<p>The last few decades have been a very exciting period in thedevelopment of micro-electronics and brought us to the brink ofimplementing entire systems on a single chip, on a hithertounimagined scale. However an unforeseen challenge has croppedup in the form of managing wires, which have become the mainbottleneck in performance, masking the blinding speed of activedevices. A major problem is that increasingly complicatedeffects need to be modelled, but the computational complexityof any proposed model needs to be low enough to allow manyiterations in a design cycle.</p><p>This thesis addresses the issue of closed form modelling ofthe response of coupled interconnect systems. Following astrict mathematical approach, second order models for thetransfer functions of coupled RC trees based on the first andsecond moments of the impulse response are developed. The2-pole-1-zero transfer function that is the best possible fromthe available information is obtained for the signal path fromeach driver to the output in multiple aggressor systems. Thisallows the complete response to be estimated accurately bysumming up the individual waveforms. The model represents theminimum complexity for a 2-pole-1-zero estimate, for this classof circuits.</p><p>Also proposed are new techniques for the optimisation ofwires in on-chip buses. Rather than minimising the delay overeach individual wire, the configuration that maximises thetotal bandwidth over a number of parallel wires isinvestigated. It is shown from simulations that there is aunique optimal solution which does not necessarily translate tothe maximum possible number of wires, and in fact deviatesconsiderably from it when the resources available for repeatersare limited. Analytic guidelines dependent only on processparameters are derived for optimal sizing of wires andrepeaters.</p><p>Finally regular tiled architectures with a commoncommunication backplane are being proposed as being the mostefficient way to implement systems-on-chip in the deepsubmicron regime. This thesis also considers the feasibility ofimplementing a regular packet-switched network-on-chip in atypical future deep submicron technology. All major physicalissues and challenges are discussed for two differentarchitectures and important limitations are identified.</p>

Le déploiement des répéteurs iso-fréquence est une solution économique pour étendre la couverture d’un émetteur principal aux zones d’ombre. Cependant, ce mode de déploiement fait apparaître le phénomène des échos radio-fréquence entre antennes d’émission et de réception du répéteur. Selon les standards, un écho aussi faible soit-il réduit le débit de la liaison radio, tandis qu’un écho fort fait courir au répéteur le risque d’endommager ses circuits électroniques, ces risques sont dûs aux ondulations de puissance créées par les échos. L’objectif de cette thèse à caractère industriel est d’étudier ce phénomène naturel en considérant des signaux provenant de différents standards des télécommunications. Cette étude permet une caractérisation des échos radio-fréquence pour mieux s’orienter vers une solution optimisée et industriellement réalisable.Nous nous sommes orientés vers la solution du traitement du signal avancé en favorisant le filtrage adaptatif pour sa rapidité de convergence et sa simplicité relative d’implantation matérielle. Les circuits reconfigurables sont retenus pour leur prix et leur souplesse. L’implantation des solutions est effectuée en virgule fixe afin de satisfaire les exigences de réactivité. Durant la mise en oeuvre de la solution anti-écho, nous avons proposé une multitude de solutions numériques souples et fiables. À partir de ce constat, notre partenaire industriel a décidé de généraliser ce mode de traitement par le développement, la fabrication et la commercialisation de répéteurs de nouvelle génération entièrement numériques. / On-frequency repeaters are a cost-effective solution to extend coverage and enhance wireless communications, especially in shadow areas. However, coupling between the receiving antenna and the transmitting antenna, called radio frequency echo, increases modulation errors and creates oscillations in the system when the echo power is high. According to the communication standards, extremely weak echoes decrease the transmission rate, while strong echoes damage electroni ccircuits because of power peaks. This thesis aims at characterizing the echo phenomenon under different modulations, and proposing an optimized solution directly integrated to industry. We have turned to digital solutions especially the adaptive because of their high convergence rate and their simplicity to be implemented. The programmable circuits are chosen for their attractive price and their flexibility. When implementing echo cancellation solution, we proposed several reliable solutions, showing that digital processing is much more beneficial. For this reason, digital solutions are generalized, and the new generation of repeaters is fully digital.

