Open Access
Add to BibSonomy
Abstract
Enhancing spectral efficiency (SE) of ultra-dense wavelength division multiplexing passive optical network (UDWDM-PON) is vital to providing broadband access for massive users. Here, we experimentally demonstrate a high SE UDWDM-PON in the C-band, based on the simplified coherent reception of 10 Gb/s 4-level pulse-amplitude modulation (PAM-4) signals. We investigate the WDM signal reception by mathematical derivation and propose to enhance the SE by adopting both intradyne detection and pulse shaping techniques. Then, both approaches are numerically evaluated, with an identification that there occurs a trade-off between SE and power budget improvements. Finally, we experimentally achieve a SE of 0.83 (bit/s)/Hz and a power budget of 25 dB for a proof-of-concept 3 × 10 Gb/s PAM-4 downstream transmission over 20 km standard single mode fiber (SSMF).
© 2023 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
The continuously increasing demand for bandwidth, driven by high-definition video, Internet of Things, wireless fronthaul, and other bandwidth-consumable applications, has stimulated the need for high-capacity broadband optical access networks [1]. Currently deployed passive optical networks (PONs) employ a time-division multiple-access (TDMA) technique, which can provide an aggregate network capacity of only 10 Gb/s [2] and rapidly evolve to a 50 Gb/s counterpart. Meanwhile, the latest ITU-T standard named next-generation PON 2 (NG-PON2), intends to employ both time- and wavelength-division multiplexed (TWDM) techniques to enhance the aggregated network capacity [3]. As a result, intensive research works have been performed on ultra-dense wavelength division multiplexing passive optical networks (UDWDM-PONs), with the help of simplified coherent detection. In comparison with previous WDM-PON, the UDWDM-PON can share the same installed optical distributed network (ODN) of TDM-PON. Moreover, the end-user is possible to occupy the full access bandwidth, instead of sharing it among all TDM-PON users [4].
Coherent detection has been widely employed in long-haul fiber optical transmission, due to its capability of complete optical field recovery, superior sensitivity, and high spectral efficiency (SE) [5]. This optical field recovery feature allows for the mitigation of transmission impairments through digital signal processing (DSP) [6]. In the context of UDWDM-PON, there are two motivations for adopting coherent detection. Firstly, its fine wavelength selectivity preserves the optical filter-free nature of the ODN, indicating a cost-effective PON deployment. Secondly, superior sensitivity is indispensable for compensating for the insertion loss of ODN and increasing the power budget of PON. However, the conventional coherent receiver is complex and expensive, which hinders its application in the cost-sensitive PON. To address such issues, several schemes of simplified coherent detection have been proposed. The key point of a simplified coherent receiver is to realize a polarization-independent reception. Transmitter-side approaches, including Alamouti coding [7] and polarization scrambling [8], are still inconvenient for UDWDM-PON. Although the Alamouti coding can greatly simplify the optical front end of a coherent receiver [9], its SE is reduced at the optical network unit (ONU) [10]. On the other hand, polarization scrambling requires an external polarization modulator with a modulation bandwidth greater than twice the symbol rate. Meanwhile, one receiver-side approach can realize a polarization-diversity analog homodyne reception, by exploiting the injection locking of a pair of externally modulated lasers [11]. However, such a scheme poses a stringent demand on the linewidth of the used laser and necessitates the complex orthogonal frequency division multiplexing (OFDM) demodulation. Alternatively, a receiver-side approach based on a polarization beam splitter (PBS) and low-complexity analog signal processing provides an economical solution for the UDWDM-PON [12]. The 3 × 3 coupler based heterodyne reception has been experimentally verified [13] and applied for the field demonstrations of UDWDM-PON [14].
