sensors Article

A Study of Dispersion Compensation of Polarization Multiplexing-Based OFDM-OCDMA for Radio-over-Fiber Transmissions Chih-Ta Yen * and Wen-Bin Chen Department of Electrical Engineering, National Formosa University, Yunlin County 632, Taiwan; [email protected] * Correspondence: [email protected] or [email protected]; Tel.: +886-5-631-5625 Academic Editors: Teen-Hang Meen, Shoou-Jinn Chang, Stephen D. Prior and Artde Donald Kin-Tak Lam Received: 7 July 2016; Accepted: 29 August 2016; Published: 7 September 2016

Abstract: Chromatic dispersion from optical fiber is the most important problem that produces temporal skews and destroys the rectangular structure of code patterns in the spectra-amplitude-coding-based optical code-division multiple-access (SAC-OCDMA) system. Thus, the balance detection scheme does not work perfectly to cancel multiple access interference (MAI) and the system performance will be degraded. Orthogonal frequency-division multiplexing (OFDM) is the fastest developing technology in the academic and industrial fields of wireless transmission. In this study, the radio-over-fiber system is realized by integrating OFDM and OCDMA via polarization multiplexing scheme. The electronic dispersion compensation (EDC) equalizer element of OFDM integrated with the dispersion compensation fiber (DCF) is used in the proposed radio-over-fiber (RoF) system, which can efficiently suppress the chromatic dispersion influence in long-haul transmitted distance. A set of length differences for 10 km-long single-mode fiber (SMF) and 4 km-long DCF is to verify the compensation scheme by relative equalizer algorithms and constellation diagrams. In the simulation result, the proposed dispersion mechanism successfully compensates the dispersion from SMF and the system performance with dispersion equalizer is highly improved. Keywords: spectral-amplitude-coding optical code-division multiple-access (SAC-OCDMA); orthogonal frequency-division multiplexing (OFDM); balanced photo-detector (BPD); multiple access interference (MAI); dispersion equalize

1. Introduction Optical code-division multiple-access (OCDMA) schemes allow multiple users in a local area network (LANs) to access the same fiber channel asynchronously without delays or the need for scheduling. Accordingly, OCDMA represents an attractive access technique for radio-over-fiber (RoF) networks [1,2]. In the RoF system, micro-cells are connected by optical fibers to radio base stations (RBSs), which are connected in turn to central control stations (CSs). Regarding the configuration of RoF networks, a linear bus or passive star optical link is generally preferable since architecture of this type lowers the cost of implementation and enables the straightforward addition of additional access points. In a RoF network with OCDMA access, different RBSs assign different codes. The advantage of the network is easily to add/remove the access points by changing the length of codes. In the past decade, spectral-amplitude-coding OCDMA (SAC-OCDMA) systems have gained increasing attention. When SAC-OCDMA systems are coupled with quasi-orthogonal code, multiple access interferences (MAI) can be eradicated in ideal conditions, rendering SAC-OCDMA systems extremely suitable for RoF systems. The SAC-OCDMA is the most appropriate way for signal

Sensors 2016, 16, 1440; doi:10.3390/s16091440

www.mdpi.com/journal/sensors

Sensors 2016, 16, 1440

2 of 16

transmitting because it is uncomplicated and the information data can be regenerated without any synchronization and can be implemented in analog intensity modulation (IM) radio links. The detailed operational theorem of the decoding scheme has been proposed in [3,4]. The other OCDMA scheme is coding in polarized domain. In spectral polarization coding architectures, light sources are divided into two mutually orthogonal states of polarization (SOPs), one vertical and one horizontal. The two SOPs are then coded and user data are uploaded to the two coded and polarized optical carriers, improving spectrum utilization and increasing the user volume in the system. Moreover, phase-induced intensity noise (PIIN) can be effectively suppressed by combining SAC-OCDMA with polarization division multiplexing (PDM) technology, thereby increasing overall system efficiency [5,6]. However, in the multi-wavelength OCDMA (MW-OCDMA) system, the decoder is sensitive to the relative temporal positioning of optical pulses transmitted simultaneously along different wavelength channels. When the optical fiber transmitting distance of the network is more than 500 m, the system performance will be significantly degraded by chromatic dispersion, which results in temporal skewing among wavelength channels [7]. To improve the distortion, the dispersion compensation scheme is needed in the MW-OCDMA system for compensating dispersion distortion. There are some researches regarding dispersion compensation, such as the following: the compensators that are used to construct a dispersion slope equalizer have been demonstrated in the form of sampled chirped fiber Bragg gratings [8], virtually imaged phased arrays [9], all-pass filters [10], waveguide grating routers with thermal lenses [11], and dispersion compensating fibers (DCFs) [12]. The length of DCFs required to compensate for the dispersion from single-mode fiber (SMF) is long and optical amplifiers will be needed because the insertion loss of DCF is high. Integrated dispersion slope equalizer of AWG-based optical CDMA is then proposed [13]. However, series connections of Mach–Zehnder Interferometers (MZIs) have serious group delay ripples and dispersion compensation is not easier to adjust precisely. Discussion of other issues, such as negative group delay (NGD) at microwave wavelengths, are also presented; the NGD is useful for the compensation of radio frequency/microwave signal delays [14,15]. Orthogonal frequency-division multiplexing (OFDM) is the fastest developing technology in the academic and industrial fields of wireless transmission. OFDM technologies are resistant to frequency-selective interference and demonstrate high transmission rate and high bandwidth usage efficiency. Thus, OFDM is widely applied in broadband communication systems, including three types of fixed-line networks, namely, high-speed digital subscriber line (HDSL), very high-speed digital subscriber line (VDSL), and asymmetric digital subscriber line (ADSL) [16]. Other common applications include digital audio broadcasting (DAB) and broadband wireless access (BWA) systems, as well as the IEEE 802.11 specifications of wireless local area network (WLAN) [17–20]. In this study, we apply several dispersion equalizer algorithms to construct an electronic dispersion compensation (EDC) equalizer of electrical domain plus with DCF of optical domain which suits for the OFDM-OCDMA with Walsh–Hadamard codes along 10 km-long SMF and 4 km-long DCF. The remainder of this paper is organized as follows. In Section 2, the configuration of the proposed OFDM-OCDMA RoF is described. In Section 3, the effect that chromatic dispersion makes to the OFDM-OCDMA RoF system is introduced. In Section 4, simulation result analysis and compensation mechanism of dispersion compensation fiber DCF and electronic dispersion compensation equalizer in the proposed system is interpreted. Brief conclusions are presented in Section 5. 2. System Description of OFDM-OCDMA RoF System Figure 1 presents a block diagram of the proposed OFDM-OCDMA network. In a system architecture, the transmitter largely consists of Nc − 1 base stations (BSs).

