Optimizing FFT Precision in Optical OFDM Transceivers

Bouziane, Rachid; Koutsoyannis, Robert; Milder, Peter; Benlachtar, Y.; Hoe, James C.; Glick, Madeleine; Killey, Robert I.

doi:10.1109/lpt.2011.2164059

Cited by 17 publications

(9 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

Section: Introductionmentioning

confidence: 99%

“…As the processing speed of a typical FPGA is constrained at hundreds of MHz compared with multi-GHz DACs/ADCs, a highly parallel and pipelined IFFT architecture is thus crucial, which, however, introduces significant difficulties in reusing the complex functions that are typically employed in low-speed OFDM systems. Investigations of the impacts of the IFFT bit resolution on the OOFDM transceiver performance have been reported [11][12][13], where the IFFT operation is treated as a "black-box" without taking into account bit resolution variations between different intermediate IFFT stages.In this paper, extensive numerical explorations are undertaken, for the first time, of the IFFT stage-dependent minimum bit resolutions required for achieving the specific transceiver performance. An optimum map of minimum bit resolutions of different IFFT stages against DAC resolutions is obtained for different signal modulation formats.…”

mentioning

confidence: 99%

See 1 more Smart Citation

IFFT Stage-dependent Minimum Bit Resolution Maps for Real-time Optical OFDM Transceivers

Zhang

Yuan

Giddings

et al. 2014

Optical Fiber Communication Conference

View full text Add to dashboard Cite

To significantly reduce the FPGA/ASIC resource usage in OOFDM transceivers, numerical identifications and experimental verifications are undertaken, for the first time, of an optimum map of minimum bit-resolutions of different IFFT stages against DAC resolutions. OCIS codes: (060.2330) Fiber optic communications, (060.4080) Modulation IntroductionOptical OFDM (OOFDM) is a promising technique for practically realizing intelligent transceivers for future elastic optical networks [1]. The inherent DSP-rich OOFDM transceivers offer salient features including, for example, automatic awareness of channel spectral characteristics, adaptability to system/network imperfections, dynamically variable capacity versus reach performance and user-controlled capability of channel add/drop multiplexing. Over the past several years, a number of real-time OOFDM implementations of transmitters [2-5], receivers [6-8] and transceivers [9,10] have been reported using FPGAs. Technically speaking, the huge FPGA logic resource usage associated with the high OOFDM DSP complexity has become one of the most significant obstacles to experimentally demonstrating real-time OOFDM transceivers having more advanced functionalities and flexibilities. For example, just the IFFT DSP algorithm alone can take >82% of the total FPGA logic resources [4]. As such several FPGAs have to be combined together to perform all necessary OOFDM DSP algorithms [6,7].From the practical application point of view, it is vital to explore effective approaches capable of minimizing, via optimizing the DSP operations, the FPGA resource usage without degrading the transceiver performance. Given the fact that the IFFT is the fundamental algorithm at the heart of the OOFDM transceivers, in this paper, special attention is, therefore, focused on the IFFT operation. As the processing speed of a typical FPGA is constrained at hundreds of MHz compared with multi-GHz DACs/ADCs, a highly parallel and pipelined IFFT architecture is thus crucial, which, however, introduces significant difficulties in reusing the complex functions that are typically employed in low-speed OFDM systems. Investigations of the impacts of the IFFT bit resolution on the OOFDM transceiver performance have been reported [11][12][13], where the IFFT operation is treated as a "black-box" without taking into account bit resolution variations between different intermediate IFFT stages.In this paper, extensive numerical explorations are undertaken, for the first time, of the IFFT stage-dependent minimum bit resolutions required for achieving the specific transceiver performance. An optimum map of minimum bit resolutions of different IFFT stages against DAC resolutions is obtained for different signal modulation formats. The high accuracy of the map is experimentally verified in a real-time FPGA platform. The map can be used to significantly reduce the FPGA logic resource usage without degrading the transceiver performance.

show abstract

Section: Introductionmentioning

confidence: 99%

mentioning

confidence: 99%

IFFT Stage-dependent Minimum Bit Resolution Maps for Real-time Optical OFDM Transceivers

