Wireless Sensor Network Solution for Sustainable Food Production

Environmental monitoring has become a key aspect in food production over the last few years. Due to their low cost, low power consumption and flexibility, Wireless Sensor Networks (WSNs) have turned up as a very convenient tool to be used in these environments where no intrusion is a must. In this work, a WSN application in a food factory is presented. The paper gives an overview of the system set up, covering from the initial study of the parameters and sensors, to the hardware-software design and development needed for the final tests in the factory facilities.

Timing analysis of an inter-cell mobility procedure for a wired/wireless PROFIBUS network

Recently, there have been a few research efforts towards extending the capabilities of fieldbus networks to encompass wireless support. In previous works we have proposed a hybrid wired/wireless PROFIBUS network solution where the interconnection between the heterogeneous communication media was accomplished through bridge-like interconnecting devices. The resulting networking architecture embraced a Multiple Logical Ring (MLR) approach, thus with multiple independent tokens, where the communication between different domains was supported by the Inter-Domain Protocol (IDP). The proposed architecture also supports mobility of stations between different wireless cells. To that hybrid wired/wireless networking architecture we have proposed a worst-case response timing analysis of the IDP, without considering inter-cell mobility (or handoff) of stations. In this paper, we advance that previous work by proposing a worst-case timing analysis of the mobility procedure.

Timing analysis of a multiple logical ring wired/wireless PROFIBUS network

Recently, there have been a few research efforts towards extending the capabilities of fieldbus networks to encompass wireless support. In previous works we have proposed a hybrid wired/wireless PROFIBUS network solution where the interconnection between the heterogeneous communication media was accomplished through bridge-like interconnecting devices. The resulting networking architecture embraced a multiple logical ring (MLR) approach, thus with multiple independent tokens, to which a specific bridging protocol extension, the inter-domain protocol (IDP), was proposed. The IDP offers compatibility with standard PROFIBUS, and includes mechanisms to support inter-cell mobility of wireless nodes. We advance that work by proposing a worst-case response timing analysis of the IDP.

Body sensor network for in-home personal healthcare

A body sensor network solution for personal healthcare under an indoor environment is developed. The system is capable of logging the physiological signals of human beings, tracking the orientations of human body, and monitoring the environmental attributes, which covers all necessary information for the personal healthcare in an indoor environment. The major three chapters of this dissertation contain three subsystems in this work, each corresponding to one subsystem: BioLogger, PAMS and CosNet. Each chapter covers the background and motivation of the subsystem, the related theory, the hardware/software design, and the evaluation of the prototype’s performance.

DISCUS : an end-to-end solution for ubiquitous broadband optical access

Fiber to the premises has promised to increase the capacity in telecommunications access networks for well over 30 years. While it is widely recognized that optical-fiber-based access networks will be a necessity in the shortto medium-term future, its large upfront cost and regulatory issues are pushing many operators to further postpone its deployment, while installing intermediate unambitious solutions such as fiber to the cabinet. Such high investment cost of both network access and core capacity upgrade often derives from poor planning strategies that do not consider the necessity to adequately modify the network architecture to fully exploit the cost benefit that a fiber-centric solution can bring. DISCUS is a European Framework 7 Integrated Project that, building on optical-centric solutions such as long-reach passive optical access and flat optical core, aims to deliver a cost-effective architecture for ubiquitous broadband services. DISCUS analyzes, designs, and demonstrates end-to-end architectures and technologies capable of saving cost and energy by reducing the number of electronic terminations in the network and sharing the deployment costs among a larger number of users compared to current fiber access systems. This article describes the network architecture and the supporting technologies behind DISCUS, giving an overview of the concepts and methodologies that will be used to deliver our end-to-end network solution. © 2013 IEEE.

