Open Access
Add to BibSonomy
Abstract
A reconfigurable optical-to-electrical signal aggregation is proposed, for the first time, using optical signal processing and photo-mixing technology. Two optically modulated quadrature phase-shift keying (QPSK) signals are aggregated into a single 16-quadrature amplitude modulation (16-QAM) signal and, simultaneously, carried over a 28-GHz millimeter wave (MMW) carrier using an optimized heterodyne beating process through a single photodiode. To demonstrate the system reconfigurability, aggregation of two optical binary phase-shift keying signals is mapped into MMW QPSK or four-level pulse amplitude modulation signals by controlling the relative phases and amplitudes, respectively, of the input signals. In addition, the aggregation of two 16-QAM signals into a 256-QAM signal and the aggregation of three QPSK signals into a 64-QAM format are achieved. Besides, we report the effect of laser phase noise on signal aggregation performance. The de-aggregation of the aggregated MMW signals is performed electrically using a successive interference cancellation algorithm. Moreover, a proof-of-concept experiment is conducted to validate the numerical simulations. Two 1-Gbaud BPSK (1 Gbps) and QPSK (2 Gbps) optical signals are optically transmitted over a 20-km single-mode fiber as MMW over fiber signals. Then, the signals are aggregated into QPSK (2 Gbps) and 16-QAM (4 Gbps) 28-GHz MMW signals, respectively. The aggregated signal is further transmitted over a 1-m wireless channel. The performance of the proposed system is evaluated using bit error rate and error vector magnitude metrics.
© 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
The ever-increasing demands for bandwidth-hungry applications, such as ultra-high definition (UHD) video streaming, the Internet of Things (IoT), big data, cloud computing, and fifth or sixth generation (5G and 6G) services are constantly increasing [1]. These demands have been met by new technologies and methods in the fields of photonics and millimeter waves (MMWs). For instance, 200 gigabit Ethernet (GE) and 400 GE have recently become standardized, and their optical components and devices are easily available [2]. In addition, future Ethernet line rates of 800 G and 1.6 T are under consideration and might become available by 2025 [3]. Besides, 5G networks have already been deployed in many countries, and the research wave and vision have been started towards 6G wireless networks, which promise 50 times the peak data rate (i.e., $\sim$ T bps) of the 5G counterpart and potential operating frequency windows of 0.245 THz up to 1.03 THz [4]. The interaction between photonics and MMW technologies is of particular interest in many commercial and military applications. Examples are photonic generation and processing of MMW signals [5–7], radio-over-fiber (RoF) systems [8,9], and photonic-based radio frequency (RF) spectrum analysis [10]. However, each of these technologies has limitations. For instance, MMW operating frequency windows suffer from atmospheric attenuation, which can be as high as 100 dB/km at the THz window [4]. Whereas, optical transmission systems experience fiber channel non-linearities resulting from the transmission of power-hungry modulation formats (i.e., M-ary quadrature amplitude modulation (QAM), M = 64 and 256). These issues limit the reach distances and spectral efficiency (SE) of MMW and photonics systems, respectively.
In this regard, many researchers have demonstrated transporting MMW signals for last-mile access networks over fiber (MMWoF), owing to the low attenuation of the fiber channel [8,9,11–13]. For example, a 25-GHz MMW signal was optically generated and transmitted over a hybrid 5-km single-mode fiber (SMF), a 2-m free-space optics (FSO), and a 3.3-m wireless channel [9]. In addition, advanced laser sources have been utilized to generate MMW signals in 28-GHz [8] and 60-GHz [12] bands, where the system performance was investigated over a hybrid channel of 10-km SMF and indoor wireless distances. In [13], the simultaneous generation and transmission of 50-GHz MMW, 100-GHz sub-Tera, and 150-GHz sub-Tera bands was demonstrated. A hybrid channel of 20-km SMF, 500-m FSO, and up to 2-m of wireless channels was built to assess the transmission success. On the other side, optical signal processing (OSP) and non-orthogonal optical data aggregation methods have attracted interest as a format conversion tool to increase SE in traditional optical networks by generating higher order modulation formats from lower order formats. In addition, they could manage modulation formats between different optical sub-networks, with rapid operation and low latency [14,15]. Therefore, integrating OSP functionalities and non-orthogonal data aggregation in MMWoF systems would enhance the SE in these systems. Specifically, OSP allows (1) operating at the optical line rate; (2) enabling a high degree of tunability owing to data transparency; and (3) processing multiple dimensions of the optical wave – such as amplitude, phase, wavelength, and polarization [15]. Moreover, optical frequency combs (OFCs) have become an enabling tool for advanced OSP functions. OFC is a source that generates many (i.e., tens to hundreds) coherent equidistance optical lines with low phase noise. OFC has been used for all-optical and photonics-based RF functions, including optical filtering, format conversion, RF spectrum analysis, and MMW generation [10–13].