Sensors 2016, 16, 1440 Sensors 2016, 16,1440 1440 Sensors 2016, 16,

3 of 16 of16 16 33of

Figure 1. Integrated configuration of orthogonal frequency-division multiplexing-optical codeFigure Integratedconfiguration configuration of orthogonal frequency-division multiplexing-optical codeFigure 1. 1. Integrated of orthogonal frequency-division multiplexing-optical code-division division multiple-access (OFDM-OCDMA) radio-over-fiber (RoF) system. division multiple-access (OFDM-OCDMA) radio-over-fiber (RoF) system. multiple-access (OFDM-OCDMA) radio-over-fiber (RoF) system.

The optical communication simulation software Optisystem 10 which was made by Optiwave The optical communication simulation software Optisystem 10 was made The communication simulation Optisystem 10which which was madeby byOptiwave Optiwave Systems optical Inc., Ottawa, Canada and was usedsoftware to simulate the proposed system architecture. Figure 2 Systems Inc., Ottawa, Canada and was used to simulate the proposed system architecture. Figure 22 Systems Inc., Ottawa, ON, Canada and was used to simulate the proposed system architecture. illustrates a simulation of the OFDM-OCDMA system with three BSs and three receivers in Figure the CS. illustrates aasimulation ofofthe system three BSs in the illustrates simulation theOFDM-OCDMA OFDM-OCDMA systemwith with BSsand andthree threereceivers receivers theCS. CS. The BSs loaded the signals of two OFDM channels ontothree the optical carriers in two in SOPs. A The BSs loaded the signals of two OFDM channels onto the optical carriers in two SOPs. A The BSs loaded the signals of two OFDM channels onto the optical carriers in two SOPs. A polarization polarization combiner was used to combine the polarized signals and transmit the combined signals polarization combiner was usedthe to combine the polarized signals and transmit the combined signals combiner was used combine signals andontransmit the combined to of anthe encoder. to an encoder. Thetoencoder codedpolarized the signals based a codeword assignedsignals to each BSs. to an encoder encoder.coded The encoder coded the on signals based on a codeword assigned eachSubsequently, of the BSs. The the signals based a codeword to each of the to BSs. Subsequently, a combiner was used to collect the codedassigned signals, and the signals were transmitted to Subsequently, a combiner was used to collect the coded signals, thetransmitted signals were to athe combiner was used to collect the coded signals, and the signalsand were to transmitted the CS. CS. the CS. BS#1 BS#1

010..

OFDM Modulator#1

OFDM Modulator#1 010.. PRBS Generator Bit rate=Bite rate Bits/s PRBS Generator Bit rate=Bite Power rate Bits/s Combiner 4x1 Power Combiner 4x1

CS CS

Mach-Zehnder Modulator Mach-Zehnder Modulator

Receiver#1 Receiver#1

Polarization Splitter Polarization Combiner Device angle=0 deg Device angle=0 deg DFB Laser Array ES Polarization Splitter Polarization Combiner Encoder#1 Frequency=193.1 THzDevice angle=0 deg Mach-Zehnder Modulator_1 Device angle=0 deg DFBFrequency Laser Array ES spacing=100 GHz Encoder#1 Frequency=193.1 THz Power=10 dBm Mach-Zehnder Modulator_1 Frequency spacing=100 GHz Power=10 dBm

decoder(v)#1

decoder(v)#1 Polarization Splitter_1 Polarization Splitter_1

BS#2 BS#2

010.. 010.. PRBS Generator_1

Electrical Gain_1 Gain=100 Electrical Gain_1 Gain=100

decoder(h)#1

OFDM Modulator#2

Electrical Gain Gain=100 Electrical Gain Gain=100

Electrical Adder Electrical Adder

decoder(h)#1

OFDM Modulator#2

Bit rate=Bite rate Bits/s PRBS Generator_1 Bit rate=Bite Power rate Bits/s Combiner 4x1

Power Combiner 4x1

Receiver#2 Receiver#2 Mach-Zehnder Modulator_2 Mach-Zehnder Modulator_2

decoder(v)#2

decoder(v)#2

Polarization Splitter_2 Polarization Combiner_1 Device angle=0 deg DFB Laser Array ES_2 Device angle=0 deg Encoder#2 Polarization Combiner_1 Frequency=193.1 THzPolarization Splitter_2 Mach-Zehnder Modulator_3 Device angle=0 deg spacing=100 GHzangle=0 deg DFBFrequency Laser Array ES_2 Device Encoder#2 Power=10 dBm Frequency=193.1 THz Mach-Zehnder Modulator_3 Frequency spacing=100 GHz Power=10 dBm

Power Combiner Power Combiner

Power Splitter Power Splitter

Polarization Splitter_4 Polarization Splitter_4

Electrical Gain_3 Gain=100 Electrical Gain_3 Gain=100

decoder(h)#2

Electrical Gain_2 Gain=100 Electrical Gain_2 Gain=100

Electrical Adder_1 Electrical Adder_1

BS#3

Receiver#3

OFDM Modulator#3

Receiver#3

010.. PRBS Generator_2 Bit rate=Bite rate Bits/s PRBS Generator_2 Combiner 4x1 Bit rate=Bite Power rate Bits/s

OFDM Modulator#3

Power Combiner 4x1

OFDM Demodulator#2 OFDM Demodulator#2

decoder(h)#2

BS#3

010..

OFDM Demodulator#1 OFDM Demodulator#1

decoder(v)#3

decoder(v)#3 Mach-Zehnder Modulator_4 Mach-Zehnder Modulator_4

Polarization Splitter_3 Polarization Combiner_2 Device angle=0 deg DFB Laser Array ES_2 Device angle=0 deg Encoder#3 Frequency=193.1 THzPolarization Splitter_3 Polarization Combiner_2 Mach-Zehnder Modulator_5 Frequency spacing=100 GHz Device angle=0 deg Device angle=0 deg DFBPower=10 Laser Array ES_2 dBm Encoder#3 Frequency=193.1 THz Mach-Zehnder Modulator_5 Frequency spacing=100 GHz Power=10 dBm

Polarization Splitter_5 Polarization Splitter_5

Electrical Gain_5 Gain=100 Electrical Gain_5 Gain=100

decoder(h)#3

Electrical Gain_4 Gain=100 Electrical Gain_4 Gain=100

Electrical Adder_2 Electrical Adder_2

OFDM Demodulator#3 OFDM Demodulator#3

decoder(h)#3

Figure 2. An architectural diagram of an OFDM-OCDMA system with three BSs. Figure 2. An architectural diagram of an OFDM-OCDMA system with three BSs. Figure 2. An architectural diagram of an OFDM-OCDMA system with three BSs.

