基于忆阻器卷积神经网络的表情识别

忆阻器具有纳米级尺寸、低功耗、类似神经突触等优点,在神经计算、图像分类等领域具有广阔的应用前景。本文提出了一种基于忆阻器卷积神经网络的面部表情识别方法,首先基于忆阻器构建了ResNet卷积神经网络,并对ResNet网络进行剪枝操作,然后将ResNet模型的所有卷积层以及全连接层的权重映射为忆阻器十字交叉阵列中忆阻器的忆导值。实验结果显示忆阻器卷积神经网络模型在FER2013数据集上的识别准确率为63.82%,在CK+数据集上的识别准确率为93.95%。相比与原卷积网路,准确率损失仅分别为0.31%和0.76%。最后测试了忆阻器的非理想特性对准确率的影响,为忆阻器神经网络的实际部署提供参考。     

基于多层特征增强的实时视觉跟踪

为了解决全卷积孪生视觉跟踪网络（SiamFC）出现相似语义信息干扰物使得跟踪目标发生漂移，导致跟踪失败的问题，设计出一种基于多层特征增强的实时视觉跟踪网络（MFESiam），分别去增强高层和浅层的特征表示能力，从而提升算法的鲁棒性。首先，对于浅层特征，利用一个轻量并且有效的特征融合策略，通过一种数据增强技术模拟一些在复杂场景中的变化，例如遮挡、相似物干扰、快速运动等来增强浅层特征的纹理特性；其次，对于高层特征，提出一个像素感知的全局上下文注意力机制模块（PCAM）来提高目标的长时定位能力；最后，在三个具有挑战性的跟踪基准库OTB2015、GOT-10K和2018年视觉目标跟踪库（VOT2018）上进行大量实验。实验结果表明，所提算法在OTB2015和GOT-10K上的成功率指标比基准SiamFC分别高出6.3个百分点和4.1个百分点，并且以每秒45帧的速度运行达到实时跟踪。在VOT2018实时挑战上，所提算法的平均期望重叠率指标超过2018年的冠军，即高性能的候选区域孪生视觉跟踪器（SiamRPN），验证了所提算法的有效性。     

Spray coating process variable and property analysis of UV-curable polyurethane acrylate coating on polycarbonate substrate

A series of spray coating experiments were conducted on an UV-curable, polyurethane-modified, acrylate-based coating formulation with the aim to control defects, coating thickness, and thickness variation. Statistical approaches including design of experiment, residual examination, analysis of variance, and t-test were used in designing the experiments and analyzing data. Viscosity of formulation, atomizing pressure, liquid feeding pressure, distance between nozzle and substrate, and travel speed of substrate were the process variables studied. The ranges of process variables that gave defect-free coating were identified and used in the subsequent experiments to determine process variables and interactions that had significant contribution to the changes in coating thickness and thickness variation. All process variables studied were found to have contribution to the change in coating thickness, but they showed no significant contribution to the variation of coating thickness. No interaction displayed significant contribution. Confirmation tests performed on extra samples prepared with varying coating thicknesses indicated a good agreement with the experimental results. Additional samples were tested for total transmittance, transmission haze, adhesion, surface roughness, hardness, scratch hardness, abrasion resistance, and durability to attack of car wash chemicals. Spray coated samples showed slight improvement in the total transmittance over the uncoated samples, while maintaining the transmission haze and exhibiting rougher surfaces. Only samples with thin coatings were found to possess sufficient adhesion to the substrate. These thin coatings gave improved hardness, scratch hardness, and durability to car wash attack to the level comparable to commercial coated polycarbonate headlamp lenses, whilst giving better abrasion resistance.     

