Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / This thesis proposes an Intelligent Alarm Processor for fault diagnosis in electrical power systems. The objective is to develop a methodology for automatic fault analysis using reported alarms from Supervisory Control and Data Acquisition (SCADA) to allow the use of diagnosis systems in large power systems. The proposal can be used in real-time decision support systems to assist control center‟s operators during the decision-making after unscheduled contingencies with relevant information to power system restoration. This work expects to contribute to the development of advanced alarm management logics that allow modifying the chronological sequence of reported alarms, event mapping and the generation of operating patterns of protection systems according to topology network. Still, mathematical programming models have been formulated as a parsimonious set covering problem to fault section estimation and identification of protective devices with improper operation. Among these models, it stands out the model that deals with integrated analysis of reported alarms, events and diagnosis that better explain the alarms. The proposed approach has been tested in different portions of the Southern Brazilian power system. The results show that alarm processing allows the practical implementation of intelligent diagnosis methods in existing supervisory systems. The proposed diagnosis methods show better performance and accurate solutions than other methods presented in literature. / Esta tese propõe um Processador Inteligente de Alarmes para diagnóstico de faltas em sistemas elétricos de potência. O objetivo é desenvolver uma metodologia para a análise automática de faltas a partir dos alarmes reportados no sistema de supervisão e aquisição de dados (SCADA) que possibilite o uso de métodos de diagnóstico em sistemas de potência de grande porte. Essa proposta pode ser empregada em sistemas de apoio à decisão em tempo real, que auxiliem operadores de centros de controle do sistema (COS) na tomada de decisão após desligamentos não programados, com informações pertinentes para o restabelecimento do sistema. O trabalho espera contribuir com o desenvolvimento de lógicas avançadas de gerenciamento de alarmes que possibilitem a reordenação cronológica dos alarmes reportados, o mapeamento dos eventos e a geração de padrões de funcionamento de sistemas de proteção de acordo à topologia da rede. Além disso, os modelos de programação matemática foram formulados como um problema de recobrimento de conjuntos parcimonioso, para estimação da seção em falta e identificação dos dispositivos de proteção com atuação indevida. Dentre esses modelos, destaca-se o modelo que analisa, de forma integrada, os alarmes reportados e determina os eventos e diagnósticos que melhor explicam os alarmes. A abordagem proposta foi testada em diferentes porções do sistema sul do sistema interligado nacional (SIN). Os resultados mostram que as rotinas desenvolvidas para o processamento de alarmes permite a implantação prática de métodos inteligentes de diagnóstico em sistemas supervisórios existentes. Os métodos propostos para diagnóstico de faltas mostraram desempenhos e precisão nos resultados superiores a outros métodos presentes na literatura.

This work proposes a hybrid system for alarm processing and fault diagnosis in electrical networks which use two methods of computational intelligence: Generalized Regression Neural Network and Genetic Algorithms. The neural network has the function of processing the set of received alarms and present as a response the characteristic(s) event(s), using for this, an elaborated knowledge based on the functional diagrams for protection and interviews with operators. Six modules were implemented for different neural components of a test system, according to their protection schemes. The output of these modules is used as input to the GA which has to do a combined analysis along with its database and provide the operator with the main protective components involved in the incident, as well as the probable causes of defects and actions to be taken in order to return the system in the shortest possible time and greater safety. For average inserted random errors of 0%, 7,73%, 15,46% and 23,19% in the received alarms, the system was able to diagnoses correctly in 100%, 93,60%, 74,26% and 48,07% of the cases respectively. It was found that the genetic algorithm improved the results obtained by neural network with good capability of generalization and condition to present multiple solutions, and the response time of the hybrid system was acceptable to the under consideration problem. / O presente trabalho propõe um sistema híbrido para processamento de alarmes e diagnóstico de faltas em redes elétricas com a utilização de dois métodos de inteligência computacional: Generalized Regression Neural Network e Algoritmos Genéticos. A rede neural tem a função de processar o conjunto de alarmes reportados e apresentar como resposta o evento(s) característico(s), utilizando-se, para isso, de um conhecimento elaborado com base nos diagramas funcionais da proteção e entrevista com operadores. Foram implementados seis módulos neurais para diferentes componentes de um sistema teste, de acordo com os seus respectivos esquemas de proteção. A saída destes módulos é utilizada como entrada para o AG que deve fazer uma análise combinatória juntamente com sua base de dados e apresentar ao operador os principais componentes de proteção envolvidos na incidência, bem como as prováveis causas do defeito e ações a serem tomadas de forma a restabelecer o sistema no menor tempo possível e com maior segurança. Para erros aleatórios médios inseridos de 0%, 7,73%, 15,46% e 23,19% nos alarmes reportados, o sistema se mostrou capaz de diagnosticar corretamente em respectivamente 100%, 93,60%, 74,26% e 48,07% dos casos. Verificou-se que o algoritmo genético melhorou os resultados obtidos pela rede neural, apresentando boa capacidade de generalização e condições de apresentação de múltiplas soluções, sendo o tempo de resposta do sistema híbrido aceitável para o problema tratado.

