Open Access
Add to BibSonomyAbstract
We demonstrate the transmission of 108-Gb/s polarization-division-multiplexed (PDM) 16-ary amplitude- and differential phase-shift-keying (16ADPSK) signal by using a non-coherent receiver. We generate the 16ADPSK signal by using a differential 8-ary phase-shift-keying (D8PSK) modulator and a phase-distortion-free amplitude-shift-keying (ASK) modulator. On the other hand, the receiver is implemented by using a delay interferometer based on a 3×3 fiber coupler and the data-aided phase-noise estimation (DAPNE) algorithm. By using these transmitter and receiver, we achieve a nearly quantum-limited receiver sensitivity in the back-to-back condition. In addition, we examine the possibility of transmitting 108-Gb/s signals on a 25-GHz grid without using the coherent detection technology. The results show that we can secure a sufficient optical-signal-to-noise (OSNR) margin after the transmission of 80-km long dispersion managed link. The achieved spectral efficiency is 4.0 bit/s/Hz.
©2009 Optical Society of America
1. Introduction
The multilevel differential phase-shift-keying (DxPSK) format has attracted significant attention due to its high spectral efficiency (SE) and high tolerance to linear and nonlinear impairments [1–4]. In addition, the information encoded in the differential phase can be detected by using a simple direct-detection receiver based on the delay interferometers (DIs) instead of the complicate coherent receiver [5,6]. Recently, there have been numerous attempts to exploit these advantages of DxPSK signals [1–4]. For example, by using the differential quadrature phase-shift-keyed (DQPSK) format together with the polarization-division-multiplexing (PDM), the transmission capacity of >20 Tb/s (with the SE as high as 3.2 bit/s/Hz) has been already demonstrated [2,3]. However, to further increase the SE (and the transmission capacity), we need to cope with the penalty caused by the stringent optical bandpass filtering as well as the crosstalk from neighboring channels (since the non-coherent detection does not have a frequency selectivity at the baseband stage, unlike the coherent detection [5]). Thus, we may need to utilize the amplitude modulation in addition to the multilevel phase modulation. In fact, there have already been some proposals of using the amplitude-shift-keying (ASK) and DxPSK formats together (ADPSK) [7–11]. However, it is not simple to generate such ADPSK signal without distortions. In addition, as we increase the number of bits per symbol, the requirement of the optical signal-to-noise ratio (OSNR) increases as well and the receiver sensitivity becomes vulnerable to the imperfections in the DIs [12–14]. Thus, in order to secure a sufficient OSNR margin for long-distance transmission, it is critical to avoid the excessive penalties caused in the transmitter and the receiver.
In this paper, we demonstrate the transmission of 54-Gb/s 16ADPSK signal and PDM 108-Gb/s signal by using a direct-detection receiver. To realize 16ADPSK signal, we used the combination of differential 8-ary PSK (D8PSK) and binary ASK formats. This signal has the highest theoretical sensitivity among all the (A)DPSK formats capable of providing 4 bit/symbol. For example, the receiver sensitivity of this signal is better than those of D16PSK and 16ADPSK (implemented by using DQPSK and 4-ary ASK formats) signals by 3.5 dB and 1.4 dB, respectively [14]. However, this signal has a relatively small tolerance to the phase noise due to the use of the D8PSK tributary [15–17]. Because of this small tolerance, most of the previous demonstrations on 4 bit/symbol have utilized the combination of DQPSK and 4-ary ASK formats [7,8]. In contrast, we overcome this problem by minimizing the deterioration of the D8PSK tributary (caused by ASK modulation) using a novel phase-distortion-free ASK modulation technique, in which the chirp and extinction ratio (ER) are adjusted independently. In addition, we implement the receiver by using an adjustment-free DxPSK demodulator, made of a 120-degree optical hybrid [18–20]. It has been reported that this demodulator is very effective in the suppression of the distortion occurred in the receiver [19,20]. We also utilize the data-aided phase-noise estimation algorithm (DAPNE) based on the multi-symbol phase estimation (MSPE) to reduce the differential phase noise [21–23]. The implemented receiver has a nearly quantum-limited sensitivity. By using this receiver, we demonstrate the error-free transmission of 108-Gb/s PDM 16ADPSK signals on a 25-GHz WDM grid. The achieved SE is 4 bit/s/Hz.
