[kand] Sähkötekniikan korkeakoulu / ELEC

Permanent URI for this collectionhttps://aaltodoc.aalto.fi/handle/123456789/13

Browse

Recent Submissions

Now showing 1 - 20 of 3616
  • Ethical Decision-Making Frameworks for Autonomous Vehicles
    (2024-12-15) Kantola, Selim
     Sähkötekniikan korkeakoulu | Bachelor's thesis
    The topic of this literature analysis thesis is frameworks that carry out decision-making through ethical algorithms, specifically in the context of autonomous vehicles (AV). A core aspect of these frameworks’ designs is the ethical basis and principles that the decision-making algorithms base their calculations on. Ethics is a broad subject with various theories, models, trends, philosophies, etc. The varying views among humans regarding ethical beliefs affects the development and end result of the ethical rules and the formation of a unified ethical system, the ethical basis, in the design phase of these frameworks. This thesis introduces ethics in a compact way, describing different ethical perspectives which these frameworks can base their decisions on, such as a modified version of utilitarianism, and the problems associated with ethical frameworks and their algorithms. The work also touches upon the cross-section between ethics and autonomous machines. The integration of ethical issues into mechanical, digital, and logical algorithms and processes is a vast topic, involving many potential problem areas. Some of these challenges are metrological errors, biases, the creation of an ethical foundation and reaching a consensus on values, the responsibility and safety of decision-making frameworks, and vague assessments and procedures in ethical decision-making algorithms and processes. Finally, throughout all the chapters, this thesis provides examples of real-life applications for these frameworks and proposes how their design process can be improved with good design practices. The final motivation is to inspire the reader in numerous vague ways resulting in emergent qualities of a future-oriented drive, be it a personal desire, high intrigue, the will to improve people’s lives, or even something greater than that.
  • Yksinkertaisen mittalaitegeometrian toteuttaminen, joka kykenee kasvattamaan ~5 nm:n hiukkasia
    (2024-12-15) Huuskonen, Uuno
     Sähkötekniikan korkeakoulu | Bachelor's thesis
    Tämä kandidaatintyö käsittelee laitteen rakentamista ja testaamista, jonka tarkoituksena on helpottaa pienten noin 5 nm:n kokoisten aerosolihiukkasten mittaamista. Aerosolien tutkimus on keskeistä ilmakehätutkimuksessa, sillä se auttaa muun muassa ilmastonmuutoksen vaikutusten ja ilmansaasteiden terveysvaikutusten kartoittamisessa. Pienten, erityisesti alle 5 nm:n, hiukkasten mittaaminen on kuitenkin hankalaa niiden luonteen takia, ja lisäksi niiden vaikutuksista ilmakehässä tiedetään suhteellisen vähän. Tässä työssä rakennetun laitteen avulla pyritään helpottamaan tämän kokoluokan aerosolien mittaamista. Työn tarkoituksena on rakentaa resursseiltaan ja monimutkaisuudeltaan helposti saavutetta laite, joka parantaisi pienten aerosolien mittaustuloksia. Laitteen toiminta perustuu mittalaitteistossa tapahtuvien hiukkashäviöiden minimoimiseen ja pienten hiukkaskokojen kasvattamiseen. Työhön kuuluu laitteen rakentamisen lisäksi sen testaus. Laite pyritään säätämään mahdollisimman tehokkaaksi vaihtelemalla sen parametreja mittausten mukaan. Tulokset vaikuttavat lupaavilta, mutta laitteessa on vielä huomattavasti parannettavaa. Työn tulosten perusteella voidaan pohtia laitteen jatkokehittämistä sekä sen käytännöllisyyttä ilmakehätutkimuksessa.
