</p><p>Det första paradigmet benämns som “indata-paradigmet” och inriktar sig på vilket data som skickas in i systemet. Det andra paradigmet går under namnet “indatamekanism-paradigmet” och behandlar hur datat skickas in i systemet, dvs fokus placeras på själva inskickarmekanismen. Genom att definiera olika</p><p>dimensioner för de två paradigmen, är det möjligt att beskriva deras utmärkande drag. Fördelen med att använda denna teoretiska konstruktion är att ett paradigm skapar ett eget teoriområde med sina egna frågeställningar och tekniker.</p><p>De problem som definierats för detta arbete är främst fokuserade på indatamekanism-paradigmet, där framtagna dimensioner tillämpas. Nya verifikationstekniker deduceras och analyseras baserat på generella mjukvarutestprinciper. Utifrån den skapade teorin skapas en testsystemarkitektur för kontrollsystemet. Sedan utvecklas ett testsystem baserat på arkitekturen samt ett praktiskt scenario med syftet att utreda systemets automationsgrad.</p><p>Den praktiska miljön för detta arbete kretsar kring en ny simulator under utveckling. Den är baserad på LabVIEW och PXI-teknik och hanterar över 200 I/O. Verkliga maskinkomponenter ansluts till denna miljö tillsammans med konstgjorda komponenter som simulerar motorn, hydralik samt en skog. Utöver detta, ansluts styrspakar och knappar för att möjliggöra användarstyrning av maskinen som simuleras.</p><p>Resultatet förknippat med detta arbete är för det första användbara verifikationstekniker. Man kan generellt säga att några av dessa tekniker är skalbara och därmed möjliga att tillämpa för ett helt system. Andra tekniker är ej skalbara, men lämpliga att applicera på en systemdelmängd som behöver testas mer utförligt.</p><p>För det andra, en arkitektur har konstruerats för kontrollsystemet baserat på ett urval av tekniker. Sist men inte minst, som en följd av ovanstående har en lyckad implementation av ett generellt testsystem utförts. Detta system implementerades med hjälp av C# och LabVIEW. Det som återstår beträffande implementationen är att utöka systemet så att alla funktioner som arkitekturen beskriver är inkluderade samt att införa resultatanalys.</p>

This work is devoted to the issue of monitoring of distributed real-time systems. In particular, it focuses on formal aspects of model-based supervision and problems which are related to it. In its first part, we present the basic properties of two well-known formal models used to model distributed systems: networks of timed automata and time Petri nets. We show that the behavior of these models can be represented with so-called branching processes. We also introduce the key conceptual elements of the supervisory system. The second part of the work is dedicated to the issue of constrained unfoldings which enable us to track causal relationships between events in a distributed system. This type of structure can be used to reproduce processes of the system on the basis of a completely unordered set of previously observed events. Moreover, we show that time constraints imposed on a system and observations submitted to the supervisory system can significantly affect a course of events in the system. We also raise the issue of parameters in time constraints. The proposed methods are illustrated with case studies. The third part of the work deals with the issue of unobservable cyclical behaviors in distributed systems. This type of behaviors leads to an infinite number of events in constrained unfoldings. We explain how we can obtain a finite structure that stores information about all observed events in the system, even if this involves processes that are infinite due to such unobservable loops. The fourth and final part of the work is dedicated to implementation issues of the previously described methods.

The employment of real time NDT systems has been spread widely last years. It is very important to control in real-time layer thicknesses of multi-layered materials during manufacturing process. In this work multi-layer thickness measurement digital signal processing methods are evaluated. Presented real-time nondestructive testing system and signal propagation model is a background for measurement algorithm analysis and development. In this work multi-layer thickness measurement problem is addressed with a different approach, by applying correlation functions on parts of the ultrasonic signal, thus reducing the overall computational complexity.