Previously, all-optical non-orthogonal data aggregation has been proposed using optical-to-optical (O/O) nonlinear devices. This includes periodically poled lithium-niobate (PPLN) [16–18], semiconductor optical amplifier (SOA) [19], and high nonlinear fiber (HNLF) [20–24]. For instance, in [16], six high-order modulation formats of up to 64-QAM were generated by aggregating low-order modulation formats in two PPLN non-linear stages, which achieved the coherence and multiplexing of aggregated channels. The system complexity was further reduced in [17] by replacing one PPLN non-linear stage with a coherent frequency comb source, where three high-order modulation schemes were considered. In [18], the authors used the PPLN to perform optical aggregation on Nyquist channels, generated using a micro-resonator and intensity modulator, to yield a 16-QAM signal. In [19], an 8-QAM format was obtained optically from independent on-off keying (OOK) and QPSK formats; the authors utilized cross-phase and gain modulation (XPM and XGM) in SOA. In [20], the authors used numerical simulations to show the feasibility of aggregating four OOK signals into a 16-QAM signal using two HNLFs in an optical loop mirror configuration. The format conversion was achieved through two conversion stages. First, the OOK signals were converted into two QPSK signals by phase modulation in one HNLF. Then, the 16-QAM signal was obtained by amplitude modulation in the second HNLF. Similarly, in [21], numerical simulations were used to aggregate three binary phase-shift keying signals into an 8-QAM signal. The aggregation was accomplished using three conversion stages and a single HNLF. In [22,23] and [24], HNLF was used with four-wave mixing (FWM) and XPM, respectively, for signal aggregation to generate 8-QAM and dual-polarization QPSK signals, respectively. Moreover, in [25], numerical and experimental investigations were developed to report spectrum sharing and combining of two asynchronous QPSK signals of 20 and 4 or 9-Gbaud speeds, carried over the same wavelength, using power division non-orthogonal multiple access (NOMA) to generate a hybrid-QAM optical signal. NOMA has been proposed for 5G systems to efficiently utilize the modulation bandwidth by exploiting users’ power superposition [26]. The recovery of the multiplexed signals can be achieved sequentially using the successive interference cancellation (SIC) method at the receiver side.
So far, the above methods have employed O/O non-orthogonal data aggregation, also known format conversion. To our knowledge, no work has yet been reported performing optic-to-electric (O/E) non-orthogonal format conversion for both data aggregation and MMW generation simultaneously in MMWoF systems. In O/O data aggregation methods, two separate systems are needed to perform the data aggregation followed by MMW generation. However, in O/E data aggregation, only one device is used to perform both functions. In this paper, we address several issues. First, we present an O/E signal aggregation and MMW communication system, at 28-GHz, to transmit aggregated data steams using spectrally efficient modulation formats. An aggregated 16-QAM format is generated over a 28-GHz MMW carrier from two optically modulated QPSK signals. Second, we demonstrate OSP reconfigurability through the generation of QPSK and PAM-4 signals from two BPSK signals, by controlling the phase and amplitude, respectively, of optical signals. In addition, we report the aggregation of two 16-QAM and three QPSK optical signals to generate MMW 256-QAM and 64-QAM signals, respectively. Third, we report the effect of laser phase noise (PN) on the aggregation performance. Finally, we conduct a proof-of-concept experiment to validate the numerical simulation. This work utilizes the tunability and reconfigurability of optical processing for efficient and cost-effective MMW transmission.
The remaining sections of this paper are organized as follows. Section 2 describes the concept and methods. Section 3 presents the theory of O/E signal aggregation and MMW generation. In Section 4, the numerical simulation setup and results are analyzed. A proof-of-concept O/E aggregation experiment is shown in Section 5. Finally, the paper conclusion is given in Section 6.
2. Concept and methods
2.1 Concept
Figure 1 shows the concept of O/E data aggregation and MMW generation. Each user on the left side of the figure generates a data signal in a low modulation format (e.g., BPSK). For each user, the optical carrier is obtained from a common OFC source. The data signals arrive at the data aggregation unit, where fiber deployment is not feasible. Aggregation and MMW transmission (e.g., at 28-GHz) are achieved in the data aggregation unit using OSP. This happens by controlling the amplitude and phase of each incoming signal and beating signals on a photodiode, to sum the signals together at the MMW carrier frequency. The Advantages of this architecture are two fold (i) generating a MMW signal by all-optical conversion where optical fiber deployment is not feasible; and (ii) increasing SE over a MMW wireless channel by aggregating different signals. The MMW output of the aggregation unit (i.e., photodiode) is then radiated over a beam-focusing antenna. At the de-aggregation unit, the signal is either processed and distributed among the receiving users, using a SIC algorithm, or it is optically modulated and re-transmitted over a fiber channel. The SIC algorithm is explained in the next sub-sections.
Fig. 1. Architecture of mobile access network with O/E data aggregation and MMW generation. OFC: optical frequency comb, MMW: millimeter wave, Agg.: aggregated, and De-agg.: de-aggregated.