  • Wireless Sensor Networks for Received Signal Strength-Based Target Localization and Tracking with Kalman Filter Data Processing
    (2024-12-15) Leino, Akseli
     Sähkötekniikan korkeakoulu | Bachelor's thesis
    Localization and tracking different targets have been and still are an important part of human society. There are countless ways to Locating and tracking targets. Each of them has their own advantages and disadvantages. In this thesis the received signal strength indicator -based locating and tracking techniques are explored. Received signal strength indicator is based on radio wave power decreasing inversely proportional to travelled distances square. Usually, radio waves power is measured in decibels, thus converting the travelled distance squared to logarithm of the travelled distance. Received signal strength indicator can be measured with a wireless sensor network. A wireless sensor network is a measurement cluster powered by small, independent, self-organizing, battery-powered nodes. In a wireless sensor network, there are two types of nodes. So called measurement nodes which measure the environment, and anchor nodes which estimate their own position. If something from the trackable target's movement modes is known, with the Kalman filter the measured data can be filtered. For this thesis, a simulation of a radio wave power decreasing in 100 meters by 100 meters room was created, to study the effects of anchor node count and positioning. Furthermore, the effectiveness of Kalman filter is briefly demonstrated in this thesis.
  • Occupant-targeted ventilation solutions and control methods
    (2024-12-10) Nordman, Axel
     Sähkötekniikan korkeakoulu | Bachelor's thesis
    Energy consumption and energy efficiency are crucial factors in modern buildings. The building sector accounts for approximately 40% of global energy consumption, half of which is used to power different heating, ventilation and air conditioning (HVAC) systems to ensure the quality of the indoor climate. Hence, there is a clear need to reduce the energy consumption of these systems, however, without compromising indoor air quality (IAQ), thermal comfort and overall wellbeing. Occupant-targeted methods for air distribution and HVAC control can play an important role in fulfilling these requirements. This thesis performs a literature review on occupant-targeted ventilation solutions and control methods introduced in commercial and educational buildings. The thesis presents information and research on advanced air distribution methods with the intention of improving traditional ventilation. This includes diffuse ceiling ventilation (DCEV), underfloor air distribution (UFAD), stratum ventilation (SV), impinging jet ventilation (IJV), wall-attached ventilation (WAV) and personalized ventilation (PV). Traditional mixing ventilation can cost-effectively be upgraded to DCEV, improving thermal comfort. PV enables personal control of supply air parameters, thus emerging as the most effective method for thermal comfort and air quality in the breathing zone. UFAD, IJV and traditional methods can also be used in combination with PV to maintain an acceptable background climate. In addition, UFAD, SV, IJV and WAV show overall promising results and could be considered favourable alternatives to traditional methods. The thesis also presents methods for energy-efficient HVAC control, where demand-controlled ventilation (DCV) is highly influential. This works on the principle of dynamically adjusting airflow rates and temperature based on current demand and occupancy to ensure an acceptable indoor climate and avoid excessive ventilation and energy consumption. CO2-based DCV is the most well-documented strategy, where indoor CO2 level is controlled based on a predefined setpoint. Model predictive control and machine learning (ML) contribute to more optimal control by predicting the CO2 level development. Another strategy is occupancy-based DCV, relying on detection or estimation of the number of occupants present. This can be achieved utilizing ML models, cameras equipped with computer vision, and Wi-Fi traffic metrics. Occupant activity and behaviour also affect the indoor climate and must be considered in HVAC control, albeit still in the early stages of development. Reinforcement learning (RL) and computer vision have been employed for predicting thermal preferences and estimating metabolic rate (MET) and clothing insulation (CLO). However, when cameras and intelligent methods are used, a balance is always required to comply with security and privacy concerns.
  • Combining Open and Closed-Loop Approaches for Task Planning in Robotics Using Language Models
    (2024-12-13) Dömötör, Albert
     Sähkötekniikan korkeakoulu | Bachelor's thesis
    Task planning for robotics has traditionally been done using methods that rely on formal descriptions of specific environments constructed by human experts and lack generalizability, posing a limit to what tasks robots can perform. With advancements in language models, a new method for robotic task planning is emerging that leverages the semantic knowledge contained in language models to plan robotic actions based on natural language conditioning. This thesis compares different methods proposed for robotic task planning with a focus on their respective strengths and shortcomings. Furthermore, a task planning framework employing a combined open and closed-loop planning approach is presented. The results indicate that 1) LLMs are capable of constructing plans in unseen environments based only on natural language conditioning 2) a combined approach to LLM-based task planning is feasible, however, careful implementation of primitives is key to effective task planning.
