Open Access
Add to BibSonomy
Abstract
A cost-efficient dual-band transmission scheme based on return-to-zero (RTZ)-mode digital-to-analog converter (DAC) and sub-Nyquist sampling analog-to-digital converter (ADC) is proposed in this paper. The scheme can increase the data rate and meanwhile halve the required sampling rate of the DAC/ADC in optical intensity modulation direct detection (IMDD) transmission system. Based on this scheme, we experimentally demonstrate a dual-band discrete Fourier transformation spread (DFT-S) OFDM intensity modulation direct detection (IMDD) system. Although the sampling rate of the DAC and ADC used in the receiver is only 5-GSa/s and there is no mixer and oscillator at the transmitter and receiver side, 22-km standard single-mode fiber (SSMF) transmission with up to 26.33-Gb/s data rate is successfully realized. The experimental results show that in the system, the first sub-band transmission based on 128-QAM mapping can achieve a bit error rate (BER) below 3.8 × 10−3. The second sub-band transmission based on 32-QAM mapping can achieve a BER below 2.4 × 10−2. The spectral efficiency (SE) of the first sub-band signal is up to 6.14 bit/s/Hz and the SE of the second sub-band signal is up to 4.39 bit/s/Hz.
© 2018 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
With the emergence of a variety of network applications, such as high-quality Internet protocol television (IPTV), video sharing, virtual reality, cloud computing, and peer-to-peer multimedia services [1,2], the end-users’ demands for bandwidth increase dramatically [3]. In the optical access networks, today’s standardized fastest time/wavelength division multiplexing-passive optical network (TWDM-PON) which supports 10 Gb/s lane rate non-return-to zero (NRZ) is expected to support beyond 20 Gb/s lane rate in the future [4–6]. To address the upgrade issue, the scheme of intensity modulation direct detection (IMDD) with advanced modulation format has been proposed and is studied intensively in the recent years. At present, it has been regarded as one of the most promising solution schemes for optical access network due to its low complexity, high stability and high spectrum utilization efficiency [7–10].
Since advanced modulation formats, such as carrier-less amplitude and phase modulation (CAP) [11,12] and orthogonal frequency-division multiplexing (OFDM) [13–17], etc., are enabled by efficient digital signal processing (DSP), in general, digital-to-analog converter (DAC)/analog-to-digital converter (ADC) is essential in corresponding optical IMDD system. The conventional implementation scheme of the beyond 20 Gb/s optical IMDD system with advanced modulation format is not cost-efficient, because the high-sampling-rate DAC/ADC is not cheap and normally needs to match a DSP chip with high performance, which limits the industrialization of the optical IMDD system with advanced modulation format at the optical access level. Fortunately, there are some effective solution schemes to increase the data rate and meanwhile decrease the requirement of the ADC's sampling rate for optical IMDD transmission system. For example, in [18–21], R. P. Giddings et al. proposed and demonstrated some dual-band transmission schemes to increase the data rate and meanwhile decrease the requirement of the DAC/ADC's sampling rate in optical OFDM IMDD system. However, in these schemes, mixer and precise oscillator should be equipped for up-conversion/down-conversion at the transmitter/receiver side. It would lead to an increase in the system complexity and the maintenance difficulty. In [22], C. H. Lin et al. proposed and demonstrated a cost-efficient receiving scheme based on sub-Nyquist sampling for high-speed optical IMDD OFDM transmission. However, the demonstration in [22] is based on high-speed digital storage oscilloscope (DSO). It doesn’t consider the factor of the serious bandwidth limitation and high-frequency fading effect of the low-sampling-rate ADC in practical applications. The factor would seriously limit the performance of the transmission of the system, especially the performance of the transmission of the high-frequency subcarriers. More importantly, the proposed scheme in [22] is different from the dual-band/multi-band scheme and still requires high-sampling-rate DAC.
