In this thesis, a second-order low pass square root domain filter (SRD filter) based on operational transconductors amplifiers (OTAs) is presented. The SRD filter consists of two translinear filters and four OTAs. Because the OTA has small voltage swings, which may violate the large signal natural of the SRD filter. We investigate the dynamic range of this compact SRD filter with different quality factor(Q). The circuit has fewer numbers of transistors and operate in low voltage, therefore, it has less power consumption and less chip area. The circuit has been fabricated with 0.35£gm CMOS technology. It operates with a supply voltage 1.5V and the biasing current varies from 10uA to 80uA. Measurement results lts show that Im/I0≥40% when the external capacitance C is 3.5pF¡B7pF and Im/I0≥53% when the external capacitance C is 3pF¡B8.5pF. The cutoff frequency of the filter can be tuned from 1.24MHz to 5.53MHz when the external capacitance C is 3.5pF¡B7pF and the cutoff frequency can be tuned from 900KHz to 4.46MHz when the external capacitance C is 3pF¡B8.5pF. The total harmonic distortion is 0.908% and the power consumption is 506£gW.

With the rapid development of information technology, more and more bandwidth is required to transmit multimedia data. Since local communication networks are moving to wireless domain, it brings up great challenges for making integrated wideband wireless front-ends suitable for these applications. RF filtering is a fundamental need in all wireless front-ends and is one of the most difficult parts to be integrated. This has been a major obstacle to the implementation of low power and low cost integrated wireless terminals. <p> Lots of previous work has been done to make integrated RF filters applicable to these applications. However, some of these filters are not designed with standard CMOS technology. Some of them are not designed in desired frequency bands and others do not have sufficient frequency bandwidth. This research demonstrates the design of a tunable wideband RF filter that operates at 3.6 GHz and can be easily changed to a higher frequency range up to 5 GHz. This filter is superior to the previous designs in the following aspects: a) wider bandwidth, b) easier to tune, c) balancing in noise and linearity, and d) using standard CMOS technology. The design employs the state-of-the-art inductor degenerated LNA, acting as a transconductor to minimize the overall noise figure. A Q-enhancement circuit is employed to compensate the loss from lossy on-chip spiral inductors. Center frequency and bandwidth tuning circuits are also embedded to make the filter suitable for multi band operations. <p> At first, a second order bandpass filter prototype was designed in the standard 0.18 ìm CMOS process. Simulation results showed that at 3.6 GHz center frequency and with a 60-MHz bandwidth, the input third-order intermodulation product (IIP3) and input-referred 1 dB compression point (P1dB) was -22.5 dBm and -30.5 dBm respectively. The image rejection at 500 MHz away from the center frequency was 32 dB (250 MHz intermediate frequency). The Q of the filter was tunable over 3000 and the center frequency tuning range was about 150 MHz. <p> By cascading three stages of second order filters, a sixth order filter was designed to enhance the image rejection ability and to extend the filter bandwidth. The sixth order filter had been fabricated in the standard 0.18 ìm CMOS process using 1.8-V supply. The chip occupies only 0.9 mm 0.9 mm silicon area and has a power consumption of 130-mW. The measured center frequency was tunable from 3.54 GHz to 3.88 GHz, bandwidth was tunable from 35 MHz to 80 MHz. With a 65 MHz bandwidth, the filter had a gain of 13 dB, an IIP3 of -29 dBm and a P1dB of -46 dBm.

