Open Access
Add to BibSonomyAbstract
We propose a new optical coherent detection scheme with a “two-tone” local light, which is also in principle insensitive to the laser phase fluctuation and is assisted by the digital signal processing technique for radio-over-fiber (RoF) systems. The main feature of our proposal is that the frequency separation of the two-tone local light is different from that of the RoF signal, which is called “offset-frequency-spaced” in this paper. First, we explain the principle of our new proposal and experimentally demonstrate the data recovery. Then, the influence of the frequency detuning between photo-detected modulated and unmodulated signals is discussed. Moreover, the transmission performance with an error vector magnitude (EVM) is evaluated for the optical coherent detection of a 10-Gbaud quadrature-phase-shift-keying RoF signal after a 20-km-long standard single-mode fiber transmission.
© 2015 Optical Society of America
1. Introduction
Owing to the rapid and widespread commercialization of mobile and Wi-Fi devices, radio services have exponentially grown in the global market. The demand for radio access has been further expanding due to the high-speed radio terminals, such as smartphones, tablets, mobile Wi-Fi routers. Annual global mobile data traffic in 2014 grew by about 1.7 times, compared with that in 2013, and more than 80% of them were mobile cloud traffic [1]. In addition, 45% of annual global mobile traffic in 2014 was offloaded onto the fixed network and the ratio is estimated to exceed 50% in a few years. Moreover, on an average, a smart device generated 22 times more traffic than a non-smart device in 2014. According to these trends of mobile traffic, the current deployment of mobile services tends towards higher bit rates with higher carrier frequencies for supporting future broadband access. From these facts and trends, there is no doubt that optical access networks for the purpose of offload will take on an important role. From the physical viewpoint, physically seamless connection between radio and optical links is highly desired, leading to the transparency of the modulation type of radio wave in the optical link. To realize such seamless connectivity on a physical level, many radio-over-fiber (RoF) technologies have been studied [2–4].
Recently, digital-signal-processing-assisted (DSP-assisted) optical coherent detection techniques and its applications have been intensively studied [5–10]. Thanks to some advanced DSP techniques, more accurate compensation of various signal distortions, such as chromatic dispersion and polarization mode dispersion, has become possible, comparing with the compensation by means of only analog signal processing. However, from the theoretical viewpoint, the system is insensitive in principle to the laser phase fluctuation, such as laser phase noise and laser frequency offset. We have proposed a laser-phase-fluctuation-insensitive optical coherent detection scheme assisted by a DSP for RoF systems [11]. In this technique, modulated and unmodulated components of RoF signal are independently mixed with a “two-tone” local light, and then two photo-detected signals are digitally processed. In this case, the phase fluctuation in both signals must have the same value and the value can be easily extracted from the unmodulated signal. If the phase fluctuation in the modulated signal is simply subtracted with that extracted from the unmodulated signal, data is recovered without any laser phase drift. Thus, the cancellation of phase fluctuation of light sources is performed.
In this paper, as a modification of our previous proposal, we propose a new laser-phase-fluctuation-insensitive optical coherent detection, which uses “two-tone” local light and DSP technique for a RoF signal. The main feature of our proposal is that the frequency separation of the two-tone local light is different from that of the RoF signal, which is called “offset-frequency-spaced” in this paper. The major difference from the previous scheme is that both the modulated and unmodulated components of RoF signal are “simultaneously” coherent-detected with a single optical coherent detector [12]. Therefore, the optical configuration is expected to be simpler than that in the previous scheme. At first, the principle of our newly proposed scheme is explained in detail. Next, the data recovery from RoF signal with 10-Gbaud quadrature-phase-shift-keying (QPSK) payload data is experimentally demonstrated for the proof-of-concept. Then, the effect of the frequency detuning between photo-detected modulated and unmodulated signals is discussed. Moreover, the transmission characteristics of the 10-Gbaud QPSK RoF signal is also evaluated with an error vector magnitude (EVM) and a bit error rate (BER).
2. Principle
Figure 1 shows the schematic of newly proposed laser-phase-fluctuation-insensitive optical coherent scheme for a RoF signal. In this scheme, it is assumed that a carrier frequency of radio-frequency-band (RF-band) or intermediate-frequency-band (IF-band) signal, frs, a separation frequency of local two-tone light, frl, and a symbol rate of modulated component, B, are given. Therefore, it can be assumed that a frequency offset between frs and frl, Δf [ = frl-frs], is constant. In addition, the transmission direction is assumed to be an uplink.