Le développement rapide de l’industrie des systèmes sans fil suscite une demande urgente pour des communications à haut débit, notamment en environnement « indoor ». Toutefois, les protocoles traditionnels de communications sont incapables de supporter de très hauts débits et, surtout, il n'y a pas de ressources spectrales disponibles à basse fréquence. Comme une alternative, la bande des 60GHz est préconisée parce qu’elle permet un débit de plus de 5Gbit/s, grâce à son large spectre (57GHz à 64 GHz). Toutefois, en pratique, l’environnement « indoor » est complexe et, dans les situations d’absence de visibilité directe, la couverture radio est difficile à assurer à cause des fortes atténuations. Afin de remédier à ce problème, l’utilisation de répéteurs est possible. Dans cette thèse, l’intérêt est porté sur les répéteurs passifs, plus simples à installer et compatibles avec une réalisation faible coût. Le cas critique de la couverture radio d’un couloir en T est choisi comme fil conducteur, tout au long de cette étude. Les solutions préconisées visent aussi la compacité, ce qui justifie l’utilisation de réflecteurs plans. A partir de la théorie des réseaux d’antennes, une structure périodique générique pour le répéteur est proposée. Elle implique deux cellules réfléchissantes élémentaires (espacées de /2), produisant des ondes en opposition de phase. Plusieurs topologies et technologies sont ensuite envisagées et étudiées pour la mise en oeuvre. La première utilise des antennes à résonateur diélectrique (DRA). Deux topologies différentes sont étudiées et comparées, le DRA à encoche et le DRA couplé à une ligne déphaseuse en circuit ouvert. Différentes stratégies de modélisation sont également étudiées en utilisant une cellule unique, un couple de 2 cellules avec phases opposées ou un petit réseau. Pour chaque topologie, un réseau canonique de 6×6 éléments est simulé. La deuxième technologie étudiée utilise des guides d’ondes rectangulaires chargés par un matériau diélectrique et courtcircuités. Des analyses paramétriques sont effectuées et un 6 ×6 réseau est conçu et optimisé pour valider la faisabilité de la solution. Finalement, une structure en guide à plaques métalliques parallèles est analysée. Dérivant de la solution précédente, elle permet une fabrication plus simple. Pour cette dernière solution, une analyse plus complète est menée incluant des considérations sur la bande passante, les performances en incidence oblique et la direction de rayonnement maximal. Enfin, un réseau à base de cet élément rainuré est fabriquée présentant une taille de 200 mm×200 mm. Des mesures expérimentales à 60GHz sont réalisées pour tester les performances du réseau. La preuve de concept est ainsi donnée et les résultats expérimentaux sont analysés. / The fast development of information and consumer electronics industries creates a pressing demand for high-speed indoor communications. Traditional communication protocols are unable to support such high transmission rate, and there are no radio bands available at lower frequency. As an alternative, 60GHz communications have sparked great attention, since it enables a maximum data transmission rate more than 5Gbit/s based on its wide unlicensed bandwidth. However, practical indoor environment is usually complicated (e.g. walls, corridors, stairs, etc.), thus the non-line-of-sight (NLOS) areas cannot be covered due to the significant attenuation. Considering the requirements for signal recovery in the NLOS environment, a planar passive repeater with endfire radiation and high directivity, low profile and low cost is investigated and designed in this thesis. Based on array theory, a periodic structure for the foreseen printed repeater is derived out. It involves two reflector elements (in the spacing of 2) to provide 0° and 180°reflection phases respectively. In order to construct a proper reflector element, four topologies are proposed and analyzed. In the first part of the study, DRA (Dielectric Resonator Antenna) technology is the main focus. Two different topologies are investigated and compared, including notched DRA and DRA coupled to phase-shift stub. Different modeling strategies are also studied using either a single cell, a couple of cells with opposite phases or a small array. For each topology, a preliminary 6×6 array is simulated. Optimization at the array level is emphasized through the whole thesis. In the second part, investigations are taken on a rectangular dielectric filled waveguide element. Parametric analyses are carried out and possible fabrication technologies are discussed. Once again, a preliminary 6×6 array is designed and optimized to validate the feasibility of such a structure. In the third part, a parallel-plate groove structure is analyzed. It evolves from the rectangular waveguide, and enjoys more simplification. Further work is taken to explore the arrays’ bandwidth, oblique incidence performance and maximum radiation direction. Detailed theoretical analysis based on the simulation results are demonstrated in the end. Finally, an array based on the groove elements is fabricated in the size of 200mm×200mm. Practical measurements for 60GHz communications in NLOS environment are designed to test the array’s performance. Analyses on the experiment results are given.