Denna studie ses som ett försök att bryta ny mark inom kriskommunikationsforskningen; vilken kan anses fokuserad på stora samhälleliga katastrofer, potentialen hos avancerade tekniska larm- och varningssystem samt massmediernas roll som meningsskapare. Mot bakgrund av detta, samt det faktum att vi lever i ett allt mer mobilt kommunikationsklimat, undersöker denna uppsats människors inställning till samhälleliga larmfunktioner; idag och imorgon. Empirin har genererats genom en rad fokuserade gruppintervjuer. Det teoretiska ramverket utgörs av Beck och Baumans tankar om ett senmodernt samhälle karaktäriserat av ökad institutionaliserad individualism. Castells och Frändeberg & Thulin får illustrera tidigare forskning och teorier kring människors användning av och inställning till nya kommunikationstekniker. Resultaten belyser en stor komplexitet i dessa frågor. Generellt sett kan dock sägas att medborgarna har ett högt men samtidigt diffust förtroende för larmnummer 112. I samband med detta kan ny mobil kommunikationsteknik tänkas ge allt fler kontaktmöjligheter, men samtidigt göra relationen mellan individ och larmfunktion allt mer komplex.

Des techniques statistiques de détection d'objet ponctuel dans le ciel ou résolu dans la mer dans des images infrarouges de veille panoramique sont développées. Ces techniques sont adaptées aux inhomogénéités présentes dans ce type d'image. Elles ne sont fondées que sur l'analyse de l'information spatiale et ont pour objectif de maîtriser le taux de fausse alarme sur chaque image. Pour les zones de ciel, une technique conjointe de segmentation et détection adaptée aux variations spatiales de la luminosité moyenne est mise en œuvre et l'amélioration des performances auxquelles elle conduit est analysée. Pour les zones de mer, un détecteur de bord à taux de fausse alarme constant en présence d'inhomogénéités et de corrélations spatiales des niveaux de gris est développé et caractérisé. Dans chaque cas, la prise en compte des inhomogénéités dans les algorithmes statistiques s'avère essentielle pour maîtriser le taux de fausse alarme et améliorer les performances de détection. / Statistical detection techniques of point target in the sky or resolved target in the sea in infrared surveillance system images are developed. These techniques are adapted to inhomogeneities present in this kind of images. They are based on the spatial information analysis and allow the control of the false alarm rate in each image.For sky areas, a joint segmentation detection technique adapted to spatial variations of the mean luminosity is developed and its performance improvement is analyzed. For sea areas, an edge detector with constant false alarm rate when inhomogeneities and grey level spatial correlations are present is developed and characterized. In each case, taking into account the inhomogeneities in these statistical algorithms is essential to control the false alarm rate and to improve the detection performance.

