SOURCE LOCALIZATION BY VIRTUAL ACOUSTIC REALITY
Maršálek, Petr; Rund, František; Štorek, Dominik
2016 - anglický This thesis presents a new algorithm for virtual sound source positioning for the purposes of rendering acoustic components of headphone-based virtual reality. The one-channel positioning method using Di erential Head-Related Transfer Function (DHRTF) is introduced. The DHRTF method utilizes localization cues (time and intensity di erences) extracted from HRTF pairs for particular positions. In contrast to the usual two-channel ltering by a HRTF pair, only one channel is processed when the DHRTF algorithm is applied. This results in one channel is being delayed and attenuated with respect to the other one, which contains the unprocessed original sound. Therefore, the nal positioned sound has the same time and level di erences as when ltered by the HRTF pair, however, with di erent signals in each channel. An introduction to the eld of virtual positioning and contemporary state-of-the-art presents the rst chapter. The theoretical concept of the DHRTF positioning method is presented and demonstrated with model examples in the second chapter. Speci c features, performance, and e ciency of the DHRTF algorithm are examined and compared to the attributes of other positioning methods (namely amplitude panning and HRTF ltering) in the last chapter. The highlights of the results delineate remarkably good spatial depth, reduction of processing requirements, and low a ection of natural sound timbre. Possible applications of the DHRTF method are in assistive systems for visually impaired, in computer games audio, and in enhancing spatialization of musical or lm mixing on digital audio workstations. Plné texty jsou dostupné v digitálním repozitáři ČVUT. SOURCE LOCALIZATION BY VIRTUAL ACOUSTIC REALITY

This thesis presents a new algorithm for virtual sound source positioning for the purposes of rendering acoustic components of headphone-based virtual reality. The one-channel positioning method ...

Maršálek, Petr; Rund, František; Štorek, Dominik
České vysoké učení technické v Praze, 2016

Analysis of disfluency in pathological speech
Čmejla, Roman; Lustyk, Tomáš
2016 - anglický Plné texty jsou dostupné v digitálním repozitáři ČVUT. Analysis of disfluency in pathological speech