  • The value of diffusion-based MRI tractography in TMS interventions
    (2025-01-10) Liimatta, Jere
     Sähkötekniikan korkeakoulu | Bachelor's thesis
    Transcranial magnetic stimulation (TMS) is a brain stimulation technique that uses an electromagnetic coil placed outside the head to activate brain activity in a targeted area. Activation of a brain region results in corresponding functions being triggered, such as hand movement in the case of the motor cortex. Tractography is a modeling technique that leverages data derived from magnetic resonance imaging (MRI) to create three-dimensional representations of brain pathways. These pathways reveal connections between different brain regions. TMS is deployed as a treatment method for conditions such as depression and chronic pain. To ensure optimal treatment outcomes, the targeting of the stimulation area needs to be tailored to the individual neuroanatomy of the patient. Navigated transcranial magnetic stimulation (nTMS) adopts neuronavigation to facilitate an enhanced version of TMS that can define the stimulation area in a three-dimensional MRI in real time. In preoperative brain mapping, nTMS can assist in locating critical functions, such as those involved in speech production, near a brain tumor that is to be resected. This bachelor’s thesis, conducted as a literature review, investigates the value of combining transcranial magnetic stimulation (TMS) withtractography in interventions. The majority of studies deployed diffusion tensor imaging (DTI), a tractography technique argued to have shortcomings. However, DTI has contributed to the advancement of both technology and research. In neurosurgery, tractography provides value by depicting critical brain pathways, needed to preserve brain functionality. In other TMS-based therapeutic applications, tractography has the potential to enhance the efficacy of treatments, particularly in cases where functional brain regions may have reorganized. The thesis concludes that integrating TMS with tractography facilitates the devel opment of more individualized treatments. However, there is a need for large-scale studies employing state-of-the-art techniques and standardization to better assess the influence of white matter connectivity on treatment outcomes.
  • Taajuusmuuttajan analyysi ORTin eri vaiheissa
    (2025-01-10) Puustinen, Sauli
     Sähkötekniikan korkeakoulu | Bachelor's thesis
    Taajuusmuuttajien käyttö on kasvanut merkittävästi viime vuosikymmeninä, mikä on lisännyt niiden luotettavuusvaatimuksia. Luotettavuuden parantamiseksi on tärkeää testata taajuusmuuttajia laboratoriossa todellisia käyttöolosuhteita simuloiden. Tässä työssä tutkittiin jatkuvan luotettavuustestauksen (ORT) kehittämistä, erityisesti sen analyysimenetelmiä eri vaiheissa. Työssä perehdyttiin taajuusmuuttajien rakenteeseen, toimintaperiaatteisiin sekä testausmenetelmiin. Nykyisessä analyysimenetelmässä havaittiin puutteita, kuten riittämätön lähtöanalyysi, datan riittämätön hyödyntäminen testauksen aikana sekä jälkitestauksen vähäisyys. Tutkimuksessa esitettiin kehitysehdotuksia, kuten vakioidun testauksen käyttöönotto ORT-testausta edeltävissä vaiheissa, datan tehokkaampi hyödyntäminen testauksen aikana sekä visuaalinen tarkastus testin jälkeen. Tutkimuksen tuloksena esitettiin monipuolinen ja vakioitu tapa analysoida taajuusmuuttajia ORT-testausten eri vaiheissa. Tämä lisää testauksen luotettavuutta ja tukee tuotteiden laadunvalvontaa.
