The aim of this work is to implement digital filters into programmable gate array. The work also includes a description of the MEMS technology, including comparisons with the technology of MEMS microphones from various manufacturers. Another part is devoted to the Sigma-delta modulation. The main section is the design and implementation of digital CIC and FIR filters for signal processing of digital microphone, including simulation and verification of properties of the proposed filter in Matlab.

Mémoire numérisé par la Direction des bibliothèques de l'Université de Montréal.

This Master’s thesis deals with digital decimation filter design for undersampling and filtering of sigma-delta ADC signal. Filter cascade is designed in Matlab according to given requirements and is then described in VHDL language aiming for minimum area. Implemented filter functionality is compared to Matlab-generated reference filters in created verification environment. Finally the design is synthesized in specified technology and verified on gate level.

The doctoral thesis deals with development and design of novel architectures of digital filters for low-power integrated circuits. The main goal was to achieve optimum parameters of digital filters with respect to the chip area, power consumption and operating frequency. The target group of the proposed architectures are application specific integrated circuits designed for signal processing from sensors using delta-sigma modulators. Three novel architectures of digital filters optimized for low-power integrated circuits are presented in the thesis. The thesis provides analysis and comparison of parameters of the new filter architectures with the parameters of architectures generated by Matlab tool. A software tool has been designed and developed for the practical application of the proposed architectures of digital filters. The developed software tool allows generating hardware description of the filters with respect to defined parameters.

The thesis aims to implement different digital filters such as finite impulse response (FIR), infinite impulse response (IIR) and cascade integrator comb (CIC) on the field-programmable gate array (FPGA) development board using hardware description language (VHDL). To this purpose, Intel’s systems integration tool Platform designer is used to convert the implementation to an IP core. The implemented FIR and IIR filters include different filter types such as lowpass, highpass, bandpass and bandstop. All the filters have a pipeline architecture as well as adjustable parameters such as filter order, frequency specifications and resolution. The coefficients of the filters are calculated according to the user's specifications. The calculated coefficients are verified using simulation. Furthermore the IP has been validated on hardware by the FPGA board MAX DE-10 lite. The IP is also analyzed regarding timing and power consumtion with good results. FIR filters of different types have been implemented and tested up to 501 taps with a coefficient width of 24 bits, which covered just below 50% of the available logic gates on the MAX 10-DE lite board with 50000 gates in total. The FIR filters have an option to be used with a Kaiser window with a maximum tap level of 51. Different IIR filters have also been implemented and tested on the hardware. However, the results have shown that the IIR filters do not perform so well, especially those of order higher than 6. One of the main reasons for this is the overflow caused by instability of the IIR.

(in English): This thesis deals with post-war fate of Josef Robotka the member of the second and third resistence and people who cooperated with him. It cannot be mentioned his activities during the Second World War in the Council of three but the key issue will be the engagement after 1945. I would describe his disenchantment with communist ideas while on Vorošilov military academy in Moscow, subsequently forced to return to Czechoslovakia and post-February intelligence cooperation with CIC. I will also try to bring the role of public authorities in the arrest, detention and trial. In the next section I would like to map the entire arrest of Robotka group, their imprisonment and trial. The following chapters will deal with Robotka's execution and the role of wives and close relatives of convicted persons. The work is based mainly on archival research, supplemented by memories of Vlasta Nováčková Jakubová who is Robotka's niece and close collaborator. Main question is how big the importance of the information that the group sent abroad was and what significance should they have in case of possible conflict.

Transcranial routing of signal (TCROS) was accomplished using completely-in-the-canal (CIC) hearing aids in 5 profoundly unilaterally hearing-impaired individuals. The functional gain realized by the participants far exceeded the gain predicted by measuring the acoustic output and real ear aided response of the hearing aids. The difference between predicted and functional gain increased with signal frequency and was attributed at least in part to mechanical vibration of the hearing aid in the external canal. Implications for fitting unilateral hearing loss using TCROS amplifying systems are discussed.