Signals transmitted from the transmitter were the cumulative signals of all active BSs. These Signals the transmitter were the cumulative signals of illustrated all active BSs. These1 signals weretransmitted transmittedfrom via optical fiber to the receiver. The receiving signals in Figure signals were transmitted via optical fiber to the receiver. The receiving signals illustrated in Figure 1

Sensors 2016, 16, 1440

4 of 16

Signals transmitted from the transmitter were the cumulative signals of all active BSs. These signals were transmitted via optical fiber to the receiver. The receiving signals illustrated in Figure 1 of star coupler are the cumulative signals of all the BSs, which can be expressed using the following equation: K

r (t) = rv (t) + rh (t) =

∑ [ Ik (t) + Qk (t)],

(1)

k =1

where t is time parameter; rv (t) and rh (t) represent the two receiving signals in the SOPs; Ik (t) and Qk (t), respectively, represent the encoded signals of I-channel and Q-channel that were modulated by BSs onto the optical carriers; and K represents the number of BSs. However, the present study used the Hadamard code. Therefore, K = Nc − 1, where Nc represents the length of the code. The CS consisted of an equal number of receivers as the number of BSs. Thus, each receiver decoded the signals based on the codeword of one BS. Finally, an OFDM demodulator was used to restore the electrical signals to digital signals. The output signals in the vertical and horizontal SOPs for Receiver #l can be expressed using Equations (2) and (3). The results of the two equations indicate that the proposed method can completely eliminate interference between BSs in ideal conditions.   Nc · I (t), k 2 rv (t) · Cl − rv (t) · C l =  0,   Nc · Q (t), k 2 rh (t) · Cl − rh (t) · C l =  0,

k=l

,

(2)

,

(3)

k 6= l k=l k 6= l

Finally, the balanced electrical signals in the two SOPs were combined to form complete OFDM signals, which were sent to the OFDM demodulator for demodulation, restoring the electrical signals to digital signals, as expressed in Equation (4). 

   Nc rv (t) · Cl − rv (t) · C + rh (t) · Cl − rh (t) · C l = [ I (t) + Qk (t)], 2 k

(4)

In the system simulations, the present study assumed that the system adopted a back-to-back method to transmit signals. Only the attenuation of the fiber channel was investigated in this section. The effects of CD are discussed in Sections 3 and 4. Moreover, the OFDM signals transmitted by the wireless transmitter were set at ideal conditions because only the non-ideal effects of the fiber channel were examined in this study. The simulation parameters are tabulated in Table 1. Table 1. The Simulation Parameters. Modulation Technique

4-QAM

Data Length Data Transmission Rate IFFT Size Upscale Frequency (fu ) Electronic Filter Bandwidth (B) Input power Number of Optical Carriers Laser Beam Width Optical Carriers Spacing AWG Bandwidth

32,768 bits 10 G bits/s 1024 7.5 GHz 6 GHz 10 dBm 4 5 MHz 100 GHz 15 GHz

Figure 3 is an architectural diagram of the OFDM modulator and optical encoder in a simulation environment. In Figure 3a; a PRBS generator was used to produce random digital signals with a length

Sensors 2016, 16, 1440

5 of 16

Sensors 2016, 16, 1440

5 of 16

of 32,768 bits and a transmission rate of 10 G (bits/s). The signals were then transmitted to the OFDM modulator to produce OFDM signals. In the OFDM modulator, the inputThe digital signals were processed processed using four-quadrature amplitude modulation (4-QAM). modulated signals were using four-quadrature amplitude The modulated signals were transmitted to the transmitted to the OFDM block modulation to produce (4-QAM). OFDM signals with an inverse fast Fourier transform OFDMsize block produce OFDM with an fast Fourier transform (IFFT) size ofconsistent 1024 bits. (IFFT) of to 1024 bits. In ordersignals to maintain theinverse wave pattern of the transformed signals In order to maintain the wave pattern of the transformed signals consistent with that of the original with that of the original signals, only the middle 512 carriers were used to load signals. The remaining signals, were only the middle 512 carriers were used to load signals. The remaining were loaded carriers loaded with zero signals. The carriers with zero signals imposedcarriers no influence on the with zero signals. The carriers with zero signals imposed no influence on the transformed signals. transformed signals. Then, a low-pass filter was used to de-noise the signals of I-channel and QThen, a low-pass wasOFDM used toblock. de-noise signals of I-channel Q-channel produced by the channel producedfilter by the Thethe two baseband signalsand were separately multiplied by OFDM block. The two baseband signals were separately multiplied by cos ( 2π f t ) and − sin ( 2π f t) u u cos(2 fut ) and  sin(2 fut ) to upscale them to 7.5 G. The spectrums of the two baseband signals to upscale them to 7.5 G. The spectrums of the two baseband signals are illustrated in Figure 4a,b. are illustrated in Figure 4a,b. The overall bandwidth of the signals is 15 G. However, the primary The overall bandwidth of the signals is 15 G. However, the primary signal range was between 5 and signal range was between 5 and 10 G, and the preceding and succeeding sections (5 G) were the guard 10 G, and the preceding and succeeding sections (5 G) were the guard intervals. intervals.

(a)

(b)

(c) Figure 3. An architectural diagram of the base station (BS) in a simulation environment: (a) the base Figure 3. An architectural diagram of the base station (BS) in a simulation environment: (a) the base station; (b) the OFDM modulator; and (c) the OCDMA encoder. station; (b) the OFDM modulator; and (c) the OCDMA encoder.

A distributed feedback laser (DFB) laser array was used to produce four light sources with a linewidth of 5 MHz and central frequencies of 193.1, 193.2, 193.3, and 193.4 THz to serve as the optical carriers (Figure 4c). The optical carriers were polarized into two different SOPs (vertical and

Sensors 2016, 16, 1440

6 of 16

A distributed feedback laser (DFB) laser array was used to produce four light sources with a linewidth of 5 MHz and central frequencies of 193.1, 193.2, 193.3, and 193.4 THz to serve as the Sensorscarriers 2016, 16, 1440 6 of 16 optical (Figure 4c). The optical carriers were polarized into two different SOPs (vertical and horizontal) using a polarization splitter. In addition, an ideal Mach-Zehnder modulator (MZM) horizontal) using a polarization splitter. In addition, an ideal Mach-Zehnder modulator (MZM) was was employed to modulate the upscaled signals of I-channel and Q-channel onto the two types of employed to modulate the upscaled signals of I-channel and Q-channel onto the two types of polarized optical carriers. All optical carriers and I-channel and Q-channel signals possess a 5 G polarized optical carriers. All optical carriers and I-channel and Q-channel signals possess a 5 G guard guard interval, as illustrated in Figure 4d,e. Figure 4f illustrates a spectral diagram of the combined interval, as illustrated in Figure 4d,e. Figure 4f illustrates a spectral diagram of the combined vertically and horizontally polarized signals using the polarization combiner. Finally, the combiner vertically and horizontally polarized signals using the polarization combiner. Finally, the combiner was used to to collect #1 and and #3 #3ofofthe theAWG AWGbased basedononcodeword codeword was used collectthe thesignals signalsfrom fromOutput Output Terminals Terminals #1 (1010) and the signals were transmitted to the receivers. Figure 4g illustrates a spectral diagram (1010) and the signals were transmitted to the receivers. Figure 4g illustrates a spectral diagramofofthe encoded signals. the encoded signals.

(a)

(b)

(c)

(d)

(e)

(f) Figure 4. Cont.

