WDM-FSO System for 5G Wireless Network Using Spatial Multiplexing in the Presence of Wireless and Optical Nonlinearities

Abstract

This chapter focuses on theoretical analysis of the transmission performance of wavelength division multiplexed (WDM) free-space optical communication (FSO) links amidst optical and wireless nonlinearities. It has been shown that, to mitigate the nonlinearities due to atmospheric turbulence and optical path, spatial multiplexing technique has been utilized, coupled with advanced modulation formats such as CS-RZ, QPSK and DSPK, and QAM. An in-depth theoretical model and simulation study have been carried out. In the presence of strong and medium turbulence, our analysis indicates that modulation formats QAM, DPSK, and QPSK perform better than normal baseband CS-RZ. The analysis further shows that using spatial multiplexing (MIMO), QAM has a 5.1 dB advantage over CS-RZ and a 1.6 dB advantage over DPSK and QPSK modulations at the FEC threshold level of 3.8 × 10⁻³. The WDM-FSO system can transmit wireless signals up to a distance of 36 kilometers at the FEC threshold. Further, we have shown that a combination of the WDM-FSO system can transport wireless signals up to 36 km at the FEC threshold 3.8 × 10⁻³. Specifically, it has been demonstrated that the spectrally efficient QAM modulation format achieves higher transmission efficiency at FEC error rate 3.8 × 10⁻³.

1. Introduction

The swift development of 5G and 6G wireless systems, together with their impending deployment, necessitates increased system capacity, enhanced spectral efficiency, and faster data rates for both front- and back-haul networks [1, 2, 3, 4, 5, 6]. The combined transfer of enormously diverse data traffic from several network planes and massively networked devices into the core network is the main difficult issue facing 5G and B5G wireless systems. This issue suggests that 5G systems must employ multiplexing, a spectrally efficient modulation format (MIMO), and techniques seamless integration with the core optical backbone transmission path. In the 5G scenario, a backhaul network needs dense cellular structures to provide data access, which can be accomplished using microwave and millimeter-wave systems [7, 8]. However, they face limitation signal due to degradation, fading, and spectrum scarcity. Uses of signal transmission through optical fiber transmission has the benefit of being transparent to wireless signals, allowing for the sub-carrier multiplexing of numerous heterogeneous wireless channels using coherent systems or optical intensity modulation (IM-DD). Consequently, a dual hybrid optical transmission combination such as optical-wireless free-space optics (FSO) for the wireless signal part, and wavelength division multiplexing (WDM) for optical path can utilize, and 5G needs its complementary strengths. High bandwidth, high capacity, and secure data transmission have recently been thoroughly examined [2, 9]. The statistical fading effect scintillation brought on by atmospheric turbulence limits FSO transmission and therefore the transmission distance. The FSO system will work in conjunction with WDM technology to get around that distance restriction. WDM technology can send combined data streams at a higher transmission rate by utilizing several optical wavelengths. Furthermore, a number of spectrally efficient modulation formats that can send greater data rates over channels with limited bandwidth can be used to increase the data rate. Modulation in formats such as QPSK, DPSK, and m-QAM provides resistance against fading and greater spectral efficiency. In the wireless path, they are usable.

Moreover, different spatial multiplexing techniques with multiple transmit and receive structures have recently been investigated experimentally and theoretically [4, 5, 6, 9, 10] in order to improve the robustness and battle the fading effect in the FSO channel. It has been demonstrated that system performance can be enhanced by using MIMO FSO in an intensity-modulated direct-detection approach. The MIMO-FSO system’s BER performance as reported in [8, 11] only takes atmospheric distortions into account when analyzing FSO transmission. However, nonlinear distortions and fading effects in both wireless and optical channels must be considered when analyzing the WDM-FSO combination with MIMO spatial multiplexing.

The analysis of system performance, including bit error rate (BER), is analyzed in this chapter using a detailed analytical model that considers both atmospheric and optical nonlinearities. The robustness of these modulation formats, which include QPSK, QAM DPSK, CS-RZ, and QAM, has been evaluated.

