Evaluation of the Digital Data Performance Transmission on a Fiber-Radio System

Abstract

This work shows a set of numerical simulations to evaluate the digital data performance transmission for a fiber-radio system. For this goal, VPI Photonics software is used to simulate the transmission of a digital signal at a bit rate of 2.4 GHz through an optical link of 25 km of single-mode standard fiber (SM-SF). Whereas MATLAB software is utilized to emulate a wireless channel considering four main phenomena inherent to this channel, such as multipath and slow fading, co-channel interference, and additive white Gaussian noise (AWGN). The modeling carried out in MATLAB considers a user in movement within its coverage area cell. The constellation diagram, bit error rate (BER), and signal-to-interference-plus-noise ratio (SINR) statistical tests are used to evaluate the performance of this approach. This proposal finds potential applications in the field of cellular telephony, the Internet of Things (IoT), or any other communication system to evaluate the quality of the delivered signal to a user in motion within a particular coverage cell.

1. Introduction

Advances in communication technologies have revolutionized the distribution and transmission of several services such as radio, HDTV, telephony, and the Internet, in an ever faster and more efficient way. Furthermore, a combination of technologies such as optical and wireless allows users to exploit the inherent advantages of each one of these technologies. In particular, fiber-radio technology uses optical fiber to transport signals to be radiated to multiple users using antennas [1]. On the one hand, the large bandwidth, immunity to electromagnetic interference (EMI), and low attenuation of the optical fiber are used in an advantageous manner [2]. On the other hand, the wireless transmission provides absolute mobility to users without the need for a connection terminal [3]. Figure 1 shows a basic fiber-radio scheme where the data are transmitted from a central site (CS) through an optical link to the remote radio unit (RRU). The wireless medium (atmosphere) is mainly affected by fading channels, AWGN [4], and co-channel interference [5], among other impairments.

Several applications of fiber-radio schemes are reported in the literature. For instance, in [6], an all-optical link with double spectral-efficient transmission and compensation of chromatic dispersion-induced power fading is proposed and demonstrated experimentally. In [7], a system for the transmission of energy through an optical fiber is proposed. In [8], the functionality of a fiber-radio scheme with different simultaneous transmission frequencies using a digital multiplexer for the selection of frequencies and channels is demonstrated. In [9], a fiber-radio scheme is characterized where a microwave signal provided by a receiving antenna is supplied to an optical modulator. In [10], a hybrid scheme with a fiber-to-the-home (FTTH) system combined with a mobile wireless application system using a double optical modulator and two different signals (in RF and baseband) is proposed. Finally, in [11], a fiber-radio scheme is used for the transmission of an analog TV signal through an optical link using a radiating cable for its wireless emission.

From the paragraphs above, the goal of this work is to evaluate the digital data performance transmission for a fiber-radio system. This evaluation is supported by a set of numerical simulations. For this task, the VPI Photonics software [12] takes the following into account: 1. a transmission of a digital signal at a bit rate of 2.4 GHz through an optical link of 25 km of SM-SF; 2. an optical external modulation scheme, and; 3. a dispersive channel. Whereas MATLAB software is utilized to emulate a wireless environment taking into account: 1. an urban scenario; 2. a user in movement displacing within a coverage cell of 500 m radius that corresponds to the maximum distance from the wireless receiver to the base station (BS), and; 3. a QPSK strategy format as well as four effects such as multipath and slow fading, co-channel interference, and AWGN. The performance evaluation of this proposal is evaluated taking account of the parameters of SINR and BER statistical tests. In summary, the main novelty of this work lies in the evaluation of fiber-radio systems at simulation level, finding potential applications in cellular telephony, the IoT, or any other mobile system. The rest of this document is organized as follows: Section 2 describes in detail the main concepts for the optical and wireless transmission arrangements; Section 3 shows the results of the numerical simulations carried out for the optical and wireless schemes; and finally, Section 4 summarizes the main conclusions of this work.

2. Optical and Wireless Transmission Schemes

Figure 2 depicts the block diagram of the fiber-radio scheme proposed in this work. The blue dotted box corresponds to the electro-optic arrangement simulated by the use of the VPI Photonics software. The simulation conditions are a sequence of bits at a BR of 2.4 GHz (sequence used commercially according to the optical communication standards), a distributed feedback (DFB) laser as an optical source emitting at 1550 nm (wavelength used in standard optical communication systems) at an optical power of 2 dBm, a Mach-Zehnder intensity modulator (MZ-IM), 25 km (length commonly used in an urban optical transmission link) of SM-SF whose chromatic dispersion parameter (D) at 1550 nm is 16.75 ps/nm⸱km. A photodetector (PD) recovers the modulated optical signal delivering an electrical signal (transmitted bits). From the library of this software, the Laser_vtms, ModulatorMZ_vtms, FiberNLS_vtm, and Photodiode_vtms modules are selected to simulate the DFB (linewidth of 10 MHz), the MZ-IM, the SM-SF (attenuation of 0.2 dB/km, chromatic dispersion parameter of 16.75 ps/nm⋅km), and the PD acting as an APD (avalanche photodiode with a responsivity of 0.9 A/W), respectively. The modules PRBS_vtms_1 and Coder_NRZ_vtms1 are used to generate the electrical sequence to be sent. Whereas the SignalAnalyzer modules are used as optical and electrical analyzers to monitor the corresponding signals step by step. Figure 3 shows the diagram corresponding to the electro-optic arrangement simulated in the VPI photonics software. The interface of this software allows the visualization of the results in a graphical manner and export to a file with a .txt extension.

