Open Access
Add to BibSonomyAbstract
Orthogonal frequency division multiple access-based passive optical network (OFDMA-PON) is considered as a strong candidate for next-generation optical access network. In intensity modulation/direct detection system, OFDMA-PON downlink transmission is relatively stable, but critical issues exist in uplink multiple access. Because of different optical paths, optical beat interference (OBI) and timing offset effect are generated, which seriously disturb signal detection. We propose optical pulse division multiplexing-based OBI reduction. By considering both the spectrum broadening effect and the time domain near orthogonality, OBI could be reduced. We demonstrate that the spectral efficiency can be improved from 0.37 to 3.8 bit/s/Hz in 1-GHz signal bandwidth.
© 2016 Optical Society of America
1. Introduction
Data-traffic requirement in both wireless and optical access networks has been explosively increased. This trend will continue to accommodate various applications such as Internet-of-Things (IoT), which is considered as one of the main applications in the fifth-generation (5G) wireless communication [1]. To support a high data traffic within a limited network resource, a spectrum-efficient transmission technique is required both in wireless and optical communications.
Orthogonal frequency division multiplexing (OFDM) is a highly spectrum-efficient parallel transmission technique that uses orthogonally overlapped subcarriers. Because each subcarrier is independent, the channel effect could easily be equalized using a simple single-tap equalizer. Further, its capacity could be maximized by employing adaptive modulation. Moreover, it is robust against multipath fading channel by inserting cyclic prefix. With these advantageous properties, OFDM became a standard of Long Term Evolution-Advanced (LTE-A) in wireless communications [2]. Furthermore, the OFDM-based passive optical network (PON) has been considered as one of the strong candidates for next-generation PON systems in optical communications [3].
In orthogonal frequency division multiple access-based PON (OFDMA-PON), a signal bandwidth could be simply allocated according to the data demand by assigning subcarriers to different optical network units (ONUs), which enable a network resource management to be flexible. In an intensity modulation/direct detection (IM/DD) system, downlink transmission of an optical OFDMA signal is relatively stable, such as the point-to-point link. However, critical issues exist in the uplink transmission because each ONU passes through different optical paths, which causes timing offset effect [4, 5] and optical beat interference (OBI) [6–14]. Timing offset would break the orthogonal condition among ONUs, which results in multiple access interference (MAI). To avoid the timing offset effect, we proposed a filter bank multicarrier (FBMC)-based asynchronous reception in an OFDMA-PON uplink transmission that could effectively mitigate MAI [5]. Whereas MAI mainly affects some boundary subcarriers, OBI is more severe because it would be generated over a whole signal bandwidth with a relatively large intensity. In an IM/DD system, multiple optical carriers from different ONUs are simultaneously detected at the optical line terminal (OLT), which generates carrier-to-carrier beat because of square-law detection. Optical beat interference is generated by convolution of power spectral density (PSD) of optical fields, and the center of OBI is located at the wavelength difference of the optical carriers [6, 7]. Although uplink signals have the same nominal wavelength, the independently generated optical carriers from different optical sources can be slightly different owing to their linewidth, which generates OBI in the low-frequency region. Even if an optical source is seeded from OLT to ONUs to obtain an identical optical source at every ONU, the modulated uplink optical carriers contain a time difference due to an optical path difference, which also generates OBI. To reduce OBI, several techniques have been proposed [8–14]. By separating the wavelength of optical sources [8–10], the center frequency of OBI migrates from a low frequency to a very high frequency corresponding to the wavelength difference. Wavelength separation could effectively reduce OBI in a signal band; however, it requires a different wavelength at every ONU, which would be wasteful in optical spectrum utilization and would be an obstacle in extending it to a wavelength division multiplexing (WDM) system. The polarization division multiplexing-based OBI reduction [11] can perfectly remove OBI because a carrier-to-carrier beat is not generated between orthogonal polarization states; however, the system becomes complex and could not support more than two ONUs. Carrier suppression with coherent detection was proposed [12] by eliminating the origin of the beat, which is a novel technique. However, the coherent system is still complex compared with the IM/DD system. For a simple technique, spectrum-broadening (SB)-based OBI reductions were proposed [13, 14]. By making a broad-linewidth optical source, OBI could be spread over a broad spectrum, which lowers the OBI level in a signal band. Thus, previous works on OBI reduction based on SB have focused only on the optical spectral width because a broader optical source could result in a lower noise level at a signal band. However, considering the transmission distance and WDM channel width, SB suffers from some limitations.