Recent government regulations for Enhanced 911 locating of wireless handsets require accuracy to within 50 and 300 meters. Two technologies under consideration are triangulation using existing wireless base stations and location using global positioning satellites (GPS). Satellite positioning is the leading candidate, however, reception of GPS signals within large buildings is difficult and considerable research is devoted to this topic. Conventional GPS receivers require line of sight to at least four satellites and, under outdoor conditions, the expected signal level is about -160 dBW. Within large buildings, detection is very difficult because there is high thermal noise and some satellite signals can be attenuated to less than -185 dBW while others can suffer little attenuation. In order to construct the pseudo-ranges necessary for position finding, the receiver must synchronize to the incoming codephase of each satellite and must operate with substantial Doppler frequency offset caused by satellite motion.<p> This thesis investigates the application of a parallel non-coherent spread spectrum synchronizer previously implemented as a very-large-scale integration (VLSI) circuit. The circuit processes one millisecond of incoming signal and uses a segmented matched filter (SMF) by which the segmentation provides some tolerance to Doppler shift. The thesis presents simulation results of averaging for tens of seconds. Through simulation, the SMF is compared with a transversal matched filter (TMF) under conditions of no Doppler shift; coherent and non-coherent integration are discussed. The simulation is conducted at 290 K (17°C) such that the Boltzmann noise is -204 dBW/Hz, with a GPS signal bandwidth of 2 MHz and signal level of -185 dBW, and the receiver input signal-to-noise ratio (SNR) is -44 dB.<p> The SMF is applied using differing segment lengths to high-sensitivity GPS data from indoor and urban simulated GPS data. The results demonstrate the SMFs ability to tolerate Doppler frequency offsets while allowing for long integration times to detect the weak GPS signals.

Many parallelization systems detect memory access conflicts across concurrent threads by disambiguating address-sets using bit-vector-based Bloom filters, which are efficient, but can report false conflicts that do not exist. Systems with lazy conflict detection often use Bloom filters unconventionally by testing sets for null-intersection via Bloom filter intersection, contrasting with the conventional approach of issuing membership queries into the Bloom filter. In this dissertation we develop much-needed theory for probability of false conflicts in Bloom filter null-intersection tests, notably demonstrating that Bloom filter intersection requires substantially larger bit-vectors to provide equivalent statistical behavior to querying. Furthermore, we recognize that our theoretical implications counter practical intuition, and thus use RingSTM to evaluate theory in practice by implementing and comparing the Bloom filter configurations. We find that despite its overheads, the queue-of-queries approach reduces execution time and is thus the most compelling alternative to Bloom filter intersection for lazy address-set disambiguation.

Many parallelization systems detect memory access conflicts across concurrent threads by disambiguating address-sets using bit-vector-based Bloom filters, which are efficient, but can report false conflicts that do not exist. Systems with lazy conflict detection often use Bloom filters unconventionally by testing sets for null-intersection via Bloom filter intersection, contrasting with the conventional approach of issuing membership queries into the Bloom filter. In this dissertation we develop much-needed theory for probability of false conflicts in Bloom filter null-intersection tests, notably demonstrating that Bloom filter intersection requires substantially larger bit-vectors to provide equivalent statistical behavior to querying. Furthermore, we recognize that our theoretical implications counter practical intuition, and thus use RingSTM to evaluate theory in practice by implementing and comparing the Bloom filter configurations. We find that despite its overheads, the queue-of-queries approach reduces execution time and is thus the most compelling alternative to Bloom filter intersection for lazy address-set disambiguation.

