Gotowa bibliografia na temat „Microphone based acoustic vector sensor”

One direct advantage to characterizing the acoustic field with a vector sensor is the ability to measure both components of acoustic energy, potential (pressure) and kinetic (particle velocity). While it is well known that the kinetic energy exceeds potential in the near-field of a source, equivalent and opposite imbalances are prevalent in the far-field due to propagation effects. For example, at normal incidence the field exhibits a distinct frequency dependent relationship that depends explicitly on the impedance of the ground reflector, including any internal reflections from sediment layers. In this paper, we examine the measured vector field, both in-air and underwater, with a specific focus the energy balance when a source is directly overhead. While the construction of a neutrally buoyant volume for use in dense media (such as water) is fairly straightforward, the application to airborne acoustics requires extremely lightweight materials. Airborne measurements are provided by an Accelerometer-based Intensity Vector Sensor (AIVS) system, which is constructed from a lightweight MEMS accelerometer and microphone embedded in a spherical, expanded polystyrene (EPS) foam volume.

An acoustic method for simultaneous condition detection, localization, and classification in air-filled pipes is proposed. The contribution of this work is threefold: (1) a microphone array is used to extend the usable acoustic frequency range to estimate the reflection coefficient from blockages and lateral connections; (2) a robust regularization method of sparse representation based on a wavelet basis function is adapted to reduce the background noise in acoustical data; and (3) the wavelet components are used to localize and classify the condition of the pipe. The microphone array and sparse representation method enhance the acoustical signal reflected from blockages and lateral connections and suppress unwanted higher-order modes. Based on the sparse representation results, higher-level wavelet functions representing the impulse response are used to localize the position of the sensor corresponding to a blockage or lateral connection with higher spatial resolution. It is shown that the wavelet components can be used to train and to test a support vector machine (SVM) classifier for the condition identification more accurately than with a time domain SVM classifier. This work paves the way for the development of simultaneous condition classification and localization methods to be deployed on autonomous robots working in buried pipes.

Irregularities in microphone distribution enrich the diversity of spatial differences to decorrelate interferences from the beamforming target. However, the large degrees of freedom of irregular placements make it difficult to analyse and optimize array performance. This article proposes fast and feasible optimal irregular array design methods with improved beamforming performance for human speech. Important geometric features are extracted to be used as the input vector of the neural network structure to determine the optimal irregular arrangements of sensors. In addition, a hyperbola design method is proposed to directly cluster microphones in the hyperbola areas to produce rich differential distance entropies and yield significant signal-to-noise ratio improvements. These methods can be easily applied to guide non-computer-aided optimal irregular array designs for human speech in acoustic scenes such as immersive cocktail party environments.

A method of determining the road surface state based on the sound intensity analysis is presented. The proposed method is designed for the passive, non-contact assessment of a road surface state, especially in dry/wet conditions. The proposed method is intended for monitoring stations utilizing low-cost hardware. The road surface state is determined from the analysis of sound intensity emitted by road vehicles passing by the sensor and recorded with an acoustic vector sensor (AVS). A frequency domain sound intensity analysis included spatial filtering to reduce the environmental interference using the designed amplitude and phase correction algorithms. A test installation in a real-world scenario, consisting of a small AVS constructed from MEMS microphones and a state-of-art optic-based sensor, was used to evaluate the proposed method. A dataset representing many road vehicles moving at different speeds through the observed road section in varying conditions (dry and wet surface) was collected. Compared with the reference data, an evaluation of the proposed method and its accuracy in determining the road surface state is presented and discussed.The project has been subsidized by the Polish National Centre for Research and Development (NCBR) from the European Regional Development Fund No. POIR.04.01.04/2019 entitled: INFOLIGHT—“Cloud-based lighting system for smart cities.”

In this work, a hybrid radio frequency (RF)- and acoustic-based activity recognition system was developed to demonstrate the advantage of combining two non-invasive sensors in Human Activity Recognition (HAR) systems and smart assisted living. We used a hybrid approach, employing RF and acoustic signals to recognize falling, walking, sitting on a chair, and standing up from a chair. To our knowledge, this is the first work that attempts to use a mixture of RF and passive acoustic signals for Human Activity Recognition purposes. We conducted experiments in the lab environment using a Vector Network Analyzer measuring the 2.4 GHz frequency band and a microphone array. After recording data, we extracted the Mel-spectrogram feature of the audio data and the Doppler shift feature of the RF measurements. We fed these features to six classification algorithms. Our result shows that using a hybrid acoustic- and radio-based method increases the accuracy of recognition compared to just using only one kind of sensory data and shows the possibility of expanding for a variety of other different activities that can be recognized. We demonstrate that by using a hybrid method, the recognition accuracy increases in all classification algorithms. Among these classifiers, five of them achieve over 98% recognition accuracy.

