Open Access
Add to BibSonomy
Abstract
A dual sideband reception scheme for radio-over fiber (RoF) links is introduced. It is shown that the new receiver can increase the performance of noise-limited systems by up to 3 dB (2.97 dB in a lab back-to-back experiment). The receiver scheme exploits the fact that current RoF links do not realize their full potential. This is because in typical RoF receivers, the radio-frequency (RF) signals are mapped back to the optical domain by means of electro-optical modulator. In this process energy typically is lost as only one of the two generated sidebands is subsequently used. The suggested receiver exploits the signal of both sidebands. The receiver scheme was subsequently tested in a full optical-RF-optical transmission link at RF carrier frequencies of 228 GHz over a free-space channel spanning distances of 1400 m for symbol rates of up to 48 Gbaud 4 QAM. Here, we could achieve SNR improvements of up to 2.6 dB.
© 2024 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
THz wireless links can bridge fiber-optical networks, particularly in scenarios, where fibers cannot be deployed and/or a flexible allocation via beam steering is needed. The workhorse in such radio-over-fiber (RoF) systems [1–4] consists of microwave photonic devices [5,6] that handle optical-to-RF [7–11] and RF-to-optical conversions. Specifically, the RF/THz receiver, a crucial element, needs to be able to receive signals with low powers at high bandwidths and translate the RF signal (back) to the optical domain.
Recently, successful THz data transmissions have been shown, achieving line-rates at and above 100 Gbit/s [12–28] for distances up to 110 m [21], a line-rate of 50 Gbit/s over 850 m [29] and even 192 Gbit/s over 115 m [30]. Greater data rates of 600 Gbit/s [31] and even 1056 Gbit/s [32] were successfully demonstrated over short-range distances of 2.8 m, and up to 3.1 m by employing multiplexing schemes such as polarization division multiplexing and/or MIMO-schemes. All these successful demonstrations rely on a direct optical-to-RF conversion using an ultra-fast uni-traveling carrier (UTC) photodiode at the transmitter side. Yet, at the receiver side the anticipated RF-to-optical conversion to continue in the fiber network was often avoided by RF-to-electrical reception [12–21,23–25,29,31–33]. Here, typically an electrical mixer based on Schottky-barrier diodes (SBD) are utilized that show high receiver sensitivities: Under low power operation, SBDs and passband RF-amplifiers showed typically low noise-equivalent powers of $1.5 \cdot {10^{ - 12}}W/\sqrt {\textrm{Hz}} $ [34] and low noise figures of ∼8-12 dB [35–37]. Therefore, system performance is limited by the added thermal noise of the active RF components. In addition, they are also limited by their intermediate-frequency bandwidth to a few tens of GHz and thus, multiple narrowband channels with high-order modulation formats were utilized to achieve high data-rates.
However, the (intermediate) down-mixing to baseband can be bypassed by a direct up-conversion from the RF domain to the optical by means of an electro-optical modulator [38–40]. The RF signal is thereby mapped to the upper and lower sideband of the optical carrier. Both sidebands carry the transmitted information and thus, one sideband is subsequently selected and received by an optical receiver. This scheme simplifies the front-end architecture and enables a broadband reception scheme at a broad range of frequencies.
The major challenge is that the modulator should operate at or beyond 100 GHz frequencies. Limitations in electro-optical bandwidth and/or modulation efficiency will lead to lower sideband powers. Consequently, the system performance and thus the achievable data-rates are limited by the amplified spontaneous emission (ASE) noise, in particular the (ASE)-LO beat noise, added by the erbium-doped fiber amplifier (EDFA) after the modulator. Recently, leveraging plasmonic modulators [40] with their >500 GHz bandwidth [41] enabled to build transparent optical-RF-optical transmission links [30,42,43] with line-rates of up to 240 Gbit/s [30] over distances of up to 115 m [30]. Again, in all these demonstrations, the ultimate reach was limited by the signal-to-noise ratio (SNR) after the RF-to-optical mixer and thus, there is a quest for novel methods to increase the receiver sensitivity.
Early on, the optical community has investigated so-called multi-aperture reception [44–53] schemes in free-space links as a technique to mitigate the atmospheric turbulence. Here, the phase-distorted wave front of the incoming EM field is collected by N apertures instead of a single aperture. After amplification and coupling back to a single-mode fiber, all N signals are coherently combined in the optical, the RF or digital domain. On average, this leads to a $10 \cdot {\log _{10}}N$ dB SNR advantage over a single aperture. The question then is if there is an option to exploit a multi-reception scheme to improve the SNR of RoF links.
In this work, we propose and demonstrate a novel reception scheme in RoF links, where the information on the upper and lower sidebands next to the optical carrier are separately received and coherently combined in the digital domain. As a result, one is capable to increase the performance of optical-RF-optical links by 3 dB (2.97 dB measured) and thus double the achievable data-rates. We also introduce a scheme, where the receiver complexity remains low by using a standard dual polarization coherent receiver to detect the two sidebands.
The manuscript is structured as follows: Section 2 describes the implementation of a fiber-RF-to-fiber link exploiting the dual sideband reception to obtain a RF-to-Fiber SNR improvement. Section 3 explains the theory behind the digital coherent combination and the resulting SNR gain. Section 4 focuses on how the coherent combination of the sidebands can be implemented with digital signal processing (DSP) and standard coherent receivers. Finally, in section 5, the novel concept is tested in experiment.
