Research Of Chaotic Dynamic Characteristics And Reservoir Computing Based On A Novel Semiconductor Nanolaser

With the rapid development of information technology,there is an urgent demands for secure and reliable communication technology and efficient information processing rate in daily life.Due to booming development of optical devices and photonics,secure optical communications system and optical information processing technology have opened up new opportunities.Because of the advantages of small size,high performance,high speed and easy integration,the novel semiconductor nanolasers have potential applications in photonic integrated circuits and system-on-a-chip technologies.Therefore,this thesis focuses on the time delay signature of the chaotic semiconductor nanolasers and the reservoir computing system combined with the neural network.The research is mainly carried out from the following three aspects:1.A chaos system based on novel nanolaser with dual-path chaotic optical injection is presented.The mathematical model of this system is established.The time delay signatures of output chaotic signals from the master nanolasers and the slave nanolaser are compared by using autocorrelation function.It is found that the time delay signature for the slave nanolaser can be better concealed than that for the master nanolasers.The time delay signature concealment is evaluated by the peak size of the autocorrelation function.The effects of feedback rate,injection rate,frequency detuning,Purcell cavityenhanced spontaneous emission factor and spontaneous emission coupling factor on the time delay signature concealment are quantified and analyzed.Meanwhile,the maximum range of controllable parameters leading to suppressed time delay signature is given.2.The time delay signature of chaotic signals based on the novel semiconductor nanolasers system with one-way ring network is investigated.The rate equation of this scheme is presented.The effects of system parameters and internal parameters on the time delay signature concealment are numerically examined by extracting the peak size of the autocorrelation function.Then the optimal ranges of the parameters are determined.The results show that this scheme can successfully conceal the time delay signatures of chaotic signals from all the semiconductor nanolasers in the system.Furthermore,the time delay signature concealment performance can be improved by introducing the dualpath chaotic optical injection into a novel nanolaser.3.In this thesis,novel application based on semiconductor nanolaser is further explored.The time series prediction performance of the reservoir computing system based on the nanolaser with phase conjugation feedback is studied.The mathematical models of semiconductor nanolaser with single time-delay phase conjugate feedback and double time-delay phase conjugate feedback are established by modifying the rate equations through introducing current modulation,respectively.The influences of controlled parameters on the prediction performance of two reservoir computing systems are quantitatively analyzed by normalized mean square error,such as feedback rate,feedback time-delay,Purcell cavity-enhanced spontaneous emission factor,spontaneous emission coupling factor.From the comparison,the prediction performance of the reservoir computing system with double time-delay phase conjugate feedback is better than that with single time-delay phase conjugate feedback because of its nonlinear dynamics enhancement.The range of related parameters are identified when the good prediction performance is realized by the system.