Grael脑电放大器与深度学习的手势实时识别研究

手势识别是人机交互的关键。为了能够更好地实现脑电信号与肌电信号的融合,精准地识别人体的运动,本文建立了一套基于Grael脑电放大器的手势动作实时检测识别的研究系统。通过Grael脑电放大器和Curry8系统采集5个通道的8种不同手势的表面肌电信号(sEMG),并对采集到的sEMG信号进行滤波去噪、滑动窗口分割以及特征提取等预处理的操作;最后采用几种常用的分类器与卷积神经网络(CNN)对不同手势的sEMG信号进行实时分类识别。结果表明CNN的识别准确率最高,能达到92.98%;对每个手势动作进行30次实时识别检测,结果显示识别延迟大概在1～1.5 s,实时识别的精度可高达90%。该系统为将来研究脑电信号与肌电信号的融合提供了一个可行的方法,在人机交互方面展现了巨大的潜力和应用空间。     

VLSI Potentiostat Array With Oversampling Gain Modulation for Wide-Range Neurotransmitter Sensing

A 16-channel current-measuring very large-scale integration (VLSI) sensor array system for highly sensitive electrochemical detection of electroactive neurotransmiters like dopamine and nitric-oxide is presented. Each channel embeds a current integrating potentiostat within a switched-capacitor first-order single-bit delta-sigma modulator implementing an incremental analog-to-digital converter. The duty-cycle modulation of current feedback in the delta-sigma loop together with variable oversampling ratio provide a programmable digital range selection of the input current spanning over six orders of magnitude from picoamperes to microamperes. The array offers 100-fA input current sensitivity at 3.4-muW power consumption per channel. The operation of the 3 mm times3 mm chip fabricated in 0.5-mum CMOS technology is demonstrated with real-time multichannel acquisition of neurotransmitter concentration     

Cortically coupled computer vision for rapid image search.

We describe a real-time electroencephalography (EEG)-based brain-computer interface system for triaging imagery presented using rapid serial visual presentation. A target image in a sequence of nontarget distractor images elicits in the EEG a stereotypical spatiotemporal response, which can be detected. A pattern classifier uses this response to reprioritize the image sequence, placing detected targets in the front of an image stack. We use single-trial analysis based on linear discrimination to recover spatial components that reflect differences in EEG activity evoked by target versus nontarget images. We find an optimal set of spatial weights for 59 EEG sensors within a sliding 50-ms time window. Using this simple classifier allows us to process EEG in real time. The detection accuracy across five subjects is on average 92%, i.e., in a sequence of 2500 images, resorting images based on detector output results in 92% of target images being moved from a random position in the sequence to one of the first 250 images (first 10% of the sequence). The approach leverages the highly robust and invariant object recognition capabilities of the human visual system, using single-trial EEG analysis to efficiently detect neural signatures correlated with the recognition event.     

基于嵌入式系统的简易逻辑分析仪设计

基于数字信号采集处理以及数字示波器存储显示原理,提出了一种简易逻辑分析仪制作方案。该系统主要由C8051F020与FPGA最小系统模块、ADC采集模块、信号衰减模块及TFT触摸显示模块组成。该设计采用单级、三级触发方式判断,可同时对8路信号进行采集、触发、存储及显示。经实验验证,该系统具有较高的测试速率、能实现多通道输入、可进行多级触发等优点。     

基于FPGA与退化YOLO的手机镜片缺陷检测系统

针对镜片缺陷检测采用图像处理法和神经网络法存在时延高、功耗高和检测缺陷类别较少等问题,设计了一种基于FPGA与退化YOLO的软硬协同检测系统。系统中使用卷积层代替YOLO网络的重排序层进行网络退化,并映射到FPGA上;采用动态量化、模块融合、双缓冲流水线、循环展开和分块等优化策略,设计可动态配置的加速IP,其中的卷积计算模块分别实现了基于Winograd和GEMM的快速卷积算法。实验结果表明,本系统的加速IP在PYNQ-Z2上获得了51.89 GOP/s的计算性能,比基于典型滑动窗口卷积计算方法的性能提高了0.76倍,加速单张图像的时延为433ms,功耗为1.07W,与Core i5-10500 CPU相比,能效是其365.27倍,实现了小型设备对手机镜片低时延、低功耗的多缺陷检测。     

基于RSVP的面向不良信息检测人机协作系统研究

针对复杂的环境背景下不良信息的快速准确检测问题,提出了基于快速序列视觉呈现( rapid serial visual presentation, RSVP)的面向不良信息检测人机协作系统。 首先利用快速佩戴便携式采集系统采集了 12 名受试者的脑电数据;然后采用 Mallat 算法提取较低维度的时频特征,使用人工神经网络(ANN)和支持向量机(SVM)两种模型分类对比;最后在训练集中引入 不同次数的叠加平均数据以改善模型的分类性能。 实验结果表明,在含有 3 个目标的 60 张图像中平均正确输出至少 2 张目 标,AUC 值达到了 0. 9。 该系统在小批量数据集、环境变化复杂的不良图像信息检测中有着良好的性能,相较于人工检测提高 了效率。     

