Collector optimization for tradeoff between breakdown voltage and cut-off frequency in SiGe HBT

As is well known, there exists a tradeoff between the breakdown voltage BVcEO and the cut-off frequency fT for a standard heterojunction bipolar transistor （HBT）. In this paper, this tradeoff is alleviated by collector doping engineering in the SiGe HBT by utilizing a novel composite of P＋ and N- doping layers inside the collector-base （CB） space-charge region （SCR）. Compared with the single N-type collector, the introduction of the thin P＋ layers provides a reverse electric field weakening the electric field near the CB metallurgical junction without changing the field direction, and the thin N layer further effectively lowers the electric field near the CB metallurgical junction. As a result, the electron temperature near the CB metallurgical junction is lowered, consequently suppressing the impact ionization, thus BVcEO is improved with a slight degradation in fT. The results show that the product of fTXBVcEo is improved from 309.51 GHz.V to 326.35 GHz.V.

Effect of Cut-Off Frequency of Butterworth Filter on Detectability and Contrast of Hot and Cold Regions in Tc-99m SPECT

In SPECT, noise is one of the major limitations that degrade image quality. To suppress the noisy signals in an image, digital filters are most commonly applied. However, in SPECT image reconstruction, selection of an appropriate filter and its functions has always remained a difficult task. In this work an attempt was made to investigate the effects of varying cut-off frequencies and in keeping the order of Butterworth filter constant on detectability and contrast of hot and cold re-gions images. A new insert simulating hot and cold regions which provides similar views in a reconstructed image was placed in the phantom’s cylindrical source tank and imaged. Tc-99m radionuclide was distributed uniformly in the phantom. SPECT data were collected in a 20% energy window centered at 140 keV by a Philips ADAC Forte dual head gamma camera mounted with a LEHR collimator. Images were generated by using the filtered backprojection technique. A Butterworth filter of order 5 with cut-off frequencies 0.35 and 0.45 cycles·cm-1 was applied. Images were examined in terms of hot and cold regions, detectability and contrast. Results show that the hot and cold regions’ detectability and contrast vary with the change of cut-off frequency. With a 0.45 cycles·cm-1 cut-off frequency, a significant enhancement in contrast of cold regions was achieved as compared to a 0.35 cycles·cm-1 cut-off frequency. Furthermore, the detectability of hot and cold regions improved with the use of a 0.45 cycles·cm-1 cut-off frequency. In conclusion, image quality of hot and cold regions affected in a different way with a change of cut-off frequency. Thus, care should be taken in selecting the filter cut-off frequency prior to reconstruction of images;particularly, when both types of regions are expected in the reconstructed image.

Feasibility of applying the lower cut-off frequency for the density radial coverage extension in EAST reflectometry measurement

The extraordinary mode(X-mode)lower cut-off frequency is proposed for use in the reflectometry diagnostic on ITER for the electron density profile measurement,which is a trade-off between extreme plasma parameters and the accessible probing frequency.In contemporary experiments,the lower cutoff frequency can be identified at the probing frequency below the electron cyclotron frequency(f_(ce)) under certain plasma conditions.We provide here,for the first time,the experimental validation of the use of the lower cut-off frequency for the density profiles via the reflectometry measurement on EAST.The corresponding group delay of the lower cut-off frequency evolves continuously with the upper one,revealing a reasonable radial coverage extension of reflectometry measurement toward the plasma core.It is concluded that the lower cut-off frequency can be used as a supplement to the upper one in the density profile inversion process,which is of particular interest in the high magnetic field and/or density discharge to extend the radial coverage of reflectometry measurement.

An Experimental Method to Determine the Cut-Off Frequency of an Acoustical Free Field in a Non-Ideal Environment

The availability of ideal conditions like anechoic chamber to characterize some sound parameters, like sound intensity and sound power necessities the determination of free field and cut off frequency measurements. In this article, full experiment was executed at Wayne State University (Detroit-Michigan), to determine the cut off frequency in all directions;the obtained results showed that the free field can be determined for a specified space. So other tests can take place in this space avoiding regions where reflections and consequently noise can be found. Upon these results tests related to noise abatement in vehicles can be done in such environment.

