Study On State Analysis And Performance Monitoring Of Optical Links Based On Chaos Theory

With the rapid progress of the optical networking technology as well as the explosive growth of various service demands, the optical network has been continually evolving toward a more complicated system with higher speed, capacity, transparency and intelligence. Traditional optical performance monitoring (OPM) technique based on spectral analysis will no longer apply to the new optical network. Thus, the OPM technique with superior performance is indispensible to ensure the service quality. New requirements for OPM drive scholars to explore new schemes. Particularly, chaos theory is thought as one of these alternatives due to the complicated nonlinear dynamics of semiconductor lasers and the optical fiber. In this thesis, the OPM scheme based on chaos theory and time series analysis has been studied. Related theoretical analysis and experiments to monitor the performance of optical links have been carried out. The main researches accomplished are as follows.(1) Based on the chaotic time series analysis, the extraction of common chaos characterization parameters has been realized, including the attractors, the maximum Lyapunov exponent, the recurrence plots and the recurrence quantity analysis. The simulated program has been validated through analyzing the chaos states of time series from some typical systems.(2) According to the correlation between chaos theory and nonlinear dynamics, factors related to the degree of physical impairments in optical links have been analyzed. The influences of launched optical power, transmission distance and dispersion compensated or not have been discussed for single-wavelength optical links, while the influences of launched optical power, number of multiplexing wavelengths, transmission distance and dispersion compensated or not have been discussed for WDM optical link.(3) OPM simulation systems apply to both single-wavelength optical links and WDM optical links have been realized based on the chaos theory and the pattern recognition. For every set of time series, the maximum Lyapunov exponent, together with the cross-correlation maximums of the time series and the phase-only filters, are considered as the input of an artificial neural network. The output corresponds to the current chaotic region of the optical signal, that is, the degree of physical impairments. The GUI of the OPM simulation system has been designed based on Matlab. (4) Experimental systems of high-speed and long distance single-wavelength optical links and WDM optical links have been set up. With the increasing launched optical power on the transmitting terminal, the influence of changing physical impairments on the chaos state of the transmission signal has been tested. The experimental results are quite consistent with the simulation results.