A static traffic assignment model combined with an artificial neural network delay model

As traffic congestion continues to worsen in large urban areas, solutions are urgently sought. However, transportation planning models, which estimate traffic volumes on transportation network links, are often unable to realistically consider travel time delays at intersections. Introducing signal controls in models often result in significant and unstable changes in network attributes, which, in turn, leads to instability of models. Ignoring the effect of delays at intersections makes the model output inaccurate and unable to predict travel time. To represent traffic conditions in a network more accurately, planning models should be capable of arriving at a network solution based on travel costs that are consistent with the intersection delays due to signal controls. This research attempts to achieve this goal by optimizing signal controls and estimating intersection delays accordingly, which are then used in traffic assignment. Simultaneous optimization of traffic routing and signal controls has not been accomplished in real-world applications of traffic assignment. To this end, a delay model dealing with five major types of intersections has been developed using artificial neural networks (ANNs). An ANN architecture consists of interconnecting artificial neurons. The architecture may either be used to gain an understanding of biological neural networks, or for solving artificial intelligence problems without necessarily creating a model of a real biological system. The ANN delay model has been trained using extensive simulations based on TRANSYT-7F signal optimizations. The delay estimates by the ANN delay model have percentage root-mean-squared errors (%RMSE) that are less than 25.6%, which is satisfactory for planning purposes. Larger prediction errors are typically associated with severely oversaturated conditions. A combined system has also been developed that includes the artificial neural network (ANN) delay estimating model and a user-equilibrium (UE) traffic assignment model. The combined system employs the Frank-Wolfe method to achieve a convergent solution. Because the ANN delay model provides no derivatives of the delay function, a Mesh Adaptive Direct Search (MADS) method is applied to assist in and expedite the iterative process of the Frank-Wolfe method. The performance of the combined system confirms that the convergence of the solution is achieved, although the global optimum may not be guaranteed.

A Static Traffic Assignment Model Combined with an Artificial Neural Network Delay Model

As traffic congestion continues to worsen in large urban areas, solutions are urgently sought. However, transportation planning models, which estimate traffic volumes on transportation network links, are often unable to realistically consider travel time delays at intersections. Introducing signal controls in models often result in significant and unstable changes in network attributes, which, in turn, leads to instability of models. Ignoring the effect of delays at intersections makes the model output inaccurate and unable to predict travel time. To represent traffic conditions in a network more accurately, planning models should be capable of arriving at a network solution based on travel costs that are consistent with the intersection delays due to signal controls. This research attempts to achieve this goal by optimizing signal controls and estimating intersection delays accordingly, which are then used in traffic assignment. Simultaneous optimization of traffic routing and signal controls has not been accomplished in real-world applications of traffic assignment. To this end, a delay model dealing with five major types of intersections has been developed using artificial neural networks (ANNs). An ANN architecture consists of interconnecting artificial neurons. The architecture may either be used to gain an understanding of biological neural networks, or for solving artificial intelligence problems without necessarily creating a model of a real biological system. The ANN delay model has been trained using extensive simulations based on TRANSYT-7F signal optimizations. The delay estimates by the ANN delay model have percentage root-mean-squared errors (%RMSE) that are less than 25.6%, which is satisfactory for planning purposes. Larger prediction errors are typically associated with severely oversaturated conditions. A combined system has also been developed that includes the artificial neural network (ANN) delay estimating model and a user-equilibrium (UE) traffic assignment model. The combined system employs the Frank-Wolfe method to achieve a convergent solution. Because the ANN delay model provides no derivatives of the delay function, a Mesh Adaptive Direct Search (MADS) method is applied to assist in and expedite the iterative process of the Frank-Wolfe method. The performance of the combined system confirms that the convergence of the solution is achieved, although the global optimum may not be guaranteed.

Context-free and context-sensitive dynamics in recurrent neural networks

Continuous-valued recurrent neural networks can learn mechanisms for processing context-free languages. The dynamics of such networks is usually based on damped oscillation around fixed points in state space and requires that the dynamical components are arranged in certain ways. It is shown that qualitatively similar dynamics with similar constraints hold for a(n)b(n)c(n), a context-sensitive language. The additional difficulty with a(n)b(n)c(n), compared with the context-free language a(n)b(n), consists of 'counting up' and 'counting down' letters simultaneously. The network solution is to oscillate in two principal dimensions, one for counting up and one for counting down. This study focuses on the dynamics employed by the sequential cascaded network, in contrast to the simple recurrent network, and the use of backpropagation through time. Found solutions generalize well beyond training data, however, learning is not reliable. The contribution of this study lies in demonstrating how the dynamics in recurrent neural networks that process context-free languages can also be employed in processing some context-sensitive languages (traditionally thought of as requiring additional computation resources). This continuity of mechanism between language classes contributes to our understanding of neural networks in modelling language learning and processing.