In this paper, we proposed a dual-band transmission scheme based on return-to-zero (RTZ) DAC and sub-Nyquist sampling ADC for optical IMDD system without mixer and precise oscillator. It can effectively halve the required sampling rate of the optical IMDD system and thereby reduce the implementation cost. Based on the scheme, we experimentally demonstrated a 26.33 Gb/s dual-band discrete Fourier transformation spread (DFT-S) OFDM transmission over a directly modulated laser (DML)-based IMDD system using 5-GSa/s DAC and 5-GSa/s ADC. The experimental results show that after 22-km SSMF transmission, the system can achieve a BER at blow 3.8 × 10−3 over the first sub-band and a BER at blow 2.4 × 10−2 over the second sub-band.
2. The principle of the proposed scheme
The proposed dual-band transmission scheme is shown in Fig. 1. At the transmitter side, two digital signals are firstly generated by digital modulation (CAP or OFDM). One of the two digital signals is sent to a NRZ-mode DAC with sampling rate at fs and another one is sent to an RTZ-mode DAC with the same sampling rate fs. Then, two analog signals can be generated correspondingly. The generated analog signal from the NRZ-mode DAC is filtered by an analog low-pass filter (LPF) with fs/2 cut-off frequency, and the generated analog signal from the RTZ-mode DAC is filtered by an analog high-pass filter (HPF) with also fs/2 cut-off frequency. After that, the two sub-band signals can be obtained and the frequency range of the two sub-band signals are from 0 to fs/2 and from fs/2 to fs, respectively. Then, the dual-band signal can be generated by combining the two filtered signals via an electrical coupler and it can be further used to driver the optical modulator for E/O conversion. At the receiver side, the optical signal is received by a photodetector (PD) for O/E conversion and the electrical dual-band signal can be recovered correspondingly. Subsequently, the recovered dual-band signal is sent to a single-ended to differential amplifier. In this way, two signals with opposite polarity but with the same information as the recovered dual-band signal can be generated correspondingly. One of them is filtered by an LPF and another one is filtered by an HPF. The LPF and HPF are identical to those at the transmitter side. As thus, the two sub-band signals can be extracted from the dual-band signal. And by using two identical ADCs with sampling rate at fs, the two sub-band signals can be captured respectively. In detail, the first low-frequency sub-band signal with frequency range from 0 to fs/2 can be captured by direct sampling of the ADC. However, due to that the frequency range of the second high-frequency sub-band signal is from fs/2 to fs and the first available Nyquist zone of the fs sampling rate ADC is limited at 0 ~fs/2, 1/2 sub-Nyquist sampling (also known as undersampling or super-Nyquist sampling) should be used to capture the second high-frequency sub-band signal. Eventually, after analog-to-digital conversion, two digital signals can be obtained and used for demodulation and data recovery. RTZ-mode DAC and ADC with undersampling are the key devices to enable the scheme. In the following part, the principle of the scheme enabled by the RTZ-mode DAC and undersampling ADC will be discussed.
Fig. 1 The diagram of the proposed dual-band transmission scheme for optical IMDD system with advanced modulation format.
At present, most of commercial DACs can be operated at NRZ mode or RTZ mode [23]. And common DAC is operated at NRZ-mode by default. In NRZ mode, the impulse response of the DACs is the NRZ impulse response as shown in Fig. 2(a). The sample period is identical with the time Ts (Ts = 1/fs, fs is the sampling rate of the ADC). Hence, ideally, the frequency response of the NRZ-mode DAC’s can be expressed as
where fout is the DAC output frequency and A0 is the amplitude factor. Nevertheless, quite a part of DACs can also support RTZ mode. In return-to-zero (RTZ) mode, the DAC output amplitude is zero for 50% of the time Ts as shown in Fig. 2(b). Then the frequency response of the RTZ-mode DAC can be expressed asFigure 2(c) shows the normalized frequency response of the NRZ-mode DAC and RTZ-mode DAC according to Eqs. (1) and (2). It can be seen that RTZ mode can effectively extend the bandwidth of the DAC from 0 ~fs/2 to 0 ~fs compared with the NRZ mode. Thus, the RTZ DAC is able to generate the second sub-band signal with frequency range from fs/2 to fs in the proposed scheme.Fig. 2 (a) The impulse response of the NRZ-mode DAC; (b) the impulse response of the RTZ-mode DAC; (c) the frequency response of the RTZ-mode DAC and NRZ-mode DAC.
