Open Access
Add to BibSonomyAbstract
We experimentally demonstrate the use of data-aided digital signal processing for format-flexible coherent reception of different 28-GBd PDM and 4D modulated signals in WDM transmission experiments over up to 7680 km SSMF by using the same resource-efficient digital signal processing algorithms for the equalization of all formats. Stable and regular performance in the nonlinear transmission regime is confirmed.
©2012 Optical Society of America
1. Introduction
Since the future internet traffic is expected to be much more dynamic than in the past, options for the flexibilization of optical networks and elastic networking have recently attracted a lot of attention [1,2]. In particular, the flexible generation and reception of modulation formats with different spectral efficiencies using a single transponder (i.e. software-defined) will allow to optimize the use of the network infrastructure under dynamic bandwidth demands [3].
While the combination of a high-speed digital-to-analog converter (DAC) with a dual-polarization in-phase- and quadrature (DP-IQ)-modulator in the transmitter and a high-speed analog-to-digital converter (ADC) with coherent reception in the receiver in principle already allows for generation and reception of different modulation formats, format-flexible and therefore resource-efficient digital signal processing (DSP) is the key technology for the practical realization of such a transponder.
In this context, data-aided equalization algorithms [4–6], have many advantages over popular blind algorithms like the constant-modulus algorithm because they work independently of the modulation format in the payload signal and provide stable performance.
In this contribution, we experimentally show the use of data-aided channel estimation in combination with resource-efficient frequency-domain equalization for the wavelength-division multiplexed (WDM) transmission of 28-GBd polarization-division multiplexed binary phase-shift keying (PDM-BPSK), polarization-switched quaternary PSK (PS-QPSK) [7–10], PDM-QPSK and PDM 8-ary quadrature amplitude modulation (PDM-8QAM) over an uncompensated ITU-T G.652 standard single-mode fiber (SSMF) link with up to 7680 km length. The different signals were generated by using a high-speed four-channel DAC at the transmitter and provided varying bit rates from 56 Gb/s to 168 Gb/s within the same optical bandwidth as well as different maximal transmission reaches. All formats carried the same header sequence and were equalized offline using the same algorithms.
2. Data-aided digital signal processing
The used header structure and the used offline algorithms are shown in Fig. 1 . The header was composed of three different sequences (BPSK and QPSK symbols) which were used for frame synchronization, carrier frequency offset (CFO) compensation and channel estimation [5,6]. Each of the sequences in the Y-polarization was a cyclic-shifted copy of the corresponding X-polarization sequence. The header was periodically repeated with a total overhead of 1.17%, corresponding to a negligible required optical signal-to-noise ratio (OSNR) penalty of less than 0.1 dB. In the DSP, the signal was first resampled to two samples per symbol. After front-end corrections [11], the accumulated chromatic dispersion (CD) was blindly estimated [12], using a block of 4096 samples and compensated in the frequency domain on the full shot within one step to reduce the length of the channel impulse response and therefore the required length of the header. The following frame synchronization was based on an autocorrelation metric [13], and used the framing sequences in the header consisting of BPSK symbols. Then, the carrier frequency offset was estimated in two data-aided stages [14]. The first stage used the header CFO synchronization sequences consisting of QPSK signals and had an acquisition range of ± 1.5 GHz. The second stage used the pilot sequences for fine adjustment. These sequences carried constant-amplitude-zero-autocorrelation (CAZAC) sequences [5], consisting of 16 QPSK symbols and were primarily used for the channel estimation. The equalizer matrix was directly calculated from the estimated channel matrix using the minimum-mean-square error criterion [5]. The following feed-forward add-and-overlap frequency-domain equalizer [15], used a block size of 1024 samples and an overlap of 25%. The carrier phase estimation for all PSK formats (including PS-QPSK) was conducted using the format-flexible block-based blind feed-forward Viterbi-Viterbi (VV) algorithm [16]. For PDM-8QAM, we used the blind-phase search (BPS) algorithm [17] with 32 test angles. We also tested this algorithm with the other formats and found a similar performance as for the VV algorithm under our experimental conditions. The used averaging block size was 32 in all cases. Since the BPS algorithm can be in principle used for all types of modulation formats including PSK, QAM and 4D modulation, it enables full format-flexible carrier phase estimation completing the format-flexibility of all algorithms of the DSP shown in Fig. 1(b). After the header removal, decision and demapping, the errors were counted and the resulting bit-error ratio (BER) was converted to a Q-factor.
