息肉与麻痹喉声源分类中非线性动力学发声系统模型研究

提出一种非线性动力学建模仿真发声系统,分类息肉和麻痹喉声源的方法,为声带疾病分类时参数选择提供了依据。首先介绍息肉和麻痹声带力学模型,耦合声门气流产生喉声源,求取喉声源频率(基频)、基频微扰;提出用庞加莱截面,分岔图对模型振动进行非线性分析;改变声带病理参数及声门下压,分析频率参数和混沌参数李雅普诺夫指数的变化。仿真实验结果表明,声带麻痹减小了发声基频,且只在一定压力范围内出现混沌振荡;息肉声带的混沌则分布在整个压力范围内。根据最大李雅普诺夫指数随声门下压变化的差异性分布,有助于识别并分类声带息肉和声带麻痹。      

高速摄影成像分析声带振动发声的前后不对称性

高速摄影成像直接观察到声带振动的前后不对称性。将11个离体狗喉声带进行发声实验,设置3组声门下压分别为10 cm H₂O,20 cm H₂O和30 cm H₂O,利用高速摄像仪和传声器,分别记录不同声门下压的声带振动图像和声信号.对高速摄影成像与同步采集的声信号基频进行定量分析和比较,基频均随声门下压的增大而增加。此外,对两种测量方法得到的基频进行相关分析比较,得到在同一声门下压下两种方法的基频相关系数均大于0.9,表明高速摄影成像得到的基频与声信号的基频具有高度相关性。高速摄影成像能直观地测量声带振动行为,对研究声带振动发声机理提供了有价值的测量手段。高速摄影获得的声带线性结构上25%,50%,75%位置处的振动幅度,显示了声带前后振动不对称且声门下压较低时振动不对称较明显。      

非对称黏性空气动力学声带模型及其病理喉声源分类

二质量块模型(SH模型)在模拟病理发声时未考虑弹性力对发声系统的影响,也未考虑黏性气流在声门闭合阶段的作用,本文提出一种非对称黏性空气动力学声带模型(ISAC模型)。对非对称振动时的附加弹性系数和弹性形变进行分析,修正质量块所受的碰撞力,随位移量变化调节原始模型中的劲度系数,模拟环甲肌和甲杓肌的张力作用;通过声门倾角变化得到声门处的气流分布,以实现声带壁上的非对称气流压力作用。该模型应用于发声病理诊断,模型仿真和病理喉声源分类识别的实验结果显示,各声门特征参数相对误差不超过1.5%,ISAC模型的加权平均误差低于SH模型,二分类识别率和细分准确率均高于SH模型。      

嗓音多频带非线性分析的声带病变识别

提出了一种嗓音多频带非线性分析的声带病变识别方法,以提高声带病变嗓音的识别率。首先采用Gammatone听觉滤波器组对嗓音信号进行滤波,求取每个频带下的最大李雅普诺夫指数;对映射到核空间的数据采用高斯最大似然度准则优化核函数,然后采用优化核主成分分析算法实现特征抽取。识别实验表明,多频带最大李雅普诺夫指数的识别率比传统的MFCC和最大李雅普诺夫指数分别有6.52%和8.45%的提高,且采用优化核主成分分析算法比传统核主成分分析算法有更好的抽取效果.将多频带非线性分析和优化核主成分分析算法结合,识别率提升至97.82%。      

Liu混沌系统的混沌分析及电路实验的研究

研究了一种新型混沌系统——Liu混沌系统的基本动力学行为以及电路实现的问题，给出了相图、庞卡莱映射、功率谱以及李雅普诺夫指数，基于李雅普诺夫指数谱和分叉图分析了系统参数对Liu混沌系统的影响.最后设计硬件电路证实了Liu混沌系统以及Liu混沌系统随系统参数变化时的各种状态的存在.给出数值仿真和电路实验的结果. 关键词： Liu混沌系统 分岔 电路实验     

一种最大李雅普诺夫指数估计的稳健算法

最大李雅普诺夫指数是诊断和描述动态系统混沌的重要参数.在深入研究相空间重构技术和轨道跟踪法的基础上,提出了一种从标量混沌时间序列中估计最大李雅普诺夫指数的新算法.该算法能够克服现有算法的不足,主要有以下三个优点:1)很高的精度;2)几乎不受噪声的影响;3)所需的计算时间和存贮空间小, 能进行在线计算.     

利用声带动力学模型参数反演方法进行病变嗓音分类

提出一种声带动力学模型参数反演方法,从发声机理角度对声带病变嗓音进行有效区分。依据声带生理组织和伯努利定律构建声带动力学模型,确定模型优化参数向量,耦合声门气流获取模型声门波;利用迭代自适应逆滤波算法获得实际嗓音声门波作为目标声门波;采用遗传优化算法提出通过匹配目标和模型声门波特征参数实现模型参数反演。实验结果表明,表征声门波的各时频域参数匹配相对误差不超过2%;依据反演所获模型参数提出去除声门下压影响的平均归一化缩放系数,克服声带非对称性特征在区分病变嗓音方面的不足,实现病理嗓音的全面有效区分。      

周期驱动下保守椭圆摆系统的混沌运动

运用广义拉格朗日方程,对周期驱动下的椭圆摆运动过程建立了模型方程.利用数值方法求解该方程,研究了该系统在周期驱动作用下不变环面的变形、破裂以及产生混沌运动的过程.并且,利用最大李雅普诺夫指数研究了驱动频率及质量参数对其混沌运动的影响.     

