Open Access
Add to BibSonomy
Abstract
The 5G mobile communication system provides ultrareliable, low-latency communications at up to 10 Gbps. However, the scale and power consumption of 5G is tremendous owing to a large number of antenna drivers required by the massive multiple-input multiple-output technique. The 6G system will require an architectural paradigm shift to resolve this problem. In this study, we propose an analog RoF downlink scheme for 6G wireless communications. The upcoming oversized base station problem is solved using photonics techniques. The antennas are driven together within the optical domain at a centralized station. The proposed system uses orbital angular momentum (OAM) beams as the generated space-division-multiplexing beams. An RF-OAM beam has a weak coupling effect between different modes, which will dramatically decrease the complexity of the signal processing. In our proof-of-concept experiment, the generated RF-OAM beam was shown to carry a 2-Gbaud OOK/BPSK signal in the Ku-band. Signals were transmitted over a 19.4-km RoF link without dispersion-induced power fading. In addition, by switching the OAM beams, a two-dimensional direction scanning was achieved.
© 2021 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
At the end of 2019, 5G networks began being deployed commercially in numerous countries around the world. A 5G network follows the technological path of previous mobile communication systems and reaches a high downlink speed of up to 10 Gbps, providing ultrareliable, low-latency communications services. Compared with 4G, 5G uses the massive multiple-input, multiple-output (MIMO) technique in dense networks with smaller cell sizes and more antennas. To support the massive MIMO approach, the system applied in the base station (BS) will inevitably be oversized and consume vast amounts of electricity because the BS contains a plethora of transceivers, antennas, and antenna drivers [1]. The amount of mobile traffic has reported an explosive growth, driven by smartphones, tablets, and video streaming [2]. The next-generation wireless communication, 6G, will be released soon. The 6G system is aimed at providing a downlink data rate of 100Gbps–1Tbps, and a latency of 1–10μs at a frequency of higher than 6GHz. However, the BS system size will force 6G to incur an architectural paradigm shift. An extremely promising candidate for a 6G system architecture is a photonic-based analog RoF link. A photonics technique features a compact size, large bandwidth, low insertion loss, and immunity to electromagnetic interference, and is expected to prevail over other designs for use in future wireless communication systems.
Owing to the demanding downlink data rate upgrade of 6G, the degree of space division multiplexing (SDM) will be much higher than in a 5G system, resulting in higher complexity of the massive MIMO processing. SDM beams with a weak coupling effect between different modes are suitable for this scenario because they have a low complexity requirement for MIMO processing [3]. Orbital angular momentum (OAM) is a fundamental physical property of electromagnetic (EM), and it’s related to the spatial distribution. An OAM-carrying beam has a unique spiral spatial phase term $\exp ({iL\theta } )$, where L is the so-called topological charge (L = 0, ±1, ±2, …), and $\theta $ is the azimuthal angle. This phase front generates a helical phase structure with a phase singularity in the center and a “doughnut”-shaped intensity as it propagates [3]. OAM beams with different modes are orthogonal to each other, and owing to this theoretical orthogonality, OAM mode-division-multiplexing (MDM) is categorized as an SDM with a weak coupling effect [4]. In addition, the theoretical orthogonality of OAM in different eigenstates makes it a new degree of freedom for communication. The OAM-MDM can significantly increase the data capacity and spectral efficiency in communication system [5–7]. Therefore, an OAM beam is a competitive candidate for a 6G radio waveform scheme.
To realize the 6G system architecture based on the OAM beam, the radio frequency (RF)-OAM beam generation system based on microwave photonics technology need to be constructed. And for the flexibility and security of this system, RF-OAM beams with different mode and radiation direction can be transmitted flexibly, so the steering and switching of OAM beams also need to be researched. Steering OAM beams can help to accurately align the transmitter and receiver [8], and the fast switching of OAM beam with different mode or radiation direction can provide secure communications [9]. To generate an OAM-carrying beam, spatial modulation to a plane beam is the first way to be considered. Based on that, lots of schemes have been proposed, such as spatial phase plate (SPP), Q-plates, geometric methods, and the circular antenna array (CAA) [10–14]. Considering the OAM mode-flexibility and the beam radiation direction-adjustability, the CAA with the antenna array structure can provide great convenience for our system. With the CAA, OAM beams with any mode if the mode $|L |< N/2$ can be generated by the same system, where the N is the number of antennas in the CAA. With the OAM beam steering theory based on the CAA, OAM beams can be steered to different directions without moving the device [15].
