Browsing by Author "Panahi, Issa M.S."
Now showing 1 - 10 of 10
- Results Per Page
- Sort Options
Item Acoustic Feedback Cancellation Techniques for Hearing Aid Applications in Noisy Environments(2018-05) Mishra, Parth; 0000-0003-2173-3070 (Mishra, P); Panahi, Issa M.S.Acoustic feedback cancellation is a problem of great interest to researchers in Hearing Aid Device (HAD) applications. Acoustic feedback effect arises due to the acoustic leakage from the loudspeaker to the input microphone. It interferes with and degrades the receiving speech/sound signals to the HADs. It is inadvertently picked up by the HAD microphone and re-amplified which can lead to the appearance of undesirable irritating “whistling” and “screeching” sounds, often referred to as howling effect. Adaptive Feedback Cancellation (AFC) techniques are widely used to eliminate the undesired acoustic feedback effect arising in the HADs due to the coupling between the speaker and the microphone of the HAD. However, the closed loop formed by the acoustic feedback path makes the spectrally colored desired speech signal highly correlated with the feedback signal leading to biased feedback path estimate. Thus, for efficient feedback cancellation, decorrelating techniques are utilized with AFC methods. This thesis proposes two novel methods which involve prefiltering and noise injection as decorrelating techniques to eliminate the acoustic feedback effect in the HADs in presence of noisy environment. These methods involve utilization of a computationally efficient Spectral Flux feature-based voice activity detector which supervises the feedback cancellation process. The proposed algorithm’s performance is objectively evaluated for realistic noisy conditions. The simulations performed for the proposed methods support the effectiveness and better performance of the proposed algorithms for the HAD applications over several earlier methods. In this work, we also propose a novel low latency smartphone-based application that demonstrates the real-time operation (i.e. frame by frame operation of input data) to cancel the negative effects of acoustic feedback. With the proposed application, we can estimate the transfer function between speaker and microphone of the smartphone in the changing room acoustics making the feedback cancellation very effective.Item Control and Diagnosis of Permanent Magnet Motors(2022-12-01T06:00:00.000Z) Li, Chen; Akin, Bilal; Fischetti, Massimo; Nourani, Mehrdad; Gardner, Matthew; Panahi, Issa M.S.Permanent magnet synchronous machines (PMSMs) have been deployed widely in recent years due to their inherent features such as high efficiency and high power density. Thanks to these merits, they are used in various applications including renewable systems, transportation, and automation systems to name a few. Considering the safety, reliability and system efficiency, these systems should be monitored and maintained carefully to avoid accidents or operation losses. Therefore, developing reliable fault diagnosis and post-fault control tools is essential. To design high performance fault diagnosis and post-fault-control algorithms, drive system requires accurate electrical parameter information and highly precise current and voltage feedback measurement. This dissertation presents a comprehensive monitoring and diagnosis techniques for inter-turn short-circuit (ITSC) fault which can accurately estimate short circuit current based on well calculated motor parameters and compensated feedback signals. Inductances are crucial parameters for electric motors. It is essential to obtain accurate electrical parameter information of a permanent magnet synchronous machines (PMSM) for high performance controller and observer design. Due to the saturation of magnetic elements, the inductances of permanent magnet motors change depending on the operating points. To solve this problem, the inductance model is analyzed carefully and an improved absolute inductance estimation is proposed based on high-frequency current signal injection. Secondly, the compensation of feedback current signal in drive system is studied. To obtain highly accurate current signal, sigma delta ADCs (SD-ADCs) are used to improve sensing resolution and signal-to-noise ratio. However, the additional latency caused by the use of digital SINC filters for demodulation becomes remarkable and degrades the performance of dynamic systems. The effects of latency on the system bandwidth and gain/phase are analyzed in detail and a Kalman filter-based latency compensation algorithm and compensation strategies are proposed. In the last of the research, a comprehensive analysis for ITSC online short circuit current estimation is proposed. By analyzing the voltage harmonics caused by ITSC fault under stationary and rotating reference frame, the relationship between voltage harmonics, short circuit current and reference voltage is established. According to this relationship, a harmonic analysis-based online short circuit estimation method is proposed. Then the proposed short circuit estimation technique is adopted for PMSM healthy condition monitoring and post-fault-control algorithm design.Item Exploring Machine Learning for