Analysis and compensation of low frequency characteristics of sensors for vibration testing

In view of the influence and harm of low frequency vibration environment on the structure of spaceflight products,a low frequency dynamic study method for piezoelectric sensor based on the dynamic system of sinusoidal pressure is proposed.This method uses a sinusoidal pressure dynamic system with two-way dual channel import and export synchronization technology to study the low frequency characteristics of a piezoelectric sensor of PCB company,and its lower cut-off frequency is 0.26 Hz.It is also studied that when the frequency of the measured vibration or shock signal is 1-200 kHz,the error range of signal positive pressure action time is 4.87%-0.03%.The dynamic compensation for the low frequency of the vibration sensor is carried out,and the compensation effect is good.

Radio-frequency transistors from millimeter-scale graphene domains

Graphene is a new promising candidate for application in radio-frequency （RF） electronics due to its excellent elec- tronic properties such as ultrahigh carrier mobility, large threshold current density, and high saturation velocity. Recently, much progress has been made in the graphene-based RF field-effect transistors （RF-FETs）. Here we present for the first time the high-performance top-gated RF transistors using millimeter-scale single graphene domain on a SiO2/Si substrate through a conventional microfabrication process. A maximum cut-off frequency of 178 GHz and a peak maximum os- cillation frequency of 35 GHz are achieved in the graphene-domain-based FET with a gate length of 50 nm and 150 nm, respectively. This work shows that the millimeter-scale single graphene domain has great potential applications in RF devices and circuits.

MHz cut-off frequency and permeability mechanism of iron-based soft magnetic composites

The lack of soft magnetic composites with high power density in MHz frequency range has become an obstacle in the efficient operation of the electrical and electronic equipments.Here,a promising method to increase the cut-off frequency of iron-based soft magnetic composites to hundreds of MHz is reported.The cut-off frequency is increased from 10 MHz to 1 GHz by modulating the height of the ring,the distribution of particles,and the particle size.The mechanism of cut-off frequency and permeability is the coherent rotation of domain modulated by inhomogeneous field due to the eddy current effect.An empirical formula for the cut-off frequency in a magnetic ring composed of iron-based particles is established from experimental data.This work provides an effective approach to fabricate soft magnetic composites with a cut-off frequency in hundreds of MHz.

基于准零刚度支撑结构的低频主动隔振平台研究

针对量子科学试验、精密制造等先进应用对亚赫兹微振动的抑制需求,研究基于准零刚度支撑结构和分离式音圈作动器的低频主动隔振技术。首先,给出一种低频主动隔振平台设计方法,考虑准零刚度结构的等效动力学特性,采用牛顿-欧拉方法建立平台的动力学模型;然后,基于所建立动力学模型分析多自由度耦合下平台的扰动传递特性并设计主动隔振算法;最后,研制低频主动隔振平台进行试验验证。试验结果表明,所开发主动隔振平台能够在三个线自由度上实现优于98.3%的微振动抑制效果,在0.1~10.0 Hz内载荷z方向微振动小于1.3×10^(-8) g。

梅尔频率倒谱系数在声带息肉手术前后嗓音分析中的价值研究

目的 本研究拟通过提取患者嗓音中的梅尔频率倒谱系数(MFCC)指标,探讨其在声带息肉手术前后嗓音分析中的临床价值。方法 回顾性分析于2018年1月—2019年8月行声带息肉手术且术前及术后1个月均行嗓音评估的患者41例,男31例,女10例;平均年龄(42.9±11.4)岁。另选取无声嘶且无声带病变的正常受试者21例作为基线对照。使用基于Python编程语言的librosa语音处理包进行MFCC特征提取,分别提取每位患者的MFCC均值,MFCC方差与MFCC标准差,使用配对样本t检验比较声带息肉手术前后上述各MFCC特征的差异。结果 声带息肉患者术后MFCC均值1.25±1.01、MFCC方差561.34±154.98及MFCC标准差21.74±4.03比术前MFCC均值6.81±2.05、MFCC方差1 019.66±295.87及MFCC标准差34.37±6.63显著下降,差异具有统计学意义(t=18.596,P=0.000;t=10.338,P=0.000;t=11.852,P=0.000)。声带息肉组患者术后1个月其MFCC均值、MFCC方差及MFCC标准差与正常受试者相比差异均无统计学意义,表明绝大部分声带息肉患者术后嗓音得到良好的恢复。结论 本研究首次探索了MFCC在声带息肉手术前后嗓音分析中的价值,MFCC各特征可作为评估声带息肉术后嗓音恢复的指标。