2. Hybrid FSO-WDM system architecture

Figure 1(a) and (b) display the system architecture that utilizes FSO to backhaul dense cellular radio access in 5G environments along with WDM-based optical core transport. The WDM enhances the data rate to approximately 40 Gb/s by multiplexing FSO channels, and they share the same wavelength and system elements. Wireless and optical transports are utilized to their full potential in the above converged architecture. DWDM employs bidirectional data transmission and capacity multiplication, However, FSO offers greater deployment and reusability flexibility. Furthermore, the 5G scenario’s radio access units (RAUs) and the FSO’s wireless portion operate in separate frequency ranges and are not impacted in a comparableway by wireless routes. The deployment requirements dictate whether or not RAUs and FSO data transfer channels are compatible, and Refs. [9, 12, 13] have already demonstrated hybrid architectures for the deployment of backhaul/front-haul networks. In addition, optical modulation is transparent to RF signals, which allows for the use of cost-effective intensity-modulation direction detection (IM-DD) in optical transmission.

Figure 1.

(a) Schematic diagram of FSO-based access network and (b) configuration of WDM-M-QAM transmission over FSO link.

Figure 1(a) illustrates the combined FSO-WDM design for signal transmission. It is hampered by a number of nonlinear distortions in the wireless, FSO, and optical paths. The electrical signal propagating through an FSO channel experiences random intensity fluctuations, which cause scintillation. The gamma-gamma distribution can be used effectively to model medium to strong turbulence. Consequently, we must consider the distortions in each of the aforementioned segments as well as the interaction of the aforementioned effects in order to assess the system performance in terms of BER and outage likelihood. The analytical model for evaluating BER performance will be presented in the next sections, considering the nonlinear distortions in every transmission segment.

3. Channel model

3.1 Radio access RF channel model

Before multiplexing at the micro base stations, the individual RF signals in the radio access link are subjected to statistical fading in the wireless path. In our study, we consider that the statistics of the RF fading channel follow a Rician distribution. The Rician channel’s probability density function (PDF) is expressed in the following manner:

$$\begin{array}{lll}{F}_{\mathit{RF}}\left(\gamma \right)& =& \left(1+k\right)exp\left(-k-\left(\frac{\gamma }{{\mathit{SNR}}_{\mathit{RF}}}\left(1+k\right)\right)\right).\\ & & \times I_0\left(2\sqrt{\left(\left(\frac{\gamma }{{\mathit{SNR}}_{\mathit{RF}}}\right)\left(K+K^2\right)\right)}\right)\mathit{d\gamma }\end{array}$$E1

I would like to know where I₀ the zeroth order Bessel function of the first kind is located. A Rayleigh fading environment is created when K = 0 is used for strong fading.

3.2 Model for FSO channel

In LOS mode, the FSO link works with the pointing error being the total distance between the beam weight center and the receiver aperture center. To obtain PDF of the pointing error hp, the model described in [14, 15] can be used, as per the model in [8, 12, 16]. According to this model, attenuation resulting from geometric spread with radial displacement r from the detector origin can be roughly represented as follows on the receiver plane at distance L from the transmitter if we consider a circular detection aperture with radius a and a Gaussian spatial intensity profile of beam waist:

$$h_p\left(r,L\right)=A_{0}exp\left[-\frac{2r^2}{w_{l_{\mathit{eq}}}^2}\right]$$E2

The fraction of power collected by the detector is called hp(.), $v=\sqrt{\mathit{\pi r}}/\sqrt{2w_l}$, $w_{l_{\mathit{eq}}}^2=w_l^2\sqrt{\mathit{\text{rerf}}\left(v\right)}/2vexp\left(-v^2\right)$ is the equivalent beam width, A = [erf (v)]², erf(.) is the error function, and is the percentage of the gathered power at r = 0. In contrast, the radial displacement r has a Rayleigh distribution. After that, the hp PDF is provided by:

$$f_{h_p}\left(h_p\right)=\frac{{\gamma }^2}{A_0^{{\gamma }^2}}{h}_p^{{\gamma }^2-1}0\le h_p\le A_0$$E3