2. Dual-sideband receiver to improve RF-to-fiber SNR
Figure 1 illustrates the RF-to-optical converter with the dual sideband receiver. An RF signal from a free-space channel is collected by a receiver (Rx) antenna. The RF signal is then fed to an RF-to-optical converter (RF-O Conv.). In the RF-O Conv., a high-speed modulator (HS-Mod) maps the RF signal to the optical domain, without any intermediate down-mixing. The inset D of Fig. 1 shows the optical spectrum after the HS-Mod, consisting of the optical carrier, the lower and the upper sideband. Typically, only one sideband is selected and subsequently used as both sidebands carry the same information. However, the discard of one sideband is equal to lost energy. Therefore, in the dual-sideband receiver scheme, both sidebands are maintained, propagated through a fiber, and fed to the dual sideband mapper (DSM). In the DSM, both sidebands are separated from each-other by a wavelength demultiplexer (DEMUX).
Fig. 1. RF-to-optical-to-baseband span with dual sideband reception. At the RF-to-optical converter (RF-O Conv.), a wireless RF data signal is received by an RF antenna. The RF signal then directly drives a high-speed modulator (HS-Mod) to encode the information onto the optical carrier with frequency ${f_{c,Rx}}$. The optical spectrum consisting of the optical carrier, the lower (LSB) and upper (USB) sidebands is illustrated in inset D. Subsequently, the optical signal is amplified and propagates through a fiber to the dual sideband receiver (Rx). Here, a wavelength division multiplexer (DEMUX) acts as a dual sideband mapper (DSM) by assigning each sideband to an individual coherent receiver (Coh. Rx.), see insets F-LSB and F-USB. Finally, the sidebands are mixed to baseband by a local oscillator laser (LSB: ${f_{c,Rx}} - {f_{RF}}$; USB: ${f_{c,Rx}} + {f_{RF}}$) and both sidebands are evaluated by an offline digital signal processing stage. Here, after-pre-processing, both sidebands are aligned and coherently combined in the digital domain. B-PD: balanced photodetector; Coh.: Coherent; DEMUX: wavelength division demultiplexer; DSM: dual sideband mapper; DSO: digital sampling oscilloscope; DSP: digital signal processing; HS-Mod: high-speed modulator; LSB: lower sidebands; OBPF: optical band pass filter; RF-O Conv: RF-to-optical converter; Rx: receiver; USB: upper sideband.
The lower sideband (LSB) and the upper sideband (USB) are then collected by individual coherent receivers (Coh. Rx), where local oscillators at optical frequencies ${\textrm{f}_{\textrm{c},\textrm{Rx}}} - {\textrm{f}_{\textrm{RF}}}$ and ${\textrm{f}_{\textrm{c},\textrm{Rx}}} + {\textrm{f}_{\textrm{RF}}}\; $respectively, down-mix them to baseband. The insets labeled F-LSB and F-USB display the spectra from the LSB Coh. Rx and USB Coh. Rx respectively. Subsequently, the baseband LSB and USB signals aligned to each-other and coherently combined to enhance the overall signal quality, resulting in a higher signal-to-noise ratio denoted as$\; \textrm{SN}{\textrm{R}_{\textrm{tot},1 + 2}}$. It is ensured that $\textrm{SN}{\textrm{R}_{\textrm{tot},1 + 2}} \ge \textrm{SN}{\textrm{R}_{\textrm{tot},1}} \wedge \; \textrm{SN}{\textrm{R}_{\textrm{tot},1 + 2}} \ge \textrm{SN}{\textrm{R}_{\textrm{tot},2}}$, where $\textrm{SN}{\textrm{R}_{\textrm{tot},1}}$ and $\textrm{SN}{\textrm{R}_{\textrm{tot},2}}$ represent the individual signal-to-noise ratios of the received LSB and USB signals, respectively.
3. Theory
In this section, we explore the potential maximum SNR gain achievable by deploying separate receivers for each of the two sidebands in a RoF link, contrasting it with the conventional approach of selecting a single sideband for reception.
Figure 2(a) illustrates the RF-to-optical-to-baseband span featuring dual sideband reception, as depicted in Fig. 1, with a specific emphasis on analyzing the noise and signal-to-noise ratio (SNR) components in the system: The transmitter sends an encoded data signal through a RF channel. Subsequently, the signal is mapped to the optical domain in the RF-to-optical converter by means of an optical modulator. The two generated sidebands LSB and USB are divided into two independent fiber channels (LSB → 1, USB → 2) and received by individual receiver (Rx-1 & Rx-2). at the receiver side. We model all three channels by additive white Gaussian noise (AWGN) channels where the noise is described by ${n_{\textrm{RF}}}$, ${n_{\textrm{F},1}}$, and ${n_{\textrm{F},2}}$. For example, the noise in the fiber channels, ${n_{\textrm{F},1}}$, and ${n_{\textrm{F},2}}$, may stem from amplified spontaneous emission (ASE) noise due to optical amplification stage(s) respectively. Assuming an ideal modulator with linear transfer function from the RF to the optical, each sideband signal after the RF-O Conv. can be described by ${r_{\textrm{opt}.}} = {r_{\textrm{RF}}} = s + {n_{\textrm{RF}}}$ and its signal-to-noise ratio (SNR) is given as
Fig. 2. SNR analysis of the RF-to-optical-converter with the dual sideband receiver (a) At the transmitter, information is sent through a common RF channel, and after mapping to the optical domain by an optical modulator, the two inherently generated sidebands are split into two individual fiber channels. All channels are modeled by additive white Gaussian noise and described by different noise sources ${n_{\text{RF}}},\; {n_{\text{F},1}},\; $and$\; {n_{\text{F},2}}$. The received sidebands signals can be aligned and combined since they carry the same information, aiming to enhance the signal-to-noise ratio (SNR) due to the uncorrelated noise between the channels. (b) SNR gain $\Delta \text{SNR}\; $plotted as a function of the $\mathrm{SNR}_{\mathrm{RF}} / \mathrm{SNR}_{\text {tot }, 1}$ for different $a=\mathrm{SNR}_{\mathrm{tot}, 2} / \mathrm{SNR}_{\mathrm{tot}, 1}$.