Coordinated voltage control of wind-penetrated power systems via state feedback control

This paper presents an optimal coordinated voltage control scheme for preserving long-term voltage stability of power systems. Linear quadratic integral (LQI) controller is employed to construct this scheme. Also, this paper considers the detailed dynamic model of doubly-fed induction generator (DFIG) wind turbine, synchronous generators, over excitation limiter (OEL) and under-load tap changer (ULTC) system, which are important elements influencing voltage stability of power systems. The proposed approach at each time instance involves following two major steps: First, the power system nonlinear equations are linearized and optimal controllers are obtained by LQI technique. Then, in the second step the system dynamic behavior is investigated via time-domain simulations by applying the attained optimal control signals at the first step. The impact of the proposed coordinated voltage control scheme is evaluated by time domain simulations on a well-known test system, under variable wind speed and fault conditions.     

窗口大小自适应的空间耦合LDPC码译码算法

空间耦合低密度奇偶校验（Spatially-Coupled Low-Density Parity-Check,SC-LDPC）码具有接近香农限性能,基于置信传播译码算法,窗口译码（Windowed Decoding,WD）能够获得较小延时的同时也存在一定的局限性。为了进一步提高WD的译码性能,对SC-LDPC码的窗口译码算法提出了提前终止译码和动态调整窗口大小相结合的改进方法。该方法监测窗口大小的动态变化及相应窗口的平均迭代次数,通过加性高斯白噪声信道下的仿真分析,与传统窗口译码相比,其误码率降低,且计算复杂度更低。     

基于观测器的无人机编队自适应容错控制

针对多无人机编队飞行中的僚机故障问题,设计了一种基于观测器的自适应容错控制方法。首先,基于领导跟随法建立了无人机编队模型及僚机故障的编队模型,并将其划分为位置子系统与偏航角子系统。其次,基于观测器技术对位置子系统中的状态和故障进行观测,并结合观测的状态和故障信息构造状态反馈控制律;然后,基于自适应方法给出偏航角子系统的控制设计方案,并用Lyapunov理论证明系统跟踪误差最终有界收敛。通过仿真,本文算法在发生故障后对系统的完全跟踪时间和稳态误差分别比基于鲁棒故障估计的方法最大降低了76%和70.3%,并且均比传统观测器的方法明显减少较大,证明了本文算法能更好的克服偏差故障带来的不利影响,有效实现四旋翼无人机群的编队飞行。     

基于心冲击图和BP神经网络的心率异常分类研究

心率变异性(Heart rate variability,HRV)被广泛用于临床自主神经系统评估和心率异常分类,传统的HRV分析基于心电图(Electrocardiogram,ECG)、光容积图(Photoplethysmography,PPG)和远程光容积图(Remote PPG,RPPG),这些方法存在诸多不足:(...     

Modified cuckoo search algorithm for short-term hydrothermal scheduling

This paper proposes a modified cuckoo search algorithm (MCSA) for solving short-term hydrothermal scheduling (HTS) problem. The considered HTS problem in this paper is to minimize total cost of thermal generators with valve point loading effects satisfying power balance constraint, water availability, and generator operating limits. The MCSA method is based on the conventional CSA method with modifications to enhance its search ability. In the MCSA, the eggs are first sorted in the descending order of their fitness function value and then classified in two groups where the eggs with low fitness function value are put in the top egg group and the other ones are put in the abandoned one. The abandoned group, the step size of the Lévy flight in CSA will change with the number of iterations to promote more localized searching when the eggs are getting closer to the optimal solution. On the other hand, there will be an information exchange between two eggs in the top egg group to speed up the search process of the eggs. The proposed MCSA method has been tested on different systems and the obtained results are compared to those from other methods available in the literature. The result comparison has indicated that the proposed method can obtain higher quality solutions than many other methods. Therefore, the proposed MCSA can be a new efficient method for solving short-term fixed-head hydrothermal scheduling problems.     

Modelling of carbon–carbon composite manufacturing processes

This paper is devoted to the modelling of technological processes of manufacturing of siliconized carbon–carbon composites. The developed model describes the changes that occur in the properties of the composites (strength, elastic moduli, shrinkage) during the technological cycle of manufacturing and also the residual stresses generated in composite structures. It is shown that the level of the residual stresses and the character of changes in the properties of carbon–carbon composites essentially differ from those of polymer–matrix composites.