神经网络在机电零件制造成本预估系统的应用研究

由于激烈的市场竞争,企业必须建立完善的设计决策支持系统,针对这种需求,本文提出了一种解决机械零件在概念设计阶段的制造成本预估问题的系统方法。在数据库的支持下,首先建立合适的神经网络模型,用已有数据对其进行训练直至其收敛,从而可以进行预估计算。     

基于485总线的自动分诊系统

本文介绍一种以单片机为核心，基于485总线的自动分诊系统：它可为病人就医提供清晰的指引：并提供统计功能，能准确地对医师的工作量进行客观的统计；还提供医师与分诊员之间的联络(对讲功能)。其价格便宜，操作简便，易于上手。     

校园网DDoS攻击防御平台建设方案研究

分布式拒绝服务攻击(DDoS)已经成为互联网最大的威胁之一。分析了校园网现状,建立了业务模型,提出了流量清洗设备的功能要求,设计了一种校园网DDoS攻击防御平台的设计方案。并分析了DDOS攻击清洗方案的流量牵引技术、触发技术、流量清洗技术与流量回注技术。     

Real-time Safety Helmet-wearing Detection Based on Improved YOLOv5

Safety helmet-wearing detection is an essential part of the intelligent monitoring system. To improve the speed and accuracy of detection, especially small targets and occluded objects, it presents a novel and efficient detector model. The underlying core algorithm of this model adopts the YOLOv5 (You Only Look Once version 5) network with the best comprehensive detection performance. It is improved by adding an attention mechanism, a CIoU (Complete Intersection Over Union) Loss function, and the Mish activation function. First, it applies the attention mechanism in the feature extraction. The network can learn the weight of each channel independently and enhance the information dissemination between features. Second, it adopts CIoU loss function to achieve accurate bounding box regression. Third, it utilizes Mish activation function to improve detection accuracy and generalization ability. It builds a safety helmet-wearing detection data set containing more than 10,000 images collected from the Internet for preprocessing. On the self-made helmet wearing test data set, the average accuracy of the helmet detection of the proposed algorithm is 96.7%, which is 1.9% higher than that of the YOLOv5 algorithm. It meets the accuracy requirements of the helmet-wearing detection under construction scenarios.     

并行算法的FP描述及其脉动化的判定

本文通过引入流及流上的递归方程,增强了FP表达并行算法的能力,有效地克服了用FP描述循环数据依赖及状态记忆的困难,文中给出了并行算法的FP描述及其可脉动化的判定定理。同时说明,许多在图上研究的脉动方法可以方便地应用到用FP描述脉动化的研究中去。     

形变冗余的改进GHT目标双层定位方法研究

经典广义Hough变换可以较好地解决非形变目标定位问题,但对于存在形变的目标定位问题存在不少困难。为解决该问题,同时考虑如何提高检测定位速度与减少存储消耗, 在粗定位与精确定位两级框架下提出基于改进GHT形变目标两层定位快速算法。粗定位过程首先利用图像的局域二进制模式的直方图特征对图像进行全局搜索,检测出目标大致范围 ;在精确定位过程中,通过建立模板图像边缘像素的R表,使待检测图像边缘像素在约束的参数范围内依据该R表进行局部搜索,并通过一个投票结果散布窗对得到的累积矩阵进行 集中化处理,达到把每一点邻域内投票结果集中在某点的目的,从而给出最后的检测结果。实验表明,本文算法能够较好的解决一定程度形变目标的定位问题,同时减少了运算时 间以及存储消耗,检测稳定性高,具有一定应用意义。     

NMST:一种基于线段树的持久性内存管理优化方法

新型非易失存储介质(Non-Volatile Memory,NVM)的出现引发了编程模型的革新。现有的基于函数库的编程模型为存储系统提供的ACID特性解决了数据一致性问题,但是在分配持久性内存时,延迟较大,不能很好地满足应用程序对动态内存分配速度的要求。针对现有函数库编程模型中存在持久化内存管理和分配低效的问题,以目前最具代表性的函数库编程模型NVML为基础,提出了一种基于线段树的持久性内存管理分配优化方法NMST;另外,针对线段树在持久性内存分配过程中维护连续空间时开销较大的问题,提出构造多粒度叶子结点的线段树的方法。实验结果表明,相比于NVML原始方法,NMST方法在分配持久性内存时使延迟降低了36.9%,而优化后的NMST方法在分配持久性内存时使延迟降低了43.6%。实验结果也证明,性能提升的大小与调用NVML函数库的程序中实际持久性内存分配的次数及粒度紧密相关。     

Dynamical characteristics of fluid-conveying microbeams actuated by electrostatic force

In this paper, the dynamic characteristics and pull-in instability of electrostatically actuated microbeams which convey internal fluids are investigated. A theoretical model is developed by considering the elastic structure, laminar flow and electrostatic field to characterize the dynamic behavior. In addition, the energy dissipation induced by the fluid viscosity is studied through analyzing the fluid–structure interactions between the laminar fluid flow and oscillating microbeam by comprehensively considering the effects of velocity profile and fluid viscosity. The results indicate that the system is subjected to both the pull-in instability and the fluid-induced instability. It is demonstrated that as the flow velocity increases, both the static pull-in voltage and the dynamic one decrease for clamped–clamped microbeams while increase for clamped-free microbeams. It is also shown that the applied voltage and the steady flow can adjust the resonant frequency. The perturbation viscous flow caused by the vibration of microbeam is manifested to result in energy dissipation. The quality factor decreases with the increment of both the mode order and flow velocity. However, when the oscillating flow dominates, the flow velocity has no obvious effect.     

Leaky‐wave antenna with high gain and wide beam‐scanning angle range based on novel SIW‐CRLH transmission line

A leaky‐wave antenna (LWA) with high gain and wide beam‐scanning angle is proposed in this article using a novel substrate integrated waveguide (SIW) composite left/right‐handed transmission line (CRLH TL). The novel SIW‐CRLH TL is analyzed and the equivalent circuit model is also provided. Considering the continuous phase constant of the balanced SIW‐CRLH TL from negative to positive values, the proposed LWA can obtain a continuous beam steering property from backward to broadside to forward. For verification, a periodic LWA, which is comprised of 10 unit cells of the balanced SIW‐CRLH TL, is fabricated and measured. The measured and simulated results agree well, showing that the proposed periodic LWA operates from has continuous beam‐scanning capabilities of about 90° from backward to forward (including the broadside) with gains of better than 10 dB within the operating band. © 2015 Wiley Periodicals, Inc. Int J RF and Microwave CAE 26:36–41, 2016.