Radio repeaters in communication systems are used to extend the coverage where the base stations cannot reach directly, such as indoors, in tunnels, and obstructed mobile reception areas. The repeaters in the downlink receive signals from base stations, amplify them, and then retransmit them. In case of this, the receiving antenna may receive the signals from the transmitting antenna of the repeater, and therefore introduce oscillation. To minimize the oscillation and keep the system stable, there should be a suitable isolation between transmitting and receiving antennas. Isolation margin estimation for wide-band signals can be achieved by applying auto-correlation. However, there are still important applications in narrow-band technologies such as Global Systems for Mobile Communication (GSM). Although the GSM is not the latest technology in communication systems and has been largely surpassed by 3G, 4G, and 5G in terms of data capabilities, it is still widely used in many parts of the world such as commercial voice communication in rural areas, safety-critical transmissions in railways and so on. The auto-correlation functions are hard to use since the auto-correlation response is generally wider than that of wide-band signals. In this project, attempts are made to find more accurate estimation methods. Channel estimation method is chosen as the central focus of this project. This project simulated methods such as frequency correction and channel estimation and evaluated the performance of the algorithms. The simulation results indicate that the channel estimation algorithm performs well when the signal-to-noise ratio is 80dB and the feedback signal power is 55dB lower than the wanted signal. It also provides a result that the estimation results are accurate when the power of the feedback signal is approximately 20dB higher than the noise signal. / Radio repeatrar i kommunikationssystem används för att utöka täckningen där basstationer inte kan nå direkt, såsom inomhus, i tunnlar och i områden med blockerad mobil mottagning. Repeatrarna i nedlänken tar emot signaler från basstationer, förstärker dem och sänder sedan ut dem igen. I detta fall kan den mottagande antennen ta emot signalerna från repeatrarnas sändande antenner och därmed introducera oscillation. För att minimera oscillationen och hålla systemet stabilt bör det finnas en lämplig isolation mellan sändande och mottagande antenner. Isoleringsmarginaluppskattning för bredbandsignaler kan uppnås genom att tillämpa autokorrelation. Det finns dock fortfarande viktiga tillämpningar inom smalbands teknologier som Global System for Mobile Communication (GSM). Även om GSM inte är den senaste tekniken inom kommunikationssystem och har i stor utsträckning överskridits av 3G, 4G och 5G när det gäller datamöjligheter, används det fortfarande i stor utsträckning i många delar av världen, såsom kommersiell röstkommunikation på landsbygden, säkerhetskritiska överföringar inom järnvägar och så vidare. Autokorrelationsfunktionerna är svåra att använda eftersom autokorrelationsresponsen generellt är bredare än för bredbandsignaler. I det här projektet görs försök att hitta mer exakta uppskattningmetoder. Kanaluppskattningsmetoden väljs som den centrala fokusen för detta projekt. Detta projekt simulerade metoder som frekvenskorrigering och kanaluppskattning och utvärderade prestandan hos algoritmerna. Simuleringsresultaten indikerar att kanaluppskattningsalgoritmen fungerar bra när signal-brusförhållandet är 80dB och feedbacksignalens effekt är så låg som 55dB. Den ger också ett resultat som visar att uppskattningsresultaten är korrekta när feedbacksignalens effekt är ungefär 20dB högre än brussignalens.

Switch Optimization in a Converged Network thesis discusses the role of the Ethernet switch in the network. It describes differences between the switch and the hub in the network and the impacts on the collision domain size. In the converged network, priority data switching according to their origin and thus the qualitative point of view is also addressed - e-mails, voice services and multimedia will each behave differently. A real element, switch L3 - a switch with services quality support - is mentioned. The tool Matlab - Simulink is briefly described and a functional simplified model of the ethernet switch was created. Furthermore, the topic was used as a theme for a laboratory task in the subject Services of Telecommunication Networks. A model for Simulink and an example of the laboratory protocol for the created task are attached to the thesis.