Bone conduction transmission and head‐shadow effects were determined with transcranial completely‐in‐the‐canal (TCCIC) CROS hearing aids. Five subjects with documented profound unilateral hearing loss and experience with traditional CROS/BICROS fittings (TCROS) were tested with a CIC hearing aid placed in their poorer ear. Peak SPL was measured at the tympanic membrane and ranged from 105–115 dB SPL at 2000 Hz. Pure‐tone crossover thresholds and functional gain tested at frequencies from 250–8000 Hz varied considerably more than the SPL measures. The pure‐tone results indicated that sensitivity in the better ear was moderately associated with functional gain across frequency. Speech recognition was then tested in the sound field in two conditions: direct (noise in the poorer ear, speech in the better ear) and indirect (noise in the better ear, speech in the poorer ear) at S/Ns of −6, 0, +6, +12, and quiet. The TCCIC fittings were more effective than TCROS aids across S/Ns, particularly in the direct condition. In the indirect condition, the two fittings performed similarly. When data were pooled across conditions, the TCCIC aids provided better word recognition than the TCROS aids, particularly for those subjects with greater sensitivity in the better ear.

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / Glyphosate [N-(phosphonomethyl) glycine] is the most widely used pesticide in the world for controlling weeds in agriculture. High amounts of this herbicide and its major metabolite, aminomethylphosphonic acid (AMPA), have been found in water due to intensive use. The Ministry of Health 2914 establishes a maximum value allowed for the sum of glyphosate and AMPA at 0.5 mg L-1. The most comum method to determine these compounds requires a derivatization step for fluorescence detection. Thus, this study aimed to develop two simple and rapid methods for the direct determination of glyphosate and AMPA in water without treatment or sample derivatization steps, by capillary ion chromatography (Capillary IC) and liquid chromatography tandem mass spectrometry (LC-MS/MS). The Capillary IC method employed a capillary IonPac AS19 column (250 x 0.4 mm, 7.5 μm), a gradient elution with KOH a conductivity detector. The LC-MS/MS method used a Pursuit XRS C18 column (150 x 2 mm ID, 5 μm), an aqueous solution of ammonium hydroxide 0.01% (v/v) as mobile phase and a triple-quadrupole MS detector, operating in a selected reaction monitoring (SRM) mode. The method validation was performed by recovery experiments, evaluating accuracy, precision, linearity, limits of detection (LOD) and of quantification (LOQ). The analytical curves for glyphosate and AMPA obtained for the concentration levels 25, 50, 100, 250 and 500 μg L-1 by Capillary IC and LC-MS/MS showed r2 >0.99 for both compounds. For spiked levels of 25, 50, 250 and 500 μg L-1 recoveries were between 70 -113%, with RSD <10% for Capillary IC and 79-105%, with RSD <14% for LC-MS/MS. The method limits of detection for both glyphosate and AMPA was 8 μg L-1 and of quantification was 25 μg L-1, in both chromatographic techniques. Both methods showed good selectivity and sensitivity. Determination by direct injection of the sample, using capillary IC or LC-MS/MS proved to be efficient, simple and cost-effective tools for analysis, making it possible to monitor at levels below the maximum allowed limits for drinking water. / O glifosato [N-(fosfonometil)glicina] é o agrotóxico mais utilizado na agricultura mundial no controle de plantas daninhas. Altas quantidades desse herbicida e seu principal metabólito, o ácido aminometilfosfônico (AMPA), vem sendo encontrados em água devido ao uso intensivo. A portaria do Ministério da Saúde nº 2914 estabelece o valor máximo permitido para a soma de glifosato e AMPA em 0,5 mg L-1. O método amplamente utilizado para a determinação desses compostos é a HPLC sendo imprescindível uma etapa de derivatização para a detecção por fluorescência. Assim, este trabalho teve como objetivo desenvolver métodos simples e rápidos para a determinação direta de glifosato e AMPA em água para consumo humano, sem etapa de tratamento de amostra ou derivatização, empregando a cromatografia iônica capilar (CIC) e cromatografia líquida acoplada à espectrometria de massas em série (LC-MS/MS). O método por cromatografia iônica capilar empregou coluna capilar IonPac AS19 (250 x 0,4 mm; 7,5 μm), gradiente de eluição com gerador de KOH e detector de condutividade. Para LC-MS/MS utilizou-se coluna Pursuit XRs C18 (150 x 2 mm d.i.; 5 μm), solução aquosa de hidróxido de amônio 0,01% (v/v) como fase móvel e detector triplo quadrupolo (TQ) operando no modo de monitoramento de reações selecionadas (SRM). A validação dos métodos foi realizada através de ensaios de recuperação, avaliando exatidão, precisão, linearidade, limites de detecção (LD) e de quantificação (LQ). As curvas analíticas para glifosato e AMPA obtidas para os níveis 25; 50; 100; 250 e 500 μg L-1 por CIC e LC-MS/MS apresentaram coeficiente de determinação (r2) >0,99 para ambos os compostos. Para as fortificações nos níveis de 25, 50, 250 e 500 μg L-1 os compostos apresentaram recuperações de 70 a 113%, com RSD <10% por CIC e recuperações foram de 79 a 105%, com RSD <14% por LC-MS/MS. Os limites de detecção do método para AMPA e glifosato foi 8 μg L-1 e o de quantificação foi de 25 μg L-1, em ambas as técnicas cromatográficas. Ambos os métodos apresentaram boa seletividade e sensibilidade. A determinação, através da injeção direta da amostra, de glifosato e AMPA utilizando cromatografia iônica capilar ou LC-MS/MS demonstrou ser uma ferramenta eficiente, simples e economicamente viável para análise, permitindo o monitoramento em níveis abaixo do valor máximo permitido em água potável.