  • Mathematical modeling of 3D mental rotation
    (2024-12-20) Kilpeläinen, Sanni
     Sähkötekniikan korkeakoulu | Bachelor's thesis
    This thesis mathematically models mental rotation. It refers to the ability to imagine an object in an orientation that differs from the perceived orientation. In particular, this thesis investigates the mental rotation of three-dimensional objects. Earlier research has suggested the time to mentally rotate such objects depends on the angle of the rotation. However, the transformations the human brain represents during mental rotation remain poorly understood. To obtain data, that captures the decision-making processes occurring during mental rotation, the Bidirectional Research AI and Neuroscience laboratory (BRAIN, Aalto) has implemented a virtual reality object discrimination experiment. This experiment allows the subjects to rotate the objects with a joystick. Assuming the subjects generate their mental process into the joystick, the joystick data provides information on the distinct number of rotations test subjects conduct per task. This thesis formulates a model that reproduces the data. The data implies the number of rotations increases with larger angular differences. However, the average number of these actions remains below two with any angular difference. As the numbers of actions correspond to the number of 90° rotations that produce the orientation difference between the objects, the thesis approaches the problem with a discrete model. It represents a cube as a graph and predicts the rotations based on the transformations of one or two faces of a cube. The thesis evaluates the model in an object discrimination task under several conditions of matching parts of two objects. The results show the model replicated the VR data the best with a condition, in which the model used two parts of an object to predict the rotation, and received information on no matching object parts. These results suggest mental rotation relies on the object symmetries. Consequently, relying on the symmetries suggests the imagined transformations during mental rotation may rather occur in discrete steps than as a continuous rotation.
  • Ionimoottoreiden toimintaperiaatteet ja hyödyt avaruustehtävissä
    (2024-12-17) Siikonen, Mikael
     Sähkötekniikan korkeakoulu | Bachelor's thesis
    Ionimoottoreiden käyttö avaruuslennoilla mahdollistaa pitkän toiminta-ajan ja korkean polttoainetehokkuuden, mikä tekee niistä keskeisen teknologian syvän avaruuden tutkimuksessa. Tämä kandidaatintyö tarkastelee sähköisen propulsion periaatteita, erityisesti ionimoottoreita, ja vertailee niiden etuja perinteisiin kemiallisiin rakettimoottoreihin. Työssä esitellään ionimoottoreiden tärkeimmät komponentit, kuten plasmageneraattori, ionikiihdytin ja neutralisointikatodi, sekä näiden teknologiset ratkaisut. Ionimoottoreiden kyky tuottaa tasaista työntövoimaa pitkien matkojen aikana tekee niistä erinomaisen valinnan avaruustehtäviin, joissa vaaditaan suuria nopeuden muutoksia. Euroopan avaruusjärjestön ja Japanin ilmailu- ja avaruustutkimusjärjestön yhteinen BepiColombo-missio toimii työssä tapausesimerkkinä. Ohjelma hyödyntää sähköistä propulsiota T6-ionimoottorien ja aurinkosähköjärjestelmän muodossa matkallaan Merkuriukseen, mikä korostaa ionimoottoreiden potentiaalia pitkillä avaruuslennoilla. Työ osoittaa, että ionimoottorit edustavat merkittävää edistysaskelta avaruustutkimuksessa. Ne tarjoavat resurssitehokkaan ja kestävän ratkaisun pitkiin avaruusmatkoihin, minkä ansiosta ne ovat keskeisessä roolissa nykyisissä ja tulevissa syvän avaruuden tutkimushankkeissa.
  • Kudoksen solupopulaatioiden tunnistaminen spatiaalisen transkriptomiikan datasta
    (2024-11-28) Pasternack, Ariel
     Sähkötekniikan korkeakoulu | Bachelor's thesis
    2000-luvun alussa kehitetyt korkean suorituskyvyn sekvensointiteknologiat mahdollistavat jopa satojen miljoonien emässekvenssien yhtäaikaisen sekvensoinnin. Yleisessä käytössä olevat RNA-sekvensointiteknologiat kuten bulk-RNA-sekvensointi sekä yksisolu-RNA-sekvensointi johtavat kuitenkin luettujen transkriptien paikkatietojen menettämiseen. Spatiaalisen transkriptomiikan avulla kudosta voidaan analysoida avaruudellisessa kontekstissa mahdollistaen geeni-ilmentymien ja solupopulaatioiden sijainnin määrittämisen kudoksessa. Spatiaalisen transkriptomiikan teknologiat tuottavat suurikokoisia ja monimutkaisia datasettejä, joiden analysoimiseen on kehitetty tehokkaita laskennallisia menetelmiä. Laskennalliset menetelmät mahdollistavat muun muassa korkeaulotteisen datan visualisoimisen ja klusteroinnin kahdessa ulottuvuudessa. Työssä tehdään katsaus yleisimmässä käytössä oleviin menetelmiin ja perehdytään erityisesti t-SNE (t-distributed Stochastic Neighbor Embedding) -menetelmään. Työn kokeellisessa osiossa sovelletaan käsiteltyjä menetelmiä munarauhasen kuorikerroksesta tuotetulle spatiaalisen transkriptomiikan datasetille. Osiossa havainnollistetaan, miten histologisen leikkeen solut voidaan jakaa populaatioittain niiden sijaintia ja geeniekspressiota hyödyntäen. Analyysissä havaitaan, että käytetyn teknologian merkittävin haasteen on heikko resoluutio, joka rajoittaa merkityksellisten soluklusterien muodostumista. Teknologia mahdollistaa kuitenkin geeniekspressiokuvioiden hahmottumisen sekä tehokkaan solutyyppien paikantamisen kudokseen markkerigeenejä hyödyntämällä.