La recherche d'une borne supérieure au temps d'exécution d'un programme est une partie essentielle du processus de vérification de systèmes temps-réel critiques. Les programmes de tels systèmes ont généralement des temps d'exécution variables et il est difficile, voire impossible, de prédire l'ensemble de ces temps possibles. Au lieu de cela, il est préférable de rechercher une approximation du temps d'exécution pire-cas ou Worst-Case Execution Time (WCET). Une propriété cruciale de cette approximation est qu'elle doit être sûre, c'est-à-dire qu'elle doit être garantie de majorer le WCET. Parce que nous cherchons à prouver que le système en question se termine en un temps raisonnable, une surapproximation est le seul type d'approximation acceptable. La garantie de cette propriété de sûreté ne saurait raisonnablement se faire sans analyse statique, un résultat se basant sur une série de tests ne pouvant être sûr sans un traitement exhaustif des cas d'exécution. De plus, en l'absence de certification du processus de compilation (et de transfert des propriétés vers le binaire), l'extraction de propriétés doit se faire directement sur le code binaire pour garantir leur fiabilité. Toutefois, cette approximation a un coût : un pessimisme - écart entre le WCET estimé et le WCET réel - important entraîne des surcoûts superflus de matériel pour que le système respecte les contraintes temporelles qui lui sont imposées. Il s'agit donc ensuite, tout en maintenant la garantie de sécurité de l'estimation du WCET, d'améliorer sa précision en réduisant cet écart de telle sorte qu'il soit suffisamment faible pour ne pas entraîner des coûts supplémentaires démesurés. Un des principaux facteurs de surestimation est la prise en compte de chemins d'exécution sémantiquement impossibles, dits infaisables, dans le calcul du WCET. Ceci est dû à l'analyse par énumération implicite des chemins ou Implicit Path Enumeration Technique (IPET) qui raisonne sur un surensemble des chemins d'exécution. Lorsque le chemin d'exécution pire-cas ou Worst-Case Execution Path (WCEP), correspondant au WCET estimé, porte sur un chemin infaisable, la précision de cette estimation est négativement affectée. Afin de parer à cette perte de précision, cette thèse propose une technique de détection de chemins infaisables, permettant l'amélioration de la précision des analyses statiques (dont celles pour le WCET) en les informant de l'infaisabilité de certains chemins du programme. Cette information est passée sous la forme de propriétés de flot de données formatées dans un langage d'annotation portable, FFX, permettant la communication des résultats de notre analyse de chemins infaisables vers d'autres analyses. Les méthodes présentées dans cette thèse sont inclues dans le framework OTAWA, développé au sein de l'équipe TRACES à l'IRIT. Elles usent elles-mêmes d'approximations pour représenter les états possibles de la machine en différents points du programme. / The search for an upper bound of the execution time of a program is an essential part of the verification of real-time critical systems. The execution times of the programs of such systems generally vary a lot, and it is difficult, or impossible, to predict the range of the possible times. Instead, it is better to look for an approximation of the Worst-Case Execution Time (WCET). A crucial requirement of this estimate is that it must be safe, that is, it must be guaranteed above the real WCET. Because we are looking to prove that the system in question terminates reasonably quickly, an overapproximation is the only acceptable form of approximation. The guarantee of such a safety property could not sensibly be done without static analysis, as a result based on a battery of tests could not be safe without an exhaustive handling of test cases. Furthermore, in the absence of a certified compiler (and tech- nique for the safe transfer of properties to the binaries), the extraction of properties must be done directly on binary code to warrant their soundness. However, this approximation comes with a cost : an important pessimism, the gap between the estimated WCET and the real WCET, would lead to superfluous extra costs in hardware in order for the system to respect the imposed timing requirements. It is therefore important to improve the precision of the WCET by reducing this gap, while maintaining the safety property, as such that it is low enough to not lead to immoderate costs. A major cause of overestimation is the inclusion of semantically impossible paths, said infeasible paths, in the WCET computation. This is due to the use of the Implicit Path Enumeration Technique (IPET), which works on an superset of the possible execution paths. When the Worst-Case Execution Path (WCEP), corresponding to the estimated WCET, is infeasible, the precision of that estimation is negatively affected. In order to deal with this loss of precision, this thesis proposes an infeasible paths detection technique, enabling the improvement of the precision of static analyses (namely for WCET estimation) by notifying them of the infeasibility of some paths of the program. This information is then passed as data flow properties, formatted in the FFX portable annotation language, and allowing the communication of the results of our infeasible path analysis to other analyses.

