Diversas pesquisas apoiam e investigam o teste de programas concorrentes, as quais objetivam, principalmente, a proposição de critérios de teste e mecanismos para execução das diferentes sincronizações entre processos ou threads. As características específicas dessas aplicações podem ocasionar diferentes tipos de defeitos, os quais, em sua maioria, não são facilmente identificados. Nesse contexto, a geração automática de dados de teste pode apoiar a atividade de teste atuando na seleção de entradas mais representativas, ou seja, aquelas com maior probabilidade de revelar defeitos. Apesar disso, poucas pesquisas abordam este tema no contexto de programas concorrentes, e as existentes não consideram aspectos importantes desse tipo de aplicação. A geração de dados de teste para programas sequenciais dispõe de uma variedade de técnicas que apoiam a seleção dos dados de teste. Essas técnicas têm sido estendidas para o contexto de programas concorrentes partindo da premissa que esses programas necessitam de abordagens mais complexas para seleção de entradas, em decorrência disso um maior custo é imposto ao teste. Considerando esse contexto, uma lacuna ainda em aberto é a avaliação das técnicas para o cenário de programas concorrentes. Neste trabalho a avaliação das técnicas foi explorada por meio da realização de estudos experimentais, os quais avaliaram diferentes técnicas de geração de dados de teste para o contexto de programas concorrentes, considerando a eficácia em revelar defeitos, cobertura de critérios e custo para atividade de teste. Os resultados obtidos demonstraram que as técnicas empregadas para programas sequenciais não atingem o mínimo esperado para este tipo aplicação. Apesar disso, as técnicas investigadas apresentaram características importantes que podem auxiliar a atividade de teste para programas concorrentes e a proposição de abordagens efetivas de geração de dados para esse contexto. / The concurrent program testing has been largely investigated with propositions of testing criteria and mechanisms, which aim mainly to testing criteria proposition and mechanisms for execution of different synchronizations. The specific characteristics of these applications can lead to the different types of faults, which, in most of cases, are not easily identified. In this context, the automatic test data generation can support the testing activity acting in selecting the most representative data tests, i.e. those most likely to reveal faults. Nevertheless, few studies address this issue in the context of concurrent programs, and the these studies do not consider important aspects of this type of application. In contrast, we can find several techniques proposed to support the test data generation for sequential programs. These techniques have been extended to the context of concurrent programs on the premise that these programs require more complex approaches for selection of test data. As a result, a greater cost for testing activity is enforced. Considering this context, a gap still open is the evaluation of techniques for the scenario of concurrent programs. In this work the evaluation of techniques was explored through experimental studies, which different techniques of test data generatoon were evaluated, considering effectiveness, testing coverage and application cost. The results showed that the generation techniques used for sequential programs used in the experimental study do not reach the minimum expected in terms of effectiveness and cost for concurrent programs. Nevertheless, the techniques investigated showed significant features that can help the proposition of effective approaches for test data generation applied to concurrent programs.

A wireless ad hoc network consists of a number of mobile nodes that temporarily form a dynamic infrastructure-less network. New routing protocols that can adapt to the frequent topology changes induced by node mobility and varying link qualities are needed. During the last decade dozens of different ad hoc routing protocols have been proposed, optimized and partially compared, mainly through simulation studies. This thesis takes an experimental approach to the evaluation of ad hoc routing protocols. We argue that real world experiments are needed in order to complement simulation studies, and to gain practical experience and insights that can provide feedback to routing protocol design and existing simulation models. For example, we discovered a performance discrepancy for the AODV protocol between real world experiments and corresponding simulation studies. This so called ``communication gray zone'' problem was explored and countermeasures were implemented. As a result we could eliminate this performance problem to a large extent. We have implemented a software-based testbed called APE to carry out efficient and systematic experimental evaluation of ad hoc routing protocols. Experiments with up to 37 participating ad hoc nodes have demonstrated APE's ability to scale efficiently and assess repeatability between test runs. APE is part of our methodology for test repeatability in a real world ad hoc routing protocol testbed. It addresses the repeatability issue induced by stochastic factors like the radio environment and node mobility. Using APE, we have performed a systematic experimental evaluation of three ad hoc routing protocols (AODV, OLSR and LUNAR). Our results show that TCP does not work satisfactorily even in very small networks with limited mobility.