  • Metaomiikkamenetelmien soveltaminen suolistomikrobiomin patogeenisten toiminnallisten tilojen selvittämiseksi
    (2024-12-08) Junninen, Tapio
     Sähkötekniikan korkeakoulu | Bachelor's thesis
    Metaomiikkamenetelmillä, kuten metagenomiikalla, metatranskriptomiikalla, metaproteomiikalla ja metametabolomiikalla, voidaan tutkia mikrobiomia eri näkökulmista. Menetelmät tarjoavat mahdollisuuden kartoittaa mm. mikrobiomin koostumusta, aktiivisia geenejä, proteiinien toimintaa ja aineenvaihduntatuotteita tarjoten kokonaisvaltaisen näkemyksen mikrobiomin vaikutuksista ihmisen terveyteen. Tämä kandidaatintyö on suoritettu kirjallisuustutkimuksena, ja sen tavoite on selvittää, miten metaomiikkamenetelmiä voidaan soveltaa suolistomikrobiomin patogeenisten toiminnallisten tilojen selvittämisessä. Työssä käytetään esimerkkinä suolistosyöpää, jonka syntyyn ja etenemiseen mikrobiomin epätasapainolla ja sen aineenvaihduntatuotteilla on todettu olevan yhteys. Tuloksena todetaan, että metaomiikkamenetelmät voivat auttaa esimerkiksi tunnistamaan biomarkkereita ja tarjoamaan uusia näkökulmia diagnoosiin sekä hoitokeinojen kehitämiseen. Menetelmien käyttöä rajoittavat kuitenkin esimerkiksi datan analyysin monimutkaisuus, näytteiden käsittelyn standardoinnin puutteet ja tietokantojen rajallisuus. Työn johtopäätöksenä todetaan, että metaomiikkamenetelmillä on merkittävä rooli suolistomikrobiomin patogeenisten tilojen ymmärtämisessä. Tulevaisuudessa niiden hyödyntäminen voi parantaa sairauksien diagnostiikkaa ja yksilöllisten hoitomuotojen kehittämistä, kunhan menetelmien optimointi ja integrointi kehittyvät. Työ osoittaa, että metaomiikkamenetelmillä on merkittävä rooli sairauksien mekanismien ymmärtämisessä ja mikrobiomin toiminnan kliinisessä soveltamisessa.