Acoustic dyadic sensors (ADSs) are a new type of acoustic sensor with higher directivity than microphones and acoustic vector sensors, which has great application potential in the fields of sound source localization and noise cancellation. However, the high directivity of an ADS is seriously affected by the mismatches between its sensitive units. In this article, (1) a theoretical model of mixed mismatches was established based on the finite-difference approximation model of uniaxial acoustic particle velocity gradient and its ability to reflect the actual mismatches was proven by the comparison of theoretical and experimental directivity beam patterns of an actual ADS based on MEMS thermal particle velocity sensors. (2) Additionally, a quantitative analysis method based on directivity beam pattern was proposed to easily estimate the specific magnitude of the mismatches, which was proven to be useful for the design of ADSs to estimate the magnitudes of different mismatches of an actual ADS. (3) Moreover, a correction algorithm based on the theoretical model of mixed mismatches and quantitative analysis method was successfully demonstrated to correct several groups of simulated and measured beam patterns with mixed mismatches.

Enhancing speech captured by distant microphones is a challenging task. In this study, we investigate the multichannel signal properties of the single acoustic vector sensor (AVS) to obtain the inter-sensor data ratio (ISDR) model in the time-frequency (TF) domain. Then, the monotone functions describing the relationship between the ISDRs and the direction of arrival (DOA) of the target speaker are derived. For the target speech enhancement (SE) task, the DOA of the target speaker is given, and the ISDRs are calculated. Hence, the TF components dominated by the target speech are extracted with high probability using the established monotone functions, and then, a nonlinear soft mask of the target speech is generated. As a result, a masking-based speech enhancement method is developed, which is termed the AVS-SMASK method. Extensive experiments with simulated data and recorded data have been carried out to validate the effectiveness of our proposed AVS-SMASK method in terms of suppressing spatial speech interferences and reducing the adverse impact of the additive background noise while maintaining less speech distortion. Moreover, our AVS-SMASK method is computationally inexpensive, and the AVS is of a small physical size. These merits are favorable to many applications, such as robot auditory systems.

Once a year, in many parts of the country, transient acoustic sources light up the sky. A fourth-of-July fireworks display provides an excellent opportunity for exploring array-based detection and direction-of-arrival determination. Multiple-boom events are difficult to process; however, less-frequent single-boom events are often sufficiently isolated in time for unambiguous interpretation. Cross-correlations between channels of a four-microphone array provide time delays for angle-of-arrival determination and measures for event detection. The distinct time-domain wave shapes and high signal-to-noise enable straightforward checks of the cross-correlation delays. In addition, a chirp sequence broadcast from a known location provides a ground-truth measurement and opportunities to implement matched filtering and envelope processing. Detection strategies can include received power, rate of change of power, normalized cross-correlation coefficient, or Fisher F-statistic. A least-squares fit by plane-wave approximation gives an estimate of the actual slowness vector, which, in turn, permits estimation of error and of local sound speed. Conversion of the slowness vector to arrival azimuth and elevation angles provides a simple exercise in 3D interpretation. The homework exercise can be as simple as determination of inter-sensor time delays or as complicated as demonstrating a scheme for automatic detection and interpretation.

School bullying is a common social problem, which affects children both mentally and physically, making the prevention of bullying a timeless topic all over the world. This paper proposes a method for detecting bullying in school based on activity recognition and speech emotion recognition. In this method, motion and voice data are gathered by movement sensors and a microphone, followed by extraction of a set of motion and audio features to distinguish bullying incidents from daily life events. Among extracted motion features are both time-domain and frequency-domain features, while audio features are computed with classical MFCCs. Feature selection is implemented using the wrapper approach. At the next stage, these motion and audio features are merged to form combined feature vectors for classification, and LDA is used for further dimension reduction. A BPNN is trained to recognize bullying activities and distinguish them from normal daily life activities. The authors also propose an action transition detection method to reduce computational complexity for practical use. Thus, the bullying detection algorithm will only run, when an action transition event has been detected. Simulation results show that the combined motion-audio feature vector outperforms separate motion features and acoustic features, achieving an accuracy of 82.4% and a precision of 92.2%. Moreover, with the action transition method, the computation cost can be reduced by half.