After propagating through the independent fiber channels, the signals can be described by
where the subscript $i = 1,2$ indicates the fiber channel, see Fig. 2(a). Theirs SNRs are given by the following:The noise ${n_{\textrm{F},1}}$ and ${n_{\textrm{F},2}}$ added in the fiber channels 1 and 2 are uncorrelated to the common noise ${n_{\textrm{RF}}}$ and to each-other, i.e. $\text{E}[{{n_{\textrm{RF}}}{n_{\textrm{F},1}}} ]= \text{E}[{{n_{\textrm{RF}}}{n_{\textrm{F},2}}} ]= \text{E}[{{n_{\textrm{F},1}}{n_{\textrm{F},2}}} ]= 0$. Furthermore, the mean value of the noises is zero, i.e., $\textrm{E}{[{{n_{\textrm{RF}}}} ]^2} = \textrm{E}$[nF,i]2 $\; = 0$. After reception, the two signals ${r_{\textrm{tot},1}}$ and ${r_{\textrm{tot},2}}$ are coherently combined to improve the overall SNR of the signal as the noise between the fiber channels 1 and 2 is uncorrelated. While the signal power is here assumed to be equal, the noise loading between the channels can differ and may lead to different received $\textrm{SN}{\textrm{R}_{\textrm{tot},i}}$ with $\textrm{SN}{\textrm{R}_{\textrm{tot},2}} = a \cdot \textrm{SN}{\textrm{R}_{\textrm{tot},1}}$ and $a \in {\mathrm{\mathbb{R}}^ + }$. Therefore, to perform “maximum ratio combining” of those two received signals ${r_{\textrm{tot},1}}$ and ${r_{\textrm{tot},2}}$, we scale signal ${r_{\textrm{tot},2}}$ by a weight factor w, and add it to signal ${r_{\textrm{tot},1}}$. The resulting signal ${r_{\textrm{tot},1 + 2}}\; $is given by
and its SNR value is given by:The maximum SNRtot,1 + 2 is obtained if the weight factor $w\; $has the value ${w_{\textrm{opt}}}$ $= \text{E}[{n_{\textrm{F},1}^2} ]/\text{E}[{n_{\textrm{F},2}^2} ]$ [54]. Inserting ${w_{\textrm{opt}}}$ in Eq. (5) gives
Using the expression for the SNRs in Eq. (1), and Eq. (3), we can reformulate Eq. (6) into the following way:
The details of the derivation can be found in the Supplement 1.
Without loss of generality, we can assume $\textrm{SN}{\textrm{R}_{\textrm{tot},1}} \ge \textrm{SN}{\textrm{R}_{\textrm{tot},2}} = a \cdot \textrm{SN}{\textrm{R}_1}$ with $0 < a \le 1$. We define the SNR gain $\mathrm{\Delta SNR}$ as the ratio between $\textrm{SN}{\textrm{R}_{\textrm{tot,1 + 2}}}$ and the best SNR, here $\textrm{SN}{\textrm{R}_{\textrm{tot,1}}}.$ Fig. 2(b) shows $\mathrm{\Delta SNR}\; $ as a function of $\textrm{SN}{\textrm{R}_{\textrm{tot},1 + 2}}/\textrm{SN}{\textrm{R}_{\textrm{tot},1}} = 1 + \text{E}[{n_{\textrm{F},1}^2} ]/\text{E}[{n_{\textrm{RF}}^2} ]$ for different values of $a = \textrm{SN}{\textrm{R}_{\textrm{tot},2}}/\textrm{SN}{\textrm{R}_{\textrm{tot},1}}$. Figure 2(b) has two main findings: First, for $n_{\textrm{RF}} \ll {{n}_{\textrm{F},\textrm{i}}}$, one can achieves the highest SNR gains. Secondly, for $n_{\mathrm{RF}} \gg n_{\mathrm{F}, 1}$, no SNR gain can be obtained and $\textrm{SN}{\textrm{R}_{\textrm{tot},{\; }1 + 2}} = \textrm{SN}{\textrm{R}_{\textrm{RF}}} = \textrm{SN}{\textrm{R}_{\textrm{tot},1}}$. We will derive the findings in the following.
$\boldsymbol{n}_{\bf{RF}} \ll \boldsymbol{n}_{\bf{F}, 1}$ implies that the SNRRF is of high quality and becomes comparatively very large and the noise of the fiber channel ${n_{\textrm{F},1}}$ dominate over the noise from shared RF channel ${n_{\textrm{RF}}}$. Here, Eq. (7) simplifies to
And the maximum achievable SNR gain $\mathrm{\Delta SNR}$ can be calculated by
Therefore, no SNR gain can be won
In theory, the power of both sidebands is equal and, thus $\textrm{SN}{\textrm{R}_1} = \textrm{SN}{\textrm{R}_2}$ and ${w_{\textrm{opt}.}} = 1$. Equation (7) simplifies to
Therefore, a theoretical SNR improvement of up to 3 dB can be achieved:
The following section discusses steps how the coherent combination of the sidebands can be realized in the digital domain by digital signal processing.