基于语音节奏差异的情感识别方法

语音情感识别在金融反欺诈等领域有着重要的应用前景,但是语音情感识别的准确率提升变得越来越困难。现有基于语谱图的语音情感识别等方法难以捕捉节奏差异特征,从而影响识别效果。文中基于语音节奏特征的差异性,提出了能量帧时频融合的语音情感识别方法。其关键是,针对语音中高能量区域进行频谱筛选,以高能语音帧的分布和时频变化来体现个体的语音节奏差异。在此基础上建立基于卷积神经网络(CNN)和循环神经网络(RNN)的情感识别模型,实现对频谱的时域和频域变化特征的提取与融合。在公开数据集IEMOCAP上进行实验,结果表明,该基于语音节奏差异的语音情感识别与基于语谱图的方法相比,在加权准确率WA和非加权准确率UA指标上分别平均提升了1.05%和1.9%;同时也表明个体的语音节奏差异对提升语音情感识别效果具有重要作用。

Comprehensive Analysis of Gender Classification Accuracy across Varied Geographic Regions through the Application of Deep Learning Algorithms to Speech Signals

This article presents an exhaustive comparative investigation into the accuracy of gender identification across diverse geographical regions,employing a deep learning classification algorithm for speech signal analysis.In this study,speech samples are categorized for both training and testing purposes based on their geographical origin.Category 1 comprises speech samples from speakers outside of India,whereas Category 2 comprises live-recorded speech samples from Indian speakers.Testing speech samples are likewise classified into four distinct sets,taking into consideration both geographical origin and the language spoken by the speakers.Significantly,the results indicate a noticeable difference in gender identification accuracy among speakers from different geographical areas.Indian speakers,utilizing 52 Hindi and 26 English phonemes in their speech,demonstrate a notably higher gender identification accuracy of 85.75%compared to those speakers who predominantly use 26 English phonemes in their conversations when the system is trained using speech samples from Indian speakers.The gender identification accuracy of the proposed model reaches 83.20%when the system is trained using speech samples from speakers outside of India.In the analysis of speech signals,Mel Frequency Cepstral Coefficients(MFCCs)serve as relevant features for the speech data.The deep learning classification algorithm utilized in this research is based on a Bidirectional Long Short-Term Memory(BiLSTM)architecture within a Recurrent Neural Network(RNN)model.

基于声纹识别技术的变压器异常状态自动诊断方法

为使监测主机在变压器设备发生异常运行时能够发出预警,实现对变压器异常状态的精准诊断,设计基于声纹识别技术的变压器异常状态自动诊断方法。按照声纹识别技术应用原理,提取变压器异常运转情况下的时域特征、频域特征与梅尔倒谱特征,根据变压器指标权重的赋值结果,计算频率复杂度参量的取值范围,联合相关变量指标,求解频谱能量分布指数表达式,完成基于声纹识别技术的变压器异常状态自动诊断方法的设计。对比实验结果表明,在上述诊断方法作用下,变压器出现异常运行状态时,电压有效值始终不会超过预设警戒值,二者差值也始终小于5.0V,符合及时预警并精准诊断的实际应用需求。

语音语法在VoiceXML中的应用

语音语法是自动语音识别的重要参考依据, 在基于VoiceXML的语音增值业务中应用越来越广泛. 介绍了VoiceXML与语法的关系, 描述了语音识别原理, 推导出了常用的Nuance语法形式逻辑表达式. 在此基础上, 构建了一个语音拨号系统, 并着重阐述了其中语音语法的应用.

