Enhancement the Nonlinear Optical Impairment of the Fiber Transmission System Based on Deep Neural Network

الموضوعات

علوم تطبيقية وتكنولوجية

الملخص

The global request for high data traffic has grown dramatically over the past few years and the future demand will outpace the capacity of the current backbone of the Internet, where a fiber-optic systems are the central infrastructure. So, performance improvement is essential for these systems and makes them more flexible and scalable to meet this growing demand require better models of fiber optic channels, complicated optimization methods, and consideration of totally new approaches to transmission, which leads to increase cost and complexity. Therefore, Independent and efficient algorithms can be considered a good solution at present to help explore possible solutions for increasing data transfer rates and resource allocation.The traditional algorithms relied on the Split-Step Fourier method (SSFM) and digital backpropagation (DBP) methods to find the approximate solution to the Non-Linear Schrödinger equation (NLSE), to process the most intrinsic challenge facing such systems which is the nonlinear impairments, but its implementation requires too many resources for signal processing as well as an accurate and high-level understanding of these systems. While Artificial Intelligence (AI) algorithms promise to provide the best solutions to the described future challenges by being utilized to determine and resolve the deficiencies in very complicated systems and complex environments without required to develop a hardware (physical) model.This thesis applied machine learning based on deep neural network (DNN) as a supervised learning method in novel contexts for coherent optical communication systems to increase their capabilities. It is proposed to build an Nonlinear Compensation (NLC) technique to compensate for the non-linear impairments that occur in optical communication systems. The system is designed and fully simulated in real-time by using MATLAB App. (V.2020a) IIIwith Optisystem App. (Ver.17.0.0) and necessary testing/optimization has been performed in three directions:Firstly, on single-polarization, coherent optical systems with 4-16 QAM modulation shapes with and without multi subcarrier OFDM, transmitted over single-carrier Standard Single-Mode Fiber (SSMF) with 60Gbps for 3000Km link distance. The results found without using OFDM of the proposed DNNs-NLC model outperforms in terms of Q-factor, BER, higher transmission distance with Q-factor equal to (~13.4 dB) at testing for 4QAM coherent optical BtB 60 Gb/s @ (100 x 30span) Km measurements.With OFDM, the proposed DNNs-NLC model outperforms in terms of Q-factor, higher transmission distance compared with ANN, LE, and IVSTF-NLE by 0, 3, and 1 dB, respectivelySecondly , on single-polarization, coherent optical systems with 16QAM modulation shapes with multi subcarrier OFDM, multi-carrier wavelength division multiplexing (8 channel WDM) with 40 GHz frequency spacing, transmitted over 3000Km SSMF link distance with totally 480Gbps bitrate. The DNNs-NLC proposed model ability was evaluated based on BER, Q-factor and correct mapper decoder. The final result obtained of the Q-factor equal to (~9 dB) at testing and outperforms compared with ANN and VSTF-NLE by <1, and 2 dB respectively Finally, on dual-polarization, coherent optical systems with 16QAM modulation shapes, multi-carrier wavelength division multiplexing (8 channel WDM) with 37.5 GHz (3nm) frequency spacing, transmitted over 3000Km SSMF link distance with totally 960Gbps bitrate.The achieved results demonstrated the efficiency of the proposed design and the significant improvement in performance, mainly in data rate and correlation range compared with the similar works.

روابط وملفات

التعريف والنوع

رقم الوثيقة

0440b88a-c656-46d4-91f2-579e1b91e1bb

رقم العقد

0

نوع الوسائط

Crawler

نوع المحتوى

الرسائل العلمية

صيغة المصدر

رسائل دكتوراة

نوع الملف

pdf text

أسماء الملفات

995415_1.pdf

بيانات النشر

ترجمة العنوان

ﺗﺤﺴﯿﻦ اﻟﻀﻌﻒ اﻟﺒﺼﺮي ﻏﯿﺮ اﻟﺨﻄﻲ ﻟﻨﻈﺎم اﻟﻨﻘﻞ اﻟﻀﻮﺋﻲ ﺑﺎﻻﻋﺘﻤﺎد ﻋﻠﻰ اﻟﺸﺒﻜﺔ اﻟﻌﺼﺒﯿﺔ اﻟﻌﻤﯿقة

ألقاب المؤلفين

[{"name_ar":"اﺑراهيم ﻋﺒﺪ ﷲ ﻣﺮداس","title_ar":"اشراف","title_en":"Supervision"},{"name_ar":"علاء حمود عبد جراح","title_ar":"اعداد","title_en":"Preparation"}]

اللغة

English

المصدر والدورية

اسم المصدر

Enhancement the Nonlinear Optical Impairment of the Fiber Transmission System Based on Deep Neural Network