3. Experimental setup
The experimental setup is shown in Fig. 2 . At the transmitter, an external cavity laser (ECL) with a linewidth of ~100 kHz was used as a light source for the probe channel at 193.4 THz (1550.115 nm). For the remaining WDM channels, distributed feedback lasers (DFB) with linewidth <10 MHz were used which were located on the 50-GHz ITU-T channel grid starting at 192.95 THz. Even and odd channels were separately modulated by two integrated DP-IQ-modulators. The drive signals for the modulators were generated by a 4-channel DAC with a sample rate of 56 GS/s, a resolution of 8 bits and a 3-dB bandwidth of about 8 GHz. A digital pre-emphasis of the drive signals was applied offline before uploading the samples onto the DAC chip. The transmitted data was a de Bruijn binary sequence of length 214. Eye diagrams of the resulting output signals of the DAC for the different formats are shown in Fig. 3 . The header BPSK and QPSK symbols were always placed on regular constellation points of the different formats except for PDM-BPSK, where the header QPSK symbols cause a zero-rail in the eye diagram of the drive signal. For PS-QPSK, d4 = d1 xor d2 xor d3 was set [7]. After modulation, even and odd channels were combined by a 50-GHz interleaver (ILV).
The combined WDM signal was launched into a recirculating fiber loop incorporating three 80-km spans of SSMF and erbium-doped fiber amplifiers (EDFA) including a loop-synchronous polarization scrambler and a programmable gain equalizing filter. The WDM spectrum at the transmitter and after 7200-km transmission is shown in Fig. 4 for PDM-BPSK modulation.
At the receiver, the OSNR could be degraded by noise loading using a variable optical attenuator (VOA). The delivered OSNR at the receiver was measured in front of the 0.4-nm channel selection filter. The local oscillator (LO) ECL had a linewidth of ~100 kHz and was superimposed with the signal in a polarization-diversity optical 90° hybrid. The outputs of the hybrid were connected to four balanced photo-detectors. The resulting signals were digitized by four analog-to-digital converters with a sample rate of 50 GS/s and a bandwidth of 20 GHz. The digital samples were then processed offline in a computer. Up to 2.5 million symbols were evaluated to ensure a sufficient number of errors for all tested BER values.
4. Results and discussion
The transmitter generated 20 WDM-channels with 50 GHz channel spacing modulated with different modulation formats, namely PDM-BPSK (56 Gb/s), PS-QPSK (84 Gb/s), PDM-QPSK (112 Gb/s) and PDM-8QAM (168 Gb/s). The measured back-to-back Q-factor per-formance of the probe channel at 1550.115 nm is shown as a function of the OSNR in Fig. 5(a) . At the hard-decision FEC limit of 3.8x10−3 (Q = 8.53 dB), implementation penalties of 2.2 dB for PDM-BPSK, 1.7 dB for PS-QPSK, 2.0 dB for PDM-QPSK, and 4.1 dB for PDM-8QAM were observed, which mainly originated from performance differences of the DAC channels. PDM-8QAM suffered most from the limited DAC bandwidth resulting in an error floor at a BER of 5x10−4 at maximum OSNR, whereas the other modulation formats performed error-free (BER < 10−6) at the maximum OSNR of 42 dB. The optical back-to-back constellation diagrams at maximum OSNR after DSP and header removal are shown in Fig. 6 and show the performance differences between the X- and Y-polarization. Please note that the received PS-QPSK constellation shown in Fig. 6 corresponds to the transmitter constellation generated by the two-level driving signals shown in Fig. 3. It can be transformed to the constellation with the well-known spot at the center by a 45° polarization rotation [7,10].
