人机智能系统思考

对人工智能当前所面临的困境进行了分析,从人工智能的研究目的不应是使电脑取代人脑而应是使人脑与电脑更完美地结合的观点出发,对人机智能系统的概念与特性进行了初步探讨。     

漆包机烘炉热能平衡模糊控制系统设计与实现

漆包机烘炉温度系统的实时性和稳定性,关系到漆包线产品质量。针对漆包机烘炉热能系统特点,提出了一种的漆包机烘炉的热能平衡模糊控制系统,该系统提出了模糊热能平衡控制、PID温度控制结合的漆包机温度控制系统的设计思路。针对漆包机烘炉结构和系统特性进行研究,根据热能平衡方程建立漆包机烘炉系统的温度模型;利用热力学相关理论知识分析热能的输入输出的主要组成部分,计算烘炉的总热量;并提出热能平衡模糊控制系统原理流程图,设计相关的软、硬件模块加以实现。试验与测试表明,该热平衡系统为烘炉内各加热区域的温度调节提供了可靠保障;使烘炉的温控能够自适应环境温度的变化;热平衡系统以节约电能为准则,最大限度地减少电能消耗,提高产能比。     

Determination of an optimal configuration of operating policies for direct-input-output manufacturing systems using the Taguchi method

Some modern manufacturing systems have workstations directly integrated with a centralized storage and handling system for work-in-process. We consider a Direct-Inlet-Output manufacturing system(DIOMS) which has a number of machine centers placed along a built-in automated storagejretrieval system(AS/RS). The storagdretrievai(S/R) machine handles parts placed on pallets for the machine centers located at either one or both sides of the AS/RS.

This paper investigates the operational aspect of DIOMS by the Taguchi method. Four operating policies including input sequencing control. dispatching rule for the S/R machine, machine center-based part type selection rule, and storage assignment policy are treated as design factors. For the performance characteristics, man flow time and throughput are adopted. The number of machine centers, the number of part types, demand rate, processing time and the rate of each part type. vertical and horizontal speed of the SIR machM and the size of a local buffer in the machine centers are considered as noise factors in generating various DIOMS environments. A robust design experiment with inner and outer orthogonal arrays are conducted by computer simulation, and an optimal configuration of operating policies is presented which consists of a combination of the level of each design factor. The validity of the optimal configuration is investigated by comparing its SN ratios with those obtained by an experiment with fiill factorial designs.     

天文瞬变源快速自动识别系统的研究与实现

大视场和高时间采样率是现代天文光学瞬变源巡天项目的两个主要发展方向,相对传统的巡天项目将会产生更大的数据量和要求更快的瞬变源识别处理速度.为满足新技术下的瞬变源识别处理要求,本文提出用基于等光度测量星像轮廓等13个新的特征参量取代原有的轮廓拟合参量;使用实际星像轮廓仿真和构建较真实的训练样本算法;加入基于实测数据分析的噪声过滤判据等方法.实现了基于随机林森算法的天文光学瞬变源自动快速识别系统.通过仿真和实测数据的测试表明:本识别系统较国际主流的同类识别算法提速约10倍,样本识别的总体正确检出率和错误检出率都基本相同,而在低信噪比处,本文的识别算法有较良好的表现.本识别系统已成功应用于我国的迷你地基广角相机阵（地基广角相机阵的先导项目）,同时,本系统对于其他天文光学瞬变源巡天项目也有着重要的应用价值.     

Analytical procedure for constructing noise contours

This paper presents an analytical procedure to develop a noise contour map of the workplace. The procedure requires the following input data: ambient noise level, noise levels generated by individual machines, and the (x, y) coordinates of the machine locations. A set of mathematical formulae is also developed to estimate the combined noise levels at predetermined locations (or points) of the workplace floor. Contour lines are then drawn to connect points having an equal noise level. Although based on estimated data, this noise contour map allows engineers to better understand the noise situations of the workplace and determine if a noise hazard is present. The analytical nature of the procedure also enables engineers to quickly construct the noise contour map and revise the map when changes occur in noise levels due to a workplace re-layout or an addition of a new noise source.Relevance to industryThe procedure for constructing a noise contour map is typically laborious and time consuming, thereby limiting its applications although its usefulness is widely recognized. With the proposed analytical procedure, manufacturing industry can quickly evaluate the workplace noise situation and determine if workers are exposed to a noise hazard. Furthermore, the effect of placing an additional machine at a specific location on the overall noise level can be predicted prior to its installation, thereby not only avoiding unnecessary costs but also enhancing workplace safety.     

