Versatile Optical Network Planning
André Richter, VPIphotonics
Planning today’s and future optical networks requires versatile, powerful, and flexible network planning tools, which capture network infrastructure and equipment information, support and automatize design tasks, and compare and archive found solutions. They should support a vendor- and technology-agnostic approach to equipment modeling and performance assessment, which permits fast and efficient configuration of scalable networks, accounts for traffic demands, and optimizes spectrum utilization and equipment cost. We present the general concept of such a tool and highlight its operation by addressing several planning scenarios. We discuss a multi-band, multi-span DWDM network accounting for impairments such as ISRS and showcase metro network topologies. In addition, we present a planning approach for incorporating optical satellite networks into existing terrestrial network designs and show how interoperability with numerical system simulations can further enhance the capabilities of such a tool.
AI Carbon footprint: how to design low complexity and sustainable ai tools
Pedro Freire, Aston University
This presentation delves into the intricacies of developing low-complexity neural network-based equalizers tailored for optical communications. It explores innovative strategies to overcome challenges encountered in the realms of training, inference, and hardware synthesis.
Modeling, Processing and Detection of Optical Signals in Advanced Communication Systems
Tianhua Xu, University of Warwick
Fiber and wireless optic communications have played significant roles in increasing the data capacity of modern transmission systems. Laser and channel nonlinear distortions in optical fiber systems and the inter-symbol-interference and the time-varying channel impulse response in optical wireless communication systems have deteriorated the detection and the demodulation of transmitted symbols from the received signals in corresponding systems. In this talk, we will discuss the performance of advanced signal processing in mitigating the laser phase noise and the fiber nonlinearities under dispersion-unmanaged and dispersion-managed long-haul optical link conditions. We will also present signal detection schemes in ultra-violet optical wireless communication systems by designing optimal linear and high-dimensional combinations of extracted, signal-related geometrical features.
GPU-Accelerated Framework for Optical Communication System Design and Analysis
Egor Sedov, Aston University
The research paper introduces a GPU-accelerated framework for optical communication system simulations, demonstrating its superior performance over traditional CPU-based methods. This framework addresses the growing complexity and computational demands in the field of optical communication systems, providing a high-performance software solution tailored to their unique requirements. The key benefits of using GPU-based implementations include inherent parallelism, scalability, power efficiency, and the ability to handle large-scale matrix operations, vector manipulations, or Fourier transforms efficiently.
The framework, built on TensorFlow and Python, is compatible with MATLAB and includes telecommunication components like the transmitter, optical channel model, receiver, and metrics evaluation block. It utilizes GPU-accelerated Split-Step Fourier Method (SSFM) implementations for efficient and accurate simulation of optical communication systems, supporting various configurations and customizations. Performance evaluation showed significant speedup with GPU, especially for larger problem sizes, enabling researchers to rapidly test hypotheses, iterate on designs, and explore a wider range of scenarios.