Čmejla, Roman; Lustyk, Tomáš
České vysoké učení technické v Praze, 2016

Automated assessment of diadochokinesis and resonance in dysarthrias associated with basal ganglia dysfunction
Čmejla, Roman; Rusz, Jan; Novotný, Michal
2016 - anglický The speech disruption caused by neurodegenerative disease is termed dysarthria. Previous research has identified dysarthria as one of the earliest and the most common subclinical manifestation associated with neurodegenerative diseases such as Parkinson’s disease (PD), Huntington’s disease (HD), multiple system atrophy (MSA) and progressive supranuclear palsy (PSP). Even though presence of articulatory and resonatory deficits represent an important part of dysarthria manifestations the most widely used methods currently available for the automatic evaluation of speech performance are focused only on the assessment of dysphonia. Therefore the aim of the thesis is to present the automatic assessment of articulatory and resonatory deficits as a useful tool enhancing current approaches of dysarthria analysis. Furthermore, the next goal is to use the objective dysarthria assessment to describe different dysarthria patterns. For this reason, we have examined articulatory deficits from two perspectives: first as the description of fast repetitive moves produced during speech diadochokinetic task (repetition of /pa/, /ta/ and /ka/ syllables) and second one as the description of the velopharyngeal function using a sustained phonation of the vowel /i/. The thesis summarizes the results of several studies and therefore of different participant datasets, including speakers diagnosed with PD, MSA, PSP, HD and healthy control (HC) speakers, which were used for the purposes of articulatory and resonatory deficits estimation. In the first part aimed at rapid articulatory movements, the fully automatic assessment of several acoustic features describing six dysarthria dimensions (i.e. voice quality, laryngeal and supralaryngeal coordination, precision of consonant articulation, tongue movements, occlusion weakening and speech timing) was proposed. Results presented in this section showed significant differences between healthy and pathological utterances and moreover revealed distinctive patterns of articulatory distortion connected with different neuropathology. The presence of these distinctive patterns has been supported by the results of the performed classification experiment, which was able to distinguish between PD vs. HC with 87% accuracy. The second part used the 1/3-octave spectrum method to describe the presence of nasal resonance as an acoustic feature reflecting increased nasality. Results of this section showed an overall increased presence of hypernasality in Parkinsonian and Huntingtonian patients. Nevertheless, significant differences were found only among the HD speakers. More importantly the intermittent character of hypernasality was found to be one of the very distinctive features present in 78% of HD speakers and moreover a significant increase in the presence of intermittent hypernasality was also documented in 68 % of MSA speakers. In general, the results presented in the thesis show that the automatic assessment of articulatory deficits may provide useful clues for the early diagnosis, monitoring of disease progression, monitoring of speech therapy efficacy and adjustment and the differential diagnosis.Poškození řeči spojené s neurodegenerativními onemocněními se nazývá dysartrie. Dřívější výzkum ukázal, že dysartrie je jedním z nejdříve se vyskytujících a nejčastějším subklinickým projevem neurodegenerativních onemocnění jako jsou Parkinsonova nemoc (PD), Huntingtonova nemoc (HD), mnohočetná systémová atrofie (MSA) a progresivní supranukleární palasa (PSP). I přestože artikulační a rezonanční obtíže patří mezi významné dysartrické projevy, nejčastěji využívané metody automatického hodnocení jsou založeny pouze na měření dysfonie. Proto je hlavním cílem této práce návrh automatického hodnocení artikulačních a rezonančních deficitů jako užitečného nástroje doplňujícího metody používané v současnosti. Navíc je dalším cílem využití navržených objektivních metod pro hodnocení dysartrie k popsání různých dysartrických profilů. Z toho důvodu byly dysartrie zkoumány ze dvou perspektiv první bylo vyšetření rychlých artikulačních pohybů založené na úloze rychlého opakování slabik /pa/,/ta/ a /ka/ a tou druhou bylo hodnocení funkce měkkého patra založené na prodloužené fonaci samohlásky /i/. Dizertační práce shrnuje výsledky několika studií, a proto také pracuje s více různými soubory dat získaných od účastníků diagnostikovaných s PD, MSA, PSP, HD a navíc s kontrolními nahrávkami zdravých dobrovolníků. V první části zaměřené na rychle artikulační pohyby je představena metodika automatického hodnocení šesti řečových dimenzí zahrnujících kvalita hlasu, koordinaci laryngeálních a supralaryngeálních artikulátorů, přesnost artikulace konsonant pohyby jazyka, slábnutí okluze, časování řeči. Výsledky prezentované v této části ukazují signifikantní rozdíly mezi zdravými a patologickými promluvami a navíc odhalili rozdílné charaktery řečových poruch u rozdílných typů neurodegenerativních onemocnění. Prezentovaný klasifikační experiment ukázal, že je možné rozčlenit de novo PD pacienty od zdravých účastníků s 88% úspěšností. V druhá část, za použití analýzy 1/3-oktávového spektra, popisuje přítomnost nasální rezonance jako akustického ukazatele reflektujícího hypernasalitu. Výsledky prezentované v této části ukazují zvýšení výskytu hypernasality u PD i HD pacientů. Nicméně pouze u pacientů diagnostikovaných s HD byly nalezeny statisticky signifikantní rozdíly. Navíc se u HD a MSA se ukázala jako významná přítomnost kolísavé hypernasality kterou jsme dokumentovali u 78 % případů HD a u 68 % případů MSA Celkově, výsledky prezentované v této dizertační práci ukazují, že automatické hodnocení artikulační a rezonančních poruch může poskytnout užitečná vodítka pro potřeby brzké a diferenciální diagnózy, monitorování progrese onemocnění, monitorování efektivity léčby a monitorování účinků řečové terapie. Plné texty jsou dostupné v digitálním repozitáři ČVUT. Automated assessment of diadochokinesis and resonance in dysarthrias associated with basal ganglia dysfunction

The speech disruption caused by neurodegenerative disease is termed dysarthria. Previous research has identified dysarthria as one of the earliest and the most common subclinical manifestation ...

Čmejla, Roman; Rusz, Jan; Novotný, Michal
České vysoké učení technické v Praze, 2016

