Hybrid transmission of PS-GS4QAM and QPSK data in the form of a PS-16QAM mm-wave signal enabled by optical asymmetrical dual-SSB modulation

Hengxin Yan; Tangyao Xie; Tangyao Xie; Xinying Li; Xinying Li; Xiaolong Pan; Xiaolong Pan; Jiahao Bi; Liye Fang; Qiuying Lin; Xiangjun Xin; Xiangjun Xin; Xiangjun Xin; Xiangjun Xin

doi:10.1364/OE.520062

1. Introduction

With the rapid growth of 5 G technology and the emergence of high-speed and high-quality data service applications such as mobile communications, high-quality Internet protocol television (IPTV), and short videos, the user demand for wireless network capacity has grown exponentially in the past decade [1–3]. The integration of high-capacity optical fiber communications and high-frequency mm-wave and terahertz (THz) wireless communications have positioned radio-over-fiber (RoF) systems as promising contenders in upcoming 6 G mobile data communications [4–6]. High-fidelity and high-capacity wireless signal transmission have proven effective [7–9], especially in some scenarios where optical fibers are unavailable. The 5 G new radio (NR) protocol has two frequency ranges (FRs): 1) FR1 from 450 MHz to 6 GHz, or Sub-6 GHz equivalently; 2) FR2 from 24 GHz to 52 GHz for millimeter wave (mm-wave) applications [10–12]. However, the large number of communication devices means that there will be many resource units. Therefore, it is necessary to simplify the system structure at a low cost.

Recently, many researchers have been studying the independent dual-single-sideband (dual-SSB) system to simplify the transmission process [13–15]. The dual-SSB system is a valuable way to replace multiple transmitters with two sideband signals: a left sideband (LSB) and a right sideband (RSB). Usually, these signals are separated using two optical bandpass filters (OBPFs) at the receiver side, and then the LSB and the RSB are detected separately. This results in a higher system structure complexity. Zhou et al. proposed a low-complexity independent dual-SSB signal transmission method based on an optical in-phase/quadrature (I/Q) modulator and single photodiode (PD) detection [16]. After 50 km of standard single-mode fiber (SSMF) transmission, the bit error rates (BERs) of the LSB and RSB are less than the hard-decision forward-error-correction (HD-FEC) threshold of 3.8 × 10⁻³. Yu et al. proposed a single PD direct detection system for dual-SSB signals without an OBPF, reducing the receiver complexity and cost [17]. However, these systems are constrained by their capacity and speed, which results in suboptimal spectrum efficiency. Apart from this, the amplitude imbalance introduced by the nonideal I/Q modulator has not been analyzed in these systems.

In recent years, probabilistic shaping (PS) has been verified to be able to reduce the gap to the Shannon capacity limit [18,19]. Compared with uniformly distributed signals, the signals after PS have reduced average energy and source entropy, leading to better system performance. Moreover, the reduced averaged energy of PS signals is meaningful for improving the system tolerance to nonlinear fiber effects and slowing the influence of optoelectronics saturation [20]. Many different PS schemes have been proposed in recent years. However, PS technology has not been introduced into the independent optical dual-SSB signal generation and detection scheme based on an I/Q modulator and a single PD.

In this paper, we propose and experimentally demonstrate an independent optical asymmetrical dual-SSB generation and detection scheme based on an I/Q modulator and single PD detection to realize hybrid transmission of probabilistic-shaping geometric-shaping 4-ary quadrature amplitude modulation (PS-GS4QAM) and quadrature phase-shift keying (QPSK) data in the form of a PS-16QAM millimeter-wave (mm-wave) signal. Moreover, we analyze the influence of amplitude imbalance introduced by the nonideal I/Q modulator and demonstrate a feasible way to reduce this imbalance. Compared with [16,17], our novel scheme introduces the PS technology into the RoF system to generate a PS-16QAM mm-wave signal and can achieve a higher transmission baud rate. Additionally, the generated mm-wave signal has a higher mm-wave carrier frequency. The simulation results show that, compared with the 16QAM scenario, the PS-16QAM scenario exhibits a ∼1.2 dB improvement at the hard-decision forward error correction (HD-FEC) threshold of 3.8 × 10⁻³_. The experiment results show that, after transmission over 5 km SSMF and 1.2 m single-input-single-output (SISO) wireless link, the BERs of LSB-PS-GS4QAM and RSB-QPSK signals can meet the HD-FEC threshold. Therefore, our scheme is expected to simplify the complexity of RoF fronthaul networks and resolve the problem of insufficient spectrum resources.