2.2 Methods of aggregation
The aggregation process is further explained in Fig. 2. We used an OFC to generate multiple equidistant optical carriers. The OFC considered here is an electro-optic OFC, which we built using an off-the-shelf laser source, intensity modulator, together with an RF signal generator [15]. This is a frequency-locked and coherent OFC in which the comb spacing can be tuned by adjusting the RF frequency. The number of obtained sub-carriers and the comb bandwidth are dependent on the RF power and modulator bandwidth, respectively. The obtained optical sub-carriers are sent to the OSP unit to select four sub-carriers (i.e., for two signals to be aggregated). Here, the sub-carriers $f_1 = f_0-2f_r$ and $f_3=f_0+f_r$ exit the OSP unit from port 1, whereas sub-carriers $f_2=f_0-f_r$ and $f_4=f_0+2f_r$ exit from port 2. Note that the selected lines are chosen such that the frequency difference between $f_1$ and $f_3$ is the same as that between $f_2$ and $f_4$. Furthermore, this frequency difference defines the carrier frequency that will be emitted as a MMW from the photodiode. Each of the carriers exiting from port 1 is modulated with low-order modulation data, representing signals from different users. The modulated data are then combined with the un-modulated lines from the lower path (port 2) in the optical modulation and sub-carriers combing unit. Therefore, at the output of this unit, each un-modulated carrier has a data signal adjacent to it. In the photodiode, a heterodyne beating process occurs, which generates the electrical MMW signal at a predefined frequency ($f_1$-$f_3$ or $f_2$-$f_4$). It is worthy noting that the first three units of the proposed approach are all photonics in which sub-carriers generation, signal reconfigurability, and optical modulation are achieved. The fourth element is the O/E photo-mixer, which performs data aggregation and MMW generation. By controlling the modulation format of the incoming signals as well as the amplitude and phase of each data channel, we can achieve different aggregation sensations. For instance, Fig. 2(b) and (d) illustrate the aggregation of optical QPSK and BPSK signals to generate electrical 16-QAM and PAM-4 signals, respectively; here, we controlled the relative amplitudes of the aggregated signals. By contrast, Fig. 2(c) shows the aggregation of two equal power optical BPSK signals to generate a QPSK signal by controlling the modulated line phases. In both cases, the generated electrical constellation points, shown in blue, are generated by the vector addition of two optical constellation points (green and orange) of different amplitudes (i.e., Fig. 2(b) and (d)) or same amplitude but different phases (i.e., Fig. 2(c)).
Fig. 2. (a) block diagram of the O/E signal aggregation and MMW generation, (b) aggregation of two optical QPSK into a MMW 16-QAM format by controlling optical amplitudes, (c) aggregation of two optical BPSK into a MMW QPSK by controlling optical phases, and (d) aggregation of two BPSK into a MMW PAM-4 scheme by controlling optical amplitudes.
2.3 Methods of de-aggregation
We used the basic non-orthogonal multiple access (NOMA) scheme with SIC for the de-aggregation of $N$ users, as illustrated in Fig. 3. The transmit signal for user $U_i$, $i = 1, 2,\ldots N$ is $x_i=\sqrt {p_i}s_i$, where $p_i$ is the transmission power and $s_i$ is the user information. The aggregated signals for the two-user scenario are superposed as $x=x_1+x_2=\sqrt {p_1}s_1+\sqrt {p_2}s_2$. NOMA allows users to share the same resources, and differentiates users by their power. The received signal at the $i^{th}$ terminal is $y_i=h_ix+n_i$, where $h_i$ is the complex channel coefficient between the $i^{th}$ terminal and the transmitter, and $n_i$ denotes the receiver’s Gaussian noise. In the SIC algorithm, the decoding process is achieved in a descending order of the received signal power. That is, the interference cancellation is performed for users with lower power to cancel the inter-user interference. After that, users of low power can detect the desired information from the updated received signal. For example, for the de-aggregation of two users, assuming $p_1>p_2$, terminal 1 decodes $x_1$ from the received signal $y_1$ without interference cancellation, since $U_1$ has the "strong" signal power and $U_2$ has the "weak" signal power. Then, terminal 2 deletes the decoded data of $U_1$ from the received signal $y_2$ to decode $x_2$, as shown in Fig. 3 (inset i). The SIC process is continued until all interference signals have been cancelled, as shown in Fig. 3 (inset ii).
Fig. 3. Basic NOMA scheme applying SIC algorithm. Inset i shows SIC for user 1 signal and inset ii shows SIC for users 1, 2,…$N$-1 signals. SIC: successive interference cancellation.