High performance microwave cavities for various circuits in the front-end of transceivers such as filters, diplexers, and oscillators have conventionally been built with rectangular or cylindrical metallic waveguides, which typically have low loss, high quality (Q) factor, and higher power handling capability. However such waveguide cavity based circuits made by traditional metal machining techniques tend to be costly, particularly for complex multiple cavity based circuits, and not well suited to high volume commercial applications and integration with planar microwave integrated circuits. As commercial transceiver applications progress toward higher microwave and millimetre-wave frequencies, the use of waveguide based circuits for compact, highly integrated transceivers is becoming feasible, along with an increasing need for cost effective batch fabrication processes for realizing complex metallic cavity circuits without sacrificing structural quality and performance. It is expected that significant advancements in both microwave performance and integration will be achieved through the development of novel technologies for realizing vertically oriented three-dimensional (3-D) structures.<p>Although improvement has been made on increasing the resonator Q factor by exploiting silicon micromachining and low-temperature cofired ceramics (LTCC) techniques, there are some drawbacks inherent to silicon cavity micromachining and LTCC technology, including non-vertical sidewalls, depth limitations, and surface roughness for the silicon resonator, and dielectric and radiation loss for LTCC resonator.<p>Polymer-based fabrication is a promising alternative to silicon etching and LTCC technologies for the batch fabrication of ultra-deep microwave cavity structures. In particular, deep X-ray lithography (XRL), as part of the LIGA process, is a microfabrication technology for precisely structuring polymers, and is increasingly being applied to RF/microwave microstructures. In addition to precise patterning capabilities, deep XRL is able to structure ultra-deep cavities due to the penetration ability of hard X-rays. Cavities of several millimetres are possible in a single lithographic exposure, and with excellent sidewall quality, including verticality near 90 degrees and surface roughness on the order of tens of nanometres. These structured polymers are subsequently used as electroforming templates for fabricating metal structures with correspondingly good sidewall quality.<p>This thesis investigates the possibility of realizing high-Q cavity resonators and filters at microwave frequencies using the LIGA microfabrication process. Finite element method (FEM) electromagnetic simulation results based on the cavity models representing different fabrication conditions show that smooth LIGA cavity structures result in promising Q improvement over silicon and LTCC structures. And the potential advantages of LIGA resonators are more dramatic with cavity height and increasing operating frequency. Deep polymer cavity structures (1.8 mm) fabricated using deep XRL demonstrate excellent sidewall verticality in the PMMA structure, with only slight shrinkage at the top surface of 8.5 2.5 mm in either lateral dimensions. This corresponds to sidewalls with verticality between 89.82o and 89.9o. The structure polymers are subsequently used as templates for metal electroforming to produce cavity resonators. The performance of the resonator is measured in a planar environment. A RT/duroid6010 soft substrate patterned with coupling structures forms the sixth side, and thus completes the cavity. Despite the rather crude test assembly for the sixth side made by clamping, the measured resonator has a high unloaded Q of 2122.2 85 at the resonant frequency of 24 GHz, indicating that LIGA cavities are especially promising for high performance applications. <p>The relatively simple, single-step lithographic exposure also facilitates extension to more structurally complicated waveguide and multiple cavity-based circuits. This research work also proposes a high performance ``split-post' 3-pole cylindrical post coupled Chebyshev bandpass filter suitable for LIGA fabrication. In addition to potentially batch fabricating such a filter lithographically by exposing the entire waveguide depth in a single exposure, the filter structures composed of three cavities with metallic multi-post coupling would be extremely difficult to fabricate using traditional machining techniques, due to the extremely fine post structure and high vertical aspect ratio required. However, these types of structures could be ideal for LIGA fabrication, which offers sub-micron features, aspect ratios of 100:1 or higher, resist thicknesses of up to 3 mm, and almost vertical and optically smooth sidewalls. Also, representative LIGA sidewall roughness is used to predict very low loss and high performance, suggesting that complicated structures with multiple resonator circuits and high internal components with high aspect ratios are possible.