2. Principle

Our optical asymmetrical dual-SSB modulation scheme generates a PS-16QAM mm-wave signal based on an I/Q modulator and one single PD, as shown in Fig. 1. In the transmitter digital signal processing (DSP), two pseudorandom binary sequences (PRBSs) of the same length are individually PS-GS4QAM and QPSK modulated. Then, the PS-GS4QAM and QPSK vector signals are individually up-sampled, raised cosine (RC) shaped, and up-converted to generate independent LSB and RSB vector signals. The schematic spectrum diagrams of the generated LSB and RSB signals are shown in Figs. 1(a) and 1(b), respectively. The LSB and RSB signals are denoted by

(1)$${E_l}(t) = {E_{LSB}}(t)\exp ( - j2\pi {f_{{S_1}}}t)$$

(2)$${E_r}(t) = {E_{RSB}}(t)\exp (j2\pi {f_{{S_2}}}t)$$

where E_LSB and E_RSB represent the PS-GS4QAM and QPSK vector signals after RC shaping, respectively. Here, ${f_{{S_1}}}$ and ${f_{{S_2}}}$ represent the carrier frequency of LSB and RSB signals. The sum of the LSB and the RSB signals is sent into a digital-to-analog conversion (DAC). The DAC output can be expressed as:

(3)$$\begin{aligned} &E(t) = {E_l}(t) + {E_r}(t)\\ &= {E_{LSB}}(t)\exp ( - j2\pi {f_{{S_1}}}t) + {E_{RSB}}(t)\exp (j2\pi {f_{{S_2}}}t)\\ &= {A_l}\exp ( - j{\omega _l}t + j{\varphi _l}) + {A_r}\exp (j{\omega _r}t + j{\varphi _r})\\ &= {A_l}\cos ( - {\omega _l}t + {\varphi _l}) + {A_r}\cos ({\omega _r}t + {\varphi _r})\\ &\quad + j{A_l}\sin ( - {\omega _l}t + {\varphi _l}) + j{A_r}\sin ({\omega _r}t + {\varphi _r}) \end{aligned}$$

where A_l and A_r denote the amplitudes of the RSB and LSB signals, respectively. ${\varphi _l}$ and ${\varphi _r}$ denote their phases. And ${\omega _l}$ and ${\omega _r}$ denote their angular frequencies.

Fig. 1. Schematic of optical asymmetrical dual-SSB modulation scheme based on an I/Q modulator and one single PD. (PRBS: pseudorandom binary sequence; RC: raised cosine; ECL: external cavity laser; PM: phase modulator; MZM: Mach–Zehnder modulator): (a) schematic diagram of the LSB signal; (b) schematic diagram of the RSB signal; (c) schematic diagram of the I/Q modulator output; (d) schematic diagram of the received signal after the PD detection.

Download Full Size | PDF

The continuous wave (CW) optical signal from the external cavity laser (ECL) at a frequency of f_c and an angular frequency of ${\omega _c}$ can be expressed as:

(4)$${E_{CW}}(t) = {E_{CW}}\exp (j2\pi {f_c}t) = {E_{CW}}\exp (j{\omega _c}t)$$

The CW optical signal is sent into the optical input port of the I/Q modulator. The optical I/Q modulator is composed of two optical couplers, two Mach–Zehnder modulators (MZMs), and one phase modulator. The optical coupler of the I/Q modulator divides the input optical signal of the I/Q modulator into two branches with the same power, as depicted in Fig. 1. The upper branch MZM and the lower branch MZM of the I/Q modulator are driven by the real part and the imaginary part of the DAC output, respectively. The phase modulator of the I/Q modulator ensures that the I branch and Q branch of the I/Q modulator maintain a 90° phase difference. The I/Q modulator works at the minimum transmission point (MITP) in our scheme. The I/Q modulator output is an optical dual-SSB signal, as shown in Fig. 1(c).

