Research On High-speed Optical Signal Monitoring Technology Based On Linear Optical Sampling

In the context of today’s information and intelligence technology,with the large-scale commercialization of 5G communication,mixed reality,artificial intelligence and other technologies that require high bandwidth for network transmission,the data-rate of fiber optic communication system is growing exponentially.High-speed optical signal analyzer is an important basic instrument in the field of fiber optic communication,which plays an irreplaceable role in the analysis of optical signals with high bandwidth and complex modulation formats,and is one of the cornerstones of the development of optical communication industry.Compared with the traditional electrical sampling optical modulation analyzer,all-optical sampling technology cleverly shifts the sampling process of high-speed optical signals from the electrical domain to the optical domain,bypassing the high-bandwidth photoelectric conversion and high-speed electrical sampling process,avoiding the limitations of the electronic bottleneck,and is the most promising technology for monitoring high-speed optical signals.Using linear optical sampling techniques to extract more accurate signal parameters for higher speed and higher order modulation formats,and to obtain eye diagrams and constellation diagrams that accurately reflect the signal quality,are the main research goals in the field of linear optical sampling at present.In this thesis,linear optical sampling(LOS)system is the background.To address the problems that the existing datarate measurement method have a very limited measurement range,a range selectable data-rate measurement method(RSDMM)based on triple frequency sampling is proposed;To address the problems that the sampling order cannot be judged by optical sampling,a sampling order judgement method based on triple frequency sampling is proposed.To address the problem that the fast Fourier transform based pulse repetition frequency extraction(FFT-PRFE)algorithm is not accurate enough,a chirp-Z transform based pulse repetition frequency extraction(CZT-PRFE)algorithm is proposed.The main research work of the thesis is as follows:(1)To address the problem that the existing data-rate measurement method have a limited measurement range,this study proposes a range selectable data-rate measurement method(RSDMM)based on triplefrequency sampling.In the process of tri-frequency sampling,a frequency selection factor(FSF)is selected to realize the selection of the measurable data-rate range,which improves the measurable range of the data-rate at the same sampling cost,and the method is transparent to the modulation format,which can be used for data-rate measurements in all modulation formats.An experimental platform for multi-frequency linear optical sampling system is built to measure the data-rate of PDM-QPSK signals from 0.8GBaud to 40.8 GBaud.The relative error of the measured baudrate is less than 0.17%while the error vector magnitude(EVM)is below 0.38.(2)To address the problem that the sampling order cannot be judged by optical sampling technology,this study proposes a sampling order judgement method based on triple frequency sampling.Since optical sampling is a under-sampling process,the inherent characteristic of undersampling causes that the sampling order cannot be judged in the process of software synchronization,resulting in the reversal of the timing of the reconstructed eye diagram.Based on the RSDMM,this study proposes a sampling order discriminant,which is the first time to realize the judgement of sampling order in optical sampling,which ensures the correct timing information of the reconstructed eye diagram later.In the experiment,the effectiveness of the sampling order judgement method is verified by multi-frequency LOS system.(3)In order to further improve the accuracy of pulse repetition frequency extraction as well as the data-rate measurement,this study proposes a chirp-Z transform based pulse repetition frequency extraction(CZT-PRFE)algorithm.By using CZT to refine the spectrum of sampled signals,clearer spectral peaks were obtained,resulting in more accurate pulse repetition frequency.Experiments results have shown that the average relative error of pulse repetition frequency extracted by the CZTPRFE is 0.00307%.Compared with FFT-PRFE,the average relative error is reduced by more than 30 times.When using the pulse repetition frequency extracted by the CZT-PRFE instead of the pulse repetition frequency read by the mode-locked pulsed laser for data-rate measurement,the average relative error is reduced by 37.63%.