With the rapid development of information technology, more and more bandwidth is required to transmit multimedia data. Since local communication networks are moving to wireless domain, it brings up great challenges for making integrated wideband wireless front-ends suitable for these applications. RF filtering is a fundamental need in all wireless front-ends and is one of the most difficult parts to be integrated. This has been a major obstacle to the implementation of low power and low cost integrated wireless terminals. <p> Lots of previous work has been done to make integrated RF filters applicable to these applications. However, some of these filters are not designed with standard CMOS technology. Some of them are not designed in desired frequency bands and others do not have sufficient frequency bandwidth. This research demonstrates the design of a tunable wideband RF filter that operates at 3.6 GHz and can be easily changed to a higher frequency range up to 5 GHz. This filter is superior to the previous designs in the following aspects: a) wider bandwidth, b) easier to tune, c) balancing in noise and linearity, and d) using standard CMOS technology. The design employs the state-of-the-art inductor degenerated LNA, acting as a transconductor to minimize the overall noise figure. A Q-enhancement circuit is employed to compensate the loss from lossy on-chip spiral inductors. Center frequency and bandwidth tuning circuits are also embedded to make the filter suitable for multi band operations. <p> At first, a second order bandpass filter prototype was designed in the standard 0.18 ìm CMOS process. Simulation results showed that at 3.6 GHz center frequency and with a 60-MHz bandwidth, the input third-order intermodulation product (IIP3) and input-referred 1 dB compression point (P1dB) was -22.5 dBm and -30.5 dBm respectively. The image rejection at 500 MHz away from the center frequency was 32 dB (250 MHz intermediate frequency). The Q of the filter was tunable over 3000 and the center frequency tuning range was about 150 MHz. <p> By cascading three stages of second order filters, a sixth order filter was designed to enhance the image rejection ability and to extend the filter bandwidth. The sixth order filter had been fabricated in the standard 0.18 ìm CMOS process using 1.8-V supply. The chip occupies only 0.9 mm 0.9 mm silicon area and has a power consumption of 130-mW. The measured center frequency was tunable from 3.54 GHz to 3.88 GHz, bandwidth was tunable from 35 MHz to 80 MHz. With a 65 MHz bandwidth, the filter had a gain of 13 dB, an IIP3 of -29 dBm and a P1dB of -46 dBm.

El crecimiento exponencial que han experimentado los sistemas de comunicaciones inalámbricos actuales durante los últimos años, ha dado lugar a la necesidad de disponer dispositivos de microondas de tamaño reducido y, a su vez, de altas prestaciones. El principal obstáculo con el que se encuentran los dispositivos basados en las tecnologías clásicas, es que dichas tecnologías no son compatibles con los procesos estándar de fabricación de circuitos integrados (IC). En este sentido, dispositivos de microondas basados en resonadores acústicos, en concreto resonadores BAW, ofrecen una solución a dicha limitación ya que son compatibles con los procesos estándar de fabricación de circuitos integrados. Por otro lado, dichos resonadores son excitados mediante una onda acústica cuya velocidad de propagación es alrededor de cuatro o cinco órdenes de magnitud menor que las ondas electromagnéticas, con lo que el tamaño del dispositivo también será menor en la misma proporción. La motivación de este trabajo viene dada por el escenario que se plantea. El documento está básicamente dividido en dos grandes bloques. El primer bloque que consta de los capítulos 2, 3 y 4 está dedicado al estudio de un resonador BAW. En primer lugar, el estudio se realiza para una estructura unidimensional, es decir, teniendo en cuenta que en el resonador solamente se propaga una onda acústica en la dirección longitudinal, con el objetivo de extraer los modelos circuitales que permiten la caracterización eléctrica de dicho resonador. Sin embargo, estos modelos no son capaces de predecir con exactitud el comportamiento eléctrico del resonador acústico ya que no contemplan efectos derivados de la propagación de ondas laterales. Así, la necesidad de incorporar una herramienta de simulación 3D se hace necesaria. Mediante el simulador 3D, el resonador se puede caracterizar por completo, así, el origen y las condiciones necesarias para la presencia de ondas laterales estacionarias pueden ser establecidas. La presencia de las ondas laterales generan efectos no deseados en la respuesta eléctrica del resonador básicamente en forma de degradación del factor de calidad y de la constante de acoplo electromecánico efectiva. Así pues, la mejora del comportamiento eléctrico del resonador se basa principalmente, en la minimización de la presencia de las ondas laterales no deseadas. Para ello se proponen dos soluciones diferentes: el apodizado y la inclusión de