基于微型呼吸传感器的高血压治疗系统

采用压阻式全桥型腹带式微型呼吸传感器检测人体的呼吸信号,研制基于慢呼吸训练的高血压治疗系统,以单片机AT89C51为核心,用全速USB接口芯片PDIUSBD12实现单片机与计算机之问的通信;采用Microsoft Visual C＋＋6．0编写上位机软件,根据生物信息反馈原理设计以呼吸信号为反馈生理量的慢呼吸引导算法,以直观的音乐信号和进度条指示受试者进行呼气和吸气。进行了志愿者实验,结果表明：本系统能够有效地引导受试者进行慢呼吸训练,受试者血压下降较为显著。     

WSN中一种基于单类1/4球体支持向量机的异常检测方案

针对数据非平稳分布而导致的异常检测差错率较高的问题,提出了一种可用单类1/4球体支持向量机模型参数表示的异常率参数动态调整自适应算法。该算法在线迭代运行,依次生成基于训练集的最优模型,最后提出一种可以迭代得到训练集异常率的参数确定算法,可发现数据集中的异常。仿真实验结果表明,与异常率固定的静态模型相比,该算法可以利用训练集生成最优模型,进而实现分类集差错率最小。     

皮肤镜黑素细胞瘤图像自适应聚类的进化寻优

图像的自动准确分割是实现黑素细胞瘤图像自动分析的关键.针对皮肤镜黑素细胞瘤图像,提出一种基于改进遗传算法和自生成神经网络(SGNN)相结合的自适应聚类分割算法.首先采用遗传算法选取一组最优的种子样本作为初始神经树;然后通过SGNN对剩余样本进行训练得到一个自生成神经森林;最后令森林中每棵树代表一个类,完成黑素细胞瘤图像的自适应聚类分割.该算法解决了SGNN对样本训练顺序敏感的问题,并能够自适应地确定类别数,聚类过程无需任何人工干预;同时根据解空间的大小设定遗传算法的初始种群规模,并在进化过程中根据个体的变化对种群规模以及交叉率和变异率等遗传控制参数进行动态调整,有效地提高了算法的运行速度.实验结果表明,文中算法稳定性好,聚类结果符合人眼判别的诊断要求.     

基于非监督学习神经网络的自动调制识别研究与实现

以非监督学习神经网络为主要研究对象,描述自组织网络的基本模型,分析传统自组织网络的训练算法,提出了一种基于自组织特征映射SOFM(Self-Organizing Feature Map)神经网络的通信信号自动调制识别方法。方法改进了训练算法中的学习率函数和邻域函数,提高了算法的收敛速度和性能,并将其应用在通信信号调制识别中。仿真实验检验基于SOFM神经网络的调制识别方法的性能,并与后向反馈(BP)神经网络加以比较,结果表明SOFM神经网络的调制识别方法具有较高的识别精度,改进后的训练算法提高了识别的有效性。     

基于反馈调整的自相关过程质量损失分析

传统的SPC要求过程观测值统计独立,而实际过程大多具有自相关性,不满足统计独立的要求。在SPC与EPC的整合框架下,分析了自相关过程的质量损失。根据模型的自相关系数(ACF)和偏相关系数(PACF)以及选择迭代算法识别模型参数;采用MMSE反馈控制器对自相关过程进行调整,并结合过程能力,对自相关过程的质量损失进行分析。结果表明,在MMSE反馈调整后,自相关过程质量损失才符合实际情况。     

基于自训练SGAN的调压器半监督故障诊断模型

针对实际工况下调压器标记样本稀缺导致故障诊断效果不佳的问题,基于自训练算法与半监督生成对抗网络(semi-supervisedgenerativeadversarialnetwork,SGAN)设计了故障诊断模型。首先,对燃气调压器一维压力信号进行预处理,得到灰度图像样本。之后,基于深度卷积生成对抗网络,设计SGAN进行特征提取,判别器采用具有共享权值的堆叠鉴别器模型。然后,设计自训练算法,使用训练好的初始分类器预测无标签样本的类别标签。最后,采用重复标记方式将满足要求的样本扩充到有标签样本集重新训练,保存最终的分类器。实验结果表明,在少量调压器标签样本的情况下,所提模型依旧具有良好的性能。     