Introdução: A enurese (EN) é uma condição clínica de etiologia multifatorial com característica de perda de urina intermitente no período noturno que promove uma dificuldade ao convívio social da criança/adolescente e pode ser afetada pela presença de comorbidades e por imaturidade do sistema nervoso na regulação central sobre o funcionamento vesical. O conjunto de eventos promotores do episódio da enurese, desde o enchimento vesical durante o sono até a deflagração do esvaziamento vesical involuntário sem acordar constitui-se no foco principal de abordagem clínico-laboratorial e terapêutica. Objetivo: Estudar o impacto de uma avaliação multidisciplinar do indivíduo com enurese noturna monossintomática (ENM) sobre a eficácia terapêutica dos métodos tradicionalmente utilizados em sua abordagem, assim como comparar, pela polissonografia (PSG), os efeitos da intervenção sobre a estrutura do sono. Método: Estudo prospectivo de crianças de 6 a 17 anos incompletos, com ENM, diagnosticada por avaliação multidisciplinar utilizando anamnese estruturada com enfoque nefropediátrico, exame clínico, diários das eliminações, ultrassonografia de rins e vias urinárias, análises laboratoriais de sangue e urina, exame neurológico, diário do sono, questionários de avaliação do sono e polissonografia, avaliação psicológica de distúrbios do comportamento pelo CBCL e da qualidade de vida pelo PedsQL 4.0 e avaliação fisioterapêutica do equilíbrio. Condições crônicas e genéticas constituíram exclusão do protocolo. O estudo foi aprovado pelo Comitê de Ética em Pesquisa do HCFMUSP de número 0649/10 e apoio da FAPESP (2011/17589-1). Das 140 crianças/adolescentes ingressantes, foram excluídos 58 (41,4%) por comorbidades não tratáveis e/ou não aderência ao protocolo, 82/140 (58,6%) pacientes com ENM foram incluídos para intervenção terapêutica em três grupos de tratamento (alarme, desmopressina e alarme com desmopressina). A resposta à intervenção terapêutica foi avaliada no período pós-intervenção imediato (após seis meses de tratamento) e tardio (após 12 meses da suspensão do tratamento). A PSG foi realizada pré e pós-tratamento para análise comparativa da estrutura do sono. Resultados: Dos 82 pacientes, com idade média de 9,5 anos (? 2,6), 62 eram do sexo masculino (75,6%). Diagnosticou-se antecedente familiar de EN em 91,1% considerando parentes de 1° e 2° graus, constipação em 81,7% e apneia leve/moderada em 40,7%. A avaliação fisioterapêutica realizada na fase pré-intervenção demonstrou presença de alteração no controle do equilíbrio nos pacientes com enurese. A associação entre os fatores reconhecidos na abordagem clínica inicial e a gravidade da enurese (% de episódios/mês) na fase préintervenção, apresentou significância para prematuridade (p=0,03). Previamente à randomização, após abordagem clínica de constipação e/ou terapia comportamental simples, verificou-se cura de 7/82 pacientes. Foram randomizados para tratamento nas três modalidades de intervenção 75 pacientes. Durante a fase de intervenção ocorreram 14/75 (18,7%) desistências, principalmente no grupo alarme (p=0,00). Verificou-se nos 61 pacientes em tratamento, sucesso inicial (resposta completa e parcial) em 56,6% do grupo alarme, 70% do grupo desmopressina e 64% do grupo combinado (p=0,26). Sucesso contínuo (SC) ocorreu em 70% do grupo Alarme, 84,2% do grupo Desmopressina e 100% do grupo combinado (p=0,21). A recidiva ocorreu em 3/20 (15%) pacientes do grupo Alarme e 1/19 (5,2%) do grupo Desmopressina. O sucesso terapêutico se associou a redução nos escores de problemas de comportamento e melhora dos escores de qualidade de vida dos pacientes. A análise comparativa do sono pré e pós-intervenção nos pacientes com e sem alarme demonstrou aumento dos microdespertares (p=0,00), diminuição da eficiência do sono (p=0,02), diminuição de N2 (p=0,00) no grupo alarme. Conclusão: A enurese é um distúrbio multifatorial que exige uma abordagem diagnóstica estruturada. No presente estudo, a abordagem terapêutica utilizando três metodologias de intervenção demonstrou eficácia semelhante. Os benefícios associados ao sucesso terapêutico são amplos, sugere-se, no entanto, que a terapia com alarme possa repercutir negativamente na estrutura do sono / Introduction: Enuresis (NE) is a clinical condition of multifactorial etiology with intermittent nocturnal urine loss characteristic at night that leads to difficulties in child / adolescent social interaction. It can be affected by comorbidities and by immaturity of the central nervous system in the regulation of bladder function. The range of events of the bedwetting episode, from bladder filling until the involuntary triggering of bladder emptying during sleep, constitutes the main focus for clinical-laboratory and therapeutic approach. Objective: To study the impact of a multidisciplinary assessment of the patient with monosymptomatic nocturnal enuresis (MNE) on the therapeutic efficacy of the interventions traditionally used in its approach, and to compare, by polysomnography (PSG), the effects of intervention on sleep structure. Method: Prospective study of children, of 6 to 17 incomplete years, with MNE diagnosed by multidisciplinary assessment, based on: pediatric nephrology - oriented structured history, clinical examination, kidney and urinary tract ultrasound, laboratory tests of blood and urine