基于近似高斯金字塔的视觉注意模型快速算法

利用输入图像的近似高斯金字塔,将经典的基于显著性的视觉注意模型改造为时空开销更小的版本,从而使其更加适合在嵌入式实时系统中实现.首先采用矩形窗口近似圆形窗口,矩形平均算子近似高斯卷积核;然后采用“先做行累加,再做列累加”的方法来实现矩形平均算子,并直接采样计算出各个特征通道的显著性分布图,该算法关于输入图像像素点个数具有线性时间复杂度;最后,还给出了在显著性分布图中抑制已提取区域显著性的快速算法.在Berkeley分割图像库上的实验结果表明,该方法极大地减小了系统实现的时空开销,且输出结果的误差在可接受范围内.提出的用矩形窗口近似圆形窗口,用矩形平均算子近似高斯卷积核的方法,还适用于其他需要在嵌入式实时系统中实现的图像处理问题.     

基于图像质量分区的指纹特征提取

低质量指纹图像的特征提取和变形指纹的匹配是当前指纹识别研究中的两个主要问题。很多算法在特征提取时不区分高、低质量区域,结果在高质量区域耗费了过多的运算时间和计算资源。本文提出了一种基于图像质量分区的指纹特征提取方法,先用一种简单的图像区域质量计算方法评价各区域的图像质量,然后对高质量区域直接从灰度图像跟踪纹线、提取节点,对低质量区域执行传统的方向计算、增强、二值化和细化后提取特征。实验结果表明,该方法不仅提高了特征提取的速度,在准确性上也有所提高。     

FTPA: Supporting Fault-Tolerant Parallel Computing through Parallel Recomputing

As the size of large-scale computer systems increases, their mean-time-between-failures are becoming significantly shorter than the execution time of many current scientific applications. To complete the execution of scientific applications, they must tolerate hardware failures. Conventional rollback-recovery protocols redo the computation of the crashed process since the last checkpoint on a single processor. As a result, the recovery time of all protocols is no less than the time between the last checkpoint and the crash. In this paper, we propose a new application-level fault-tolerant approach for parallel applications called the Fault-Tolerant Parallel Algorithm (FTPA), which provides fast self-recovery. When fail-stop failures occur and are detected, all surviving processes recompute the workload of failed processes in parallel. FTPA, however, requires the user to be involved in fault tolerance. In order to ease the FTPA implementation, we developed Get it Fault-Tolerant (GiFT), a source-to-source precompiler tool to automate the FTPA implementation. We evaluate the performance of FTPA with parallel matrix multiplication and five kernels of NAS Parallel Benchmarks on a cluster system with 1,024 CPUs. The experimental results show that the performance of FTPA is better than the performance of the traditional checkpointing approach.     

雪崩光电探测器的击穿效应模型分析

基于CMOS工艺制备了空穴触发的Si基雪崩探测器(APD),基于不同工作温度下器件的击穿特性,建立空穴触发的雪崩器件的击穿效应模型。根据雪崩击穿模型和击穿电压测试结果,拟合曲线得到击穿电场与温度的关系参数(dE/dT),器件在250～320 K区间内,击穿电压与温度是正温度系数,器件发生雪崩击穿为主,dV/dT=23.3 mV/K,其值是由倍增区宽度以及载流子碰撞电离系数决定的。在50～140 K工作温度下,击穿电压是负温度系数,器件发生隧道击穿,dV/dT=-58.2 mV/K,其值主要受雪崩区电场的空间延伸和峰值电场两方面因素的影响。     

CuI Encapsulated within Single-Walled Carbon Nanotube Networks with High Current Carrying Capacity and Excellent Conductivity

High current carrying capacity and high conductivity are two important indicators for materials used in microscale electronics and inverters. However, it is challenging to obtain high conductivity and high current carrying capacity at the same time since high conductivity requires a weakly bonded system to provide free electrons, while high current carrying capacity requires a strongly bonded system. In this paper, CuI@SWCNT networks by filling the single-walled carbon nanotubes (SWCNTs) with CuI is ingeniously prepared. CuI@SWCNT shows good stability due to the confinement protection of SWCNTs. Through the host-guest hybridization, CuI@SWCNT networks exhibit a current carrying capacity of 2.04 × 10⁷ A cm⁻² and a conductivity of 31.67 kS m⁻¹. Their current carrying capacity and conductivity are significantly improved compared with SWCNT. The Kelvin probe force microscopy measurements show a drop of surface potential energy after SWCNT filled with CuI, indicating that the CuI guest molecules regulate the position of the Fermi level of SWCNTs, increasing carrier concentration, achieving high conductivity and high current carrying capacity. This study offers ideas and solutions for the regulation of high-performance carbon tube networks, which hold great promise for future applications in carbon-based electronic devices.     

