Optical independent sideband (ISB) signals can be generated by exploiting one external In-phase/Quadrature (I/Q) modulator. Our theoretical analysis shows crosstalk between the two ISB (right and left side) signals can attribute to two main imperfections: amplitude difference and phase unmatched in I/Q data. To reduce the impact of crosstalk between the two ISB signals, we propose three schemes. The first is precise phase match of the I and Q data. The second has been made possible by setting different frequencies for the left sideband (LSB) and the right sideband (RSB) signals, and the last is achieved by adding Multiple-Input Multiple-Output (MIMO) equalization digital signal processing (DSP) at the receiver side. Our experimental results have shown that these schemes can improve the performance of ISB signals. In our experimental system we designed dual ISB system with different modulation formats in two sidebands. Precise phase match can bring a ∼2.2dB improvement at BER of 1×10−2 and a ∼4.3dB improvement at BER of 1×10−3 for 16-ary quadrature-amplitude-modulation (16QAM) and quadrature-phase-shift-keying (QPSK) signals, respectively, in 4Gbaud with carrier frequency of 36GHz system. The BER of 4Gbaud 16QAM ISB signal at 30GHz and 4Gbaud QPSK ISB signal at 38GHz can reach hard-decision forward-error-correction (HD-FEC) when the input power is larger than −5.5 and −7.4dBm respectively in different frequencies system. For 4Gbaud with carrier frequency of 36GHz system, the BER of 16QAM signal and QPSK signal reduce ∼2.1 and ∼2.2dB at HD-FEC after using MIMO. In addition, MIMO can further improve the performance of the matched phase system or the system with different frequencies.
© 2020 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
Nowadays, due to the increasing popularity of different mobile broadband applications such as high-quality Internet protocol television (IPTV), virtual reality, visual call and video sharing [1,2], the demand for radio access networks (RANs) dramatically increases . Future networks are expected to provide broadband services and support mobility simultaneously , . In order to support higher data rates in the future mobile and wireless systems, radio-over-fiber (ROF) technology gets much attention because it combines the advantages of broadband of optical communication and mobility of wireless communication [6–12]. For the practical implementation of the ROF technology, it is of importance to generate optical mm-wave signals with high performance, low cost and flexible modulation formats [13,14]. By using one external modulator, we can generate single-sideband (SSB) [15–19], double-sideband (DSB) [14,20–22] and optical-carrier-suppression (OCS) [14,22].
There is a new scheme using two independent sidebands to carry two independent signals. Compared to the other scheme mentioned above, the twin-SSB system has many advantages. It nearly doubles the transmission capacity comparing with the SSB by using two sidebands, left-sideband (LSB) and right-sideband (RSB).
Many studies on this twin-SSB system were proposed. M. Chen et al. using an optical IQ modulator (IQM) to optimize optical carrier-to-signal power ratio (CSPR), which further improve receiver sensitivity of spectrally-efficient guard-band direct-detection optical orthogonal frequency-division multiplexing (OFDM) with twin single-side-band (SSB) modulation technique . R. Deng et al. mitigate the time-varying carrier frequency offset (CFO) effect and simultaneously realize down-conversion by a blind carrier recovery method . J. Shi et al. proposed a MIMO-Volterra equalizer to mitigate the interference and nonlinearity penalty of the twin-SSB signal . L. Zhang et al. experimentally demonstrated a spectrally efficient signal-signal beating interference (SSBI) cancellation system based on the guard-band twin-SSB-DMT .
In this paper, we propose a dual-ISB scheme. In the dual-ISB scheme, two sidebands carrying different information can not only have different modulation formats but also different carrier frequencies. This scheme introduces additional flexibility.
However, the dual-ISB scheme still has some shortcomings as the twin-SSB system, of which the most important is their crosstalk. Hardware imperfections, especially the unbalanced inputs of two signals in an I/Q modulator, can cause crosstalk in the signals [27–30]. For example, different drive voltages or phase mismatched could cause the two ISB signals to interfere with each other. It seriously affects the practical application of dual-ISB system. Y. Wang et al. used joint image-cancellation and nonlinearity-mitigation algorithms to solve the crosstalk and non-linear effect together .
In this work, we propose three new methods to mitigate the crosstalk between RSB and LSB. Our experimental results demonstrate that the three proposed methods effectively improve the performance of dual-ISB signals.