المحتوى والصفحات

عدد الصفحات

0

ترجمة الملخص

ﻧﻤﺎ اﻟﻄﻠﺐ اﻟﻌﺎﻟﻤﻲ ﻋﻠﻰ ﺣﺮﻛﺔ اﻟﺒﯿﺎﻧﺎت اﻟﻀﺨﻤﺔ ﺑﺸﻜﻞ ﻛﺒﯿﺮ ﺧﻼل اﻟﺴﻨﻮات اﻟﻤﺎﺿﯿﺔ وﺳﯿﺘﺠﺎوز اﻟﻄﻠﺐ ﺣﯿﺚ ﺗﺸﻜﻞ أﻧﻈﻤﺔ اﻷﻟﯿﺎف اﻟﻀﻮﺋﯿﺔ اﻟﺒﻨﯿﺔ اﻟﺘﺤﺘﯿﺔ ﻟﻺﻧﺘﺮﻧﺖ، اﻟﻤﺴﺘﻘﺒﻠﻲ ﻗﺪرة اﻟﻌﻤﻮد اﻟﻔﻘﺮي اﻟﺤﺎﻟﻲ اﻟﻤﺮﻛﺰﯾﺔ ﻟﮫ. ﻟﺬﻟﻚ، ﯾﻌﺪ ﺗﺤﺴﯿﻦ اﻷداء أﻣﺮا ﺿﺮورﯾﺎ ﻟﮭﺬه اﻷﻧﻈﻤﺔ وﺟﻌﻠﮭﺎ أﻛﺜﺮ ﻣﺮوﻧﺔ وﻗﺎﺑﻠﺔ ﻟﻠﺘﻄﻮﯾﺮ ﻟﺘﻠﺒﯿﺔ ھﺬا اﻟﻄﻠﺐ اﻟﻤﺘﺰاﯾﺪ. ﺗﺘﻄﻠﺐ ھﺬه اﻟﻘﯿﻮد ﻧﻤﺎذج أﻓﻀﻞ ﻟﻘﻨﻮات اﻷﻟﯿﺎف اﻟﻀﻮﺋﯿﺔ وأﻧﻈﻤﺔ اﻹرﺳﺎل ﻣﻊ أﻓﻀﻞ اﻟﻄﺮق ﻟﻠﺘﺤﺴﯿﻦ واﻟﻤﻌﺎﻟﺠﺔ، ﻣﻤﺎ ﯾﺆدي ھﺬا إﻟﻰ زﯾﺎدة اﻟﺘﻜﻠﻔﺔ واﻟﺘﻌﻘﯿﺪ. ﻟﺬﻟﻚ، ﯾﻤﻜﻦ اﻋﺘﺒﺎر اﻟﺨﻮارزﻣﯿﺎت اﻟﻤﺴﺘﻘﻠﺔ واﻟﻔﻌﺎﻟﺔ ﺣﻼً ﺟﯿﺪًا ﻓﻲ اﻟﻮﻗﺖ اﻟﺤﺎﻟﻲ ﻟﻠﻤﺴﺎﻋﺪة ﻓﻲ ﺗﺤﺴﯿﻦ ﻛﻔﺎءة ھﺬه اﻷﻧﻈﻤﺔ. اﻋﺘﻤﺪت اﻟﺨﻮارزﻣﯿﺎت اﻟﺘﻘﻠﯿﺪﯾﺔ ﻋﻠﻰ طﺮﯾﻘﺔ (SSFM) Split-Fourier ﺗﻘﺴﯿﻢ ﻓﻮرﯾﺮ وطﺮق اﻻﻧﺘﺸﺎر اﻟﻌﻜﺴﻲ اﻟﺮﻗﻤﻲ (DBP) ﻣﻦ ﺧﻼل إﯾﺠﺎد اﻟﺤﻞ اﻟﻤﻘﺪر ﻟـ "ﻣﻌﺎدﻟﺔ ﺷﺮود ﻧﻐﺮ ﻏﯿﺮ اﻟﺨﻄﯿﺔ (NLSE)" ، ﻟﻤﻌﺎﻟﺠﺔ اﻟﺘﺤﺪي اﻷﻛﺜﺮ ﺟﻮھﺮﯾﺔ اﻟﺬي ﯾﻮاﺟﮫ ﻣﺜﻞ ھﺬه اﻷﻧﻈﻤﺔ وھﻮ اﻻﻧﺤﻄﺎط ﻏﯿﺮ اﻟﺨﻄﻲ ، ﻟﻜﻦ ﺗﻨﻔﯿﺬه ﯾﺘﻄﻠﺐ اﻟﻜﺜﯿﺮ ﻣﻦ اﻟﻤﻮارد ﻟﻤﻌﺎﻟﺠﺔ اﻹﺷﺎرات ﺑﺎﻹﺿﺎﻓﺔ إﻟﻰ ﻓﮭﻢ دﻗﯿﻖ وﻋﺎﻟﻲ اﻟﻤﺴﺘﻮى ﻟﮭﺬه اﻷﻧﻈﻤﺔ. ﺑﯿﻨﻤﺎ ﺗﻌﺪ ﺧﻮارزﻣﯿﺎت اﻟﺬﻛﺎء اﻻﺻﻄﻨﺎﻋﻲ (AI) ﺑﺘﻘﺪﯾﻢ أﻓﻀﻞ اﻟﺤﻠﻮل