The green dotted box in Figure 2 corresponds to the block implemented in MATLAB. This block is fed with the data previously obtained from the electro-optical system. Initially, a normalization and debugging process for these data is carried out and subsequently modulated in QPSK format with a 5 GHz carrier signal to satisfy the Nyquist theorem. This format is chosen due to its robustness and because it is currently used in 4.5G and 5G technologies. The resulting modulated signal emulates the signal radiated from the antenna to the environment. Knowing that any propagated electromagnetic signal is subjected to several inherent phenomena typical of this environment, this work considers four main phenomena such as multipath and slow fading, co-channel interference [5], and AWGN [4]. The black dotted box emulates the environment through which the radiated signal travels and where the effects previously mentioned are included. The maximum distance considered between the antenna and the receiver is 500 m. Finally, the resulting electrical signal is demodulated to recover the original data digital signal and displayed graphically.

In wireless networks, the coverage area for a mobile user consists of a hexagonal distribution because it provides an equidistant array of antennas, is symmetrical, and supplies full coverage. In this work, a particular case is analyzed involving a wireless receiver U1 situated within one of the three sectors in the central base station (BS0) cell as is depicted in Figure 4.

An discrete-event Monte Carlo simulation is carried out to evaluate the performance of the U1 user inside the coverage area. The central cell is surrounded by the nearest co-channel interfering bases (BS1 to BS6) [5]. Mathematically, the modulated signal is treated as follows [13,14]

$$Prx=\frac{yh\ell }{d^{\mu }}=\frac{y h({10}^{\raisebox{1ex}{$\zeta $}\!\left/ \!\raisebox{-1ex}{$10$}\right.}) }{d^{\mu }}$$

(1)

where

Prx is the received power for U1.

y is the total signal emitted by BS0.

d is the distance between BS0 and U1 (the user position is uniformly distributed inside the coverage area).

µ is the propagation loss exponent with values between 2.5 and 6, or even higher, depending on the type of environment. As an urban environment is being considered, then µ = 4 [15].

h is a Rayleigh random variable whose models caused fading by multiple signal paths (fast fading) with standard deviation Δ [4].

$\ell$ is a log-normal random variable (${10}^{\raisebox{1ex}{$\zeta $}\!\left/ \!\raisebox{-1ex}{$10$}\right.}$) that represents shadowing losses (slow fading). It is defined as a function of the Gaussian random variable ζ, with mean 0 and standard deviation σ [13,14].

The performance of the forward link of the RF system is evaluated in terms of the SINR [13,14]

$$SINR=\frac{Prx}{z_i+CCI}$$

(2)

where $z_i$ is defined as the AWGN signal [4] with a standard deviation σ, and $CCI$ is the additive co-channel interference signal [5]. Thus, $CCI$ is modeled as

$$CCI={{\displaystyle \sum }}_{j=1}^{6}Pr{x}_j={{\displaystyle \sum }}_{j=1}^6\frac{y_j h_j{\ell }_j}{d_j^{\mu }}={{\displaystyle \sum }}_{j=1}^6\frac{y_j h_j({10}^{\raisebox{1ex}{${\zeta }_j$}\!\left/ \!\raisebox{-1ex}{$10$}\right.}) }{d_j^{\mu }}$$

(3)

where j = 1, …, 6 is the number of each neighbor base station.

Thus, the total signal x received by U1 is given by [4,5]

$$x=Prx+\left(z_i\right)+CCI$$

(4)

The SINR, expressed in decibels, is given by [13,14]

$$SINR \left[dB\right]=10log\left(SINR\right)$$

(5)

The parameters considered for the performance evaluation in terms of the SINR using the Monte Carlo simulation are listed in

Table 1. Simulation parameters considered for an urban zone [13,14,15].

Simulation Parameters
µ factor	4 (adimensional value)
Slow fading ζ mean	0 dB
Slow fading σ	8 dB
Fast fading mean	0 dB
Fast fading Δ	8 dB

.

3. Simulation Results

This section is subdivided into two sections; initially, the results corresponding to the optical transmission using the VPI Photonics software, and subsequently, the wireless transmission results obtained using MATLAB.

3.1. Optical Transmission Results

These results correspond to the electro-optic arrangement of Figure 3. Figure 5a shows an example of a proposed sequence to validate our proposal (010110111001010101). This digital waveform is generated and visualized on the virtual generator and oscilloscope of the VPI Photonics software interface, respectively. To avoid bit losses, a space of three “0” values and a start value of “1” bit are added at the beginning of the sequence to maintain its integrity. This signal is optically modulated by the MZ-IM and transmitted through the optical fiber. Finally, the PD delivers the corresponding electrical signal shown in Figure 5b where the presence of some spurious peaks attributed to the high frequency at which the system operates is observable. The procedure to cancel these peaks is discussed in Section 3.2.