When there is different amplitude or relative time shift between two driving signals of an I/Q modulator, there will be a crosstalk of different sidebands.
Figure 1 depicts the scheme of the system with the situations we mention above, in which p consists the power ratio originated from both amplitude difference and time/phase shift. The crosstalk signal would bring interference in dual ISB signal system.
Assuming that the amplitude ratio of I and Q is (1-k), in which k is a small value. The output optical signal at sideband from modulator can be expressed as:
Here, we use a positive value ω to donate the sideband angle frequency of the sideband vector signal. The complex parameter A represents the transmitter data carried by sideband. A can employ single-carrier or Orthogonal Frequency Division Multiplexing (OFDM) vector quadrature-amplitude-modulation (QAM) modulation. It means that there will be a crosstalk signal at –ω after the modulation. The amplitude ratio of power of crosstalk carrier and regular carrier is k/2. The power ratio between the two received carriers is:
When a relative time shift T or unmatched phase ωT exists between two driving signals on the I/Q modulator, the output of the modulator can be expressed as:
Since T is a small value,. = 1, sin(ωT)= ωT, the generated optical signals can be expressed as:
Equation (4) indicates that a signal will be generated at -ω, and the power ratio between two received carriers is:
For ωT = 0.05, the ratio is
In this system, the two sideband signals can be express as:
Here, we use two complex parameters, X and Y, to donate the transmitted data carried by LSB and RSB vector signals, respectively. Also, X and Y can employ the same or different vector QAM modulation format.
As shown in Fig. 2, we first add the real and imaginary parts of the LSB and RSB vector signals, respectively, and then we use the real and imaginary summation to drive the I/Q modulator. I/Q modulation introduces amplitude difference or time shift between two driving signals, and the received signals could be:
To solve the crosstalk between LSB and RSB, we propose three different methods.
First, precise matching phase and amplitude of I and Q data could balance two input signals, which can reduce the crosstalk between LSB and RSB.
An obvious way to deal with this problem is adjusting phase imbalance to balance the amplitudes of the two signals, as shown in Fig. 3. The balanced I and Q signals will result in a smaller power ratio between crosstalk noise and the interfered signal.
The advantage of method 1 is its low complexity. There is no need to add more algorithms to DSP, and adjustment of the PAA is very simple. The disadvantage of method 1 is the increased cost. PAA is expensive and uncommon in optical commination systems.
Second, we could set different frequencies for LSB and RSB signals.
Figure 4 shows that, if the frequencies of two signals, ω1 and -ω2, are different, the crosstalk noise would appear at frequency of -ω1 and ω2. The crosstalk will not overlap with the received signal, and thus, the interference can be reduced. The received data can be therefore expressed as:
There will be a frequency span of ω2-ω1 between the signal and crosstalk noise. Thus, for the situations the two signals have a wide enough frequency interval, the received signal will be less affected.
The method 2 can completely resolve crosstalk. As long as the frequency interval is large enough, no image signal will appear. Thus the crosstalk could be resolved. And this method can increase the flexibility of the system by letting the two sidebands select different frequencies. However, the large frequency separation would limit the frequency selection of ISB signal.
Third, because the origin of image signals are ISB signals, a 2×2 MIMO system can cascade after the interleaver to mitigate this kind of crosstalk.
As shown at Fig. 5, after filtered by the interleaver, the two output signals enter the MIMO arithmetic unit. And then, the MIMO unit would separate the ISB signals and image signals. The updated equations of the tap coefficients (hxx) are as follows:
In this way the MIMO can successfully mitigate the crosstalk in the dual ISB system.
The method 3 can mitigate the crosstalk by adding mature MIMO algorithm to the DSP without changing existing system and redesigning signals. But, since method 3 requires two received signals to be used as the input of MIMO together, two signals could not transmit to two different locations. It limits its application scenarios.
3. Experimental setup
Figure 6 shows the experimental setup for the generation of 4Gbaud 16QAM signal and QPSK signal based on one I/Q modulator. We simultaneously generate the 16QAM signal (LSB) and QPSK signal (RSB), i.e., The output signal SE which described in Eq. (3) and Eq. (4) generated in the digital domain using MATLAB before it is uploaded to the 92-GSa/s digital-to-analog converter (DAC) with 3-dB bandwidth of 22 GHz. Here, the pseudo-random-binary- sequence (PRBS) length of the QPSK and 16QAM signals are 29−1 and 210−1, respectively. The imaginary part of SE and the real part of SE signals from the 92GSa/s DAC through phase adjustable adapters (PAA) to precise match phase. After being amplified by a 40GHz electric amplifier (EA), the I (In-phase data) and Q (Quadrature data) signals are used to drive an I/Q modulator with 30-GHz optical bandwidth.
