Pedro E. D. Cruz

Exhaustive gaussian approach for performance evaluation of direct-detection OFDM systems employing square and cross QAM

An exhaustive Gaussian approach (EGA) is proposed to evaluate, through numerical simulation, the bit error ratio (BER) of direct-detection orthogonal frequency-division multiplexing systems employing square and cross quadrature amplitude modulation. Excellent agreement between the BER estimates from the direct error counting (DEC) and the EGA is shown for different levels of optical signal-to-noise ratio and signal distortion. It is shown that the EGA requires about three orders of magnitude less in computation time than the DEC method for BER levels around 10-6, with the difference getting higher for lower BER levels.

Impact of inter-band crosstalk due to nonlinear fibre transmission on the performance of direct-detection single-sideband MB-OFDM metro networks

The impact of inter-band crosstalk due to nonlinear fibre transmission on the performance of direct-detection single-sideband multi-band orthogonal frequency division multiplexing metro networks is evaluated through numerical simulation for one optical channel. Distortion powers with single-band and with five bands are compared. The number of independent signal runs required to achieve accurate performance estimates is assessed. An average over 16 independent signal runs is shown enough to guarantee an error vector magnitude (EVM) discrepancy not exceeding 0.5 dB from the average EVM obtained with 32 independent signal runs. Inter-band crosstalk due to nonlinear transmission of 5 bands over 240 km of standard single-mode fibre limits the average power at the fibre input to levels not exceeding -2.5 dBm.

Impact of the band selector detuning on DD multi-band OFDM virtual-carrier assisted metro network performance

The impact of the band selector (BS) detuning on multi-band (MB) orthogonal frequency division multiplexing (OFDM) signals performance employing virtual-carriers in direct-detection networks is assessed through numerical simulation. Two, three and four-band MB-OFDM systems operating at 42.8 Gb/s are optimized for a Gaussian BS and its performance is compared with the one of a higher selectivity (2-nd order super Gaussian) BS. It is shown that the MB-OFDM system presents a maximum error vector magnitude (EVM) penalty of 1 dB for a BS frequency detuning range of 1.5 GHz, and a EVM difference lower than 1 dB is obtained between the two considered BS.

Experimental Assessment of 10 Gbit/s SSB OFDM IM-DD Systems Performance Employing Raised-Cosine Based Symbol Precoding

The performance improvement of 10 Gbit/s single-sideband orthogonal frequency division multiplexing intensity-modulated direct-detection systems achieved by reducing the signal peak-to-average power ratio (PAPR) with a symbol precoding technique based on the raised-cosine pulse, without and with bandwidth expansion, is evaluated experimentally. The impact of pre-emphasis (PE), necessary to mitigate the setup frequency limitations, on the PAPR reduction achieved by symbol precoding is investigated. At the input of the optical modulator, PAPR reductions of ~ 3 dB with PE and without PE are achieved using the precoding technique without band excess; with precoding with band excess of 25%, PAPR reductions of ~ 4 dB with PE and 4.5 dB without PE are obtained. Due to precoding, a bit error ratio (BER) improvement of almost two orders of magnitude is shown along 150 km standard single mode fiber transmission, for BERs around 10 -4 . In addition, it is also shown that precoding enables similar BER improvement with 75 km fiber propagation when nonlinear fiber effects are moderate.

Theoretical Analysis of the Four-Wave Mixing Effect in Virtual Carrier-Assisted DD MB-OFDM Ultradense WDM Metropolitan Networks

A theoretical method (TM) to evaluate the four-wave mixing (FWM) power in multiband orthogonal frequency division multiplexing (MB-OFDM) ultradense wavelength division multiplexing (WDM) metropolitan networks using virtual carriers (VCs) and direct detection is proposed and validated numerically. It considers a band spacing of 3.125 GHz and 10-Gb/s data rate per band. A simplified TM (STM) to obtain the FWM power for MB-OFDM systems employing high-dispersion fiber is also proposed to relax the computation complexity. Both methods consider that FWM products generated from the mixing between two VCs and an OFDM band are dominant. Analytical expressions for the maximum allowed average power of the worst band (the band most impaired by FWM) are also derived. The TM and STM enable estimating the average power level per WDM channel corresponding to the threshold error vector magnitude (EVM) (threshold average power-TAP), where the threshold EVM is the maximum allowed EVM due to FWM, which induces a 2-dB EVM penalty to the EVM at the forward error correction threshold. Results show that, when a total data bit rate of 450 Gb/s is transmitted (nine WDM channels with five bands each) along the standard single-mode fiber (SSMF), the discrepancy between the numerical simulation-estimated and TM-estimated TAPs does not exceed 0.5 dB. By defining the upper bound of the power limit (UBPL) as the TAP obtained with 17 WDM channels, the discrepancy between the TM-estimated and STM-estimated UBPLs does not exceed 1 dB for transmission along up to 10 SSMF spans and for channel gaps up to 22.81 GHz.

Loss budget of ultra-dense 10 Gb/s Per-user guaranteed direct-detection MB-OFDM metro-access networks

An ultra-dense direct-detection single-sideband multi-band orthogonal frequency division multiplexing system with 10 Gb/s per-user guaranteed is presented and its performance is evaluated through numerical simulation. The highest loss budget for a channel with 25 GHz bandwidth and 5 bands spaced by 3.125 GHz is assessed. An highest loss budget of 20.5 dB is obtained when filtering limitations are not taken into account. When they are taken into account, a reduction on the highest loss budget of 2 dB is shown.

Experimental demonstration of PAPR reduction in 10 Gbit/s SSB-OFDM direct-detection systems through symbol precoding

The performance improvement of single-sideband orthogonal frequency division multiplexing (OFDM) signals in 10 Gbit/s direct-detection systems for metropolitan telecommunication networks, by reducing the peak-to-average power ratio (PAPR) with a symbol precoding technique is assessed experimentally. The symbol precoding technique achieves a 3 dB PAPR reduction without adding excess bandwidth to the OFDM signal. A bit error ratio improvement of more than order of magnitude is experimentally shown along 150 km standard single mode fiber transmission.

Performance Evaluation of Optical Noise-Impaired Multi-band OFDM Systems Through Analytical Modeling

The performance evaluation through analytical modeling of virtual carrier-assisted direct-detection single-sideband multi-band orthogonal frequency division multiplexing (MB-OFDM) systems dominantly impaired by amplified spontaneous emission noise is presented. An analytical expression is obtained for the bit error ratio (BER) of these systems, using the analytical relationship between the electrical signal-to-noise ratio and the optical signal-to-noise ratio. In order to analyse the accuracy of the analytical expression, the BER estimates provided by the analytical model are compared with the ones obtained using numerical simulation and the exhaustive Gaussian approach. Excellent agreement between the BER estimates is verified when distortion does not affect significantly the performance of the MB-OFDM signal.

Analytical modelling for performance evaluation of distortion-impaired direct-detection single-sideband optical OFDM systems employing virtual-carriers

An analytical model (AM) for performance evaluation of distortion-impaired direct-detection single-sideband optical orthogonal frequency-division multiplexing systems employing virtual-carriers (VCs) is presented. The error vector magnitude (EVM) is used as figure of merit for performance evaluation, and the effectiveness of the AM is verified through comparison of its EVM estimates with the ones obtained using numerical simulation (NS). Discrepancies between the EVM estimates of the AM and NS not exceeding 1 dB are verified for a VC to band power ratio of 12 dB and for VC to band gaps exceeding half of the signal bandwidth.