4. DSP realization of the dual-sideband receiver - coherent combining
Prior to the digital combination, several digital signal processing (DSP) steps must be applied on the individual sidebands, see Fig. 3. After matched filtering, the clock is synchronized by using a clock recovery algorithm [55]and thus any sub-symbol skew between the signals is removed. The remaining delay between the sidebands is an integer number of symbol times ${n_T}$. To compensate for the frequency response of the channel feed-forward equalizers are used where the filter coefficients are blindly trained based on a constant modulus algorithm (CMA) and radius-directed equalizer (RDE), both with a steepest gradient descent cost function. As each sideband is mixed down by a free-running local oscillator, there will be different time-varying laser frequency offsets $\mathrm{\Delta }f.{\; }$Therefore, each sideband’s carrier frequency is separately recovered using a 4th power algorithm. Afterwards the complex sidebands are de-skewed with respect to each other by cross-correlation of the signal’s real and imaginary part. This way not only the inter-symbol skew is removed, but also the conjugate relation between the upper and the lower sidebands ${E_{\textrm{USB}}} ={-} E_{\textrm{LSB}}^\mathrm{\ast }$ is corrected, such that the lower sideband is flipped onto the upper sideband, see inset D in Fig. 1. In a step, a coherent combining algorithm [52] aligns the phases of the sidebands with respect to each-other, followed by equal gain combination, see Fig. 3. Subsequently, the phase of the combined signal is recovered by a blind-phase-search (BPS) algorithm. Finally, a T/2-spaced feed-forward (FF) equalizer based on data-aided least-mean square algorithm removes residual bandwidth limitations. The combined signal is evaluated in terms of SNR, bit-error-ratio (BER), error vector magnitude (EVM) and modulation error ratio (MER).
Fig. 3. Digital signal processing steps to coherently combine the two sidebands digitally. BER: bit-error ratio; CMA: constant modulus algorithm; Comb.: combined; Est.: estimated; FF: feed-forward; RDE: radius-directed equalizer; SNR: signal-to-noise ratio.
5. Experimental results
The experimental configuration of the RF-optical-to-baseband span is illustrated Fig. 4(a). After the RF-O Conv., the signal is noise loaded to emulate different channel conditions. In the noise loader, a wavelength selected switch (WSS) suppresses the carrier, corrects for slight power-differences between upper and lower sideband and attenuates both signals to simulate different sideband powers. Then, noise is loaded by amplification with an EDFA. Subsequently, the signal is propagated through a fiber and fed to the new dual sideband receiver. To overcome equipment limitations, the number of required Coh. Rx was reduced by mapping the LSB to the x-polarization and USB to the y-polarization in the dual sideband mapper (DSM) and receiving both sidebands by a single dual-polarized Coh. Rx.
Fig. 4. Demonstration of digital coherent combining in a RoF transmission link (a) Experimental setup. Here, the channel was noise loaded to investigate receiver performance at low powers. Under the assumption of single-polarized and polarization-constant RF signals, the number of needed coherent receivers (Coh. Rx) was reduced by (b) mapping the lower and upper sidebands to the x- and y-polarizations respectively and received both sidebands with a standard dual polarization Coh. Rx. In Exp. 1, the dual sideband Rx was tested in a back-to-back experiment at RF carrier frequencies ${f_{\text{RF}}}$ of $30\; \text{GHz}$ and a wireless distance d of 0 m. SNR vs. power per sideband (PSB) of the lower (LSB), upper (USB) and combined (comb.) sidebands for a (c) 16 GBaud 4 QAM signal and (d) 32 GBaud 4 QAM signal. The shot noise limit of the combined signal as a solid curve, theoretical expected SNR of the combined signal in case of 3-dB win as a dashed line. The inset in (c) shows the measured optical spectrum at position D. The digital received spectrum of the LSB (up) and the USB (down) for the 32 GBaud 4QAM signal is illustrated in inset of (d). Here, on can clearly see the bandwidth limitations of the 35 GHz HS-Mod. (e) Exp. 2: fiber-RF-fiber data transmission link at a carrier frequency of 228 GHz over a free-space channel spanning a distance of 1400 m. The figure shows the SNR gain as a function of symbol rates for 4 QAM signal. The inset shows the constellation diagrams, calculated SNRs and BERs at symbol rates of 32 GBaud for the LSB, USB and the combined signal (LSB+USB = comb.). B-PD: balanced photodetector; DEMUX: wavelength division demultiplexer; DSM: dual sideband mapper; OBPF: optical band pass filter; PC: polarization controller; PBS: polarization beam splitter; PSB: power per sideband; WSS: wavelength selective switch.
The implementation of the dual sideband receiver is illustrated in Fig. 4(b). First, a wavelength division demultiplexer (DEMUX) separates the two sidebands. Then, the upper and the lower sideband are projected to the x- and y-polarization respectively using a polarization controller and a polarization beam splitter. Subsequently, the signals are orthogonally recombined using a polarization beam combiner. Two local oscillators with optical frequencies of ${f_{\textrm{c},\textrm{Rx}}} + {f_{\textrm{RF}}}$ and ${f_{\textrm{c},\textrm{Rx}}} - {f_{\textrm{RF}}}{\; }$in the respective polarizations for the lower and upper sideband are fed to coherent receiver. Here, each LO mixes down its respective sideband in the corresponding polarization to baseband. Afterwards, both signals are captured by a digital sampling oscilloscope (DSO) and the digitized signals enter the offline digital signal processing stage.
To verify that the dual sideband reception scheme brings a 3 dB advantage over the conventional single sideband reception the setup was tested in two scenarios, Exp. 1 with ${f_{\textrm{RF}}} = 30\; \textrm{GHz}\; $in a back-to-back link with $d = 0\; \textrm{m}$ free-space distance and Exp. 2 with ${f_{\textrm{RF}}} = 228\; \textrm{GHz}\; $over a free-space distance of $d = 1400\; \textrm{m}$.
