中心博士生邹秀婷的工作——“Photonic analog to-digital converter powered by a generalized and robust convolutional recurrent autoencoder”（具有泛化能力和鲁棒性的卷积循环自动编码器赋能的光子模数转换系统）的相关成果近期被Optics Express期刊接收，该工作得到了国家重点研发计划（2019YFB2203700）、国家自然科学基金（61822508）的部分资助。该工作通过搭建一个卷积循环自动编码器（CRAE）来解决光子模数转换系统（PADC）中的时间失配问题。实验验证了CRAE对PADC系统时间失配补偿的有效性。另外实验结果表明，不同于大多数神经网络，CRAE具有很好的泛化能力和鲁棒性，它对未训练过的系统状态、未训练过的失配量依然具有非常好的补偿效果。例如，当失配量高达137 ps 时， CRAE可以将信号的无杂散动态范围（SFDR）从-3 dBc 提升至31.6 dBc。我们提出的方法有望解决各种通道交织系统中的时间失配问题，实现高性能的下一代信息系统。另外，我们提出的卷积神经网络（CNN）、循环神经网络（RNN）、自动编码器结合的混合神经网络架构CRAE，也为提高神经网络泛化能力提供了一种思路。
摘要: We propose a convolutional recurrent autoencoder (CRAE) to compensate for time mismatches in a photonic analog to-digital converter (PADC). In contrast of other neural networks, the proposed CRAE is generalized to untrained mismatches and untrained categories of signals as well as robust to system states. We train the CRAE using mismatched linear frequency modulated (LFM) signals with mismatches of 35 ps and 57 ps under one system state, it can effectively compensate for mismatches of both LFM and Costas frequency modulated signals with untrained mismatches under different system states. When the mismatches increase from 35 ps to 137 ps, the spur-free dynamic range (SFDR) of the unpowered PADC reduces from 10.2 dBc to -3.0 dBc. However, for the CRAE-powered PADC, its SFDR is always over around 31.6 dBc.
论文链接