Microprocessors are getting smaller, cheaper and faster. Every day, it is easier to embed more functionality into a smaller space. Embedded processors have become pervasive, and as time goes on, more and more functions that were once implemented with analog circuitry or with electromechanical assemblies are being realized with microcontrollers, ADCs and DACs. Many of these assemblies that are being supplanted by the microprocessor are controlling dynamic processes, which is a good thing, because the microprocessor coupled with the right software is often the superior device.
上传时间： 2016-07-20
上传用户：kunger
In a world where consumers,businesses and organizations are demanding faster,amerter and more seamlessly converged information technology and communization services,concepts such as Driverless Cars,combining advanced automatic controls,artficial intelligences and automotive transport technologies are soon expected to become a commercial realty.At the same time,new blueprints such as industrial internet and industry 4.0 are driving the increased digitization of entire industries and economic sectors,boosting efficiencies and productivity.
标签： ZTE-M-ICT
上传时间： 2020-06-01
上传用户：shancjb
As we enter the next millennium, there are clear technological patterns. First, the electronic industry continues to scale microelectronic structures to achieve faster devices, new devices, or more per unit area. Secondly, electrostatic charge, electrostatic discharge (ESD), electrical overstress (EOS) and electromagnetic emissions (EMI) continue to be a threat to these scaled structures. This dichotomy presents a dilemma for the scaling of semiconductor technologies and a future threat to new technologies. Technological advancements, material changes, design techniques, and simulation can fend off this growing concern – but to maintain this ever-threatening challenge, one must continue to establish research and education in this issue.
标签： ESD-Phenomena-and-the-Reliability
上传时间： 2020-06-05
上传用户：shancjb
Thepredecessorvolumeofthisbookwaspublishedin1996.Intheyears since then, some things have changed and some have not. Two of the things that have not changed are the desire for better models and faster simulations. I performed the original simulations on my “hyperfast” 133-MHz computer! At the time, I thought if I could just getafastercomputer,allofourSPICEproblemswouldbehistory,right? TodayIamsimulatingonacomputerthathasa2.6-GHzprocessorwith 512 MB of RAM, and I would still say that simulations run too slow. The computer technology has evolved, but so have the models. In 1996 wewereperformingsimulationson100-kHzpowerconverters,whereas today I routinely see 1- and 2-MHz power converters.
标签： Switch-Mode Simulation Supply Power
上传时间： 2020-06-07
上传用户：shancjb
红外图像检测技术因具有非接触、快速等优点，被广泛应用于电力设备的监测与诊断 中，而对设备快速精确地检测定位是实现自动检测与诊断的前提。与普通目标的可见光图像相比， 电力设备的红外图像可能存在背景复杂、对比度低、目标特征相近、长宽比偏大等特征，采用原 始的 YOLOv3 模型难以精确定位到目标。针对此问题，该文对 YOLOv3 模型进行改进：在其骨干 网络中引入跨阶段局部模块；将路径聚合网络融合到原模型的特征金字塔结构中；加入马赛克 （Mosaic）数据增强技术和 Complete-IoU（CIoU）损失函数。将改进后的模型在四类具有相似波纹 外观结构的电力设备红外图像数据集上进行训练测试，每类的检测精度均能达到 92%以上。最后， 将该文方法的测试结果与其他三个主流目标检测模型进行对比评估。结果表明：不同阈值下，该 文提出的改进模型获得的平均精度均值优于 faster R-CNN、SSD 和 YOLOv3 模型。改进后的 YOLOv3 模型尽管在检测速度上相比原 YOLOv3 模型有所牺牲，但仍明显高于其他两种模型。对 比结果进一步验证了所提模型的有效性。
上传时间： 2021-10-30
上传用户：