Where $\gamma =w_{\mathit{zeq}}/2{\sigma }_s$ is the ratio between the equivalent beam radius and the standard deviation of the pointing error displacement at the receiver. The nth moment of ha is defined as $E\left[h_p^n\right]={\int }_0^{A_0}{h}_p^n{f}_{h_p}\left(h_p\right){\mathit{dh}}_p$ which can be obtained as

$$E\left[h_p^n\right]=\frac{A_0^n{\gamma }^2}{n+{\gamma }^2}$$E4

3.3 Model for atmospheric turbulence

The signal’s degradation is caused by the statistical random fluctuations caused by strong and weak turbulence in the atmosphere.

The gamma-gamma model is a model that is commonly employed and can handle both weak and strong turbulence regimes. The PDF that has been provided contains the received irradiance;

$$p\left(I\right)=\frac{2{\left(\mathrm{ab}\right)}^{\frac{a+b}{2}}}{\mathrm{\Gamma }\left(\mathrm{a}\right)\mathrm{\Gamma }\left(\mathrm{b}\right)}{I}^{\frac{a+b-2}{2}}{K}_{a-b}\left(2\sqrt{\mathit{abI}}\right),\mathrm{I}>0$$E5

where x is the normalized signal intensity scintillation, Γ(.) is gamma function, and Kn(.) is the modified Bessel function.

$$\begin{array}{lll}a& =& {\left[exp\left\{\frac{0.49{\sigma }^2}{{\left(1+0.18{\mathrm{d}}^2+0.56{\sigma }^{12/5}\right)}^{7/6}}\right\}-1\right]}^{-1}\\ b& =& {\left[exp\left\{\frac{0.51{\sigma }^2{\left(1+0.69{\sigma }^{12/5}\right)}^{-5/6}}{1+0.9d^2+0.62d^2{\sigma }^{12/5}}\right\}-1\right]}^{-1}\end{array}$$E6

${\sigma }^2=0.5C_n^2{k}^{7/6}{L}^{11/6}$ and $d=\sqrt{\left(\pi D^2/2\mathit{\lambda L}\right)}$, D(m) is the diameter of the receiver collecting lens aperture,$\lambda \left(m\right)$ is the wavelength, and L(m) is the link distance, and the parameter responsible for the refractive index structure is $C_n^2\left(m^{-2/3}\right)$. The scintillation index S.I can be expressed in terms of a and b as follows.

$$S.I=\frac{1}{a}+\frac{1}{b}+\frac{1}{\mathit{ab}}$$E7

The PDF can be used to express the combined effects of atmospheric turbulent and the pointer error.

$$f_h\left(h\right)=\int f_{h_p}\left(h_p\right).p\left(I\right)\mathit{dI}$$E8

Following the approach in [14, 17, 18], the combined conditional PDF is represented as;

$$f_h\left(h\right)=\frac{\mathit{ab\varphi }}{h_{l2}{A}_0\mathrm{\Gamma }\left(a\right)\mathrm{\Gamma }\left(b\right)}{G}_{1.3}^{3.0}\left[\frac{\mathit{ab}}{A_0{h}_{l2}}h\left|{}_{{\varphi }^2-1,a-1,b-1}^{{\varphi }^2}\right|\right]$$E9

where G (.) can I find the Meijer G function.