In this paper, we propose an optical pulse division multiplexing (OPDM)-based OBI reduction technique that considers not only the SB effect but also the near-orthogonal condition among optical carriers. In the IM/DD-based OFDMA uplink multiple access, we experimentally demonstrated that the proposed technique could simply reduce OBI and effectively minimize the OBI level to ensure signal quality. In the IM/DD OFDMA-PON uplink transmission within a single nominal wavelength, after a 20-km single-mode fiber (SMF) transmission, the spectral efficiency (SE) of the received uplink signals could be improved from 0.37 to 3.8 bit/s/Hz in a 1-GHz signal bandwidth using OPDM compared with SE of the SB technique without considering a near-orthogonal condition.
2. Schematics
Figure 1 shows the basic structure of an IM/DD-based optical OFDMA-PON system. Each ONU is independently modulated at a different location, and the signal bandwidth of the ONUs is assigned by allocating subcarriers. Although the ONUs have the same nominal wavelength, the generated optical carriers of each ONU become different because of the linewidth caused by random generation of laser diode (LD). The different uplink optical carriers are simultaneously received at the OLT after SMF transmission. Owing to the square-law detection of a photo diode (PD), multiple optical carriers generate a carrier-to-carrier beat, which generates the so-called OBI. The OBI is generated by the wavelength difference among optical carriers as the center, and the shape of the OBI is formed by convolution among the PSDs of the optical fields. If the uplink signals have the same nominal wavelength, OBI will be generated at a low frequency by forming a Lorentzian shape, as shown in Fig. 1. With a relatively large intensity in a signal band, the OBI disrupts proper detection of the uplink signals. Even if an optical source generated from a single LD is seeded from the OLT to ONUs, OBI will be generated in the same manner because the uplink optical carriers have a time delay due to an optical path difference among ONUs.
Fig. 1 Illustration of OBI generation in IM/DD-based OFDMA-PON uplink multiple access within a single nominal wavelength.
To reduce the OBI, a SB technique can be used by considering that a broad light source can spread the convolution of PSD over a broad spectrum with a reduced peak level. Figure 2 shows a SB-based OBI reduction, especially a radio frequency (RF) clipping tone (CT)-based broadening in a source seeding system [14]. The LD for the seeding source is directly modulated by RF tone at OLT. Essentially, intensity modulation has nonlinear process even in a linear modulation region of LD, which generates not only RF tone but also its harmonics. The amplitude of harmonics would be increased by modulating RF tone in a clipping region of LD, which makes the optical spectrum to be broadened. The broadened seed source is transmitted to the ONUs, and it is independently modulated by their OFDMA signal using a reflective optical modulator such as the reflective semiconductor optical amplifier (RSOA). After modulation, the uplink signals are passively combined without synchronization, which causes timing offset effect due to optical path difference. At OLT, the uplink multiple access signals are simultaneously detected, and they generate OBI, as shown in Figs. 2(a)–2(c). Without an RF CT-based SB, OBI would be generated at the low-frequency region with a high peak level, as shown in Fig. 2(a). If the RF tone is linearly modulated (m < 1, m is the modulation depth) at an out-of-signal band, partial OBI will be up-converted to RF tone because of carrier-to-tone beat while lowering the OBI peak level at a signal band, as shown in Fig. 2(b). In case of large signal modulation (m > 1), not only the RF tone but also the harmonics are strongly modulated due to signal clipping. Because large harmonics contribute to a beating process, the OBI would be up-converted to every harmonics, which could spread the OBI over a broad spectrum. As shown in Fig. 2(c), the OBI can be spread like an additive white noise by an RF CT with sufficient modulation depth.
Fig. 2 RF CT-based spectrum broadening and received RF spectra. (a) Without RF CT. (b) With linearly modulated RF tone (m < 1). (c) With nonlinearly modulated RF tone (m > 1).