On the other hand, the SE improvement of UDWDM-PON is crucial, in order to realize both multiple-access and large capacity within the limited low-loss transmission window of standard single mode fiber (SSMF). Although traditional digital coherent receivers can achieve high SE by the use of advanced modulation formats [15–17], the inherent low SE of UDWDM-PONs based on the 3 × 3 coupler simplified coherent reception presents a major challenge, due to the heterodyne detection and the incapable reception of phase-modulated signal. A 4 × 10 Gb/s WDM-PON by the use of on-off keying (OOK), is experimentally demonstrated with a power budget of 55 dB and an SSMF reach of 110 km. However its SE is only 0.1 (bit/s)/Hz [18]. Furthermore, a SE of 0.16 (bit/s)/Hz has been reported for the 6.25 GHz spaced 8 × 1 Gb/s UDWDM-PON downstream transmission [19], leading to the realization of both a power budget of 50 dB and an SSMF reach of 70 km. Recently, we have demonstrated an 8 × 10 Gb/s UDWDM-PON with a channel spacing of 20 GHz over the C-band, offering up to a SE of 0.5 (bit/s)/Hz [20]. The power budget of 29 dB is achieved after the 10 Gb/s 4-level pulse-amplitude modulation (PAM-4) signal is transmitted over 25 km SSMF. Although the use of PAM-8 signals can enhance the SE, its lower noise margin results in additional sensitivity penalty.
In the current submission, we propose to further increase the SE of UDWDM-PON by the intradyne reception and pulse shaping techniques. The rest of the paper is organized as follows. Section 2 theoretically investigates WDM signal reception utilizing the simplified coherent receiver and provides possible approaches to enhance the SE. Section 3 provides numerical validation of the proposed approaches. Finally, we experimentally demonstrate the effectiveness of both intradyne reception and pulse shaping, by achieving a SE of 0.83 (bit/s)/Hz and a power budget of 25 dB, when a 3 × 10 Gb/s PAM-4 signal is experimentally transmitted over 20 km SSMF.
2. Operation principle
Figure 1 depicts the schematic diagram of polarization-independent simplified coherent receivers. The received signal is split into two orthogonal linear polarization components by a PBS. The horizontal linear polarization component undergoes a 90-degree polarization rotation and is then combined with the vertical linear polarization component with the same state of polarization (SOP) of the local oscillator (LO), by the use of a symmetrical 3 × 3 coupler. After coherent mixing, the three-channel optical signals are converted to the corresponding electrical signals with three single-end photodetectors (SPDs). By squaring and summing the electrical currents from the three channels, the received signal can be obtained. Finally, with the help of a low-pass filter (LPF), we can suppress the polarization-dependent interference, enabling the realization of polarization-independent coherent detection.
Fig. 1. Schematic diagram of polarization-independent simplified coherent receivers based on symmetric 3 × 3 couplers. TIA: trans-impedance amplifier.
Without the loss of generality, we conduct an analysis of WDM signal reception using a simplified coherent receiver, focusing on three adjacent channels. The middle channel is under investigation for the reception. The transmitted signal can be expressed as
The direct-detection and signal-signal beating terms have been reasonably ignored, due to their significantly smaller magnitudes in comparison with the signal-LO beating term. The received signal S(t) is derived by squaring and summing three currents. Frequency domain components that are distant from the baseband have been ignored, as they will be effectively suppressed by an electrical LPF. These frequency domain components include polarization-dependent interference from adjacent channels, as it is located at 2Δυ1 or 2Δυ3. Following a series of reasonable simplifications, the S(t) expression can be finally obtained
To assess transmission impairments during the reception of WDM signals, we undertake an analysis of the S(t) expression. $r_2^2(t)$ is the intensity of the baseband signal at the middle channel to be received. $r_2^2(t)$ sinφsin(4πfoffsett+ψ+π/6) represents polarization-dependent interference, and it’s a polarization-dependent spurious copy of the baseband signal $r_2^2(t)$ translated at 2foffset. The polarization-dependent interference can be suppressed by an LPF with a suitable bandwidth [12], because it is spectrally separated from the baseband signal $r_2^2(t)$, under the condition of heterodyne detection (frequency offset foffset is greater than the bandwidth of baseband signal B). $r_2^1(t)+ r_3^2(t)+2r_1(t)\;\;r_3(t)cos(4\pi {f_{offset}}t)$ represents the inter-channel-interference (ICI), due to the reception of adjacent channels. Considering the filtering effect of the SPD, the channel closer to the LO in the frequency domain introduces more significant crosstalk. Consequently, $r_3^2(t)$ contributes dominantly to the ICI.