基于MFCC与CNN的机械故障声音自动识别

针对机械故障自动识别问题,提出一种结合梅尔频率倒谱系数(Mel Frequency Cepstral Coefficient,MFCC)与一维卷积神经网络(Convolutional Neural Networks,CNN)的机械故障声音自动识别方法,并通过实验验证该方法的有效性。实验结果表明,该方法在机械故障声音识别中具有较高的准确率、精确率及召回率,能够有效识别故障案例。

UFMC类语音系统同步方案的设计与实现

端到端语音加密通信系统中,密文语音数据在移动通信网中可能面临的失真、失码等问题,利用通用滤波多载波(Universal Filtered Multi Carrier,UFMC)技术的类语音调制特性可实现信道的抗压缩传输。为进一步增强其抗干扰性能,克服同步偏差的影响,提出一种UFMC类语音时频联合同步方案。该方案首先将Zadoff⁃Chu序列搭载至UFMC类语音系统的合适载波频段上合成新型类语音同步信号,而后根据UFMC类语音算法滤波器长度应为奇数的特点,构造出具有重复结构的类语音同步信号,使其可凭借自身冗余特性先后完成对时延、频偏的估计。在AMR⁃NB122kb/s声码器模型下的仿真结果表明:所提同步方案能精准定位解调起点,可较好满足UFMC类语音系统同步需求。

A Preliminary Receiver Operating Characteristic Analysis on Voice Handicap Index Results of the Greek Voice-Disordered Patients

Objective(s): The cutting-edge assessment of voice disorders includes objective and subjective methods in the daily clinical practice. The latter assessment is usually performed through the administration of self-reported questionnaires. Voice Handicap Index (VHI) is one of the most widely used tools both in clinical practice and in research level. This tool-questionnaire was employed in this research along with the Voice Evaluation Template (VEF). In turn, the aim of this study was to analyse and produce the cut-off points of VHI for voice-disordered patients in Greece by using Receiver Operating Characteristic Curves (ROC Curves). Methods: Sixty-three participants (40 non-dysphonic and 23 with different types of dysphonia) were classified by ENT (Ear, Nose, and Throat) doctors and SLPs (Speech-Language Pathologists). The Hellenic VHI along with the translated Greek version of the VEF was administered to the subjects of this research. Results: The voice-disordered subjects exhibited higher overall VHI scores (in total and in its 3 subdomains) compared to the control group. Statistical significant differences were found between dysphonic and non-dysphonic participants for all VHI’s construct domains. The cut-off point of VHI total score was estimated at the value of 14.50 (sensitivity: 0.870, 1-specificity: 0.000). Moreover, the cut-off points of the three subdomains were computed as 7.50 for functional (sensitivity: 0.783, 1-specificity: 0.000), 8.50 for physical (sensitivity: 0.739, 1-specificity: 0.000) and 8.50 for emotional domain (sensitivity: 0.783, 1-specificity: 0.050). Conclusion: The preliminary statistical and ROC data analysis of VHI concluded that by using this type of assessment method, populations with or without voice disorders (in Greece) can be distinguished. Albeit this tool is a non-interventional method it could consequently offer an adequate screening and monitoring capability.

Emotion Detection by Analyzing Voice Signal Using Wavelet

Emotion is such a unique power of human trial that plays a vital role in distinguishing human civilization from others. Voice is one of the most important media of expressing emotion. We can identify many types of emotions by talking or listening to voices. This is what we know as a voice signal. Just as the way people talk is different, so is the way they express emotions. By looking or hearing a person’s way of speaking, we can easily guess his/her personality and instantaneous emotions. People’s emotion and feelings are expressed in different ways. It is through the expression of emotions and feelings that people fully express his thoughts. Happiness, sadness, and anger are the main medium of expression way of different human emotions. To express these emotions, people use body postures, facial expressions and vocalizations. Though people use a variety of means to express emotions and feelings, the easiest and most complete way to express emotion and feelings is voice signal. The subject of our study is whether we can identify the right human emotion by examining the human voice signal. By analyzing the voice signal through wavelet, we have tried to show whether the mean frequency, maximum frequency and Lp values conform to a pattern according to its different sensory types. Moreover, the technique applied here is to develop a concept using MATLAB programming, which will compare the mean frequency, maximum frequency and Lp norm to find relation and detect emotion by analyzing different voices.

