Abstract
Transmission systems that use advanced complex modulation schemes and digital coherent receivers, have been driving the growth of optical communication networks for nearly a decade. In this fact, new strategies for network diagnosis and management are developed. In this paper, we propose and experimentally demonstrate a cost-effective technique that not only exhibits optical performance monitoring functionalities but also enables simultaneous symbol rate and modulation format identification. It is implemented for both dual-polarized quadrature phase shift keying and 16-quadrature amplitude modulation at various symbol rates. Our approach is based on artificial neural networks (ANNs) in conjunction with asynchronous amplitude histogram (AAH) evaluation. Our ANN-based pattern recognition algorithm, with all of the features of AAHs, facilitates both symbol rate and modulation format classification at high data rates. Experimental investigation demonstrates accurate classification of advanced modulation formats, in the presence of various link impairments. The latter include chromatic dispersion (CD) varying from 200 to 1600 ps/nm, differential group delay between 10 and 70 ps, phase noise varying from 10 kHz to 10 MHz and amplified spontaneous emission noise. The results of the aforementioned method demonstrate multiimpairment monitoring over wide monitoring ranges. At high symbol rates, after a precise extraction of monitoring parameters (i.e., F_{\text{CD}} and F_{\text{OSNR}}), our approach achieved a minimum dispersion parameter close to 0.2 dB and a maximum performance parameter of 6 dB. Our technique is robust and transparent for both the symbol rates and the modulation formats.