un anillo en el perímetro del electrodo superior. La primera solución consiste en diseñar el electrodo superior de manera que no existan caras paralelas. Con ello, los patrones de resonancia que se generan hacen que la onda deba viajar una distancia mayor con la correspondiente atenuación debida a las pérdidas del material. Sin embargo, aunque la amplitud de dichas ondas se ve claramente decrementada, el número de patrones resonantes aumenta respecto al caso convencional, y también el número de modos no deseados. Esto finalmente se traduce en una degradación del factor de calidad del resonador. La segunda solución consiste en forzar unas condiciones de contorno determinadas que hacen que las ondas laterales no puedan propagarse. Con esto, la respuesta del resonador aparece libre de modos no deseados con lo que el factor de calidad del resonador es mucho mayor que lo que se puede conseguir mediante el apodizado. El segundo gran bloque de este trabajo está dedicado a la aplicación de los resonadores acústicos al diseño de filtros. Este bloque se ha centrado básicamente en topologías en las que los resonadores acústicos están conectados eléctricamente, y en particular en topologías tipo"ladder". Este tipo de filtros presenta una selectividad alta debido a la presencia de un par de ceros de transmisión, pero a su vez un pobre rechazo fuera de banda debido al comportamiento capacitivo de dichos resonadores. Así, en primer lugar se propone una metodología de diseño con expresiones cerradas, en un primer lugar considerando negligible la presencia de los electrodos metálicos, y realizando alguna modificación en el proceso de diseño para incluir dichos efectos. Por otro lado, con el objetivo de mejorar las prestaciones del filtro fuera de banda, se propone como solución modificar los resonadores acústicos mediante la presencia de elementos reactivos (capacidades y bobinas) ya sea en serie o en paralelo. La modificación de dichos resonadores se da principalmente en la posición de las frecuencias de resonancia. Así, si en una topología ladder se incluyen resonadores modificados y no modificados, se genera un nuevo par de ceros de transmisión en la respuesta que hacen que el rechazo fuera de banda sea mayor. El último capítulo de este trabajo está dedicado al diseño de un filtro con una respuesta en transmisión que presenta dos bandas de paso. La topología está basada en la topología ladder clásica con la diferencia que, en lugar de tener una celda elemental formada por un resonador serie y paralelo, se tiene una celda elemental formada por dos resonadores serie y dos paralelos. Con esto, una banda de transmisión viene dada por la interacción de un resonador serie y uno paralelo, mientras que la segunda banda se da por la interacción del segundo par de resonadores. Con el método de diseño que se propone, cada una de las bandas puede ser diseñada a la frecuencia que se desee. Finalmente, se muestra el diseño de un filtro dual para la aplicación de GPS L1/L5 y Galileo E5a/E5b. / The exponential growth in wireless communication systems in recent years has been due to the requirements of small high performance microwave devices. The main limitation of microwave devices based on traditional technologies is that these technologies are not compatible with the manufacturing process of standard integrated circuits (IC). Microwave devices based on acoustic resonators, and bulk acoustic wave (BAW) resonators in particular, overcome this limitation since this technology is compatible with standard IC technologies. Furthermore, acoustic resonators are excited by means of an acoustic wave with a propagation velocity around four or five times lower than the propagation velocity of electromagnetic waves, and the resulting size of the device is therefore also lower in the same proportion. This is the incentive behind this study. The study is basically divided into two blocks. The first block comprises the chapters number 2, 3 and 4, and is devoted to the study of the BAW resonator. First, the BAW resonator is studied in its one-dimensional form, i.e. considering that only a mechanical wave is propagated in the thickness dimension, in order to obtain the equivalent circuits which enable the electrical characterization of the BAW resonator. However, these models do not take into account effects due to lateral waves. The need for a 3D simulator tool therefore becomes evident. Using the 3D simulator, the electrical behaviour of the BAW resonator can be completely characterized, and the boundary conditions required as well as the origin of lateral standing waves can therefore be stated. The presence of lateral standing waves entails the degradation of the electrical performance of the BAW resonator, mainly in terms of the quality factor and the effective electromechanical coupling constant. The improvement in the electrical performance of the BAW resonator is therefore mainly based on its ability to minimize the presence of unwanted lateral modes. Two different solutions are proposed to that end: apodization and the presence