基于多粒度对抗训练的鲁棒跨语言对话系统

跨语言对话系统是当前国际研究的热点和难点. 在实际的应用系统搭建中, 通常需要翻译引擎作为不同语言之间对话的桥梁. 然而, 翻译引擎往往是基于不同训练样本构建的, 无论是所在领域, 还是擅长处理语言的特性, 均与对话系统的实际应用需求存在较大的差异, 从而导致整个对话系统的鲁棒性差、响应性能低. 因此, 如何增强跨语言对话系统的鲁棒性对于提升其实用性具有重要的意义. 提出了一种基于多粒度对抗训练的鲁棒跨语言对话系统构建方法. 该方法首先面向机器翻译构建多粒度噪声数据, 分别在词汇、短语和句子层面生成相应的对抗样本, 之后利用多粒度噪声数据和干净数据进行对抗训练, 从而更新对话系统的参数, 进而指导对话系统学习噪声无关的隐层向量表示, 最终达到提升跨语言对话系统性能的目的. 在公开对话数据集上对两种语言的实验表明, 所提出的方法能够显著提升跨语言对话系统的性能, 尤其提升跨语言对话系统的鲁棒性.     

基于时间序列分析的水位噪声信号仿真

汽包水位的控制关系到整个联合循环电厂的运行效率和运行安全.由于水位信号中往往包含着噪声信号,因此水位噪声信号对水位控制系统的影响在进行控制系统设计和仿真时都必须加以考虑.然而已有的大多数余热锅炉仿真模型都没有考虑汽包水位噪声的影响.该文采用时间序列分析的方法,对现场采集的汽包水位噪声信号进行了分析,采用Burg方法建立了相应的水位噪声信号AR模型;该模型从功率谱上对现场数据进行了匹配模拟,使用该模型作为噪声模块在时域上成功地重建了汽包水位噪声信号,为开发余热锅炉汽包水位先进控制策略奠定了基础.     

Sorption characterization and actuation of a gas-gap heat switch

Cryogenic gas-gap heat switches are rendered compact by using a sorption pump as actuating device. The sorption pump adsorbs gas when it is cooled and releases gas when it is heated. Upon desorption, the released gas lies in the gap between two blocks and increases the conduction heat transfer between them; in the reverse case - when the gas is adsorbed - the gap space between the two blocks behaves as an insulator. The temperatures at which there is sufficient desorption and adsorption of gas for actuating the ON an OFF states depend on several parameters of the gas-adsorbent system. The adsorption characteristics of helium gas on several carbon adsorbents were studied at pressures ranging from 50 Pa to 120 kPa and temperatures from 10 to 100 K. One of the carbons was successfully used in a heat switch at 6 K under different configurations in order to access different actuation characteristics. The experimental helium adsorption equilibrium data for each carbon were successfully fitted by a semi-empirical adsorption isotherm model. A thermal model of the heat switch, whose actuation can be tailored upon specifications, was developed based on the adsorption data. The thermal model was validated by reproducing the experimental data for two different configurations of a heat switch prototype.     

运输类大型飞机噪声适航测试方法应用

对研发过程中的新舟700系列飞机在内的涡桨运输类客机的噪声适航方法进行分析,对其瞬时声压级、有效感觉噪声级等算法进行研究.使用丹麦B&K公司的噪声分析产品PLUSE声学测试分析系统,采用基于静态测试预测噪声级的算法,“以机代台”对发动机静态数据进行分析.利用涡桨类飞机竞争机型庞巴迪DHC-8-400飞机已有的噪声数据,使用C#语言和Visual Studio 2013平台开发预测系统.最后利用SPSS软件进行置信区间的计算,验证数据的符合性.方法及竞争机型噪声数据可为今后上市的涡桨类运输飞机提供噪声适航研发依据.     