2. Transmitter and receiver for 54-Gb/s (108-Gb/s with PDM) 16ADPSK signal
2.1 Transmitter
Figure 1 shows the schematic diagram of the 16ADPSK transmitter. We used an external-cavity laser (ECL) operating at 193.6 THz as a light source. The 3-dB linewidth was 170 kHz. The output of this laser was first modulated in the D8PSK format by using a combination of a DQPSK modulator and a phase modulator.
We then applied the ASK modulation. Since the D8PSK tributary has a small phase tolerance, it is critical to minimize the parasitic phase distortion added by the ASK modulation. However, even a chirp-free Mach-Zehnder (MZ) intensity modulator can cause a small amount of the phase distortion depending on the power splitting ratio between the two arms in the MZ modulator (i.e., extinction ratio (ER) of the modulator). For example, Fig. 2(a) shows the phasor trajectories of the output signal of the MZ modulator calculated for the case when the modulator’s ER was 16 dB. In this calculation, we assumed that the phase deviations given in the two arms of the MZ modulator were perfectly balanced. Thus, the degradation of the ER was caused by the imbalance of the power splitting ratio of these two arms. In Fig. 2(a), the open circles show the operating points when we set the power ratio of the mark and space levels (i.e., on-off ratio) for the ASK modulation to be 3.9 dB. When the ER is infinite, the trajectory of the output signal is on the real axis and the operation is chirp-free. On the other hand, when the ER has a finite value of 16 dB, the trajectory on the phasor diagram becomes elliptic as shown in Fig. 2(a). Therefore, the phase is rotated during the transition between mark and space, and this phase error results in the distortion on the D8PSK tributary. Figure 2(b) shows the angle of the induced phase error calculated as a function of the ER of the modulator. The result shows that an unrealistically high ER is needed to reduce the phase error to a few degrees.
To solve this problem, we used a novel phase-distortion-free ASK modulation by using a QPSK modulator, where one of the MZ modulators was used for the data modulation with zero bias, while the other MZ modulator was used for the dc offset, as shown in Fig. 3(a). By adjusting the phase between the signals in two arms of the modulator, we could cancel out the parasitic phase deviations. (For the adjustment of this phase, we used the differential phase monitor developed in [19,20].) Figure 3(b) shows the differential phasor of this phase-distortion-free ASK modulation obtained by applying a CW signal to the modulator. The constellations were split into three amplitude levels depending on the combinations of the adjacent symbols’ status (mark-mark, mark-space/space-mark, and space-space levels) [24]. The constellations were aligned linearly and the additional phase distortion induced during the ASK modulation was negligible.
Fig. 2. (a) Trajectories of the complex electric field, E, of a conventional chirp-free MZ modulator with a finite ER of 16 dB. Open circles indicate the operating points of mark and space when the on-off ratio was set to 3.9 dB. (b) Phase errors (i.e., angle between the mark and the space on the phasor diagram caused by the finite ER of the MZ modulator).