3. Theory and operation
This section discusses the theory and operation of the O/E signal aggregation and MMW generation. First, an external distributed feedback (DFB) laser generating a continuous wave (CW) optical signal, of an average power $P_0$ and an emission frequency of $f_0$ THz, is connected to a mach Zehnder modulator (MZM) to generate an OFC signal. Assume that the CW laser output is given as:
Then, the output of MZM can be written as [27]:4. Numerical simulation
4.1 Simulation setup
The simulation setup is shown in Fig. 4. The setup was developed using VPITransmissionMaker Ver. 11.1. It comprises four main parts: OFC, OSP, MMWoF, and data aggregation and signal analysis. The OFC was built using a CW optical laser source with an emission frequency of $f_0=193.1$ THz and a single intensity modulator driven by an RF signal of $f_r$ = 28-GHz. This generates equi-spaced optical sub-carriers spaced at 28-GHz spacing, as shown in Fig. 4(a). The OSP was built using a bank of optical band pass filters, attenuators, and optical phase shifters to control the phase and amplitude of each sub-carrier. For two-signal aggregation, four sub-carriers were selected by the OSP and fed into the the MMWoF part, as shown in Fig. 4(b). Here, left first-order and right second-order sub-carriers were used for data modulation, whereas left second-order and right first-order sub-carriers remained un-modulated. The data sub-carriers were modulated separately using two IQ modulators, with two different 1-Gbaud multilevel electrical signals. Then, the optimized phase and amplitude optical signals were combined to form a MMWoF signal as shown in Fig. 4(c). Finally, the O/E aggregation and MMW generation were achieved using a high-speed photodiode. The aggregated high-order multilevel signal was generated over a MMW carrier of 28-GHz. Figure 4(d) shows the electrical spectrum of the obtained aggregated signal. To reconstruct the original individual signals, SIC de-aggregation process was executed offline using MATLAB. Then, signal analysis was performed using the bit error rate (BER) and error vector magnitude (EVM) measures.
Fig. 4. Simulation setup of the O/E data aggregation and MMW generation. (a) OFC output, (b) OSP output spectrum, (c) formed MMWoF signal, and (d) obtained RF spectrum of aggregated signal. LD: laser diode, IM: intensity modulator, OFC: optical frequency comb, OSP: optical signal processing, OBPF: optical band pass filter, OA: optical attenuator, OP: optical phase adjust, PN: phase noise, PM: phase modulator, AWGN: additive white Gaussian noise, MMWoF: millimeter wave over fiber, IQM: optical IQ modulator, OC: optical coupler, EDFA: Erbium doped fiber amplifier, PD: photodiode, SIC: successive interference cancellation, EVM: error vector magnitude, and BER: bit error rate.
4.2 Aggregation performance
Two base-band BPSK signals of 1-Gbaud are generated and modulated at frequencies of $f_2=193.072$ THz and $f_4=193.156$ THz. Then, the optical modulated BPSK signals are passed through phase shift elements to achieve a phase difference of $\pi /2$ between the two copies. After that, the un-modulated lines at frequencies of $f_1=193.044$ THz and $f_3 =193.128$ THz are power controlled and combined with the modulated sub-carriers. The entire MMWoF signals are aggregated to generate a QPSK signal of 2 Gbps, as shown in Fig. 5(a), over a 28-GHz MMW carrier. Figure 5(b) illustrates the aggregation of two optical BPSK signals of different amplitudes and zero phase shift, which generated a PAM-4 signal of 4 Gbps. It is worth noting that the PAM-4 signal is obtained after adjusting the power ratio between the two optical BPSK at 85%.In addition, two optical QPSK(16-QAM) signals of 1-Gbaud are generated and modulated at the same frequencies of $f_0-f_r$ and $f_0+2f_r$ THz. The power ratio between the two signals is adjusted at 85%(95%) to obtain an electrical aggregated 16-QAM(256-QAM) signal of 4 Gbps(8 Gbps), with equi-distances constellation diagrams. Figure 5(c) and (d) show the obtained constellation diagrams of the generated 16-QAM and 64-QAM, respectively. Furthermore, we succeeded in aggregating three signals, as shown in Fig. 5(e), using three optical QPSK signals modulated at optical sub-carriers of $f_0-f_r$, $f_0+2f_r$, and $f_0+5f_r$ THz. This step produced a MMW 64-QAM of 6 Gbps. It is noteworthy that a SIC algorithm was used for signal de-aggreagation, as shown in Fig. 5(b) to (d), whereas simple two correlators with orthogonal basis functions and two thresholds were used to obtain the de-aggregated signals shown in Fig. 5(a). In Fig. 6, we evaluated the aggregation performance by plotting BER and EVM versus the optical received power of the de-aggregated signals. It is evident that the BER values of the BPSK and QPSK signals, de-aggregated from the less challenging modulation schemes (i.e. less affected by system noise) such as QPSK and 16-QAM, are strongly reduced, with an increase in the received optical power. Whereas, the BER values of BPSK, QPSK, and 16-QAM signals, de-aggregated from the more challenging modulation formats (i.e. highly affected by system noise) such as PAM-4, 64-QAM and 256-QAM, are gradually reduced, with an increase in the received optical power. For –23 dBm received power, the BER values are lower than the forward error correction (FEC) limit of 3.8 $\times$ $10^{-3}$ and record 1$\times$ 10$^{-3}$ and 3$\times$ 10$^{-10}$ for BPSK formats de-aggregated from PAM-4 and QPSK aggregated signals, respectively. In addition, the de-aggregated signals attain 1$\times$ 10$^{-6}$ and 3$\times$10$^{-3}$ for QPSK formats that de-aggregated from 16-QAM and 64-QAM signals, respectively. However, they exceed the FEC limit for the 16-QAM format de-aggregated from the 256-QAM signal. This highlights the potential of realizing flexible multiformat candidate integrated photonics and wireless architectures.