Policy makers in education and international development have lately gravitated toward the randomized controlled trial (RCT)—an evaluation design that randomly assigns a sample of people or households into an intervention group and a control group in order to measure the differential effect of the intervention—as a means to determine program impact. As part of federal regulations, the U.S. Department of Education and the U.S. Agency for International development explicitly declared a preference for the RCT. When advocating for adopting the RCT model as the preferred evaluation tool, policy makers point to the success story of medical trials and how they revolutionized medicine from Medieval charlatanry to a modern life-saving discipline. By presenting a more nuanced account of the role of the RCT in medical history, however, this study finds that landmark RCTs were accompanied with challenges, Evidence-Based Medicine had rightful critics, and opportunistic biases in drug trials apply equally to education policy and international development. This study also examines the recent privileged role of the RCT in education and international development, concluding that its initial promise was not entirely born out when put into practice, as the national Reading First Initiative exemplifies. From a comparative perspective, the RCT movements also encountered major RCT critics, whose voices were not initially heard. These voices, however, seem to have contributed to a swing of the pendulum away from RCT primacy back towards greater methodological pluralism. A major conclusion of this study is that policy makers should exercise great caution when using RCTs as a policy evaluation tool. This conclusion is arrived at via considering RCT biases, challenges, and limited generalizability; understanding its interpretive-qualitative components; and broadening the overall methodological repertoire to better enable evaluations of macro-policy interventions. / text

Diversas pesquisas apoiam e investigam o teste de programas concorrentes, as quais objetivam, principalmente, a proposição de critérios de teste e mecanismos para execução das diferentes sincronizações entre processos ou threads. As características específicas dessas aplicações podem ocasionar diferentes tipos de defeitos, os quais, em sua maioria, não são facilmente identificados. Nesse contexto, a geração automática de dados de teste pode apoiar a atividade de teste atuando na seleção de entradas mais representativas, ou seja, aquelas com maior probabilidade de revelar defeitos. Apesar disso, poucas pesquisas abordam este tema no contexto de programas concorrentes, e as existentes não consideram aspectos importantes desse tipo de aplicação. A geração de dados de teste para programas sequenciais dispõe de uma variedade de técnicas que apoiam a seleção dos dados de teste. Essas técnicas têm sido estendidas para o contexto de programas concorrentes partindo da premissa que esses programas necessitam de abordagens mais complexas para seleção de entradas, em decorrência disso um maior custo é imposto ao teste. Considerando esse contexto, uma lacuna ainda em aberto é a avaliação das técnicas para o cenário de programas concorrentes. Neste trabalho a avaliação das técnicas foi explorada por meio da realização de estudos experimentais, os quais avaliaram diferentes técnicas de geração de dados de teste para o contexto de programas concorrentes, considerando a eficácia em revelar defeitos, cobertura de critérios e custo para atividade de teste. Os resultados obtidos demonstraram que as técnicas empregadas para programas sequenciais não atingem o mínimo esperado para este tipo aplicação. Apesar disso, as técnicas investigadas apresentaram características importantes que podem auxiliar a atividade de teste para programas concorrentes e a proposição de abordagens efetivas de geração de dados para esse contexto. / The concurrent program testing has been largely investigated with propositions of testing criteria and mechanisms, which aim mainly to testing criteria proposition and mechanisms for execution of different synchronizations. The specific characteristics of these applications can lead to the different types of faults, which, in most of cases, are not easily identified. In this context, the automatic test data generation can support the testing activity acting in selecting the most representative data tests, i.e. those most likely to reveal faults. Nevertheless, few studies address this issue in the context of concurrent programs, and the these studies do not consider important aspects of this type of application. In contrast, we can find several techniques proposed to support the test data generation for sequential programs. These techniques have been extended to the context of concurrent programs on the premise that these programs require more complex approaches for selection of test data. As a result, a greater cost for testing activity is enforced. Considering this context, a gap still open is the evaluation of techniques for the scenario of concurrent programs. In this work the evaluation of techniques was explored through experimental studies, which different techniques of test data generatoon were evaluated, considering effectiveness, testing coverage and application cost. The results showed that the generation techniques used for sequential programs used in the experimental study do not reach the minimum expected in terms of effectiveness and cost for concurrent programs. Nevertheless, the techniques investigated showed significant features that can help the proposition of effective approaches for test data generation applied to concurrent programs.