The deterministic generation and manipulation of quantum states has attracted much interest ever since the rise of quantum mechanics. Large-scale, distributed quantum states are the basis for novel applications such as quantum communication, quantum remote sensing, distributed quantum computing or quantum voting protocols. The necessary infrastructure will be provided by distributed quantum networks, allowing for quantum bit processing and storage at single nodes. Quantum states of light then allow for inter-node transmission of quantum information. Transmission losses in optical fibers may be overcome by quantum repeaters, the quantum equivalent of classical signal amplifiers. The fragility of quantum superposition states makes building such networks very challenging. Hybrid solutions combine the strengths of different physical systems: Efficient quantum memories can be realized using alkali atoms such as rubidium. Leading in the deterministic generation of single photons and polarization entangled photon pairs are semiconductor InAs/GaAs quantum dots grown by the Stranski-Krastanov method. Despite remarkable progress in the last twenty years, complex quantum optical protocols could not be realized due to low degree of entanglement, low brightness and broad wavelength distribution. In this work, an emerging family of epitaxially grown GaAs/AlGaAs quantum dots obtained by droplet etching and nanohole infilling is studied. Under pulsed resonant two-photon excitation, they emit single pairs of entangled photons with high purity and unprecedented degree of entanglement. Entanglement fidelities up to f = 0.94 are observed, which are only limited by the optical setup or a residual exciton fine structure. The samples exhibit a very narrow wavelength distribution at rubidium memory transitions. Strain tuning is applied via piezoelectric actuators to allow for reversible fine-tuning of the emission frequency. In a next step, active feedback is employed to stabilize the frequency of single photons emitted by two separate quantum dots to an atomic rubidium standard. The transmission of a rubidium-based Faraday filter serves as the error signal for frequency stabilization. A residual frequency deviation of < 30MHz is achieved, which is less than 1.5% of the quantum dot linewidth. Long-term stability is demonstrated by Hong-Ou-Mandel interference between photons from the two quantum dots. Their internal dephasing limits the expected visibility to V = 40%. For frequency-stabilized dots, V = (41 ± 5)% is observed as opposed to V = (31 ± 7)% for free-running emission. This technique reaches the maximally expected visibility for the given system and therefore facilitates quantum networks with indistinguishable photons from distributed sources. Based on the presented techniques and improved emission quality, pivotal quantum communication protocols can now be implemented with quantum dots, such as transferring entanglement between photon pairs. Embedding quantum dots in a dielectric antenna ensures a bright emission. For the first time, entanglement swapping between two pairs of photons emitted by a single quantum dot is realized. A joint Bell measurement heralds the successful generation of the Bell state Ψ+ with a fidelity of up to (0.81 ± 0.04). The state's nonlocal nature is confirmed by violating the CHSH-Bell inequality with S = (2.28 ± 0.13). The photon source is tuned into resonance with rubidium transitions, facilitating implementation of hybrid quantum repeaters. This work thus represents a major step forward for the application of semiconductor based entangled photon sources in real-world scenarios.

In this work, the interference sensitivity of an FMCW radar has been evaluated by studying the impact on a simulated detection chain. A commercially available FMCW radar was first characterized and its properties then laid the foundation for a simulation model implemented in Matlab. Different interference methods have been studied and a selection was made based on the results of previous research. One method aims to inject a sufficiently large amount of energy in the form of pulsed noise into the receiver. The second method aims to deceive the radar into seeing targets that do not actually exist by repeating the transmitted signal and thus giving the radar a false picture of its surroundings. The results show that if it is possible to synchronize with the transmitted signal then repeater jamming can be effective in misleading the radar. In one scenario the false target even succeeded in hiding the real target by exploiting the Cell-Averaging CFAR detection algorithm. The results suggests that without some smart countermeasures the radar has no way of distinguishing a coherent repeater signal, but just how successful the repeater is in creating a deceptive environment is highly dependent on the detection algorithm used. Pulsed noise also managed to disrupt the radar and with a sufficiently high pulse repetition frequency the detector could not find any targets despite a simulated object in front of the radar. On the other hand, a rather significant effective radiated power level was required for the pulse train to achieve any meaningful effect on the radar, which may be due to an undersampled signal in the simulation. It is therefore difficult based on this work to draw any conclusions about how suitable pulsed noise is in a non-simulated interference context and what parameter values to use.