From the prospective of the time domain, in fact, the signal generated by RTZ-mode DAC can be equivalent to the signal generated by NRZ-mode DAC with inserting zeros as shown in Fig. 3(a). In theory, in the operation of zero insertion, the zero-inserted signal xz(n) is obtained by inserting a zero between every two samples of a discrete-time signal x(n). Hence, xz(n) can be expressed as
The spectrum of the zero-inserted signal is related to the spectrum of the discrete-time signal X(f) [24] byEquation (4) illustrates that the spectrum of the zero-inserted signal Xz(f) is a frequency-scaled version of the spectrum of the original signal X(f). The base-band spectrum of the zero-inserted signal is composed of 2 repetitions of the based band spectrum of the original discrete-time signal. In other words, the signal after inserting zeroes carries two copies of the frequency-domain information of the original discrete-time signal.Fig. 3 (a) The equivalent model of the signal shaping process of the RTZ-mode DAC; (2) the spectrum evolution of the proposed dual-band transmission scheme.
With filtering via appropriate high pass filter (normalized cut-off frequency is about 0.5), one of the two copies in the low-frequency part can be removed and the high-frequency part can be still used for signal demodulation. This is the principle of the second sub-band signal generation in the proposed scheme. In signal processing, undersampling (1/2 sub-Nyquist sampling) is a technique where one samples a bandpass-filtered signal at a sample rate below its Nyquist rate (twice the upper cutoff frequency), but is still able to reconstruct the signal. Consider that the undersampled signal yu(m) which is obtained by undersampling a discrete-time signal y(m) can be expressed as y(2m), the spectrum of the undersampled signal will be related to the spectrum of the original discrete-time signal Y(f) [24] by
Obviously, by making a comparison between Eqs. (4) and (5), it can be found that the principle of undersampling is the opposite of the principle of the signal shaping process of the RTZ-mode DAC. Hence, the spectrum evolution of the second sub-band signal can be illustrated as shown in Fig. 3(b). This is the principle of the generation and reception of the second sub-band signal in the proposed dual-band transmission scheme.3. Experimental setup
Based on the proposed dual-band transmission scheme, a dual-band DFT-S OFDM IMDD transmission system is experimentally demonstrated. The experimental setup of the transmission system is established as shown in Fig. 4. At the transmitter side, a 10-bit arbitrary waveform generator (AWG) (Tektronix 7122C) integrated with NRZ-mode DAC is set at 5 GSa/s and used as the generator of the first sub-band DFT-S OFDM signal. And an FPGA equipped with a 12-bit@5-GSa/s RTZ-mode DAC are used as the generator of the second sub-band DFT-S OFDM signal. The DSP flowchart for generating the two sub-band OFDM signals is shown in Fig. 4(a). The DSP flowchart can be divided into 7 stages as follows: (1) pseudo-random binary sequence (PRBS) generation; (2) QAM mapping. Noted that, 128-QAM mapping is applied for the generation of the first sub-band signal and 32-QAM mapping is applied for the generation of the second sub-band signal; (3) 234-point DFT; (4) Hermitian symmetry, which is used to realize the generation of the real-valued OFDM signal; (5) 512-point inverse fast Fourier transform (IFFT); (6) 16-point cyclic prefix (CP) addition; (7) clipping, the clipping ratio is 11 dB; (8) training symbol (TS) insertion. The detailed parameters of the two sub-band OFDM generation is shown in Table 1.