Aplicação de um modelo de programação linear inteira mista (PLIM) na expansão da rede de distribuição de energia elétrica em Angola

Em Angola, apenas cerca de 30% da população tem acesso à energia elétrica, nível que decresce para valores inferiores a 10% em zonas rurais mais remotas. Este problema é agravado pelo facto de, na maioria dos casos, as infraestruturas existentes se encontrarem danificadas ou não acompanharem o desenvolvimento da região. Em particular na capital angolana, Luanda que, sendo a menor província de Angola, é a que regista atualmente a maior densidade populacional. Com uma população de cerca de 5 milhões de habitantes, não só há frequentemente problemas relacionados com a falha do fornecimento de energia elétrica como há ainda uma percentagem considerável de municípios onde a rede elétrica ainda nem sequer chegou. O governo de Angola, no seu esforço de crescimento e aproveitamento das suas enormes potencialidades, definiu o setor energético como um dos fatores críticos para o desenvolvimento sustentável do país, tendo assumido que este é um dos eixos prioritários até 2016. Existem objetivos claros quanto à reabilitação e expansão das infraestruturas do setor elétrico, aumentando a capacidade instalada do país e criando uma rede nacional adequada, com o intuito não só de melhorar a qualidade e fiabilidade da rede já existente como de a aumentar. Este trabalho de dissertação consistiu no levantamento de dados reais relativamente à rede de distribuição de energia elétrica de Luanda, na análise e planeamento do que é mais premente fazer relativamente à sua expansão, na escolha dos locais onde é viável localizar novas subestações, na modelação adequada do problema real e na proposta de uma solução ótima para a expansão da rede existente. Depois de analisados diferentes modelos matemáticos aplicados ao problema de expansão de redes de distribuição de energia elétrica encontrados na literatura, optou-se por um modelo de programação linear inteira mista (PLIM) que se mostrou adequado. Desenvolvido o modelo do problema, o mesmo foi resolvido por recurso a software de otimização Analytic Solver e CPLEX. Como forma de validação dos resultados obtidos, foi implementada a solução de rede no simulador PowerWorld 8.0 OPF, software este que permite a simulação da operação do sistema de trânsito de potências.

Celorico Wireless

Mestrado em Engenharia Informática

Suunnitelma Ethernet-pohjaisen TCP/IP-verkon toteuttamiseksi prosessitietoliikenteen tarpeisiin Helsingin Energiassa

Prosessitietoliikenneverkko toimii mm. energian tuotannon, siirron ja jakelun käytönvalvontajärjestelmien tiedonsiirron alustana. Tietoliikenneverkkojen rooli osana yrityksen operatiivisia prosesseja kasvaa järjestelmäuusintojen myötä ja järjestelmien tehokas hyödyntäminen vaatii luotettavat sekä nopeat tietoverkot. Uusittavien järjestelmien tietoliikenneratkaisut perustuvat pääsääntöisesti Ethernetpohjaiseen TCP/IP-verkkotekniikkaan. Tällä hetkellä prosessijärjestelmien tiedonsiirtoyhteydet kenttälaitteiden ja valvomojärjestelmien välillä on pääsääntöisesti toteutettu PCM-siirtojärjestelmällä tai kiinteän yhteyden modeemeilla. Käytössä oleva PCM-tekniikka on elinkaarensa loppupuolella ja tiedonsiirtonopeudet sekä liityntärajapinnat eivät täytä nykyaikaisten prosessijärjestelmien tiedonsiirto¬vaatimuksia. Työssä käytiin läpikeskeisimmät Ethernet-tekniikkaan ja TCP/IP-protokollaperheeseen liittyvät standardit ja verkon aktiivilaitteet. Vaatimusmäärittelyissä selvitettiin Helsingin Energian eri liiketoimintayksiköiden tarpeita sekä energia-alan lainsäädännön vaatimuksia prosessitietoliikenneverkolle. Tarpeiden ja vaatimusten pohjalta verrattiin kolmea eri vaihtoehtoista verkkotekniikkaa: kytkintekniikkaa, reititintekniikkaa ja MPLS-tekniikkaa. Vertailun perusteella valitulla verkkotekniikalla tehtiin suunnitelma avoimesta, valmistajasta riippumattomasta standardeihin perustuvasta yritystason Ethernet-pohjaisesta TCP/IP prosessiverkkoratkaisusta. Johtopäätöksissä pohditaan valitun ratkaisun käytettävyyttä, tietoturvaa sekä toimenpiteitä niiden kehittämiseksi.