We set the on-off ratio of the ASK modulation to be 3.9 dB, at which the theoretical limits of the receiver sensitivities for D8PSK and ASK tributaries became identical (so that we could achieve the best BER performance [14]). The data signals were 13.5-Gb/s pseudo random bit sequence (PRBS) of length 215-1. These signals were applied to the modulators with different bit delays for decorrelation. Figures 4(a) and 4(b) show the measured differential phasor diagrams with and without the ASK modulation (i.e., 40.5-Gb/s D8PSK and 54-Gb/s 16ADPSK signals), respectively. Owing to the use of the proposed phase-distortion-free ASK modulation technique, the generated 16ADPSK signal had good symmetry and low phase distortion. (Note that the slight asymmetry was observed due to the limited bandwidth of the phasor monitor [20].) For the quantitative evaluation of this phase distortion, we measured the standard deviation of the phase noise (which represented the blurriness of the symbols in the phase direction on the constellation diagram) for the D8PSK and 16ADPSK signals [25]. These standard deviations were measured to be 3.5 and 4.1 degrees for the D8PSK and 16ADPSK signals, respectively. In this way, the increase in the phase deviation due to the ASK modulation was negligibly small. (Note that, from these values, the deviation of the excessive phase distortion induced by the ASK modulation was estimated to be 2.1 degree.) From this result, we concluded that the additional phase distortion caused by the proposed ASK modulation technique was negligible. The intensity eye diagram of the 16ADPSK signal shown in Fig. 4(b) was slightly blurred due the imperfection of the driving waveform (caused by the impedance mismatching in the driver amplifier and the modulator).
Fig. 4. (a) Differential phasor of 40.5-Gb/s D8PSK tributary. (b) 54-Gb/s 16ADPSK signal. (left) measured differential phasor. (right) intensity eye diagram
2.2 Receiver
For the detection of the 16ADPSK signal, we used a single DI made of a 3×3 coupler (120-degree hybrid) and two Faraday rotator mirrors (FRMs) (for the polarization-independent operation), as shown in Fig. 5 [18–20]. The delay time between two delay arms was set to be identical to a symbol period T of 74 ps. Using the output currents of three 30-GHz PIN photodiodes, I 1~I 3, we obtained the delayed self-homodyne signals with different phase retardations determined by the 3×3 coupler. We then detected these signals by using a 50-Gs/s digital storage oscilloscope and demodulated them by off-line processing.
We first obtained the I-Q components of the differential phasor by using the coordinate transformation from the tri-phase non-orthogonal system (i.e., I 1~I 3) to the orthogonal system [18–20]. We then estimated the phase delay of the DI by using the phase-delay estimation algorithm (which utilized the asymmetry of the differential phasor trajectory) and compensated the drift of the phase delay by rotating the measured phasor [18–20]. Thus, this differential phase demodulator was completely adjustment-free and its operation was wavelength- and polarization-independent. In addition, we could easily compensate for all the imperfections of the optical components such as the imbalances of phase retardations and splitting ratios of the 3×3 coupler during the coordinate transformation. Consequently, we could minimize the effect of the distortion occurred in the receiver. For the detection of the ASK tributary, a 12-GHz PIN-FET receiver was used. We then recovered the clock signal by utilizing the residual clock component in one of the output of the DI, and re-sampled both the detected differential phasor and the intensity signal by using the recovered clock signal as a time base.
For the D8PSK tributary, we applied a T/2-spaced feed-forward equalizer (FFE) with 30 complex-valued taps. Following this procedure, we applied a novel data-aided phase noise estimation (DAPNE) algorithm [21], which was obtained by modifying the multi-symbol phase estimation (MSPE) algorithm [22,23]. The phase noise in the differential phasor of the k-th symbol, Δϕk, consisted of the phase noises of two adjacent symbols in the original electric field, ϕk-ϕk -1. In DAPNE, we first estimated the differential phase noise Δϕk-1~Δϕk-N of the past N-symbols by using the past decision data, and then estimated ϕ k-1 by integrating Δϕ k-1~Δϕk-N. If the estimation of ϕ k-1 was accurate, we could reduce the differential phase noise on the k-th symbol just by shifting the differential phase of the k-th symbol by ϕ k-1. The advantage of using this algorithm is the fact that we can use the fixed-tap Wiener filter for the estimation of ϕ k-1 from Δϕ k-1~Δϕk-N [21]. Since this estimation is optimum in the mean-square criteria, the performance of the noise suppression is good, and the error propagation due to incorrect decisions hardly occurs. For the ASK tributary, we applied a conventional 140-tap, T/2-spaced FFE.
Fig. 6. Back-to-back sensitivities of 54-Gb/s and 108-Gb/s PDM 16ADPSK receivers. The dotted lines show the quantum-limited sensitivity.