Microwave Methods for Interferometric Measurements
Hoffmann, Karel; Hudec, Přemysl; Příhoda, Milan
2016 - anglický Interferometrická mikrovlnná měření mohou najít uplatnění v celé řadě aplikací jako jsou měření malých fázových rozdílů, měření vysokých impedancí nebo nedestruktivní testování. Ačkoliv jsou interferometrická měření dobře známá například z optiky, v mikrovlnné technice mají svá specifika. Jak ukazují různé práce publikoavné v nedávné době, je zde několik nevyřešených problémů týkající se interferometrických měření. Jedním z takových problémů je přítomnost systematických chyb, která je v řadě prací opomíjena a systémy jsou považovány za ideální. S tím souvisí i fakt, že nebyly vyvinuty vhodné kalibrační techniky a vhodné kalibračních standardy pro taková měření. V případě přesného měření fází zcela chybí metodika pro absolutní měření, protože doposud publikované práce používaly interferometrické systémy jen pro relativní měření. Cílem této disertační práce je vyřešit problém absolutního měření pro přesné měření fází, analyzovat systematické chyby použitých systémů a navrhnout vhodnou kalibrační/korekční techniku pro odstranění těchto systematických chyb. Tyto cíle jsou v práci systematicky řešeny, od základních rozborů a analýz systémů s idealizovanými parametry, přes rozbor systematických chyb, díky němuž bylo možné navrhnout zcela speciální kalibrační techniku, která využívá frekvenčně závislých modelů pro kalibrační standardy a vyhodnocuje měřená data přes celé kmitočtové pásmo zároveň narozdíl od běžných kalibračních technik. Navržená kalibrační technika byla experimentálně ověřena pomocí interferometrického systému ve frekvenčním pásmu 8÷10GHz. Výsledky prokázaly, že navržená technika umožňuje kalibraci interferometrického systému pro přesné měření fází. Dále bylo ukázáno, že nejhorší případ nejistoty měření s takto zkalibrovaným systémem je téměř dvakrát menší v porovnání s přímým měřením pomocí vektorového obvodového analyzátoru.The microwave interferometric measurements can be used in the several applications such as measurement of the small phase differences, measurement of the extreme impedances or near-field sensing. Although the interferometric measurements are well-known from the optics but there are specific aspects of such a measurement in the microwaves. As was shown in the recently published works, there are several problems regarding these measurements, which haven’t been solved yet. One of the problems is the existence of the systematic errors which were in the most cases completely neglected and the interferometric system was assumed as ideal. It is very closely related to the fact that there is no suitable calibration technique as well as calibration standards for such a systems. Finally, for the case of precise phase measurement or measurement of small phase differences there is lack of the method for the absolute measurement. Works published until now utilized interferometric measurement only for the relative phase measurement. The main aim of this thesis is to find an approach for the absolute measurement of the phase, to analyze the effect of the systematic errors and to design a suitable calibration technique to eliminate these systematic errors. These goals are successively analyzed and solved. Starting with the analysis of the idealized systems through the systematic errors analysis which helped to design new calibration technique, which utilize the parametric frequency dependant model and moreover contrary to the ordinary calibration techniques it evaluate the measured data simultaneously over the entire bandwidth. This technique was experimentally verified using the interferometric system in the frequency range 8 ÷ 10GHz. The results prove that the proposed technique enables the precise interferometric measurement of the phase. Additionally it was shown that the worst case of the uncertainty using the calibrated interferometric system is roughly twice smaller in comparison with the direct VNA measurement. Klíčová slova: Interferometrické měření; měření fází; kalibrační techniky; vektorový obvodový analyzátor Plné texty jsou dostupné v digitálním repozitáři ČVUT. Microwave Methods for Interferometric Measurements

Interferometrická mikrovlnná měření mohou najít uplatnění v celé řadě aplikací jako jsou měření malých fázových rozdílů, měření vysokých impedancí nebo nedestruktivní testování. Ačkoliv jsou ...

Hoffmann, Karel; Hudec, Přemysl; Příhoda, Milan
České vysoké učení technické v Praze, 2016

