Deep Reservoir Computing Based On Cascaded Coupled Semiconductor Lasers And Its Applications

The delay-based photon reservoir computing（RC）is a novel computing method with simple structure,high computational efficiency and strong learning ability.Using this method,the trajectories and synchronizations of optical chaos can be well predicted,and the chaotic signals can be accurately distincted.The delay-based RC is expected to solve the key problems of chaotic synchronization in secure communication,such as low quality and instability,to achieve high-quality optical chaos synchronization.Although the delay-based photon RC with a single hidden layer structure has good simplicity in learning of model parameters,but it can not effectively extract and learn the high-dimensional feature information of time-series data with multi-time scale and complex dynamic.Combining the advantages of the current mainstream deep learning neural network in learning of multi-level feature information,the project explores the possibility of combining the simple and efficient features of the delay-based photon RC with the deep learning neural network model.For this purpose,this project generalizes the cascaded optically pumped spin-VCSELs to apply in the construction of a delay-based deep photonic RC system with multi-layer,and explores the application of the delay-based deep photonic RC system in multi-channel optical chaotic secure communications.For these research contents,the major works of this project are presented as follows:（1）Two cascade coupling modes（unidirectional coupling and bidirectional coupling）are used to construct a four-layer deep RC system based on the cascade coupled optically-pumped spin-VCSEL.Under these two coupling modes,the chaotic x-PC and y-PC emitted by the driving optically pumped spin-VCSEL（D-Spin-VCSEL）,as two learning targets,are predicted by utilizing the four-layer reservoirs.In different parameter spaces,it is further explored that the outputs of the double sub-reservoirs in each layer are respectivel synchronized with the chaotic x-PC and y-PC emitted by the D-Spin-VCSEL.The memory capacities（MCs）for the double sub-reservoirs in each layer are even further investigated.The results show that under two coupling modes,the predictions of the double sub-reservoirs with higher-layer for these two targets have smaller errors,denoting that the higher-layer double sub-reservoirs possess better predictive learning ability.Under the same system parameters,the outputs of the higher-layer dual parallel reservoirs are better synchronized with two chaotic PCs emitted by the D-Spin-VCSEL,respectively.The larger MCs can also be obtained by the higher-layer double reservoirs.In particular,compared with the four-layer RC system under unidirectional coupling,the four-layer RC system under bidirectional coupling shows better predictive ability.The chaotic synchronizations predicted by each layer double sub-reservoirs under bidirectional coupling can be obtained higher qualities than those under unidirectional coupling.By the optimization of the system parameters,the outputs of the fourth-layer double sub-reservoirs are almost completely synchronized with the chaotic x-PC and y-PC emitted by the D-Spin-VCSEL,respectively.See chapter 3 for this part.（2）In this work,we utilize four parallel reservoirs to model the chaotic dynamics of the output four polarization components（PCs）from a driving quantum-dot（QD）spin-VCSEL.Here,the four parallel reservoirs are implemented using the four PCs of a reservoir QD spin-VCSEL.High-quality chaos synchronizations of four pairs of PCs can be realized by using the four parallel reservoirs based on a reservoir QD spin-VCSEL.Under these high-quality synchronizations,we successfully implement four-channel secure communications with 4×100Gb/s 16QAM messages under guaranteeing.Then,the performances of the bit error ratios for four decoding messages under different parameters are further discussed.The results show that all bit error ratios via different parameters are less than 7×10^-3,denoting that high-quality data-transmissions can be potentially obtained in the system.Moreover,it is proved that our proposed multi-channel optical chaotic communication scheme has the same level of security as the traditional schemes.Our findings show that the delay-based optical reservoir computing based on a QD spin-VCSEL provides an effective method for realization of multi-channel optical secure communication.See chapter 4 for this part.