Sensors 2016, 16, 1440

7 of 16 Sensors 2016, 16, 1440

7 of 16

Sensors 2016, 16, 1440

7 of 16

(g) Figure 4. A spectral diagram of the signals at the various nodes of the BS: (a) the upscaled signals in Figure 4. A spectral diagram of the signals at the various nodes of the BS: (a) the upscaled signals in I-channel; (b) the upscaled signals in Q-channel; (c) the optical carrier; (d) the modulated, vertically I-channel; (b) the upscaled signals in Q-channel; (c) the optical carrier; (d) the modulated, vertically polarized(g) signals; (e) the modulated, horizontally polarized signal; (f) the combined signals; and (g) polarized signals; (e) the modulated, horizontally polarized signal; (f) the combined signals; and (g) the the encoded signals. encoded signals. Figure 4. A spectral diagram of the signals at the various nodes of the BS: (a) the upscaled signals in I-channel; (b) the upscaled signals in Q-channel; (c) the optical carrier; (d) the modulated, vertically

Figure 5 illustrates an architectural diagram of a receiver in a simulation environment. First, a Figure 5 illustrates anthearchitectural diagrampolarized of a receiver inthe a simulation environment. First, polarized signals; (e) modulated, horizontally signal; (f) combined signals; and (g) polarization splitter was used to split the received signals into vertically and horizontally polarized the encoded signals. a polarization splitter was used to split the received signals into vertically and horizontally polarized signals. Then, the two polarized signals were transmitted to the decoder for decoding. Figure 6a,b signals. Then, the two polarized signals were transmitted to the decoder for decoding. Figure 6a,b illustrate spectral diagrams of athe two polarized signals.environment. First, a Figure 5 illustrates an the architectural diagram of receiver in a simulation illustrate the spectral diagrams of the two polarized signals. In the decoder (using the decoder for the vertical SOPhorizontally as an example), an AWG was used to filter polarization splitter was used to split the received signals into vertically and polarized In the decoderthe (using the decoder for the vertical SOP as an example), an AWG was used optical carrier and the attached OSSBtosignals. The signals were transmitted signals. Then, the two polarized signals were transmitted the decoder for decoding. Figure 6a,b to the photodetectors to filter the optical carrier and the attached OSSB signals. The signals were transmitted to the to convert optical electrical data. Then, the signals were balanced using the BS codeword. illustrate the spectral diagrams of thedata twointo polarized signals. photodetectors to convert optical data into electrical data. Then, the signals were balanced using In the decoder (using the #1 decoder for the the vertical SOP assignals an example), AWG was used to Combiner collected digital fromanPhotodetector #1filter and the Photodetector #3 BS codeword. Combiner #1 attached collected the#2digital signals from Photodetector #1the and Photodetector #3 the optical carrier OSSB signals. The signals were transmitted to photodetectors C1 ), the ( and and Combiner collected the digital signals from Photodetector #2 and Photodetector #4 ( C1 (C1 ),to and Combiner collected the digital signals Photodetector #2using and Photodetector #4 (C1 ). convert optical#2 data into electrical data. Then, thefrom signals were balanced the BS codeword. ). The signals of the two combiners were subtracted from one another to eliminate MAI, as illustrated Combiner #1 two collected the were digitalsubtracted signals from Photodetector and Photodetector #3 The signals of the combiners from one another to #1 eliminate MAI, assimilar illustrated in of the decoder in Figure 5b. The operations of the decoder for the horizontal SOP were to those C ( ), and Combiner #2 collected the digital signals from Photodetector #2 and Photodetector #4 ( C Figure 15b. The operations the decoder the horizontal SOPthe were similar to thoseof ofthe the signals decoder 1 for theofvertical SOP. for Figure 6c,d illustrate spectral diagrams infor I-channel and Q). The signals the two combiners were subtracted from one another to eliminate MAI, as illustrated the vertical SOP. of Figure 6c,d illustrate the spectral diagrams of the signals in I-channel and Q-channel channel magnified at 100×, respectively. The signals were magnified 100× before they were in Figure of to the decoder fordemodulator the horizontal SOP100 were similar those ofinthe decoder andtoQ-channel were magnified at5b. 100The ×, operations respectively. The signals were magnified ×after before they were transmitted transmitted the OFDM because theto signals I-channel for the vertical SOP. Figure 6c,d illustrate the spectral diagrams of the signals in I-channel and Q- onto the OFDM demodulator because after the signals in I-channel andsignal Q-channel modulated modulated onto the optical carriers and encoded, powerwere decreased due to insertion loss, which channel magnified at 100×, respectively. The signals were magnified 100× before they were the optical carriers drastically and encoded, signal power decreased due to insertion loss, which drastically reduced the power of the signals once they were detected by the receivers. Therefore, the transmitted to the OFDM demodulator because after the signals in I-channel and Q-channel were reduced the power detected of the signals once they were weremagnified detected by the Therefore, detected electrical signals 100× to receivers. fully display the signalthe spectrum in the oscilloscope modulated onto the optical carriers and encoded, signal power decreased due to insertion loss, which box and to compensate for the loss of signal power during insertion loss andbox transmission process. electrical signals were magnified 100 × to fully display the signal spectrum in the oscilloscope and drastically reduced the power of the signals once they were detected by the receivers. Therefore, the In the OFDM demodulator block, a down-converter was used to downscale the balanced signals to compensate for the loss of signal power during insertion loss and transmission process. detected electrical signals were magnified 100× to fully display the signal spectrum in the oscilloscope to baseband OFDM signals. The downscaled signals were transmitted to a built-in OFDM In the demodulator down-converter was used toloss downscale the balanced signals to box andOFDM to compensate for theblock, loss ofasignal power during insertion and transmission process. demodulation blocka for demodulation, a 4-QAM demodulation basebandInOFDM signals. The downscaled signals were transmitted totoa built-in demodulation the OFDM demodulator block, down-converter wasand usedthen to downscale theOFDM balanced signals block to restore the signals to digital as illustrated Figure 5c. to Finally, two signals combined to form baseband OFDM signals. The downscaled signals in were transmitted to the a the built-in OFDM blocktofor demodulation, and then tosignals, a 4-QAM demodulation block restore signals towere digital complete OFDM signals, which were transmitted to the OFDM demodulator for demodulation block for demodulation, and then to a 4-QAM demodulation block to restore the signals, as illustrated in Figure 5c. Finally, the two signals were combined to form complete OFDM demodulation. 6e ainspectral diagram ofthe the complete signals.6etoillustrates signals to digital signals, asillustrates illustrated Figure 5c. Finally, signalsOFDM were combined form signals, which wereFigure transmitted to the OFDM demodulator fortwo demodulation. Figure a complete OFDM signals, which were transmitted to the OFDM demodulator for demodulation. spectral diagram of the complete OFDM signals. Figure 6e illustrates a spectral diagram of the complete OFDM signals.

(a) (a) Figure 5. Cont.

Sensors 2016, 16, 1440 Sensors 2016, 16, 1440

8 of 16 8 of 16

Sensors 2016, 16, 1440

8 of 16

(b)

(b)

(c)

(c)

Figure 5. An architectural diagram of a receiver in a simulation environment: (a) the architectural Figure 5.5. An diagram of in aasimulation environment: (a) the architectural Figure Anofarchitectural architectural diagram of aareceiver receiverfor inthe simulation (a)(c) the diagram the receiver; (b) the OCDMA decoder vertical SOP environment: (decoder(v)); and thearchitectural OFDM diagram of the the receiver; receiver;(b) (b)the the OCDMA decoder the vertical (decoder(v)); the diagram OCDMA decoder for for the vertical SOP SOP (decoder(v)); and (c)and the (c) OFDM demodulator.