Fast multidimensional directed information analysis

Multidimensional directed information (MDI) analysis is a signal processing method to quantify and visualize the causality of multichannel time series in the form of information flow. MDI analysis is defined as conditional mutual information and needs large calculations. Although MDI is used for electroencephalogram (EEG) analysis, large computation time is a problem. MDI can be calculated without direct probability calculations, assuming that the multichannel time series has Gaussian profiles. However, the amount of calculation increases exponentially with increase in the number of channels. Such large calculations have prevented practical use of MDI analysis in medical fields such as clinical EEG analysis in which many multidimensional time series need to be processed. In this paper, we propose a new calculation approach to drastically decrease the calculation time of MDI analysis. The proposed method makes it possible to decrease the calculation time exponentially for multichannel time series that can be approximated with multidimensional autoregressive models. © 2012 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.     

Time-interleaved SAR ADC design with background calibration

In this article, a low power time-interleaved SAR (TI-SAR) ADC is presented. Background calibration is used to improve the linearity of the ADC. Offset, gain, and capacitor mismatches between interleaved channels are calibrated by postprocessing the ADC output. Besides, a novel trimming-based calibration algorithm is used to calibrate the timing mismatches between channels. The proposed calibration algorithm is more power-efficient compared with most of its counterparts. The ADC consists of 18 parallel channels, a reference channel with two dummy channels, and a channel for timing calibration. The timing calibration channel is clocked only when the reference channel samples. The dummy channels are utilized to equalize the input load over time as they sample one after another to fill the gap where the reference channel does not sample. There is no need for any other dummy channels for timing calibration channel since it has low kickback noise over input driver. Each parallel channel operates at 111 ms/s while the reference channel runs at 105 ms/s. The aggregate sampling speed of the converter is 2 GS/s, and 52-dB SNDR is accomplished near Nyquist frequencies.     

智能电网大数据流式处理方法与状态监测异常检测

针对智能电网大数据流的实时性、易失性、无序性等特点,提出智能电网大数据的实时流处理框架,实现数据收集、数据缓冲与流式计算,满足状态监测异常检测与用电数据分析等快速处理需要。通过采集系统节点监听数据源变化并实时收集数据,利用消息订阅模式对数据进行缓冲,解决数据采集与流式计算速度不一致的问题。提出一种基于Storm的状态监测数据流滑动窗口处理方法,在规定时间内分批处理状态监测数据流,保证数据的连续计算,通过阈值判断进行异常检测。实验结果表明,在集群规模一定的条件下,适当地改变工作进程数以及执行器线程的并发数设置,可以增大滑动窗口的元件吞吐量,提高状态监测异常检测的实时处理效率。     

基于模糊粗糙集和神经网络的短期负荷预测方法

针对采用神经网络进行电力系统短期负荷预测时其网络输入变量的选择是影响预测效果的关键问题，该文提出使用模糊粗糙集理论解决这一问题：对采集到的信息进行特征提取、形成决策表；利用模糊粗糙集理论进行属性约简、去除冗余信息；用得到的属性作为BP网络的输入进行训练预测。该方法既全面考虑了影响负荷预测的历史时间序列、气象等各种因素，为合理地选择神经网络的输入变量提供了一种新的方法，又避免了由于输入变量过多而导致神经网络拓扑结构复杂、训练时间长等不足。计算实例表明，文中提出的方法是有效且可行的。     

EEG topography recognition by neural networks

Electroencephalography (EEG) pattern-recognition studies were carried out using EEG topography (readiness potential, or RP, spatiotemporal patterns) generated the moment before voluntary movements of muscles. RPs generated prior to pronouncing syllables and controlling a joystick were studied by experiments and simulation. The spatiotemporal patterns of RPs were measured by multichannel surface electrodes pasted on the subject's scalp. Backpropagation neural networks were used for RP pattern recognition. The results show that RPs generated prior to syllable pronouncement contain some information about those syllables, and that RPs generated prior to joy stick movements contain information on the direction of intended movement. They also show that neural networks can be used to recognize EEG information and so create a new type of man-machine interface for data input.     

基于多种仪器总线多通道示波器校准系统

示波器是一种重要的电子设备,在生产测试中广泛应用.目前绝大多数的计量单位对示波器仪器校准采用的传统手动校准方案,少数的单位使用基于GPIB总线的半自动校准系统完成单一通道的校准.介绍了基于多种仪器总线的多通道示波器校准系统,采用VISA的软件架构支持GPIB、TCP/IP、USB、RS232仪器总线,利用NI的PXI开...     

Epileptic seizure detection.