Fig. 5 (a) Measured back-to-back Q-factor at a symbol rate of 28 GBd for PDM-BPSK (black) and PS-QPSK (red) on the left hand side as well as PDM-QPSK (green) and PDM-8QAM (blue) on the right hand side. The solid lines show the theoretical Q-factor for an AWGN channel. (b) Measured Q-factor at a symbol rate of 28 GBd as a function of the transmission length over standard single-mode fiber for PDM-BPSK (black), PS-QPSK (red), PDM-QPSK (green), and PDM-8QAM (blue).
Fig. 6 Optical back-to-back constellation diagrams for the X- and Y-polarizations after DSP and header removal in the case of maximum OSNR.
Transmission measurements were conducted for the optimum span launch power of −2 dBm per channel. The Q-factor of the probe channel versus the transmission length is presented in Fig. 5(b). PDM-BPSK and PS-QPSK showed a comparable performance and enabled a maximum reach of 7680 km and 6720 km, respectively, with a Q-factor above the FEC-limit. For large Q-factors the curves intersect and PS-QPSK performed better than PDM-BPSK at Q-factors larger than 11 dB which was attributed to the error floor for PDM-BPSK already seen in the back-to-back curves shown in Fig. 3a. Using PDM-QPSK, a maximum reach of 4320 km was achieved, whereas PDM-8QAM was limited to 960 km. These results show that the same data-aided DSP enables stable equalization at any tested transmission length for any of the tested formats, i.e. for PDM-PSK, PDM-QAM and 4D formats like PS-QPSK.
In Fig. 7 , the measured Q factor as a function of the launch power per channel for a transmission length close to maximum is shown. As stated above, the optimum launch power close to −2 dBm is independent of the modulation format. This independence is in good agreement with recent studies on the nonlinear performance of different modulation formats [18].
Fig. 7 Measured Q-factor at a symbol rate of 28 GBd as a function of the launch power per channel for close-to-maximum-length transmission over standard single-mode fiber for PDM-BPSK (black), PS-QPSK (red), PDM-QPSK (green), and PDM-8QAM (blue).
5. Conclusions
We experimentally demonstrated the use of data-aided equalization algorithms for resource-efficient and format-flexible DSP in WDM transmission experiments over up to 7680 km. Using the same algorithms and the same header sequences for different PDM and 4D formats enabled bit-rate-, spectral-efficiency- and reach-variable transmission from 56 Gb/s to 168 Gb/s with spectral efficiencies from 1 to 3 bit/s/Hz and transmission reaches from 960 km to 7680 km.
Acknowledgments
This work was funded by the German Federal Ministry of Education and Research under the grant 01BP12402.
References and links
1. O. Gerstel, M. Jinno, A. Lord, and S. J. Yoo, “Elastic Optical Networking: A new dawn for the optical layer?” IEEE Commun. Mag. 50(2), s12– s20 (2012). [CrossRef]
2. M. Angelou, K. Christodoulopoulos, D. Klonidis, A. Klekamp, F. Buchali, E. Varvarigos, and I. Tomkos, “Spectrum, cost and energy efficiency in fixed-grid and flex-grid networks,” in Proc. Opt. Fiber Commun. Conf. (2012), paper NM3F.4.
3. M. Eiselt, B. Teipen, K. Grobe, A. Autenrieth, and J.-P. Elbers, “Programmable modulation for high-capacity networks,” in Proc. European Conf. Opt. Commun. (2011), paper Tu.5.A.5.
4. R. Dischler, “Experimental comparison of 32- and 64-QAM constellation shapes on a coherent PDM burst mode capable system,” in Proc. European Conf. Opt. Commun. (2011), paper Mo.2.A.6.