Notably, the I/Q modulator's nonideal property and unmatched transmission paths from the DAC to the I/Q modulator will result in the amplitude imbalance of the optical dual-SSB signal. Thus, a crosstalk between different sidebands can occur. The crosstalk can lead to interference and distortion in the optical dual-SSB signal, affecting the system's overall performance. When the amplitude ratio of the driving dual-SSB signal's real part and imaginary part is (1-k), with k being a small value, the optical dual-SSB signal generated from the I/Q modulator can be expressed as

(5)$$\begin{aligned} &{E_{I/Q}}(t) = {E_{CW}}(t)\left\{ \begin{array}{l} {J_1}(\alpha {A_r})[{(1 - k)\cos ({\omega_r}t + {\varphi_r}) + j\sin ({\omega_r}t + {\varphi_r})} ]\\ + {J_{ - 1}}(\beta {A_l})[{(1 - k)\cos ( - {\omega_l}t + {\varphi_l}) + j\sin ( - {\omega_l}t + {\varphi_l})} ]\end{array} \right\}\\ &= {E_{CW}}\exp (j{\omega _c}t)\left\{ \begin{array}{l} {J_1}(\alpha {A_r})\left[ {(1 - \frac{k}{2})\exp (j{\omega_r}t + j{\varphi_r}) - \frac{k}{2}\exp ( - j{\omega_r}t - j{\varphi_r})} \right]\\ + {J_{ - 1}}(\beta {A_l})\left[ {(1 - \frac{k}{2})\exp ( - j{\omega_l}t + j{\varphi_l}) - \frac{k}{2}\exp (j{\omega_l}t - j{\varphi_l})} \right] \end{array} \right\}\\ &= {E_{CW}}\left\{ \begin{array}{l} {J_1}(\alpha {A_r})\left[ {(1 - \frac{k}{2})\exp (j({\omega_c} + {\omega_r})t + j{\varphi_r}) - \frac{k}{2}\exp (j({\omega_c} - {\omega_r})t - j{\varphi_r})} \right]\\ + {J_{ - 1}}(\beta {A_l})\left[ {(1 - \frac{k}{2})\exp (j({\omega_c} - {\omega_l})t + j{\varphi_l}) - \frac{k}{2}\exp (j({\omega_c} + {\omega_l})t - j{\varphi_l})} \right] \end{array} \right\} \end{aligned}$$

where ${J_n}({\cdot} )$ is the Bessel function of the first kind and order n, while α and β are the modulation coefficients of the I/Q modulator at the operating frequency of ${f_{{S_1}}}$ and ${f_{{S_2}}}$, respectively. Equation (5) shows that if the amplitude of the optical dual-SSB signal is imbalanced, crosstalk will occur.

At the receiver side, the imbalanced optical dual-SSB signal is directly fed into the PD. According to the PD’s square law characteristic, the photocurrent generated from the PD can be expressed as