Novel planar balanced bandpass filter with wideband common‐mode suppression and in‐band common‐mode noise absorption

This paper presents a novel planar balanced bandpass filter (BPF) with wideband common mode (CM) noise suppression and in‐band CM noise absorption using coupled lines (CLs) with short‐circuited stubs to realize high selectivity and wideband differential mode (DM) filtering performance. Two one‐quarter wavelength stubs loaded with grounded resistors are introduced to realize wideband CM noise suppression. Thus, CM noise can be suppressed under a certain level at all frequencies. Four resistors are used to achieve CM noise absorption by dissipating the CM noise into heat, which can avoid the noise being reflected to the communication system and realize a wide absorption bandwidth with 90% absorption efficiency. For demonstration, an absorptive balanced BPF operating at 3.5 GHz with wide 3‐dB fractional bandwidth (FBW) of 79.43% is fabricated and experimentally validated. It is worth noted that the absorptive balanced BPF can realize broadband CM noise suppression from 0 to 8 GHz, and the CM noise is well absorbed more than 10 dB from 2.41 to 4.63 GHz. Besides, wideband CM noise absorption with 90% efficiency from 2.51 to 4.60 GHz is realized, which indicates potential applications in improving the performance of the balanced radio frequency (RF) circuits. Good agreements between the simulated and measured results are observed.     

Automatic normalization of short texts by combining statistical and rule-based techniques

Short texts are typically composed of small number of words, most of which are abbreviations, typos and other kinds of noise. This makes the noise to signal ratio relatively high for this specific category of text. A high proportion of noise in the data is undesirable for analysis procedures as well as machine learning applications. Text normalization techniques are used to reduce the noise and improve the quality of text for processing and analysis purposes. In this work, we propose a combination of statistical and rule-based techniques to normalize short texts. More specifically, we focus our attention on SMS messages. We base our normalization approach on a statistical machine translation system which translates from noisy data to clean data. This system is trained on a small manually annotated set. Then, we study several automatic methods to extract more general rules from the normalizations generated with the statistical machine translation system. We illustrate the proposed methodology by conducting some experiments with a SMS Haitian-Créole data collection. In order to evaluate the performance of our methodology we use several Haitian-Créole dictionaries, the well-known perplexity criteria and the achieved reduction of vocabulary.     

基于偏微分方程的稠密视差图获取方法

立体匹配是计算机视觉领域中的一个重要的热门研究课题,为了获得性能更优的稠密视差图,通过把偏微分方程理论运用于机器视觉中,提出了一种新的基于能量函数获取稠密视差图(disparity map)的方法,并首先分析了匹配点对在不同相对位置下对匹配项产生的影响;接着提出了适用于视差图的各向异性的热扩散方程,它不仅继承了Alvarez定义的正则项对初始视差图内部平滑和保持边缘不连续的特性,还通过引入图像的噪声屏蔽函数和二阶方向导数来分别控制对应视差图中不同区域的扩散速度和角点处的扩散方向;最后通过定义的正则项和匹配项来构造新的能量函数,并把基于区域匹配算法得到的视差图作为初始值,再利用最速下降法求解相应的最小能量泛函。实验结果表明,无论从视觉效果上,还是重构深度图的判别上,该新算法都取得了更优的性能。     

噪声在慢变系统中的随机Chay神经元模型的自共振

研究了白噪声在慢变系统中的Chay神经元模型的共振．被加到慢变量中的白噪声相当于钙离子的波动,这是更具有生理意义的．随着噪声强度的增加,发现在某一优化噪声水平处,信噪比经过极大值,这意味自共振的发生．同时,也证实了钙离子的波动可以引起信号的探察和转换．     

一种基于高斯混合模型的视频监控运动物体提取系统

现有的视频监控系统必须要依靠人对监控图像的实时观察才能发挥作用。设计并实现了一种基于高斯混合模型的自适应的视频监控运动物体提取系统,使用高斯混合模型对监控场景进行建模,利用对象形状的空间连通性和最小像素尺寸约束去除噪声影响,从而实时地分离出前景中的运动物体并对其进行追踪和计数,仿真实验的结果证明了系统的有效性。     