  • Time-Sensitive Networking: Time Synchronization in Wi-Fi 7
    (2024-12-15) Räisänen, Elias
     Sähkötekniikan korkeakoulu | Bachelor's thesis
  • Yliopiston runkoverkon toteuttaminen EVPN-tekniikalla
    (2024-11-27) Mustajoki, Tiitus
     Sähkötekniikan korkeakoulu | Bachelor's thesis
    EVPN (engl. Ethernet Virtual Private Network) on Ethernet-protokollaa käyttävä virtuaalinen erillisverkko, jota voidaan hyödyntää suuren skaalan datakeskuksissa ja niiden välisessä liikenteessä. EVPN-tekniikan avulla informaatiota pystytään kujettamaan tehokkaasti niin datakeskusten sisällä kuin niiden välilläkin. Tapoja toteuttaa datakeskusten välinen runkoverkko EVPN-tekniikan avulla on kuitenkin paljon. Tämän työn tarkoitus on selvittää, mikä menetelmä soveltuu parhaiten Aalto-yliopiston käyttötarkoituksiin. Tähän liittyen työssä tarkastellaan myös erilaisia datakeskusten arkkitehtuureja. Lisäksi EVPN-tekniikkaan liittyy vahvasti underlay- ja overlay-verkot, joiden merkitystä työssä myös avataan. Tämän kirjallisuustutkimuksen aikana löydettiin ratkaisuja erilaisiin tilanteisiin. Vaikka kohdedatakeskukset ovat tunnettuja, vaikuttaa oikean menetelmän valintaan silti moni muuttuja. Erilaiset prioriteetit ja olemassa oleva laitteisto voivat tukea tietyn menetelmän valintaa. Datakeskusten arkkitehtuurien tapauksessa eri datakeskuksissa voi olla jopa kannattavaa käyttää eri arkkitehtuureja.
  • Ice Detection for Small Lakes with Satellite Imagery and Machine Learning
    (2024-12-08) Arjanne, Juho
     Sähkötekniikan korkeakoulu | Bachelor's thesis
    The role of satellite imagery in the continuous surface classification of the Earth has grown rapidly through the start of the 21st century. Advancements in machine learning and satellite imaging technology have facilitated the realisation of automated ice cover detection. Whilst sea ice detection has garnered great interest from researchers, the detection of lake ice has amassed considerably less focus. Specifically, few studies have evaluated ice detection methods for small lakes, which, albeit less prominent than larger water bodies, possess environmental and economic significance. This bachelor's thesis compares combinations of prevalent satellite imagery technologies and machine learning methods to find the optimal combination for small-lake ice detection. To discern the prevalent machine learning methods and imaging technologies, the trends and results of previous ice detection studies were explored. Synthetic aperture radar (SAR) and multispectral imaging (MSI) were identified as the two major imaging technologies applied to ice detection. For the machine learning methods, support vector machines (SVMs) and convolutional neural networks (CNNs) were found to be widely employed for ice detection research. The open availability and resolution of the SAR and MSI data provided by the Sentinel-1 and Sentinel-2 pairs of satellites led to their selection as the data source for an ice detection experiment detailed in this thesis. The experiment was performed by constructing several machine learning models for each combination of the chosen satellite imagery technologies and machine learning methods. The models were trained and initially assessed on data from two Canadian lakes. The global applicability of the models was then evaluated with a test set of images from lakes, which were not sources for the training set. The results of the experiment showed a significant difference in water-ice classification performance between the SAR and MSI models. Perfect validation accuracies were achieved with the MSI models, while the highest validation accuracy reached with a SAR model was 89.29%. For the test set, a CNN trained with MSI data attained the highest accuracy of 98.57%. The most accurate classification performance for SAR models was acquired with an SVM model, which managed a 73.20% accuracy. While the models trained on SAR data were only qualified for local indicative ice predictions, the test results suggest the CNN trained on MSI data to be globally applicable for small-lake ice detection.
  • Q-oppiminen ja töiden reititysongelma kahdelle palvelimelle
    (2024-12-15) Ala-Turkia, Henry
     Sähkötekniikan korkeakoulu | Bachelor's thesis
    Datakeskusten tehokas toiminta on keskeinen tekijä modernissa tietotekniikassa, niiden suorituskyky vaikuttaa suoraan monien yritysten ja palveluiden luotettavuuteen. Yksi keskeisistä haasteista on töiden reititysongelma, jossa tavoitteena on jakaa tehtävät optimaalisesti palvelimille niin, että palvelun viive minimoidaan ja järjestelmän tehokkuus maksimoidaan. Tässä työssä keskitytään kahden palvelimen reititysongelmaan heterogeenisessa ympäristössä, jossa palvelimien palvelunopeudet voivat vaihdella. Työssä hyödynnetään Q-oppimista, joka on malliton vahvistusoppimisen algoritmi, reitityspolitiikan optimointiin. Q-oppimisen avulla opittua politiikkaa vertaillaan perinteisiin heuristiikoihin, kuten JSQ- ja SED-politiikoihin, sekä satunnaistettuihin politiikkoihin. Tulosten perusteella arvioidaan, kuinka hyvin Q-oppiminen soveltuu heterogeenisten ympäristöjen reititysongelman ratkaisemiseen.