Fig. 4 The experimental setup of the dual-band DFT-S OFDM transmission system with the proposed scheme. (a) The DSP flowchart of the DFT-S OFDM transmitter; (b) The DSP flowchart of the DFT-S OFDM receiver. Synchro.: Synchronization. CP remov.: CP removal; Channel Est.: Channel Estimation; Channel Eq.: Channel Equalization; BER ana.: BER analyzer.
Table 1. The parameters for the DFT-S OFDM generation
Then, the generated signal from the AWG is filtered by an analog low-pass filter (LPF) (Mini-Circuit VLF-2250) and the generated signal from the FPGA equipped with a 12-bit DAC is filtered by an analog high-pass filter (HPF) (Mini-Circuit VHF-2700A+). By using an electrical coupler, the two filtered signals are combined to generate the dual-band signal. After that, the dual-band signal is amplified by an electrical amplifier (EA-1, 13-dB gain). The peak-to-peak voltage of the amplified dual-band signal is about 1.8 V. Then, the amplified signal is used to driver a DML for E/O conversion. The central wavelength of the DML is about 1559.2 nm and the modulation bandwidth of the DML is 10 GHz. With 22-km SSMF transmission, an erbium-doped optical fiber amplifier (EDFA) and an adjustable optical attenuator (ATT) are used to control the power of the received optical signal. Then, a PD is used to receive the optical signal for O/E conversion. In such a way, the electrical dual-band signal can be recovered by the PD. Afterwards, a single-ended to differential electrical amplifier (EA-2, 6-dB gain) is used to amplify the electrical dual-band signal. The output of the amplifier is two differential electrical signals. One of them is filtered by an LPF (Mini-Circuit VLF-2250) with a direct current (DC) block and the first sub-band signal will be recovered. The other one is filtered by an HPF (Mini-Circuit VHF-2700A + ) and the second sub-band signal will be recovered. Eventually, the two sub-band signals are captured by two receivers which consist of two VC707 FPGA boards and two 10-bit@5-GSa/s ADCs. The captured signals are then uploaded to a personal computer for demodulation and BER analyzer by offline processing. The DSP flowchart of the OFDM demodulation is shown in Fig. 4(b). It consists of the following steps: (1) Timing Synchronization; (2) CP removal; (3) 512-point FFT; (4) channel estimation, the estimation method is based on intra-symbol frequency averaging (ISFA) [25,26]; (5) channel equalization; (6) 234-point IDFT; (7) QAM demapping; (8) BER analyzer.
4. Experimental results and discussions
In the experiment, the practical power spectral density (PSD) evolution of the signal at the transmitter side is firstly recorded by a 12.5-GSa/s DSO and it is shown in Figs. 5(a)–5(e). Figure 5(a) shows the normalized PSD of the signal generated by the NRZ-mode DAC of the AWG. Figure 5(b) shows the normalized PSD of the signal generated by the RTZ-mode DAC. Obviously, the signal generated by the 5-GSa/s RTZ-mode DAC suffers from a lower power attenuation in the second Nyquist zone (2.5 ~5 GHz). Figure 5(c) shows the normalized PSD of the first sub-band OFDM signal which is generated by filtering the signal from the NRZ-mode DAC via the analog LPF. Figure 5(d) shows the normalized PSD of the second sub-band OFDM signal which is generated by filtering the signal from the RTZ-mode DAC via the analog HPF. After using electrical coupler, the two sub-band signals will be merged and the normalized PSD of the merged dual-band signal is shown in Fig. 5(e). At the receiver side, the PSD evolution of the signal is recorded by the DSO and the 5-GSa/s ADC and the results are shown in Figs. 5(f)–5(j). Figure 5(f) shows the normalized PSD of the received dual-band signal. Figures 5(g) and 5(h) show the normalized PSD of the received first sub-band signal and second sub-band signal, respectively. Then, with 5-GSa/s ADCs, the two sub-band signals are captured and the normalized PSD of the captured signals are shown in Figs. 5(i) and 5(j). As shown in the figures, there is a clock noise which is located at 1.25 GHz, and it will affect the performance of the two sub-band transmission. Hence, to mitigate the effect, the 128th sub-carrier of the two sub-band OFDM signals is set as null sub-carrier as shown in Tab. 1.