The results of Exp. 1 reported in Fig. 4(c) and (d) show an SNR advantage of up to 2.97 dB in experiment. The exact results of a 16 GBaud 4QAM experiment are presented in (c). The plot shows the SNR as a function of power per sideband (PSB) for the lower sideband (LSB), upper sideband (USB), and the measured combined signal. The dashed line corresponds to the combined signal in theory, while the solid line represents the shot noise limit of the combined signal calculated using the formula $\eta {P_s}/({hfR} )\; $ [56]. The shot noise limit is here employed as it establishes the upper boundary for the ASE-LO beat noise limit, which is calculated using the formula $SNR = \eta {P_s}/({{n_{\textrm{sp}}}hfR} )$ [56], considering an ideal optical amplifier with a spontaneous emission factor ${n_{\textrm{sp}}} \simeq 1$. In these equations, $\eta $=1 represents the quantum efficiency, $hf$ is the photon energy, R denotes the symbol rate in the Nyquist limit, and ${P_s}\; $is the allocated power of the both the upper and the lower sidebands, i.e. ${P_s} = {P_{\textrm{LSB}}} + {P_{\textrm{USB}}}$. The inset displays the measured optical spectrum after the high-speed modulator (HS-Mod) at position D. The HS-Mod has a 3 dB electrical bandwidth of 35 GHz.
Looking at the figure, we observe, that the combined signal can achieve up to 2.9 dB gain in SNR. At higher sideband powers, the SNR gain decreases due to limitations imposed by the SNR of the shared RF channel SNRRF, as explained in our Theory section and shown in Fig. 2(b). In this sideband-power regime, the system is constrained by the noise added in the transmitter well as the quantization noise of the AWG and DSO. In this scenario, the uncorrelated ASE-LO beat noise less and less matters. Consequently, the noise terms in both sidebands are correlated, and the SNR gain will converge to zero. The experimental configuration of the transmitter in the back-to-back experiment is shown in Fig S1(a) of the Supplement 1.
Remarkably, in the very low SNR scenarios depicted in Fig. 4 (c) and (b), the single sideband receivers encounter difficulties in digital signal processing, specifically at the carrier phase recovery stage. However, the dual sideband receiver scheme manages to successfully decode the signals. Using the digital coherent combination, one is capable to successfully transmit and detect a 16 GBaud 4 QAM signal even with powers as low as -55 dBm per sideband. The combined signals then are only 1.5 dB off from the ideal shot noise limit while operation with only one sideband would result in a penalty of 4.4 dB. To confirm the setup also works despite bandwidth limitations of the 35 GHz HS-Mod, higher symbol rates were tested in the same setup, see Fig. 4 (d). The inset of Fig. 3(d) illustrates the digital received spectra of the 32 GBaud 4 QAM signal. The result shows that utilization of coherent combination can lead to SNR advantages of up to 2.97 dB.
The results of a full fiber-RF-fiber link experiment (Exp. 2) over a distance of 1400 m at a carrier frequency of 228 GHz are given in Fig. 4(e). The results disclose that SNR gains of up
to 2.6 dB at 16 GBaud were achieved by digital combination of the 4 QAM signals. The experimental setup of the transmitter side is illustrated in Fig S1(b) of the Supplement 1. The configuration of the full link, in particular of the transmitter side, were presented and discussed in detail in [30]. From Fig. 4(e) one can see that a penalty is introduced when going to higher symbol rates. This penalty is clearly not a limitation due to the modulator, as we used a an unpackaged plasmonic Mach-Zehnder modulator with over >500 GHz bandwidth. The penalty can rather be explained by the fact, that the SNR of the received RF signal $\textrm{SN}{\textrm{R}_{\textrm{RF}}}$ decreases with higher speed and the power of the received signal over this distance is limited. This ultimately upper-limits the achievable SNR gain by the dual sideband reception. The inset of Fig. 4(e) displays the constellation diagrams, SNR and BER of the USB, LSB and the combined signals at a symbol rate of 32 GBaud. The calculated error vector magnitude (EVM) and modulation error ratio (MER) of the combined signals are for the 32 GBaud 36.1% and 8.9 dB, respectively. The improvement of the constellation diagrams by digital combination can clearly be seen. Ultimately, this led to a successful transmission and reception of 48 GBaud with BER of 1.2·10−2 below the 20%-OH SD-FEC limit.
6. Conclusion
In summary, we have introduced a dual sideband reception scheme for RoF links. Here both sidebands, which are generated by mapping the RF data signal to the optical domain in the electro-optical modulator, are maintained, separately received and finally digitally combined, leading to a potential SNR improvement of up to 3 dB. In a back-to-back experiment (Exp. 1), we could measure a 2.97 dB SNR improvement for a 16 GBaud 4 QAM signal at a RF carrier frequency of 30 GHz, hereby coming as close as 1.5 dB to the shot noise limit. In a second experiment (Exp. 2), we tested the dual sideband reception scheme in a full optical-RF-optical transmission link at carrier frequencies of 228 GHz over a free-space channel spanning distances of 1400 m for symbol rates of up to 48 Gbaud 4 QAM. Here, we could achieve SNR improvements of up to 2.6 dB. In all experiments, the dual side band receiver was implemented using a standard dual polarization coherent receiver to detect the two sidebands, effectively maintaining low system complexity.
Funding
Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung (206021_198113); Armasuisse Science and Technology; H2020 Research and Innovation Programme (822030); H2020 European Research Council (670478).
Disclosures
The authors declare no conflicts of interests.
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.
Supplemental document
See Supplement 1 for supporting content.