The diploma thesis deals with the thermal-humidity stress of the wooden wall. The work is focused on the connection of the wall to the base structure of the building. Three details were chosen. The work was focused on detail with the most common structure of an external wall used in the Czech Republic on the composition with a supporting structure made of KVH columns, which are filled with mineral insulation. This construction is covered with plate elements. The insulation from the exterior is made of ETICS with expanded polystyrene thermal insulation. This detail was assessed in the software. To compare the results calculated by real-time software, an experimental model was made, which was subjected to experimental measurements. Part of the diploma thesis is a comparison of detail stress under different boundary design conditions, with or without anchoring.

The purpose of this research was to evaluate the Capuchin API launched by Sony Ericsson at Lund, Sweden in 2008. The Capuchin API bridges Adobe’s Flash graphics and effects with JSR support from Java ME. We evaluated Capuchin API with regard to its suitability for a Mobile TV application. We tested this API in Ericsson’s TV lab where we had access to live TV streams and online multimedia resources by implementing a Mobile TV client. This test application was named “Min TV”, in English: “My TV”. Using Capuchin in the Ericsson TV lab environment has shown that it has some benefits, but also has many drawbacks. The Flash developer can be used to create an animated user interface and Java developers can do complex programming. At this early stage Capuchin technology is not mature enough, nor is it suitable for Mobile TV client development. Only after Sony Ericsson adds features such as soft keys, easier debugging of Flash Lite standalone applications, test emulator support in the software development kit, and more data communication methods than string and number, only then it will be a suitable technology for Mobile TV applications. Ericsson’s current Mobile TV application client was built using a framework called ECAF, which supports a graphics frontend and Java ME as backend. We compared ECAF and Min TV with respect to parameters such as: flexibility, performance, memory footprint, code size, and cost of skinning. (All these parameters are explained in detail in the methodology chapter.) As a possible future technology for Mobile TV, we evaluated a number of different presentation/graphics technologies including HECL, SVG Tiny, MIDP 3.0, .NET Compact Framework, etc. Moreover, we examed if a pure Flash Lite client application is a viable solution for Mobile TV. The comparison of different presentation technologies showed that Java ME is a comprehensive platform for mobile development offering all the necessary support from third party graphical user interface makers. .NET CF also looks like a good option for development with the scaled down capabilities for different programming languages supported using CLR. / Syftet med denna forskning var att utvärdera Capuchin API lanserades av Sony Ericsson i Lund, Sverige 2008. Den Capuchin API broar Adobe Flash grafik och effekter med JSR stöd från Java ME. Vi utvärderade Capuchin API med avseende på dess lämplighet för ett mobil-tv ansökan. Vi testade detta API i Ericssons TV lab där vi hade tillgång till TV-strömmar och online multimediaresurser genom en mobil-TV-klient. Detta test ansökan hette "Min TV", på engelska: "My TV". Använda Capuchin i Ericsson TV lab miljö har visat att det har vissa fördelar, men också många nackdelar. Flash-utvecklare kan användas för att skapa en animerad användargränssnitt och Java utvecklare kan göra komplexa programmering. På detta tidiga stadium Capuchin tekniken inte mogen, det är inte heller lämpliga för mobil-TV-klient utveckling. Först efter Sony Ericsson lägger till detaljer såsom mjuka nycklar, enklare felsökning av Flash Lite fristående program, testa emulator stöd i Software Development Kit, och mer data kommunikationsmetoder än string och antal, först då kommer det att vara en lämplig teknik för mobil-TV-program . Ericssons nuvarande mobil-tv ansökan klient byggdes med hjälp av en ram som kallas ECAF, som stöder en grafiska gränssnittet och Java ME som backend. Vi jämförde ECAF och min TV med avseende på parametrar såsom flexibilitet, prestanda, minne fotavtryck kod storlek och kostnaden för avhudning. (Alla dessa parametrar förklaras i detalj i den metod kapitel.) Som en möjlig framtida teknik för mobil-TV Vi utvärderade ett antal olika presentation / grafik teknik inklusive HECL, SVG Tiny, MIDP 3.0,. NET Compact Framework, etc. Dessutom har vi examed om en ren Flash Lite klientprogrammet är en hållbar lösning för mobil-TV. Jämförelsen mellan olika presentation teknik visade att Java ME är en övergripande plattform för mobila utvecklingen erbjuder allt nödvändigt stöd från tredje part grafiskt användargränssnitt beslutsfattare. . NET CF också ser ut som ett bra alternativ för utvecklingen med ned kapacitet för olika programspråk som stöds med hjälp av CLR.