(6)$$\begin{aligned} &{i_{PD}}(t) = \frac{1}{2}[{(1 - \frac{k}{2})^2}J_1^2(\alpha {A_r}) + {(\frac{k}{2})^2}J_1^2(\alpha {A_r}) + {(1 - \frac{k}{2})^2}J_{ - 1}^2(\beta {A_l}) + {(\frac{k}{2})^2}J_{ - 1}^2(\beta {A_l})]\\ &+ \frac{1}{2}[J_1^2(\alpha {A_r})\cos (2{\omega _c}t + 2{\omega _r}t + 2{\varphi _r}) + J_1^2(\alpha {A_r})\cos (2{\omega _c}t - 2{\omega _r}t - 2{\varphi _r})\\ &+ J_{ - 1}^2(\beta {A_l})\cos (2{\omega _c}t - 2{\omega _l}t + 2{\varphi _l}) + J_{ - 1}^2(\beta {A_l})\cos (2{\omega _c}t + 2{\omega _l}t - 2{\varphi _l})]\\ &+ {J_1}(\alpha {A_r}){J_{ - 1}}(\beta {A_l})[{(1 - \frac{k}{2})^2}(\cos (2{\omega _c}t + {\omega _r}t - {\omega _l}t + {\varphi _r} + {\varphi _l}) + \cos ({\omega _r}t + {\omega _l}t + {\varphi _r} - {\varphi _l}))\\ &+ {(\frac{k}{2})^2}(\cos (2{\omega _c}t - {\omega _r}t + {\omega _l}t - {\varphi _r} - {\varphi _l}) + \cos ({\omega _r}t + {\omega _l}t + {\varphi _r} - {\varphi _l}))\\ &+ (1 - \frac{k}{2})(\frac{k}{2})(\cos (2{\omega _c}t - {\omega _r}t - {\omega _l}t - {\varphi _r} + {\varphi _l}) + \cos ({\omega _r}t - {\omega _l}t + {\varphi _r} + {\varphi _l}))\\ &+ (1 - \frac{k}{2})(\frac{k}{2})(\cos (2{\omega _c}t + {\omega _r}t + {\omega _l}t + {\varphi _r} - {\varphi _l}) + \cos ({\omega _r}t - {\omega _l}t + {\varphi _r} + {\varphi _l}))]\\ &+ J_1^2(\alpha {A_r})(1 - \frac{k}{2})(\frac{k}{2})[\cos (2{\omega _c}t) + \cos (2{\omega _r}t + 2{\varphi _r})]\\ &+ J_{ - 1}^2(\beta {A_l})(1 - \frac{k}{2})(\frac{k}{2}){[\cos (2{\omega _c}t) + \cos ( - 2{\omega _l}t + 2{\varphi _l})]_{}} \end{aligned}$$

In Eq. (6), the first term is toward the direct current component, so it is negligible. Additionally, terms with ${\omega _c}$ need to be omitted due to the frequency of f_c being higher than the operating frequency range of the PD. Furthermore, the desired signal is located at ${\omega _r} + {\omega _l}$, so the term at ${\omega _r} - {\omega _l}$ can be filtered out. Therefore, Eq. (6) can be approximated as

(7)$$\begin{aligned} &i(t) \approx {J_1}(\alpha {A_r}){J_{ - 1}}(\beta {A_l})[{(1 - \frac{k}{2})^2}\cos ({\omega _r}t + {\omega _l}t + {\varphi _r} - {\varphi _l})\\ &+ {(\frac{k}{2})^2}\cos ({\omega _r}t + {\omega _l}t + {\varphi _r} - {\varphi _l})]\\ &+ J_1^2(\alpha {A_r})(1 - \frac{k}{2})(\frac{k}{2})\cos (2{\omega _r}t + 2{\varphi _r})\\ &+ J_{ - 1}^2(\beta {A_l})(1 - \frac{k}{2})(\frac{k}{2})\cos ( - 2{\omega _l}t + 2{\varphi _l}) \end{aligned}$$

It is easy to find that the optical dual-SSB signal’s amplitude imbalanced will introduce interference. The first term is the desired mm-wave signal that can be extracted from the detected photocurrent. However, if ${f_{{S_1}}}$ is equal to ${f_{{S_2}}}$, the second and third terms will overlap with the desired mm-wave signal, and interference will affect the signal quality, as shown in Fig. 2(a). Therefore, we should set different frequencies for LSB and RSB signals to reduce interference. Figure 2(b) shows that, if ${f_{{S_1}}}$ and ${f_{{S_2}}}$ are different, crosstalk noise will appear at frequencies of ${f_C} + {f_{{S_1}}}$ and ${f_C} - {f_{{S_2}}}$. After PD detection, the interference will not overlap with the desired mm-wave signal and can be reduced.

Fig. 2. Schematic diagram of the nonideal I/Q modulator output and the generated mm-wave signal after PD detection in the scenarios of (a)${f_{{S_1}}} = {f_{{S_2}}} = {f_S}$, and (b)${f_{{S_1}}} \ne {f_{{S_2}}}$. ① and ③ are the LSB and RSB signals, respectively. ②and ④ are the crosstalk noise introduced by the amplitude imbalance of the optical dual-SSB signal.

Download Full Size | PDF

Additionally, the phase of the received desired mm-wave signal is the difference between the LSB and RSB signal phases, and its amplitude is related to J₁(αA_r)J_-1(βA_l). Although the modulation coefficients α and β of the RSB and LSB are different, the signal amplitude proportion of different constellation points for each sideband remains unchanged. Thus, after PD detection, the desired mm-wave signal is in PS-16QAM format at the receiving end. The constellations of the LSB-PS-GS4QAM, RSB-QPSK, and PS-16QAM are illustrated in Fig. 3. The color correspondence reflects the mapping relationship between the received PS-16QAM signal and the dual-SSB signal.

