Combined Novel Gate Level Model and Critical Primary Input Sharing for Genetic Algorithm Based Maximum Power Supply Noise Estimation

A gate level maximum power supply noise （PSN） model is defined that captures both IR drop and di/dt noise effects. Experimental results show that this model improves PSN estimation by 5.3% on average and reduces computation time by 10.7% compared with previous methods. Furthermore,a primary input critical factor model that captures the extent of primary inputs＇ PSN contribution is formulated. Based on these models,a novel niche genetic algorithm is proposed to estimate PSN more effectively. Compared with general genetic algorithms, this novel method can achieve up to 19.0% improvement on PSN estimation with a much higher convergence speed.

Realization of Toffoli gate operation using four-level atoms in cavity QED system

This paper proposes a scheme for realization of a three-qubit Toffoli gate operation using three four-level atoms by a selective atom-field interaction in a cavity quantum electrodynamics system. In the proposed protocol, the quantum information is encoded on the stable ground states of atoms, and atomic spontaneous emission is negligible as the large atom-avity detuning effectively suppresses the spontaneous decay of the atoms. The influence of the dissipation on fidelity and success probability of the three-qubit Toffoli gate is also discussed. The scheme can also be applied to realize an N-qubit Toffoli gate and the interaction time required does not rise with increasing the number of qubits.

A scheme for conditional quantum phase gate via bimodal cavity and a∧-type three-level atom

We propose a scheme to implement a two-qubit conditional quantum phase gate for the intracavity field via a single three-level ∧-type atom driven by two modes in a high-Q cavity. The quantum inforraation is encoded on the Fock states of the bimodal cavity. The gate＇s averaged fidelity is expected to reach 99.8%.

Algorithm of Dynamic Operation Process of Hydraulic Automatically Operated Canals with Constant-Downstream Level Gates

On the basis of analysis the governing process of downstream water level gates AVIO and AVIS, a mathematical model for simulation of dynamic operation process of hydraulically automated irrigation canals instalIed with AVIO and AVIS gates is presented, the main point of this rnathematical model is firstly applying a set of unsteady flow equations （St. Venant equations here） and treating the condition of gate movement as its dynamic boundary, and then deeoupling this interaction of gate movement with the change of canal flow. In this process, it is necessary to give the gateg open-loop transfer function whose input is water level deviation and output is gate discharge. The result of this simulation for a practical reach has shown it has satisfactory accuracy.

Generation of an N -qubit phase gate via atom–cavity nonidentical coupling

A scheme for approximate generation of an N-qubit phase gate is proposed in cavity QED based on nonidentical coupling between the atoms and the cavity. The atoms interact with a highly detuned cavity-field mode, but quantum information does not transfer between the atoms and cavity field, and thus the cavity decay is negligible. The gate time does not rise with an increase in the number of qubits. With the choice of a smaller odd number l （related to atom-cavity coupling constants）, the phase gate can be generated with a higher fidelity and a higher success probability in a shorter time （the gate time is much shorter than the atomic radiative lifetime and photon lifetime）. When the number of qubits N exceeds certain small values, the fidelity and success probability rise slowly with an increase in the number of qubits N. When N→∞, the fidelity and success probability infinitely approach 1, but never exceed 1.

A scheme for the implementation of unconventional geometric phase gates with trapped ions

We propose a scheme for the realization of unconventional geometric two-qubit phase gates with two identical two-level ions, In the present scheme, the two ions are simultaneously illuminated by a standing-wave laser pulse with its pulse frequency being tuned to the ionic transition. The gate operation time can be much shorter, making the system robust against decoherence. In addition, we choose the appropriate experimental parameters to construct the geometric phase gate in one step, and thus avoid implementing the pure geometric single qubit operation.