A received RoF signal shown in Fig. 1(a) has two components, which are separated by frs at the center of fcs with the same phase noise, ϕcs(t): One is an unmodulated carrier at fcs + frs/2, the other is a signal modulated with an uplink data of radio signal, θ(t), at fcs-frs/2. Here, note that the unmodulated carrier and modulated signal components are phase-locked. The two-tone local light shown in Fig. 1(b) has also two components, which are separated by frl at the center of fcl with the same phase noise, ϕcl(t): Both of them are unmodulated sinusoidal carriers at fcl + frl/2 and fcl-frl/2 as local references. For the purpose of optical coherent detection, fcl-frl/2 is set to be close to fcs-frs/2. In other words, fcl-fcs-Δf/2 is set to be close to zero. In addition, the frequency offset, Δf [ = frl-frs], is assumed to be set to be larger than a symbol rate, B, so that the modulated signal and the unmodulated carrier can be easily separated in a following frequency de-multiplexer. All optical frequency components are input into one optical coherent detector to generate beating signals as shown in Fig. 1(c), where the higher-frequency beating signals are not taken into account because they do not contribute to the following processing. Then, the modulated signal and the unmodulated carrier appear at fcl-fcs-Δf/2 and fcl-fcs + Δf/2, and are de-multiplexed as shown in Figs. 1(d) and 1(e)-1, respectively. Here, a differential phase noise between lasers in the transmitter and the receiver, ϕcl(t)-ϕcs(t), must be the same value in both the modulated signal and the unmodulated carrier components. The differential phase noise and the laser frequency offset, fcl-fcs, can be easily extracted from the unmodulated carrier component down-shifted by Δf as shown in Figs. 1(e)-2. In the same manner as that in Ref [11], the extracted information of the differential phase noise and the laser frequency offset between two lasers can be used to cancel out those in the modulated signal for the simple phase subtraction. Finally, only data, θ(t), should be obtained as shown in Fig. 1(f), which is insensitive to both the differential phase noise and the laser frequency offset between two lasers. Thus, the cancellation of phase fluctuation of light sources is performed. In this scheme, the optical configuration can be simpler than that in the previous scheme because the number of optical coherent detector becomes small.
3. Experimental demonstrations and discussion
3.1 Experimental setup
Figure 2 shows the experimental setup for demonstrating the cancellation of the laser phase fluctuation with our proposed scheme. In a transmitter, which emulates a RoF signal generator, a 10-Gbaud QPSK 25-GHz-band RoF signal was generated as a test signal in the similar manner used in Ref [11]. Therefore, the frequency separation of frequency components of RoF signal, frs, was 25 GHz. The linewidth of transmitting laser was more than 10 MHz (coherence control on mode, Anritsu MG9541A) [11]. As an alternative technique, a reversely modulated optical single sideband scheme will be also applicable for emulating the RoF signal generator, which can stably generate a RoF signal with the high modulation power efficiency [10]. The generated RoF signal was transmitted over a 20-km-long standard single-mode fiber (SMF). In a receiver, a received RoF signal was optically amplified with an Erbium-doped fiber amplifier (EDFA) and input into an optical coherent detector via a polarization controller (PC). An optical attenuator (ATT) just before the receiver was used to control the received optical power. At the same time, an offset-frequency-spaced two-tone local light from a two-tone light source was also put into the optical coherent detector via another EDFA and another PC. The two-tone local light was generated with a double sideband with the suppressed carrier (DSB-SC) modulation technique, which was driven by a local frequency of 18 GHz, frl/2, from an electrical local oscillator. As an initial setting, the frequency separation of the two-tone local light, frl, was 36 GHz, which has an 11-GHz detuning frequency from that of the RoF signal. The linewidth of local laser was less than 100 kHz (coherence control off mode, Agilent HP81682A) [11]. To maximize the beating signal, all state-of-polarization of optical signals were manually controlled by using PCs. A pair of photo-detected in-phase and quadrature-phase (I/Q) signals was directly captured with a real-time digital storage oscilloscope (DSO) and they were digitally processed. In the offline processing, at first, the frequency de-multiplexing was performed. Next, the differential phase noise and the laser frequency offset were extracted from the unmodulated carrier. Then, those in the modulated signal were cancelled out with the extracted ones by the simple phase subtraction. It should be noted that these cancellation process does not include any frequency and phase estimation function. After that, a symbol recovery was performed and an error vector magnitude (EVM) was estimated from the recovered symbol.