To measure the BER of the 16ADPSK signal, we counted the bit errors of the D8PSK and ASK tributaries simultaneously. The data length was 5×105 samples. We repeated this measurement (~6 times) until we obtained the sufficient resolution of BER. The open circles and squares in Fig. 6(a) show the OSNR sensitivities for D8PSK and ASK tributaries, respectively, measured in back-to-back condition without using FFEs and DAPNE. This figure showed that the receiver sensitivity of the D8PSK tributary was close to the quantum limit, whereas there was a large penalty for the ASK tributary due to the blurred eye. However, by applying FFEs and DAPNE, the receiver sensitivities were significantly improved as shown by the filled circles and squares. The total BER of 54-Gb/s 16ADPSK, which is the ratio of the number of bit errors occurred in all tributaries to the total number of the transmitted bits, is shown by the open circles in Fig. 6(b). The required OSNR for BER=10-3 was measured to be 18.6 dB. It was only 1.6 dB away from the quantum limit, which was calculated by assuming the use of a conventional DI without MSPE. This sensitivity is almost identical to that of the 16QAM signal achieved by using a coherent receiver in a recent experiment [26], although the difference between the theoretical limits of non-coherent 16ADPSK and coherent 16QAM is ~4 dB [12,14]. We also measured the receiver sensitivity for the 108-Gb/s PDM 16ADPSK signal by demultiplexing the PDM signal using a manually adjusted polarization controller. The degradation in the receiver sensitivity caused by the PDM was measured to be 3.2 dB. The sensitivity difference between the PDM channels was less than 0.4 dB.
3. 108-Gb/s PDM 16ADPSK transmission experiment on 25-GHz WDM grid
By using the PDM 16ADPSK transmitter and receiver described above, we examined the feasibility of transmitting the 108-Gb/s signal on a 25-GHz WDM grid without using the coherent detection technology. Figure 7 shows the experimental setup. To evaluate the effect of the tight optical bandpass filtering in this experiment, we first measured the BER performance of a single channel PDM 16ADPSK signal traversed through two 25/50-GHz interleavers (ILs). The open and filled circles in Fig. 8 show the BER performances measured with and without using these ILs, respectively, for (a) D8PSK tributary, (b) ASK tributary, and (c) 16ADPSK signal. The results showed that the penalty caused by the stringent filtering was very small (<1 dB). This excellent tolerance was attributed to the compact spectrum of the 16ADPSK signal. In fact, the measured optical spectrum of the 16ADPSK signal was almost identical to that of the DQPSK signal with the same baud rate, and the shape of the main lobe of the 16ADPSK signal’s spectrum was maintained even after passing through the IL as shown by the curves a and b in Fig. 9.
We also performed a WDM transmission experiment by adding two neighboring channels obtained from a different PDM 16ADPSK transmitter (indicated by ‘Crosstalk’ in Fig. 7). We used two ECLs operating at 193.575 and 193.625 THz, and prepared 16ADPSK signals by using a different pattern generator (pattern length: 215-1). After PDM, we added these channels through the odd port of the IL. The curve c in Fig. 9 shows the spectrum of all three WDM signals. Since we used the serial phase modulation technique with multi-level electronic signals for the generation of the D8PSK portion of the crosstalk channels, the bandwidth of the crosstalk channels became slightly wider than that of the test channel. At the receiver, we filtered out the test channel at 193.600 THz by using the IL and the 0.25-nm Gaussian-profile OBPF after the preamplifier.