一种新的基于强化学习改进SAR的无人机路径规划

搜索和救援优化算法(SAR)是2020年提出的模拟搜救行为的一种元启发式优化算法,用来解决工程中的约束优化问题.但是, SAR存在收敛慢、个体不能自适应选择操作等问题,鉴于此,提出一种新的基于强化学习改进的SAR算法(即RLSAR).该算法重新设计SAR的局部搜索和全局搜索操作,并增加路径调整操作,采用异步优势演员评论家算法(A3C)训练强化学习模型使得SAR个体获得自适应选择算子的能力.所有智能体在威胁区数量、位置和大小均随机生成的动态环境中训练,进而从每个动作的贡献、不同威胁区下规划出的路径长度和每个个体的执行操作序列3个方面对训练好的模型进行探索性实验.实验结果表明, RLSAR比标准SAR、差分进化算法、松鼠搜索算法具有更高的收敛速度,能够在随机生成的三维动态环境中成功地为无人机规划出更加经济且安全有效的可行路径,表明所提出算法可作为一种有效的无人机路径规划方法.     

时分制多道生理遥测仪

本机是用来遥测被试者的多项生理参数和所处环境的一些物理参数的仪器。它可用于医学生理学的研究及某些场合的医务监督。系统采用 PDM—FM 调制形式,机内设有快变交换子和两个慢变交换子(即主、付交换子),配接相应的传感放大器,即可同时遥测六路心电、脑电、振动之类的快变信号和十三路呼吸、温度之类的慢变信号。发送端输入信号幅度为±50mv,解调器输出为±2.5v。慢变信道安插在帧同步脉冲的位置,以便能充份运用系统。为了提高系统的同步可靠性,在接收端采用了惯性帧同步电路和锁相环。此外,发射机载频采用声表面波调频振荡器,它具有频率较稳、频偏较大、频谱较纯的特点。系统均方根误差为±2%,遥测距离为100米。     

基于变分自编码器的不平衡样本异常流量检测

随着机器学习技术的快速发展,越来越多的机器学习算法被用于攻击流量的检测与分析,然而攻击流量往往只占网络流量中极小的一部分,在训练机器学习模型时存在训练集正负样本不平衡的问题,从而影响模型训练效果.针对不平衡样本问题,文中提出了一种基于变分自编码器的不平衡样本生成方法,其核心思想是在对少数样本进行扩充时,不是对全部进行扩充,而是分析这些少数样本,对其中最容易对机器学习产生混淆效果的少数边界样本进行扩充.首先,利用KNN算法筛选出少数类样本中与多数类样本最近的样本;其次,使用DBSCAN算法对KNN算法筛选出的部分样本进行聚类处理,生成一个或多个子簇;然后,设计变分自编码网络模型,对DBSCAN算法区分出的一个或多个子簇中的少数类样本进行学习扩充,并将扩充后的样本加入原有样本中用于构建新的训练集;最后,利用新构建的训练集来训练决策树分类器,从而实现异常流量的检测.选择召回率和F1分数作为评价指标,分别以原始样本、SMOTE生成样本、SMOTE改进方法生成样本和文中所提方法生成样本为训练集进行对比实验.实验结果表明,在4种异常类型中,采用所提算法构造训练集训练的决策树分类器在召回率和F1分数上都有提升,F1分数相比原始样本及SMOTE方法最高提升了20.9％.     

基于模糊神经网络火灾探测信号处理方法

针对火灾探测信号的特点,建立了火灾探测系统模型及用于处理火灾信号的模糊神经网络计算模型.利用神经网络构造模糊系统,用神经网络的自学习和自适应能力自动调整模糊系统参数,用改进的BP算法对网络进行学习和训练.根据国家标准试验火数据进行网络的学习和测试,系统误差小于试验火标准误差要求,表明了算法的有效性和可行性.     