OFDM demodulator. demodulator.

(a)

(b)

(a)

(b)

(c)

(d)

(c)

(d) Figure 6. Cont.

Sensors Sensors2016, 2016,16, 16,1440 1440

Sensors 2016, 16, 1440

99of of16 16

9 of 16

(e)

(e) Figure 6. Spectrograms of each node signal: (a) vertical SOP signal at the receiver; (b) horizontal SOP Figure 6. Spectrograms of each node signal: (a) vertical SOP signal at the receiver; (b) horizontal SOP Figure of each signal: (a) vertical SOP balance signal at the receiver;and (b) horizontal SOP signal 6.atSpectrograms the receiver; (c) node I-channel signal after detection amplification; signal at the receiver; (c) I-channel signal after balance detection and amplification; signal at the receiver; (c) I-channel signal after detection (d) Q-channel (d) Q-channel signal after balance detection andbalance amplification; andand (e) amplification; the complete OFDM signal at (d) Q-channel signal after balance detection and amplification; and (e) the complete OFDM signal at signal after balance detection and amplification; and (e) the complete OFDM signal at the receiver. the receiver. the receiver.

In In order orderto toaddress addressproblems problemsof oftransmission transmissiondistance, distance, optic optic fiber fiberwas wasused usedas asthe thetransmission transmission In order to address problems of transmission distance, optic fiber was used as the transmission medium. medium. However, this section only investigates the optic fiber fiber simulation simulation results results after after adding adding medium. However, this section only investigates the optic fiber simulation results after adding attenuation attenuation factors; factors; the the simulation simulation results results after after adding adding chromatic chromatic dispersion dispersion (CD) (CD) parameters parameters are are attenuation factors; the simulation results after adding chromatic dispersion (CD) parameters are analyzed analyzedin indetail detailin innext nextsections, sections,and andaacompensation compensationmechanism mechanismisisalso alsoproposed. proposed. analyzed in detail in next sections, and a compensation mechanism is also proposed. To Todetermine determineaccuracy, accuracy,the thepresent presentstudy studyemployed employedMatlab Matlabto toprocess processthe thedigital digitalsignals signals of of the the To determine accuracy, the present study employed Matlab to process the digital signals of the transmitter through 4-QAM demodulation in an access environment. However, transmittermodulated modulated through 4-QAM demodulation in ansimulation access simulation environment. transmitter modulated through 4-QAM demodulation in an access simulation environment the usable data length data was 32,768 limited memory, suggesting that when using a single witha However, the usable lengthdue wasto32,768 due to limited memory, suggesting that whenBS using However, the usable data length was 32,768 due to limited memory, suggesting that when using a predetermined parameters, the receivers of the OFDM-OCDMA system obtained the minimum single BS with predetermined parameters, the receivers of the OFDM-OCDMA system obtained bit the single BS with predetermined parameters, the receivers of the OFDM-OCDMA system obtained the −5 . error rate (BER) measured value is limitedvalue in 3 × minimum bit error rate (BER) measured is 10 limited in 3 × 10−5. minimum bit error rate (BER) measured value is limited in 3 × 10−5. Figure Figure77isisBER BERcurve curvediagrams diagramsat atdifferent differenttransmission transmissiondistances. distances. Regarding Regarding the the transmission transmission Figure 7 is BER curve diagrams at different transmission distances. Regarding the transmission distance, distance, keeping keeping the the input input laser laser source source power power fixed fixed at at 10 10dBm dBmand andsetting setting the theattenuation attenuation factor factor in in distance, keeping the input laser source power fixed at 10 dBm and setting the attenuation factor in the the optic optic fiber fiberat at0.2 0.2dB, dB,the theBER BERincreases increaseswith withincreased increaseddistance. distance. The The BER BER for forsystem system architecture architecture the optic fiber at 0.2 dB, the BER increases with increased distance. The BER for system architecture with withaadata datalength lengthof of32,768 32,768was was“0” “0”when whenthe thetransmitted transmitteddistance distanceisissmaller smallerthan than80 80km. km. The The cluster cluster with a data length of 32,768 was “0” when the transmitted distance is smaller than 80 km. The cluster points points in in the theconstellation constellationdiagram diagram also alsobecome becomegradually graduallymore moreblurred. blurred. In In addition, addition, calculations calculations points in the constellation diagram also become gradually more blurred. In addition, calculations using using the the Matlab Matlab software software indicated indicated that that the the receiver receiver could could completely completely restore restore the the received received signals signals using the Matlab software indicated that the receiver could completely restore the received signals without errors when the transmission distance was shorter than 80 km. without errors when the transmission distance was shorter than 80 km. without errors when the transmission distance was shorter than 80 km.

Figure 7. Curve diagram of bit-error-rate (BER) in relation to different transmission distances. Figure 7. Curve diagram of7.bit-error-rate (BER) relation to different distances. Figure Curve diagram of in bit-error-rate (BER) in transmission relation to different transmission distances.

Sensors 2016, 16, 1440

10 of 16

3. Influence of Dispersion on an OFDM-OCDMA System Dispersion is a type of physical effect caused by different group velocities that generate different time delays when optical carrier waves with different wavelengths are transmitted through an optic fiber. Three main types of dispersion exist, namely modal dispersion, CD, and polarization dispersion. CD can further be divided into material dispersion and waveguide dispersion. In multimode fibers (MMFs), modal dispersion is greater than are material dispersion and waveguide dispersion, making modal dispersion the principal type of dispersion in MMFs. Regarding SMFs, modal dispersion does not exist because only one mode is being transmitted in the optic fiber; only CD and polarization dispersion exist. This study adopts SMF as the transmission medium. Furthermore, the influence of CD in a multi-wavelength system is severer than that of polarization dispersion; therefore, simulation analysis is performed in this section after introducing the dispersion parameters. The CD is caused by different time delays produced when light with different wavelengths are transmitted in an optic fiber at different velocities. CD can also be divided into material dispersion and waveguide dispersion, and is generally referred to as dispersion. Because actual light sources are not made up of pure monochromatic light and the SMF refraction index varies among light of different wavelengths, the time delay is dissimilar for light of different wavelengths, causing a broadening of the output light pulse. The output pulse expansion is also greater when the optic fiber is longer, which not only causes interference among the data and increases the BER, but also limits the system’s transmission rate and bandwidth. The material dispersion-induced broadening per unit length can be expressed as in Equation (5). ∆τm = | Dm (λ)| ∆λ, L

