Iris Flower Recognition Using Four-Channels Ring Photonic Reservoir Computing Based on Unidirectional Coupled VCSELs
Iris Flower Recognition Using Four-Channels Ring Photonic Reservoir Computing Based on Unidirectional Coupled VCSELs
复制标题
DOI：
10.1109/jqe.2023.3292064
复制
发表时间：
2023-10
影响因子：
2.5
通讯作者：
Qian Yu;S. Xiang;Xingxing Guo;Ya Hui Zhang;Biling Gu;Yanan Han;Z. Song;Yue Hao
中科院分区：
文献类型：
--
作者：
Qian Yu;S. Xiang;Xingxing Guo;Ya Hui Zhang;Biling Gu;Yanan Han;Z. Song;Yue Hao
提出了一种基于单向耦合垂直腔面发射激光器(VCSEL)的环形四通道水库计算方法(RC)，并进行了数值验证。在这里，四路环路首次由四个VCSEL实现。将该RC系统用于解决虹膜花卉分类问题，最高分类正确率达到100%。此外，还考虑了偏置电流、外注入强度、频率失谐、反馈强度、耦合强度以及虚拟节点数对基于VCSELs的环形四通道RC识别性能的影响。提出的基于VCSEL的环形四通道RC系统对于提高RC系统解决更复杂问题的能力具有重要意义。
In this paper, a ring Four-channels reservoir computing (RC) based on unidirectional coupled vertical cavity surface emitting lasers (VCSELs) is proposed and demonstrated numerically. Here, the ring four channels are realized by four VCSELs for the first time. Such proposed RC system is employed to solve the Iris flower classification task, and the highest classification accuracy achieved at 100%. Besides, the effects of bias currents, external injection strength, frequency detuning, feedback strength, couple strength, as well as the number of virtual nodes on the recognition performance of the ring Four-channels RC based on VCSELs are also considered. The ring four-channel RC system based on VCSEL proposed may be of great significant for improving the RC system’s ability to solve more complex problems.