  • Brouwer Fixed Point Theorem
    (2024-12-14) Bergqvist, Eliel
     Sähkötekniikan korkeakoulu | Bachelor's thesis
    Fixed points are points, where a functions output and input match. Fixed point theorems state conditions for the existence of fixed points. These theorems are fundamental results in mathematical analysis, and they have significant and long reaching consequnces in a wide range of areas of pure and applied mathematics. The focus of this thesis is on two important fixed point theorems: the Banach and the Brouwer fixed point theorems. Banach’s fixed point theorem states that contraction mappings defined on a complete metric spaces, have a unique fixed point. Brouwer’s fixed point theorem ensures the existence of fixed points but not their uniquness. However it holds for all continuous mappings that are defined on compact convex sets, which is often a less strict condition to meet. Traditionally the proof of Brouwer’s fixed point theorem is based on algebraic topology. This thesis will however present an elementary proof that avoids these complications. We will also introduce two major consequences of Banach and Brouwer fixed point theorems: the Picard-Lindelöf theorem and Nash’s existence theorem.
  • Enhancement of audio quality at low frequencies with missing fundamental phenomenon
    (2024-11-27) Viitanen, Ari
     Sähkötekniikan korkeakoulu | Bachelor's thesis
    By leveraging a psychoacoustic phenomenon, digital signal processing can create the perception of bass frequencies without the corresponding physical sound waves. This enables the enjoyment of deep bass tones and rhythmic elements in music or sound effects, even when using small speakers incapable of reproducing related low frequencies. The phenomenon, known as the "missing fundamental", occurs within the auditory system. When a series of sinusoids forming a harmonic series is heard, it is perceived as being associated with the fundamental frequency, despite the absence of the fundamental harmonic itself. The virtual bass technique synthesizes upper harmonic series into the audible range, generating the perceptual illusion of bass frequencies that are otherwise inaudible. This literature research investigates the audio quality at low frequencies enhanced by four virtual bass systems, utilizing near-linear devices, a phase vocoder, and a hybrid approach for harmonic synthesis. The hybrid system separates tonal components, transients, and noise to be processed each with an optimal harmonic synthesis algorithm. The findings indicate an estimated enhancement of 33% in bass perception by the hybrid system, as evaluated through listening tests. Additionally, a significant observation was made with a 3-inch loudspeaker, which, despite its inability to reproduce frequencies below 110 Hz, achieved a perceptual bass quality comparable to a 5-inch speaker capable of reproducing frequencies as low as 68 Hz when the audio signal was pre-processed using virtual bass technology. The results suggest that virtual bass can be effectively implemented in digital audio playback systems, such as small speakers, laptops, and portable devices. Significant potential remains for advancing virtual bass technology. Machine learning and the increasing computational power of compact devices could soon enable near-real-time operation, improving synthesised bass realism. Crucially, standardised quality assessment methods are needed for virtual bass to provide objective evaluations and guide development.
  • OFDM-signaalin leikkaaminen ja leikkaamisen vaikutus systeemin suorituskykyyn
    (2024-12-12) Uusitalo, Olavi
     Sähkötekniikan korkeakoulu | Bachelor's thesis
    Tämä kandidaatintyö esittelee OFDM-signaalin leikkaamisen ja leikkaamisen vaikutuksen systeemin suorituskykyyn. Työ on tehty kirjallisuustutkimuksena, ja sen tavoitteena on selvittää, miten OFDM-signaalia leikataan, miksi leikkaaminen on tarpeellista ja mitä vaikutuksia leikkaamisella on signaaliin ja OFDM-systeemin suorituskykyyn. Tutkimus on toteutettu lukemalla useita erityisesti OFDM-signaalin leikkaamista käsitteleviä tutkimusraportteja. Tutkimuksessa havaitaan, että OFDM-signaaliin muodostuu väistämättä suuria tehohuippuja, jotka ovat ongelmallisia erityisesti lähettimen tehovahvistimessa. Näiden tehohuippujen hallinta on tarpeellista, ja signaalin leikkaaminen osoittautuu toimivaksi tehohuippujen hallinnan menetelmäksi. Signaalin leikkauksella havaitaan olevan sivuvaikutuksia, jotka edellyttävät leikkaamisen jälkeistä suodatusta ja mahdollisesti bittivirheiden hallintaa. Leikkaamisen vaikutuksen systeemin suorituskykyyn todetaan olevan tasapainoilua leikkaustavan ja sen sivuvaikutusten hallinnan välillä.