Nonlinear cascade control of an electro-hydraulic actuator with large payload variation

Electro-hydraulic actuators have been widely used in industrial production, but the unknown variable payload seriously affects its position control accuracy. Therefore, a radial basis function neural network disturbance observer is designed to estimate the lumped disturbance force through strong online learning ability in the absence of force sensor. Besides, a nonlinear cascade controller with double loop structure is proposed in this paper. A global fast terminal sliding mode control method is firstly applied in the outer loop position system, which can eliminate chattering and improve convergence speed comparing to traditional sliding mode control. The inner loop force system adopts a backstepping control method to calculate the actual input of the whole system. Theoretical analysis indicates that the proposed controller is stable even if existing time-variant disturbance. Moreover, three comparative controllers are designed and tested in both simulations and experiments. Comparative results show that the developed method has absolute average errors of 1.14 and 0.49 mm in different position tracking, which means more satisfactory tracking performance compared to the contrast controllers.     

基于离散分数阶傅里叶变换的二维跳频通信系统及性能分析

传统跳频(FH)通信技术具有抗干扰能力强、截获概率低等优点,广泛应用在军民领域。针对检测传统跳频的手段越来越成熟,信息易被截获的问题,该文借鉴正交频分复用(OFDM)系统框架,提出一种基于离散分数阶傅里叶变换(DFrFT)的时宽与起始频率跳变的分数阶跳频(FrFT-FH-VTFB)系统,设计了一种新的系统框架,实现信息隐蔽传输的同时,通过DFrFT的工程实现规避传统跳频工程应用中跳速受频率合成器限制的问题。该系统通过两组不同伪随机序列选取时宽与起始频率跳变的Chirp基信号,实现系统参数的多维变换,打破系统的周期特性。此外,建立了系统发送与接收两端数学模型,并在此基础上推导了系统在白噪声信道下的理论误码率。仿真结果表明,该文所设计的系统有较好的抗衰落性能;且功率谱淹没在噪声之下,时频域特征无明显周期特性,有较好的隐蔽性。     

基于弹性秘密共享的多方门限隐私集合交集协议

$ (t, n) $门限隐私集合交集协议, 指$ N $个参与者各自拥有大小为$ n $的隐私集合, 在不泄露自身隐私信息的前提下, 如果各参与者交集数量大于门限值$ t $, 则参与各方能够获得交集信息, 其有广泛的应用, 如指纹识别、在线拼车、相亲网站等. 然而现有门限隐私集合交集协议大多针对两方参与者进行研究, 对多方门限隐私集合交集协议的研究仍存在许多挑战, 现有的多方门限隐私集合交集协议使用全同态加密等开销较大的公钥算法, 尚没有有效实现. 针对上述问题, 结合弹性秘密共享、布隆过滤器提出两种有效的多方门限隐私集合交集协议, 并首次仿真实现了协议. 首先, 设计一种新的布隆过滤器构造方法, 将弹性秘密共享生成的份额与参与方的集合元素相对应, 通过查询布隆过滤器获取的秘密子份额能否重构出正确秘密来判断各方交集是否达到门限值, 有效防止交集基数的泄露. 设计的第1个协议避免使用开销较大的公钥算法, 当设置安全参数$ \lambda $为128, 集合大小为$ {2^{14}} $, 门限值为$ 0.8n $时, 在三方场景下协议在线阶段的时间成本为191 s. 此外, 为了能在半诚实模型下抵抗至多$ N - 1 $个敌手合谋, 在第1个协议基础上结合不经意传输设计一种该协议的变体, 相同条件下, 在线阶段时间成本为194 s. 最后通过安全证明, 证明上述协议在半诚实模型下是安全的.     

Classification of Imagined Speech EEG Signals with DWT and SVM

With the development of human–computer interaction technology, brain–computer interface (BCI) has been widely used in medical, entertainment, military, and other fields. Imagined speech is the latest paradigm of BCI and represents the mental process of imagining a word without making a sound or making clear facial movements. Imagined speech allows patients with physical disabilities to communicate with the outside world and use smart devices through imagination. Imagined speech can meet the needs of more complex manipulative tasks considering its more intuitive features. This study proposes a classification method of imagined speech Electroencephalogram (EEG) signals with discrete wavelet transform (DWT) and support vector machine (SVM). An open dataset that consists of 15 subjects imagining speaking six different words, namely, up, down, left, right, backward, and forward, is used. The objective is to improve the classification accuracy of imagined speech BCI system. The features of EEG signals are first extracted by DWT, and the imagined words are classified by SVM with the above features. Experimental results show that the proposed method achieves an average accuracy of 61.69%, which is better than those of existing methods for classifying imagined speech tasks.