Fig. 3. Ideal constellations and mapping relationship of (a) QPSK, (b)PS-GS4QAM, and (c)PS-16QAM.

Download Full Size | PDF

Compared with normal uniform constellations, the constellations after PS have reduced average energy and source entropy, leading to better system performance in near Shannon performance. In our scheme, the outermost ring of the received PS-16QAM signal is related to the (−3,0) direction of the LSB. Similarly, the innermost ring of the received 16QAM signal is related to the (1,0) direction of the LSB. Suppose that the (−3,0) probability is reduced, reducing the probability of the outermost ring of the received signal. Similarly, if the (1,0) probability is enhanced, the probability of the innermost ring of the received signal will increase. Figures 4(a)–4(c) show the probability distributions for the QPSK, PS-GS4QAM, and PS-16QAM vector signals, respectively.

Fig. 4. The constellation of probability distributions for (a) QPSK, (b)PS-GS4QAM, and (c)PS-16QAM. The bars represent the probability for each symbol.

Download Full Size | PDF

Notably, PS-GS4QAM has the same amplitude proportion on different circles as dose PS-16QAM. Therefore, when PS-GS4QAM is subject to the Maxwell-Boltzmann (MB) distribution, the received PS-16QAM is also subject to the MB distribution. In this paper, the PS-GS4QAM dataset exhibits a distribution that follows the MB distribution, which can be expressed as follows

(8)$$\begin{array}{c}{P_X}({x_i}) = \frac{1}{{\sum\nolimits_{k = 1}^{k = 4} {{e^{ - v{{|{{x_k}} |}^2}}}} }}{e^{ - v{{|{{x_i}} |}^2}}}\\ ({x_1} = 1;{x_2} ={-} 1 + 2j;{x_3} ={-} 1 - 2j;{x_4} ={-} 3) \end{array}$$

where v is a scaling factor that can be adjusted to change the information entropy. In this paper, v is set to 0.2.

3. Simulation setup and results

To validate the advantage of our proposed scheme, we set a simulation based on our proposed scheme to generate a 4 Gbaud-PS16QAM/16QAM mm-wave signal at a carrier frequency of 35 GHz. The dual-SSB signal is composed of a 4 Gbaud LSB-PS-GS4QAM/LSB-GS4QAM signal at a carrier frequency of −15 GHz and a 4 Gbaud RSB-QPSK signal at a carrier frequency of 20 GHz. The DSP algorithm at the transmitter end is shown in Fig. 1. The DAC has a sampling rate of 64 GSa/s. A tunable ECL generates an optical carrier with a linewidth of 100 kHz and an emission frequency of 193.1 THz. After being transmitted through 10 km SSMF, a variable optical attenuator (VOA) is employed to control the received optical power. After PD detection, a traditional DSP algorithm separates and recovers the synthesized signal at the receiving end.

The simulation results are presented in Fig. 5, illustrating the relationship between the BER and the received optical power (ROP) for the 4 Gbaud LSB and 4 Gbaud RSB. The results show that, compared with the 16QAM scenario, the PS-16QAM scenario exhibits ∼1.2 dB improvement at the HD-FEC threshold. Therefore, the results indicate that due to the introduction of the PS technology and optical asymmetrical dual-SSB modulation, our proposed scheme performs better than the previous methods [16,17] and is expected to transmit higher baud rates and longer distances.

Fig. 5. BER versus power into PD.