and daily bladder and intestinal elimination diaries; neurological examination, sleep diary and sleep questionnaires and PSG; psychology evaluation using CBCL and PedsQL 4.0 questionnaires; physiotherapic evaluation of balance. The study was approved by the HCFMUSP Ethics Committee in Research of number 0649/10 and was supported by Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) Grant # 2011/17589-1. Of the 140 children/adolescents entering the study, 58 (41.4%) were excluded because of non - treatable comorbidities and / or nonadherence to study protocol, 82/140 (58.6%) patients with MNE were included for therapeutic intervention in three treatment groups (alarm, desmopressin and alarm with desmopressin). The response to therapeutic intervention was evaluated in the immediate and late post-intervention period, respectively after six of treatment and after 12 months of posttreatment follow-up. Pre and post intervention PSG were compared to evaluate therapy impact on the structure of sleep. Results: Of 82 patients, 62 were male (75.6%) with mean age of 9.5 years (± 2.6). Family history of NE was diagnosed in 91.1% of first and second degree relatives, constipation in 81.7% and mild/moderate apnea in 40.7%. The physical therapy preintervention evaluation identified an alteration in the balance control of the MNE patients. In the pre-intervention phase, an association of prematurity with severity of enuresis (% of episodes/month) (p = 0.03) was identified. Enuresis cure was achieved, prior to randomization, after clinical approach to constipation and / or simple behavioral therapy, by 7/82 patients, 75 patients were randomized to treatment with three modes of intervention. During the intervention phase, a dropout rate of 14/75 (18.7%) patients was verified, especially in the alarm group (p = 0.00). Initial success (complete and partial response) was observed in 56.6% patients of the alarm group, 70% of the desmopressin group and 64% of the combined group (p = 0.26). Continued success (SC) occurred in 70% patients of the alarm group, 84.2% of desmopressin group and 100% of the combined group (p = 0.21). Recurrence occurred in 3/20 (15%) patients of the alarm group and 1/19 (5.2%) of desmopressin group. Therapeutic success was associated with a reduction in scores for behavioral problems and with improvement of patients\' quality of life scores. The comparative analysis between pre and post-intervention sleep structure parameters, in patients treated with and without alarm, showed increased arousals (p=0.00), decreased sleep efficiency (p = 0.02), decreased N2 (p = 0.00) in the alarm group. Conclusion: Enuresis is a multifactorial disorder that requires a structured diagnostic approach. In the present study, three therapeutic intervention methodologies demonstrated similar efficacy. The benefits associated with treatment success are multiple; the present study data suggest, however, that alarm therapy can have a negative effect on sleep structure

Les systèmes d'alarmes jouent un rôle critique dans la bonne exploitation des installations industrielles modernes. Cependant, dans la plupart de ces systèmes, les alarmes ne sont pas toujours traitées correctement par les opérateurs car il y a régulièrement beaucoup trop d’alarmes à gérer, notamment lors des avalanches d’alarmes. Pour réduire le flux d'alarmes, notre approche consiste à détecter des redondances entre alarmes qui pourraient être supprimées. Pour atteindre cet objectif, nous recherchons dans un premier temps les adjacences fréquentes entre les alarmes contenues dans un historique. Ceci est réalisé en adaptant l’algorithme de recherche de motifs fréquents AprioriAll. Nous explorons également une seconde méthode consistant à trouver des précédences systématiques. Pour les découvrir, nous identifions les relations de domination et de mutuelle dépendance contenues dans l’historique des alarmes. Pour faciliter l’analyse experte, les relations découvertes sont traduites sous la forme d’un réseau de Pétri.Ces deux méthodes sont ensuite confrontées à un historique d’alarmes industriel fourni par General Electric. Les résultats obtenus montrent que nos deux méthodes permettent une réduction globale du flux d'alarmes qui est plus importante durant les avalanches. / Alarm systems play an important role for the safe and efficient operation of modern industrial plants. However, in most of industrial alarm systems, alarm flows cannot always be correctly managed by the operators as they often turn into alarm floods, sequences of numerous alarms occurring in a short period of time. To reduce the alarm flows, this report focuses on detection of redundant alarms that could be removed. This objective is met by, first, looking for frequent adjacency in the alarm log. To identify them, the frequent pattern mining algorithm AprioriAll is adapted. Another way to find potentially redundant alarms is to look for systematic predecessors. To discover them, dominations and mutual dependencies contained in the alarm log are identified. To ease this analysis, the discovered relations are depicted in the form of Petri nets.Both those methods are then tested against an industrial alarm log made available by General Electric. The results show that both methods allow a reduction of the alarm flow, with the biggest reduction being during alarm floods.