(5)

where the suffix m specifies material dispersion, L is the optic fiber length, ∆τm is the pulse expansion time caused by material dispersion, as all light emission sources emit nonmonochromatic light within a specific wavelength range, and Dm (λ) is the material dispersion coefficient. Waveguide dispersion is related to the fundamental mode group velocity and the parameter V, and V is also related to the light source and wavelength. Because light sources have specific frequency spectra, the value of V is different when light of different wavelengths is transmitted in the optic fiber, causing the group velocity to differ as well. This further causes the time delay with which each wavelength reaches the end of the optic fiber to differ, inducing a broadening phenomenon of the output signal. Waveguide dispersion differs from material dispersion, in which waveguide dispersion is related to the group velocity and V, but the group velocity of material dispersion is related to the refraction index. The pulse broadening per unit length caused by waveguide dispersion is expressed in Equation (6), which is similar to Equation (5). ∆τw = | Dw (λ)| ∆λ, L

(6)

where the suffix w specifies material dispersion, ∆τw is the pulse expansion time caused by material dispersion and Dw (λ) is the waveguide dispersion coefficient, which is related to the characteristics of the optic fiber. The general literature collectively refers to the effects caused by material and waveguide dispersion as dispersion; thus, the total dispersion per unit length can be written as Equation (7). ∆τCD = | Dm (λ) + Dw (λ)| ∆λ, L

(7)

where ∆τCD is the output pulse expansion time caused by dispersion. As previously described, different wavelengths of light reach the receiver at distinct times. Therefore, when the OFDM signal has been carried onto the optical carrier wave, each subcarrier signal

Sensors 2016, 16, 1440

11 of 16

Sensors 2016, 16, 1440

11 of 16

Figure 8 shows that if no cyclic prefix is added, the receiver will lack part of the signal when it retrieves the signal. Furthermore, if the receiver retrieves other OFDM signals, an intersymbol of OFDM respectively an greatly optical reducing carrier wave, and willeffectiveness. cause the signals oncolors each interference (ISI) effect represents is generated, the system’s The carried different subcarrier to reach the receiver at different times after transmission through the optic fiber. mean different frequencies of modulated OFDM signal. By adding a cyclic prefix in the OFDM Figure 8the shows that if nodata cyclic added,can thebe receiver will lack part the signal when technology, OFDM signal lossprefix at theisreceiver compensated and theofinfluence of ISI on it retrieves the be signal. Furthermore, if the study, receiver retrieves other aOFDM an intersymbol the signal can suppressed. In a related scholars utilized single signals, optical carrier wave to interference (ISI)data, effectand is generated, greatly the system’s effectiveness. The different colors transmit OFDM they deduced thereducing cyclic prefix time length required by OFDM signals to mean different frequencies of ISI modulated OFDM because signal. By cyclic prefix in the utilizes OFDM suppress dispersion-induced [21]. However, theadding presenta system principally technology, the OFDM signal data loss at data, the receiver can be the alone influence of ISI multiple optical carrier waves to transmit the method of compensated adding a cyclicand prefix cannot be on the canISI be and suppressed. In a relatedinterference. study, scholars utilized a single optical carrier wave used tosignal suppress dispersion-induced to transmit OFDM deduced the prefix was timeadopted length required by OFDM signals In general, the data, use ofand DCFthey to compensate forcyclic dispersion in the proposed system. In to suppress dispersion-induced ISI [21]. However, because theDCF, present addition, because most of the dispersion is suppressed by the thesystem length principally of the cyclicutilizes prefix multiple carrier waves to transmit data, the method of adding cyclic prefix cannot added to optical the OFDM signal in this system architecture only needs to be aidentical to thealone length used be in used to suppress ISI and dispersion-induced interference. general OFDM technology (e.g., 1/4 of the inverse fast Fourier transform (IFFT) size).

Figure Figure 8. 8. Schematic Schematic diagram diagram of of dispersion-induced dispersion-induced ISI ISI effect. effect.

4. Simulation Result Analysis and Compensation Mechanism In general, the use of DCF to compensate for dispersion was adopted in the proposed system. In addition, becauseanalysis most of the dispersionon is the suppressed by results the DCF, of the cyclicsystem prefix In this section, is conducted simulation ofthe thelength OFDM-OCDMA added to the OFDM signal in this system architecture only needs to be identical to the length used in after adding the dispersion parameters, and the compensation ability of the proposed EDC equalizer general OFDM technology (e.g., 1/4 parameters of the inverse Fourier in transform is investigated. The SMF simulation arefast presented Table 2. (IFFT) size). 4. Simulation Result Analysis and Compensation Mechanism Table 2. SMF simulation parameters. In this section, analysis is conducted on the simulation results Wavelength 1550 nm of the OFDM-OCDMA system after adding the dispersion parameters, and the compensation ability of the proposed EDC equalizer is Attenuation 0.2 dB/km investigated. The SMF simulation parameters are presented in ×Table Dispersion (DSMF) 17 ps/nm km 2.

Dispersion slope

0.056 ps/nm2 × km

Table 2. SMF simulation parameters.

The parameters in Table 2 were selected mainly by referring to the SMF specifications set by Wavelength 1550 nm the International Telecommunication Union and the0.2specification to the SMF-28-J9 sold by Attenuation dB/km Thorlabs [22,23]. Dispersion (DSMF ) 17 ps/nm × km 2 × km Dispersion slope of 0.056 ps/nmsignal Figure 9 is a receiver constellation diagram an optical after having been subjected to 10 km SMF transmission and dispersion influence. The unit a.u. in Figure 9 means arbitrary units. AfterThe calculation, the in BER of this constellation diagram foundto to the be 0.485. The transmitted parameters Table 2 were selected mainly bywas referring SMF specifications setoptical by the signal causes the receiver constellation diagram to become blurred, making it impossible to modulate International Telecommunication Union and the specification to the SMF-28-J9 sold by Thorlabs [22,23]. the increased rate (Figure 9). Figure 9 error is a receiver constellation diagram of an optical signal after having been subjected to 10 km SMF transmission and dispersion influence. The unit a.u. in Figure 9 means arbitrary units. After calculation, the BER of this constellation diagram was found to be 0.485. The transmitted optical signal causes the receiver constellation diagram to become blurred, making it impossible to modulate the increased error rate (Figure 9).

Sensors 2016, 16, 1440 Sensors 2016, 16, 1440

12 of 16 12 of 16

Figure Figure9. 9. Receiver Receiver constellation constellation diagram diagram after after10 10km kmof ofSMF SMFtransmission. transmission.