The SE of WDM transmission system is commonly defined as the ratio between the bit-rate per channel and the channel spacing [22]. Reducing the channel spacing leads to the SE enhancement, but simultaneously introduces serious ICI, which results in a reduction of power budget. Therefore, there exists a trade-off between SE and power budget. To enhance SE while providing a sufficient power budget, we decide to adopt intradyne detection, where the frequency offset foffset is smaller than the bandwidth of baseband signal B [23]. Meanwhile, the root-raised cosine (RRC) pulse shaping is proposed by the use of a simplified coherent receiver in the UDWDM-PON. Since reducing the frequency offset can shift the CH3 away from the LO in the frequency domain, the ICI can be suppressed, together with the maintenance of both the power budget and the higher SE. However, the operation of intradyne detection brings another problem of stronger polarization-dependent interference. Thus, the optimization of frequency offset is crucial. Furthermore, since the RRC pulse shaping technique can compress the bandwidth of baseband signal, it is helpful to mitigate the polarization-dependent interference, under the condition of intradyne detection.
3. Simulation
In order to investigate the transmission performance of the proposed UDWDM-PON with the help of the simplified coherent receiver, we carry out several numerical simulations. An 8 × 10Gbps PAM-4 transmission is simulated over 20 km SSMF over the C band. The center frequency of the middle channel is fixed at 193.1THz, and the local oscillator (LO) frequency is configured as 193.1THz + foffset, for the ease of its reception. Focusing on the characterization of the middle channel can intuitively demonstrate the worst performance of the multi-channel downstream transmission, thereby assessing the overall transmission performance. Moreover, the frequency domain response of C-band intensity modulation signal after a 20 km SSMF transmission exhibits the first frequency dip at 13.6 GHz [24], beyond the bandwidth of 5GBaud PAM-4 signal. Thus, the impact of dispersion on the PAM-4 signal is trivial. To mitigate nonlinear transmission impairments, it is initially set the launch power at −5 dBm in the simulation. Reducing the SPD bandwidth can be effective in mitigating the ICI, but it may introduce the inter-symbol-interference (ISI), due to the bandwidth constraint effect. To address such a challenge, we can choose a smaller frequency offset to reduce the SPD bandwidth of B + foffset. Consequently, it becomes imperative to optimize both the frequency offset and SPD bandwidth concurrently. In pursuit of achieving high SE, we perform optimization for both frequency offset and SPD bandwidth, when the received optical power (ROP) is -26 dBm. As shown in Fig. 3(a), when the optimal value is 3.5 GHz for the frequency offset and the SPD bandwidth is 5 GHz, the optimization efforts yield a SE of 0.74 b/(Hz·s), under a minimum channel spacing of 13.5 GHz.
Fig. 3. (a) Channel spacing versus different SPD bandwidth and frequency offset. (b) Received sensitivity penalty under different roll-off factors. (c) Channel spacing versus different SPD bandwidths and frequency offsets when the roll-off factor is 0.4.