By creating a broad-linewidth optical source, the OBI could be spread over a broad spectral range while lowering the OBI level in a signal band. Although a broader optical source could result in a lower noise level, the spectral width of the optical source should be limited within a WDM channel to be employed as a technique in a WDM system. Thus, we propose an OPDM-based OBI reduction technique to overcome the limitation of the SB technique. In RF CT-based SB, the broadened optical source can have not only a broad linewidth but also time/frequency regularity. However, previous works on the SB technique only focused on spectral width without considering its regularity. Thus, a broad light source without regularity, such as an amplified spontaneous emission sliced optical source, is different from a CT-based broadened source even though it has a similar spectral width. To maximize the effect of OBI reduction and ensure signal quality of service (QoS), the proposed OPDM considers the regularity of RF CT- based SB. Figure 3 shows the proposed technique in an IM/DD OFDMA-PON uplink multiple access based on source seeding system for colorless ONUs. For the seeding source, the RF CT directly modulates the LD to extend the optical spectrum, which can up-convert the OBI to an RF tone frequency and spread the OBI over a broad spectrum after uplink transmission, as shown in Fig. 2(c). In the time domain, the RF CT creates optical intensity variation according to its frequency, as shown in Fig. 3(1). This optical source is termed as optical pulse train in this work because of a periodic large intensity variation, although it is not a perfectly an on–off variation. Each ONU modulates the seeded optical pulse train using their own OFDMA signal for uplink transmission [Fig. 3(2)], which changes the envelop of the optical pulse train [Fig. 3(3)]. This process can be considered to be the same as optical sampling because signal replica is generated at the tone frequency and harmonics. Thus, the RF tone frequency should be two times larger than the signal bandwidth at the minimum according to the Nyquist sampling theorem to avoid aliasing. In the proposed method, staggered arrangement is required before the optical combination of multiple uplink signals to optimize the signal QoS. In other words, the peaks and valleys of the optical pulse train transmitted from different ONUs should be alternatively overlapped, as shown in Fig. 3(4); the condition is termed as ‘near orthogonality’ in this work.
Mathematically, the orthogonality between two time-domain signals can be defined as followings.
Assuming that the seeded optical source is an ideal rectangular pulse train, it can be expressed as
where T is a period of repetition and is a rectangular pulse train to support m-ONUs. The optical pulse train is generated at OLT and then seeded to each ONUs. For simplicity, we assumed m = 2 as illustrated in the schematics.where is a reflected rectangular pulse train without uplink signal modulation from nth ONU to OLT, and d is delay between ONUs due to optical path difference. Because optical pulse trains are simultaneously generated at OLT before source seeding, and are equal to each other except a delay. After OFDMA signal modulation, the uplink signals from two ONUs could be expressed aswhere and represent an uplink OFDMA signal and a seeded pulse train with uplink signal modulation of nth ONU, respectively. According to the definition in Eq. (1), if , then orthogonal condition could be satisfied. In this case, when , then , which could make rectangular pulse trains to be perfectly staggered while satisfying orthogonality. Thus, by controlling delay between ONUs, orthogonal condition can be adjusted. In this work, this condition is termed as ‘near orthogonality’ instead of ‘orthogonality’. Because the proposed system used RF CT-based optical pulse modulation, the seeded source was non-ideal rectangular pulse train and not perfectly on-off; therefore,. Nevertheless, orthogonal condition could be nearly satisfied because; thus, the condition is termed as near orthogonality. This condition is satisfied when the peaks of optical pulse train transmitted from ONUs are alternatively appeared. The term ‘OPDM’ is used when optical pulse trains from different ONUs are combined under this condition.The proposed OPDM can overcome the limitation of the SB technique by considering both broad spectrum and near orthogonality. The proposed method can simply and effectively reduce the OBI when a single nominal wavelength is used. Moreover, as expressed in Eq. (2), the number of overlaid ONU without OBI can be increased by changing the pattern of the optical pulse train. In this case, the number of ONU will be limited due to response of LD because a higher frequency of RF tone is required to support the increased number of ONU. Nevertheless, it could make PON split ratio to be increased by several times without additional time scheduling by hybridizing with a legacy time division multiplexing (TDM) system.