3.4 Combined model of radio access RF and FSO

Our belief is that there is no statistical correlation between the fading in the radio access path and the atmospheric turbulence in the FSO link. Conditional PDFs can be used to represent the channel state to assess system performance,

$$P\left(\psi \right)=F_{\mathit{RF}}\left(\gamma \right)\cdot f_h\left(h\right)$$E10

3.5 Distortion due to optical nonlinearities

The signal that has faded due to both RF fading and FSO atmospheric turbulence is sent through optical fiber. Different optical carriers modulate the intensity of the faded signals, which are then multiplexed to form a WDM signal. The SSMF is used to transmit each WDM signal through intensity-modulated direct detection (IM-DD). Evaluation of the system performance of the WDM-FSO involves evaluating the effects of different optical nonlinearities encountered by the signal through the SMF. After being converted to the optical carrier frequency (FC), the WDM signal for N channels can be written as:

$$S\left(t\right)=\sum _{i=0}^{N-1}{s}_i\left(t\right)=\sum _{i=0}^{N-1}{u}_i\left(t\right)\cdot exp\left[j{\omega }_i\left(t\right)+{\varphi }_i\left(t\right)\right]$$E11

The optical power output from the laser diode is proportional to the modulating signals as

$$P\left(t\right)=P_t\left(1+\sum _{i=0}^{N-1}{m}_{i}s\left(t\right)+a_3{\left[\sum _{i=0}^{N-1}{m}_{i}s\left(t\right)\right]}^3\right)$$E12

The average transmitted optical power, a₃, and mi are the average OMI per channel, which is Pt. The equations govern the small signal intensity modulation (IM) and phase modulation (PM) as they propagate through the fiber, taking into account both chromatic dispersion and SPM effect.

$$\frac{\partial {\tilde{p}}_N}{\partial z}={\beta }_2{\omega }^2{P}_0\tilde{\varphi }$$E13

$$\frac{\partial \tilde{\varphi }}{\partial z}=-\left[\frac{{\beta }_2{\omega }^2}{4P_0}+\gamma exp\left(-\mathit{\alpha z}\right)\right]{\tilde{p}}_N$$E14

where ${\tilde{p}}_N={\tilde{p}}_{N\left(z,\omega \right)}$, $\tilde{\varphi }=\tilde{\varphi }\left(z,\omega \right)$ is the frequency of angular modulation, t is the time in a frame moving at the fiber input, and z is the distance. $\alpha$ is the coefficient of power attenuation, $\gamma$ is the nonlinearity coefficient owing to, and ${\beta }_2$ is the dispersion coefficient to nonlinear Kerr effect. where ${\tilde{p}}_N={\tilde{p}}_{N\left(z,\omega \right)}$ and $\tilde{\varphi }=\tilde{\varphi }\left(z,\omega \right)$ are the Fourier transforms of the normalized IM and PM terms $p_N\left(z,t\right)$ and $\varphi \left(z,t\right)$, respectively $\omega =2\mathit{\pi f}$. The transmission of signals may experience IM-IM and PM-IM conversions due to small variations in the modulated optical signal caused by fiber dispersion and non-linear effects.

Fiber transfer function that incorporates PM-IM and IM-IM effects can be characterized by [17, 19, 20],

$$H_F\left(z,\omega \right)=C_{\mathit{IM}-\mathit{IM}}\left(z,\omega \right)+\frac{H_{\mathit{PM}}\left(\omega \right)}{2}{C}_{\mathit{PM}-\mathit{IM}}\left(z,\omega \right)$$E15

where

$$C_{\mathit{PM}-\mathit{IM}}=-\frac{2\mathit{\alpha \beta }}{\mathit{\nu \gamma }{P}_0}sin\left(\frac{\mathit{\alpha \nu z}}{2}\right)=2sin\left(\frac{{\beta }_2{\omega }^{2}z}{2}\right)$$E16

and

$$C_{\mathit{IM}-\mathit{IM}}=cos\left(\frac{\mathit{\alpha \nu z}}{2}\right)=cos\left(\frac{{\beta }_2{\omega }^{2}z}{2}\right)$$E17

$$H_{\mathit{PM}}\left(\omega \right)=2P_0{\tilde{\varphi }}_{\mathit{\text{in}}}\left(\omega \right)/{\tilde{p}}_{\mathit{\text{in}}}\left(\omega \right)$$E18

For directly modulated laser,

$$H_{\mathit{PM}}\left(\omega \right)={\alpha }_c\left(1-{\mathit{jf}}_c/f\right)$$E19

4. Derivation of BER expression

In this section, we present analytical BER expressions that take into account both optical nonlinearities and the effects of RF fading and atmospheric turbulence on the FSO path. The BER will be presented without any diversity first. The analytical expressions for BER with spatial multiplexing will be introduced in the future.