The filled squares in Fig. 8 show the BER sensitivities measured in the back-to-back condition. The OSNR penalty for the D8PSK tributary was not severe, while the ASK tributary suffered a large penalty due to the crosstalk. The solid and dotted curves in Fig. 10 show the optical spectra of the received signal measured after the IL at the receiver with and without neighboring channels, respectively. The results showed that a significant amount of the neighboring channels remained after the IL. (To mitigate the crosstalk-induced penalty, it might be necessary to improve the roll-off characteristics of the IL filter by using the cascaded ILs as in [3].) Nevertheless, the receiver sensitivity of the PDM 16ADPSK signal was excellent since the ASK tributary carried only one-fourth of the information. The required OSNR for BER=1.8×10-3 (i.e., the threshold for enhanced forward-error correction (EFEC) code) was measured to be 24.4 dB. We transmitted 3-channel PDM 16ADPSK signals over 80-km long standard single-mode fiber (SSMF). The launched power was 0 dBm per channel. At the receiver end, we used Erbium-doped fiber amplifiers (EDFAs) with a dispersion compensation fiber (DCF) module. Figure 11 shows the differential phasor diagrams for X-and Y- polarizations measured at OSNR=30 dB. Although the symbols were slightly blurred due to the stringent optical bandpass filtering, the constellations were still clear. The open squares in Fig. 8 show the measured BER as a function of the OSNR after the 80-km long transmission. The OSNR was adjusted by changing the input power to the EDFA after the DCF module (as indicated by ‘X’ in Fig. 7) by using an attenuator. (Note that the OSNR of the received signal without this attenuation was >34 dB.) As shown in Fig. 8(c), the OSNR penalty caused by the transmission was not significant. As a result, a sufficient OSNR margin could be secured even after the transmission. In this way, the SE of 4.0 bit/s/Hz could be achieved by using non-coherent receivers. To our knowledge, this result represents the first demonstration of the 100-Gb/s transmission on a 25-GHz WDM grid without using the coherent detection technique.
Fig. 9. Measured optical spectra of 108-Gb/s PDM 16ADPSK signals (resolution: 1.2 GHz). (a) Single channel without filtering, (b) single channel filtered by IL, and (c) 3 WDM channels.
Fig. 10. Optical spectra of the received signal measured after passing through the 0.25-nm OBPF. (a) 3-channel WDM transmission. (b) Single channel transmission.
Fig. 11. Differential constellations of 108-Gb/s PDM 16ADPSK signal measured after the transmission of 80-km long SSMF.
4. Summary
We have demonstrated the 54-Gb/s (and 108-Gb/s with PDM) 16ADPSK transmission system based on direct-detection receivers. At the transmitter, we generated the 16ADPSK signal by using the D8PSK and binary ASK modulators. To suppress the phase distortion caused by the ASK modulation, we used a novel phase-distortion-free ASK modulation, in which the chirp and the ER could be controlled independently. For the detection of 16ADPSK signal, we employed the adjustment-free M-ary DPSK receiver composed of a 3×3 coupler. In addition, the data-aided phase-noise estimation algorithm (DAPNE) (which was an extension of MSPE algorithm) was used to improve the receiver sensitivity. The required OSNR (0.1-nm resolution) to achieve BER=10-3 was measured to be 18.6 dB, which was close to the quantum-limited sensitivity within 1.6 dB and almost identical to that of the recently reported coherent 16QAM receiver. We attributed this outstanding performance to the high quality of the 16ADPSK modulation technique and the enhanced receiver sensitivity achieved by using DAPNE. By using the proposed 16ADPSK transmitter and receiver, we experimentally evaluated the feasibility of transmitting 100-Gb/s WDM signals on a 25-GHz grid without using the coherent detection technology. The results showed that the penalty of the tight filtering was small due to the compact spectrum of the 16ADPSK signal. We also demonstrated that a sufficient OSNR margin could be secured even after the transmission of 80-km long SSMF. To our knowledge, this result represents the first demonstration of the SE of 4.0 bit/s/Hz achieved without using the coherent detection technique.
Acknowledgement
We thank Prof. Hoon Kim of National University of Singapore for valuable discussions. This work was supported by the IT R&D program of MKE/IITA, [2008-F017-02, 100Gbps Ethernet and optical transmission technology development] and the Brain Korea 21 Project, KAIST.
