The last few years have seen the start of a unique change in microphones for consumer devices such as smartphones or tablets. Almost all analogue capacitive microphones are being replaced by digital silicon microphones or MEMS microphones. MEMS microphones perform differently to conventional analogue microphones. Their greatest disadvantage is significantly increased self-noise or decreased SNR, while their most significant benefits are ease of design and manufacturing and improved sensitivity matching. This thesis presents research on speech processing, comparing conventional analogue microphones with the newly available digital MEMS microphones. Specifically, voice activity detection, speaker diarisation (who spoke when), speech separation and speech recognition are looked at in detail. In order to carry out this research different microphone arrays were built using digital MEMS microphones and corpora were recorded to test existing algorithms and devise new ones. Some corpora that were created for the purpose of this research will be released to the public in 2013. It was found that the most commonly used VAD algorithm in current state-of-theart diarisation systems is not the best-performing one, i.e. MLP-based voice activity detection consistently outperforms the more frequently used GMM-HMM-based VAD schemes. In addition, an algorithm was derived that can determine the number of active speakers in a meeting recording given audio data from a microphone array of known geometry, leading to improved diarisation results. Finally, speech separation experiments were carried out using different post-filtering algorithms, matching or exceeding current state-of-the art results. The performance of the algorithms and methods presented in this thesis was verified by comparing their output using speech recognition tools and simple MLLR adaptation and the results are presented as word error rates, an easily comprehensible scale. To summarise, using speech recognition and speech separation experiments, this thesis demonstrates that the significantly reduced SNR of the MEMS microphone can be compensated for with well established adaptation techniques such as MLLR. MEMS microphones do not affect voice activity detection and speaker diarisation performance.

Since the micro-electro mechanical system microphone, MEMS microphone, has the advantages of superior sound quality, low power consumption, higher temperature resistance and anti-noise ability in used. The researchers therefore have studied the functions of MEMS microphone since 1980s. The MEMS microphones is applied as the part of 3G mobile phone in the market. Though the functions of microphone are improved by manufacturing process technique and new material designed, this study tends to provide a new, low-cost and rapid design idea to gain the performance in chamber of microphone. Taguchi method and BEASY software, which is boundary element method, are combined to evaluate the results of the design in sound field. Taguchi method is a famous method in industrial design to find out relations between system parameters and chamber size. BEASY is a tool for sound field analysis in the research. The result from Taguchi method appears the sound pressure level gain about 2.2 dB to 2.4 dB due to the change of microphone chamber size only. It is also interested in studying the optimization design for position of microphone. It is displayed that the location of port is closer to the boundary of chip will also increase about 0.3 dB to 0.6dB sound pressure level in sound field. The higher frequency of sound source will also create larger sound pressure level at two corners on the port.

To achieve the miniaturization and high performance of the mobile phone, notebook, hearing aid and personal digital assistant (PDA), many researchers focus on the developing a new-type microphone with very small dimension, high quality and low manufacturing cost utilizing MEMS technology. By using the surface and bulk micromachining technologies, this thesis designed and fabricated a capacitive MEMS microphone with a polyimide bcakplate microstructure. The main processing steps adopted in this study include five photolithoghaphies and seven thin-film depositions. A MEMS-based microphone with an only 2¡Ñ2 mm2 sensing area of the floating Si3N4/Poly-Si/Si3N4 membrane and a 2 £gm-height gap distance between the top and bottom electrodes was implemented and characterized. Measured in a special isolated-box and under 1 kHz audio frequency, a -60.3 dB/Pa sensitivity (deducted the 22.6 dB output gain of the pre-amplifier) and a 51 dB signal to noise ratio (SNR) of the implemented MEMS microphone can be obtained as the biasing voltage only about 3 volts. The very low driving voltage, moderate SNR and sensitivity demonstrated in this work keep abreast with the results of many outstanding research laboratories in the world.