In-memory column-store database systems are state of the art for the efficient processing of analytical workloads. In these systems, data compression as well as vectorization play an important role. Currently, the vectorized processing is done using regular SIMD (Single Instruction Multiple Data) extensions of modern processors. For example, Intel’s latest SIMD extension supports 512-bit vector registers which allows the parallel processing of 8× 64-bit values. From a database system perspective, this vectorization technique is not only very interesting for compression and decompression to reduce the computational overhead, but also for all database operators like joins, scan, as well as groupings. In contrast to these SIMD extensions, NEC Corporation has recently introduced a novel pure vector engine (supercomputer) as a co-processor called SX-Aurora TSUBASA. This vector engine features a vector length of 16.384 bits with the world’s highest bandwidth of up to 1.2 TB/s, which perfectly fits to data-intensive applications like in-memory database systems. Therefore, we describe the unique architecture and properties of this novel vector engine in this paper. Moreover, we present selected in-memory column-store-specific evaluation results to show the benefits of this vector engine compared to regular SIMD extensions. Finally, we conclude the paper with an outlook on our ongoing research activities in this direction.

The exploitation of data as well as hardware properties is a core aspect for efficient data management. This holds in particular for the field of in-memory data processing. Aside from increasing main memory capacities, in-memory data processing also benefits from novel processing concepts based on lightweight compressed data. To speed up compression as well as decompression, an active research field deals with the specialization of these algorithms to hardware features such as vectorization using SIMD instructions. Most of the vectorized implementations have been proposed for 128 bit vector registers. However, hardware vendors still increase the vector register sizes, whereby a straightforward transformation to these wider vector sizes is possible in most-cases. Thus, we systematically investigated the impact of different SIMD instruction set extensions with wider vector sizes on the behavior of straightforward transformed implementations. In this paper, we will describe our evaluation methodology and present selective results of our exhaustive evaluation. In particular, we will highlight some challenges and present first approaches to tackle them.

Τα Δίκτυα Αισθητήρων αποτελούνται από ένα μεγάλο αριθμό μικρών αυτόνομων συσκευών, που αλληλεπιδρούν με το άμεσο περιβάλλον τους μέσω αισθητήρων, συλλέγουν δεδομένα και τα προωθούν προς ένας σταθερό, συνήθως, κέντρο ελέγχου, με αναμεταδόσεις στους ενδιάμεσους κόμβους. Η διαδικασία αυτή έχει ως αποτέλεσμα τη μεγάλη κατανάλωση ενέργειας στις συσκευές, ιδιαίτερα σε αυτές που βρίσκονται κοντά στο κέντρο ελέγχου, αφού πρέπει να αναμεταδίδουν και τα δεδομένα που φτάνουν από το υπόλοιπο δίκτυο προς το κέντρο ελέγχου. Για την επίτευξη μιας πιο ισορροπημένης και αποδοτικής διαδικασίας συλλογής δεδομένων, τα τελευταία χρόνια έχει υιοθετηθεί μια νέα προσέγγιση, όπου το κέντρο ελέγχου είναι κινητό. Η βασική ιδέα είναι ότι το κέντρο ελέγχου διαθέτει σημαντικά και εύκολα ανανεώσιμα αποθέματα ενέργειας, επομένως μπορεί να κινείται στην περιοχή όπου έχει αναπτυχθεί το δίκτυο αισθητήρων, αναλαμβάνοντας να συλλέξει τα δεδομένα από τους κόμβους με πολύ μικρό κόστος. Ωστόσο, η μετάδοση των δεδομένων μπορεί να παρουσιάζει σημαντικές