The continuous-wavelength (CW) lightwave at 1551.528 nm is generated by an external cavity laser (ECL) with < 100-kHz linewidth and 14 dBm output power, and it is used as the optical input of the I/Q modulator. The output optical signal of the I/Q modulator is −10dBm. Here, the total optical power loss of 24dB after passing the I/Q modulator is caused by the insertion loss and the limited 30-GHz optical bandwidth of the I/Q modulator. Moreover, although the output amplitude of the 92-GSa/s DAC is up to 800 mV at 10GHz, it is only less than 100 mV at ∼40GHz. A 40GHz Erbium doped fiber amplifier (EDFA) is employed to boost the optical power to be 12dBm, and the DC-bias of I/Q modulator is adjusted carefully to ensure that the central optical carrier has a proper power which close to that of the LSB and RSB.
After 10km SSMF transmission without optical dispersion compensation, the dual ISB optical signal is divided into two parts by a 50/100GHz interleaver. 75GHz photodetectors (PDs) are used to detect the two signals, respectively. The variable optical attenuators (VOA) are used before the PDs to adjust the input power into the PD to measure BER. After amplified by linear broadband amplifiers 40GHz with 35dB gain, the signals are finally captured by a 160Ga/s real-time digital storage oscilloscope (OSC) with 65GHz bandwidth. The received data is recovered by offline DSP with heterodyne detection, including down conversion to baseband, constant modulus algorithm (CMA) for QPSK and constant modified modulus algorithm (CMMA) for 16QAM, 2×2 MIMO, frequency offset estimation, phase offset estimation, decision-directed least mean square (DD-LMS) equalization for 16QAM, and BER calculation [32–36].
Figure 7(a) demonstrates the optical spectrum of received 4Gbaud 16QAM signal at 15GHz and QPSK signal at 25GHz dual-ISB optical mm-wave signal without optimized time shifting. It clearly shows that the image signals appeared and would interfere with the signals, so the image signals will generate crosstalk noise if they are too close to the ISB signal. In fact, the image signal of right ISB signal (RSB) will affect the left ISB signal (LSB). If the frequency of the image signal of RSB is far away from the LSB, there will be no crosstalk. However, if the frequencies of image signal and the ISB are close or the same, the crosstalk will be large. We set different frequencies of the two ISB signals to make the image signals or interference noise clear. It is helpful for us to adjust the phase imbalance, otherwise, it is hard to find the interference noise and the optimized time shifting. As shown in Fig. 7(b), the image signals (interference noise) have been reduced over 10dB when the time shifting is adjusted to an optimized point. In this situation, the two ISB signals are more independent of each other. However, due to the amplitude of the driven signals and the different response of the I/Q port of the modulator, we cannot completely suppress this interference. Consequently, the image signal or crosstalk signal is still visible in Fig. 7(b). The interference signals will still degrade the performance of the ISB signals, although they are not severe.
4. Experimental result
4.1 Precise matching phase and amplitude of I and Q data
To test the first method, we tested the performance of 36GHz 16QAM signal (LSB) and QPSK signal (RSB) with and without the matched phase imbalance.
We set the two signals at the same frequency of 36GHz with the most optimized settings of PAA. Figure 8 shows the optical spectra of the 4Gbaud 16QAM and QPSK signals at 36GHz before and after optical interleaver. It is clearly seen that the two signals are separated completely by passing the optical interleaver. Because the RSB and LSB signals have the same frequency, the image signals are not visible. But we believe that the image signals should be small because we have optimized the phase imbalance.
Figure 9 gives the measured BER versus the input power into PD for 4Gbaud 16QAM and QPSK signals at 36GHz with and without the optimized phase imbalance. We can find that, both signals perform better when the phase imbalance is matched. For the QPSK signal, the matched phase imbalance can bring a ∼4.3dB improvement at BER of 1×10−3. As for the 16QAM signal, the matched phase imbalance can bring a ∼2.2dB improvement at BER of 1×10−2. It proves that a suitable phase imbalance in this system will bring a significant improvement.