3. Experiments
Figure 4 shows the experimental setup to demonstrate the OPDM-based OBI reduction in an IM/DD OFDMA-PON uplink transmission. A distributed feedback laser diode (DFB-LD) was used as a single optical seeding source. Because we proposed an OPDM-based OBI reduction technique by considering a broad linewidth and near-orthogonal condition, LD was directly modulated by RF CT for SB. The vector signal generator (VSG) generated a 2.7-GHz RF tone, which satisfied the Nyquist sampling theorem because the used uplink signal bandwidth was 1 GHz. The optical source was split and seeded to the ONUs by passing through different optical paths. To break coherence and create an optical path difference, an additional 5-km SMF was inserted. The uplink signals were independently modulated at each separate ONU with the best effort using the RSOA for a colorless system. Polarization controllers (PCs; PC1 and PC2) were used to optimize the modulation efficiency before the RSOAs considering the polarization sensitivity of the RSOA. The electro-optic conversion was individually optimized by controlling the bias point and modulation depth. After the optical modulation, additional PCs (PC3 and PC4) were used to maximize the OBI by matching the polarization state. Thus, we can consider that polarization effect was excluded, and the OBI reduction totally depended on the OPDM effect in this experiment. To verify the performance variation according to near-orthogonal condition, an optical delay line was inserted at ONU1 before a 3-dB coupler. The uplink signals passed through different optical paths and were then passively combined at the optical 3-dB coupler without synchronization. The optical bandpass filter (OBPF) emulated the WDM multiplexer to verify WDM expandability. The 3-dB bandwidth of the used OBPF was 50 GHz. After a 20-km SMF transmission, the multiple access signals were directly detected at the PD.
OFDMA uplink signals were generated using an offline process. The number of subcarriers was 128, and each ONU was assigned 63-subcarriers (2nd–64th for ONU1 and 66th–128th for ONU2) except the guard subcarrier (1st and 65th). Even though a single AWG generated two synchronized OFDMA signals for ONUs, timing offset effect would be generated between ONUs due to RF/optical path difference. In our previous work [5], we experimentally demonstrated that the OFDM sidelobes were effectively suppressed by employing FBMC in an OFDMA-PON uplink multiple access, which could effectively mitigate the MAI among ONUs. To avoid the timing offset effect, FBMC-based sidelobes suppression was employed in this work [5, 15]. Thus, this experiment only focused on the OBI effect by excluding the timing offset effect. The performance was evaluated according to near-orthogonal condition in terms of channel error vector magnitude (EVM) based on preamble. Adaptive modulation was employed based on preamble feedback information. The preamble was modulated by four quadrature amplitude modulations (4-QAM) before setting the bit-loading profile. Bit-loading criteria was to assure quality of received signal by assuming forward error correction (FEC). At the receiver, individual fast Fourier transform (FFT)-based asynchronous reception [4, 5] was used on each ONU to avoid inter-symbol-interference. After the FFT, a single-tap equalizer was used to equalize the channel effects based on the preamble at the receiver side.
4. Results and discussions
Figure 5(a) shows the spectrum of RF tone before optical modulation, which directly drives LD after amplification. Figures 5(b)–5(d) show the received RF spectra without uplink OFDMA signal modulation, which well matches those shown in Figs. 2(a)–2(c). Without the RF CT-based SB, OBI is generated at a low frequency, forming a Lorentzian shape, as shown in Fig. 5(b). If the RF tone directly modulates the LD, the OBI can be up-converted to the RF tone frequency, as shown in Fig. 5(c), because both optical carriers and RF tone participate in the beating process, which can lower the OBI level at the low-frequency side. Furthermore, if a directly modulated RF tone is sufficiently strong to create clipping, larger harmonics will be generated compared with that in a small-signal modulation case. Thus, OBI could be up-converted to not only an RF tone but also harmonics, which can lower the OBI level as well as form a flattened spectrum shape, as shown in Fig. 5(d). Therefore, sufficient RF signal power is required in an RF CT-based SB to reduce the OBI level in a signal bandwidth.