System-on-Chip Design Using High-Level Synthesis Tools

This paper addresses the challenges of System-on-Chip designs using High-Level Synthesis (HLS). HLS tools convert algorithms designed in C into hardware modules. This approach is a practical choice for developing complex applications. Nevertheless, certain hardware considerations are required when writing C applications for HLS tools. Hence, in order to demonstrate the fundamental hardware design concepts, a case studyis presented. Fast Fourier Transform (FFT) implementation in ANSI C is examined in order to explore the important design issues such as concurrency, data recurrences and memory accesses that need to be resolved before generating the hardware using HLS tools. There are additional language constraints that need to be addressed including use of pointers, recursion and floating point types.

Generation of four-atom cluster states in thermal cavity and implementing remote controlled not gate

We propose an experimentally feasible scheme for preparing a four-atom cluster state in a thermal cavity. In the scheme, the cavity field is only virtually excited and the photon-number-dependent part in the effective Hamiltonian is cancelled so that the system is insensitive to the cavity decay and the thermal field. At the same time, the scheme can be generalized to prepare n-atom cluster states with the success probability 100%. In addition, using the four-atom cluster state, we also propose a simpler scheme for implementing a remote-controlled not gate （CNOT） without the Bell states measurement.

Research on Night Vision System Based on Range-Gated Imaging

A design of low-light-level night vision system is described,which can image objects selectively in the specific space. The system can selectively image some objects in specific distances,meanwhile ignore those shelters on the way of observation by combining an intensifying charge coupled device(ICCD) with a near infrared laser assisted in vision,whose operation wavelength matches with the photocathode of the image tube,and adopting the gated mode and adjustable time-delay. A semiconductor laser diode of 100 W in peak power is chosen for illumination. The laser and the image tube operate in 150 ns pulse width and 2 kHz repeat frequency. Some images of different objects at the different distances within 100 m can be obtained clearly,and even behind a grove by using a sampling circuit and a delay control device at 100 W in peak power of semiconductor laser diode,150 ns in pulse width of laser and image tube,2 kHz in repeat frequency.

六垛南闸下游引河泥沙淤积特性数值分析

为了明确六垛南闸下游引河的淤积特性,基于Mike软件数值计算求解了该闸下游引河的水沙运动,对闸下引河冲淤进行了预测分析,并通过闸下实测水位和预测水位进行对比验证了计算模型的正确性。结果表明:各泄流量时下游引河的水流均为顺岸流动,流速分布规律相同;相同流量时沿程流速整体呈增大趋势。闸下引河的冲淤特性呈现出潮位越低,冲淤量越大的特点。建议六垛南闸开闸冲淤应尽量选择入海口潮位<0.3m时进行,同时可按照10d一个周期进行开闸冲淤。研究成果可为江苏省同类沿海挡潮闸防淤减淤提供参考。

结合多粒度视图动态融合的多模态方面级情感分析

为了解决以往多模态方面级情感分析研究中存在的特征提取不充分、数据噪声未被有效处理以及多模态数据中的复杂交互被忽视等问题,提出了一种多粒度视图动态融合模型(multi-granularity view dynamic fusion model,MVDFM)。从粗粒度和细粒度两个视角,对文本和图像数据进行向量化编码,以便充分捕捉数据特征,增强模型信息表达能力;提取文本、图像的多粒度视图特征,并设计动态门控自注意力机制,对细粒度级的文本、图像视图进行降噪,进一步保证特征提取质量;为了挖掘不同粒度上多视图之间的互补性和一致性,提出一种三视图分解高阶池化机制,对多粒度视图特征进行两阶段动态融合,得到最终的目标方面词情感极性。实验结果表明,该模型在公共数据集Twitter-2015和Twitter-2017上的准确率和F1值分别达到了78.69%、74.48%以及72.77%、71.61%,相较于最优基线模型分别提升了0.55、0.88个百分点,以及1.67、2.45个百分点。说明该方法能够充分利用多模态数据中包含的深层语义信息,并有效挖掘与目标方面词相关的重要信息,从而提高方面级情感预测效果。