καθυστερήσεις. Συλλογή δεδομένων με προσαρμοστικούς χρόνους αναμονής: Στην παρούσα διατριβή αναπτύχθηκαν πρωτόκολλα ελέγχου της κίνησης ενός κέντρου ελέγχου σε δίκτυο αισθητήρων με ανομοιογενή ανάπτυξη των κόμβων αισθητήρων, με στόχο την αποδοτική, ως προς την ενέργεια και τον χρόνο παράδοσης, συλλογή των δεδομένων. Πιο συγκεκριμένα, αρχικά παρουσιάζεται ένα πρωτόκολλο με βάση το οποίο το κέντρο ελέγχου διαιρεί νοητά το δίκτυο σε περιοχές τις οποίες και επισκέπτεται διαδοχικά, σταματώντας σε κάθε περιοχή για ένα συγκεκριμένο χρονικό διάστημα, ώστε να συλλέξει τα δεδομένα. Προτείνουμε δύο τρόπους κίνησης του κέντρου ελέγχου, ντετερμινιστικό και τυχαίο. Στην τυχαία κίνηση, η επιλογή της επόμενης περιοχής την οποία θα επισκεφτεί το κέντρο ελέγχου γίνεται με τυχαίο τρόπο, εισάγοντας όμως ένα όρο μεροληψίας, έτσι ώστε να προτιμούνται περιοχές που έχουν δεχτεί λιγότερες επισκέψεις. Επιπλέον η μέθοδός μας αποφασίζει το χρόνο παύσης σε κάθε περιοχή λαμβάνοντας υπόψιν κάποιες βασικές παραμέτρους του δικτύου, όπως τα αρχικά αποθέματα ενέργειας των κόμβων αισθητήρων και την πυκνότητα της κάθε περιοχής, έτσι ώστε να παραμένει περισσότερο χρόνο σε περιοχές με μεγαλύτερη πυκνότητα, άρα και μεγαλύτερη ποσότητα πληροφορίας. Με τον τρόπο αυτό επιτυγχάνεται η γρήγορη κάλυψη όλου του δικτύου, καθώς επίσης και η δίκαιη εξυπηρέτηση των επιμέρους περιοχών του δικτύου. Προσαρμοστικοί τυχαίοι περίπατοι Στη συνέχεια, μελετάται η χρήση τυχαίων περιπάτων κατά την κίνηση του κέντρου ελέγχου σε δίκτυα αισθητήρων με στόχο την επίτευξη ενός ικανοποιητικού σημείου ισορροπίας μεταξύ κατανάλωσης ενέργειας και καθυστέρησης στην παράδοση των μηνυμάτων. Για την ικανοποίηση του στόχου αυτού, προτείνουμε τρεις νέους τυχαίους περιπάτους, τους α) Τυχαίος Περίπατος με Αδράνεια, κατά τον οποίο το κινούμενο αντικείμενο τείνει να διατηρεί την ίδια κατεύθυνση στην κίνησή του όσο ανακαλύπτει κόμβους αισθητήρων που δεν έχει επισκεφτεί και αλλάζει την κατεύθυνσή του όταν φτάνει σε κόμβους που έχει ξαναεπισκεφτεί, β) Explore-and-Go, κατά τον οποίο το κινούμενο αντικείμενο τείνει να εκτελεί μια Brownian κίνηση γύρω από την περιοχή του όσο υπάρχουν κόμβοι που δεν έχουν δεχτεί επίσκεψη, γ) Curly Random Walk, όπου το κινούμενο αντικείμενο διαπερνάει όλη την περιοχή του δικτύου ξεκινώντας από το κέντρο και επεκτείνοντας την κίνησή του με συνεχόμενες κυκλικές κινήσεις προς τα έξω. Για την εφαρμογή των τυχαίων περιπάτων χρησιμοποιούμε ένα νοητό πλέγμα ώστε να καλύπτουμε την περιοχή του δικτύου αισθητήρων• οι περίπατοι κινούνται πάνω στους κόμβους του πλέγματος. Αν και στις περισσότερες περιπτώσεις οι τυχαίοι περίπατοι μελετώνται σε Gn,p και Grid γράφους, τα δίκτυα αισθητήρων μοντελοποιούνται με μεγαλύτερη ακρίβεια χρησιμοποιώντας το μοντέλο των Random Geometric Graphs (RGG), εφόσον έτσι αναπαρίσταται καλύτερα η χωρική εγγύτητα του δικτύου. Οι παραπάνω τυχαίοι περίπατοι δεν δίνουν τα επιθυμητά αποτελέσματα όταν τρέχουν σε RGG. Έτσι οδηγηθήκαμε στο σχεδιασμό ενός νέου τυχαίου περιπάτου, του γ-Stretched Random Walk, η βασική ιδέα του οποίου είναι να μεροληπτεί υπέρ της επίσκεψης των πιο μακρινών γειτόνων του τρέχοντος κόμβου έτσι ώστε να μειώσει στο ελάχιστο τις επικαλύψεις στις επισκέψεις. Αλγόριθμοι που λαμβάνουν υπόψιν την ηλεκτρομαγνητική ακτινοβολία στο δίκτυο: Εκτός από τη μελέτη της κίνησης του κέντρου ελέγχου σε δίκτυα αισθητήρων, στη διατριβή αυτή παρουσιάζεται μια πρώτη προσπάθεια μελέτης θεμάτων σχετικά με την επίγνωση της εκπομπή ακτινοβολίας σε περιβάλλοντα όπου λειτουργούν πολλαπλά ετερογενή ασύρματα δίκτυα. Ως ακτινοβολία σε ένα σημείου του τρισδιάστατου χώρου καλούμε τη συνολική ποσότητητα ηλεκτρομαγνητικής ακτινοβολίας που