In this paper, we propose an analog RoF downlink scheme for 6G wireless communications. RF-OAM beams are generated, transmitted, steered, and switched for a system demonstration. The proposed analog RoF architecture has a detailed realization in terms of the centralized station (CS) and BS. A CS contains a complex optically controlled phase shift array and WDM signal sources. A BS contains only a simple CAA and DWDM devices. Owing to a specially designed double-sideband (DSB) modulation, modulated optical signals can be transmitted over an arbitrary length of fiber without the chromatic dispersion (CD)-induced power fading problem, with the intrinsic advantages of polarization independence, a high output power, and a possible frequency multiplication. Then a 2-Gbaud OOK/BPSK signal is successfully transmitted via the proposed RoF link, which marks the first development of an RoF communication system for RF-OAM beams with signal transmission in the Ku-band (12–18GHz) over 19.4km. Experimental verification of the above functions was achieved, and the results agree well with theory.
2. Principles
Figure 1 shows the framework of the proposed analog RoF downlink for a 6G system as compared with 5G. The red line indicates the optical fibers, and the blue line indicates the electrical cables. The 5G base station connects directly to the core network and detects the optical signal from the core network as the baseband electrical signal. The massive MIMO technique supports beam forming and a MIMO signal transmission by controlling dozens or hundreds of antennas. Each antenna requires an independent phase shifter and an electrical amplifier, resulting in a tremendous framework size and power consumption for the base station. The 6G RoF downlink uses micro-photonics techniques. The signal up-converter can be moved from the BS to the CS, and a large number of antenna drivers can be also moved to the CS and replaced by a cheap single optical spectrum processor (OSP) such as a WaveShaper. The link framework size is dramatically decreased. Furthermore, the BS contains only a wavelength division de-multiplexer, antennas, and the corresponding photodetectors.
Fig. 1. Framework of 5G system downlink and proposed analog RoF link for 6G system. (OSP: optical spectrum processor)
Following the framework shown in Fig. 1, the analog RoF link is constructed, as seen in Fig. 2. This link consists of three modules, i.e., the CS, BS, and receiving terminal. The CS part includes a WDM optical carrier selection in the block shown in Fig. 3(a), a special DSB modulation with the benefit of fiber dispersion immunity in the block shown in Fig. 3(b), and a programmable WDM phase shifter array in the block shown in Fig. 3(c). In the BS part, an erbium doped fiber amplifier (EDFA) is employed to lift the optical power, several photoelectric detectors (PDs) are used to recover the RF signals and a CAA is arranged to generate the RF-OAM beam. The receiving terminal is consisting of two receiving antennas, a vector network analyzer (VNA) and an oscilloscope, by which waveforms and signals are received and analyzed.
Fig. 2. Proposed analog RoF link for 6G wireless communication. The system transmits, steers, and receives microwave photonic RF-OAM beams.
There are several functions can be realized successively. The first is the generation, steering and switching of RF-OAM beams, and this process is signed by the red solid lines in Fig. 2. The second is the transmission and detection of broadband signals, and this process is signed by the red dotted lines in Fig. 2.
First, let’s define three states for the optical switch: state “1” respects connection to the wavelength-division-multiplexing (WDM) Laser Group 1, state “2” respects connection to the WDM Laser Group 2, and state “3” is the state of automatically switching between state “1” and “2”. Set the optical switch in state “1”, and optical signals containing N lights with different wavelengths radiated from the WDM Laser Group 1 are routed to the optical port of the polarization-multiplexing dual-drive Mach-Zehnder modulator (PM-DMZM). After that, an broadband RF signal generated by the VNA is DSB-modulated onto the optical signal. The modulated signal is then sent to a WDM phase shifter array, where the designed phase shifts are added by wavelength of carrier. The intensities of modulated signals and designed added phase shifts are depicted in the spectrum chart in Fig. 2, in which the red arrows respect intensities and the blue short lines respect the added phase shifts.
The signal after shifting phase is routed to a long-distance standard single mode fiber (SSMF) which connecting the CS and BS. As we all know, the SMF will cause the CD-induced power fading problem when transmitting a DSB signal. To overcome this problem, a specific DSB modulation scheme is proposed.
