针对电子战条件下,通信信号易受压制干扰的问题,提出了一种基于动态学习率深度自编码器(dynamic learning rate deep Auto Encoder,DLr-DAE)的信道编码算法来提高系统抗压制干扰性能。首先对输入未编码信号进行预处理,将原始输入信号转...针对电子战条件下,通信信号易受压制干扰的问题,提出了一种基于动态学习率深度自编码器(dynamic learning rate deep Auto Encoder,DLr-DAE)的信道编码算法来提高系统抗压制干扰性能。首先对输入未编码信号进行预处理,将原始输入信号转换为单热矢量;随后使用训练数据样本集,用非监督学习方法训练深度自编码器,基于随机梯度下降法(SGD)更新网络参数,利用指数衰减函数,在迭代次数和网络损失函数值变化过程中动态微调学习率,减少网络迭代循环次数,避免收敛结果陷入局部最优点,从而获得面向电子战环境的信道编码深度学习网络。仿真结果表明,相比现有深度学习编码算法,该算法在取得同等误码率时,抗噪声压制干扰性能最大可提升0.74 d B。展开更多
The construction of the SSC-Linac has made significant progress in 2019.The first beam commissioning of SSC-Linac as an injector of the Separated Sector Cyclotron(SSC)was performed succssfully with ^(40)Ar^(7+) beam a...The construction of the SSC-Linac has made significant progress in 2019.The first beam commissioning of SSC-Linac as an injector of the Separated Sector Cyclotron(SSC)was performed succssfully with ^(40)Ar^(7+) beam and the particle energy of 5.98 MeV/u was obtained at the exit of SSC on December 17,2019.展开更多
文摘针对电子战条件下,通信信号易受压制干扰的问题,提出了一种基于动态学习率深度自编码器(dynamic learning rate deep Auto Encoder,DLr-DAE)的信道编码算法来提高系统抗压制干扰性能。首先对输入未编码信号进行预处理,将原始输入信号转换为单热矢量;随后使用训练数据样本集,用非监督学习方法训练深度自编码器,基于随机梯度下降法(SGD)更新网络参数,利用指数衰减函数,在迭代次数和网络损失函数值变化过程中动态微调学习率,减少网络迭代循环次数,避免收敛结果陷入局部最优点,从而获得面向电子战环境的信道编码深度学习网络。仿真结果表明,相比现有深度学习编码算法,该算法在取得同等误码率时,抗噪声压制干扰性能最大可提升0.74 d B。
文摘The construction of the SSC-Linac has made significant progress in 2019.The first beam commissioning of SSC-Linac as an injector of the Separated Sector Cyclotron(SSC)was performed succssfully with ^(40)Ar^(7+) beam and the particle energy of 5.98 MeV/u was obtained at the exit of SSC on December 17,2019.