Zhang

Yuan

Giddings

et al. 2014

Optical Fiber Communication Conference

View full text Add to dashboard Cite

show abstract

“…Despite the variation in the constellation sizes between QPSK, 16-QAM, and 64-QAM, the EVMs achieved close proximity due to the normalization factor of the EVM relative to the individual constellation sizes; see Eq. (10) [32,34]. However, a larger constellation size requires a lower EVM to achieve the same bit error rate compared to a smaller constellation size [32,34].…”

Section: Best Case and Worst Case Transmission Conditionmentioning

confidence: 99%

“…(10) [32,34]. However, a larger constellation size requires a lower EVM to achieve the same bit error rate compared to a smaller constellation size [32,34].…”

Section: Best Case and Worst Case Transmission Conditionmentioning

confidence: 99%

Experimental Full Duplex Simultaneous Transmission of LTE Over a DWDM Directly Modulated RoF System

Kanesan¹,

Ng²,

Ghassemlooy³

et al. 2013

J. Opt. Commun. Netw.

View full text Add to dashboard Cite

Abstract-In this paper, we experimentally demonstrate the seamless integration of full duplex system frequency division duplex (FDD) long-term evolution (LTE) technology with radio over fiber (RoF) for eNodeB (eNB) coverage extension. LTE is composed of quadrature phase-shift keying (QPSK), 16-quadrature amplitude modulation (16-QAM) and 64-QAM, modulated onto orthogonal frequency division multiplexing (OFDM) and single-carrier-frequency division multiplexing for downlink (DL) and uplink (UL) transmissions, respectively. The RoF system is composed of dedicated directly modulated lasers for DL and UL with dense wavelength division multiplexing (DWDM) for instantaneous connections and for Rayleigh backscattering and nonlinear interference mitigation. DL and UL signals have varying carrier frequencies and are categorized as broad frequency spacing (BFS), intermediate frequency spacing (IFS), and narrow frequency spacing (NFS). The adjacent channel leakage ratio (ACLR) for DL and UL with 64-QAM are similar for all frequency spacings while cross talk is observed for NFS. For the best case scenario for DL and UL transmissions we achieve error vector magnitude (EVM) values of ∼2.30%, ∼2.33%, and ∼2.39% for QPSK, 16-QAM, and 64-QAM, respectively, while for the worst case scenario with a NFS EVM is increased by 0.40% for all schemes.

show abstract

“…The required resolution of a digital analog converter (D/A) and an analog digital converter (A/D) for OFDM systems is reported in [7]. The authors of [8] discussed the relationship between fast Fourier transform (FFT) size and calculation precision, and they suggested that reducing the precision of the FFT and inverse FFT (IFFT) allows us to reduce the power of the DSP [9]. However, there is no precision management approach for OFDM-PON systems according to transmission distance designed to reduce power consumption.…”

Section: Introductionmentioning

confidence: 99%

Numerical Analysis of Dynamic SNR Management by Controlling DSP Calculation Precision for Energy-Efficient OFDM-PON

Kimura

Nakamura

Kimura

et al. 2012

IEEE Photon. Technol. Lett.

View full text Add to dashboard Cite

We propose a novel energy-saving technique using a variable precision digital signal processor for an orthogonal frequency division multiplexing (OFDM)-passive optical network. Our technique minimizes energy consumption while satisfying the bit error rate requirements by controlling the calculation precision of fast Fourier transform (FFT) and inverse FFT when generating an OFDM signal. We clarify the feasibility of our proposed technique numerically for QPSK and 16 quadraticamplitude modulation signals. The simulation results indicate that calculation precision can be reduced according to the transmission distance.Index Terms-Digital signal processing, energy reduction, fiber optics communications, orthogonal frequency division multiplexing passive optical network (OFDM-PON).

show abstract

Optimizing FFT Precision in Optical OFDM Transceivers

Cited by 17 publications

References 5 publications

IFFT Stage-dependent Minimum Bit Resolution Maps for Real-time Optical OFDM Transceivers

IFFT Stage-dependent Minimum Bit Resolution Maps for Real-time Optical OFDM Transceivers

Experimental Full Duplex Simultaneous Transmission of LTE Over a DWDM Directly Modulated RoF System

Numerical Analysis of Dynamic SNR Management by Controlling DSP Calculation Precision for Energy-Efficient OFDM-PON

Contact Info

Product

Resources

About