As seen in Fig. 2(b), the RF signal is evenly divided into two signals, and these two signals are routed to two different ports of the PM-DMZM. Signals along two orthogonal polarizations correspond to a pair of orthogonal trigonometric functions owing to the 90° electrical phase shifter (EPS) before the PM-DMZM. When the signals are detected in the PD, their CD-induced amplitude terms can be transferred into the phase according to the trigonometric formula, making the output signal immune to the CD-induced power fading [16]. Supposing the wavelength of the ${n^{th}}$ light in the optical signal radiated by Laser Group 1 is ${\omega _n}$, the wavelength of the RF signal is, and the designed phase shifts are added to the signals to obtain a fixed phase shift ${\pm} {\varphi _n}$ on the upper or lower sidebands, then the optical fields along the polarization-multiplexed x- and y-directions after the CS can be expressed as:
In the BS, after power-amplified by EDFA, the optical signal is demultiplexed by the de-wavelength division multiplexer (DWDM). Ignoring the CD term and the frequency-doubled terms, the photo-current of each channel in(t) demodulated by the photodetector (PD) is
The CD-induced power fading terms are transferred to the phases of the photo-current. Therefore, a flat response of the power to the frequency is achieved.
So, N broadband RF signals without CD-induced power fading and with designed shifted phases are recovered. Route them to corresponding antennas in CAA, and broadband RF-OAM beam with designed mode can be generated. In the receiving terminal, a removable receiving antenna is arranged in a certain distance to record signals, and send them to VNA to analyzer and get the waveform.
As mentioned above, the usage of the CAA to generate OAM beams can provide mode and radiation direction flexibility. As seen in Fig. 3, a CAA containing N antenna elements evenly distributed among a circle. Define the center of CAA as the original point, and the location of the nth antenna element is ${\theta _n}\textrm{ = }2\pi (n - 1)/N$. To generate an OAM beam with mode L, N equal-amplitude signals with different phase should be routed to corresponding antennas of the CAA, and the difference between phases of signals touted to adjacent antennas should be equal and is described as $\Delta \varphi \textrm{ = }2\pi L/N$. So, to generate an OAM beam with mode L, the phase fed to nth antenna element should be ${\varphi _n}\textrm{ = }2\pi (n - 1)L/N$. The OAM beam steering also relies on the phase difference of the CAA. A variable steering angle can be realized by feeding signals with designed phases to antennas of the CAA. To realize the steering angle of ${\theta _{}}$, the phase of signal fed to the ${n^{th}}$ antenna should be:
Fig. 4. The simulated diagram of RF-OAM beams with L=1. Diagrams in the upper line are the OAM beam without steering, and that in the lower line are the OAM beam with steering angle $\theta \textrm{ = }{5^o}$. (a) and (e) show the 2-D intensity pattern of RF-OAM beams, and (b) and (f) are the corresponding 1-D curves at the red dotted lines in (a) and (e). (c) and (g) show the 2-D phase pattern of RF-OAM beams, and (d) and (h) are the corresponding 1-D curves. Sizes are in normalized units at the wavelength $\lambda$ of RF beams.
After that, set the optical switch in state “2”, optical signal containing lights with another group of wavelengths is radiated by WDM Laser Group 2. After the specific DSB modulation, the modulated signal is routed to the phase shifter array to get another group of shifted phases, which is related to a different OAM mode and different steering angle. Another broadband RF-OAM beam with different mode and radiation direction is generated. Then set the optical switch in state “3”, the switching between two broadband RF-OAM beams with different OAM mode and radiation directions can be realized.
Till now, the generation, steering and switching of broadband RF-OAM beams have been realized, and the next is to research the signal transmission by this link.
First, generate a broadband OOK/BPSK signal by the arbitrary waveform generator (AWG), then, modulate it onto the single-frequency RF signal generated by the VNA with the multiplier. After that, modulate this modulated signal onto the optical signal by the specific DSB modulation. So, the optical signal carrying OOK/BPSK signal is obtained. After shifting phase in phase shifter array, long-distance transmission in SSMF and power-amplified in EDFA, RF-signals carrying OOK/BPSK signal are recovered by PDs, and routed to corresponding antenna in CAA. Finally, an RF-OAM beams carrying OOK/BPSK signal is generated. In the receiving terminal, two immovable antennas are parallel located after the CAA to receive the signals and send them to an oscilloscope to recover the OOK/BPSK signal.
3. Results
Proof-of-concept experiments are performed following Fig. 2. The first one is the experiment of broadband RF-OAM beam-generation, steering and switching, and the other is the OOK/BPSK signal transmission.