惩罚复杂诊断系统混合噪音的模糊小波分类机

针对复杂故障诊断系统特征数据中具有高斯、幅值较大、奇异点等混合噪音, 设计一种分段式损失函数, 构造基于小波基函数的小波核函数, 由此得到一种输出为模糊数的模糊小波ν-支持向量分类机, 即模糊鲁棒小波ν-支持向量分类机(FRWν-SVC). 它可以有效地压制故障特征时序的多种噪音和奇异点, 具有很强的鲁棒性, 而且它比标准模糊小波ν-支持向量分类机(FWν-SVC)具有更简洁的对偶优化问题. 最后进行了汽车装配线故障诊断的实例分析, 结果表明基于FRWν-SVC的故障诊断模型是有效可行的.     

Sound field analysis and simulation for fluid machines

Fluid machines, such as pumps, fans, and internal combustion engines, are widely used in duct systems for air-conditioning, cooling, ventilating, heat releasing, and dust collecting. Vibration and noise will be created when the fluid machine works with fans at various rotating speeds. Noise problems associated with fan installation are a concern in fluid machines. Methods to analyze the sound field and simulation of fan installation are, therefore, important for the design to reduce the noise output from fluid machines. The sound field is simulated by using the boundary element method (BEM), which is a numerical technique to reduce the boundary integral equations using the fundamental solution and Green’s transfer functions, for sound field analysis in this paper. For the sound field analysis, the geometry of the fluid machines (the axial fan and centrifugal fan) and the acoustic properties are modeled in Beasy software based on the boundary element method technology. The 1/1 octave frequency band from 63 Hz to 8 kHz ranges are selected on sound field analysis. The sound pressure of the fan and the motor in each octave band are calculated by the parameters of rotating speed, flow volume, horse power and number of blades used. The results show that there is a high level sound pressure inside the housing of the axial fan due to the sound source located there. The higher sound pressure level is observed on both the inlet and outlet. The results for the centrifugal fan are the higher the frequency, the heavier energy that is found to radiate around the sound source.     

Recognition of consonant-vowel (CV) units under background noise using combined temporal and spectral preprocessing

This paper proposes hybrid classification models and preprocessing methods for enhancing the consonant-vowel (CV) recognition in the presence of background noise. Background Noise is one of the major degradation in real-time environments which strongly effects the performance of speech recognition system. In this work, combined temporal and spectral processing (TSP) methods are explored for preprocessing to improve CV recognition performance. Proposed CV recognition method is carried out in two levels to reduce the similarity among large number of CV classes. In the first level vowel category of CV unit will be recognized, and in the second level consonant category will be recognized. At each level complementary evidences from hybrid models consisting of support vector machine (SVM) and hidden Markov models (HMM) are combined for enhancing the recognition performance. Performance of the proposed CV recognition system is evaluated on Telugu broadcast database for white and vehicle noise. The proposed preprocessing methods and hybrid classification models have improved the recognition performance compared to existed methods.     

Induction machine magnetic noise: Impact of a new stator magnetic circuit design

The main topic of this paper is to examine the impact on induction machine magnetic noise of a new stator magnetic circuit design. This new design was developed in order to increase AC machine energetic performance by assembling non-segmented shifted grain-oriented steel sheets which make it possible to reduce iron losses. Experimental results on induction machine efficiency increase are presented and justified using reluctance network simulations. An analysis, finally, shows and explains how this structure behaves considering acoustic magnetic noise in a Pulse Width Modulation supply case.     

门机抓梁自动检测系统研究

介绍水电站门机抓梁自动检测系统,分析了穿销位移传感器的抓梁水平传感器的工作原理,讨论了系统的硬件设计和软件编程,给出了ＰＬＣ程序的流程图,指出灵活运用各种数字信号处理方法可极大地优化系统的性能。     

微细线材缠绕机的可视化控制

设计了一种新的微细线材缠绕机可视化控制系统,该设备用旋转编码器检测主轴的旋转速度,PLC检测传感器信号控制步进电机的进给,以编码器发出一定数目的脉冲数为一个计时单位,步进电机的跟随频率适时发生变化,驱动过线架以一定的速比要求做往复运动。很好地解决了设备中存在的噪声问题。触摸屏实现人机对话,各个控制参数一目了然。