Table 2. Width of pulses sent and recovered.

No. of Pulse	Sent (Seconds)	Recovered (Seconds)	Widening (Seconds)
1	0.320 × 10−9	0.390 × 10−9	0.07 × 10−9
2	0.310 × 10−9	0.380 × 10−9	0.07 × 10−9
3	0.730 × 10−9	0.800 × 10−9	0.07 × 10−9
4	1.14 × 10−9	1.190 × 10−9	0.05 × 10−9
5	0.330 × 10−9	0.390 × 10−9	0.06 × 10−9
6	0.320 × 10−9	0.400 × 10−9	0.08 × 10−9
7	0.330 × 10−9	0.390 × 10−9	0.06 × 10−9
8	0.790 × 10−9	0.780 × 10−9	0.01 × 10−9
9	0.400 × 10−9	0.400 × 10−9	0.00 × 10−9
10	1.13 × 10−9	0.800 × 10−9	0.33 × 10−9

tabulates the width values of the pulses sent and recovered to quantify the widening that occurs as a consequence of the influence of the chromatic dispersion exhibited by the SM-SF. As was previously mentioned, these data are exported in a file with a .txt extension.

3.2. Wireless Transmission Results

The saved data of the optical transmission stage are used to validate the wireless transmission. This data undergoes a debugging and normalization process to remove spurious peaks; in this process, the first three “0” bits and the “1” bit are removed. To eliminate the spurious peaks, the signal is subjected to a zero-crossing function, in which the mean value of the received signal is taken as the threshold normalizing the signal. Before modulation and transmission, a BR correction is also carried out, and this result is shown in Figure 6.

This sequence of bits is modulated in QPSK format with a 5 GHz carrier signal. The modulated signal is shown in Figure 7a, whereas Figure 7b corresponds to the recovered transmitted signal affected by fast and slow fading, AWGN, propagation losses, and co-channel interference.

Figure 8a,b show the constellation diagrams corresponding to the transmitted and received signal, respectively.

Figure 9 corresponds to the demodulated QPSK signal. Note the similitude of this graph with the results shown in Figure 5a.

Because noise, co-channel interference, and fading impairments are factors that cause loss of information, their levels must be considered to evaluate adequate wireless reception. For this reason, statistical tests are carried out to verify the performance of the wireless stage. The distance from U1 to BS0 varies within a range from 1 to 500 m within the corresponding sector and its position determines the affectation of the co-channel interference. The noise, interference, and fading levels are modeled as a function of the standard deviation values of the random variables that determine them. This way of modeling allows different noise levels to be proposed, which is the factor with the greatest possible variation [13,14]. Considering that co-channel interference and attenuation are modeled by random variables, and assuming that the U1 user is moving within its corresponding sector, the result of the simulation is averaged in terms of SINR and BER for different noise levels. To establish the appropriate confidence interval of the wireless simulation to obtain reliable statistical data, 1000, 10,000, 30,000, 60,000, and 100,000 repetitions are performed. Therefore, the observation was made that within the interval of 60,000 to 100,000 times, the difference between the results is negligible. Thus, 60,000 repetitions is established as the confidence interval.

Figure 10 illustrates graphically how the noise level influences the BER percentage within the procedure carried out. One can observe that from an AWGN standard deviation of $1\times {10}^{-11}$ practically all the information is recovered; this means that the system is functional with noise levels equal to or less than this value. Figure 11 shows the SINR (dB) resulting from the same AWGN standard deviation levels. At the measure that the AWGN parameter is decreased, the SINR parameter is increased. Finally, in Figure 12 we observe that the highest values of SINR cause lower values of BER.

4. Conclusions

Using a set of numerical simulations, the digital data performance transmission for a fiber-radio system was successfully evaluated. All the parameters considered in the numerical simulations correspond to standard values used in optical and wireless commercial systems. VPI Photonics software was used to simulate the transmission of a digital signal at a bit rate of 2.4 GHz through an optical link of 25 km of SM-SF. At this point, it is very important to remark that the effect of the chromatic dispersion was not a significant influence on the transmission through the optical fiber. MATLAB software was utilized to emulate a wireless channel considering four main phenomena inherent to this channel, such as multipath and slow fading, co-channel interference, and AWGN. Moreover, this modeling considered a user in movement within its coverage area cell. The constellation diagram and BER and SINR statistical tests were used to evaluate the performance of this approach. Results demonstrated good robustness of the system, which remains functional even in the presence of noise. In addition, all possible positions of the reference user inside the coverage area were considered. The chosen QPSK format provided good robustness for the wireless transmission. In summary, this numerical simulation finds potential applications in the evaluation of fiber-radio systems, in particular, cellular telephony, the IoT, or any other mobile system.