Download Full Size | PDF

4. Experimental setup and results

The experiment setup of our novel scheme generates a 5 Gbaud PS-16QAM mm-wave signal from an optical asymmetrical dual-SSB signal based on direct PD detection, as shown in Fig. 6. The redundancy ratio introduced by the PS technology is about 0.1, and therefore, the net bit rate of the 5 Gbaud PS-16QAM is ∼19 Gbit/s (5 Gbaud × 4× (1-0.1) = 19 Gbit/s). The dual-SSB signal is composed of a 5 Gbaud RSB-QPSK at a carrier frequency of 20 GHz and 5 Gbaud LSB-PS-GS4QAM at a carrier frequency of −15 GHz. The ECL generates the CW with a linewidth of 100 kHz, an emission frequency of 193.4 THz, and an output power of 15 dBm. At the transmission end, two PRBSs are generated, with the data length of the two PRBSs being 2¹⁵. The RC filter has a roll-off factor of 0.2. The arbitrary waveform generator (AWG) has a sampling rate of 100 GSa/s. The I/Q modulator has a 3 dB bandwidth of 40 GHz. Figure 6(a) shows the optical spectrum of the I/Q modulator output. Due to the amplitude of the generated PS-GS4QAM higher than that of the generated QPSK during the transmitter DSP, the power of the left-side band is higher than that of the right-side band after OCS modulation. The optical signal of the I/Q modulator output is amplified by an Erbium-doped fiber amplifier (EDFA) and then sent into a transmission link. The transmission link comprises a 5 km SSMF and a 1.2 m SISO wireless link. After transmission through the 5 km SSMF, a VOA is employed to achieve precise control over the signal intensity by adjusting the optical power entering the PD. Then, a PD with a 3 dB bandwidth of 70 GHz directly detects the optical signal and generates a 35 GHz electrical mm-wave signal. In the wireless link, the 35 GHz electrical mm-wave signal is firstly boosted by an electrical amplifier (EA) with a gain of 24 dB and an operating frequency range of 60 kHz - 50 GHz. One pair of horn antennas transmits the electrical mm-wave signal over the 1.2 m wireless link. The antennas have a gain of 25 dBi and a bandwidth of 20∼40 GHz. After an EA boosts the received signal, the received signal is captured by a digital storage oscilloscope (OSC) with a 256 GSa/s sampling rate and 50 GHz electrical bandwidth. Figure 6(b) shows the mm-wave signal spectra captured by the OSC.

Fig. 6. Experiment setup of the proposed scheme: (a) the optical spectrum of the I/Q modulator output, (b) the electrical spectrum of the received PS-16QAM mm-wave signals by OSC. EA, electrical amplifier; SSMF, standard single-mode fiber; VOA, variable optical attenuator; EDFA, Erbium-doped fiber amplifier; OSC, oscilloscope; CMMA, cascaded multimodal algorithm; BPS, blind phase search; DDLMS, decision-directed least mean square.

Download Full Size | PDF

At the receiving end, a traditional DSP algorithm separates and recovers the synthesized signal. The DSP algorithm comprises downconversion, resampling, cascaded multimodal algorithm (CMMA), blind phase search (BPS), decision-directed least mean square (DDLMS), and demapping. Ultimately, the restored signal is evaluated by quantitatively assessing the error performance of the LSB and RSB by calculating the BER. The CMMA algorithm has 19 taps with a step factor of 0.001. The BPS is implemented with 64 test phases.

Figure 7 shows the BER performance versus the power into the PD for the 5 Gbaud LSB-PS-GS4QAM signal and the 5 Gbaud RSB-QPSK signals in three different scenarios: back-to-back (BTB), 1.2 m wireless transmission only, as well as both 5 km SSMF transmission and 1.2 m wireless transmission. The results indicate that after transmitting over 5 km of SSMF and 1.2 m SISO wireless link, the BER of each sideband signal can be below the HD-FEC threshold value of 3.8 × 10⁻³. It is evident that as the power into the PD increases, the BER of each sideband signal decreases. For the back-to-back (BTB) scenario, for a power of −7 dBm, the BER of each sideband is lower than the HD-FEC threshold. Compared with the BTB scenario, the 1.2 m wireless scenario introduces a ∼1 dB power penalty. Compared with the wireless scenario, both 5 km SMF transmission and 1.2 m wireless transmission introduce a ∼2 dB power penalty.

Fig. 7. BER versus power into the PD in the scenarios of BTB, wireless, and SMF & wireless.