In this study, radial basis function (RBF) neural networks are used to identify seizure or preseizure states. As input to the RBF networks the study used raw EEG data, coefficients from a Fourier transform, and wavelet decomposition of the raw data. An RBF network consists of an input layer, a single hidden layer, and an output node. The use of half-second windows of raw data as input demonstrates the ability of the RBF network to learn differences in the patterns of ictal and interictal EEG data without feature extraction. Wavelet decomposition of the narrow window of raw data improves performance while transformation of a wider window, up to about five seconds, improves it even further. The ability of wavelet decomposition to transform five seconds of raw data into a vector of manageable length without substantial loss of relevant information makes it an effective tool for preprocessing EEG data     

基于宽频功率测量系统的高压电机能效计量检测平台设计

当前设计的高压电机能效计量检测平台的检测耗时较长,存在效率低和检测准确率较差的问题,因此提出基于宽频功率测量系统的高压电机能效计量检测平台设计。硬件部分设计了绝缘电阻信号采集模块、温度信号采集模块和局部放电信号采集模块,通过继电器开关控制继电器的闭合,避免了电流换向带来的测量误差,提高了电阻检测精度。在软件设计中实现了登陆窗口功能、管理员功能、历史数据查询功能、状态显示功能和报警功能。计算输入信息参数,采用宽频功率系统及开关滤波器组实现基波与谐波的分开测量,得到输出参数,实现高压电机能效的计量和检测。实验结果表明,所提方法设计的检测平台的检测耗时较短、检测结果准确率高,具有一定的实际应用性。     

基于Android系统的电力谐波检测仪

基于目前电力系统电能质量现状设计了一款基于Android系统的便携式电力谐波检测仪。介绍了检测仪硬件实现方案和软件设计方法。数据采集模块基于16位,6通道同步采样芯片AD7606,上层Cortex-A9平台搭载Android操作系统,完成检测系统配置,实现了数据处理、数据存储、人机界面控制功能。采用加窗FFT算法完成数据处理,降低了栅栏效应和频谱泄漏造成的检测误差。测试结果表明,该检测仪能够提取电能谐波信息,满足电力谐波检测要求。     

Artificial neural nets for K-complex detection

An explorative study was initiated to determine whether artificial neural nets (ANNs) can be used to detect K-complexes in EEGs (electroencephalograms). K-complexes are relatively large waves with a duration of between 500 and 1500 ms often seen during sleep stage 2. Sleep spindles (bursts of rhythmic activity with a frequency of 12 to 16 Hz) are almost always observed in the neighborhood of K-complexes. The data and methods used to analyze K-complex are described. In all cases, a multilayer backpropagation ANN was used. The number of input nodes and hidden layers varied. Two different strategies were used to prepare the input to the ANN, and results for both are presented. The results indicate that the neural net approaches used are not adequate for the detection of K-complexes.     

多通道直流大电流比较仪的稳定性

为了解决多通道直流大电流比较仪的通道封闭问题,对多通道直流大电流比较仪的稳定性进行了研究。设计的多通道直流大电流比较仪的每个通道铁心间有一段空气隙,以消除铁心的剩磁;每个通道构成一个独立的直流比较仪,各通道采用双重磁检测传感原理与方法,提高了稳定性;基于劳斯判据可分析线性系统的稳定性,根据劳斯阵列的第一列元素符号为正值可判断多通道直流大电流比较仪的闭环系统是否稳定。计算结果表明:设计的多通道直流大电流比较仪,各通道运行稳定,不会出现通道封闭的情况。     

基于ATMega8单片机的数字式温度传感器采集模块

本文介绍了一种应用于冰箱、空调等产品进行温度检测的数据采集模块,该模块以单片机ATMEGA8为处理器,管理8个通道温度传感器DSl8820的数据采集、4个通道的开关量输入和4个通道的开关量输出。关键词：ATMega8;DSl8820;温度测量;数据采集模块     

一种新型远距离无线双向通信检测系统研究

为满足测井过程中需要实时准确获得井下温度压力等相关数据,以及测试人员对井下工作系统的远距离实时操作的需求,设计了一种新型远距离无线双向通信检测系统。该系统由井下模块与地面模块组成。针对井下通信信道复杂,上行下行信号需在同一信道中传输,信号衰减严重且信噪比低的难点,设计了基于时频分复用协议的低频电磁场信道,并采用小波去噪等方法解决微弱信号测量的难题。试验表明,该系统在实地工况下首次成功实现套管井中的双向通信,对测井等远距离检测技术发展具有积极意义。     

基于PLC的多路温度监控系统

摘要介绍了一种性能价格比较高、又能充分利用PLC特点的多路温度监控系统。在无需增加模拟量输入模块数量的情况下，使系统处理温度的通道函数由4路扩展到48路甚至更多，并能实现恒温控制功能。重点阐述了系统的硬件构成、各部分的主要作用及系统软件的设计过程。