Dans le domaine des systèmes critiques, l'analyse des temps d'exécution des programmes est nécessaire pour planifier et ordonnancer au mieux différentes tâches et par extension pour dimensionner les systèmes. La durée d'exécution d'un programme dépend de divers facteurs comme ses entrées ou le matériel utilisé. Or cette variation temporelle pose problème dans les systèmes temps-réel dans lesquels il est nécessaire de dimensionner précisément les temps processeur alloués à chaque tâche, et pour cela, connaître leur temps d'exécution au pire cas. Au sein de l'équipe TRACES à l'IRIT, nous cherchons à calculer une borne supérieure à ce temps d'exécution au pire cas qui soit la plus précise possible. Pour cela, nous travaillons sur le graphe de flot de contrôle d'un programme qui représente un sur-ensemble des ses exécutions possibles, que nous accompagnons d'annotations sur des comportements spécifiques du programme susceptibles de réduire la sur-approximation de notre estimation. Dans les outils destinés au calcul du temps d'exécution au pire cas des programmes, les annotations sont habituellement exprimées et intégrées grâce à des langages d'annotation spécifiques. Nous proposons d'utiliser des automates appelés automates d'annotation de flot en lieu et place de ces langages, afin de fonder non seulement l'expression, mais également l'intégration d'annotations dans l'analyse sur des bases formelles. Nous présentons ces automates enrichis de contraintes, de variables et d'une hiérarchie et nous montrons comment ils supportent les divers types d'annotations utilisés dans le domaine de l'analyse du temps d'exécution au pire cas. Par ailleurs, l'intégration des annotations dans une analyse se fait habituellement par l'association de contraintes numériques au graphe de flot de contrôle. Les automates que nous présentons supportent cette méthode mais leur expressivité offre également de nouvelles possibilités d'intégration basées sur le dépliage du graphe de flot de contrôle. Nous présentons des résultats expérimentaux issus de la comparaison de ces deux méthodes qui montrent comment le dépliage de graphe peut améliorer la précision de l'analyse. A terme, ce gain de précision dans l'estimation du temps d'exécution au pire cas permettra de mieux exploiter le matériel sans faire courir de risques à l'utilisateur ou au système. / In the domain of critical systems, the analysis of execution times of programs is needed to schedule various task at best and by extension to dimension the whole system. The execution time of a program depends on multiple factors such as entries of the program or the targeted hardware. Yet this time variation is an issue in real-time systems where the duration is required to allow correct processor time to each task, and in this purpose, we need to know their worst-case execution time. In the TRACES team at IRIT, we try to compute a safe upper bound of this worst-case execution time that would be as precise as possible. In order to do so, we work on the control flow graph of a program that represents an over-set of its possible executions and we combine this structure with annotations on specific behaviours of the program that might reduce the over-approximation of our estimation. Tools designed to compute worst-case execution times of programmes usually support the expression and the integration of annotations thanks to specific annotation languages. Our proposal is to replace these languages with a type of automata named flow fact automata so that not only the expression but also the integration of annotations in the analysis inherit from the formal basis of automata. Based on these automata enriched with constraints, variables and a hierarchy, we show how they support the various annotation types used in the worst-case execution time domain. Additionally, the integration of annotations in an analysis usually lead to associate numerical constraint to the control flow graph. The automata presented here support this method but their expressiveness offers new integration possibilities based on the partial unfolding of control flow graph. We present experimental results from the comparison of these two methods that show how the graph unfolding can improve the analysis precision. In the end, this precision gain in the worst-case execution time will ensure a better usage of the hardware as well as the absence of risks for the user or the system itself.

Les systèmes temps réel (STR) sont au coeur de machines souvent jugés critiques pour lasécurité : ils en contrôlent l’exécution afin que celles-ci se comportent de manière sûre dans le contexte d’un environnement dont l’évolution peut être imprévisible. Un STR n’a d’autre alternative que de s’adapter à son environnement : sa correction dépend des temps de réponses aux stimuli de ce dernier.Il est couramment admis que le formalisme des réseaux de Petri temporels (RdPT) est adapté àla description des STR. Cependant, la modélisation de systèmes simples, ne possédant que quelquestˆaches périodiques ordonnancées de façon basique se révèle être un exercice souvent complexe.En premier lieu, la modélisation efficace d’une gamme étendue de politiques d’ordonnancementsse heurte à l’incapacité des RdPT à imposer un ordre d’apparition à des évènements concurrentssurvenant au même instant. D’autre part, les STR ont une nette