Fig. 5. Received constellation, at -14 dBm received optical power, of the individual channels, the aggregated signals, and the de-aggregated signals (a) QPSK aggregation of two BPSK users, (b) PAM-4 aggregation of two BPSK users, (c) 16-QAM aggregation of two QPSK users, (d) 256-QAM aggregation of two 16-QAM users, and (e) 64-QAM aggregation of three QPSK users.
4.3 Effect of laser PN on data aggregation
In the previous section, we considered a low phase noise analysis, which is the case of OFC. However, phase noise may occur, causing a penalty, which we investigate in this section. The phase error can happen owing to phase offset, frequency offset, and laser phase noise. This effect and can be compensated to enable extracting the in-phase and quadrature information. Phase estimation is a commonly used method in wireless and optical systems to compensate for phase noise. Conventionally, algorithms like the Viterbi-Viterbi and maximum likelihood phase estimation assume that the data is embedded in white Gaussian noise (WGN) and corrupted by a quasi-constant phase offset [28]. These algorithms estimate this constant phase offset out of the WGN by averaging over a window with several symbols. As the phase error can change dynamically from symbol to symbol due to the strong laser phase noise [28], special algorithms must be developed. In this paper, we model the effects of phase noise in the system by modulating the phase of the modulated sub-carriers with a random Gaussian noise.
We aggregated two optical QPSK channels into a MMW 16-QAM signal while introducing various PN effects. The PN is simulated by adding an optical white noise source, generated by a phase modulator and WGN source (shown as a dashed box in Fig. 4), to the data sub-carriers, with noise power densities ranging from $1\times 10^{-12}$ to $10\times 10^{-12}$ W/Hz. First, the system performance is represented in terms of signal constellations for aggregated signals with a PN effect, aggregated signals after PN compensation, and de-aggregated signals, as shown in Fig. 7, for different power densities. Then, the Viterbi-Viterbi algorithm is implemented at the receiver side to mitigate the laser PN in the aggregated signals. The EVM and BER performances are compared in Fig. 8 for 16-QAM and QPSK signals after aggregation and de-aggregation, respectively, for various PN densities. At the FEC limit, the aggregated(de-aggregated) 16-QAM(QPSK) signal displayed 0.4(0.3) and 2.1(2) dB penalties at PN power densities of $5\times 10^{-12}$ and $10\times 10^{-12}$ W/Hz, respectively, with respect to the low PN power density of $1\times 10^{-12}$ W/Hz. Figure 9 illustrates the BER and EVM results of the de-aggregated QPSK signals after PN correction using the Viterbi-Viterbi algorithm. More than an order of magnitude enhancement in the BER values is evident for the highest PN power density of $10\times 10^{-12}$ W/Hz.
Fig. 7. Received constellation of the individual QPSK channels, aggregated 16QAM signals, and the de-aggregated channels for different PN power densities.
Fig. 8. BER and EVM performance of aggregated 16-QAM and de-aggregated QPSK signals with the effect of PN.
Fig. 9. BER and EVM performance of QPSK signal after de-aggregation with the effect of phase noise and phase noise correction algorithm.
5. Experimental setup and results
Figure 10 shows the experimental setup of the O/E data aggregation and MMW generation system. The OFC circuit we employed is built using an off-the-shelf laser source, optical modulator, and RF analog signal generator. A 1550-nm distributed feedback (DFB) fiber laser source (NKT-Photonics Koheras ADJUSTIK) with a low phase noise and an Hz-range linewidth is used to generate a 15-dBm output power continuous wave (CW) signal. A polarization controller (PC) is engaged to maintain signal polarization. The CW laser signal is applied to a 40-GHz (EOspace AZ-DV5-40-PFU-SFU-LV) C-band mach Zehnder intensity modulator (IM) driven by a 28-GHz RF signal, which is generated by an analog signal generator (Keysight E8257D) with a measured phase noise of -100 dBc/Hz at 1-kHz offset, which provided negligible effect on the system performance. The RF source output power is adjusted such that the IM output spectrum is as shown in Fig. 10 inset P1, where the middle sub-carrier is suppressed. The OFC signal is sent to a liquid crystal on Silicon (LCoS) wave selective switch (WSS, Finisar 4000S), which controls the phase and amplitude of each sub-carrier – that is, the WSS represents the OSP module. The WSS first port (lower arm) carries sub-carriers (–2) and (+1) which remain un-modulated, as shown in Fig. 10 inset P2. Whereas, the second output port of the WSS (upper arm) carries sub-carriers (–1) and (+2), which are amplified using Amonics Erbium dipped fiber amplifier (EDFA-1, Amonics AEDFA-BO-23-B-FA), to compensate for the WSS and IM insertion losses. Then, the amplified sub-carriers are used for optical modulation. An arbitrary waveform generator (AWG, Keysight M8195A) is used to generate the various 1-Gbaud in-phase quadrature signals, obtained via pseudo-random binary sequence (PRBS) with a length of $2^{11}-1$ and root raised cosine (RRC) pulse shaping filter of 0.35 roll-off factor (i.e., 1.35 Ghz signal bandwidth), to modulate both sub-carriers (–1) and (+2) using an IQ modulator (IQM, Fujitsu FTM7977HQA), as shown in Fig. 10 inset P3. Then, the modulated signals are applied to a dispersion-compensating fiber (DCF) to de-correlate both optical sub-carriers. A 50:50 optical coupler (OC) is used to combine the modulated and un-modulated sub-carriers to form a MMWoF optical signal, as shown in Fig. 10 inset P4. In addition, a PC is used at the DCF output to maintain the mode coherency of the beat-tones. The MMWoF signal is amplified using EDFA-2 (Amonics AEDFA-PA-40-B-FA) and transmitted over 20-km SMF before being applied to a high-speed 70-GHz photodiode (Finisar XPDV3120) for the hetrodyne beating process. Note that EDFA-2 amplifies both aggregated signals simultaneously, however it retains the features (i.e., phase and amplitude differences) imposed on each sub-carrier by the WSS. Furthermore, a tunable optical attenuator (Agilent N7764A) is used to control the received optical power. The aggregated and MMW generated signals are amplified using a low noise amplifier (QuinStar QLW-24403336-J0) and transmitted over a 1-m wireless channel using two horn antennas (SAGE SAR-2507-28-S2). It should be noted that the wireless channel distance is limited by the constraints of the available indoor laboratory space. Then, the MMW signals are received by a high bandwidth digital storage oscilloscope (Keysight DSOX 932048), which has a bandwidth of 32-GHz and a speed of 80-GSa/sec. The received signals are analyzed using vector signal analyzer software (Keysight VSA 89600). Several digital signal processing (DSP) steps are performed before signal demodulation. These include signal down-conversion, carrier and clock recovery, baseband filtering with an RRC filter of 0.35 roll-off factor, and adaptive equalization for linear distortions compensation.
Fig. 10. Experimental setup of the O/E aggregated system. LD: laser diode, IM: intensity modulator, OFC: optical frequency comb, PC: polarization controller, WSS: wave selective switch, EDFA: Erbium doped fiber amplifier, AWG: arbitrary waveform generator, IQM: optical IQ modulator, DCF: dispersion compensating fiber, OC: optical coupler, VOA: variable optical attenuator, SMF: single-mode fiber, PD: photodiode, and DSO: digital storage oscilloscope.
The results of experimental aggregation are shown in Fig. 11. We report two aggregation scenarios here. That is, 1-Gbaud QPSK (2 Gbps) and 16-QAM (4 Gbps) MMW modulation formats over 28-GHz are generated by the optical aggregation of two optical BPSK and QPSK signals, respectively. Specifically, for the aggregated QPSK signal, the phase difference of the modulated sub-carriers (i.e., individual BSPK signals) was adjusted to $\pi$/2 via the WSS. By contrast, for 16-QAM data aggregation, the amplitude of the modulated sub-carriers (i.e., individual QPSK signals) was controlled by the WSS to obtain symmetric square modulation. Fig. 11(a) and (b) show the aggregation performance for 20-km of SMF and 1-m of wireless channel in terms of the measured EVM and the calculated BER values versus the optical received power for the aggregated QPSK and 16-QAM signals. The back-to-back (BtB) traces show a negligible penalty for the transmission channel. It is worth mentioning that the difference in receiver sensitivity between the simulation and experimental work was the result of noise stress that was added by optical devices, such as EDFAs and PD, as well as electrical equipment, such as AWG and DSO. For the QPSK aggregated signal, the BER exceeded the FEC limit at a receiver sensitivity of $\sim$–8.2 dBm. The receiver sensitivity was degraded to $\sim$–2.2 dBm for the aggregated 16-QAM signal, at the FEC limit, hence there was a power penalty of 6 dB. Figure 11(c) and (d) show the aggregated constellations at a received power of –5 to –8 dBm for QPSK and 3 to –2 dBm for 16-QAM signals, respectively. It is to be noted that the received constellation points showed a marginal phase noise effect owing to the low linewidth of the used laser source.
Fig. 11. Performance of aggregated 1-Gbaud QPSK and 16-QAM MMW signals, with 20-km SMF and 1-m wireless channel, in terms of (a) calculated BER, (b) measured EVM, (c) aggregated QPSK constellations, and (d) aggregated 16-QAM constellations. The back-to-back (BtB) BER and EVM results are shown for comparison.