The Nyquist pulse shaping plays a crucial role in reducing the signal bandwidth, while avoiding inter-symbol interference (ISI) at the decision point. In practice, the RRC filter is commonly employed both at the transmitter and receiver to realize the pulse shaping. The simplified coherent receiver employs a root-raised cosine filter, instead of the low-pass filter, to enable the matched filtering and achieve a polarization-independent reception. As a result, the bandwidth of the baseband signal is confined within 0.5B(1+α), where α represents the roll-off factor. As for 5 GBaud PAM-4 signals, the bandwidth of the baseband signal can be restricted within a range of 2.5∼5 GHz, when α varies from 0 to 1. Figure 3(b) illustrates the penalty of receiver sensitivity with respect to various roll-off factors. The system has the best sensitivity performance, when the roll-off factor is 0.5. The sensitivity penalty is less than 0.5 dB, when the roll-off factor is 0.4 and the bandwidth of baseband signal becomes 3.5 GHz. A comparison of frequency spectra between the pulse-shaped signal with a roll-off factor of 0.4 and an original signal without the pulse shaping is indicated in Fig. 3(b). In the following simulation, a roll-off factor of 0.4 is chosen, due to the significant reduction of signal bandwidth, while maintaining an acceptable sensitivity penalty. Based on the optimal roll-off factor, Fig. 3(c)shows the joint optimization of both frequency offset and SPD bandwidth at the ROP of −26 dBm. It is observed that the optimal values are identified with 3 GHz for the frequency offset and 5 GHz for the SPD bandwidth, under the condition of the pulse-shaped signal. This configuration achieves a minimum channel spacing of 12.8 GHz, resulting in higher SE in comparison with the case without the pulse shaping. The simplified coherent receiver operates in the heterodyne regime for the pulse-shaped signal, when the frequency offset is 3.5 GHz, thereby undergoing less polarization-dependent interference. When a frequency offset of 3 GHz is applied, intradyne detection enables higher SE.
In the previous results, the launch power and the ROP are fixed at −5dBm and −26dBm, respectively, providing only a 21 dB power budget. The launch power can be further enhanced, in order to maximize the power budget. However, this will inevitably result in the nonlinear impairments. As for the UDWDM-PON, the dominant factor contributing to nonlinear impairment penalty is the four-wave mixing (FWM) effect, due to the narrow channel spacing [25]. Based on the optimization above, the optimal frequency offset is 3.5 GHz for the PAM-4 signals without the pulse shaping, while the optimal value of 3 GHz is chosen for the pulse-shaped signals. For both scenarios, the SPD bandwidth is fixed to 5 GHz. To evaluate the impact of the FWM effect on the performance of receiver sensitivity, we carry out a numerical simulation to identify the power budget and the optimal launch power, as shown in Fig. 4(a). It is observed that both the optimal launch power and the power budget become better with the SE reduction, due to the mitigation of the FWM effect under the wide channel spacing. 29 dB power budget can be achieved for a spectral efficiency of 0.625 b/(Hz·s). Reducing the channel spacing leads to a penalty of power budget. Figure 4(b) shows the launch power in relevant to the power budget. For the PAM-4 signals with the pulse shaping and without pulse shaping, the channel spacing is set to 14 GHz and 13 GHz, respectively. It is seen that, eight-channel downstream transmission can achieve 25 dB and 23 dB power budgets, respectively. In comparison with the three-channel downstream transmission, eight-channel transmission experiences additional an 2 dB penalty of power budget, due to the serious FWM effect with the growing channels.
Fig. 4. (a) Power budget and optimal launch power versus spectral efficiency. (b) Power budget versus launch power.