3.2 Experimental results
To confirm our concept, some electrical spectra were calculated with the captured signal after the photo-detection by means of the fast Fourier transform (FFT) algorithm. Figure 3 shows the calculated electrical spectra for the back-to-back case. It is noted that the FFT was used only for the spectrum calculation and it was never used in the cancellation process. Figure 3(a) shows the generated beating signal just after the optical coherent detection. As expected, it had the modulated signal and the unmodulated carrier and corresponded to Fig. 1(c). As shown in Figs. 3(b) and 3(c), the modulated signal and unmodulated carrier were successfully separated, which correspond to Figs. 1(d) and 1(e)-1, respectively. In Fig. 3(b), an undesired spurious frequency component at around 11 GHz, which was not originated from the unmodulated carrier component, was observed. However, this component will not seriously affect the system performance because a digital low pass filter used in the demodulation process can easily remove it. Because fcl and fcs are practically unknown values, it is noted that it is hard to exactly make fcl-fcs-Δf/2 be zero. Therefore, as shown in Fig. 3(c), the unmodulated carrier component was extracted by using a digital filter with the bandwidth of 1 GHz at the center frequency of Δf [ = 11GHz], covering the fluctuation range of laser frequency offset, fcl-fcs. To match the laser frequency offset to that of the modulated signal, fcl-fcs-Δf/2, the frequency of the unmodulated carrier was down-shifted by Δf [ = 11GHz], as shown in Fig. 3(d), which corresponds to Fig. 1(e)-2. In this case, as expected, fcl-fcs-Δf/2 was not zero due to the unpredictable fluctuation of laser frequency offset. It is noted in theory that this value must be the same value as that of the center frequency of modulated signal shown in Fig. 3(b). Finally, by subtracting the phase extracted from unmodulated carrier shown in Fig. 3(d) from that of the separated modulated signal shown in Fig. 3(b), only the data was obtained. The spectrum is shown in Fig. 3(e), which corresponds to Fig. 1(f).
Fig. 3 Calculated electrical spectra: (a) generated beating signal, (b) separated modulated signal, (c) separated unmodulated carrier, (d) separated unmodulated carrier down-shifted by Δf, and (e) recovered data after laser-phase-fluctuation cancellation.
Figure 4 plots the measured symbol constellations of modulated signal before and after the cancelation of the laser phase fluctuation, which correspond to the signals shown in Figs. 1(c) and 1(e). The measurement was performed for the back-to-back case and for no optical attenuation (ATT = 0 dB) case. As shown in Fig. 4(a), it can be seen that the symbol constellation is fluctuated causing huge phase drift due to the laser phase fluctuation. After the laser-phase-fluctuation cancelation, we can see from Fig. 4(b) that the symbol constellation is successfully carried out and is clearly separated. This shows that the laser phase fluctuation was successfully canceled out. In this case, the EVM was about 9.3%rms. Here, in our process of laser-phase-fluctuation cancellation between Figs. 4(a) and 4(b), no filter for eliminating additive noises was applied to the QPSK-modulated signal component, where the additive noises were independent of our focusing laser phase fluctuation. Therefore, in general, it is considered from the theoretical point of view that the signal-to-noise power ratios before and after the cancellation should be the same.
Fig. 4 Measured symbol constellations of modulated signal: (a) before and (b) after laser-phase-fluctuation cancelation.
We also investigated the detuning frequency dependency of measured EVM. Figure 5 shows the result. The measurement was also performed for the back-to-back case and for no optical attenuation (ATT = 0 dB). Here, a low-pass Bessel filter with the 3-dB bandwidth of 7.5 GHz was used in our digital signal processing. In the frequency range of more than 9 GHz, the EVMs of around 10%rms are stably obtained. In the frequency range of less than 8 GHz, on the other hand, the EVMs become worse and are fluctuated. This result suggests that it will be better that the detuning frequency should be set to be more than 9 GHz for 10-Gbuad QPSK RoF signal in our experimental setup.
Moreover, we investigated the transmission characteristics before and after 20-km-long SMF transmission. Figures 6(a)-6(d) show the measured EVM as a function of the received optical power for the detuning frequencies of 9, 10, 11, and 12 GHz, respectively. From Fig. 6, we can see that the EVM becomes better as the received optical power becomes higher for all cases. In addition, the EVM of less than 12.4%rms, which corresponds to the theoretical BER of less than 10−15, was achieved for all cases. Here, the EVMs after the 20-km-long SMF transmission become slightly worse than those for the back-to-back case, which is probably caused by the chromatic fiber dispersion effect because no compensation of the fiber dispersion was carried out in this experiment.