Understanding Formal Arrow-Connected Diagrams and Free-form Sketches
Hlaváč, Václav; Průša, Daniel; Bresler, Martin
2016 - anglický Drawing has been a natural way for humans to express their thoughts and ideas since ancient times. People got used to create and understand illustrations. Many wellde ned formal notations have evolved including technical drawings, musical scores, and various diagrams. On the other hand, people use free-form sketches or ad hoc intuitive notations often. Electronic devices equipped with a touch screen take part in our daily lives and make it is easier to create and share drawings. Naturally, it is more and more desired to have methods for automatic recognition and understanding of drawings. It is much easier and desired in the case of formal drawings consisting of well de ned entities. A computer system can work further with such recognized drawing { beautify the drawing, rearrange a diagram, perform some actions according to a diagram, or manufacture something according to a technical drawing. On the other hand, it is extremely di cult and potentially unnecessary in the case of free-form sketches. It is often enough to provide the user with tools for easier creation and manipulation with sketches. The reason is that these sketches are meant to be understood by humans only and the computer system serves for their creation, storage, and sharing. Obviously, we can classify drawing into formal drawings and free-from sketching. This thesis deals with two tasks: recognition of formally de ned diagrams and segmentation of object of interest in free-form sketches. These two apparently di erent topics are strongly related. We assume that the drawing is created on an electronic device and thus it is in form of a sequence of strokes rather than a raster image. We propose a recognition framework for arrow-connected diagrams. We introduce a model for recognition by selection of symbol candidates, based on evaluation of relations between candidates using a set of predicates. It is suitable for simpler structures, in which the relations are explicitly given by symbols, arrows in the case of diagrams. Knowledge of a speci c diagram domain is used. The two domains are owcharts and nite automata. We created a benchmark database of diagrams from these two domains. Although the individual pipeline steps are tailored for these, the system can be readily adapted for other domains. The recognition pipeline consists of the following major steps: text/non-text separation, symbol segmentation, symbol classi cation, and structural analysis. We performed a comparison with state-of-the-art methods for recognition of owcharts and nite automata and veri ed that our approach outperforms them. We also analysed our system thoroughly and identi ed most frequent causes of recognition failures. The situation is more complicated in the case of a free-from sketching where the user can draw and write anything freely. We cannot expect any particular structure. We can often nd a combination of pictorial drawing with more structured gures and text in form of labels. The classical segmentation by recognition can not be used here. In some cases, the understanding of a sketch might be di cult even for humans. Speci cally, the identi cation of an individual object in a drawing may di er from user to user, case to case. Recent research showed that a single linkage agglomerative clustering of strokes with trainable distance function can be used for segmentation of objects from prede ned symbol classes in formal drawings. The proper distance function can be learned from annotated data. It requires a lot of data and time. We propose an approach combining several pre-trained distance functions for particular structures together to segment object in free-form sketches. We show that the desired segmentation can be achieved in many cases by combining together distance functions trained for very general object types like rows, columns, words, or compact images. We also show that the best combination of distance functions can be found from a very limited data in real time. We propose a segmentation approach, which estimates the optimal combination of clustering distance functions from an initial selection of one object. It results in segmentation of objects, which have similar characteristics to the initial one. Based on this approach, we designed a selection tool bringing additional functionality allowing to select and manipulate the segmented objects seamlessly. The method is suitable for fast rearrangement of sketches during collaborative content creation (brainstorming). Plné texty jsou dostupné v digitálním repozitáři ČVUT. Understanding Formal Arrow-Connected Diagrams and Free-form Sketches

Drawing has been a natural way for humans to express their thoughts and ideas since ancient times. People got used to create and understand illustrations. Many wellde ned formal notations have ...

Hlaváč, Václav; Průša, Daniel; Bresler, Martin
České vysoké učení technické v Praze, 2016

Distributed Signal Processing in Wireless Sensor Networks for Diagnostics
Šmíd, Radislav; Neužil, Jan
2016 - anglický This work introduces the on-board vibrational condition monitoring of aircrafts powerplants by its original novel approach of distributed WSN based vibrational condition monitoring. Proposed approach allows to employ computationally intensive methods of vibra- tional signal processing and methods of condition monitoring in computationally weak wireless sensor network. It introduces the fault detection method based on novelty detection which is executed directly in wireless sensor nodes. The novel framework of WSN condition monitoring system with its key attributes Adaptivity, Recon gurability and Three phases of operation enables the immediate fault detection capability while providing long-term monitoring. The proposed methods and framework were evaluated by the means of experiment on the small turbojet engine. Plné texty jsou dostupné v digitálním repozitáři ČVUT. Distributed Signal Processing in Wireless Sensor Networks for Diagnostics

This work introduces the on-board vibrational condition monitoring of aircrafts powerplants by its original novel approach of distributed WSN based vibrational condition monitoring. Proposed ...

Šmíd, Radislav; Neužil, Jan
České vysoké učení technické v Praze, 2016

OnGIS: Ontology-Based Geospatial Data Integration and Retrieval
Kouba, Zdeněk; Křemen, Petr; Šmíd, Marek
2016 - anglický Querying geospatial data from multiple heterogeneous sources backed by different management technologies poses an interesting problem in the data integration and the subsequent result interpretation. I propose two ways of entering complex spatial queries in a user-friendly way and broker techniques for answering a query by finding relevant data sources (from a catalogue) capable of answering the query, eventually splitting the query and finding relevant data sources for the query parts, when no single source suffices. For expressing data source capabilities, I describe each source with a set of prototypical queries that are algorithmically arranged into a lattice, which makes searching efficient. The proposed algorithms leverage GeoSPARQL query containment enhanced with OWL 2 QL semantics. Using GeoSPARQL as both the querying language and the language for expressing capabilities gives the flexibility to incorporate many heterogeneous geospatial data sources, no matter what storage, data model, or terminology they use. All parts of the design are presented in a prototype called OnGIS, which implements the proposed methods of entering a user query, searching for the relevant data sources, and translating the query to requests the underlying data source technologies understand. Plné texty jsou dostupné v digitálním repozitáři ČVUT. OnGIS: Ontology-Based Geospatial Data Integration and Retrieval

Querying geospatial data from multiple heterogeneous sources backed by different management technologies poses an interesting problem in the data integration and the subsequent result ...

