Font Size: a A A

Studies On Key Technologies And Applications Of Signal Processing In Elastic Optical Networks

Posted on:2017-07-13Degree:DoctorType:Dissertation
Country:ChinaCandidate:D S WangFull Text:PDF
GTID:1318330518996789Subject:Electronic Science and Technology
Abstract/Summary:PDF Full Text Request
With the development of the information society, the amount of global data traffic is exploding, and the network bandwidth is significantly improved, which puts forward higher requirements for the traditional optical network bearing capacity. In the existing wavelength switched optical networks, the channel interval is fixed and the exchange granularity is too large, which lead to the waste of spectrum resources, the lack of network flexibility, and poor adaptive ability. In presence of the above problems, elastic optical network based-on flexible raster utilizes variable bandwidth optical transceiver and optical switching mechanism to allocate spectrum resources flexibly and efficiently according to service needs,which has become one of the most protective research hotspots of optical network. Elastic network mainly adopts multi-dimensional higher-order optical modulation, requires the corresponding signal processing technologies with the characteristics of high speed, energy saving,flexibility and intelligence. But, in the present elastic optical network, the traditional signal processing technologies still faces such problems as electronic speed bottleneck, the excessive energy consumption, single function, low efficiency, high complexity, no transparence of signal formats and bandwidth etc., which make it cannot meet the needs of elastic optical network. Therefore, the exploration and research of signal processing technology in the elastic optical network is of great significance,and is also a difficulty and a hotspot of current research in optical networks.This dissertation focuses on the difficulties and requirements of optical signal processing technologies in elastic optical network, and explores the optical signal processing technologies in elastic optical switching network and digital signal processing technologies in elastic optical transmission network. Several technical schemes are proposed. The main innovations of this paper are as follows:Firstly, three kinds of elasticity reconstruction of optical add-drop multiplexing (ROAMD) structure with optical signal processing function based on domestic tunable bandwidth wavelength selective optical switch(TB-WSS) are proposed. The optical signal processing module uses four-wave mixing (FWM) effect in semiconductor optical amplifier (SOA),which can realize a variety of optical signal processing functions.Experiments show that based on the module wavelength conversion of differential phase shift keying (DPSK), quadrature phase shift keying(QPSK) and 16 quadrature amplitude modulation (16QAM) can be realized, and can realize the conversion range up to 1THz with 12.5GHz conversion step.Secondly, multicast scheme based on optical spilitter has many problems such as high power cost and frequent wavelength conflict. In this dissertation, a wavelength division multiplexing (WDM) multicast scheme based on multi-pumps FWM effect in SOA is proposed. Experimental results show that the scheme supports of the 1 to 6 and 1 to 10 WDM multicast for 12.5Gbaud QPSK signal, and also supports 1 to 7 WDM(FWDM) multicast specifically for the elastic optical network. It can realize 1 to 3 and 1 to 6 WDM multicast for dual-channel QPSK signals.The experimental results show that all the generation multicast signals have optical signal-to-noise ratio (OSNR) between 13 to 25 dB in above multicast schemes, and most of the signals have good performances to satisfy the actual communication demands.Thirdly, focusing on the problems such as single function and low efficiency of traditional optical signal processing, a scheme about format conversion and wavelength conversion of dual DQPSK (2×12.5Gbaud) to DPSK based on FWM effect in the SOA is proposed. The experimental results show that the scheme can effectively support the format conversion and wavelength conversion of dual-channel signals. Compared with the original signal, DSPK signal after conversion has the strong ability to resist damage with -0.5dB power cost. Two other new generation DQPSK signals have the power cost of less than 1.0 dB.Fourthly, a scheme of three-channel DPSK signal (3×12.5Gbaud)exclusive or (XOR) operation based on FWM effect in quantum dots SOA(QD - SOA) is proposed. The experimental results show that the scheme can realize parallel three signals XOR operation and realize their WDM multicast function at the same time. The bit error rate (BER) of all signals is below 10-9.Fifthly, aiming at nonlinear phase noise (NLPN) problem in elastic optical transmission, bringing machine learning algorithms to digital signal processing of elastic optical transmission, an algorithms of low complexity nonlinear phase noise suppression based on support vector machine(SVM)is proposed. Simulation results show the algorithm can suppress the NLPN of BPSK, QPSK, 8PSK and 16QAM modulation in elastic optical transmission, which has more obvious impacts for high order signal. It can increase system tolerance by 2.8 dB of 100 Gbps 16QAM signal.Sixthly, aiming at the problems of the difficulty of dynamic damage compensation in the traditional digital signal processing algorithm and the bad adaptive ability in the elastic optical transmission, a multiple physical damage compensation algorithm based on back propagation artificial neural network (BP-ANN) is proposed. Simulation results show the algorithm can effective compensate gaussian white noise, light source phase noise, I/Q branch imbalance, nonlinear phase noise, and other physical damages, can to broaden the tolerance limit of light source line to 170 kHz in 100 Gbps 16 QAM system, increase nonlinear tolerance by 2.7 dB and extend the maximum transmission distance to 240 km.Seventhly, focusing on the problems of too long training time in the traditional digital signal processing algorithm and dependency on compensation link types, an algorithm with free training and elastic optical transmission hybrid compensation link based on k nearest neighbor (KNN)is proposed. The simulation analyzes nonlinear compensation effects in the dispersion-shifted link, the dispersion management link and dispersion unmanagement link. Simulation results show the algorithm has strong link damage adaptability and is suitable for mixing link of elastic optical transmission. In three kinds of typical links, the system nonlinear tolerance can be increased by 1.7 dB, 1.0 dB and 1.7 dB.
Keywords/Search Tags:Elastic optical network, optical signal processing, digital signal processing, machine learning, semiconductor optical amplifier, four wave mixing, high order modulation, nonlinear compensation, artificial neural network, support vector machine
PDF Full Text Request
Related items