基于篇章级信息特征增强的中文突发事件检测

事件检测旨在检测句子中的触发词并将其分类为特定的事件类型。针对目前大多数的中文事件检测方法存在篇章内句子间相互依赖信息利用不充分的问题,提出了基于篇章级信息特征增强的中文突发事件检测方法。首先,通过BERT(Bidirectional Encoder Representations from Transformers)预训练语言模型编码字向量,使用平均池化将字向量转换成句子表示信息;其次,利用双向门控循环神经网络(Bidirectional gated recurrent neural network,Bi-GRU)学习句子内和篇章内的上下文隐层信息,接着通过注意力机制(Attention Mechanism)分别得到句子级信息和由若干句子间的相互依赖信息生成的篇章级信息;最后,将篇章级信息融合到句子级信息上后与字向量拼接,再使用条件随机场完成对句子中触发词的识别和标注。实验结果表明,该方法有效的提升了中文突发事件检测效果,F1值达到79.95%。

面向GaN驱动的高噪声抗扰度电平位移电路

在高压GaN半桥栅驱动系统应用中,需要通过电平位移电路来实现信号在不同电压域之间的转换。为了保证转换过程中的信号完整性,设计了一种面向GaN驱动的高噪声抗扰度电平位移电路。在半桥开关节点电压发生快速切换时,针对电路内部大寄生电容节点充放电导致输出误翻转的问题,采用交叉耦合方式抑制共模噪声电流传递,实现了较高的噪声抗扰度。另外,采用电压-电流转换技术提高了抗负压能力。基于0.8μm 600 V高压BCD工艺进行电路设计。仿真结果表明,该电平位移电路平均传输延时为5.62 ns,dV/dt噪声抗扰度为200 V/ns,在6 V电源电压下允许开关节点负压低至-4.5 V。

基于异构平台的卷积神经网络加速系统设计

在计算和存储资源受限的嵌入式设备上部署卷积神经网络,存在执行速度慢、计算效率低、功耗高的问题。提出了一种基于异构平台的新型卷积神经网络加速架构,设计并实现了基于MobileNet的轻量化卷积神经网络加速系统。首先,为降低硬件资源消耗以及数据传输成本,采用动态定点数量化和批标准化融合的设计方法,对网络模型进行了优化,并降低了加速系统的硬件设计复杂度;其次,通过实现卷积分块、并行卷积计算、数据流优化,有效提高了卷积运算效率和系统吞吐率。在PYNQ-Z2平台上的实验结果表明,此加速系统实现的MobileNet网络推理加速方案对单幅图像的识别时间为0.18 s,系统功耗为2.62 W,相较于ARM单核处理器加速效果提升了128倍。

基于四电平编码与弱光通信的极低功耗传光式电流互感器

针对传光式电流互感器高压侧现有的母线取能和激光供能方式存在供电死区、能量转换效率低、使用寿命不足等问题,提出一种基于四电平编码与弱光通信的极低功耗光电传输系统方案。首先,设计三态门电路将信号采集模块的采样时钟和数据信号叠加为一个四电平信号。然后,利用超高亮度且极低功耗的LED将四电平信号转换为光信号并通过光纤传输至低压侧。最后,在低压侧通过高灵敏度雪崩光电二极管探测器进行弱光检测后输入解码电路实现信号解码。将极低功耗光电传输系统应用于传光式电流互感器中进行实验验证,测量结果满足0.2级精度要求,系统高压侧的平均功耗可降低至3.91 mW,保证测量精度的同时大幅降低了功耗。由于高压侧的功耗极低,故可采用大容量电池或小功率光伏系统等可靠供能方案,从而解决了高压侧的供能问题,提高了系统的稳定性和可靠性。

水库水位的VMD-CNN-GRU混合预测模型

水库水位预测为其运营、防洪、水资源调度管理提供了重要决策支持.准确可靠的预测对水资源的优化管理起着至关重要的作用.针对水库水位数据的非线性、不稳定性以及复杂的时空特性,提出一种融合自适应变分模态分解(VMD)、卷积神经网络(CNN)和门控循环单元(GRU)的混合水库水位预测模型.VMD通过对水位序列进行分解消除噪声,CNN用于有效提取水位数据的局部特征,GRU用于提取水位数据的深层时间特征.以葠窝水库日水位为例,与多个相关模型对比分析,结果表明:精度方面,新模型在选取的评价指标上均表现最佳;运算效率方面,本文选择的GRU与长短时记忆网络(LSTM)相比,运算效率显著提高.新模型预测的高精度、高运算效率更能满足实际水库水位实时调度的需求.