While many of the currently used network intrusion detection systems (NIDS) employ signature-based approaches, there is an increasing research interest in the examination of anomaly-based detection methods, which seem to be more suited for recognizing zero-day attacks. Nevertheless, requirements for their practical deployment, as well as objective and reproducible evaluation methods, are hereby often neglected. The following thesis defines aspects that are crucial for a practical evaluation of anomaly-based NIDS, such as the focus on modern attack types, the restriction to one-class classification methods, the exclusion of known attacks from the training phase, a low false detection rate, and consideration of the runtime efficiency. Based on those principles, a framework dedicated to developing, testing and evaluating models for the detection of network anomalies is proposed. It is applied to two datasets featuring modern traffic, namely the UNSW-NB15 and the CIC-IDS-2017 datasets, in order to compare and evaluate commonly-used network intrusion detection methods. The implemented approaches include, among others, a highly configurable network flow generator, a payload analyser, a one-hot encoder, a one-class support vector machine, and an autoencoder. The results show a significant difference between the two chosen datasets: While for the UNSW-NB15 dataset several reasonably well performing model combinations for both the autoencoder and the one-class SVM can be found, most of them yield unsatisfying results when the CIC-IDS-2017 dataset is used. / Obwohl viele der derzeit genutzten Systeme zur Erkennung von Netzwerkangriffen (engl. NIDS) signaturbasierte Ansätze verwenden, gibt es ein wachsendes Forschungsinteresse an der Untersuchung von anomaliebasierten Erkennungsmethoden, welche zur Identifikation von Zero-Day-Angriffen geeigneter erscheinen. Gleichwohl werden hierbei Bedingungen für deren praktischen Einsatz oft vernachlässigt, ebenso wie objektive und reproduzierbare Evaluationsmethoden. Die folgende Arbeit definiert Aspekte, die für eine praxisorientierte Evaluation unabdingbar sind. Dazu zählen ein Schwerpunkt auf modernen Angriffstypen, die Beschränkung auf One-Class Classification Methoden, der Ausschluss von bereits bekannten Angriffen aus dem Trainingsdatensatz, niedrige Falscherkennungsraten sowie die Berücksichtigung der Laufzeiteffizienz. Basierend auf diesen Prinzipien wird ein Rahmenkonzept vorgeschlagen, das für das Entwickeln, Testen und Evaluieren von Modellen zur Erkennung von Netzwerkanomalien bestimmt ist. Dieses wird auf zwei Datensätze mit modernem Netzwerkverkehr, namentlich auf den UNSW-NB15 und den CIC-IDS- 2017 Datensatz, angewendet, um häufig genutzte NIDS-Methoden zu vergleichen und zu evaluieren. Die für diese Arbeit implementierten Ansätze beinhalten, neben anderen, einen weit konfigurierbaren Netzwerkflussgenerator, einen Nutzdatenanalysierer, einen One-Hot-Encoder, eine One-Class Support Vector Machine sowie einen Autoencoder. Die Resultate zeigen einen großen Unterschied zwischen den beiden ausgewählten Datensätzen: Während für den UNSW-NB15 Datensatz verschiedene angemessen gut funktionierende Modellkombinationen, sowohl für den Autoencoder als auch für die One-Class SVM, gefunden werden können, bringen diese für den CIC-IDS-2017 Datensatz meist unbefriedigende Ergebnisse.