Research On In-Band Optical Signal-to-Noise Ratio Monitoring Techniques In Optical Fiber Communication System

With the sharp growth of data traffic demand,optical fiber communication systems are currently developing in the direction of higher speed,higher spectral efficiency,larger capacity,longer transmission distance,and dynamic reconfiguration.To ensure the reliability and stability of the operation of optical fiber communication systems,the role of optical performance monitoring(OPM)techniques is particularly important.Among various OPM parameters,the optical signal-to-noise ratio(OSNR)parameter which is directly related to the bit error rate can accurately reflect the quality of optical signals.Therefore,accurate OSNR monitoring techniques have always been indispensable in the field of optical fiber communications.However,with the narrower channel spacing of dense wavelength division multiplexing(DWDM)systems and the widespread deployment of reconfigurable optical add/drop multiplexers,the traditional out-of-band OSNR monitoring technique based on linear interpolation is no longer applicable.Thereupon,for high-speed optical fiber communication systems and dynamic optical networks,the OSNR monitoring techniques have to change from out-of-band to in-band.Besides,to flexibly monitor the quality of optical signals at any node across the fiber link,OSNR monitoring needs to be spread throughout optical networks,including receiver destination nodes and intermediate nodes.Based on the above background,this thesis has carried out a series of researches on the in-band OSNR monitoring techniques,and the main innovations are as follows:(1)An in-band OSNR monitoring technique based on the frequency domain linear frequency modulation(LFM)signal applied at receiver destination nodes is proposed and experimentally verified.In this technique,the LFM signal with extremely low power is firstly superposed on the spectral range of the data signal at the transmitter,and then employs the LFM signal exhibits energy aggregation property to effectively separate signal power and noise power when performing fractional Fourier transform with the optimum order at the receiver,thereby achieving in-band OSNR monitoring.Finally,a 10 GBaud QPSK coherent optical communication system is set up for experimental researches,experimental results show this technique still has a high OSNR monitoring accuracy without occupying additional spectrum resources.(2)An in-band OSNR monitoring technique based on Gaussian process regression(GPR)applied at intermediate nodes is proposed and experimentally verified.This technique is a novel machine learning method.It first sweeps the signal by adjusting the center wavelength of a wide-band tunable optical bandpass filter with a constant step,and then utilizes GPR in conjunction with optical power measurements at different wavelength components of the signal measured by a low-speed photodetector,thereby achieving in-band OSNR monitoring.Finally,a 9×10GBaud QPSK coherent optical communication system and 9×32GBaud PDM-QPSK,PDM-16 QAM,and PDM-64 QAM coherent optical communication systems are set up for experimental researches respectively,experimental results show this technique has the following advantages: higher monitoring accuracy;wider monitoring range;strong robustness to chromatic dispersion,polarization mode dispersion,nonlinear effect,cascade filtering effect,and optical amplifier type;transparency to modulation formats;compatibility with symbol rates;no prior knowledge of link configuration required;no measurement of transmitter spectrum required;low cost;fast response time;distributed monitoring.Therefore,this technique can realize OSNR monitoring at intermediate nodes which is suitable for dynamic reconfigurable high-speed DWDM optical fiber communication systems and has wide practical application potential and huge development prospects.