This master thesis discusses design and realization of a security device for vehicles. This device can determine its actual location, discover unauthorized trespass and send information about this to a distant master system too. This distant master system can show the exact position of the security device on a map. The hardware and software for this security device and the software for the master system are described in this paper.

Les réseaux de télécommunications deviennent de plus en plus complexes, notamment de par la multiplicité des technologies mises en œuvre, leur couverture géographique grandissante, la croissance du trafic en quantité et en variété, mais aussi de par l’évolution des services fournis par les opérateurs. Tout ceci contribue à rendre la gestion de ces réseaux de plus en plus lourde, complexe, génératrice d’erreurs et donc coûteuse pour les opérateurs. On place derrière le terme « réseaux autonome » l’ensemble des solutions visant à rendre la gestion de ce réseau plus autonome. L’objectif de cette thèse est de contribuer à la réalisation de certaines fonctions autonomiques dans les réseaux de télécommunications. Nous proposons une stratégie pour automatiser la gestion des pannes tout en couvrant les différents segments du réseau et les services de bout en bout déployés au-dessus. Il s’agit d’une approche basée modèle qui adresse les deux difficultés du diagnostic basé modèle à savoir : a) la façon d'obtenir un tel modèle, adapté à un réseau donné à un moment donné, en particulier si l'on souhaite capturer plusieurs couches réseau et segments et b) comment raisonner sur un modèle potentiellement énorme, si l'on veut gérer un réseau national par exemple. Pour répondre à la première difficulté, nous proposons un nouveau concept : l’auto-modélisation qui consiste d’abord à construire les différentes familles de modèles génériques, puis à identifier à la volée les instances de ces modèles qui sont déployées dans le réseau géré. La seconde difficulté est adressée grâce à un moteur d’auto-diagnostic actif, basé sur le formalisme des réseaux Bayésiens et qui consiste à raisonner sur un fragment du modèle du réseau qui est augmenté progressivement en utilisant la capacité d’auto-modélisation: des observations sont collectées et des tests réalisés jusqu’à ce que les fautes soient localisées avec une certitude suffisante. Cette approche de diagnostic actif a été expérimentée pour réaliser une gestion multi-couches et multi-segments des alarmes dans un réseau IMS. / While modern networks and services are continuously growing in scale, complexity and heterogeneity, the management of such systems is reaching the limits of human capabilities. Technically and economically, more automation of the classical management tasks is needed. This has triggered a significant research effort, gathered under the terms self-management and autonomic networking. The aim of this thesis is to contribute to the realization of some self-management properties in telecommunication networks. We propose an approach to automatize the management of faults, covering the different segments of a network, and the end-to-end services deployed over them. This is a model-based approach addressing the two weaknesses of model-based diagnosis namely: a) how to derive such a model, suited to a given network at a given time, in particular if one wishes to capture several network layers and segments and b) how to reason a potentially huge model, if one wishes to manage a nation-wide network for example. To address the first point, we propose a new concept called self-modeling that formulates off-line generic patterns of the model, and identifies on-line the instances of these patterns that are deployed in the managed network. The second point is addressed by an active self-diagnosis engine, based on a Bayesian network formalism, that consists in reasoning on a progressively growing fragment of the network model, relying on the self-modeling ability: more observations are collected and new tests are performed until the faults are localized with sufficient confidence. This active diagnosis approach has been experimented to perform cross-layer and cross-segment alarm management on an IMS network.