References
1. D. Wake, A. Nkansah, and N. J. Gomes, “Radio Over Fiber Link Design for Next Generation Wireless Systems,” J. Lightwave Technol. 28(16), 2456–2464 (2010). [CrossRef]
2. C. Lim, A. Nirmalathas, M. Bakaul, P. Gamage, K. L. Lee, Y. Yang, D. Novak, and R. Waterhouse, “Fiber-Wireless Networks and Subsystem Technologies,” J. Lightwave Technol. 28(4), 390–405 (2010). [CrossRef]
3. J. Beas, G. Castanon, I. Aldaya, A. Aragon-Zavala, and G. Campuzano, “Millimeter-Wave Frequency Radio over Fiber Systems: A Survey,” IEEE Commun. Surv. Tutorials 15(4), 1593–1619 (2013). [CrossRef]
4. C. Lim and A. Nirmalathas, “Radio-Over-Fiber Technology: Present and Future,” J. Lightwave Technol. 39(4), 881–888 (2021). [CrossRef]
5. D. Marpaung, J. Yao, and J. Capmany, “Integrated microwave photonics,” Nat. Photonics 13(2), 80–90 (2019). [CrossRef]
6. C. H. Cox III, Analog optical links: theory and practice. Cambridge University Press, 2006.
7. A. Wakatsuki, T. Furuta, Y. Muramoto, T. Yoshimatsu, and H. Ito, “High-power and broadband sub-terahertz wave generation using a J-band photomixer module with rectangular-waveguide output port,” in 2008 33rd International Conference on Infrared, Millimeter and Terahertz Waves, 2008, pp. 1–2.
8. T. Nagatsuma, H. Ito, and T. Ishibashi, “High-power RF photodiodes and their applications,” Laser Photonics Rev. 3(1-2), 123–137 (2009). [CrossRef]
9. N. Argyris, G. Giannoulis, K. Kanta, N. Iliadis, C. Vagionas, S. Papaioannou, G. Kalfas, D. Apostolopoulos, C. Caillaud, H. Debrégeas, N. Pleros, and H. Avramopoulos, “A 5 G mmWave Fiber-Wireless IFoF Analog Mobile Fronthaul Link With up to 24-Gb/s Multiband Wireless Capacity,” J. Lightwave Technol. 37(12), 2883–2891 (2019). [CrossRef]
10. A. Delmade, C. Browning, T. Verolet, J. Poette, A. Farhang, H. H. Elwan, R. D. Koilpillai, G. Aubin, F. Lelarge, A. Ramdane, D. Venkitesh, and L. P. Barry, “Optical Heterodyne Analog Radio-Over-Fiber Link for Millimeter-Wave Wireless Systems,” J. Lightwave Technol. 39(2), 465–474 (2021). [CrossRef]
11. G. Kalfas, C. Vagionas, A. Antonopoulos, E. Kartsakli, A. Mesodiakaki, S. Papaioannou, P. Maniotis, J. S. Vardakas, C. Verikoukis, and N. Pleros, “Next Generation Fiber-Wireless Fronthaul for 5 G mmWave Networks,” IEEE Commun. Mag. 57(3), 138–144 (2019). [CrossRef]
12. S. Koenig, D. Lopez-Diaz, J. Antes, F. Boes, R. Henneberger, A. Leuther, A. Tessmann, R. Schmogrow, D. Hillerkuss, and R. Palmer, “Wireless sub-THz communication system with high data rate,” Nat. Photonics 7(12), 977–981 (2013). [CrossRef]
13. H. Shams, T. Shao, M. J. Fice, P. M. Anandarajah, C. C. Renaud, F. V. Dijk, L. P. Barry, and A. J. Seeds, “100 Gb/s Multicarrier THz Wireless Transmission System With High Frequency Stability Based on A Gain-Switched Laser Comb Source,” IEEE Photonics J. 7(3), 1–11 (2015). [CrossRef]
14. T. Nagatsuma, Y. Fujita, Y. Yasuda, Y. Kanai, S. Hisatake, M. Fujiwara, and J. Kani, “Real-time 100-Gbit/s QPSK transmission using photonics-based 300-GHz-band wireless link,” in 2016 IEEE International Topical Meeting on Microwave Photonics (MWP), 2016, pp. 27–30.
15. X. Yu, S. Jia, H. Hu, M. Galili, T. Morioka, P. U. Jepsen, and L. K. Oxenløwe, “160 Gbit/s photonics wireless transmission in the 300-500 GHz band,” APL Photonics 1(8), 081301 (2016). [CrossRef]
16. X. Pang, S. Jia, O. Ozolins, X. Yu, H. Hu, L. Marcon, P. Guan, F. Da Ros, S. Popov, G. Jacobsen, M. Galili, T. Morioka, D. Zibar, and L. K. Oxenløwe, “260 Gbit/s photonic-wireless link in the THz band,” in 2016 IEEE Photonics Conference (IPC), 2016, pp. 1–2.
17. X. Pang, S. Jia, O. Ozolins, X. Yu, H. Hu, L. Marcon, P. Guan, F. D. Ros, S. Popov, G. Jacobsen, M. Galili, T. Morioka, D. Zibar, and L. K. Oxenløwe, “Single Channel 106 Gbit/s 16QAM Wireless Transmission in the 0.4 THz Band,” in Optical Fiber Communication Conference, 2017, p. Tu3B.5.
18. V. K. Chinni, P. Latzel, M. Zégaoui, C. Coinon, X. Wallart, E. Peytavit, J. F. Lampin, K. Engenhardt, P. Szriftgiser, M. Zaknoune, and G. Ducournau, “Single-channel 100 Gbit/s transmission using III–V UTC-PDs for future IEEE 802.15.3d wireless links in the 300 GHz band,” Electron. Lett. 54(10), 638–640 (2018). [CrossRef]
19. A. Udalcovs, S. Jia, L. Zhang, O. Ozolins, X. Pang, D. Kong, X. Yu, S. Xiao, S. Popov, and J. Chen, “107.1-Gbps net-rate transmission over a joint 51km-fibre-and-10.7 m-wireless link for terahertz radio access networks,” in 45th European Conference on Optical Communication (ECOC 2019), 2019, pp. 1–4.
20. X. Li, J. Yu, L. Zhao, W. Zhou, K. Wang, M. Kong, G.-K. Chang, Y. Zhang, X. Pan, and X. Xin, “132-Gb/s Photonics-Aided Single-Carrier Wireless Terahertz-Wave Signal Transmission at 450 GHz Enabled by 64QAM Modulation and Probabilistic Shaping,” in 2019 Optical Fiber Communications Conference and Exhibition (OFC), 2019, pp. 1–3.