4.2 Using different frequencies for LSB and RSB
To verify the second method, we adjust the QPSK signal to different frequencies while the frequency of 16QAM remains at 30GHz. Thus, the frequency span between the image signals and ISB signals can be adjusted.
Figure 10 shows the measured BER of 4GHz versus the input power into PD for 4Gbaud QPSK signal at three different frequencies. In our test interval, the 38GHz signal performs best, which can be attributed to the largest frequency span between the image signal and ISB signal in this case. So, the larger frequency span can weak the interference caused by the image signals.
The received optical spectrum of 4Gbaud 16QAM signal at 30GHz and QPSK signal at 38GHz are shown in Fig. 11. It is clearly seen that the crosstalk noise is reduced because of the large frequency span between the image signals and ISB signals. Figure 12 shows the optical spectra of the output signals, although the image signal doesn’t disappear, they cannot interfere with the ISB signals.
The measured BER versus the input power into PD for the 4Gbaud 16QAM signal at 30GHz and QPSK signal at 38 GHz are depicted in Fig. 13. In this case, QPSK signal and 16QAM signal can reach the HD-FEC threshold (3.8×10−3) with ∼ −7.4 and −5.6dBm input power. The inserts give the constellations for 4Gbaud 30GHz 16QAM signal with input power of −4dBm and 4Gbaud QPSK signal at 38GHz with input power of −6dBm, and almost no bit error is recorded within the collected data.
4.3 Using MIMO at the receiving end to mitigate crosstalk
MIMO can effectively deal with the problem of crosstalk noise at the receiver side. As shown at Fig. 14, without matched phase imbalance or frequency span, the MIMO can reduce the BER effectively. It can bring 3.8dB improvement at BER of 1×10−3 for the QPSK signal, and 1.9dB improvement at BER of 1×10−2 for the 16QAM signal. The insets compare the constellation for two signals with and without MIMO equalization. It is obvious that after MIMO, the constellation become much clearer.
Because there are still some minor image signals to get some crosstalk even the phase is matched as shown in Fig. 6(b). MIMO can also be used in the phase matched system and different frequencies with very small difference system. Figure 14 shows the effect of MIMO in the two systems we mentioned above. In Fig. 14(a), MIMO brings ∼0.6dB improvement in the 16QAM signal at BER of 1×10−2. Because QPSK is more tolerant to crosstalk compared to 16QAM, the effect of MIMO is not obvious for QPSK signal. As shown in Fig. 15(b), when the frequency span is set to 2GHz, the MIMO can bring ∼0.8dB improvement for 16QAM signal at BER of 1×10−2.
With unmatched I/Q data including phase and amplitude on the I/Q modulator, the crosstalk exists between RSB and LSB vector mm-wave signals. We have experimentally demonstrated three methods to mitigate the crosstalk. The first one is compensating the unmatched phase between two inputs of I/Q modulation by adjusting the phase imbalance. The second one is generating two single-band signals at different frequencies. The third is using MIMO DSP at the receiving end. Then we designed three experimental systems to verify these schemes. We found that the matched phase imbalance can improve the performance of 4Gbaud 36GHz 16QAM and QPSK signals at 36GHz. And 4Gbaud 16QAM signal at 30GHz and QPSK signal at 38GHz perform better comparing with the 4Gbaud 36GHz 16QAM and QPSK signals. Finally, the MIMO can further improve the performance of the dual-ISB system, which can be used in matched delay system and the same or different frequencies system.
National Natural Science Foundation of China (61571063, 61720106015, 61835002); Natural Science Foundation of Beijing Municipality (3182028).
The authors declare no conflicts of interest.