Old concrete buttress dams are sensitive to cracking if exposed to large temperature variations. The cracks can make dams sensitive to failure, depending on the size and location of the cracks. These problems can be overcome by lowering the temperature variations and stabilizing the dams. Insulation walls can be built to lower the temperature variations, and the area inside the insulation wall can be climate controlled to ensure a constant temperature. Stabilizing measures could be installing tendons, anchoring monoliths to the foundation or to keep parts of the monolith together. However, the best way to make sure the dam is functioning as expected is to monitor the behavior of the dam through different sensors. The sensors should be connected to some kind of dam monitoring software, which can indicate whether the dam is going to fail in a near future. For this to work, some kind of alert and alarm values has to be determined. The main purpose for this project is to develop a finite element model that could be used to simulate the real behavior of a concrete buttress dam and predict the future behavior of the dam. This makes it possible to determine alert and alarm values for monitoring equipment installed on the dam. Some steps are necessary to be able to create a finite element model representing the real behavior and to predict the future behavior of a dam. A first step is calibration of the model against real measurements, and during the calibration process it is important to evaluate the predictions made. A second step is to determine the normal variation in the behavior of the dam. A last step is to define suitable alert and alarm values, where the alert values are based on the normal variation of the dam and the alarm values are based on failure analyses. The results show that it is possible to calibrate a finite element model with sufficient accuracy in order for it to be used for predictions of the dams behavior. The results show two failure modes of the concrete buttress dam which deviate from previous research, where post-tensioned tendons were not included. From the results, information is given about where to place sensors to be able to capture a failure, how well the finite element model is calibrated, and what the alarm values should be. Furthermore, the results show that the evaluation of predictions made in the calibration process is of utmost importance to achieve a model representing the real behavior. / Gamla betonglamelldammar är känsliga för sprickbildning om de utsätts för stora temperaturvariationer. Sprickor kan göra dammarna känsliga för brott, beroende på storlek och placering av sprickorna. Dessa problem kan övervinnas genom att sänka temperaturvariationerna och stabilisera dammarna. Isoleringsväggar kan byggas för att sänka temperaturvariationerna, och området innanför isoleringsväggen kan klimatkontrolleras för att säkerställa en konstant temperatur. Stabiliserande åtgärder skulle kunna vara att installera spännkablar, förankring av monoliten till berggrunden eller att hålla ihop delar av monoliten. Emellertid är det bästa sättet att se till dammen fungerar som förväntat för att övervaka beteendet hos dammen genom olika sensorer. Givarna borde anslutas till någon form av programvara för dammövervakning, som kan indikera om dammen kommer att gå till brott inom en snar framtid. För att detta ska fungera måste någon form av mjuka och hårda larmvärden bestämmas. Huvudsyftet för detta projekt är att skapa en finit elementmodell som kan användas för att simulera det verkliga beteendet hos en betonglamelldamm och förutsäga framtida beteende av dammen för att kunna bestämma mjuka och hårda larmvärden för vald övervakningutrustning på dammen. Några steg är nödvändiga för att kunna skapa en finit elementmodell som representerar det verkliga beteendet och göra det möjligt att förutsäga det framtida beteendet av en damm. Ett första steg är kalibrering av modellen mot riktiga mätningar och under kalibreringsprocessen är det viktigt att utvärdera predikterade värden. Ett andra steg är att bestämma den normala variationen av dammens beteende. Ett sista steg är att definiera lämpliga värden för mjuka och hårda larmvärden, där de mjuka värdena baseras på dammens normala variation och de hårda larmvärdena på brottsanalyser. Resultaten visar att det är möjligt att kalibrera en finit elementmodell med tillräckligt god noggrannhet att den kan användas för prediktering av dammens beteende. Resultaten visar två brottmoder av betonglamelldammen som skiljer sig från tidigare studier där spännkablar inte hade inkluderats. Från resultaten ges information om var sensorer ska placeras för att kunna fånga ett brott, hur väl finita elementmodellen kalibrerats, och vilka de mjuka och hårda larmvärdena bör vara. Dessutom visar resultaten att utvärderingen av predikteringar som gjorts i kalibreringsprocessen är av yttersta vikt för att uppnå en modell som representerar det verkliga beteendet.

The ability to communicate referentially has been historically viewed as being uniquelly human. However, with the increasing amount of studies discussing the communicative faculty of animals in the context of the evolution of human language, there is now growing evidence that this ability is present in many animal species, too. Although we know a lot about mimetic abilities of domesticated African grey parrots' (Psittacus erithacus) and their competence to use human words in a referential way, we know very little about the elements of referentiality in their natural vocalization. Our goal in this work was to find, whether and in which context is functionally referential vocalization in this species of birds present. We were determining its presence by the experiment, in which we have exposed four captured African grey parrot, in the ownership of FHS UK Prague, to stimuli that in other animal species usually elicit functionally referential vocalization: predators and prefered type of food. We analysed the behaviour and vocalizations of the tested birds by analytic softwares Interact, respectively Soundforge 0.8. We find out, that despite the fact that birds reaponded to our experimental objects appropriatelly (i.e. "predators" caused more fear and the like), they give the major amout of...