4.1 BER without spatial multiplexing in the single input-single output (SISO) scenario

The signal to noise ratio (SNR) and distortion ratio (Distortion) ratio in the optical channel should be evaluated before evaluating the BER. After transmission through the optical path, the faded signal from the RF-FSO path becomes worse because of optical fiber dispersion and the nonlinear Kerr effect. Clipping noise is occurring in the signal due to the rapid amplitude fluctuations of the multiplexed RF signal that exceed the laser liner threshold. The equation above shows that the required signal experiences multiple distortions caused by fiber dispersion and phase modulation, such as SPM. The signal that was received after the fiber transmission is a description.

$$y\left(t\right)=h_{\mathit{\text{fiber}}}\left(t\right)\otimes s\left(t\right)+n_d\left(t\right)+n_{\mathit{\text{AWGN}}}\left(t\right)+n_I\left(t\right)$$E20

On each side, apply a Fourier transform to

$$Y\left(\omega \right)=H_{\mathit{\text{fiber}}}\left(\omega \right)\cdot S\left(\omega \right)+N_d\left(\omega \right)+N_G\left(\omega \right)+N_I\left(\omega \right)$$E21

After a few mathematical steps, using Eqs. (14) and (15), the received signal-to-noise ratio can be evaluated as

$$\mathit{\text{SNDR}}=\overline{\gamma }=\frac{\alpha P_{\mathit{RX}}}{{\sigma }_d^2+{\sigma }_{\mathit{NLI}}^2+{\sigma }_G^2+{\sigma }_I^2}$$E22

$$\text{where}{\sigma }_d^2=m_i^2{\left[f_m\left(D\left(\omega \right){\mathit{\eta \lambda }}^2/c\right)\right]}^2$$E23

$$\text{and}{\sigma }_{\mathit{NLI}}^2={\left[\frac{1}{2}m\cdot \frac{{\lambda }^2}{2\mathit{\pi c}}\cdot D\cdot k\cdot n_2\cdot \frac{P_0}{A_{\mathit{eff}}}\cdot {\left(\omega \right)}^2\cdot z^2\left(L\right)\right]}^2\cdot N_{\mathit{CSO}}$$E24

N_CSO is the number of second-order signal-interference beating terms. The clipping noise can model as Poisson process and can be evaluated using [16]

$${\sigma }_{\mathit{\text{clip}}}^2={\sigma }_I^2=\frac{4\overline{\tau }}{3}{\pi }^{-3/2}{m}_{\mathit{\text{total}}}^{5}exp\left(-1/{\mathrm{m}}_{\mathit{\text{total}}}^2\right){\left(\rho {\mathrm{P}}_{\mathrm{r}}\right)}^2$$E25

Gaussian noise arises from the ASE noise generated by the optical amplifier in the transmission path. By using, it is possible to evaluate the BER of an optical signal transmitted in the presence of Gaussian noise and optical nonlinear distortion for different modulations.