Preliminary Receiver Operating Characteristic Analysis on Voice Handicap Index of Laryngeal Inflammation in Greek Patients

Objective(s): Laryngeal inflammations lead to voice disorders. Medical conditions such as chronic laryngitis, gastroesophageal reflux, laryngopharyngeal reflux, Reinke edema and/or vocal folds hemorrhage, result in diverse symptoms including chronic cough, throat cleaning and dysphonia (e.g. hoarseness). In turn, the dysphonic symptoms can be evaluated via subjective and objective procedures. The objective procedures usually include self-perceived questionnaires like the Voice Handicap Index (VHI). Studies reported that VHI can distinguish objectively dysphonic and non-dysphonic populations using the cut-off points of Receiver Operating Characteristic Curves. The purpose of this study was to calculate the cut-off points for individuals exhibiting voice symptoms which had been developed from laryngeal inflammatory diseases in Greece. Methods: One hundred and twelve participants (90 non-dysphonic and 22 dysphonic) filled in the Hellenic Voice Handicap Index (VHI) and the Greek translated version of Voice Evaluation Template (VEF) were administrated. All subjects were evaluated by an Otolaryngologist and a Speech-Language Pathologist. Results: The group with voice disorders exhibited higher VHI scores (in total and in its 3 subdomains) compared to non-dysphonic subgroup. Statistical significant differences were found for all VHI’s total cut-off point of 19.50 (sensitivity: 0.882, 1-specificity: 0.011) and for its three subdomains [functional 6.50 (sensitivity = 0.636, and 1-specificity = 0.022);physical 9.50 (sensitivity = 0.636, and 1-specificity = 0.000);emotional 6.50 (sensitivity = 0.455, and 1-specificity = 0.133)]. Conclusion: The preliminary results showed that VHI could discriminate individuals having voice disorders from laryngeal inflammations. The Voice Handicap Index can be used as a primary health care tool and a self-monitoring procedure in acute and sub-acute phases of the laryngeal inflammation.

数字示波器语音识别

示波器广泛应用于大学物理实验教学中,相较于传统的模拟示波器,数字示波器将输入模拟信号转换为数字信号进行显示。设计了一个新的使用数字示波器进行人声音识别的综合性实验。实验中把压电陶瓷片贴于测试者的喉部靠近声带处,声带的振动转换为电信号后,使用数字示波器解析测试者的声音特征。在此基础上建立了含多个测试者的语音样本数据库,通过对于测试者声音频率的比对实现了对特定测试者的语音识别。该实验拓宽了示波器的实验教学内容,取得了良好的教学效果。

基于改进MFCC融合特征及FA-PNN的驾驶员路怒情绪识别

现今关于驾驶员路怒情绪识别方法中语音特性分析相对较少,该研究以路怒情绪为研究对象,利用模拟驾驶系统建立数据集,通过分析驾驶员语音的频谱特征,将时域中短时能量及短时过零率特征参数和改进Mel频率倒谱系数(Mel frequency cepstral coefficients,MFCC)特征参数融合构成特征参数向量,利用萤火虫算法(firefly algorithm,FA)优化PNN神经网络(probabilistic neural networks)并构建识别模型,实现驾驶员路怒情绪的识别。实验结果表明,在相同神经网络下,改进MFCC融合特征提取方法相比传统MFCC特征提取方法具有更好的抗噪性。同时,FA-PNN模型的识别准确率为93.0%,相比传统PNN模型提高了11个百分点;F1-Score值为0.9328,提高了0.1047。该研究论证了语音信号处理技术对驾驶员路怒情绪识别的可行性,为汽车主动安全驾驶预警研究提供了新方法。