1. 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. Multimedia 19(7), 1431–1441 (2017). [CrossRef]
2. 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]
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 (2015). [CrossRef]
4. I. S. Amiri, S. E. Alavi, S. M. Idrus, A. S. M. Supaat, J. Ali, and P. P. Yupapin, “W-Band OFDM transmission for radio-over-fiber link using solitonic millimeter wave generated by MRR,” IEEE J. Quantum Electron. 50(8), 622–628 (2014). [CrossRef]
5. X. Li, J. Xiao, and J. Yu, “Long-distance wireless mm-wave signal delivery at W-band,” J. Lightwave Technol. 34(2), 661–668 (2016). [CrossRef]
6. L. Chen, J. Xiao, and J. Yu, “Application of volterra nonlinear compensation in 75-GHz mm-wave fiber-wireless system,” IEEE Photonics J. 9(1), 1–7 (2017). [CrossRef]
7. F. Li, X. Li, J. Zhang, J. Yu, and Z. Cao, “Fiber-wireless-fiber link for DFTspread OFDM signal transmission at W-band,” IEEE Photonics Technol. Lett. 27(12), 1273–1276 (2015). [CrossRef]
8. D. Wake, A. Nkansah, and N. Gomes, “Radio Over Fiber Link Design for Next Generation Wireless Systems,” J. Lightwave Technol. 28(16), 2456–2464 (2010). [CrossRef]
9. M. Sauer, A. Kobyakov, and J. George, “Radio Over Fiber for Picocellular Network Architectures,” J. Lightwave Technol. 25(11), 3301–3320 (2007). [CrossRef]
10. L. Chen, H. Wen, and S. Wen, “A Radio-Over-Fiber System With a Novel Scheme for Millimeter-Wave Generation and Wavelength Reuse for Up-Link Connection,” IEEE Photonics Technol. Lett. 18(19), 2056–2058 (2006). [CrossRef]
11. C. Browning, H. Elwan, E. Martin, S. Duill, J. Poette, P. Sheridan, A. Farhang, B. Cabon, and L. Barry, “Gain-Switched Optical Frequency Combs for Future Mobile Radio-Over-Fiber Millimeter-Wave Systems,” J. Lightwave Technol. 36(19), 4602–4610 (2018). [CrossRef]
12. A. Deb and N. Kjærgaard, “Radio-over-fiber using an optical antenna based on Rydberg states of atoms,” Appl. Phys. Lett. 112(21), 211106 (2018). [CrossRef]
13. S. E. Alavi, M. R. K. Soltanian, I. S. Amiri, M. Khalily, A. S. M. Supa’at, and H. Ahmad, “Towards 5G: A Photonic Based Millimeter Wave Signal Generation for Applying in 5G Access Fronthaul,” Sci. Rep. 6(1), 19891 (2016). [CrossRef]
14. X. Li, J. Yu, J. Zhang, J. Xiao, Z. Zhang, Y. Xu, and L. Chen, “QAM vector signal generation by optical carrier suppression and precoding techniques,” IEEE Photonics Technol. Lett. 27(18), 1977–1980 (2015). [CrossRef]
15. J. Yu, Z. Jia, L. Yi, Y. Su, G. K. Chang, and T. Wang, “Optical millimeter-wave generation or up-conversion using external modulators,” IEEE Photonics Technol. Lett. 18(1), 265–267 (2006). [CrossRef]
16. Y. Tong, C. Chow, G. Chen, C. Peng, C. Yeh, and H. Tsang, “Integrated Silicon Photonics Remote Radio Frontend (RRF) for Single-Sideband (SSB) Millimeter-Wave Radio-Over-Fiber (ROF) Systems,” IEEE Photonics J. 11(2), 1–8 (2019). [CrossRef]
17. G. H. Smith, D. Novak, and Z. Ahmed, “Technique for optical SSB generation to overcome dispersion penalties in fibre-radio systems,” Electron. Lett. 33(1), 74–75 (1997). [CrossRef]
18. J. Ma and W. Zhou, “Joint influence of the optical carrier-to-sideband ratio and guard band on direct-detection SSB-OOFDM system,” IEEE Photonics J. 7(5), 1 (2015). [CrossRef]
19. Q. Ding, M. Wang, H. Mu, Y. Tang, J. Zhang, B. Wu, and T. Li, “Full-duplex broadcast RoF-WDM-PON with self-coherent detection and photonic frequency up/down-conversion using SSB pilot-carrier,” Opt. Commun. 427(15), 54–60 (2018). [CrossRef]
20. S. Yaakob, S. M. Idrus, M. Z. Abdul Kadir, M. S. Salleh, R. Mohamad, C. B. M. Rashidi, Z. Zan, and M. A. Mahdi, “Link Budget Analysis for Dual Sideband Optical Carrier Suppression RoF System,” 2018 Progress in Electromagnetics Research Symposium (PIERS-Toyama), 1–4, DOI: 10.23919/PIERS.2018.8598031 (2018).