δέχεται το σημείο αυτό. Έτσι, καταρχάς μελετάμε σε αναλυτικό επίπεδο την ακτινοβολία σε διάφορες γνωστές τοπολογίες (τυχαίες, πλέγματα) και κατόπιν επικεντρώνουμε το ενδιαφέρον μας στην εύρεση ενός μονοπατιού ελάχιστης ακτινοβολίας το οποίο ακολουθείται από κάποιο άτομο που κινείται στην περιοχή που καλύπτεται από ένα ασύρματο δίκτυο αισθητήρων. Προτείνουμε τρεις ευρετικές μεθόδους για την εύρεση του μονοπατιού καθώς το άτομο κινείται, ενώ υπολογίζουμε και την οffline λύση χρησιμοποιώντας τον αλγόριθμο ελάχιστου μονοπατιού. Κατόπιν, εξετάζουμε το θεμελιώδες πρόβλημα της διάδοσης των δεδομένων σε ασύρματα δίκτυα αισθητήρων, προσπαθώντας τόσο να παραμείνει γρήγορη η διαδικασία παράδοσης των μηνυμάτων, παράλληλα όμως και η συνολική ηλεκτρομαγνητική ακτινοβολία που παράγεται από τις συνεχείς ασύρματες μεταδόσεις να διατηρηθεί σε χαμηλά επίπεδα. Αυτό επιτυγχάνεται αρχικά χρησιμοποιώντας κάποιες άπληστες ευρετικές μεθόδους που όμως λαμβάνουν υπόψιν την ακτινοβολία. Επιπλέον, οι μέθοδοι αυτοί συνδυάζονται με μεθόδους που πραγματοποιούν back-off στο χρόνο, χρησιμοποιώντας τοπικές ιδιότητες του δικτύου (όπως ο αριθμός γειτόνων, η απόσταση από το κέντρο ελέγχου), έτσι ώστε «απλωθεί» κατά κάποιο τρόπο η ακτινοβολία τόσο ως προς το χρόνο αλλά και ως προς το χώρο. Τα προτεινόμενα πρωτόκολλα αξιολογήθηκαν πειραματικά μέσω προσομοίωσης, χρησιμοποιώντας ποικίλες τιμές για βασικές παραμέτρους του δικτύου και σύγκρινοντάς τα με σχετικές υπάρχουσες ευρέως αποδεκτές μεθόδους. Συστημικές Εφαρμογές: Τέλος, στη διατριβή παρουσιάζονται κάποιες συστημικές εφαρμογές ασύρματων δικτύων αισθητήρων σε κτίρια. Συγκεκριμένα, η πρώτη εφαρμογή αναλαμβάνει σε περίπτωση ανίχνευσης φωτιάς, την εύρεση του ελάχιστου μονοπατιού μακριά από το σημείο όπου έγινε η ανίχνευση. Επιπλέον, παρέχει καθοδήγηση στους ενοίκους του κτιρίου (οι οποίοι μοντελοποιούνται από ένα κινούμενο ρομπότ) έτσι ώστε να εγκαταλείψουν με ασφάλεια το κτίριο. Η επόμενη εφαρμογή παρουσιάζει τη δυνατότητα της απρόσκοπτης διασύνδεσης αυτοματισμών έξυπνων κτιρίων, αποτελούμενων από ενσωματωμένα συστήματα, στο διαδίκτυο και την αφαιρετικοποίησή τους ως απλά web services. Η προσέγγιση αυτή έχει στόχο την δημιουργία ενός ευέλικτου, εύκολα κλιμακώσιμου συστήματος που είναι προσβάσιμο και ελεγχόμενο απομακρυσμένα. Η προσέγγιση που ακολουθήθηκε και παρουσιάζεται στην παρούσα διατριβή περιλαμβάνει την ανάπτυξη ενός αριθμού αισθητήρων μέσα σε ένα κτίριο, οι οποίοι αποκτούν IPv6 διεύθυνση ώστε να είναι προσβάσιμοι διαδικτυακά, ενώ παράλληλα διασυνδέονται με ηλεκτρικές συσκευές του κτιρίου για σχηματισμό αυτοματισμών. Τέλος αναπτύχθηκε μία web εφαρμογή για απομακρυσμένη διαχείριση του δικτύου και του κτιρίου γενικότερα. / Wireless Sensor Networks consist of a large number of small, autonomous devices, that are able to interact with their environment by sensing and collaborate to fulfill their tasks, as, usually, a single node is incapable of doing so; and they use wireless communication to enable this collaboration. The collected data is disseminated to a static control point – data sink in the network, using node to node - multi-hop data propagation. However, sensor devices consume significant amounts of energy in addition to increased implementation complexity, since a routing protocol is executed. Also, a point of failure emerges in the area near the control center where nodes relay the data from nodes that are farther away. Recently, a new approach has been developed that shifts the burden from the sensor nodes to the sink. The main idea is that the sink has significant and easily replenishable energy reserves and can move inside the area the sensor network is