Fig. 5 RF spectra of the received signal without uplink signal modulation. (a) RF tone (without amplification) before optical modulation (b) OBI without SB. (c) Spectrum broadening with RF tone (m < 1). (d) Spectrum broadening with RF CT (m > 1).
A sufficiently large RF CT could broaden optical spectrum of seeding source. Figures 6(a) and 6(b) show optical spectra of seed source at OLT before and after RF CT modulation, respectively. In case of Fig. 6(b), a frequency of the modulated RF tone was 2.7 GHz and an input RF power was + 19.5 dBm. Even though a relatively large RF power is required to make a broadened optical source, ONU power efficiency is not a critical issue because the seeding sources are commonly generated at OLT (or remote node). As shown in Fig. 6(b), a broadened 10-dB optical spectrum was about 0.2 nm (25 GHz) by employing RF CT. It is corresponded to a required broadened optical linewidth to spread OBI over a broad spectral range with a flat-spectrum shape by using RF CT-based SB.
Although a proper SB could reduce the OBI level in a signal band, this process is not sufficient to ensure quality of uplink signals, as obviously shown in Figs. 7(a)–7(d). Figures 7(a)–7(d) show the RF spectra of the received signal to demonstrate the importance of a near-orthogonal condition. Figures 7(a) and 7(b) show the spectra without an uplink signal. Figure 7(a) shows that the OBI could be effectively reduced and flattened using OPDM, which considers both SB and time-domain near orthogonality. Figure 7(b) shows the RF spectra of OPDM and SB without considering near orthogonality (SB only). Although the RF signal equally modulated the LD using the same modulation depth in both cases, the received noise level was greatly different; at low frequency, the difference was approximately 20 dB. Figures 7(c) and 7(d) show the received RF spectra of the OPDM case and a SB case, respectively, when the uplink signals were modulated. In both cases, because the RF CT acted similar to optical sampling, the uplink signal replica is observed at the CT frequency. No aliasing can be observed in Figs. 7(c) and 7(d) because the uplink signal bandwidth was 1 GHz and CT was 2.7 GHz to adhere to the Nyquist sampling theorem. Due to the RF tones, modulation efficiency of the RF CT-based system would be degraded compared to that of a pure optical source (without RF CT)-based single ONU transmission. Moreover, a broadened optical spectrum would reduce spectral efficiency of WDM compared to that of a single ONU transmission. Nevertheless, it is impossible to obtain meaningful signal quality due to OBI when multiple ONUs simultaneously transmit signals within a single wavelength. Thus, the critical OBI could be effectively controlled with a small sacrifice of modulation efficiency. The modulation efficiency of the uplink signal is determined by the RSOA, and the noise level caused by the OBI would not affect the modulation efficiency. Therefore, the modulation levels of the uplink signal become equal in both cases, as shown in Figs. 7(c) and 7(d). However, the OBI noise level changes depending on the near-orthogonal condition, and the increased noise level reduces the signal-to-noise ratio, which degrades the signal quality. Thus, SB without OPDM is not sufficient for OBI reduction because the signal QoS cannot be satisfied without satisfying the near-orthogonal condition.
Fig. 7 RF spectra of the received signal. (a) OBI and OPDM without uplink signal modulation. (b) Spectrum broadening without near orthogonality (SB only) and OPDM when the uplink signal was not modulated. (c) OPDM with uplink signal modulation. (d) Spectrum broadening without near orthogonality when the uplink signal was modulated.