21. C. T. Lin, J. Chen, P. T. Shih, W. J. Jiang, and S. Chi, “Ultra-High Data-Rate 60 GHz Radio-Over-Fiber Systems Employing Optical Frequency Multiplication and OFDM Formats,” J. Lightwave Technol. 28(16), 2296–2306 (2010). [CrossRef]
22. X. Li, J. Xiao, Y. Xu, and J. Yu, “QPSK Vector Signal Generation Based on Photonic Heterodyne Beating and Optical Carrier Suppression,” IEEE Photonics J. 7(5), 1–6 (2015). [CrossRef]
23. M. Chen, M. Peng, H. Zhou, Z. Zheng, X. Tang, and L. Maivan, “Receiver sensitivity improvement in spectrally-efficient guard-band twin-SSB-OFDM using an optical IQ modulator,” Opt. Commun. 405, 259–264 (2017). [CrossRef]
24. R. Deng, J. Yu, J. He, M. Chen, Y. Wei, L. Zhao, Q. Zhang, and X. Xin, “Twin-SSB-OFDM transmission over heterodyne W-band fiber-wireless system with real-time implementable blind carrier recovery,” J. Lightwave Technol. 36(23), 5562–5572 (2018). [CrossRef]
25. J. Shi, Y. Zhou, Y. Xu, J. Zhang, J. Yu, and N. Chi, “200-Gbps DFT-S OFDM using DD-MZM-based twin-SSB with a MIMO-volterra equalizer,” IEEE Photonics Technol. Lett. 29(14), 1183–1186 (2017). [CrossRef]
26. Y. Wang, J. Yu, H. C. Chien, X. Li, and N Chi., “Transmission and direct detection of 300-Gbps DFT-S OFDM signals based on O-ISB modulation with joint image-cancellation and nonlinearity-mitigation,” ECOC 42nd Dusseldorf, Germany, 2016, 978-3-8007-4274-5
27. X. Chen and J. Yao, “Wavelength Reuse in an RoF Link Based on CS-DSB, Coherent Detection and DSP,” IEEE Photonics Technol. Lett. 29(12), 975–978 (2017). [CrossRef]
28. X. Chen, P. Dong, S. Chandrasekhar, K. Kim, B. Li, H. Chen, A. Adamiecki, A. Gnauck, and P. Winzer, “Characterization and Digital Pre-compensation of Electro-optic Crosstalk in Silicon Photonics I/Q Modulators,” ECOC 42nd Dusseldorf, Germany, 2016, 978-3-8007-4274-5
29. Z. Khan, E. Zenteno, P. Händel, and M. Isaksson, “Digital Predistortion for Joint Mitigation of I/Q Imbalance and MIMO Power Amplifier Distortion,” IEEE Trans. Microwave Theory Tech. 65(1), 322–333 (2017). [CrossRef]
30. M. Petit and A. Springer, “Analysis of a Properness-Based Blind Adaptive I/Q Filter Mismatch Compensation,” IEEE Trans. Wireless Commun. 15(1), 781–793 (2016). [CrossRef]
31. L. Zhang, T. Zuo, Q. Zhang, J. Zhou, E. Zhou, and G. N Liu, “150-Gb/s DMT over 80-km SMF transmission based on spectrally efficient SSBI cancellation using guard-band twin-SSB technique,” ECOC 42nd Dusseldorf, Germany, 2016, 978-3-8007-4274-5
32. J. Yu, “Photonics-Assisted Millimeter-Wave Wireless communication,” IEEE J. Quantum Electron. 53(6), 1–17 (2017). [CrossRef]
33. X. Li, Z. Dong, J. Yu, N. Chi, Y. Shao, and G. K. Chang, “Fiber wireless transmission system of 108-Gb/s data over 80-km fiber and 2×2 MIMO wireless links at 100GHz W-Band frequency,” Opt. Lett. 37(24), 5106–5108 (2012). [CrossRef]
34. X. Li, J. Yu, Z. Dong, J. Zhang, N. Chi, and J. Yu, “Investigation of interference in multiple-input multiple-output wireless transmission at W band for an optical wireless integration system,” Opt. Lett. 38(5), 742–744 (2013). [CrossRef]
35. X. Li, J. Yu, J. Zhang, Z. Dong, F. Li, and N. Chi, “A 400G optical wireless integration delivery system,” Opt. Express 21(16), 18812–18819 (2013). [CrossRef]
36. J. Yu, X. Li, and N. Chi, “Faster than fiber: over 100-Gb/s signal delivery in fiber wireless integration system,” Opt. Express 21(19), 22885–22904 (2013). [CrossRef]