In the first experiment, the frequency of broadband RF signal generated by VNA (Agilent, 8722ES) is from 15GHz to 25GHz. The number of lights in one group of WDM laser and antennas in CAA are all $N = 4$. Lights radiated from WDM Laser Group 1 (Coherent Solutions, Mtp 1000&LaserBlade) are in 1548.79 nm, 1550.4 nm, 1551.999 nm, and 1553.720 nm, and WDM Laser Group 2 (ID Photonics, CoBrite DX4) radiating lights are in 1548.305 nm, 1549.85 nm, 1551.435 nm, and 1553.115 nm. A 2×1 optical switch (Sercalo, SLBU4 × 4-9N-LC/PC) is employed to select and switch two optical channels. The specific DSB modulation is finished in the PM-DMZM (Fujitsu FTM7980EDA), whose half-wave voltage is ${V_\pi } = 8.2v$, and the direct-current (DC) bias voltage is set in the orthogonal offset point (6.51v in x-polarization path/6.31v in y-polarization path) by adding/subtracting ${{{V_\pi }} / 2}$ in the minus point (2.41v in x-polarization path/10.41v in y-polarization path).
As seen in Fig. 5, the modulated signals recorded by an optical spectrum analyzer (Advantest, Q8384) are shown by the blue and golden curves, of which the 8 carries are around -17dBm with flatness less than 1.5dBm, and 16 sidebands are about -28dBm with flatness also less than 1.5dBm. After modulated, the optical signals is sent to the OSP (WaveShaper 4000s), where the designed shifted phases are added as depicted in the red short line in Fig. 5. The phase- shifted signal is then transmitted by a 19.4km SSMF to the BS. In the BS, after power-lifted by the EDFA (Conquer, KG-EDFA-P), the optical signal is demultiplexed by the DWDM, of which the channel spacing is about 200GHz. Then signals from the 4 out-ports of DWDM are further routed to corresponding PDs (Optilab, LR-30), where the RF signals are recovered and sent to CAA. Finally, RF-OAM beams are generated and transmitted by the CAA.
Fig. 5. The measured intensities of modulated signals and experimental configuration of the added phase shift in the OSP. The golden curves sign the intensities of signals in channel 1, and the blue curves sign intensities of signals in channel 2. The red short lines sign the shifted phases that in channel 1 can generate OAM beam with $L ={-} 1$, while that in channel 2 generates OAM beam with $L = 1$, and beams can be steered by adjusting the phase shift as described in (4). The gray curves respect the boundary of channels of DWDM.
First, to verify the immunity to CD-induced power fading of the system, we set the optical switch in state “1”, and measure the signals after PDs. Figure 6 shows the measured RF signals. As seen in Fig. 6(a), flat phase-frequency responses at over 15–25GHz of four RF signals recovered by corresponding PDs are shown, and these signals will be routed to corresponding antenna element of the CAA. By adjusting the DC bias voltage to the x-polarization path orthogonal offset point (6.51v in x-polarization path and 10.41v in y-polarization path) and y-polarization path orthogonal offset point (2,41v in x-polarization path and 6.31v in y-polarization path), comparisons of power-frequency responses between common DSB signals and proposed DSB signal are conducted, as seen in Fig. 6(b). The common DSB signals, including the x- or y- polarization DSB signals represented by the purple dotted and short blue lines, respectively, may suffer from power fading points (PFPs) because of the CD. Through a CD-immune DSB modulation, a flat power response is obtained as depicted by the red line.
Fig. 6. Measured signals after PD in BS: (a) phase-frequency response and (b) normalized amplitude-frequency response through different paths.
To further compare the influence to OAM beam generation of these two DSB modulation, an RF-signal in 15.8GHz, where one PFP of the x-polarization path is located, is modulated onto the optical signal by these two DSB modulation. The generated RF-OAM beams with $L ={-} 1$ are shown in Fig. 7, which is recorded by a pitch antenna, about 35 cm away from CAA, controlled by the stepper slide rail placed along the horizontal direction. In Figs. 7(a) and 7(b), the 2-D intensity and phase curves of OAM beam modulated by proposed DSB modulation are shown, and features of the OAM beam are clearly. In Fig. 7(b), intensity and phase curves of OAM beam modulated by common DSB modulation are shown, and the features are missing. Clearly, CD-induced power fading has a huge impact on the beamforming. Thus, our proposed modulation method is significant.
