Utilização de regressão logística para mitigação de não linearidades em sistemas de comunicações ópticas coerentes digitais

• Autor(a) principal: Paula Júnior, Rômulo Aparecido de
• Data de Publicação: 2019
• Tipo de documento: Trabalho de conclusão de curso
• Idioma: por
• Título da fonte: Repositório Institucional da UNESP
• Texto Completo: http://hdl.handle.net/11449/217446
• topic: Comunicações ópticas; Inteligência artificial; Telecomunicações

In this work we studied a method for compensating nonlinearities based on artificial intelligence for digital coherent long-reach passive optical networks. The transmission of an optical signal modulated utilizing 16-QAM format through a link, whose model considers both linear and nonlinear effects, was computationally simulated. In such systems, noise and attenuation limit the range. Unfortunately, increasing transmission power is not a feasible solution, because it causes the excitation of nonlinearities as a result the interation between the signal and the medium. The results show that, due to the nonlinear phase rotation caused by the Kerr effect, the conventional detection by maximum likelihood presents low performance. The implemented computational technique, denominated logistic regression, is able to draw complex decision boundaries that fit the received distorted constellations. By applying the aforementioned algorithm, a decrease of up to 38% of the optimal value of the symbol error rate in relation with the maximum likelihood detection can be achieved. These results reveal that, by the compensation of such effects, it’s possible to improve the bandwidth-length product.