Verkkotuotteen käyttöönoton kehittäminen tuotetiedonhallinnan keinoin

Tiivistelmä Asiakas kokee tuotteen laadun omien odotusten ja oman tietämyksen kautta. Teknisessä verkkoratkaisuissa asiakas ei aina näe laatuun vaikuttavia eri komponentteja. Tällöin asiakas kokee kaikkien esille tulevien ongelmien johtuvan yrityksen tarjoamasta palvelusta. Yritys itse aiheuttaa myös ongelmia heikolla tiedonkululla yksiköiden välillä. Työssä kuvataan verkkoratkaisun käyttöönotossa esille tulevia ongelmia. Asioita tarkastellaan sekä asiakkaan että yrityksen näkökulmasta. Lisäksi tarkastellaan tuotetiedonhallinnan eri työkalujen mahdollisuuksia ongelman ratkaisemisessa. Lopputulokseksi saatiin malli, jonka avulla sisäistä tiedonkulkua parantamalla edistetään tiedon saatavuutta yrityksessä. Parantuneen tiedonkulun seurauksena voidaan tuottaa parempaa laatua sekä nopeuttaa ongelmien ratkaisua.

LTE tulevaisuuden mobiiliverkkoratkaisuna

Tässä kandidaatintyössä käytiin läpi kuinka LTE soveltuu tulevaisuuden mobiiliverkkoratkaisuksi. Ensiksi selvitettiin sen kehityshistoriaa ja vaatimuksia. Seuraavaksi käytiin läpi LTE:n teknistä puolta, kuten radiotekniikkaa, taajuusjakoa ja antenneja. Työssä tehtiin katsaus myös uuteen SAE verkkoarkkitehtuuriin. Teknillisen puolen jälkeen työssä käytiin läpi LTE:n kilpailijoita, sekä tutustuttiin tarkemmin sen pahimpaan haastajaan WiMAX:iin. Näiden kahden tekniikan välillä tehtiin myös vertailu. Lopuksi katsottiin tulevaisuuden trendejä ja mobiilin datasiirron kehittymistä sekä tutustuttiin LTE:n seuraavaan versioon LTE-Advanced:iin.

Reinforcement learning in dendritic structures

The discovery of binary dendritic events such as local NMDA spikes in dendritic subbranches led to the suggestion that dendritic trees could be computationally equivalent to a 2-layer network of point neurons, with a single output unit represented by the soma, and input units represented by the dendritic branches. Although this interpretation endows a neuron with a high computational power, it is functionally not clear why nature would have preferred the dendritic solution with a single but complex neuron, as opposed to the network solution with many but simple units. We show that the dendritic solution has a distinguished advantage over the network solution when considering different learning tasks. Its key property is that the dendritic branches receive an immediate feedback from the somatic output spike, while in the corresponding network architecture the feedback would require additional backpropagating connections to the input units. Assuming a reinforcement learning scenario we formally derive a learning rule for the synaptic contacts on the individual dendritic trees which depends on the presynaptic activity, the local NMDA spikes, the somatic action potential, and a delayed reinforcement signal. We test the model for two scenarios: the learning of binary classifications and of precise spike timings. We show that the immediate feedback represented by the backpropagating action potential supplies the individual dendritic branches with enough information to efficiently adapt their synapses and to speed up the learning process.

Reinforcement learning in dendritic structures

The discovery of binary dendritic events such as local NMDA spikes in dendritic subbranches led to the suggestion that dendritic trees could be computationally equivalent to a 2-layer network of point neurons, with a single output unit represented by the soma, and input units represented by the dendritic branches. Although this interpretation endows a neuron with a high computational power, it is functionally not clear why nature would have preferred the dendritic solution with a single but complex neuron, as opposed to the network solution with many but simple units. We show that the dendritic solution has a distinguished advantage over the network solution when considering different learning tasks. Its key property is that the dendritic branches receive an immediate feedback from the somatic output spike, while in the corresponding network architecture the feedback would require additional backpropagating connections to the input units. Assuming a reinforcement learning scenario we formally derive a learning rule for the synaptic contacts on the individual dendritic trees which depends on the presynaptic activity, the local NMDA spikes, the somatic action potential, and a delayed reinforcement signal. We test the model for two scenarios: the learning of binary classifications and of precise spike timings. We show that the immediate feedback represented by the backpropagating action potential supplies the individual dendritic branches with enough information to efficiently adapt their synapses and to speed up the learning process.