MEMS microphones have been a research topic for the last two and half decades. The state-of-the-art comprises surface mount MEMS microphones in laptops, mobile phones and tablets, etc. The popularity and the commercial success of MEMS microphones is largely due to the steep cost reduction in manufacturing afforded by the mass scale production with microfabrication technology. The current MEMS microphones are de-signed along the lines of traditional microphones that use capacitive transduction with or without permanent charge (electret type microphones use permanent charge of their sensor element). These microphones offer high sensitivity, stability and reasonably at frequency response while reducing the overall size and energy consumption by exploiting MEMS technology. Conceptually, microphones are simple transducers that use a membrane or diaphragm as a mechanical structure which deflects elastically in response to the incident acoustic pressure. This dynamic deflection is converted into an electrical signal using an appropriate transduction technique. The most popular transduction technique used for this application is capacitive, where an elastic diaphragm forms one of the two parallel plates of a capacitor, the fixed substrate or the base plate being the other one. Thus, there are basically two main elements in a microphone { the elastic membrane as a mechanical element, and the transduction technique as the electrical element. In this thesis, we propose and study novel design for both these elements. In the mechanical element, we propose a simple topological change by introducing slits in the membrane along its periphery to enhance the mechanical sensitivity. This simple change, however, has significant impact on the microphone design, performance and its eventual cost. Introduction of slits in the membrane makes the geometry of the structural element non-trivial for response analysis. We devote considerable effort in devising appropriate modeling techniques for deriving lumped parameters that are then used for simulating the system response. For transduction, we propose and study an FET (Field Effect Transistor) coupled micro-phone design where the elastic diaphragm is used as the moving (suspended) gate of an FET and the gate deflection modulated drain current is used in the subthreshold regime of operation as the output signal of the microphone. This design is explored in detail with respect to various design parameters in order to enhance the electrical sensitivity. Both proposed changes in the microphone design are motivated by the possibilities that the microfabrication technology offers. In fact, the design proposed here requires further developments in MEMS technology for reliably creating gaps of 50-100 nm between the substrate and a large 2D structure of the order of a few hundred microns in diameter. In the First part of the thesis, we present detailed simulations of acoustic and squeeze lm domain to understand the effect slits could bring upon the behaviour of the device as a microphone. Since the geometry is nontrivial, we resort to Finite element simulations using commercial packages such as COMSOL Multiphysics and ANSYS in the structural, acoustic and Fluid-structure domains to analyze the behaviour of a microphone which has top plate with nontrivial geometry. On the simulated Finite element data, we conduct low and high frequency limit analysis to extract expressions for the lumped parameters. This technique is well known in acoustics. We borrow this technique of curve Fitting from the acoustics domain and apply it in modified form into the squeeze lm domain. The dynamic behaviour of the entire device is then simulated using the extracted parameters. This helps to simulate the microphone behaviour either as a receiver or as a transmitter. The designed device is fabricated using MEMSCAP PolyMUMPS process (a foundry Polysilicon surface micromachining process). We conduct vibrometer (electrostatic ex-citation) and acoustic characterization. We also study the feasibility of a microphone with slits and the issues involved. The effect of the two dissipation modes (acoustic and squeeze lm ) are quantified with the experimentally determined quality factor. The experimentally measured values are: Resonance is 488 kHz (experimentally determined), low frequency roll-off is 796 Hz (theoretical value) and is 780 Hz as obtained by electrical characterization. The first part of this thesis focusses on developing a comprehensive understanding of the effect of slits on the performance of a MEMS microphone. The presence of slits near the circumference of the clamped plate cause reduction in its rigidity. This leads to an increase in the sensitivity of the device. Slits also cause pressure equalization between the top and bottom of the diaphragm if the incoming sound is at relatively low frequencies. At this frequency, also known as the lower cutoff frequency, the microphone's response starts dropping. The presence of slits also changes the radiation impedance of the plate as well as the squeeze lm damping below the plate. The useful bandwidth of the microphone changes as a consequence. The cavity formed between the top plate and the bottom fixed substrate increases the stiffness of the device significantly due to compression of the trapped air. This effect is more pronounced here because unlike the existing capacitive MEMS microphones, there is no backchamber in the device fabricated here. In the second part of the thesis, we present a novel subthreshold biased FET based MEMS microphone. This biasing of the transistor in the subthreshold region (also called as the OFF-region) offers higher sensitivity as compared to the above threshold region (also called as the ON-region) biasing. This is due to the exponentially varying current with change in the bias voltage in the OFF-region as compared to the quadratic variation in the ON-region. Detailed simulations are done to predict the behaviour of the device. A lumped parameter model of the mechanical domain is coupled with the drain current equations to predict the device behaviour in response to the deflection of the moving gate. From the simulations, we predict that the proposed biasing offers a device sensitive to even sub-nanometer deflection of the flexible gate. As a proof of concept, we fabricate fixed-fixed beams which utilize CMOS-MEMS fabrication. The process involves six lithography steps which involve two CMOS and the remaining MEMS fabrication. The fabricated beams are mechanically characterized for resonance. Further, we carry out electrical characterization for I-V (current-voltage) characteristics. The second part of the thesis focusses on a novel biasing method which circumvents the need of signal conditioning circuitry needed in a capacitive based transduction due to inbuilt amplification. Extensive simulations with equivalent circuit has been carried out to determine the increased sensitivity and the role of various design variables.

In this paper, a new real-time approach for audio recognition of waterbird species in noisy environments, based on a Texas Instruments DSP, i.e. TMS320C6713 is proposed. For noise estimation in noisy water bird's sound, a tonal region detector (TRD) using a sigmoid function is introduced. This method offers flexibility since the slope and the mean of the sigmoid function can be adapted autonomously for a better trade-off between noise overvaluation and undervaluation. Then, the features Mel Frequency Cepstral Coefficients post processed by Spectral Subtraction (MFCC-SS) were extracted for classification using Support Vector Machine classifier. A development of the Simulink analysis models of classic MFCC and MFCC-SS is described. The audio recognition system is implemented in real time by loading the created models in DSP board, after being converted to target C code using Code Composer Studio. Experimental results demonstrate that the proposed TRD-MFCC-SS feature is highly effective and performs satisfactorily compared to conventional MFCC feature, especially in complex environment.