$$p_{\mathit{CS}-\mathit{RZ}}=Q\left(\sqrt{I^2\raisebox{1ex}{$\overline{\gamma }$}\!\left/ \!\raisebox{-1ex}{$2$}\right.}\right);\mathrm{CS}‐\mathrm{RZ}\text{case}$$E26

$$p_{\mathit{\text{DPSK}}}=\frac{1}{2}exp\left(-I^2\overline{\gamma }\right);\text{DPSK\hspace{0.17em}case}$$E27

$$\begin{array}{ll}& P_{s,16-\mathit{QAM},\mathit{Av}}=\frac{\left(1-M^{-1/2}\right)}{N{log}_2\left(M\right)}\\ & \times \sum _{i=0}^{N-1}\underset{0}{\overset{\infty }{\int }}\left[2\mathit{erfc}\left(\sqrt{\frac{\overline{\gamma }\left(I\right)}{2\left(M-1\right)}}\right)-\left(1-\sqrt{M^{-1}}\right){\mathit{erfc}}^2\left(\sqrt{\frac{\overline{\gamma }\left(I\right)}{2\left(M-1\right)}}\right)\right]f_1\left(I\right)\mathit{dI};\mathrm{QAM}‐\text{case}\end{array}$$E28

To evaluate the bit error rate (BER) in the free-space optical communication (FSO), the BERs mentioned in (26)–(28) must now be averaged across the can be PDF of the gamma-gamma log-normal channel and expressed as

$$p_e=\underset{0}{\overset{\infty }{\int }}p\left(I^2\overline{\gamma }\right)f\left(I\right)\mathit{dI}$$E29

where $p\left(I^2\overline{\gamma }\right)$ is BER of the optical signal transmitted in the additive white Gaussian noise (AWGN) channel and $\overline{\gamma }$ the average ratio of electrical signal-to-noise are represented respectively.

4.2 BER with spatial multiplexing (multi input-multiple output-MIMO case)

We are presently concentrating on FSO links that have a variety of spatial features, such as M transmit and N receive apertures. The M N routes are where the laser output signals are transmitted. The signal received by the nth receiver is given by.

$$r_n=\eta u_n\sum _{m=1}^M{I}_{\mathit{mn}}+g_n;n=1,2,\dots \dots \dots \dots N;u_n\in 0,1$$E30

Let us assume that the EGC technique is used to combine optical signals with equal gain factors on each branch when receiving signals are combined. The receiver’s output can be described as

$$r=\sum _{n=1}^N{r}_n=\eta u_n\sum _{n=1}^N\sum _{m=1}^M{I}_{\mathit{mn}}+g_n$$E31

The instantaneous electrical signal-to-noise ratio (SNR) is a term that can be defined for an EGC receiver as follows:

$$r_{\mathit{EGC}}={\left(\sum _{m=1}^M\sum _{n=1}^N\sqrt{{\gamma }_{\mathit{mn}}}\right)}^2$$E32

In the event that both transmit and receive diversity are applied, the most effective decision metric for on-off keying is given by:

$$p\left(r|1,I_{\mathit{mn}}\right)=\frac{exp\left(-\frac{1}{2{\sigma }_g^2}{\sum }_{n=1}^N{r}_n^2\right)}{{\left(2{\mathit{\pi \sigma }}_g^2\right)}^{\frac{N}{2}}}$$E33

$$p\left(r|0,I_{\mathit{mn}}\right)=\frac{exp\left(-\frac{1}{2{\sigma }_g^2}{\sum }_{n=1}^N{\left(r_n-\eta {\sum }_{m=1}^M{I}_{\mathit{mn}}\right)}^2\right)}{{\left(2{\mathit{\pi \sigma }}_g^2\right)}^{\frac{N}{2}}}$$E34

Assuming 0, 1 equiprobable, $p_e\left(o|I_{\mathit{mn}}\right)=p_e\left(1|I_{\mathit{mn}}\right)$ and averaging over the fading coefficients, we obtain p_e as,

$${\left[P_e\right]}_{\mathit{FSO}-\mathit{\text{Optical}}}=\underset{I}{\int }{f}_I\left(I\right)Q\left(\frac{\eta I_0}{2\mathit{MN}{\sigma }_g}\sqrt{{\sum }_{n=1}^N{\left({\sum }_{m=1}^M{e}^{2x_{\mathit{mn}}}\right)}^2}\right)\mathit{dI}$$E35

Eqs. (26)–(28) are now available for numerical evaluation of the BER performance of different modulation formats with SDRT in the gamma-gamma distributed turbulent channel.