Download Full Size | PDF

When the power into the PD is located in the range of −9 dBm to −5 dBm, the BER curves of the BTB and 1.2 m wireless transmission exhibit similarity. Compared with the BTB, for 1.2 m wireless transmission, the electrical signal is amplified by the EA after PD detection and then sent to the transmit antenna. Therefore, the EA can compensate for the path loss introduced by the wireless transmission. Different from the ordinary asymmetric single-sideband (ASSB) scheme, the LSB and RSB light waves of the beating frequency here need to maintain the synchronization of the encoding information to achieve the generation of the PS-16QAM signal. Compared with 1.2 m wireless transmission, for 5 km SSMF and 1.2 m wireless communication, when the BER meets the HD-FEC threshold, the power into the PD is 2 dB higher. The phenomenon is caused by the wavelength walk-off effect between LSB and RSB induced by fiber dispersion. Apart from this, the imperfect characteristics of the optoelectronic devices also influence the generated PS-16QAM signal quality by PD detection. Thus, it is necessary to adopt an effective method to compensate for the signal amplitude imbalance introduced by the I/Q modulator. Meanwhile, it is essential to introduce the PS technology in the system to reduce the distortion introduced by the SSMF and optoelectronic devices.

5. Conclusion

In this paper, we propose and experimentally demonstrate a scheme to generate a low-complexity independent asymmetrical dual-SSB signal based on the probabilistic shaping (PS) technology to decrease the influence of the optical dual-SSB signal’s amplitude imbalance and to enhance the system performance in low SNR. After the PD direct detection, the mm-wave signal is in the PS-16QAM format, generated from the LSB-PS-GS4QAM signal and the RSB-QPSK signal. Therefore, the demodulation scheme does not require dividing the dual-SSB signal into two channels by a pair of OBPFs. In contrast, using a conventional DSP algorithm can restore the information of the dual-SSB signal. The simulation results indicate that, compared with LSB-GS4QAM, LSB-PS-GS4QAM results in an improvement of ∼1.2 dB in the BER of the HD-FEC threshold. For the controlled experiment, the RSB signal in QPSK format remains unchanged. However, the RSB-QPSK also achieves an improvement of ∼1.2 dB in the BER of the HD-FEC threshold when the LSB is in PS-GS4QAM format. The experimental results demonstrate that the BERs of an independent 5 Gbaud LSB-PS-GS4QAM at −15 GHz and 5 Gbaud RSB-QPSK at 20 GHz can achieve low BER transmissions below 7% HD-FEC after 5 km of SSMF and 1.2 m of SISO wireless link. The BERs for the LSB-PS-GS4QAM and RSB-QPSK vector signals are 3 × 10⁻³ and 3.3 × 10⁻³ at power into the PD of −4dBm, respectively. Our proposed scheme can tolerate the I/Q modulator nonideal property and performs well in a low SNR. These results indicate that our scheme has the potential to offer a cost-effective and high-quality solution for mobile fronthaul in 5 G NR networks operating at Q-band mm-waves.

Funding

National Key Research and Development Program of China (2023YFB2806100); National Natural Science Fund for Excellent Young Scientists Fund Program (Overseas) (3050013532305); National Natural Science Foundation of China (62305026).

Disclosures

The authors declare no conflicts of interest.

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. F. Wang, R. Gao, Z. Li, et al., “OAM mode-division multiplexing IM/DD transmission at 4.32 Tbit/s with a low-complexity adaptive-network-based fuzzy inference system nonlinear equalizer,” Opt. Lett. 49(3), 430–433 (2024). [CrossRef]

2. K. Wang, T. Song, Y. Wang, et al., “Evolution of short-range optical wireless communications,” J. Lightwave Technol. 41(4), 1019–1040 (2023). [CrossRef]

3. E. Ruggeri, A. Tsakyridis, C. Vagionas, et al., “A 5 G Fiber wireless 4Gb/s WDM fronthaul for flexible 360° coverage in V-Band massive MIMO small cells,” J. Lightwave Technol. 39(4), 1081–1088 (2021). [CrossRef]

4. J. Ding, J. Yu, W. Li, et al., “High-speed and long-distance photonics-aided terahertz wireless communication,” J. Lightwave Technol. 41(11), 3417–3423 (2023). [CrossRef]

5. W. Li, J. Yu, J. Ding, et al., “23.1-Gb/s 135-GHz wireless transmission over 4.6-Km and effect of rain attenuation,” IEEE Trans. Microwave Theory Techn. 71(11), 5004–5018 (2023). [CrossRef]

6. F. Wang, Y. Wang, W. Li, et al., “Echo state network based nonlinear equalization for 4.6 km 135 GHz D-band wireless transmission,” J. Lightwave Technol. 41(5), 1278–1285 (2023). [CrossRef]

7. W. Zhou, X. Yang, W. Li, et al., “Demonstration of PDM-2048QAM W-band signal delivery over 4.6 km wireless transmission employing one bit DSM,” in Optical Fiber Communication Conference (OFC), Technical Digest Series, paper W2B.25 (2023).