一种最大李雅普诺夫指数估计的稳健算法

最大李雅普诺夫指数是诊断和描述动态系统混沌的重要参数.在深入研究相空间重构技术和轨道跟踪法的基础上,提出了一种从标量混沌时间序列中估计最大李雅普诺夫指数的新算法.该算法能够克服现有算法的不足,主要有以下三个优点:1)很高的精度 2)几乎不受噪声的影响 3)所需的计算时间和存贮空间小,能进行在线计算 关键词：     

一类四维超混沌系统的界及同步的研究

文章研究了一类四维连续自治系统,在一定条件下它只有一个平衡点,却表现出复杂的动力学行为. 文章首先分析了系统的平衡点及李雅普诺夫指数;其次对该系统的界进行了估计,给出了界的表达式.通过设计一个控制器,研究了该四维超混沌系统的完全同步,并做出了相应的数值模拟. 关键词： 超混沌系统 李雅普诺夫指数 混沌系统的界 同步     

An aeroacoustic approach to phonation

A fluid mechanical, or aeroacoustic, point of view is followed to study possible sources of sound during phonation. Concentration is on two features of the vocal tract during phonation: abrupt area change from the glottis to the vocal tract and the finite length of the vocal tract. With these features, a source of sound distinct from the volume velocity source can be identified and a preliminary account of its effect on the acoustic field given. This source of sound is an oscillating force resulting from an interaction of rotational fluid motion with itself. Because of the schematic nature of the geometry of the model used here, this source may be considerably modified in actual phonation. It is concluded that specification of volume velocity is not enough to specify the source during phonation, even neglecting source-tract interaction.     

Pressure-flow relationships during phonation in the canine larynx

Measurements of air pressure and flow were made using an in vivo canine model of the larynx. Subglottic pressures at varying flow rates were taken during phonation induced by laryngeal nerve stimulation. Results showed that during constant vocal fold stiffness, subglottic pressure rose slightly with increased air flow. The larynx in the in vivo canine model exhibited a flow-dependent decrease in laryngeal airway resistance. Increasing flow rate was associated with an increase in frequency of phonation and open quotient, as measured glottographically. Results from this experiment were compared with a theoretical two-mass model of the larynx and other theoretical models of phonation. The influence of aerodynamic forces on glottal vibration is explained by increased lateral excursion of the vocal folds during the open interval and shortening of the closed interval during the glottal cycle.     

Measures of phonation type in Hmong

This study examines measures of glottal flow for vowels of Hmong, a Southeast Asian language which uses breathy and normal phonation contrastively. A software inverse filter was used to recover glottal airflow from oral airflow recordings. Properties of glottal flow measured in the time domain were glottal pulse symmetry and relative closed-phase duration. In the frequency domain, measures of spectral tilt and the amplitude difference between F0 and H2 were applied to discrete Fourier transforms (DFTs) of the glottal flow waveforms. Spectral tilt could not be reliably measured for many tokens. For the other measures, values were available for all tokens and were compared across phonation types. Flow pulse symmetry is not significantly different for breathy and normal-voice vowels. On the other hand, prominence of the fundamental relative to the second harmonic is a very significant correlate of the breathy/normal distinction, as is the relative closed-phase duration. These results are considered in light of an existing model of the voice source.     

Physical mechanisms of phonation onset: a linear stability analysis of an aeroelastic continuum model of phonation

In an investigation of phonation onset, a linear stability analysis was performed on a two-dimensional, aeroelastic, continuum model of phonation. The model consisted of a vocal fold-shaped constriction situated in a rigid pipe coupled to a potential flow which separated at the superior edge of the vocal fold. The vocal fold constriction was modeled as a plane-strain linear elastic layer. The dominant eigenvalues and eigenmodes of the fluid-structure-interaction system were investigated as a function of glottal airflow. To investigate specific aerodynamic mechanisms of phonation onset, individual components of the glottal airflow (e.g., flow-induced stiffness, inertia, and damping) were systematically added to the driving force. The investigations suggested that flow-induced stiffness was the primary mechanism of phonation onset, involving the synchronization of two structural eigenmodes. Only under conditions of negligible structural damping and a restricted set of vocal fold geometries did flow-induced damping become the primary mechanism of phonation onset. However, for moderate to high structural damping and a more generalized set of vocal fold geometries, flow-induced stiffness remained the primary mechanism of phonation onset.     

Modeling the role of nonhuman vocal membranes in phonation

Although the mammalian larynx exhibits little structural variation compared to sound-producing organs in other taxa (birds or insects), there are some morphological features which could lead to significant differences in acoustic functioning, such as air sacs and vocal membranes. The vocal membrane (or "vocal lip") is a thin upward extension of the vocal fold that is present in many bat and primate species. The vocal membrane was modeled as an additional geometrical element in a two-mass model of the larynx. It was found that vocal membranes of an optimal angle and length can substantially lower the subglottal pressure at which phonation is supported, thus increasing vocal efficiency, and that this effect is most pronounced at high frequencies. The implications of this finding are discussed for animals such as bats and primates which are able to produce loud, high-pitched calls. Modeling efforts such as this provide guidance for future empirical investigations of vocal membrane structure and function, can provide insight into the mechanisms of animal communication, and could potentially lead to better understanding of human clinical disorders such as sulcus vocalis.     

The association of laryngoceles with ventricular phonation

Laryngoceles represent dilatations of the laryngeal saccule that may extend internally into the airway, or externally through the thyrohyoid membrane. Unilateral laryngoceles are uncommon clinical entities and bilateral laryngoceles are rare. Certain activities like glass blowing and playing a wind instrument are associated with laryngocele development, as is laryngeal carcinoma in the ventricular area. This case describes development of bilateral laryngoceles in a patient who chronically uses ventricular phonation during speech. The pathogenesis involves repetitive elevation of intralaryngeal pressure during false vocal cord approximation, exposing the ventricles to abnormally high air pressures. The pathogenesis in this case, as well as in laryngoceles associated with occupational or anatomic risk factors, is discussed.