21. T. Harter, C. Füllner, J. N. Kemal, S. Ummethala, J. L. Steinmann, M. Brosi, J. L. Hesler, E. Bründermann, A. S. Müller, W. Freude, S. Randel, and C. Koos, “Generalized Kramers–Kronig receiver for coherent terahertz communications,” Nat. Photonics 14(10), 601–606 (2020). [CrossRef]
22. C. Castro, R. Elschner, T. Merkle, C. Schubert, and R. Freund, “100 Gb/s Real-Time Transmission over a THz Wireless Fiber Extender Using a Digital-Coherent Optical Modem,” in Optical Fiber Communication Conference (OFC) 2020, 2020, p. M4I.2.
23. J. Tebart, M. Steeg, F. Exner, A. Czylwik, and A. Stöhr, “Frequency-Scalable Coherent Radio-Over-Fiber Architecture for 100 Gbit/s Wireless Transmission,” URSI Radio Sci. Lett. 2, 5 (2021). [CrossRef]
24. J. Ding, W. Li, Y. Wang, J. Zhang, M. Zhu, W. Zhou, and J. Yu, “124.8-Gbit/s PS-256QAM Signal Wireless Delivery Over 104 m in a Photonics-Aided Terahertz-Wave System,” IEEE Trans. Terahertz Sci. Technol. 12(4), 409–414 (2022). [CrossRef]
25. W. Li, J. Yu, Y. Wang, J. Ding, F. Wang, B. Zhu, X. Ji, Y. Tan, T. Zheng, J. Liu, W. Zhou, F. Zhao, and J. Yu, “High Spectral Efficiency Photonics-Based Polarization Multiplexing OFDM Signal Delivery at W-Band,” J. Lightwave Technol. 41(12), 4010–4016 (2023). [CrossRef]
26. F. Boes, T. Messinger, J. Antes, D. Meier, A. Tessmann, A. Inam, and I. Kallfass, “Ultra-broadband MMIC-based wireless link at 240 GHz enabled by 64GS/s DAC,” in 39th International Conference on Infrared, Millimeter, and Terahertz waves (IRMMW-THz), 2014, pp. 1–2.
27. S. Jia, M.-C. Lo, L. Zhang, O. Ozolins, A. Udalcovs, D. Kong, X. Pang, R. Guzman, X. Yu, S. Xiao, S. Popov, J. Chen, G. Carpintero, T. Morioka, H. Hu, and L. K. Oxenløwe, “Integrated dual-laser photonic chip for high-purity carrier generation enabling ultrafast terahertz wireless communications,” Nat. Commun. 13(1), 1388 (2022). [CrossRef]
28. S. Jia, X. Pang, O. Ozolins, X. Yu, H. Hu, J. Yu, P. Guan, F. Da Ros, S. Popov, G. Jacobsen, M. Galili, T. Morioka, D. Zibar, and L. K. Oxenløwe, “0.4 THz Photonic-Wireless Link With 106 Gb/s Single Channel Bitrate,” J. Lightwave Technol. 36(2), 610–616 (2018). [CrossRef]
29. W. Li, J. Yu, B. Zhu, J. Zhang, M. Zhu, F. Zhao, T. Xie, K. Wang, Y. Wei, X. Yang, B. Hua, M. Lei, Y. Cai, W. Zhou, and J. Yu, “Photonics-assisted 320 GHz THz-band 50 Gbit/s Signal Outdoor Wireless Communication over 850 Meters,” in 2023 Optical Fiber Communications Conference and Exhibition (OFC), 2023, pp. 1–3.
30. Y. Horst, T. Blatter, L. Kulmer, et al., “Transparent Optical-THz-Optical Link at 240/190 Gbit/s over 5/115 m Enabled by Plasmonics,” Journal of Lightwave Technology, 2022.
31. S. Jia, L. Zhang, S. Wang, W. Li, M. Qiao, Z. Lu, N. Idrees, X. Pang, H. Hu, X. Zhang, L. K. Oxenloewe, and X. Yu, “2×300 Gbit/s Line Rate PS-64QAM-OFDM THz Photonic-Wireless Transmission,” Journal of Lightwave Technology, p. 1, 2020.
32. X. Li, J. Yu, L. Zhao, K. Wang, C. Wang, M. Zhao, W. Zhou, and J. Xiao, “1-Tb/s Millimeter-Wave Signal Wireless Delivery at D-Band,” J. Lightwave Technol. 37(1), 196–204 (2019). [CrossRef]
33. Y. Wang, W. Li, J. Ding, J. Zhang, M. Zhu, F. Zhao, M. Wang, and J. Yu, “Integrated High-Resolution Radar and Long-Distance Communication Based-on Photonic in Terahertz Band,” J. Lightwave Technol. 40(9), 2731–2738 (2022). [CrossRef]
34. J. L. Hesler and T. W. Crowe, “NEP and responsivity of THz zero-bias Schottky diode detectors,” in 2007 Joint 32nd International Conference on Infrared and Millimeter Waves and the 15th International Conference on Terahertz Electronics, 2007, pp. 844–845.
35. A. Gadallah, M. H. Eissa, T. Mausolf, D. Kissinger, and A. Malignaggi, “A 300-GHz Low-Noise Amplifier in 130-nm SiGe SG13G3 Technology,” IEEE Microw. Wireless Compon. Lett. 32(4), 331–334 (2022). [CrossRef]
36. A. Tessmann, A. Leuther, H. Massler, V. Hurm, M. Kuri, M. Zink, M. Riessle, and R. Lösch, “High-gain submillimeter-wave mHEMT amplifier MMICs,” in 2010 IEEE MTT-S International Microwave Symposium, 2010, pp. 53–56.
