End-to-end Learning for Optical Fiber Communication with Data-driven Channel Model
Mingliang Li, Danshi Wang, Qichuan Cui, Zhiguo Zhang, Linhai Deng, Min Zhang
Abstract
An end-to-end autoencoder is proposed for optical fiber communication system, which is based on the data-driven channel modeled by BiLSTM technique. The data-driven channel not only solve the problem about back-propagation but reflect the complex channel characteristic. The proposed can effectively reduce the dependence on expert knowledge and obtains the lower BER than conventional method.
Topics & Concepts
Channel (broadcasting)Computer scienceAutoencoderOptical fiberFiberEnd-to-end principleData modelingElectronic engineeringTelecommunicationsDeep learningArtificial intelligenceEngineeringMaterials scienceDatabaseComposite materialOptical Network TechnologiesWireless Signal Modulation ClassificationAdvanced Photonic Communication Systems