Fig. 6 Measured EVMs versus received optical power after 20-km-long SMF transmission and back-to-back: (a) 9-GHz detuning, (b) 10-GHz detuning, (c) 11-GHz detuning, and (d) 12-GHz-detuning.
To verify that the symbol decision worked well, we measured BERs before and after 20-km-long SMF transmission. The measured BER for the detuning frequency of 11 GHz is shown in Fig. 7. From Fig. 7, it can be seen that the measured BERs are better as the received optical power becomes higher in the received optical power range of up to around −37 dBm. And no error occurred in the received optical power range of more than −37 dBm, since the effective number of bits was about 88,000 for measuring the BER due to the limited memory size of DSO. Although the BER after the 20-km-long SMF transmission was slightly worse than that for the back-to-back case, no serious degradation due to the fiber transmission was observed. This shows that the symbol decision was correctly performed and the fiber-optic transmission up to 20 km was successfully performed.
3.3 discussion
In the above-mentioned experiments, we digitally implemented a band elimination filter (BEF) in order to ideally remove the strong pilot-tone component from the signal component in the de-multiplexing process. Figure 8 shows the actual schematic of de-multiplexing process for the pilot-tone elimination. In the de-multiplexing, a captured signal gets copied. One is used to extract a pilot tone. The extracted pilot tone is used to eliminate the same amount of pilot tone in the other component with a simple complex amplitude subtraction. With this configuration, a function of BEF is equivalently achieved, enabling de-multiplexing operation. Thus, the demultiplexing is performed. In this discussion, we investigate the pilot-tone cleaning effect by using the BEF.
To verify the effect of BEF, as a reference, we measured constellations for the worst case, in which the pilot tone was not eliminated from the signal component. To eliminate the fiber dispersion effect, this measurement was performed for the back-to-back case. Figure 9 shows the measured symbol constellations of modulated signal before and after cancelling the laser phase fluctuation without and with the BEF. In this measurement, the detuning frequency, Δf, was 11 GHz. The experiment was carried out under the same condition for the measurement of constellation shown in Fig. 4 as just described. As shown in Figs. 9(a) and 9(c), it can be seen that the symbol constellation is fluctuated due to the laser phase fluctuation. After the laser-phase-fluctuation cancelation, we can see that the symbol constellation is successfully compensated as shown in Figs. 9(b) and 9(d). Here, as expected, it can be verified that the constellations shown in Figs. 9(c) and 9(d) are almost the same constellations shown in Figs. 4(a) and 4(b), respectively. However, the residual carrier, which corresponds to the unmodulated carrier component, obviously affects the constellation without the BEF as shown in Fig. 9(b). It is considered that the residual carrier component appeared with the bias at the complex amplitude of QPSK constellation points. Therefore, the constellation with the BEF (Fig. 9(d)) is more clearly separated than that without the BEF (Fig. 9(b)).
Fig. 9 Measured constellations: (a) before cancellation without BEF, (b) after cancellation without BEF, (c) before cancellation with BER, and (d) after cancellation with BEF.
We also investigated the estimated EVMs as a function of received optical power with and without the BEF, as shown in Fig. 10. Here, the BER for the case of using the BEF is the same one as that in the back-to-back case of Fig. 6(c). As can be seen, for the case of no BEF, no improvement of EVM was observed in the received optical power range of more than −32 dBm. Thus, it can be seen that the EVM was improved by more than 17%rms in the received optical power range of more than −30 dBm. In the same condition, we also measured the BERs as a function of received optical power with and without the BEF, as shown in Fig. 11. Also in this case, as mentioned in Sec. 3.2, it was confirmed that the symbol decision was correctly performed. From Fig. 11, it can be seen that the received power at the BER of 10−6 was improved by more than 4 dB. These results clearly show that the pilot-tone cleaning effect by using the BEF is obtained and important for enhancing the system performance.