6. Conclusion
An optical-to-electrical (O/E) aggregation and MMW generation method is proposed and demonstrated through numerical simulation and experimental demonstration. Using the proposed method, we demonstrated aggregating two and three optical data streams, having data rates of up to 4 Gbps in lower formats, into a MMW higher format signal on a 28-GHz carrier, with data rates of up to 8 Gbps. In addition, the aggregated MMW signal was de-aggregated in simulation with the SIC algorithm and streamed back to the various users. We studied the performance of the proposed method in terms of BER, EVM, and constellation diagrams for the aggregated and de-aggregated data. The results showed that the effect of laser PN degraded the aggregation performance. However, compensation techniques can enhance the BER performance by more than an order of magnitude. Proof-of-concept experiments successfully demonstrated the aggregation of two signals to generate higher order schemes, such as QPSK and 16-QAM modulations. This work paves the way for the future potential aggregation of multiple signals to generate higher SE formats, such as 64- and 256-QAM. Furthermore, thanks to the flexibility of photonic systems, the proposed approach can be extended to higher frequency bands, such as the next-generation communication bands (i.e., V- and W-bands). Hence, developing such a system would reinforce the feasibility and potential use of photonics-based non-orthogonal data aggregation in 5G and 6G applications.
Funding
National Plan for Science, Technology and Innovation (MAARIFAH), King Abdulaziz City for Science and Technology Kingdom of Saudi Arabia, Award Number (2-17-02-001-0009).
Disclosures
The authors declare that there are no conflicts of interest related to this article.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. Cisco white paper, “Cisco annual Internet report,” https://www.cisco.com/c/en/us/solutions/executive-perspectives/annual-internet-report/index.html. Last accessed: 19-04-2022.
2. P. J. Winzer, D. T. Neilson, and A. R. Chraplyvy, “Fiber-optic transmission and networking: the previous 20 and the next 20 years Invited,” Opt. Express 26(18), 24190–24239 (2018). [CrossRef]
3. D. Rodgers, “2020 Ethernet alliance roadmap,” https://ethernetalliance.org/blog/2020/05/06/2020-ethernet-alliance-roadmap/. Last accessed: 19-04-2022.
4. H. Tataria, M. Shafi, A. F. Molisch, M. Dohler, H. Sjöland, and F. Tufvesson, “6G wireless systems: vision, requirements, challenges, insights, and opportunities,” Proc. IEEE 109(7), 1166–1199 (2021). [CrossRef]
5. Y. Li, A. Rashidinejad, J.-M. Wun, D. E. Leaird, J.-W. Shi, and A. M. Weiner, “Photonic generation of W-band arbitrary waveforms with high time-bandwidth products enabling 3.9 mm range resolution,” Optica 1(6), 446–454 (2014). [CrossRef]
6. H. Shams, M. J. Fice, L. Gonzalez-Guerrero, C. C. Renaud, F. van Dijk, and A. J. Seeds, “Sub-THz wireless over fiber for frequency band 220–280 GHz,” J. Lightwave Technol. 34(20), 4786–4793 (2016). [CrossRef]
7. F. Saadaoui, M. Fathallah, A. M. Ragheb, M. I. Memon, H. Fathallah, and S. A. Alshebeili, “Optimizing OSSB generation using semiconductor optical amplifier (SOA) for 5G millimeter wave switching,” IEEE Access 5, 6715–6723 (2017). [CrossRef]
8. M. Z. M. Khan, Q. Tareq, A. M. Ragheb, M. A. Esmail, and S. A. Alshebeili, “Bidirectional MMWoF-wireless convergence system based on a 1610 nm L-band quantum-dash laser,” Opt. Express 29(17), 27493–27507 (2021). [CrossRef]
9. D.-N. Nguyen, J. Bohata, M. Komanec, S. Zvanovec, B. Ortega, and Z. Ghassemlooy, “Seamless 25 GHz transmission of LTE 4/16/64-QAM signals over hybrid SMF/FSO and wireless link,” J. Lightwave Technol. 37(24), 6040–6047 (2019). [CrossRef]
10. Z. Tang, D. Zhu, and S. Pan, “Coherent optical RF channelizer with large instantaneous bandwidth and large in-band interference suppression,” J. Lightwave Technol. 36(19), 4219–4226 (2018). [CrossRef]
11. Q. Tareq, A. Ragheb, E. Alkhazraji, M. Esmail, S. Alshebeili, and M. Khan, “Hybrid 28 GHz MMW over fiber-wireless QPSK transmission system based on mid L-band external injection-locked quantum-dash laser comb source,” Opt. Fiber Technol. 64, 102553 (2021). [CrossRef]
12. D. Dass, A. Delmade, L. Barry, C. G. Roeloffzen, D. Geuzebroek, and C. Browning, “Flexible V-band mmwave analog-RoF transmission of 5G and WiGig signals using an InP-SiN integrated laser module,” in 2021 International Topical Meeting on Microwave Photonics (MWP), (2021), pp. 1–4.