The The main main reason reasoncausing causingthe theOFDM-OCDMA OFDM-OCDMAsystem system receiver receiverconstellation constellationdiagram diagram to tobecome become blurred and concentrated is that this system utilizes multiple optical carrier waves to transmit blurred and concentrated is that this system utilizes multiple optical carrier waves to transmit the the OFDM carrier waves of different wavelengths that that are transmitted through an optic OFDMsignal, signal,and andoptical optical carrier waves of different wavelengths are transmitted through an fiber reach thereach receiver differentattimes because of dispersion. In addition, in the opticwill fiber will theatreceiver different times because of dispersion. In OFDM-OCDMA addition, in the system, the OFDM signalsthe entrained in eachentrained optical carrier arecarrier also influenced by dispersion, OFDM-OCDMA system, OFDM signals in eachwave optical wave are also influenced causing the signals to reach the receiver at inconsistent times. Therefore, when the receiver by dispersion, causing the signals to reach the receiver at inconsistent times. Therefore, performs when the signal collection balance detection on the detection two polarization states, the phasestates, signals the receiver performsand signal collection and balance on the two polarization theof phase I-channel and Q-channel will be destroyed because of frequency dispersion. In this type of situation, signals of the I-channel and Q-channel will be destroyed because of frequency dispersion. In this type simply using EDC equalizer cannot compensate for the signal. This of situation, simply using technology EDC equalizer technology cannot compensate fordemonstrates the signal. that This compensation cannot be performed directly on the electrical signal when transmitting optical signals demonstrates that compensation cannot be performed directly on the electrical signal when in an OFDM-OCDMA system.inTherefore, compensation must Therefore, first be performed in the initial transmitting optical signals an OFDM-OCDMA system. compensation muststage first of be the optical transmission domain; the dispersion that could not be fully compensated is subsequently performed in the initial stage of the optical transmission domain; the dispersion that could not be compensated using is EDC equalizer technology. fully compensated subsequently compensated using EDC equalizer technology. Regarding the influence Regarding the influence of ofdispersion, dispersion, this this study studyadopted adopted DCF DCF to tocompensate compensate for fordispersion dispersion in in the the optical opticaldomain. domain. The The simulation simulation parameters parametersare arepresented presentedin inTable Table3.3. Table 3. DCF simulation parameters. Table 3. DCF simulation parameters. Wavelength Wavelength Attenuation Attenuation Dispersion (DDCF ) Dispersion (DDCF) Dispersion slope

Dispersion slope

1550nm nm 1550 0.25dB/km dB/km 0.25 −38.5 ps/nm × km −38.5 ps/nm ×2 km −0.12 ps/nm × km −0.12 ps/nm2 × km

The The parameters parameters in in Table Table33were wereselected selectedprimarily primarily by by referring referring to to the the specification specification to to the the DCF DCF with model number DCF38 sold by Thorlabs [24]. In addition, the system’s total dispersion with model number DCF38 sold by Thorlabs [24]. In addition, the system’s total dispersionamount amount can canbe beobtained obtainedusing usingEquation Equation(8). (8).

DSMF LSMF +  DDCF , , Dall D =allDSMF · L SMF · DCF L DCF DCF L

(8) (8)

where system’s totaltotal dispersion amount; DSMF and SMF and DCF dispersion Dallis the DSMFDDCF whereDall is the system’s dispersion amount; andareDthe DCF are the SMF and DCF values, respectively; LSMF is the SMF transmission distance; and L DCF is the required DCF length. dispersion values, respectively; LSMF is the SMF transmission distance; and LDCF is the required In addition, assuming a situation where DCF cannot completely compensate for the dispersion, DCF length. setting the SMF and DCF lengths to respectively 10 km and 4 km and subsequently utilizing In addition, assuming a situation where DCF cannot completely compensate Equation (8) for calculation, the system’s total dispersion amount can be foundfor to the be dispersion, 16 ps/nm. setting the SMF and DCFobtained lengths to respectively km andin4Figure km and utilizing The constellation diagram after simulation10 is shown 10, subsequently and the BER is 0.5 in Equation (8) for calculation, the system’s total dispersion amount can be found to be 16 ps/nm. The this situation. constellation diagram obtained after simulation is shown in Figure 10, and the BER is 0.5 in this situation.

Sensors 2016, 16, 1440 Sensors Sensors2016, 2016,16, 16,1440 1440

13 of 16 1313ofof1616

Figure set 4 km). Figure10. 10.Receiver Receiverconstellation constellationdiagrams diagramsafter afterDCF DCFcompensation compensation(DCF (DCFlength lengthis setto km). Figure 10. Receiver constellation diagrams after DCF compensation (DCF length isisset toto44km).

Regarding where the dispersion has not been completely compensated, the insertion Regardingsituations situationswhere wherethe thedispersion dispersion has not been completely compensated, insertion Regarding situations has not been completely compensated, thethe insertion of ofofpilot symbols was used totoconduct optic fiber channel estimation and compensate for the phase pilot symbols was used conduct optic fiber channel estimation and compensate for the pilot symbols was used to conduct optic fiber channel estimation and compensate for the phasephase shift. shift. the signal was to be ininEquation (9). shift.First, First, thereceived received wasassumed assumed beasasexpressed expressed Equation (9). First, the received signal signal was assumed to be astoexpressed in Equation (9). rikr xxik hhk nn (9) ik ik, , (9) ik ik k rik = xik ⊗ hk + nik , (9) where wherei irepresents representsthe theith ithdata datasymbol, symbol,k kisisthe thekth kthsubcarrier, subcarrier, rikrik isisthe thereceived receivedsignal, signal, xikxik isisthe the where i represents the ith data symbol, k is the kth subcarrier, rik is the received signal, xik is the transmitted transmitted signal, signal, hhk isis the the fiber fiber channel channel effect, effect, and and nn random noise. noise. Subsequently, Subsequently, aa ik ik isis random transmitted signal, hk isk the fiber channel effect, and nik is random noise. Subsequently, a least-square least-square algorithm was to estimate frequency response ininthe channel, asasshown inin least-square wasused used estimatethe the frequency thefiber fiber channel, shown algorithm wasalgorithm used to estimate thetofrequency response in theresponse fiber channel, as shown in Equation (10), ˆ ˆ Equation where equals FFT hhk k . . where Hˆ ik(10), equals FFT{H hH Equation (10), where kik}ik. equals FFT 1 −1 −1 11 −111 ˆH ˆikXik Hˆ ik = = ·(X N )= ) H  XX HikHikX+ (Xikik ·H X·XikRikR R H Hikik+ NN Xik X1 N  Nik ·, N , ik , ik ikX ik ik ik)  H ik ik

ik

ik

ik

ik

ik

ik

ik

ik

ik

(10) (10) (10)

Regarding Regarding the pilot symbol arrangement, comb-type arrangement was adopted; pilot symbols Regardingthe thepilot pilotsymbol symbolarrangement, arrangement,aaacomb-type comb-typearrangement arrangementwas wasadopted; adopted;pilot pilotsymbols symbols were were periodically inserted on the subcarriers and then removed at the receiver after they had been wereperiodically periodicallyinserted insertedon onthe thesubcarriers subcarriersand andthen thenremoved removedat atthe thereceiver receiverafter afterthey theyhad hadbeen been subjected channel interference, and the calculation was performed. the without subjected interference, and the calculation was performed. ForFor the without pilot subjectedtotochannel channel interference, and the calculation was performed. Forsubcarriers thesubcarriers subcarriers without pilot the method was totoestimate the frequency response. This symbols, the interpolation method was used toused estimate the channel’s frequency response. This study pilotsymbols, symbols, theinterpolation interpolation method wasused estimate thechannel’s channel’s frequency response. This study adopted linear totoconduct The primarily adopted a linearaainterpolation methodmethod to conduct channel estimation. The schematic diagram studyprimarily primarily adopted linearinterpolation interpolation method conductchannel channelestimation. estimation. Theschematic schematic diagram 11 and can using isdiagram shownisin Figurein11 and the interpolation value canvalue be obtained using Equation (11). isshown shown inFigure Figure 11 andthe theinterpolation interpolation value canbe beobtained obtained usingEquation Equation(11). (11).

f(x) f(x)

f(x f(x2)2) f 1f(x) 1(x) f(x f(x1)1)

xx1 1

xx

xx 22

xx

Figure Figure11. 11.Schematic Schematicdiagram diagramof ofthe thelinear linearinterpolation interpolationmethod. method. Figure 11. Schematic diagram of the linear interpolation method.