Evaluating mobile apps for breathing training: The effectiveness of visualization

Deep and slow breathing exercises can be an effective adjunct in the treatment of stress, anxiety, post-traumatic stress disorder, chronic pain and depression. Breathing techniques are traditionally learned in courses with trainers and/or with materials such as audio CDs for home practice. Recently, mobile apps have been proposed as novel breathing training tools, but to the best of our knowledge no research has focused on their evaluation so far. In this paper, we study three different designs for breathing training apps. The first employs audio instructions as in traditional training based on audio CDs, while the other two include visualizations of the breathing process, representative of those employed in current breathing training apps. We carry out a thorough analysis, focusing on users’ physiological parameters as well as subjective perception. One visualization produces better results both objectively (measured deepness of breath) and subjectively (users’ preferences and perceived effectiveness) than the more traditional audio-only design. This indicates that a visualization can contribute to the effectiveness of breathing training apps. We discuss which features could have allowed one visualization (but not the other) to obtain better results than traditional audio-only instructions.     

基于示波法的血管硬度无创检测系统

在示波法测量血压的原理基础上,设计和开发了血压与血管硬度的无创检测系统。该系统以MOTOROLA单片机MC68HC908JL8为控制核心,在血压测量过程中,从采集到的血压波形信号中高精度地提取血管硬度的有效信息,并通过大量的临床实验数据的分析,建立相应的血管硬度的模型,实现了在无创血压测量的同时,检测出血管硬度指数、心率等多种心血管参数。临床实验表明,该检测系统能够较好地反映人体的血管健康状况。     

Quantitative feedback theory with applications in welding

Two different control systems were developed using quantitative feedback theory for weld quality improvement. The first application to be discussed was designed for the resistance pinch welding process and uses electrode displacement and force as feedback parameters. A correlation between weld quality and electrode displacement is established for constant electrode force. The system is capable of producing repeatable welds of consistent quality, with wide variations in weld parameters. This is the first time feedback control has been successfully applied to pinch welding. The second example presented in this paper was developed for penetration control of the gas-tungsten-arc welding process. The feedback signal is obtained by measuring the amount of light emitted from the back side of the weld. Welds of constant penetration have been demonstrated in tests with travel speeds varying from 1.5 to 6 inches per minute and with 200 per cent changes in part thickness.     

模糊PID在冷轧机厚度控制系统上的应用

为解决现有轧机液压AGC系统响应慢、精度不高和工作效率低的问题,将现有液压PID控制系统改进成模糊PID自整定控制系统。首先在分析轧制过程中的变形确定轧制阻力,以及控制环节的传递关系的基础上,建立了工作装置和液压控制系统联合仿真模型,并且在轧机厚度控制系统的数学模型基础上进行了参数优化,利用AMESim和Mat]lab/Simulink联合仿真得出了轧件入出口板厚曲线,并对液压缸PID控制系统和模糊PID自整定控制系统进行对比分析。经研究表明:普通PID控制系统无法对出口板厚进行有效实时调整,而模糊PID自整定控制系统除对板厚能实时调整外,其控制精度大大提高,最大板厚误差仅为0.1 mm,改进后的系统响应快,响应时间为0.4 s。     

A model-based fuzzy logic controller with Kalman filtering fortracking mean arterial pressure

This paper proposes a new noninvasive measurement method for tracking the tendency of mean arterial pressure (MAP) in the radial artery. The designed system consists of a tonometer, a microsyringe device, and a model-based fuzzy logic controller. The modified flexible diaphragm tonometer is to detect the continuous blood pressure waveform and vessel volume pulse. A precise mathematical model describing the interaction between the tonometer and artery is derived. To reach accurate measurement without distortion, a model-based fuzzy logic control system is designed to compensate the change of MAP by applying a counter pressure on the tonometer chamber through the microsyringe device. The proposed control system consists of a linear predictor, a Kalman filter, and a synthetic fuzzy logic controller (SFLC). Simulation results show that, for the real physiologic MAP with changing rates up to 20 or -20 mm-Hg/minute, the model-based SFLC can beat-to-beat adjust the tonometer's chamber pressure to follow the tendency of MAP accurately     

The effect of hyperoxia on submaximal exercise with the self-contained breathing apparatus