  • Exploring neural alignment in response to political statements
    (2024-12-11) Mäntysaari, Matias
     Sähkötekniikan korkeakoulu | Bachelor's thesis
    The paradoxical thinking intervention is an experimental paradigm in social psychology, which aims to facilitate co-operation in intractable conflicts. In the intervention, participants are subjected to information that is consistent with their beliefs, yet extreme or exaggerated. This is supposed to make the participant question their own views. The changing of attitudes during the intervention can be examined by observing the inter-subject neural alignment of participants via functional brain imaging. This thesis aims to explore the correlation between neural data, collected with magnetoencephalography (MEG), and explicit attitudes measured via questionnaires. The experimental data in this thesis is from a paradoxical thinking intervention, where supporters of the Social Democratic Party of Finland (SDP) listened to confrontational statements about the National Coalition Party (NCP). While listening to the statements in the MEG machine, participants wrote down their agreement level with each statement. This thesis examines neural alignment between the SDP supporters by finding inter-subject correlation patterns from the MEG data with multi-set canonical correlation analysis (MCCA). The dataset of each participant is first reduced in dimensionality with principal component analysis (PCA). Then, MCCA finds the canonical components that capture most correlation between the transformed datasets. Lastly, the best MCCA component is compared with the behavioural attitude measures collected before and during the intervention. The data from manipulation and control groups is analysed separately and the results are compared between the groups. The results from the mathematical analysis indicate no significant correlation between neural alignment and behavioural measures of attitude change. Neural alignment does not differ significantly between the manipulation and control groups. Small sample size (22 participants) decreases the significance of these results. The polarisation level between the parties chosen for the intervention may not match the requirement for the paradoxical thinking method. The statements in the manipulation group may be too far from the political inclination of the participants. The results present a requirement for further research on the neural basis of the paradoxical thinking intervention.
  • How Artificial Intelligence Modifies the Functionality of Smart Homes
    (2024-09-06) Lavonen, Leila
     Sähkötekniikan korkeakoulu | Bachelor's thesis
    The purpose of smart homes is to make living more efficient by facilitating everyday tasks, leaving more time for other areas of interest. Communication in smart homes is simpler with AI through hand gestures, voice, and text commands. AI together with Machine Learning (ML) can automate smart home functions, such as lights and temperature, by analysing and learning from the resident's behaviour. Thus, the resident does not have to adjust the values manually, and over time, the smart home system learns to anticipate the resident's intentions. This Bachelor's thesis is a literature review that discusses how Artificial Intelligence (AI) shapes the functionality of smart homes and what are the advantages and disadvantages of these changes. The material mainly consists of scientific studies, articles, and conference publications. This thesis presents the concept of smart homes, as well as their evolution and qualities that need further development. In addition, this thesis explores the interaction between AI and humans, and AI’s characteristics and future in smart homes. The aim of the thesis is to determine the effects of AI on smart homes and their residents. This thesis examines AI in smart homes on a general level, looking at the effects from different perspectives. This thesis pays little to no attention to healthcare and entertainment systems in research and instead concentrates on building maintenance. In the future, AI-driven smart homes could potentially operate completely autonomously, while being personalised, connected, and sustainable. These features can enable a fully automated system that could interact directly with the user's brain as well as smart cities. However, the adaptation of smart homes is an issue that requires improvements that are achievable by inspecting different interaction modes and fixing usability issues. Thus, fewer and fewer users would forget the smart features of their homes and abandon them.