13. H.-H. Lu, C.-Y. Li, W.-S. Tsai, P.-S. Chang, Y.-T. Chen, C.-X. Liu, T. Ko, and Y.-Y. Lin, “Simultaneous transmission of 5G MMW and sub-THz signals through a fiber-FSO-5G NR converged system,” J. Lightwave Technol. 40(8), 2348–2356 (2022). [CrossRef]
14. H. Wang, L. Pan, and Y. Ji, “All-optical aggregation and de-aggregation of 4× BPSK-16QAM using nonlinear wave mixing for flexible optical network,” IEEE J. Sel. Top. Quantum Electron. 27(2), 1–8 (2021). [CrossRef]
15. A. E. Willner, A. Fallahpour, K. Zou, F. Alishahi, and H. Zhou, “Optical signal processing aided by optical frequency combs,” IEEE J. Sel. Top. Quantum Electron. 27(2), 1–16 (2021). [CrossRef]
16. M. R. Chitgarha, S. Khaleghi, Z. Bakhtiari, M. Ziyadi, O. Gerstel, L. Paraschis, C. Langrock, M. M. Fejer, and A. E. Willner, “Demonstration of reconfigurable optical generation of higher-order modulation formats up to 64 QAM using optical nonlinearity,” Opt. Lett. 38(17), 3350–3353 (2013). [CrossRef]
17. M. R. Chitgarha, S. Khaleghi, M. Ziyadi, A. Almaiman, A. Mohajerin-Ariaei, O. Gerstel, L. Paraschis, C. Langrock, M. M. Fejer, J. Touch, and A. E. Willner, “Demonstration of tunable optical generation of higher-order modulation formats using nonlinearities and coherent frequency comb,” Opt. Lett. 39(16), 4915–4918 (2014). [CrossRef]
18. A. Fallahpour, F. Alishahi, K. Zou, Y. Cao, A. Almaiman, A. Kordts, M. Karpov, M. H. P. Pfeiffer, K. Manukyan, H. Zhou, P. Liao, C. Liu, M. Tur, T. J. Kippenberg, and A. E. Willner, “Demonstration of tunable optical aggregation of QPSK to 16-QAM over optically generated nyquist pulse trains using nonlinear wave mixing and a Kerr frequency comb,” J. Lightwave Technol. 38(2), 359–365 (2020). [CrossRef]
19. M. Uetai, H. Kishikawa, and N. Goto, “Modulation format conversion from OOK and QPSK to 8QAM using XPM and XGM in an SOA,” in 2017 Conference on Lasers and Electro-Optics Pacific Rim, (Optical Society of America, 2017), p. s1582.
20. G. Huang, Y. Miyoshi, A. Maruta, Y. Yoshida, and K.-I. Kitayama, “All-optical OOK to 16-QAM modulation format conversion employing nonlinear optical loop mirror,” J. Lightwave Technol. 30(9), 1342–1350 (2012). [CrossRef]
21. Q. Li, X. Yang, and J. Yang, “All-optical aggregation and de-aggregation between 8QAM and BPSK signal based on nonlinear effects in HNLF,” J. Lightwave Technol. 39(17), 5432–5438 (2021). [CrossRef]
22. H. Liu, H. Wang, and Y. Ji, “Simultaneous all-optical channel aggregation and de-aggregation for 8QAM signal in elastic optical networking,” IEEE Photonics J. 11(1), 1–8 (2019). [CrossRef]
23. B. Zhang, H. Zhang, C. Yu, X. Cheng, Y. K. Yeo, P.-K. Kam, J. Yang, H. Zhang, Y.-H. Wen, and K.-M. Feng, “An all-optical modulation format conversion for 8QAM based on FWM in HNLF,” IEEE Photonics Technol. Lett. 25(4), 327–330 (2013). [CrossRef]
24. Y.-H. Wen and K.-M. Feng, “A simple NRZ-OOK to PDM RZ-QPSK optical modulation format conversion by bidirectional XPM,” IEEE Photonics Technol. Lett. 27(9), 935–938 (2015). [CrossRef]
25. P. Liao, K. Zou, H. Zhou, A. Fallahpour, N. Hu, Y. Cao, A. Almaiman, F. Alishahi, C. Bao, M. Tur, and A. E. Willner, “Flexible spectrum sharing of two asynchronous phase-shift keying signals using power division multiplexing,” Opt. Lett. 45(5), 1176–1179 (2020). [CrossRef]
26. Z. Wei, D. W. K. Ng, J. Yuan, and H.-M. Wang, “Optimal resource allocation for power-efficient mc-noma with imperfect channel state information,” IEEE Trans. Commun. 65(9), 3944–3961 (2017). [CrossRef]
27. K. Li, Y. Chen, Y. Huang, Y. Li, Y. Han, J. Fu, and J. Yu, “16-QAM vector terahertz-wave signal generation by simple MZM-based optical frequency comb,” IEEE Photonics J. 13(1), 1–8 (2021). [CrossRef]
28. M. Seimetz, “Performance of coherent optical square-16-QAM-systems based on IQ-transmitters and homodyne receivers with digital phase estimation,” in National Fiber Optic Engineers Conference, (Optical Society of America, 2006), p. NWA4.