Automated Diagnosis in the Presence of Missing and Corrupted Data(December 2023) Apalak, Merve 1991-; Hamlen, Kevin; Kiasaleh, Kamran; Panahi, Issa M.S.; Nourani, Mehrdad; Tamil, LakshmanThe advances in electronic health records (EHRs) and machine learning (ML) algorithms have brought a new perspective to biomedical sciences and medical practice. This has enabled and improved research on automated diagnostics, data-driven disease categorization and personalized treatments. Researchers and healthcare providers have already welcomed the recent advancements. However, the transition into practice happens gradually due to the challenges in the field. Even though patients’ records are being recorded as EHRs in hospital systems, it is necessary to thoroughly analyze, process, and annotate the data before employing it for prediction problems. Each chapter of this dissertation highlights the obstacles that have emerged when implementing learning models on clinical data, particularly in the application domain of sepsis prediction. The first part of the thesis proposes a strategy to alleviate the poor prediction performances caused by irregularly-spaced and incompletely observed databases. The proposed method employs Conditional Generative Adversarial Networks (GANs), with Long Short-Term Memory (LSTM) networks serving as both the generator and discriminator, conditioned on class labels. Experimental results show that while the proposed framework profitably identifies long-term temporal dependencies and exploits the missing patterns, it also delivers highly notable performance results. The second part of this dissertation focuses on non-invasive, computationally efficient, and continuous patient monitoring in intensive care units (ICUs) using single-lead electrocardiogram (ECG) signals for the early prediction of sepsis. We develop a continuous early sepsis detection algorithm utilizing two databases; in particular, the Medical Information Mart for Intensive Care (MIMIC-III) Clinical Database and MIMIC-Waveform Database. We carry out a systematic approach to selecting ECG segments with superior quality that are recorded in highly dynamic ICU environments. Moreover, since we are approaching the early sepsis prediction as a supervised time series classification, we evaluate the model performance by implementing Temporal Convolutional Networks (TCN). It is discovered that hear rate variability (HRV) demonstrates considerable decelerations for sepsis patients, and the HRV characteristics of adults can be a valuable indicator for continuous sepsis monitoring in an ICU. Finally, this research work adds to the field of early sepsis detection by providing an annotated continuous waveform database from the MIMIC-Waveform Database, which is made accessible to the public.Item Real-Time Single and Dual-Channel Speech Enhancement on Edge Devices for Hearing Applications(2021-04-26) Shankar, Nikhil; Panahi, Issa M.S.Speech Enhancement (SE) is an important module in the signal processing pipeline for hearing applications and it helps enhance the comfort of listening. Many single and dualmicrophone SE techniques have been developed by researchers over the last few decades. In this thesis, novel single and dual-channel SE techniques have been proposed and are implemented on edge devices as an assistive tool for hearing applications. The smartphone is considered as the processing platform for real-time implementation and testing. In this work, both statistical signal processing and deep learning algorithms are proposed for SE. Firstly, we compare different two-channel beamformers for SE. Later, the Minimum Variance Distortionless Response (MVDR) beamformer assisted by a voice activity detector (VAD) is used as a Signal to Noise Ratio (SNR) booster for the SE method. Deep neural network architectures comprising of convolutional neural network (CNN) and recurrent neural network (RNN) layers are proposed in this thesis for real-time SE. Finally to filter out background noise, the SE gain estimation for noisy speech mixture is smoothed along the frequency axis by a Mel filter-bank, resulting in a Mel-warped frequency-domain gain estimation. In comparison with existing SE methods, objective assessment and subjective results of the developed methods indicate substantial improvements in speech quality and intelligibility.Item Real-Time Speech Processing Algorithms for Smartphone Based Hearing Aid Applications(2021-04-26) Shreedhar Bhat, Gautam; Panahi, Issa M.S.Signal processing algorithms are extensively used in hearing aid applications to improve the quality and intelligibility of speech. The hearing aid device (HAD) signal processing pipeline consists of several key modules that help to improve the perception for hearing-impaired listeners. In this dissertation, novel speech processing algorithms have been proposed that can be used in smartphone-based hearing aid (HA) setup. Every chapter of this dissertation concentrates on the individual modules of the signal processing pipeline in HADs. The first algorithm is developed for speech enhancement (SE) to suppress the background noise. A voice activity detector (VAD) to classify the incoming signal into speech or noise