4.3 Spatial multiplexing (multiple input-multiple output-MIMO case) and RF fading can be combined to form BER

The signals on the RF access path experience fading, which we have already mentioned and modeled using Rician fading as described in Eq. (10), and the overall BER can be evaluated by averaging [Pe]FSO-Optical over Rician fading and expressed as,

$$P_e=\underset{0}{\overset{\infty }{\int }}{\left[P_e\right]}_{\mathit{FSO}-\mathit{\text{Optical}}}{P}_{\psi }\mathit{d\psi }$$E36

5. Results and discussions

BER expressions, described in Section 4, were used to perform numerical analysis and simulation of the Hybrid WDM-FSO system with and without spatial diversity. The OPTIWAVE platform and Monte-Carlo method were utilized to carry out the system simulation, with the Monte-Carlo method being used. The system parameters used for simulations are as given in Table 1. The received spectrum of a four-channel WDM optical signal paired with a faded radio frequency signal in the radio access and the FSO path over a transmission distance of L = 36 km is shown in Figure 2(a) and (b). In Figure 2(c),(d),(e), and (f), the eye diagrams for 16-QAM, DPSK, CS-RZ, and QPSK are displayed in their respective eye diagrams. The wireless signal can be transported by the combination of WDM and FSO, as demonstrated.

Parameters	Description	Refs.	Proposed work
(Rb)	Bit rate	[15, 16, 18, 21]	1.25-10 Gb/s
(λ)	Optical wavelength	[12, 16]	1550 nm
α	Fiber attenuation	[15, 18, 21]	0.2 dB per km
	Length of FSO link	[15, 18, 21, 22]	0.6 km
	Atmospheric attenuation	[12, 16]	0.2 dB per km
L	Distance of fiber path	[12, 16]	20–60 km
Sth	Density of thermal noise	[15, 18, 21]	16 pA/sqrt(Hz)
ρ	Photodiode responsivity,	[15, 18, 21]	0.8 A/W
D	SSMF parameters	[15, 18, 21, 23, 24]	17 ps/(km nm)
S	SSMF parameters	[15, 18, 21]	0.0578 ps/nm^2 km
α	SSMF parameters	[15, 18, 21]	0.2 (dB/km)
γ	Nonlinear coefficient	[15, 16, 18, 21]	1.31 (1/(W km),
(OMI)	Optical modulation index	[15, 16, 18, 21, 24]	0.7
K	Rician factor for wireless channel	[15, 16, 18, 21, 23, 25]	4

Table 1.

System parameters used for the proposed work.

Figure 2.

(a) 16-QAM Rx optical spectrum, (b) DPSK Rx optical spectrum, (c) 16-QAM eye diagram, (d) DPSK eye diagram, (e) CS-RZ eye diagram, and (f) QPSK eye diagram.

Figure 3 displays the BER performance of three modulation formats: (i) QAM with EGC receiver diversity, (ii) QPSK, and (iii) CS-RZ. The performance of the above modulation formats can be improved by the use of diversity technique, as shown in Figure 3. Using diversity at the FEC threshold of 3.8 × 10⁻³, CS-RZ achieves a performance gain of 2.1 dB, as shown in Figure 3(a). For (i) QAM and (ii) QPSK, the diversity gain is 5.1 dB and 2.3 dB, respectively, at the same BER level. This is because QAM functions better under medium-turbulence atmospheric scintillation and has a higher spectral efficiency than the traditional CS-RZ-based QQK modulation. Figure 4 shows the performance of QAM signals in the presence of optical nonlinearities under different turbulence conditions. It demonstrates that BER of 10⁻² may be reached even in very turbulent environments, guaranteeing little QOS. However, the system’s transmission performance is greatly impacted by the effect of optical nonlinearities. Figure 5 shows the combined impacts of air turbulence, optical nonlinearities, and RF fading. Figure 5(a) and (c) show what has been noticed at due to the combined effects of optical nonlinearities, RF, and atmospheric fading, the QAM and QPSK signals degrade by 1.7 and 1.9 dB, respectively, to an FEC threshold of 3.8 × 10⁻³. This demonstrates that the QAM signal is more resilient to the combined effects of fading and optical nonlinearities. Figure 6 shows the overall system performance of the WDM-FSO system coupled with RF access. In this case, a four-channel WDM system has been studied. Figure 6 shows the BER performance for the poorest channels, channels 1 and 4, in comparison to OSNR. The accompanying image illustrates how the employment of spatial diversity strategy in conjunction with WDM can achieve BER 3.8 × 10⁻³ at a transmission distance of 36 km.