Fig. 7. Power and phase of RF-OAM beam $L ={-} 1$ generated by (a, b) CD-immune DSB modulation and (c, d) common DSB modulation.
To verify the OAM beam-steering function, the broadband RF-OAM beams with $L = 1$is steered at approximately 2.53° and 4.53°, respectively, by changing the shifted phases following (4). Figure 8 shows the 2-D intensity and phase curves of the RF-OAM beam with and without steering in different frequency point. The steering is clearly both in intensity pattern and in phase pattern. Due to the long measure time, there are some delay in different frequency.
Fig. 8. 2-D intensity and phase patterns of RF-OAM beams with (a) and (b): $\theta = {0^o}$, and (c) and (d): $\theta = {4.53^o}$.
To verify the switching of RF-OAM beams, the optical switch is set in state “3”, and the phases shifted by the OSP is shown in Fig. 5. A detection antenna is in a fixed point, about 35 cm away from CAA, to record the phase of signal, and send it to VNA to analyze. As seen in Fig. 9, the switching of phases appears at the period of 20 ms, which sign the switching of RF-OAM beams with different mode and steering direction are realized.
Finally, the OOK/BPSK signal is successfully transmitted by the proposed downlink. The 2GBaud OOK/BPSK signal generated by the AWG (M9505A) is modulated to the RF signal generated by VNA, and the optical switch is set in state “1”. After modulated, phase shifted, transmitted, power-amplified, demultiplexed, and recovered, the RF-signals carrying OOK/BPSK signal are routed to the CAA, and an RF-OAM beam with L=1 carrying OOK/BPSK signal is generated. Two patched antennas are parallel located about 35 cm away from CAA to receive the Ku-band RF signals and send them to an oscilloscope (Agilent DSO-X 93204A) to analyze. Using a digital down-conversion and adaptive phase compensation for all antennas channel, the signal constellations are recovered. Figure 10 shows the bit error rate (BER) curve varied with Eb/N0, which reflects the signal to noise ratio (SNR) in power. Blue curve shows the BER of the BPSK signal, while the red one shows the OOK signal. The two curves have around 3 dB offset in Eb/N0, which fit the theory. The insets are corresponding signal constellations at the BER around 1 × 10−3.
4. Summary
In conclusion, we proposed an analog RoF downlink scheme for future wireless communications (6G). Owing to the intrinsic orthogonality of OAM, the OAM beam is employed to be as the data carrier and the generated fixed mapping SDM beams to reduce the amount of MIMO computation and latency compared to 5G system. The analog RoF downlink based on the OAM beam is constructed, containing the CS and BS. The CS is for data processing, and the BS is for beam forming. Thus, the oversized workload of BS of 5G system has been paradigm shifted to the CS by photonics techniques. A long-distance SMF is used to connect the BS and CS. Specific DSB modulation is proposed to avoid the CD-induced power fading. To increase the flexibility and reliability of the link, OAM beam steering and switching is researched. Verification experiments are conducted, and RF-OAM beams generation, steering and switching are realized. The RF-OAM beam with $L ={-} 1$is generated by this link, and the broadband RF-OAM beams with $L = 1$ is steered to different angles. Switching of two RF-OAM beams with different mode and steering direction is realized at the period of 20 ms. Finally, the OOK/BPSK signal is successfully transmitted by this proposed downlink. We hope the proposed scheme can solve 5G’s problem and apply to the next generation wireless communications.
Funding
National Natural Science Foundation of China (61690195, 61801038, 61821001, U1831110).
Disclosures
The authors declare no conflicts of interest.
References
1. J. G. Andrews, S. Buzzi, W. Choi, S. Hanly, A. Lozano, A. C. K. Soong, and J. C. Zhang, “What Will 5G Be?” IEEE J. Select. Areas Commun. 32(6), 1065–1082 (2014). [CrossRef]