of a perimetric ring on the top of the metal electrode. The former solution consists of designing the top electrode in such a way that non-parallel edges are found. The resonant paths thereby become larger and the resonant modes thus become more attenuated due to the material losses. However, although the strength of these modes is lower, more resonant modes are present since there are more possible resonant paths. This finally leads to the degradation of the quality factor of the BAW resonator. The latter solution consists of including a thickened edge load on the top of the metal electrode, which leads to boundary conditions in which lateral waves cannot propagate through the structure. By doing this, the electrical response of the BAW resonator is spurious-free and the quality factor obtained is thus higher than if the apodization solution is used. The second block of this thesis is devoted to the application of BAW resonator to the microwave filter design, and particularly to filters based on electrically connected BAW resonators, as in the case of ladder-type filters. This type of filter presents a very high selectivity due to the presence of a pair of transmission zeros, but a poor out-of-band rejection due to the natural capacitor divider. The design procedure using closed-form expressions is therefore presented first, with the effect of the metal electrodes considered negligible, and this procedure subsequently modified in order to include these effects. In order to improve the filter performance out-of-band, the proposed solution consists of modifying the BAW resonators with the presence of reactive elements (capacitors and inductances) in series or a shunt configuration. By doing so, the modification is directly related with the allocation of the resonant frequencies. The modified and non-modified BAW resonators are therefore connected in a ladder-type topology including a new pair of transmission zeros, making the out-of-band rejection higher. The final chapter of this study focuses on the design of a dual-band filter based on BAW resonators. The proposed topology is based on the conventional ladder-type topology but instead of having an elemental cell comprising a single BAW resonator in series and a shunt configuration, two series and two shunt BAW resonators are now present. By doing this, one of the transmission bands is related with one series and shunt BAW resonator and the second transmission band is due to the other pair of resonators. With the proposed design procedure, each of the transmission bands can be allocated to the desired frequencies. Finally, the design procedure is applied to the GPS L1/L5 and Galileo E5a/E5b applications.

The first objective of this research has always been carry out a study of visual techniques SLAM (Simultaneous localization and mapping), specifically the type monovisual, less studied than the stereo. These techniques have been well studied in the world of robotics. These techniques are focused on reconstruct a map of the robot enviroment while maintaining its position information in that map. We chose to investigate a method to encode the points by the inverse of its depth, from the first time that the feature was observed. This method permits efficient and accurate representation of uncertainty during undelayed initialization and beyond, all within the standard extended Kalman filter (EKF).At first, the study mentioned it should be consolidated developing an application that implements this method. After suffering various difficulties, it was decided to make use of a platform developed by the same author of Slam method mentioned in MATLAB. Until then it had developed the tasks of calibration, feature extraction and matching. From that point, that application was adapted to the characteristics of our camera and our video to work. We recorded a video with our camera following a known trajectory to check the calculated path shown in the application. Corroborating works and studying the limitations and advantages of this method.

The increased market for navigation, localization and mapping system has encouraged the research to dig deeper into these new and challenging areas. The remarkable development of computer soft- and hardware have also opened up many new doors. Things which more or less where impossible ten years ago are now reality. The possibilities of using a mathematical approach to compensate for the need of expensive sensors has been one of the main objectives in this thesis. Here you will find the basic principles of localization of indoor UAVs using particle filter (PF) and Octomaps, but also the procedures of implementing 2D scanmatching algorithms and quaternions. The performance of the algorithms is evaluated using a high precision motion capture system. The UAV which forms the basis for this thesis is equipped with a 2D laser and an inertial measurement unit (IMU). The results show that it is possible to perform localization in 2D with centimetre precision only by using information from a laser and a predefined Octomap.