The e^ffects of hyperoxia on submaximal exercise with the self-contained breathing apparatus (SCBA) were studied in 25 males. Each participant completed a graded exercise test for the determination of ventilatory threshold (VT) and then a submaximal practice trial with a normoxic gas mixture. The normoxic (20.93 ± 0.22% O₂ ; SUB₂₁) and hyperoxic (40.18 ± 0.73% O₂; SUB₄₀) submaximal trials were then administered in a random order. All exercise tests were completed on separate days while wearing firefighting gear and the SCBA. Compared with SUB₂₁, hyperoxia significantly reduced minute ventilation ([Vdot]_E ), mask pressure (P_mask), heart rate, blood lactate concentration, perceived exertion, and perceived breathing distress. As expected, hemoglobin saturation remained higher (p<0.05) during SUB₄₀. The reductions in both [Vdot]_E and P_mask with hyperoxia imply a reduction in the work of breathing during exercise. Total gas consumption was 10.3 ± 8.1% lower during SUB₄₀ when compared to SUB₂₁, another finding that has significant practical implications for occupational safety.     

The specificity of endurance training on muscular power and muscle fibre size

In an effort to further examine the specificity of training concept, six female student volunteers (mean age=25.3 years) were endurance trained three times a week for a period of twelve weeks. The training consisted of thirty minutes of continuous running per workout. Identical test batteries were administered prior to (pre) and immediately following (post) the treatment period. Each battery of tests consisted of three measures of muscular leg power (vertical jump test, 18-29 m sprint, and the Margaria Step Test), and two measures of cardiovascular endurance (Cooper's 12 minute run, and V?O₂ max during treadmill running). A muscle biopsy from the lateral aspect of the gastrocnemius muscle was also included. The biopsy specimens were stained for myosin ATPase. Percentages of fast twitch (FT) and slow twitch (ST) muscle fibres as well as mean individual FT and ST fibre areas were determined. The training produced significant changes in the exercising subjects for thel2-minute run (p=0.006)and VO₂max(p=0.003). No changes were found in the tests of muscular power or individual fibre areas. These data indicate that the endurance training program was of sufficient intensity to effect an increase in cardiovascular fitness. However, these improvements in endurance were not accompanied by any significant changes in muscular leg power or individual muscle fibre size.     

预失真多合体功率放大器ACLR与反馈通道带宽的关系

针对TI公司数字预失真芯片GC5322和Xilinx公司的数字预失真知识产权核(IP Core),考虑输入九载波TD-SCDMA信号,测试了数字预失真多合体功率放大器的反馈通道带宽与邻道及次邻道泄漏功率比(ACLR)的关系。测试结果表明,当反馈通道带宽大于信号带宽1.7倍时,再增加反馈通道带宽,ACLR的改善量变化不明显,在±1.3 dB范围内波动;当反馈通道带宽小于信号带宽1.7倍以下时,再降低反馈通道带宽,ACLR的改善量恶化明显。该结果对工程实现时选取最低反馈通道带宽,降低系统对模数转换器(ADC)的指标要求和器件成本,具有重要的参考和应用价值。     

A self-adjusting active compliance controller for multiple robots handling an object

This paper presents the concept and experimental validation of a self-adjusting active compliance controller for n robots handling its compliant behaviour concerning partly unknown flexible object. The control strategy is based on the decomposition of the 6n-dimensional position/force space and includes a feedforward and feedback level. The feedforward level contains motion coordination, force distribution of external forces, creation of internal forces, and an additional loop adding the elastic displacements due to the applied forces to the planned robot positions. The feedback level is organized in the form of an active compliance control law. For adjusting the controller to the, in general, unknown flexible behaviour, which in practice is the main problem of the controller design, a quasi-static model of the system is derived for different contact cases of the object and a procedure is presented, which by use of this model is capable of determining the compliance of the considered system and therefore of adjusting the controller. Experiments with two puma-type robots have been conducted to show the applicability of the self-adjusting control strategy. The task has been to grasp and move an unconstrained object. It is shown, that the system can adjust the control parameters to the unknown system compliance and that the control performance is improved considerably.