is developed. Signal enhancement techniques like blind source separation and dereverberation are developed. The algorithms are developed using conventional and supervised learning techniques. Objective and subjective evaluations are conducted for each of the proposed techniques to show substantial improvements in speech quality and intelligibility.Item Signal Processing Algorithms for Smartphone-Based Hearing Aid Platform; Applications and Clinical Testing(2022-05-01T05:00:00.000Z) Tokgoz, Serkan; Panahi, Issa M.S.; Hoyt, Kenneth; Thibodeau, Linda M.; Kiasaleh, Kamran; Tamil, LakshmanDigital signal processing algorithms are widely utilized in hearing aid applications to im- prove the quality of speech. The signal processing pipeline for speech involves several crucial components that enhance hearing-impaired people’s listening. This thesis covers the devel- opment of novel methods that can be used in the speech processing pipeline and clinical testing. Each chapter of the dissertation focuses on the components of the speech process- ing pipeline for smartphone-based hearing aid setup. The first algorithm, speech source localization (SSL), which identifies the direction of the talker of interest using multiple mi- crophones, is discussed. A speaker identification method is proposed as an assistive system to the pipeline, and it can be used to boost the overall system’s performance. A clinical testing system is developed to evaluate the new signal processing algorithms. An approach is developed to use multiple microphones of iPhone simultaneously for real-time low-latency audio applications. Real-time integration of several signal processing modules that appear in digital hearing aids is developed as smartphone apps. Subjective evaluations are conducted for the proposed methods to show noticeable improvements and compared with state of the art methods. Additionally, the implementation of the proposed methods is explained on smartphones and computers.Item Smartphone Based Multi-Channel Dynamic-Range Compression for Hearing Aid Research and Noise-Robust Speech Source Localization Using Microphone Arrays(2019-05) Hao, Yiya; Panahi, Issa M.S.Dynamic-range compression (DRC) is one technique of audio signal processing which maps the wide dynamic range of audio into the narrow range. In most of hearing aid devices (HADs), DRC plays as an important role to help hearing impaired people to listen to the speech with better quality and intelligibility. Compared to single-channel DRC, multi-channel DRC additionally improves the audio quality and intelligibility by setting appropriate adjustment parameters for every frequency band. Since most HADs are costly and unaffordable, fitting multi-channel DRC into smartphones becomes an outstanding solution since most people own smartphones. In this dissertation, three types of multi-channel DRCs has been proposed including their real-time implementations on smartphones. Speech Source Localization (SSL) (or Direction of Arrival (DOA) estimation) is another topic in this dissertation, which identifies the direction of the talker of interest in a noisy environment using multiple fixed microphones (known as a Microphone Array). This Proposal mainly covers only DRC, DOA and real-time implementationItem Sound Source Localization for Improving Hearing Aid Studies Using Mobile Platforms(December 2021) Kucuk, Abdullah; Panahi, Issa M.S.; Ntafos, Simeon; Busso, Carlos; Nourani, Mehrdad; Nosratinia, AriaMicrophone array is one of the powerful techniques that enables to apply effective signal processing algorithms to systems. One of the critical application areas of microphone array is sound source localization (SSL), which refers to identify the speaker of interest using a microphone array. SSL can be used as a preprocessing technique to boost up the entire system efficiency. Recent studies show that smartphones can be an efficient assistive device for hearing aid devices because of smartphones’ powerful hardware and software components. Also, Deep Learning (DL) has shown a considerable performance increase in audio signal processing. DL based SSL using the direction of arrival estimation (DOA) methods for two and eight microphone array structures and the distance estimation methods using a single microphone are proposed in this work. The performance of the proposed methods are evaluated in several realistic noisy conditions, reverberations using real-recorded data. Another contribution of this work is to present real-time implementations of the DL based methods on edges devices, i.e., smartphones, tablets.Item Stray Magnetic Flux Based Condition Monitoring Techniques for Permanent Magnet Synchronous Motors(2021-08-01T05:00:00.000Z) Gurusamy, Vigneshwaran; Akin, Bilal; Yen, I-Ling; Panahi, Issa M.S.; Gardner, Matthew; Nourani, MehrdadPermanent Magnet Synchronous Motors (PMSM) are widely used in various applications from home appliances to electric vehicles due to its advantages like high efficiency, high power density etc. The market share of this type of motor is increasing significantly in the recent decade. In order to ensure their reliable operation in safety critical system, condition monitoring tools are essential. The