4. Experimental results and discussions
We experimentally verify the proposed UDWDM-PON with high SE, based on a simplified coherent receiver, as shown in Fig. 5(a). Considering the dominate contribution of adjacent channels to ICI, we carry out a proof-of concept experiment based on three wavelength channels. At the optical line terminator (OLT), the downstream transmitter employs three distributed feedback (DFB) lasers as its light sources. The operation wavelength of the central channel is fixed at 1550 nm, while the operation wavelength of two adjacent channels can be adjusted according to the experimental setting of channel spacing. The central channel and adjacent channels are modulated by two Mach-Zehnder modulators (MZMs), respectively. To ensure that adjacent channels carry the uncorrelated data, two independent pseudorandom bit sequences (PRBSs) are utilized to generate 10 Gb/s PAM-4 signals. The length of each PRBS sequence is 214−1. An arbitrary waveform generator (AWG) running at 92 GSa/s generates two independent 10 Gb/s PAM-4 signals. After being amplified by electrical amplifiers (EAs), the electrical waveforms are used to drive two MZMs, respectively. After being combined with a 2 × 1 optical coupler, the PAM-4 signals are introduced into the SSMF. At the ODN, a variable optical attenuator (VOA) is used to emulate the insertion loss of a power splitter and adjust the ROP. In the absence of wavelength division de-multiplexing, three-channel PAM-4 signals are directly introduced to the ONU and detected by the simplified coherent receiver. The LO power is set at 6 dBm, and its operation wavelength is manually adjusted to achieve either heterodyne or intradyne detection. The output ports of the 3 × 3 coupler are independently connected to three positive-intrinsic-negative-photodiodes (PIN-PDs) with trans-impedance amplifiers (TIA) having a 3-dB bandwidth of 5 GHz. To emulate electrical signal processing, a real-time oscilloscope (RTO) operating at 40 GSa/s is used to record three photodetected signals, respectively. Then three received signals are individually squared, summed, and filtered by offline signal processing, in order to recover the electrical PAM-4 signals. When the pulse shaping technique is employed, the DSP flowcharts at the transmitter and receiver side are indicated in Fig. 5(b) and Fig. 5(c), respectively. The signal is resampled to two samples per symbol and then processed with the RRC filter at both the transmitter and receiver. The low-pass filter in the simplified coherent receiver is replaced with the RRC filter, achieving both matched filtering and polarization-independent reception.
Fig. 5. (a) Experimental Setup of PAM-4 UDWDM-PON based on simplified coherent receivers. (b) Transmitter DSP process (c) Receiver DSP process. EA: electronic amplifier.
Initially, we perform an experimental investigation on single-channel 10 Gb/s PAM-4 back-to-back(B2B) transmission, when only one DFB laser is powered at the OLT. The impact of frequency offset on the BER performance is examined, by finely tuning the wavelength of the LO laser with a fixed LO power. Figure 6(a) indicates that, the BER is optimal when the frequency offset ranges from 4 GHz to 5 GHz. For the single-channel B2B transmission, we confirm that the frequency offset of 5 GHz is optimal for the heterodyne detection. However, a frequency offset of 3.5 GHz is chosen for the intradyne detection, in order to increase the SE for the UDWDM-PON downstream transmission.
Fig. 6. (a) BER performance versus frequency offset. (b) BER performance versus channel spacing for intradyne detection and heterodyne detection.
Next, we conduct the UDWDM-PON downstream transmission to determine the minimum channel spacing for both heterodyne and intradyne detection. The launch power per channel is fixed at −1 dBm. As shown in Fig. 6(b), in order to realize the power budget of 27 dB, our experimental results indicate that intradyne detection and heterodyne detection can achieve channel spacings of 14 GHz and 16 GHz, respectively, leading to the SE enhancement from 0.625 (bit/s)/Hz to 0.714 (bit/s)/Hz.
As shown in Fig. 7(a), we further investigate the performance of Nyquist pulse shaping with different roll-off factors for single-channel 5 GBaud PAM-4 downstream transmission, under the B2B transmission. It is observed that, the BER performance starts to degrade significantly under a roll-off factor of 0.1. When the roll-off factor of Nyquist pulse shaping is about 0.4, the system achieves the optimal BER performance. Additionally, the eye diagram of the pulse-shaped signal with a roll-off factor of 0.4 is presented in Fig. 7(a). The upper level of PAM-4 eye diagram is noisy, because the major noise sources of the simplified coherent receiver comes from both the shot noise and the relative intensity noise (RIN) of the LO laser. In light of this phenomenon, unequally-spaced PAM-4 modulation signaling can provide an enhancement of receiver sensitivity [26].
Fig. 7. (a) Optimization of roll-off factor. (b) BER performance versus frequency offset. (c) BER performance versus channel spacing. (d) SE versus power budget for different works.