2. Cisco, “Cisco Annual Internet Report (2018-2023)”, March. 2020, white paper at Cisco.com.
3. A. E. Willner, H. Huang, Y. Yan, Y. Ren, N. Ahmed, G. Xie, C. Bao, L. Li, Y. Cao, Z. Zhao, J. Wang, M. P. J. Lavery, M. Tur, S. Ramachandran, A. F. Molisch, N. Ashrafi, and S. Ashrafi, “Optical communications using orbital angular momentum beams,” Adv. Opt. Photonics 7(1), 66–106 (2015). [CrossRef]
4. Y. Yan, G. Xie, M. Lavery, H. Huang, N. Ahmed, C. Bao, Y. Ren, Y. Cao, L. Li, Z. Zhao, A. Molisch, M. Tur, M. Padgett, and A. Willner, “High-capacity millimetre-wave communications with orbital angular momentum multiplexing,” Nat. Commun. 5(1), 4876 (2014). [CrossRef]
5. J. Wang, J. Y. Yang, I. M. Fazal, N. Ahmed, Y. Yan, H. Huang, Y. Ren, Y. Yue, S. Dolinar, M. Tur, and A. E. Willner, “Terabit free-space data transmission employing orbital angular momentum multiplexing,” Nat. Photonics 6(7), 488–496 (2012). [CrossRef]
6. F. Tamburini, E. Mari, A. Sponselli, F. Romanato, T. Bo, A. Bianchini, L. Palmieri, and C. G. Someda, “Encoding many channels in the same frequency through radio vorticity: first experimental test,” New J. Phys. 14(11), 118002 (2012). [CrossRef]
7. Z. Wang, N. Zhang, and X. C. Yuan, “High-volume optical vortex multiplexing and demultiplexing for free-space optical communication,” Opt. Express 19(2), 482–492 (2011). [CrossRef]
8. X. L. Gao, M. Y. Zhao, M. T. Xie, M. Lei, X. Song, K. Bi, Z. Zheng, and S. Huang, “2D optically controlled radio frequency orbital angular momentum beam steering system based on a dual-parallel Mach–Zehnder modulator,” Opt. Lett. 44(2), 255 (2019). [CrossRef]
9. L. Rego, K. Dorney, N. Brooks, Q. Nguyen, C. Liao, J. Román, D. Couch, A. Liu, E. Pisanty, M. Lewenstein, L. Plaja, H. Kapteyn, M. Murnane, and C. Hernández-García, “Generation of extreme-ultraviolet beams with time-varying orbital angular momentum,” Science 364(6447), eaaw9486 (2019). [CrossRef]
10. Y. Pan, X. Gao, Z. Ren, X. Wang, C. Tu, Y. Li, and H. Wang, “Arbitrarily tunable orbital angular momentum of photons,” Sci. Rep. 6(1), 29212 (2016). [CrossRef]
11. M. Malinauskas, A. Zukauskas, S. Hasegawa, Y. Hayasaki, V. Mizeikis, R. Buividas, and S. Juodkazis, “Ultrafast laser processing of materials: from science to industry,” Light: Sci. Appl. 5(8), e16133 (2016). [CrossRef]
12. G. Mindaugas, I. Maksym, B. Martynas, M. Aidas, T. Viktorija, G. Titas, C. Ausra, R. Kipras, V. Mikas, S. Valerijus, and K. Peter G, “Toward the generation of broadband optical vortices: extending the spectral range of a q-plate by polarization-selective filtering,” J. Opt. Soc. Am. B 35(1), 190–196 (2018). [CrossRef]
13. X. Wang, A. Kuchmizhak, E. Brasselet, and S. Juodkazis, “Dielectric geometric phase optical elements fabricated by femtosecond direct laser writing in photoresists,” Appl. Phys. Lett. 110(18), 181101 (2017). [CrossRef]
14. S. M. Mohammadi, L. K. S. Daldorff, J. E. S. Bergman, R. L. Karlsson, T. Bo, K. Forozesh, and T. D. Carozzi, “Orbital Angular Momentum in Radio—A System Study,” IEEE Trans. Antennas Propag. 58(2), 565–572 (2010). [CrossRef]
15. G. Xie, Z. Zhe, Y. Yan, L. Long, Y. Ren, N. Ahmed, Y. Cao, A. J. Willner, C. Bao, and W. Zhe, “Demonstration of Tunable Steering and Multiplexing of Two 28 GHz Data Carrying Orbital Angular Momentum Beams Using Antenna Array,” Sci. Rep. 6(1), 37078 (2016). [CrossRef]
16. M. Lei, Z. Zheng, J. Qian, M. Xie, Y. Bai, X. Gao, and S. Huang, “Broadband chromatic-dispersion-induced power fading compensation for radio-over-fiber links based on Hilbert transform,” Opt. Lett. 44(1), 155 (2019). [CrossRef]
