基于IPv6的无线Mesh网络切换技术研究

如何提高无线Mesh网络的切换效率,同时保障其服务质量一直是一个很具有挑战性的问题,它直接关系到无线Mesh网络是否能够被广泛应用。基于此,提出将IPv6技术引入到无线Mesh网络中,使得网络中的每个设备都有一个唯一的IPv6地址和响应的MAC地址,实现移动终端的快速切换,提高了无线Mesh网络的切换效率。     

热管冷却技术在Yb:YAG激光放大器中的应用

为克服水流冷却在固体激光热管理中结构复杂、稳定性差、能耗大等缺点,设计了热管冷却的Yb:YAG激光放大器.分别通过理论和实验验证了热管冷却技术的可行性.首先,建立了三维、瞬态的有限元模型,模拟结果表明,对于6 kW、1.2 ms的泵浦能量,1 Hz重频下的热管冷却和水流冷却散热能力接近;其次,实验测量了Yb:YAG放大器的输出能量,两种冷却方式在1Hz下的输出能量基本相同,热管冷却的能量稳定性更好.通过优化热管参数,放大器的工作频率、输出功率得到进一步提高.     

高功率全固态激光器抽运耦合技术进展

综述了目前获得高功率激光输出的板条、薄片激光器以及运行于热容模式固体激光器抽运耦合方式的进展,分析了这几种固体激光器的抽运耦合技术路径,比较了各类抽运耦合方式在克服热效应方面的优缺点,展望了固体激光器抽运耦合技术的发展趋势。     

基于序列图像的阴影检测研究

针对由于入射光的照射,目标阴影与背景有明显的区别,而被当做目标的问题进行了研究。分析了单幅图像的阴影检测方法,提出了序列图像阴影检测方法,建立了序列图像高度测量模型;采用相邻序列图像对应点匹配的约束条件,降低了匹配的复杂度;研究了序列图像的特征点匹配,获得平面目标的高度信息表示;基于阴影的二维表面属性,设计了利用空间三维信息中的高度信息检测阴影的算法,可以对任意形状的阴影进行检测。通过试验表明,该方法简单,适用面广,可以有效地检测阴影。     

一种改进的混沌序列去噪方法

针对混沌时间序列的噪声平滑,分析了其信噪比与关联维数的关系,并结合局部投影方法的邻域选取问题,提出了一种改进的去噪方法。该方法对于含有噪声的混沌信号选定一个邻域半径值,计算去噪后序列的关联维数,确定该选定的邻域半径值是否准确并进行调整,最终找到合适值并进行多次局部投影去噪,从而得到去噪后的纯净序列。用改进后的方法对含噪声的Lorenz序列和对股市序列进行对比预测仿真,仿真结果表明该方法能够有效地选取邻域半径,进而改善局部投影方法的去噪效果,取得更好的预测效果。     

Fabrication and characterization of AlGaN/GaN HEMTs with high power gain and efficiency at 8 GHz

State-of-the-art AlGaN/GaN high electron mobility structures were grown on semi-insulating 4H-SiC substrates by MOCVD and X-band microwave power high electron mobility transistors were fabricated and characterized.Hall mobility of 2291.1 cm2/(V·s) and two-dimensional electron gas density of 9.954 × 1012 cm-2 were achieved at 300 K.The HEMT devices with a 0.45-μm gate length exhibited maximum drain current density as high as 1039.6 mA/mm and peak extrinsic transconduct-ance of 229.7 mS/mm.The fT of 30.89 GHz and fmax of 38.71 GHz were measured on the device.Load-pull measurements were performed and analyzed under (-3.5,28) V,(-3.5,34) V and (-3.5,40) V gate/drain direct current bias in class-AB,respectively.The uncooled device showed high linear power gain of 17.04 dB and high power-added efficiency of 50.56％ at 8 GHz when drain biased at (-3.5,28) V.In addition,when drain biased at (-3.5,40) V,the device exhibited a saturation output power dens-ity up to 6.21 W/mm at 8 GHz,with a power gain of 11.94 dB and a power-added efficiency of 39.56％.Furthermore,the low fmax/fT ratio and the variation of the power sweep of the device at 8 GHz with drain bias voltage were analyzed.     

基于多层感知器的外辐射源雷达多帧联合检测

针对复杂杂波背景下“低慢小”目标检测问题,发展了一种基于多层感知器(multi-layer perceptron,MLP)的检测器.该检测器联合多场距离多普勒(range-Doppler,RD)谱作为观测空间,在每一场RD谱上取沿距离维滑窗的n个(通常取n=3)单元构造观测向量.为评估该检测器实际性能,开展了无人机探测...     

Selected-Mapping Peak-to-Average Power Reduction Method with Orthogonal Pilot Sequences in Underwater Acoustic OFDM System Without Side Information

A novel Selected-mapping (SLM) Peak-to-average power ratio (PAPR) reduction scheme requires no Side information (SI) in Underwater acoustic (UWA) OFDM system is proposed. In the proposed scheme, every distinct phase sequence is represented by a unique Orthogonal comb pilot sequence (OPS), and the orthogonal properties of the OPSs are used to distinguish the index of phase sequences at the receiver. Therefore, the proposed scheme does not need to reserve bits for transmitting SI, so that the data rate can be raised. Simulation results show that the PAPR reduction performance has almost 0.5dB gains comparing to the Conventional SLM (C-SLM) scheme and the Bit error ratio (BER) performance is approximately the same as the SLM scheme with perfect SI. Field experimental results also demonstrate that the proposed scheme can differentiate phase sequences, therefore significantly enhance the quality of the UWA OFDM communication system.     

一种RNN-T与BERT相结合的端到端语音识别模型

端到端语音识别模型由于结构简单且容易训练,已成为目前最流行的语音识别模型.然而端到端语音识别模型通常需要大量的语音-文本对进行训练,才能取得较好的识别性能.而在实际应用中收集大量配对数据既费力又昂贵,因此其无法在实际应用中被广泛使用.本文提出一种将RNN-T(Recurrent Neural Network Trans...     

基于YOLOv5s的人脸是否佩戴口罩检测

为了抗击新冠肺炎,加快复产复工的速度,本文提出了一种判断人脸是否佩戴口罩的方法.该方法可以通过摄像头捕获人脸,对人脸是否佩戴口罩进行检测,及时提醒人们带好口罩.使用预处理好的9800张图片作为数据集,并对YOLOv5s算法进行离线训练,生成最终的模型;利用该模型对摄像头捕捉的画面进行分析,检测人脸是否佩戴有口罩.该算法在测试集上的精确率(precision)、召回率(recall)和平均精度(mAP)分别为78.1％、87％和53.5％,高于YOLOv3和YOLOv4检测算法的检测结果.