4. Conclusion
We have proposed the new laser-phase-fluctuation-insensitive optical coherent detection scheme with offset-frequency-spaced two-tone local light and DSP technique for a RoF signal. In the experiment for proof-of-concept, the principle of our scheme was verified and the data recovery from RoF signal with 10-Gbaud QPSK payload data without any serious laser phase fluctuation was successfully demonstrated. In the investigation of the influence of the frequency detuning between photo-detected modulated and unmodulated signals, the detuning frequency should be set to be more than 9 GHz for 10-Gbuad QPSK RoF signal in our experimental setup. Moreover, we demonstrated the 20-km-long SMF transmission of a 10-Gbaud QPSK RoF signal after a. As a result, the EVM of less than 12.4%rms, which corresponds to the theoretical BER of less than 10−15, was achieved and it was also proved that the symbol decision was correctly performed. Furthermore, the pilot-tone cleaning effect by using a BEF was verified. From the above, it can be concluded that our proposed optical coherent detection scheme is feasible and useful for RoF systems.
Acknowledgment
Authors wish to thank Dr. G.-W. Lu and Dr. A. Kanno of NICT for lending their optical modulators in the experiment. Authors also thank Dr. Yamamoto of NICT for his encouragement.
References and links
1. Cisco White Paper, “Cisco visual networking index: Global mobile data traffic forecast update, 2014-2019,” (Cisco, 2015).
2. M. Weiss, A. Stöhr, F. Lecoche, and B. Charbonnier, “27 Gbit/s photonic wireless 60 GHz transmission system using 16-QAM OFDM,” in 2009 IEEE International Topical Meeting on Microwave Photonics (MWP, 2009) pp. 1–3.
3. X. Pang, A. Caballero, A. Dogadaev, V. Arlunno, R. Borkowski, J. S. Pedersen, L. Deng, F. Karinou, F. Roubeau, D. Zibar, X. Yu, and I. T. Monroy, “100 Gbit/s hybrid optical fiber-wireless link in the W-band (75-110 GHz),” Opt. Express 19(25), 24944–24949 (2011). [CrossRef] [PubMed]
4. A. Kanno, P. Tien Dat, T. Kuri, I. Hosako, T. Kawanishi, Y. Yasumura, Y. Yoshida, and K. Kitayama, “Quadrature-phase-shift-keying radio-over-fiber transmission for coherent optical and radio seamless networks,” IEICE Trans. Electron. E96-C(2), 156–162 (2013). [CrossRef]
5. D.-S. Ly-Gagnon, K. Katoh, and K. Kikuchi, “Unrepeatered optical transmission of 20 Gbit/s quadrature phase-shift keying signals over 210 km using homodyne phase-diversity receiver and digital signal processing,” IEEE Electron. Lett. 41(4), 206–207 (2005). [CrossRef]
6. E. Yamazaki, M. Tomizawa, and Y. Miyamoto, “100-Gb/s optical transport network and beyond employing digital signal processing,” IEEE Commun. Mag. 50(2), s43–s49 (2012). [CrossRef]
7. S. L. Jansen, I. Morita, T. C. W. Schenk, N. Takeda, and H. Tanaka, “Coherent optical 25.8-Gb/s OFDM transmission over 4,160-km SSMF,” J. Lightwave Technol. 26(1), 6–15 (2008). [CrossRef]
8. F. Li, Z. Cao, X. Li, Z. Dong, and L. Chen, “Fiber-wireless transmission system of PDM-MIMO-OFDM at 100 GHz frequency,” J. Lightwave Technol. 31(14), 2394–2399 (2013). [CrossRef]
9. Z. Cao, F. Li, Y. Liu, J. Yu, Q. Wang, C. W. Oh, Y. Jiao, N. C. Tran, H. P. A. van den Boom, E. Tangdiongga, and A. M. J. Koonen, “61.3-Gbps hybrid fiber-wireless in-home network enabled by optical heterodyne and polarization multiplexing,” J. Lightwave Technol. 32(19), 3227–3233 (2014). [CrossRef]
10. Z. Cao, J. Yu, F. Li, L. Chen, Q. Shu, Q. Tang, and L. Chen, “Energy efficient and transparent platform for optical wireless networks based on reverse modulation,” IEEE J. Sel. Areas Comm. 31(12), 804–814 (2013). [CrossRef]
11. T. Kuri, T. Sakamoto, G.-W. Lu, and T. Kawanishi, “Laser-phase-fluctuation-insensitive optical coherent detection scheme for radio-over-fiber system,” J. Lightwave Technol. 32(20), 3803–3809 (2014). [CrossRef]
12. T. Kuri, T. Sakamoto, and T. Kawanishi, “Laser-phase-fluctuation-insensitive offset-frequency-spaced two-tone optical coherent detection scheme with digital-signal-processing technique for radio-over-fiber systems,” Proc. SPIE 9387, 93870D (2015). [CrossRef]