Based on the aforementioned experimental results, we select a roll-off factor of 0.4 for conducting the three-channel downstream transmission over 20 km SSMF. The relationship between frequency offset and BER is shown in Fig. 7(b). A frequency offset of 3 GHz is set under the condition of intradyne detection. Furthermore, Fig. 7(c) illustrates the relationship between channel spacing and BER. The adoption of pulse shaping and intradyne detection results in a reduction of the minimum channel spacing to 12 GHz. Consequently, the SE further increases from 0.714 (bit/s)/Hz to 0.833 (bit/s)/Hz. Since the launch power is −1dBm, a power budget of 25 dB is in-hand. Figure 7(d) presents the relationship between SE and power budget, when recent UDWDM-PON results are summarized.
5. Conclusions
We have experimentally demonstrated an UDWDM-PON with high SE, by the use of the simplified coherent detection with a power budget of 25 dB and an SSMF reach of 20 km. By employing both intradyne detection and pulse shaping, we have reduced the channel spacing to 12 GHz, leading to a SE of 0.833 (bit/s)/Hz. When the SE has increased from 0.5 (bit/s)/Hz to 0.833 (bit/s)/Hz, resulting in a 67% increase in the number of users over the C-band. The system has a lower SPD bandwidth requirement, due to the smaller receiver bandwidth required for intradyne detection, in comparison with the heterodyne detection. It is ideally desired to provide broadband access for more users, with the help of wavelength-to-end user technique.
Funding
National Natural Science Foundation of China (62025502); Guangdong Introducing Innovative and Entrepreneurial Teams of “The Pearl River Talent Recruitment Program” (2021ZT09X004).
Disclosures
The authors declare that there are no conflicts of interest related to this article.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. P. Miguelez, “What applications are driving higher capacity in access?” 2018 Opt. Fiber Commun. Conf. Expo. OFC 2018 - Proc.1–3 (2018).
2. ITU-T, “ITU-T G.987.1 10-Gigabit-capable passive optical networks (XG-PON): General requirements,” International Telecommunication Union, Geneve, Switzerland (2016).
3. ITU-T, “ITU-T Recommendation G.989.1 40-Gigabit-capable passive optical networks (NG-PON2): General requirements,” International Telecommunications Union, (2013).
4. J. D. Reis, A. Shahpari, R. Ferreira, et al., “Terabit+ (192 × 10 Gb/s) nyquist shaped UDWDM coherent PON with upstream and downstream over a 12.8 nm band,” J. Lightwave Technol. 32(4), 729–735 (2014). [CrossRef]
5. K. Kikuchi, “Fundamentals of coherent optical fiber communications,” J. Lightwave Technol. 34(1), 157–179 (2016). [CrossRef]
6. S. J. Savory, “Digital coherent optical receivers: Algorithms and subsystems,” IEEE J. Sel. Top. Quantum Electron. 16(5), 1164–1179 (2010). [CrossRef]
7. M. S. Erkilinç, D. Lavery, K. Shi, et al., “Polarization-insensitive single-balanced photodiode coherent receiver for long-reach WDM-PONs,” J. Lightwave Technol. 34(8), 2034–2041 (2016). [CrossRef]
8. I. N. Cano, A. Lerín, V. Polo, et al., “Polarization independent single-PD coherent ONU receiver with centralized scrambling in udWDM-PONs,” Eur. Conf. Opt. Commun. ECOC 1, 1–3 (2014). [CrossRef]
