1,721,012 research outputs found
End-to-end Deep Learning for VCSEL’s Nonlinear Digital Pre-Distortion
Full text linkWe propose a novel optimization method for a Neural Network based Digital Pre-Distorter (DPD), applied in Intensity Modulation-Direct Detection transmission systems leveraging Multi-Modal Fiber and Vertical-Cavity Surface-Emitting Laser. We train the DPD using End-to-end Deep Learning of the optical link, together with a Direct Learning Approach leveraging experimental measurements for modeling the transmission channel. The optimization considers VCSEL amplitude constraints, the use of an FFE at the receiver side, and the presence of a receiver non-flat Colored Gaussian Noise (CGN). We verify our optimized DPD on an experimental setup transmitting a 92 Gbps PAM-4 modulated signal. We achieve, for BER=0.01, a performance gain of more than 1 dB in terms of Optical Path Loss with respect to the best performing non-pre-distorted scenario
MMF-based Data Center Interconnect using Commercial Coherent Transceivers
Full text linkIn this manuscript we present an experimental analysis of a multi-mode fiber (MMF) link based on coherent detection. We propose the use of a commercial coherent transceiver to increase the transmission speed of a short-reach optical communication system over 300 m OM3 fiber. The performance of the system is studied in terms of bit rate as a function of the power budget margin (PBM) defined as the extra attenuation that can be introduced on the optical path. We investigate polarization multiplexed (PM) QPSK and 16-QAM modulation formats achieving up to 400 Gbps net bit rate transmission. Moreover, we analyze the impact of the lateral offset introduced by the connectors between MMF segments, showing decreasing PBM for increasing lateral offset. Nevertheless, PBM in excess of 28 dB for PM-QPSK modulation at 100G and 200G, 23 dB for 200G PM-16QAM modulation and 16 dB for 400G PM-16QAM, shows that commercial coherent transceivers can be used on MMF links up to much higher bit rates than those achieved by current VCSEL+direct detection based systems, provided that connections along the MMF have connectors connectors with offsets in the 3 μm to 6 μm range
Nonlinear Pre-distortion through a Multi-rate End-to-end Learning Approach over VCSEL-MMF IM-DD Optical Links
Full text linkWe experimentally demonstrate a nonlinear digital pre-distorter for PAM-M shaping in VCSEL+MMF IM-DD links able to operate at a generic baud rate using a fractional sample-per-symbol Neural Network. We focus on efficient and practical multi-rate operation, signal amplitude constraints, and linear equalizer at the receiver
Experimental VCSEL Digital Twin modeling for net 100 Gb/s/λ nonlinear Digital Pre-Distortion
Full text linkWe experimentally model a VCSEL-based optical
transmitter for high speed intra data center interconnects using
a convolutional neural network digital twin. The device is able to
effectively reproduce the VCSEL linear and nonlinear distortions
on PAM4 signals transmitted at 107.2 Gbps, thus enabling the
optimization of nonlinear VCSEL-MMF digital pre-distorters
An Analytical Model for Performance Estimation in Modern High-Capacity IMDD Systems
Full text linkIn this article, we propose an analytical model to estimate the signal-to-noise ratio (SNR) and then the Bit Error Rate (BER) at the output of a receiver adaptive equalizer in intensity modulation and direct detection (IMDD) optical transmission systems affected by optoelectronic bandwidth limitations, chromatic dispersion (CD), quantization noise, relative intensity noise (RIN), shot noise and thermal noise. We consider that the proposed model is a powerful tool for the numerical design of strongly band-limited IMDD systems using receiver equalization, as it happens in most of modern and future M-PAM solutions for short reach and access systems. We develop the model as an extension of a previously presented one, and then we test its accuracy by sweeping the main parameters of a 4-PAM-based communication system, such as RIN coefficient, extinction ratio (ER), accumulated CD, equalizer type and memory. Our findings show a remarkable agreement between time-domain simulations and analytical results, with SNR discrepancies below 0.1 dB in most cases, for both feed-forward and decision-feedback equalization. Moreover, we tested our model predictions against experimental measurements, confirming its accuracy
A Multi-Rate Approach for Nonlinear Pre-Distortion Using End-to-End Deep Learning in IM-DD Systems
Full text linkModern intra-data center (IDC) interconnects leverage robust and low-cost intensity modulation (IM) and direct detection (DD) optical links, based on multimode fibers (MMFs) and vertical-cavity surface-emitting lasers (VCSELs). Current solutions, based on on-off keying (OOK) modulations, reach up to 25-50 Gbps per lane over nearly 100 meters. The actual target for IDCs is to increase VCSEL-MMF links capacity up to 100 Gbps, using PAM-4 on the same devices. To counteract the consequent linear and nonlinear distortions affecting the transmitted signals, an effective solution is to exploit digital signal processing (DSP). In this manuscript, we propose a novel method to optimize a nonlinear artificial neural network (ANN) digital pre-distorter (DPD), based on End-to-end (E2E) learning, that, trained jointly with a Feed-Forward Equalizer (FFE), fulfills physical amplitude constraints and handles different ratio between the sampling rates incurring along with an optical IM-DD system. We indeed propose an E2E ANN system operating simultaneously at different sampling frequencies. We moreover propose in our training method a substitution to the time-domain injection of the receiver noise in the system with an additive regularization term in the FFE gradient loss. We experimentally show the advantages of our proposed DPD comparing the bit error rate (BER) performance against the same scenario without DPD. We assess the gain in terms of Gross Bit Rate and Optical Path Loss (OPL), at given BER targets, for different fiber lengths
TDECQ optimization of VCSEL-MMF nonlinear digital pre-distorters using end-to-end learning
Full text linkWe optimize nonlinear Digital Pre-Distorters for VCSEL-MMF links using an End-to-end (E2E) learning architecture focused on TDECQ IEEE specifications for 100 Gbps/lambda. We experimentally demonstrate that our E2E training improves the TDECQ performance by more than 0.8 dB compared to Direct Learning
Signal and Raman Pump Launch Power Optimization in a C+L+S+E System Using Fast Power Profile Estimation
No full textWe speed up signal and pump spatial power profile calculation with ISRS and backward
Raman amplification, demonstrating a 40x computational efficiency increase in the optimization of a 1000km C+L+S+E link by means of a GN/EGN model closed-form
Non-Linear Phase Noise Mitigation over Systems using Constellation Shaping
Full text linkThis paper presents a modified soft-decoding strategy, which improves performance in the presence of strong phase noise. This can substantially increase the reach of systems that are severely affected by phase noise, generated by fiber non-linear Kerr effect. This strategy is applied to two different experimental scenarios employing constellation shaping, which is known to generate strong non-linear phase noise. In the first experiment, we show that the strategy significantly improves the performance of probabilistically shaped (PS) 64 quadratic-amplitude modulation (QAM) over low-dispersion fibers. In the second experiment, the strategy is used to optimize the position of the points of a 32-QAM constellation (geometrical shaping). This optimized constellation is then compared to standard 32-QAM and PS 64-QAM over standard single-mode fiber. Also in this case, the modified strategy is able to give significant reach gains
- …