一种多任务联合训练的机器阅读理解模型

在机器阅读理解任务中,如何在包含不可回答问题的情况下提高答案的准确性是自然语言处理领域的一项重要挑战.虽然基于深度学习的机器阅读理解模型展现出很好的性能,但是这些模型仍然存在抽取特征冗余、语义信息不全面、问题分类任务和答案抽取任务耦合性不强的问题.为了解决以上问题,本文提出一种结合门控机制和多级残差结构的多任务联合训练模型GMRT(Gated Mechanism and Multi-level Residual Structure for Multi-task Joint Training),以提升机器阅读理解任务中答案预测的准确性.GMRT构建门控机制来筛选交互后的关联特征,从而控制信息的流动.采用多级残差结构分别连接注意力机制和门控机制,保证每个阶段都保留原始语义信息.同时,通过边缘损失函数对问题分类任务和答案抽取任务联合训练,确保预测答案过程中任务之间的强耦合性.在SQuAD2.0数据集上的实验结果表明,GMRT模型的EM值和F1值均优于对比模型.

基于FPGA和行折叠的稀疏矩阵向量乘优化

稀疏矩阵向量乘(SpMV)是科学与工程计算中的一个关键内核。由于稀疏矩阵中不规则的数据分布和SpMV计算中不规则的访存操作,SpMV在多核CPU和GPU等设备上的性能与这些设备的理论峰值还具有较大差距。现有的CPU和GPU由于在架构上受到限制,导致它们无法很好地利用稀疏矩阵的特殊结构来加速SpMV计算,而现场可编程门阵列(FPGA)可以通过自定义电路实现高效的并行运算,能够更好地处理稀疏矩阵的计算和存储问题。基于FPGA提出了一种SpMV优化方法,该优化方法基于高级综合的流式处理引擎,采用了一种自适应多行折叠的SpMV优化策略。该方法通过行折叠减少了处理引擎中零元的无效存储和计算,从而提升了基于FPGA的SpMV计算性能。实验结果表明,相比于现有的FPGA实现方案,设计的基于行折叠优化的数据流引擎实现了最高1.78倍和平均1.15倍的加速。

结温导向的牵引变流器寿命优化控制

结温是影响绝缘栅双极型晶体管(IGBT)寿命的主要因素。为了提高地铁两电平牵引变流器寿命,提出一种降低结温的异步牵引电机双矢量模型预测转矩控制(DVMPTC)策略。将传统DVMPTC的第二个电压矢量选择范围缩小在与第一个电压矢量不切换或仅切换一次的范围,降低IGBT开关损耗的同时减小系统计算量;在代价函数中约束IGBT及其反并联二极管的损耗,动态加入损耗因子,并赋予权重系数,使得最优矢量的选择兼顾控制性能与降低损耗。通过仿真分析,本文所提方法相较于传统基于分段调制算法的直接转矩控制策略,降低了结温及其波动,提高了牵引变流器寿命。

连续时间门限完全服务两级轮询系统性能分析

为区分业务优先级以及提高系统公平性与稳定性,在简化求解过程的基础上,提出一种连续时间门限完全服务两级轮询系统。系统状态由马尔科夫链得出并根据其建立数学模型,系统的平均循环周期、平均排队队长等性能参数由对数学模型的概率母函数进行求导所得,将求导结果与仿真实验进行对比,验证理论分析的准确性。将该系统与单级系统、其它两级系统进行对比,验证了该系统在保证业务优先级的同时更具公平性与稳定性。