Figure 3.

SNR vs. BER for mild turbulence with receive diversity for (a) CS-RZ, (b) 16-QAM and (c) QPSK.

Figure 4.

QAM performance under different turbulence condition.

Figure 5.

(a) 16-QAM received optical power vs. BER under ONL and turbulence conditions DPSK (b) and QPSK (c).

Figure 6.

OSNR versus BER can be represented by different WDM channels through WDM-FSO.

The performance evaluation of various modulation schemes including the line-coded CS-RZ system in terms of system penalty is summarized in Tables 2 and 3. It has been observed that, the line-coded CS-RZ scheme has suffered a system penalty of 4 dB in the presence of turbulence of the FSO link at FEC threshold at 1.0 × 10⁻⁴. The QPSK system has suffered less system penalty of 0.5 dB due to robustness of modulation in the presence of link turbulence. However, it has been observed that 16-QAM has suffered slightly higher system penalty of 1.2 dB. This comes at the cost of higher spectral efficiency and higher symbol rate. The DPSK has been observed to have system penalty of 2 dB. The four-channel WDM-FSO as proposed in our study suffers a system penalty of 0.4 dB as demonstrated in Figure 6.

	1	2	3	4
Technique used	m-QAM	QPSK	CE-RZ	DPSK
Data rate	10 Gb/s	10 Gb/s	10 Gb/s	10 Gb/s
Net capacity	40 Gb/s	20 Gb/s	10 Gb/s	20 Gb/s
Max. FSO link range reported	600 m	800 m	1000 m	500 m
Efficiency of Quantum	0.8	0.8	0.8	0.8
Channels no.	4	4	1	4
Target BER	10^−4	10^−4	10^−4	10^−4

Table 2.

Performance comparison of hybrid WDM-FSO optical fiber link parameters.

S. no	Modulation schemes	System penalty SNR (dB)
1	CS-RZ	4.1
2	DPSK	3.0
3	QPSK	0.5
4	16-QAM	1.2

Table 3.

Performance comparison of hybrid WDM-FSO optical fiber link in terms of system with turbulence.

6. Conclusions

In this study, we have investigated and analyzed the performance of a hybrid WDM-FSO system through detailed theoretical modeling and simulation, considering the interplaying effects of RF fading, atmospheric turbulence, and optical nonlinearities. Despite the possibility of optical nonlinearities causing additional degradation effects, the optimal combination of EGC spatial multiplexing and WDM can lead to desirable system performance at FEC BER 3.8 × 10⁻³. The system penalties for OSNR degradation in various line coding and modulation formats have been investigated. We found that 16 QAM offers higher spectral efficiency and bit rate, but marginal degradation under turbulent FSO conditions. This enables the WDM-FSO combination to transport wireless signals up to 36 km. To obtain higher bit rate wireless access, the analytical model mentioned above can be further extended with various other diversity techniques and DWDM.

Acknowledgments

The authors wish to thank the Ministry of Electronics and Information Technology (MeitY), Government of India, New Delhi for providing infrastructure support for conducting the above study.