Moreover, by adjusting the ATT to change the received optical power (ROP), the BER performance of the system is measured and the result is shown in Fig. 6. When the ROP is beyond about −2.5 dBm, the BER of the first sub-band transmission is less than 3.8 × 10−3, the hard decision forward error correction (HD-FEC) limit (7% overhead). When the ROP is beyond about −3 dBm, the BER of the first sub-band transmission is less than 2.4 × 10−2, the soft decision forward error correction (SD-FEC) limit (20% overhead). Meanwhile, when the ROP is 2 dBm, the received constellations of the two sub-band signals are recorded and the constellation samples are shown in Figs. 6(b) and 6(c). Although the QAM level of the first sub-band is much higher than that of the second sub-band, the transmission performance of the first sub-band transmission is better than that of the second sub-band transmission as shown in Fig. 6(a). On the one hand, although the 5-GSa/s RTZ-mode DAC can effectively extend the bandwidth from 0~2.5 GHz to 0~5 GHz, the second Nyquist zone from 2.5~5 GHz still suffer from relatively high attenuation compared with the first Nyquist zone from 0~2.5 GHz, thus affecting the performance of the second sub-band transmission. On the other hand, because the second sub-band signal is captured by undersampling, it will also suffer from a relatively high attenuation due to the bandwidth limitation and high-frequency fading effect of the ADC. This is the reason that the first sub-band transmission outperforms the second sub-band transmission. Nevertheless, by aggregating the BER of the two sub-band transmission, it can be found that the overall system can still achieve a relatively low BER as shown in Fig. 6(a). The data rate of the first sub-band and the second sub-band are up to about 15.36 Gb/s (234/256*2.5*7*512/(512 + 16)*99/100) and about 10.97 Gb/s (234/256*2.5*5*512/(512 + 16)*99/100), respectively. The spectral efficiency (SE) of the first sub-band is about 6.14 bit/s/Hz (15.36/2.5) and the SE of the second sub-band is about 4.39 bit/s/Hz (10.97/2.5). Hence, the data rate of the overall system is about 26.33 Gb/s. After considering the overhead of the FEC, the system can still achieve a data rate of 23.06 Gb/s (15.36*93% + 10.97*80%).
Fig. 6 (a) The BER performance of the system versus ROP; (b) the recorded constellation sample of the first sub-band DFT-S OFDM signal when the ROP is 2 dBm; (c) the recorded constellation sample of the second sub-band DFT-S OFDM signal when the ROP is 2 dBm.
5. Conclusion
In this paper, a dual-band transmission scheme is proposed and it can be used to halve the required sampling rate of the ADC/DAC and meanwhile increase the data rate in optical IMDD system with advanced modulation format. Based on the scheme, a 26.33 Gb/s dual-band DFT-S OFDM IMDD system is experimentally demonstrated in this paper. In the experiment, with 22-km SSMF transmission, the first sub-band and the second sub-band can achieve a BER of below than HD-FEC limit and SD-FEC limit, respectively, when the ROP is beyond about −2.5 dBm. It preliminarily verifies the effectiveness and feasibility of the proposed dual-band transmission scheme.
Funding
National Natural Science Foundation of China (61775054); National Natural Science Foundation of China (61377079); Science and Technology Project of Hunan Province (2016GK2011).