References and links
1. D. van den Borne, S. L. Jansen, E. Gottwald, P. M. Krummrich, G. D. Khoe, and H. de Waardt, “1.6-b/s/Hz spectrally efficient transmission over 1700 km of SSMF using 40×85.6-Gb/s POLMUX-RZ-DQPSK,” J. Lightwave Technol. 25(1), 222–232 (2007). [CrossRef]
2. H. Masuda, A. Sano, T. Kobayashi, E. Yoshida, Y. Miyamoto, Y. Hibino, K. Hagimoto, T. Yamada, T. Furuta, and H. Fukuyama, “20.4-Tb/s (204×111 Gb/s) transmission over 240 km using bandwidth-maximized hybrid Raman/EDFAs,” in Optical Fiber Communication Conference and Exposition and The National Fiber Optic Engineers Conference (OFC/NFOEC 2007) (Optical Society of America, Washington, DC, 2007), paper PDP20. [PubMed]
3. A. H. Gnauck, G. Charlet, P. Tran, P. J. Winzer, C. R. Doerr, J. C. Centanni, E. C. Burrows, T. Kawanishi, T. Sakamoto, and K. Higuma, “25.6-Tb/s C+L-band transmission of polarization-multiplexed RZ-DQPSK signals,” in Optical Fiber Communication Conference and Exposition and The National Fiber Optic Engineers Conference (OFC/NFOEC 2007) (Optical Society of America, Washington, DC, 2007), paper PDP19. [PubMed]
4. T. Ito, S. Fujita, E. L. T. de Gabory, S. Shioiri, K. Fukuchi, Y. Aono, Y. Yano, and M. Nishino, “Precise analysis of transmission impairments of Pol-Mux 110Gb/s RZ-DQPSK with automatic Pol-Dmux using straight 2,000-km SMF line,” in Proceedings of European Conference of Optical Communication (ECOC 2008), paper We.1.E.6.
5. K. Kikuchi, “Phase-diversity homodyne detection of multilevel optical modulation with digital carrier phase estimation,” IEEE J. Sel. Top. Quantum Electron. 12(4), 563–570 (2006). [CrossRef]
6. M. Nakazawa, M. Yoshida, K. Kasai, and J. Hongou, “20 Msymbol/s, 64 and 128 QAM coherent optical transmission over 525 km using heterodyne detection with frequency-stabilized laser,” Electron. Lett. 42(12), 710–712 (2006). [CrossRef]
7. K. Sekine, N. Kikuchi, S. Sasaki, S. Hayase, C. Hasegawa, and T. Sugawara, “40Gbit/s, 16-ary (4bit/symbol) optical modulation/demodulation scheme,” Electron. Lett. 41(7), 430–432 (2005). [CrossRef]
8. M. Serbay, T. Tokle, P. Jeppesen, and W. Rosenkranz, “42.8 Gbit/s, 4 bits per symbol 16-ary inverse-RZ-QASK-DQPSK transmission experiment without Polmux,” in Optical Fiber Communication Conference and Exposition and The National Fiber Optic Engineers Conference on CD-ROM (OFC/NFOEC 2006) (Optical Society of America, Washington, DC, 2006), paper OThL2.
9. N. Kikuchi, K. Mandai, K. Sekine, and S. Sasaki, “Incoherent 32-level optical multilevel signaling technologies,” J. Lightwave Technol. 26(1), 150–157 (2008). [CrossRef]
10. T. Tokle, M. Serbay, J. B. Jensen, W. Rosenkranz, and P. Jeppesen, “Advanced modulation formats for transmission systems,” in Optical Fiber Communication Conference and Exposition and The National Fiber Optic Engineers Conference (OFC/NFOEC 2008) (Optical Society of America, Washington, DC, 2008), paper OMI1. [CrossRef] [PubMed]
11. N. Kikuchi and S. Sasaki, “Sensitivity improvement of incoherent multilevel (30-Gbit/s 8QAM and 40-Gbit/s 16QAM) signaling with non-Euclidean metric and MSPE (multi symbol phase estimation),” in Optical Fiber Communication Conference and Exposition and The National Fiber Optic Engineers Conference on CD-ROM (OFC/NFOEC 2009) (Optical Society of America, Washington, DC, 2009), paper OWG1. [PubMed]