ﻟﻠﺘﺤﺪﯾﺎت اﻟﻤﺴﺘﻘﺒﻠﯿﺔ اﻟﻤﻮﺻﻮﻓﺔ ﻣﻦ ﺧﻼل اﺳﺘﺨﺪاﻣﮭﺎ ﻟﺘﺤﺪﯾﺪ وﺣﻞ أوﺟﮫ اﻟﻘﺼﻮر ﻓﻲ اﻷﻧﻈﻤﺔ اﻟﻤﻌﻘﺪة ﻟﻠﻐﺎﯾﺔ واﻟﺒﯿﺌﺎت اﻟﻤﻌﻘﺪة دون اﻟﺤﺎﺟﺔ إﻟﻰ ﺗﻄﻮﯾﺮ ﻧﻤﻮذج ﻟﻸﺟﮭﺰة )اﻟﻤﺎدﯾﺔ.( طﺒﻘﺖ ھﺬه اﻷطﺮوﺣﺔ اﻟﺘﻌﻠﻢ اﻵﻟﻲ ﻋﻠﻰ أﺳﺎس اﻟﺸﺒﻜﺔ اﻟﻌﺼﺒﯿﺔ اﻟﻌﻤﯿﻘﺔ (DNN) ﻛﻄﺮﯾﻘﺔ ﺗﻌﻠﻢ ﺗﺤﺖ اﻹﺷﺮاف ﻟﻸﻧﻈﻤﺔ اﻟﺒﺼﺮﯾﺔ ﻟﺮﻓﻊ ﻣﺴﺘﻮى أداﺋﮭﺎ. اﻗﺘﺮﺣﺖ ھﺬه اﻟﺪراﺳﺔ ﺑﻨﺎء ﺗﻘﻨﯿﺔ NLC ﻟﻠﺘﻌﻮﯾﺾ ﻋﻦ اﻻﻧﺤﻄﺎط ﻏﯿﺮ اﻟﺨﻄﻲ اﻟﺬي ﯾﺤﺪث ﻓﻲ أﻧﻈﻤﺔ اﻻﺗﺼﺎﻻت اﻟﺒﺼﺮﯾﺔ. ﺗﻢ ﺗﺼﻤﯿﻢ اﻟﻨﻈﺎم وﻣﺤﺎﻛﺎﺗﮫ ﺑﺎﻟﻜﺎﻣﻞ ﻣﻊ ﺗﻄﺒﯿﻖ Optisystem. )اﻹﺻﺪار ﻓﻲ اﻟﻮﻗﺖ اﻟﻔﻌﻠﻲ ﺑﺎﺳﺘﺨﺪام ﺗﻄﺒﯿﻖ (V.2020a) MATLAB. (17.0.0 وان اﻻﺧﺘﺒﺎر / اﻟﺘﺤﺴﯿﻦ ﺗﻢ إﺟﺮاؤه ﻓﻲ ﺛﻼﺛﺔ اﺗﺠﺎھﺎت: أوﻻً: ﻓﻲ أﻧﻈﻤﺔ ﺑﺼﺮﯾﺔ ﻣﺘﻤﺎﺳﻜﺔ أﺣﺎدﯾﺔ اﻻﺳﺘﻘﻄﺎب ﻣﻊ 16-4 ﺗﺸﻜﯿﻞ QAM ﻣﻊ وﺑﺪون OFDM ﻣﺘﻌﺪد اﻟﻤﻮﺟﺎت اﻟﺤﺎﻣﻠﺔ اﻟﻔﺮﻋﯿﺔ ، ﯾﺘﻢ ﻧﻘﻠﮫ ﻋﺒﺮ اﻷﻟﯿﺎف أﺣﺎدﯾﺔ اﻟﻮﺿﻊ اﻟﻘﯿﺎﺳﯿﺔ أﺣﺎدﯾﺔ اﻟﻤﻮﺟﺔ (SSMF) ﺑﺴﺮﻋﺔ 60 ﺟﯿﺠﺎﺑﺖ ﻓﻲ اﻟﺜﺎﻧﯿﺔ ﻟﻤﺴﺎﻓﺔ ارﺗﺒﺎط ﺗﺒﻠﻎ 3000 ﻛﯿﻠﻮ ﻣﺘﺮ.

كلمات الباحثين

Optical Communication, Deep Neural Networks, Machine Learning, Nonlinear Compensation, MATLAB Simulation

إشراف وإعداد

الإشراف

اﺑراهيم ﻋﺒﺪ ﷲ ﻣﺮداس

الإعداد

علاء حمود عبد جراح

الاقتباسات الببليوغرافية