5. M. Kuschnerov, M. Chouayakh, K. Piyawanno, B. Spinnler, E. de Man, P. Kainzmaier, M. S. Alfiad, A. Napoli, and B. Lankl, “Data-aided versus blind single-carrier coherent receivers,” IEEE Photon. J. 2(3), 387–403 (2010).
6. F. Pittalà, F. N. Hauske, Y. Ye, N. G. Gonzalez, and I. T. Monroy, “Data-aided frequency-domain 2×2 MIMO equalizer for 112 Gbit/s PDM-QPSK coherent transmission systems,” in Proc. Opt. Fiber Commun. Conf. (2012), paper OM2H.4.
7. M. Karlsson and E. Agrell, “Which is the most power-efficient modulation format in optical links?” Opt. Express 17(13), 10814–10819 (2009). [CrossRef] [PubMed]
8. M. Sjödin, P. Johannisson, H. Wymeersch, P. A. Andrekson, and M. Karlsson, “Comparison of polarization-switched QPSK and polarization-multiplexed QPSK at 30 Gbit/s,” Opt. Express 19(8), 7839–7846 (2011). [CrossRef] [PubMed]
9. D. S. Millar, D. Lavery, S. Makovejs, C. Behrens, B. C. Thomsen, P. Bayvel, and S. J. Savory, “Generation and long-haul transmission of polarization-switched QPSK at 42.9 Gb/s,” Opt. Express 19(10), 9296–9302 (2011). [CrossRef] [PubMed]
10. J. K. Fischer, L. Molle, M. Nölle, D.-D. Groß, C. Schmidt-Langhorst, and C. Schubert, “Experimental investigation of 84-Gb/s and 112-Gb/s polarization-switched quadrature phase-shift keying signals,” Opt. Express 19(26), B667–B672 (2011). [CrossRef] [PubMed]
11. I. Fatadin, S. Savory, and D. Ives, “Compensation of quadrature imbalance in an optical QPSK coherent receiver,” IEEE Photon. Technol. Lett. 20(20), 1733–1735 (2008). [CrossRef]
12. R. A. Soriano, F. N. Hauske, N. G. Gonzalez, Z. Zhang, Y. Ye, and I. T. Monroy, “Chromatic dispersion estimation in digital coherent receivers,” J. Lightwave Technol. 29(11), 1627–1637 (2011). [CrossRef]
13. K. Shi and E. Serpedin, “Coarse frame and carrier synchronization of OFDM systems: a new metric and comparison,” IEEE Trans. Wirel. Comm. 3(4), 1271–1284 (2004). [CrossRef]
14. T. M. Schmidl and D. C. Cox, “Robust frequency and timing synchronization for OFDM,” IEEE Trans. Commun. 45(12), 1613–1621 (1997). [CrossRef]
15. A. Leven, N. Kaneda, and S. Corteselli, “Real-time implementation of digital signal processing for coherent optical digital communication systems,” IEEE J. Sel. Top. Quantum Electron. 16(5), 1227–1234 (2010). [CrossRef]
16. A. J. Viterbi and A. M. Viterbi, “Nonlinear estimation of PSK-modulated carrier phase with application to burst digital transmission,” IEEE Trans. Inf. Theory 29(4), 543–551 (1983). [CrossRef]
17. T. Pfau, S. Hoffmann, and R. Noé, “Hardware-efficient coherent digital receiver concept with feedforward carrier recovery for m-QAM constellations,” J. Lightwave Technol. 27(8), 989–999 (2009). [CrossRef]
18. A. Carena, V. Curri, G. Bosco, P. Poggiolini, and F. Forghieri, “Modeling of the impact of nonlinear propagation effects in uncompensated optical coherent transmission links,” J. Lightwave Technol. 30(10), 1524–1539 (2012). [CrossRef]