References and links
1. J. Mei, K. Li, A. Ouyang, and K. Li, “A Profit Maximization Scheme with Guaranteed Quality of Service in Cloud Computing,” IEEE Trans. Comput. 64(11), 3064–3078 (2015). [CrossRef]
2. Y. Li, G. Yang, Y. Zhu, X. Ding, and X. Sun, “Unimodal Stopping Model-Based Early SKIP Mode Decision for High-Efficiency Video Coding,” IEEE Trans. Multimed. 19(7), 1431–1441 (2017). [CrossRef]
3. Y. Chen, R. Hu, Q. Yang, M. Luo, S. Yu, and W. Li, “Two orthogonal carriers assisted 101-Gb/s dual-band DDO-OFDM transmission over 320-km SSMF,” Opt. Express 23(9), 12065–12071 (2015). [CrossRef] [PubMed]
4. J. L. Wei, K. Grobe, C. Wagner, E. Giacoumidis and H. Griesser, “40 Gb/s lane rate NG-PON using electrical/optical duobinary, PAM-4 and low complex equalizations,” in proceeding of OFC 2016.
5. X. Li, S. Zhou, F. Gao, M. Luo, Q. Yang, Q. Mo, Y. Yu and S. Fu, “4×28 Gb/s PAM4 long-reach PON using low complexity nonlinear compensation,” in proceeding of OFC 2017.
6. J. Yu, M. F. Huang, D. Qian, L. Chen, and G.-K. Chang, “Centralized Lightwave WDM-PON Employing 16-QAM Intensity Modulated OFDM Downstream and OOK Modulated Upstream Signals,” IEEE Photonics Technol. Lett. 20(18), 1545–1547 (2008). [CrossRef]
7. M. Chen, J. He, Q. Fan, Z. Dong, and L. Chen, “Experimental Demonstration of Real-Time High-Level QAM-Encoded Direct-Detection Optical OFDM Systems,” J. Lightwave Technol. 33(22), 4632–4639 (2015). [CrossRef]
8. M. Chen, J. He, and L. Chen, “Real-time optical OFDM long-reach PON system over 100 km SSMF using a directly modulated DFB laser,” J. Opt. Commun. Netw. 6(1), 18–25 (2014). [CrossRef]
9. J. Shi, J. Zhang, N. Chi, and J. Yu, “Comparison of 100G PAM-8, CAP-64 and DFT-S OFDM with a bandwidth-limited direct-detection receiver,” Opt. Express 25(26), 32254–32262 (2017). [CrossRef]
10. J. He, L. Shi, L. Deng, M. Cheng, M. Tang, S. Fu, M. Zhang, P. P. Shum, and D. Liu, “Novel design of N-dimensional CAP filters for 10 Gb/s CAP-PON system,” Opt. Lett. 40(10), 2409–2412 (2015). [CrossRef] [PubMed]
11. L. Shi, D. Li, J. He, L. Deng, M. Cheng, M. Tang, S. Fu, M. Zhang, P. P. Shum, and D. Liu, “Experimental demonstration of a 10 Gb/s non-orthogonal multi-dimensional CAP-PON system based on the ISI and CCI cancellation algorithm,” Opt. Lett. 41(17), 3988–3991 (2016). [CrossRef] [PubMed]
12. J. Zhang, J. Yu, F. Li, N. Chi, Z. Dong, and X. Li, “11 × 5 × 9.3Gb/s WDM-CAP-PON based on optical single-side band multi-level multi-band carrier-less amplitude and phase modulation with direct detection,” Opt. Express 21(16), 18842–18848 (2013). [CrossRef] [PubMed]
13. J. Zhou and Y. Qiao, “Low-peak-to-average power ratio and low-complexity asymmetrically clipped optical orthogonal frequency-division multiplexing uplink transmission scheme for long-reach passive optical network,” Opt. Lett. 40(17), 4034–4037 (2015). [CrossRef] [PubMed]