Figures 8(a) and 8(b) show the experimental results according to a subcarrier index by varying the optical pulse timing offset. The optical pulse timing offset was varied by controlling the optical delay line at the ONU1 side before combining the multiple uplink signals. The delay line was controlled at an interval of 25 ps from 0 to 700 ps. Figure 8(a) shows that the channel EVM exhibited a periodic property depending on the optical timing offset among the ONUs. Because the near-orthogonal condition is determined by the intensity variation of an optical pulse, the channel EVM is influenced by this regularity. A near-orthogonal condition can be satisfied when the peaks of an optical pulse of one ONU is overlaid with the valley of an optical pulse of another ONU. The pulse train period (T) is decided by frequency of RF CT (f). Unless there is a specially intended pulse pattern, a period of pulse train T equals 1/f, which means delay cycle is 1/f. In our experiment, a pulse peak appeared every 370 ps because the RF CT was generated at 2.7 GHz, which corresponds to the cyclic property of the channel EVM. If the higher frequency of CT is used, the cycle will be shortened. This cyclic property is only related with RF CT, in other words, the delay is not related with uplink OFDM symbol rate. Thus, once optical link is set and frequency of RF CT is fixed, OPDM condition will be unchanged. To maximize transmission capacity, adaptive modulation was employed in this experiment, which was based on channel EVM. Thus, the bit-loading profile shown in Fig. 8(b) has the same property as that shown in Fig. 8(a) according to the optical pulse timing offset. When a near-orthogonal condition is satisfied, the number of total loaded bits can be maximized while increasing the transmission capacity and SE. On the other hand, when the peaks of one ONU are overlaid with those of another ONU, the capacity and SE would be minimized due to an increased noise level, which corresponds to the worst SB case shown in Fig. 7(d).
Fig. 8 Experimental results according to the subcarrier index by varying the optical pulse timing offset. (a) Channel EVM. (b) Bit-loading profile.
Figure 9 shows the SE of the uplink signals according to the optical pulse timing offset. The received performance under OBI is similar to SE of the worst case of SB, which corresponds to the minimum value of Fig. 9. Under OBI effect, subcarriers cannot load even one bit (SE < 1 bit/s/Hz) [14]. The SE without SB is not included in Fig. 9 because OBI effect is unrelated with optical timing offset. The peaks and valleys of the graph are obvious, and the cycle well matches with the optical pulse period. Actually, SB itself could improve SE of received signal due to a stabilized signal fluctuation by spreading OBI over a broad spectrum and lowering the noise level in a signal band. Even though a proper SB was employed for OBI reduction, in the 1-GHz signal bandwidth, the total SE varied from 0.37 to 3.8 bit/s/Hz according to a near-orthogonal condition. That is, the possible variation of SE in [14] can be well controlled and the following SE maximization can be achieved by adapting near-orthogonal condition with the proposed OPDM technique. Thus, the proposed OPDM-based OBI reduction is essential in ensuring the signal QoS and optimizing the SE using the SB technique.
5. Conclusion
We have proposed an OPDM-based OBI reduction with FBMC-based asynchronous reception in OFDMA-PON uplink where OBI and timing offset would be generated by multiple ONUs. By considering both spectrum-broadening effect and near-orthogonal condition, OBI could be effectively reduced, and the noise level could be minimized, which could improve the signal EVM and SE. The improvement of the proposed OPDM-based OBI reduction was experimentally demonstrated in an IM/DD-based OFDMA-PON multiple uplink transmission within a single nominal wavelength after a 20-km SMF transmission. By employing the proposed technique in OFDMA-PON, OBI could be simply reduced by overcoming the limitation of the spectrum-broadening technique, and it enables the system to be extended to a WDM system with colorless ONUs. Furthermore, it can support and hybridize a conventional TDM system, which increases the split ratio by several times without optical TDM scheduling. Thus, we believe that the proposed OPDM-based technique will be very useful for OBI reduction in an IM/DD-based OFDMA-PON uplink transmission.
Funding
ICT R&D programs of MSIP/IITP, Republic of Korea. (R0101-16-0086)
References and links
1. G. Wunder, P. Jung, M. Kasparick, T. Wild, F. Schaich, Y. Chen, S. Brink, I. Gaspar, N. Michailow, A. Festag, L. Mendes, N. Cassiau, D. Kténas, M. Dryjanski, S. Pietrzyk, B. Eged, P. Vago, and F. Wiedmann, “5GNOW: Non-orthogonal, asynchronous waveforms for future mobile applications,” IEEE Commun. Mag. 52(2), 97–105 (2014). [CrossRef]
2. E. Dahlman, S. Parkvall, and J. Skold, 4G: LTE/LTE-Advanced for Mobile Broadband, 2nd ed. (Academic, 2015), Ch. 1.
3. N. Cvijetic, D. Qian, J. Hu, and T. Wang, “Orthogonal frequency division multiple access PON (OFDMA-PON) for colorless upstream transmission beyond 10 Gb/s,” IEEE J. Sel. Areas Comm. 28(6), 781–790 (2010). [CrossRef]
4. J. von Hoyningen-Huene, H. Griesser, M. H. Eiselt, and W. Rosenkranz, “Asynchronous signal reception in OFDMA-PON-uplink,” in Advanced Photonics 2013, Advanced Photonics Congress (Optical Society of America, 2013), paper SP4D.2.