8. C. Ranaweera, C. Lim, Y. Tao, et al., “Design and deployment of optical x-haul for 5 G, 6 G, and beyond: progress and challenges [Invited],” J. Opt. Commun. Netw. 15(9), D56–D66 (2023). [CrossRef]

9. T. Xie, X. Xin, L. Fang, et al., “SC and OFDM hybrid coherent optical transmission scheme based on 1-bit bandpass delta-sigma modulation,” Opt. Express 32(7), 11337–11345 (2024). [CrossRef]

10. H. Yang, S. Wang, D. Peng, et al., “Optically powered 5 G WDM fronthaul network with weakly-coupled multicore fiber,” Opt. Express 30(11), 19795–19804 (2022). [CrossRef]

11. L. Zhao, B. Sang, Y. Tan, et al., “Transmission of 1024-QAM OFDM at 28 GHz radio frequency using 5 G millimeter wave phased array antenna,” IEEE Trans. Microwave Theory Techn. 70(9), 4211–4217 (2022). [CrossRef]

12. J. Bi, T. Xie, D. Guo, et al., “Dual vector millimeter-wave signal generation based on optical carrier suppression modulation and direct detection with one photodetector,” Opt. Lett. 49(5), 1205–1208 (2024). [CrossRef]

13. X. Li, M. Zhu, Z. Xing, et al., “Asymmetric direct detection of twin-SSB signals,” Opt. Lett. 45(4), 844–847 (2020). [CrossRef]

14. H. Song, C. Huang, H. Li, et al., “Asymmetric dual-SSB modulation for photonic co-frequency mm-wave signals generation and DSP-free receiver,” Opt. Lett. 46(17), 4366–4369 (2021). [CrossRef]

15. Y. Zhou, J. Ming, L. Wang, et al., “Independent dual-single sideband lower-IF RF receiver based on a single photodetector,” Opt. Lett. 47(20), 5317–5320 (2022). [CrossRef]

16. Y. Zhou, J. Xiao, C. Zhao, et al., “Simplified twin-single-sideband direct detection system separating left and right sideband signals using a digital signal processing algorithm instead of optical bandpass filters,” Opt. Express 30(1), 619–628 (2022). [CrossRef]

17. C. Yu, X. Sheng, R. Gao, et al., “Experimental demonstration of twin-single-sideband signal detection system based on a single photodetector and without optical bandpass filter,” Opt. Express 30(21), 37341–37349 (2022). [CrossRef]

18. M. Kong, K. Wang, J. Ding, et al., “640-Gbps/Carrier WDM transmission over 6,400 km based on PS-16QAM at 106 Gbaud employing advanced DSP,” J. Lightwave Technol. 39(1), 55–63 (2021). [CrossRef]

19. S. Xing, G. Li, A. Sun, et al., “Demonstration of PS-QAM based flexible coherent PON in burst-mode With 300 G peak rate and ultra-wide dynamic range,” J. Lightwave Technol. 41(4), 1230–1239 (2023). [CrossRef]

20. X. Wu, J. Zhang, A. P. T. Lau, et al., “C-band 100-GBaud PS-PAM-4 transmission over 50-km SSMF enabled by FIR-filter-based pre-electronic dispersion compensation,” Opt. Express 31(11), 17759–17768 (2023). [CrossRef]

Hybrid transmission of PS-GS4QAM and QPSK data in the form of a PS-16QAM mm-wave signal enabled by optical asymmetrical dual-SSB modulation

Abstract

1. Introduction

2. Principle

3. Simulation setup and results

4. Experimental setup and results

5. Conclusion

Funding

Disclosures

Data availability

References

Data availability

Cited By

Figures (7)

Equations (8)

Optics Express