9. I. B. Kovacs, M. S. Faruk, and S. J. Savory, “A Minimal Coherent Receiver for 200 Gb/s/λ PON Downstream With Measured 29 dB Power Budget,” IEEE Photonics Technol. Lett. 35(5), 257–260 (2023). [CrossRef]
10. M. S. Faruk, H. Louchet, M. S. Erkılınç, et al., “DSP algorithms for recovering single-carrier Alamouti coded signals for PON applications,” Opt. Express 24(21), 24083 (2016). [CrossRef]
11. B. Schrenk and F. Karinou, “Simple laser transmitter pair as polarization-independent coherent homodyne detector,” Opt. Express 27(10), 13942 (2019). [CrossRef]
12. E. Ciaramella, “Polarization-independent receivers for low-cost coherent OOK systems,” IEEE Photonics Technol. Lett. 26(6), 548–551 (2014). [CrossRef]
13. M. Artiglia, M. Presi, F. Bottoni, et al., “Polarization-Independent Coherent Real-Time Analog Receiver for PON Access Systems,” J. Lightwave Technol. 34(8), 2027–2033 (2016). [CrossRef]
14. M. Presi, M. Artiglia, F. Bottoni, et al., “Field-Trial of a High-Budget, Filterless, λ-to-the-User, UDWDM-PON Enabled by an Innovative Class of Low-Cost Coherent Transceivers,” J. Lightwave Technol. 35(23), 5250–5259 (2017). [CrossRef]
15. M. Luo, J. Li, T. Zeng, et al., “Real-time coherent UDWDM-PON with dual-polarization transceivers in a field trial,” J. Opt. Commun. Netw. 11(2), A166–A173 (2019). [CrossRef]
16. M. Luo, D. Wu, W. Li, et al., “100 Gb/s (4 × 25 Gb/s) real-time coherent UDWDM-PON with a large power budget,” J. Opt. Commun. Netw. 12(2), A204–A213 (2020). [CrossRef]
17. J. Li, M. Luo, Z. He, et al., “Demonstration of 128 × 100-Gb / s Real-Time Coherent UDWDM- PON with > 35-dB Power Budget,” 2022 Opt. Fiber Commun. Conf. Exhib.1–3 (2022).
18. E. Ciaramella, M. Rannello, F. Bottoni, et al., “4×10 Gb/s coherent WDM-PON system over 110 km of Single Mode Fibre and with 55 dB ODN power budget,” Int. Conf. Transparent Opt. Networks 2016-Augus, 1–4 (2016).
19. M. Presi, R. Corsini, M. Artiglia, et al., “Ultra-dense WDM-PON 6.25 GHz spaced 8×1 Gb/s based on a simplified coherent-detection scheme,” Opt. Express 23(17), 22706 (2015). [CrossRef]
20. J. Zhou, L. Gan, C. Chen, et al., “8 × 10 Gb/s downstream PAM-4 transmission for cost-effective coherent WDM-PON application,” J. Lightwave Technol. 39(9), 2837–2846 (2021). [CrossRef]
21. C. Xie, P. J. Winzer, G. Raybon, et al., “Colorless coherent receiver using 3 × 3 coupler hybrids and single-ended detection,” Opt. Express 20(2), 1164–1171 (2012). [CrossRef]
22. P. J. Winzer and R. J. Essiambre, “Advanced optical modulation formats,” Optical Fiber Telecommunications VB: Systems and Networks 94(5), 23–93 (2008). [CrossRef]
23. F. Derr, “Coherent Optical QPSK Intradyne System: Concept and Digital Receiver Realization,” J. Lightwave Technol. 10(9), 1290–1296 (1992). [CrossRef]
24. K. Zhang, Q. Zhuge, H. Xin, et al., “Performance comparison of DML, EML and MZM in dispersion-unmanaged short reach transmissions with digital signal processing,” Opt. Express 26(26), 34288 (2018). [CrossRef]
25. R. Kaler and R. S. Kaler, “Investigation of four wave mixing effect at different channel spacing,” Optik 123(4), 352–356 (2012). [CrossRef]
26. J. Zhou, L. Gan, Y. Chen, et al., “Unequally spaced PAM-4 signaling enabled sensitivity enhancement of a simplified coherent receiver applied in a UDWDM-PON,” Opt. Express 30(20), 35369 (2022). [CrossRef]