deployed, in order to acquire the data collected by the sensor nodes at very low energy cost. However, the need to visit all the regions of the network may result in large delivery delays. Data collection with biased stop times: In this work we have developed protocols that control the movement of the sink in wireless sensor networks with non-uniform deployment of the sensor nodes, in order to succeed an efficient (with respect to both energy and latency) data collection. More specifically, we first propose a protocol, where the sink partitions the network area in equal square regions and then performs a network traversal by visiting each area sequentially. Also, it pauses in each area for a certain amount of time, in order to collect the data. Two network traversal methods are proposed, a deterministic and a random one. When the sink moves in a random manner, the selection of the next area to visit is done in a biased random manner depending on the frequency of visits of its neighbor areas. Thus, less frequently visited areas are favored. Moreover, our method locally determines the stop time needed to serve each region with respect to some global network resources, such as the initial energy reserves of the nodes and the density of the region, stopping for a greater time interval at regions with higher density, and hence more traffic load. In this way, we achieve accelerated coverage of the network as well as fairness in the service time of each region. Besides randomized mobility, we also propose an optimized deterministic trajectory without visit overlaps, including direct (one-hop) sensor-to-sink data transmissions only. Adaptive random walks: Afterwards, in order to achieve satisfactory energy-latency trade-offs the use of random walks for the sink' s motion pattern is studied. Towards this direction three new random walks evaluated on a grid overlaying the wireless sensor network are proposed. The first one is the Random Walk with Inertia where the sink tends to keep the same direction as long as it discovers new nodes, while changing direction when it encounters already visited ones. The second one is the Explore-and-Go Random Walk, where as long as there are undiscovered nodes on the nearby sub-regions of the network it tends to make a Brownian-like motion until all this area is covered. When no new sensors are discovered, it performs a more or less straight-line walk in order to move to a different, possibly unvisited area. The last one is the Curly Random Walk where the sink traverses the network area beginning from the center and expanding its traversal to the entire network area with consecutive circular-like moves. In random walk studies the Gn,p and Grid graph models are well established. However, wireless sensor networks are more accurately modeled via Random Geometric Graphs (RGG), as RGG better capture certain characteristics of WSN's such as link existence dependencies of neighbouring nodes due to geometric proximity. The above mentioned random walks do not behave well on this particular graph model, thus a new random walk was defined, the so called γ-stretched random walk. Its basic idea is to favour visiting distant neighbours of the current node towards reducing node overlap. Radiation-aware algorithms: Except for the issue of mobility in wireless sensor networks, in this work we also attempt (probably for the first time from a distributed networking perspective) to investigate the aspect of electromagnetic radiation in modern and future heterogeneous wireless networks. We call “radiation” at a target elementary surface the total amount of electromagnetic