tendance à être constitués de caractéristiques récurrentes, autorisant une modélisation par composants. Or les RdPT ne sont guèreadaptés à une utilisation compositionnelle un tant soit peu générale. Afin de résoudre ces deuxproblèmes, nous proposons dans cette thèse Cifre – en partenariat entre Airbus et le Laas-Cnrs– d’étendre les RdPT à l’aide de deux nouvelles relations, les relations d’inhibition et de permission,permettant de spécifier de manière plus fine les contraintes de temps.Afin de cerner un périmètre clair d’adéquation de cette nouvelle extension à la modélisation dessystèmes temps réel, nous avons défini Pola, un langage spécifique poursuivant deux objectifs :déterminer un sous-ensemble des systèmes temps réel modélisables par les réseaux de Petri temporelsà inhibitions/permissions et fournir un langage simple à la communauté temps réel dont lavérification, idéalement automatique, est assurée par construction. Sa sémantique est donnée par traduction en réseaux de Petri temporels à inhibitions/permissions. L’explorateur d’espace d’états de laboite à outils Tina a été étendu afin de permettre la vérification des descriptions Pola / Real time systems (RTS) are at the core of safety critical devices : they control thedevices’ behavior in such a way that they remain safe with regard to an unpredictable environment. ARTS has no other choices than to adapt to its environment : its correctness depends upon its responsetime to the stimuli stemming from the environment.It is widely accepted that the Time Petri nets (TPN) formalism is adapted to the description ofRTS. However, the modeling of simple systems with only a few periodic tasks scheduled according toa basic policy remains a challenge in the worst case and can be very tedious in the most favorable one.First, we put forward some limitations of TPN regarding the modeling of a wide variety of schedulingpolicies, coming from the fact that this formalism is not always capable to impose a givenorder on events whenever they happen at the same time. Moreover, RTS are usually constituted of thesame recurring features, implying a compositional modeling, but TPN are not well adapted to sucha compositional use. To solve those problems we propose in this Cifre thesis – in partnership withAirbus and the Laas-Cnrs – to extend the formalism with two new dual relations, the forbid andallow relations so that time constraints can be finely tuned.Then, to assess this new extension for modeling of real time systems, we define Pola, a specificlanguage aimed at two goals : to determine a subset of RTS which can be modeled with forbid/allowtime Petri nets and to provide a simple language to the real time community which, ideally, can bechecked automatically. Its semantics is given by translation into forbid/allow Time Petri nets. Thestate space exploration tool of the Tina toolbox have been extended so that it can model check Poladescriptions.

Made available in DSpace on 2014-12-17T14:55:41Z (GMT). No. of bitstreams: 1 AlvaroMA_DISSERT.pdf: 1811875 bytes, checksum: d1a8b9710060f420383a8d715381bfb9 (MD5) Previous issue date: 2009-12-23 / Embedded systems are widely spread nowadays. An example is the Digital Signal Processor (DSP), which is a high processing power device. This work s contribution consist of exposing DSP implementation of the system logic for detecting leaks in real time. Among the various methods of leak detection available today this work uses a technique based on the pipe pressure analysis and usesWavelet Transform and Neural Networks. In this context, the DSP, in addition to do the pressure signal digital processing, also communicates to a Global Positioning System (GPS), which helps in situating the leak, and to a SCADA, sharing information. To ensure robustness and reliability in communication between DSP and SCADA the Modbus protocol is used. As it is a real time application, special attention is given to the response time of each of the tasks performed by the DSP. Tests and leak simulations were performed using the structure of Laboratory of Evaluation of Measurement in Oil (LAMP), at Federal University of Rio Grande do Norte (UFRN) / Os sistemas embarcados est?o amplamente difundidos atualmente. Um exemplo ? o Digital Signal Processor (DSP), que ? um dispositivo com alto poder de processamento. A contribui??o deste trabalho consiste na implementa??o em DSP da l?gica de um sistema de detec??o de vazamentos em tempo real. Dentre os v?rios m?todos de detec??o de vazamentos existentes atualmente este trabalho se desenvolve utilizando uma t?cnica baseada na an?lise da press?o no duto e que utiliza Transformada Wavelet e Redes Neurais. Nesse contexto o DSP, al?m de realizar o processamento digital do sinal de press?o, tamb?m comunica-se com um Global Positioning System (GPS), que auxilia na localiza??o do vazamento e com um sistema supervis?rio, disponibilizando informa??es para este. Para garantir robustez e confiabilidade na comunica??o entre DSP e sistema supervis?rio ? utilizado o protocolo Modbus. Como trata-se de uma aplica??o de tempo real, uma aten??o especial ? dada ao tempo de resposta de cada uma das tarefas realizadas pelo DSP. Os testes e simula??es de vazamentos foram realizados utilizando a estrutura do Laborat?rio de Avalia??o de Medi??o em Petr?leo (LAMP), da Universidade Federal do Rio Grande do Norte (UFRN)