Kouba, Zdeněk; Křemen, Petr; Šmíd, Marek
České vysoké učení technické v Praze, 2016

Demand Side Management System for Optimizing Operation of Power Grids with Renewable Energy Sources
Havel, Petr; Malík, Ondřej
2016 - anglický Plné texty jsou dostupné v digitálním repozitáři ČVUT. Demand Side Management System for Optimizing Operation of Power Grids with Renewable Energy Sources

Havel, Petr; Malík, Ondřej
České vysoké učení technické v Praze, 2016

Modelling of impacts of CO2 auctions on the district heating sector
Knápek, Jaroslav; Vecka, Jiří
2016 - anglický This doctoral thesis is describing major aspects of district heating technology and its position within energy sector in the context of the Czech Republic and EU as well and the most important district heating features. According to acknowledged European methodology, district heating networks delivering heat from CHP process could be regarded as one of the most efficient solution for covering heat needs. The core of the thesis is devoted to description and modelling the factors influencing the heat market, legislation factors could be seen as most important with potential to severely influence the market conditions. Optimization model of differential NPV was chosen as optimal solution for comparing the development in business-as-usual situation and under the new requirements and each relevant legislation effect is transformed into risk factor influencing future heat prices. Model outcomes are formulated in the two legislation and three CO2 price scenarios and confirm non-symmetrical effects on the heat market participants arising from the size of installations. There are various approaches how this distortion could be addressed. Among possible solutions indirect carbon tax based tool was identified as the most appropriate way how to remedy distortions on the heat market towards inclusion of CO2 costs into price of fuels for installations outside EU ETS. This solution could be easily realized by using existing legislation tools and application of possible remedial tool on the heat market is described. Results of these analyses support the idea of necessity to introduce this type of tool as soon as possible in order to avoid undue competition distortions on the heat market in the Czech Republic. Plné texty jsou dostupné v digitálním repozitáři ČVUT. Modelling of impacts of CO2 auctions on the district heating sector

This doctoral thesis is describing major aspects of district heating technology and its position within energy sector in the context of the Czech Republic and EU as well and the most important ...

Knápek, Jaroslav; Vecka, Jiří
České vysoké učení technické v Praze, 2016

Scaling in vehicle platoons
Šebek, Michal; Hurák, Zdeněk; Herman, Ivo
2016 - anglický This thesis deals with analysis of scaling in autonomous vehicle platoons in which vehicles keep xed distance to their neighbors. The vehicles are modelled as linear single-input single-output systems of arbitrary order. In order to control themselves, the vehicles use information from their nearest neighbors their predecessor and successor. The states used for coupling are mainly position and velocity, but other states are allowed too. The errors to neighbors in these states can be weighted di erently, hence the control law is asymmetric. Using the tools from distributed control, properties of platoons are analyzed. A comprehensive overview of the properties of platoons when identical asymmetry in all states is given. With the help of a newly derived product form of a transfer function in a network system, the steady-state gain, stability, string stability and particularly H1 norms are analyzed. The most important aspect specifying the scaling rate is the number of integrators in the open loop. For one integrator in the open loop the scaling of the H1 norm is quadratic for symmetric control and linear for asymmetric control. For two and more integrators the scaling is cubical for symmetric control and exponential for asymmetric. Since there is no good control for two integrators in the open loop with identical asymmetry, symmetric coupling in position and asymmetric in velocity is proposed. Such control is, at least for the cases analyzed in the thesis, superior to both completely symmetric and completely asymmetric control. It has similar convergence time as asymmetric control but it still keeps the bounded control e ort as the symmetric control does. Klíčová slova: vehicle platoons; distributed control; scaling; asymmetry; di erent asymmetries Plné texty jsou dostupné v digitálním repozitáři ČVUT. Scaling in vehicle platoons

This thesis deals with analysis of scaling in autonomous vehicle platoons in which vehicles keep xed distance to their neighbors. The vehicles are modelled as linear single-input single-output ...

Šebek, Michal; Hurák, Zdeněk; Herman, Ivo
České vysoké učení technické v Praze, 2016