14. Y. Xiao, Z. Wang, J. Cao, R. Deng, Y. Liu, J. He, and L. Chen, “Time–Frequency Domain Encryption With SLM Scheme for Physical-Layer Security in an OFDM-PON System,” J. Opt. Commun. Netw. 10(1), 46–51 (2018). [CrossRef]
15. Y. Qin and J. Zhang, “Novel toggle-rate based energy-efficient scheme for heavy load real-time IM-DD OFDM-PON with ONU LLID identification in time-domain using amplitude decision,” Opt. Express 25(14), 16771–16782 (2017). [CrossRef] [PubMed]
16. Q. W. Zhang, E. Hugues-Salas, Y. Ling, H. B. Zhang, R. P. Giddings, J. J. Zhang, M. Wang, and J. M. Tang, “Record-high and robust 17.125 Gb/s gross-rate over 25 km SSMF transmissions of real-time dual-band optical OFDM signals directly modulated by 1 GHz RSOAs,” Opt. Express 22(6), 6339–6348 (2014). [CrossRef] [PubMed]
17. C. C. Wei, H. C. Liu, C. T. Lin, and S. Chi, “Analog-to-Digital Conversion Using Sub-Nyquist Sampling Rate in Flexible Delay-Division Multiplexing OFDMA PONs,” J. Lightwave Technol. 34(10), 2381–2390 (2016). [CrossRef]
18. R. P. Giddings, E. Hugues-Salas, S. Ben-Ezra and J. M. Tang, “First experimental demonstration of real-time dual-band optical OFDM transmission at 17.5Gb/s over an EML-based 25km SSMF IMDD system using 4GS/s DAC/ADCs,” in proceeding of ECOC 2012.
19. R. P. Giddings, E. Hugues-Salas, and J. M. Tang, “Experimental demonstration of record high 19.125 Gb/s real-time end-to-end dual-band optical OFDM transmission over 25 km SMF in a simple EML-based IMDD system,” Opt. Express 20(18), 20666–20679 (2012). [CrossRef] [PubMed]
20. Q. W. Zhang, E. Hugues-Salas, R. P. Giddings, M. Wang, and J. M. Tang, “Experimental demonstrations of record high REAM intensity modulator-enabled 19.25Gb/s real-time end-to-end dual-band optical OFDM colorless transmissions over 25km SSMF IMDD systems,” Opt. Express 21(7), 9167–9179 (2013). [CrossRef] [PubMed]
21. Q. W. Zhang, E. Hugues-Salas, Y. Ling, H. B. Zhang, R. P. Giddings, J. J. Zhang, M. Wang, and J. M. Tang, “Record-high and robust 17.125 Gb/s gross-rate over 25 km SSMF transmissions of real-time dual-band optical OFDM signals directly modulated by 1 GHz RSOAs,” Opt. Express 22(6), 6339–6348 (2014). [CrossRef] [PubMed]
22. C. H. Lin, C. T. Lin, C. C. Wei, and S. Chi, “DFT/IDFT-free receiving scheme for spread-OFDM signals employing low-sampling-rate ADCs,” Opt. Express 25(22), 27750–27757 (2017). [CrossRef] [PubMed]
23. K. Ajay and O. Geir, “High-Speed DACs ease transmitter designs,” Microw. RF 49(8), 66–71 (2010).
24. R. E. Crochiere and L. R. Rabiner, “Interpolation and decimation of digital signals—A tutorial review,” Proc. IEEE 69(3), 300–331 (1981). [CrossRef]
25. Q. Yang, N. Kaneda, X. Liu, and W. Shieh, “Demonstration of Frequency-Domain Averaging Based Channel Estimation for 40-Gb/s CO-OFDM With High PMD,” IEEE Photonics Technol. Lett. 21(20), 1544–1546 (2009). [CrossRef]
26. X. Liu and F. Buchali, “Intra-symbol frequency-domain averaging based channel estimation for coherent optical OFDM,” Opt. Express 16(26), 21944–21957 (2008). [CrossRef] [PubMed]