quantity (in terms of energy or power density) it is exposed to. Thus, we first evaluate, both mathematically and by simulation, the radiation in well known sensor network topologies (random, grid) and then focus on the minimum radiation path problem of finding low radiation trajectories for a person moving in a sensor network. We propose three online heuristics and then we identify the (offline) optimum path given by the shortest paths' algorithm. Afterwards, we focus on the fundamental problem of efficient data propagation in wireless sensor networks, trying to keep latency low while maintaining at low levels the radiation cumulated by wireless transmissions. We first propose greedy and oblivious routing heuristics that are radiation aware. We then combine them with temporal back-off schemes that use local properties of the network (e.g. number of neighbours, distance from sink) in order to “spread” radiation in a spatio-temporal way. Al the proposed protocols were evaluated via simulation, in diverse network settings and comparatively to related state of the art solutions. Systems and applications: Finally, in this work we present two applications of wireless sensor networks in buildings. More specifically, the first application, in the event of a fire inside a monitored building, uses the information from the deployed sensor network in order to find the shortest safest path away from the emergency and provides navigation guidance to the occupants (modelled by a mobile robot), in order to safely evacuate the building. The second application addresses networked embedded systems enabling the seamless interconnection of smart building automations to the Internet and their abstractions as web services, using the latest technologies based on IPv6, such as 6LOWPAN, COAP and RESTLess Architecture.

<p>A <i>wireless ad hoc network </i>consists of a number of mobile nodes that temporarily form a dynamic infrastructure-less network. New routing protocols that can adapt to the frequent topology changes induced by node mobility and varying link qualities are needed. During the last decade dozens of different ad hoc routing protocols have been proposed, optimized and partially compared, mainly through simulation studies.</p><p>This thesis takes an experimental approach to the evaluation of ad hoc routing protocols. We argue that real world experiments are needed in order to complement simulation studies, and to gain practical experience and insights that can provide feedback to routing protocol design and existing simulation models. For example, we discovered a performance discrepancy for the AODV protocol between real world experiments and corresponding simulation studies. This so called ``communication gray zone'' problem was explored and countermeasures were implemented. As a result we could eliminate this performance problem to a large extent. </p><p>We have implemented a software-based testbed called APE to carry out efficient and systematic experimental evaluation of ad hoc routing protocols. Experiments with up to 37 participating ad hoc nodes have demonstrated APE's ability to scale efficiently and assess repeatability between test runs. APE is part of our methodology for test repeatability in a real world ad hoc routing protocol testbed. It addresses the repeatability issue induced by stochastic factors like the radio environment and node mobility. Using APE, we have performed a systematic experimental evaluation of three ad hoc routing protocols (AODV, OLSR and LUNAR). Our results show that TCP does not work satisfactorily even in very small networks with limited mobility.</p>