Generation of large fifth-order nonlinear phase shifts by use of lossless χ(2) cascaded nonlinear processes

It is shown that the cascaded fifth-order nonlinear phase shifts will increase with energy loss in the cascaded processes. Essentially different from the multi-photon absorption accompanied with inherent material nonlinearities, the loss of fundamental wave in a cascaded process is controllable and suppressible. By introducing difference frequencies generated from the reaction between the fundamental and its second harmonic after the cascaded processes, the fundamental wave can be free of energy loss, while the large cascaded fifth-order nonlinear phase shift is maintained.     

一种新型级联码

提出一种由Turbo与纠正多个随机错误的循环码构成的新型级联码,旨在用纠正多个随机错误的循环码纠正Turbo不能纠正的错误,进一步降低误码率.仿真结果表明:信噪比超过一定值后,此级联码的误码率明显低于传统的Turbo码.     

多孔硅湿敏特性研究

应用阳极氧化技术在单晶硅片上生长一层多孔硅（ＰＳ）薄膜，用真空镀膜方法分别在其上下两面蒸发一层适当厚度的金属铝作为电极，设计了两种电极图形，制成二端结构的敏感元件．在不同湿度环境下，测出其电容值，得到了多孔硅ＲＨ－Ｃ湿敏特性曲线．通过不同材料、不同工艺条件的试验研究发现：多孔硅湿敏元件的湿敏特性、响应时间与其孔隙率、膜厚等参数有关．从而，确定了适合于敏感应用的材料导电类型、电导率、工艺条件和元件结构     

应用级联神经网络预测供热锅炉次日小时热负荷的初步研究

通过对供热锅炉房热负荷的分析，建立了基于两个BP网络的级联神经网络(CNN)。相关性分析表明，可将时间序列负荷数据作纵横向分离，横向相关系列负荷可作为CNN前一BP子网络的输入数据，纵向相关系列负荷可作为CNN后一BP子网络的输入数据。前一BP子网络用于小时负荷的初始预测，其预测结果加入后一BP子网络的输入系列，实现对负荷的精确预测。按照此模型，建立了某一印染厂锅炉房次日小时蒸汽负荷的CNN预测模型。程序运行结果表明该模型在预测时足够准确可靠。     

跳频通信中纠错编码技术的设计

本文给出了一种适用于跳频通信信道的级联码编译码器的具体设计方法,即以GOLAY码为内码,GF(16)域上的RS(15,10,6)码为外码的级联码,并对级联码的误码性能进行了模拟,这是一种实用可行的方案.     

Review of public-key cryptosystem based on the error correcting code

The paper review the public-key cryptosystems based on the error correcting codes such as Goppa code, BCH code, RS code, rank distance code, algebraic geometric code as well as LDPC code, and made the comparative analyses of the merits and drawbacks of them. The cryptosystem based on Goppa code has high security, but can be achieved poor. The cryptosystems based on other error correcting codes have higher performance than Goppa code. But there are still some disadvantages to solve. At last, the paper produce an assumption of the Niederreiter cascade combination cryptosystem based on double public-keys under complex circumstances, which has higher performance and security than the traditional cryptosystems.     

级联光栅结合Sagnac环的可调谐光纤激光器

提出并验证了一种单-双波长可调谐掺铒光纤激光器。利用级联光纤布拉格光栅(Cascaded Fiber Bragg gratings,Cascaded FBGs)结合Sagnac环结构所产生的复合滤波效应,实现较高精细度滤波,并通过调节环内偏振控制器(Polarization Controller,PC),引入双折射效应,得到波长可调谐的光纤激光器。基于耦合模理论并使用传输矩阵法对该结构的传输特性进行了分析,在此基础上搭建实验系统,验证了理论分析的正确性。实验结果表明:通过调节PC,激光器输出激光的波长范围约为1 555.644～1 556.112 nm,双波长间隔的可调范围约为0.108～0.452 nm,单-双波长的边模抑制比(SMSR)均高于40 dB;在稳定性测试中,输出单-双波长激光的波长最大漂移量小于0.008 nm。该方法具有结构简单、调谐方便、易于实现且精细度较高的优点,可应用于密集波分复用及全光通信系统等领域。     

直接出射白光的新型发光二极管

介绍了实现出射白光发光二极管的几种方法,包括波长转化产生白光,多有源区级联合成白光以及单个有源区直接出射白光.并从芯片结构、材料选取、白光形成机理以及器件性能等四个方面对这几种方法进行了分析比较.最后指出了实现直接出射白光的发光二极管存在的问题及今后的研究重点.     

基于级联卷积神经网络的前列腺磁共振图像分类

针对深度学习训练成本高,以及基于磁共振图像的前列腺癌临床诊断需要大量医学常识且极为耗时的问题,本文提出了一种基于级联卷积神经网络（Convolutional Neural Network,CNN）和磁共振图像的前列腺癌（Prostate Cancer,PCa）自动分类诊断方法,该网络以Faster-RCNN作为前网络,对前列腺区域进行提取分割,用于排除前列腺附近组织器官的干扰;以基于ResNet改进的网络结构CNN40_bottleneck作为后网络,用于对前列腺区域病变进行分类.后网络由瓶颈结构串联组成,其中使用批量标准化（Batch Normalization,BN）、全局平均池化（Global Average Pooling,GAP）进行优化.实验结果证明,本文方法对前列腺癌诊断结果较好,而且缩减了训练时间和参数量,有效降低了训练成本.     

Design study for the cascaded HGHG experiment based on the SDUV-FEL

Cascading stages of high-gain harmonic generation(HGHG) free electron laser(FEL) is a promising way to produce fully coherent X-ray radiation.As a test facility for modern FEL R&D,the Shanghai deep ultraviolet FEL(SDUV-FEL) is now under upgrading for the cascading two stages of HGHG experiment.Since the energy of the electron beam is as low as about 185 MeV after upgrade,the total harmonic number of this two stages HGHG is only 2×2,and the wavelength of the final radiation is 196.5 nm which is the 4th harmonic of the 786 nm seed laser.With help of three-dimensional simulation codes,design studies on the FEL physics for the cascaded HGHG experiment are present based on the parameters of the upgraded SDUV-FEL facility.It is found from the simulation results that the part of the electron beam which has been used in the first stage can still generate powerful radiation in the radiator of the second stage,and this radiation will be difficult to be separated from the radiation generated by the fresh part of the electron beam.To overcome this problem,a novel method based on the energy spectrum of the electron beam is proposed in this paper to demonstrate the "fresh bunch" technique.