This thesis studies how the mechanical design of a surveillance camera affects the acoustic performance, mainly in terms of the frequency response within the human hearing range. During the project, the mechanical characteristics that affect frequency response were investigated by measuring the camera’s audio behavior in an anechoic chamber. A theoretical and adaptable acoustic model was built in COMSOL to simulate the frequency response of the sound path. Measurement and simulation results were compared to identify critical aspects of the mechanical design and adjust accordingly for better acoustic performance.

Voice commanded systems (VCS) have been proved to be vulnerable to signal in- jections mimicking voice commands and explored security flaws in market avail- able products for the time of each respective study. Signal injection caused with the help of amplitude modulated ultrasonic waves (being known as DolphinAt- tacks - DA) were proved to work on several such devices in 2017. In 2019, another study were also successful in achieving signal injections using modulated laser also known as LightCommands (LC). This thesis has investigated the occurring circumstances which enables such injections. Simulations and laboratory trials have shown a thermoacoustic origin enabling LC to be injected and the response differs with respect to microphones physical size. DA utilizes the non-linearity of microphones and more linear microphones have indeed been shown to withstand DAs better and physical parameters have been shown to indicate how DA may be optimized for successful injections. The results have been used to provide ideas on how a VCS system can be designed to be more resilient. / Röststyrda System har visat sig vara sårbara mot signalinjektioner som härmar röstkommandon och utnyttjar kryphål hos produkter som fanns på marknaden i samtid när studierna som först tog upp kryphålen publicerades. Signalinjek- toner inducerade med hjälp av amplitudmodulerat ultraljud (känt som Dolphi- nAttacks - DA) bevisades fungera på flertalet enheter år 2017. 2019 visade en annan studie framgång med signalinjektion genom modulerad laser, även känt som LightCommands (LC). Detta examensarbete har utrett de bakomliggande faktorer som möjliggör sådana injektioner. Simuleringar och laboratorieexperi- ment har visat att termoakustiska effekter möjliggör LC med resultat som beror på mikrofoners fysiska storlek. DA nyttjar ickelinjäritet hos mikrofoner och linjä- rare mikrofoner har visat sig stå emot DA bättre och det har visat sig att DA kan optimeras för bättre lyckade injektioner. Resultaten har används för att bidra till idéer och resonemang från föregående studier på hur lösningar mot LC och DA skulle kunna implementeras och göra mikrofoner och dess tillhörande system tåligare mot sådana angrepp.

Thesis discusses the possibility of using MEMS microphones in measuring systems. Describes the characteristics of MEMS components and shows possible realization of analog to digital signal convertor when a microphone with analog output is used. Design of the amplifier is made with respect to low noise and low power consumption. Also is shown the possibility of using antialliasing filter as microphone frequency response correction at the same time.

<p>Condition monitoring supported with artificial intelligence, cloud computing, and industrial internet of things (IIoT) technologies increases the feasibility of predictive maintenance (PdM). However, the cost of traditional sensors, data acquisition systems, and the information technology expert knowledge required to inform and implement PdM challenge the industry. This thesis proposes a hybrid condition monitoring system (CMS) architecture consisting of a distributed, low-cost IIoT-sensor solution. The CMS uses micro-electro-mechanical system (MEMS) microphones for data acquisition, edge computing for signal preprocessing, and cloud computing, including artificial neural networks (ANN) for higher-level information processing. The higher-level information processing includes condition detection and time-based prediction capabilities to inform PdM strategies. The system’s feasibility is validated using a testbed for reciprocating linear-motion axes.</p>

This diploma thesis deals with the design and implementation of an electronic module for acoustic detection. The module has the task of detecting a predetermined acoustic signals through them learned classification model. The module is used mainly for security purposes. To identify and classify the proposed model using machine learning techniques. Given the possibility of retraining for a different set of sounds, the module becomes a universal sound detector. With acoustic sound using the digital MEMS microphone, for which it is designed and implemented conversion filter. The resulting system is implemented into firmware microcontroller with real time operating system. The various functions of the system are realized with regard to the possible optimization (less powerful MCU or battery power). The module transmits the detection results of the master station via Ethernet network. In the case of multiple modules connected to the network to create a distributed system, which is designed for precise time synchronization using PTP protocol defined by the IEEE-1588 standard.