effective fault diagnosis and condition monitoring system reduces downtime and unplanned maintenance which saves both time and money. Though several motor attributes like current, vibration, temperature etc. can be used for diagnosis of health condition of an electric motor, stray magnetic flux around the motor housing has rich information about the health status of PM motors due to the use of high energy magnets. This dissertation presents condition monitoring techniques which are developed to address three major issues in PMSM. Stator fault is the most common electrical fault which begins as an inter-turn failure. It is essential to identify this failure at its initial stage to prevent the catastrophic damage caused by stator fault. It is shown that the fundamental component of stray magnetic flux may result in fault negative error under certain operating conditions. To overcome this problem, third harmonic component is analyzed comprehensively and proposed as a reliable fault metric throughout the entire range of operation to detect and locate the fault. Secondly, the common mechanical fault in electrical motors, bearing damage is studied. It is shown that the magnetic field of PM motors exhibit asymmetry due to the manufacturing imperfections of magnet. A stray magnetic flux-based bearing fault detection method is proposed by leveraging this asymmetry. The proposed method is analyzed extensively in both simulations and experiments. A comparative study between motor current based detection and the proposed method is also performed. In the last part of the research, a stray magnetic flux based the temperature estimation of permanent magnet is proposed. The stray magnetic flux around the motor can be correlated to the magnet temperature. However, it also is affected by the permeability variation of iron core and magnetic field generated by the stator current. A compensation co-efficient is proposed to compensate the mentioned effects and obtain the permanent magnet component of stray magnetic flux from the measured data. Then the proposed scheme is adopted to estimate the temperature of magnets online under dynamic operating conditions.Item Toward Measuring Tissue-level Mechanical Stresses in Branching Embryonic Epithelia(2022-05-01T05:00:00.000Z) Patil, Lokesh S; Varner, Victor; Panahi, Issa M.S.; Cogan, Stuart; Schmidtke, David; Gleghorn, JasonTree-like networks form the basic architecture for many organs in the body. In the developing embryo, these structures are shaped by a process known as branching morphogenesis, in which a simple epithelial tube is sculpted into a ramified network via a sequence of iterative branching events. Various chemical signaling mechanisms have been implicated, but in the embryonic lung mechanical forces have also been shown to regulate airway branching morphogenesis. However, how these chemical and physical mechanisms coordinate together to sculpt the functional form of lung is still unclear. This has predominantly been because of the highly interdisciplinary nature of the problem and lack of tools and methods to probe the effect of mechanics on tissue growth. One of the ways to probe tissue mechanics is by estimating patterns of mechanical stresses around the tissue. Many different methods have been developed in past two decades for quantifying mechanical stresses, but most are limited to 2D cellular level experiments. Furthermore, majority of these studies assume the underlying material to have a linear response to deformations generated by cells. Biological systems however, especially during lung development exhibit large deformations and material response can become highly non-linear. This dissertation attempts to bridge this gap by creating a Traction Force Microscopy (TFM) pipeline towards estimating patterns of mechanical stresses around branching embryonic epithelia in 3D matrices. To this end, we used mesenchyme free culture assay which has been previously used to study branching morphogenesis of embryonic lung tissues. We then modified this assay for ex vivo culture of isolated embryonic airway epithelial explants by suspending fluorescent microspheres within the surrounding gel. We demonstrate that these beads can be tracked over the course of many hours to generate a spatial deformation field. Using bead tracking data, we show that there is significant inward (towards epithelium) gel movement during events of epithelial branching suggestive of ECM remodeling near epithelium surface. We also used this tracking data to compute spatiotemporal patterns of strain and stress. Mechanical stress computation however requires knowledge of the mechanical properties of underlying substrate and existing data on the mechanical properties of Matrigel are highly inconsistent and limited to linear small-strain measurements. We thus performed multi-axial deformation mechanical testing of Matrigel to characterize the finite-deformation behavior of this extracellular matrix (ECM) material. These mechanical tests were then combined with quantitative measurements of the deformation fields around cultured embryonic airway epithelial explants to estimate the mechanical stresses exerted during airway branching.