5. S. Y. Jung, S. M. Jung, H. J. Park, and S. K. Han, “Mitigation of timing offset effect in IM/DD based OFDMA-PON uplink multiple access,” Opt. Express 23(11), 13889–13898 (2015). [CrossRef] [PubMed]
6. N. K. Shankaranarayanan, S. D. Elby, and K. Y. Lau, “WDMA/Subcarrier-FDMA Lightwave Networks: Limitations due to Optical Beat Interference,” J. Lightwave Technol. 9(7), 931–943 (1991). [CrossRef]
7. T. H. Wood and N. K. Shankaranarayanan, “Operation of a passive optical network with subcarrier multiplexing in the presence of optical beat interference,” J. Lightwave Technol. 11(10), 1632–1640 (1993). [CrossRef]
8. L. A. Neto, G. B. Farias, N. Genay, S. Menezo, B. Charbonnier, P. Chanclou, and C. A. Berthelemot, “On the limitations of IM/DD WDM-FDMA-OFDM PON with single photodiode for upstream transmission,” in Optical Fiber Communication Conference and Exposition, and the National Fiber Optic Engineers Conference 2012, OFC/NFOEC Technical Digest (Optical Society of America, 2012), paper OW4B.1.
9. W. Q. Jin and J. M. Tang, “Experimental investigations of wavelength spacing and colorlessness of RSOA-based ONUs in real-time optical OFDMA PONs,” J. Lightwave Technol. 30(16), 2603–2609 (2012). [CrossRef]
10. I. Cano, M. C. Santos, V. Polo, F. X. Escayola, and J. Prat, “Dimensioning of OFDMA PON with non-preselected independent ONUs sources and wavelength-control,” Opt. Express 20(1), 607–613 (2012). [CrossRef] [PubMed]
11. D. H. Kim, P. Y. Xing, Y. Y. Won, S. J. Park, and S. K. Han, “Bidirectional 1.25Gb/s colorless RSOA based WDM-PON using suppressed optical carrier and polarization beam splitter,” in Optical Fiber Communication Conference/National Fiber Optic Engineers Conference 2008, OSA Technical Digest (CD) (Optical Society of America, 2008), paper JThA97. [CrossRef]
12. D. Qian, N. Cvijetic, J. Hu, and T. Wang, “A novel OFDMA-PON architecture with source-free ONUs for next-generation optical access networks,” IEEE Photonics Technol. Lett. 21(17), 1265–1267 (2009). [CrossRef]
13. S. L. Woodward, X. Lu, T. E. Darcie, and G. E. Bodeep, “Reduction of optical-beat interference in subcarrier networks,” IEEE Photonics Technol. Lett. 8(5), 694–696 (1996). [CrossRef]
14. S. M. Jung, S. M. Yang, K. H. Mun, and S. K. Han, “Optical beat interference noise reduction by using out-of-band RF clipping tone signal in remotely fed OFDMA-PON link,” Opt. Express 22(15), 18246–18253 (2014). [CrossRef] [PubMed]
15. M. Bellanger, D. LeRuyet, D. Roviras, M. Terré, J. Nossek, L. Baltar, Q. Bai, D. Waldhauser, M. Renfors, T. Ihalainen, A. Viholainen, T. H. Stitz, J. Louveaux, A. Ikhlef, V. Ringset, H. Rustad, M. Najar, C. Bader, M. Payaro, D. Katselis, E. Kofidis, L. Merakos, A. Merentitis, N. Passas, A. Rontogiannis, S. Theodoridis, D. Triantafyllopoulou, D. Tsolkas, D. Xenakis, M. Tanda, T. Fusco, M. Huchard, J. Vandermot, A. Kuzminskiy, F. Schaich, P. Leclair, and A. Zhao, “FBMC physical layer: a primer,” http://www.ict-phydyas.org.