37. A. Tessmann, A. Leuther, S. Wagner, H. Massler, M. Kuri, H. P. Stulz, M. Zink, M. Riessle, and T. Merkle, “A 300 GHz low-noise amplifier S-MMIC for use in next-generation imaging and communication applications,” in 2017 IEEE MTT-S International Microwave Symposium (IMS), 2017, pp. 760–763.
38. T. P. McKenna, J. A. Nanzer, and T. R. Clark, “Photonic downconverting receiver using optical phase modulation,” in 2014 IEEE MTT-S International Microwave Symposium (IMS2014), 2014, pp. 1–3.
39. C. Yang, X. Li, J. Xiao, N. Chi, and J. Yu, “Fiber-wireless integration for 80 Gbps polarization division multiplexing− 16QAM signal transmission at W-band without RF down conversion,” Microwave Opt. Technol. Lett. 57(1), 9–13 (2015). [CrossRef]
40. Y. Salamin, W. Heni, C. Haffner, Y. Fedoryshyn, C. Hoessbacher, R. Bonjour, M. Zahner, D. Hillerkuss, P. Leuchtmann, and D. L. Elder, “Direct conversion of free space millimeter waves to optical domain by plasmonic modulator antenna,” Nano Lett. 15(12), 8342–8346 (2015). [CrossRef]
41. M. Burla, C. Hoessbacher, W. Heni, C. Haffner, Y. Fedoryshyn, D. Werner, T. Watanabe, H. Massler, D. L. Elder, L. R. Dalton, and J. Leuthold, “500 GHz plasmonic Mach-Zehnder modulator enabling sub-THz microwave photonics,” APL Photonics 4(5), 056106 (2019). [CrossRef]
42. Y. Salamin, B. Baeuerle, W. Heni, F. C. Abrecht, A. Josten, Y. Fedoryshyn, C. Haffner, R. Bonjour, T. Watanabe, M. Burla, D. L. Elder, L. R. Dalton, and J. Leuthold, “Microwave plasmonic mixer in a transparent fibre–wireless link,” Nat. Photonics 12(12), 749–753 (2018). [CrossRef]
43. S. Ummethala, T. Harter, K. Koehnle, Z. Li, S. Muehlbrandt, Y. Kutuvantavida, J. Kemal, P. Marin-Palomo, J. Schaefer, A. Tessmann, S. K. Garlapati, A. Bacher, L. Hahn, M. Walther, T. Zwick, S. Randel, W. Freude, and C. Koos, “THz-to-optical conversion in wireless communications using an ultra-broadband plasmonic modulator,” Nat. Photonics 13(8), 519–524 (2019). [CrossRef]
44. I. Goldstein, P. A. Miles, and A. Chabot, “Heterodyne measurements of light propagation through atmospheric turbulence,” Proc. IEEE 53(9), 1172–1180 (1965). [CrossRef]
45. D. L. Fried, “Aperture Averaging of Scintillation,” J. Opt. Soc. Am. 57(2), 169–175 (1967). [CrossRef]
46. D. Fink and S. N. Vodopia, “Coherent detection SNR of an array of detectors,” Appl. Opt. 15(2), 453–454 (1976). [CrossRef]
47. P. Gatt, T. P. Costello, D. A. Heimmermann, D. C. Castellanos, A. R. Weeks, and C. M. Stickley, “Coherent optical array receivers for the mitigation of atmospheric turbulence andspeckle effects,” Appl. Opt. 35(30), 5999–6009 (1996). [CrossRef]
48. A. R. Weeks, J. Xu, R. R. Phillips, L. C. Andrews, C. M. Stickley, G. Sellar, J. S. Stryjewski, and J. E. Harvey, “Experimental verification and theory for an eight-element multiple-aperture equal-gain coherent laser receiver for laser communications,” Appl. Opt. 37(21), 4782–4788 (1998). [CrossRef]
49. A. Belmonte, “Capacity of coherent laser downlinks,” J. Lightwave Technol. 32(11), 2128–2132 (2014). [CrossRef]
50. D. M. Boroson, S. R. Bryan, V. M. Daniel, A. B. Dennis, K. Farzana, M. K. Joseph, S. Zoran, and M. C. Donald, “Overview and results of the Lunar Laser Communication Demonstration,” Proc. SPIE 8971, 89710S (2014). [CrossRef]
51. J. Ma, K. Li, L. Tan, S. Yu, and Y. Cao, “Performance analysis of satellite-to-ground downlink coherent optical communications with spatial diversity over Gamma-Gamma atmospheric turbulence,” Appl. Opt. 54(25), 7575–7585 (2015). [CrossRef]
52. D. Geisler, T. Yarnall, M. Stevens, C. Schieler, B. Robinson, and S. Hamilton, “Multi-aperture digital coherent combining for free-space optical communication receivers,” Opt. Express 24(12), 12661–12671 (2016). [CrossRef]
53. D. Geisler, T. Yarnall, C. Schieler, M. Stevens, B. Robinson, and S. Hamilton, “Experimental demonstration of multi-aperture digital coherent combining over a 3.2-km free-space link,” Proc. SPIE 10096, 100960C (2017). [CrossRef]
54. D. G. Brennan, “Linear Diversity Combining Techniques (in English),” Proc. IRE 47(6), 1075–1102 (1959). [CrossRef]
55. A. Josten, B. Baeuerle, E. Dornbierer, J. Boesser, D. Hillerkuss, and J. Leuthold, “Modified godard timing recovery for non integer oversampling receivers,” Appl. Sci. 7(7), 655 (2017). [CrossRef]
56. K. Kikuchi and S. Tsukamoto, “Evaluation of Sensitivity of the Digital Coherent Receiver,” J. Lightwave Technol. 26(13), 1817–1822 (2008). [CrossRef]