12. J. G. Proakis, Digital Communications, 4th ed., (McGraw-Hill, New York, 2001).
13. J. M. Kahn and K.-P. Ho, “Spectral efficiency limits and modulation/detection techniques for DWDM systems,” IEEE J. Sel. Top. Quantum Electron. 10(2), 259–272 (2004). [CrossRef]
14. Y. Han and G. Li, “Theoretical sensitivity of direct-detection multilevel modulation format for high spectral efficiency optical communications,” IEEE J. Sel. Top. Quantum Electron. 12(4), 571–580 (2006). [CrossRef]
15. M. Serbay, C. Wree, and W. Rosenkranz, “Experimental investigation of RZ-8DPSK at 3×10.7Gb/s,” in IEEE LEOS annual meeting, paper WE3 , pp. 483–484, Sydney, Australia (2005). [CrossRef]
16. J. B. Jensen, T. Tokle, C. Peucheret, and P. Jeppesen, “Transmission of multilevel 60 Gbit/s polarization multiplexed RZ-D8PSK using only 10 Gbit/s equipment,” in Optical Fiber Communication Conference and Exposition and The National Fiber Optic Engineers Conference (OFC/NFOEC 2007) (Optical Society of America, Washington, DC, 2007), paper OWM1. [CrossRef] [PubMed]
17. Y. Han and G. Li, “Sensitivity limits and degradations in OD8PSK,” IEEE Photon. Technol. Lett. 17(3), 720–722 (2005). [CrossRef]
18. Y. Takushima, H. Y. Choi, and Y. C. Chung, “Adjustment-free DxPSK receiver based on single delay interferometer using 120-degree optical hybrid,” in Optical Fiber Communication Conference and Exposition and The National Fiber Optic Engineers Conference on CD-ROM (OFC/NFOEC 2009) (Optical Society of America, Washington, DC, 2009), paper OMM2. [PubMed]
19. Y. Takushima, H. Y. Choi, and Y. C. Chung, “Plug-and-play phasor monitor for DxPSK signals based on single delay-interferometer using a 3×3 optical coupler,” in Optical Fiber Communication Conference and Exposition and The National Fiber Optic Engineers Conference on CD-ROM (OFC/NFOEC 2008) (Optical Society of America, Washington, DC, 2008), paper OThW4. [PubMed]
20. Y. Takushima, H. Y. Choi, and Y. C. Chung, “Measurement of differential phasor diagram of multilevel DPSK signals by using an adjustment-free delay interferometer composed of a 3×3 optical coupler,” J. Lightwave Technol. 27(6), 718–730 (2009). [CrossRef]
21. Y. Takushima, H. Y. Choi, and Y. C. Chung, “Enhanced sensitivity of DxPSK receiver by using data-aided phase noise estimation algorithm,” in Opto-Electronics and Communications Conference (OECC 2009) , paper WV3.
22. H. Leib, “Data-aided noncoherent demodulation of DPSK,” IEEE Trans. Commun. 43(2), 722–725 (1995). [CrossRef]
23. X. Liu, “Generalized data-aided multi-symbol phase estimation for improving receiver sensitivity in direct-detection optical m-ary DPSK,” Opt. Express 15(6), 2927–2939 (2007). [CrossRef] [PubMed]
24. N. Kikuchi, “Inter-Symbol Interference (ISI) Suppression Technique for Optical Binary and Multilevel Signal Generation,” J. Lightwave Technol. 25(8), 2060–2068 (2007). [CrossRef]
25. Y. Takushima, H. Y. Choi, and Y. C. Chung, “Quality monitoring of DxPSK signals by using differential phasor diagram,” to appear in IEEE Photon. Technol. Lett.
26. P. J. Winzer and A. H. Gnauck, “112-Gb/s polarization-multiplexed 16-QAM on a 25-GHz WDM grid,” in Proceedings of European Conference of Optical Communication (ECOC 2008), paper Th.3.E.5.