Sensors 2016, 16, 1440

14 of 16

Sensors 2016, 16, 1440

14 of 16

f ( x2 ) − f ( x ) (11) f 1 ( x ) = f ( x1 ) + f  x2   f  x1 1 ( x − x1 ) f1 ( x)  f ( x1 )  (11) x2 − x1  x  x1  , x2  x1 In Equation (11), the point slope form is utilized to calculate the estimated numerical value Equation the point slope form is utilized the estimated numerical value betweenIntwo points.(11), Furthermore, the estimation resultsto arecalculate typically more accurate the smaller the between two points. Furthermore, the estimation results typically more accurate theissmaller interval between two data points is in situations when theare linear interpolation method used. the interval twoa data points is in situations when the linear interpolation method is used. Figurebetween 12 shows receiver constellation diagram after 10 km of SMF transmission and 4 km Figure 12 shows a receiver constellation diagram after 10 km of SMF transmission andOne 4 kmpilot of of DCF transmission, with no optical amplifier added and after electrical equalization. DCF transmission, with no optical amplifier added and after electrical equalization. One pilot symbol symbol is inserted at intervals of 64 subcarriers. In addition, a cyclic prefix with a length of 1/32 of is inserted at intervals of 64 subcarriers. In addition, a cyclic prefix with a length of 1/32 of the IFFT the IFFT amount is added to suppress ISI. A comparison of Figures 10 and 12 indicates that after amount is added to suppress ISI. A comparison of Figures 10 and 12 indicates that after compensation compensation through linear interpolation equalization, the cluster points in the constellation diagram through linear interpolation equalization, the cluster points in the constellation diagram became became relatively converged and exhibited no phase shift. In addition to adopting linear interpolation relatively converged and exhibited no phase shift. In addition to adopting linear interpolation technology, this study also investigated three common equalizer technologies, namely piecewise cubic technology, this study also investigated three common equalizer technologies, namely piecewise spline interpolation, piecewise cubic cubic Hermite interpolation, and FFT Onthe theother other cubic spline interpolation, piecewise Hermite interpolation, and FFTinterpolation. interpolation. On hand, the constellation diagrams using the linear, piecewise cubic spline, and piecewise cubic Hermite hand, the constellation diagrams using the linear, piecewise cubic spline, and piecewise cubic interpolation methods for channelforestimation are highly similar equalization, and they Hermite interpolation methods channel estimation are highlyafter similar after equalization, andpossess they identical effectiveness, as shown as byshown Figureby 12a–c. possess identical effectiveness, Figure 12a–c.

(a)

(b)

(c)

(d)

Figure 12. Constellation diagrams after estimation and equalization of fiber channels using different Figure 12. Constellation diagrams after estimation and equalization of fiber channels using different interpolation methods: (a) linear interpolation; (b) piecewise cubic spline interpolation; (c) piecewise interpolation methods: (a) linear interpolation; (b) piecewise cubic spline interpolation; (c) piecewise cubic Hermite interpolation; and (d) fast Fourier transform (FFT) interpolation. cubic Hermite interpolation; and (d) fast Fourier transform (FFT) interpolation.

Conversely, the constellation diagram obtained by using the FFT interpolation method is Conversely, the constellation diagram obtained by using the FFT interpolation method is relatively relatively diverged, as indicated by Figure 12d. The main reason for these results is that signals are diverged, as indicated by Figure 12d. The main reason for these results is that signals are subjected to

Sensors 2016, 16, 1440

15 of 16

influence from different factors in different transmission channels. For example, in a wireless channel, the signal is affected by environmental factors and interference from barriers, while a signal transmitted in an optic fiber is influenced by interference from dispersion and other factors. Perhaps using the FFT interpolation method can achieve excellent effectiveness in wireless channels, but the same effectiveness might not be attainable in fiber channels; therefore, the appropriate interpolation method should be selected according to various channel effects. In addition, because of the assumption that the DCF could not completely suppress the dispersion, the received signal will be influenced by small amounts of dispersion. That is why the pilot symbol insertion spacing was increased to 64, and a cyclic prefix with the length 32 (generally 1/4 of the IFFT size) was used to compensate for the dispersion-induced ISI. Regarding the interpolation methods, because the system effectiveness obtained by using linear, piecewise cubic spline, and piecewise cubic Hermite interpolation was identical, the linear interpolation method was adopted as an equalizer technology in the OFDM-OCDMA compensated scheme. The main reason for selecting the linear interpolation method for channel estimation in this study was because the associated calculation amount is relatively small. In Figure 12, we can find that proposed dispersion equalizer scheme of electrical domain plus with DCF of optical domain is effectively suppressing CD effect and hence improving system performance. 5. Conclusions This study investigated the dispersion effects in an OCDMA system combined with a microwave optic fiber network using OFDM and polarization multiplexing technology. In the system architecture presentation, the encoding and decoding architecture utilized arrayed waveguide grating to code the frequency domain amplitude. Subsequently, the design parameters of the arrayed waveguide grating were changed to change the bandwidth and frequency and thus enable the system to transmit two different signal formats (i.e., unilateral and bilateral optical carriers). In addition, polarization multiplexing technology was added into the BS and receiver architecture to suppress interference caused by phase-induced intensity noise and thus enhance the system’s overall effectiveness. In addition, when the two signals were separately transmitted through the optic fiber in a situation where only influence from attenuation factors was considered and the BER equaled 3 × 10−5 , the transmission distance could reach 80 km. When the long distance transmission dispersion effect was added, the signal could no longer be restored after a 10-km transmission. In this study, an ideal effect was initially investigated, and DCF plus with electrical equalizer was used to compensate for dispersion. After the dispersion effect was fully compensated using the DCF, the simulation results indicated that the signal could be restored effectively and that the BER equaled zero when the transmitted data length was 32,768. When the DCF is unable to completely compensate for the dispersion, the received signal will still be affected by small amounts of dispersion and therefore the cluster points in the constellation diagram will exhibit phase shifting, causing the system’s effectiveness to decline. In this situation, the present system utilized OFDM equalizer technology to insert pilot symbols to perform estimation and compensation on the fiber channel. The simulation results showed that this method can effectively suppress interference of the signal from remaining dispersion when the DCF could not attain complete compensation. The method also enables effective restoration of the transmitted signal and enhancement of overall system effectiveness. Acknowledgments: This study was supported in part by the Ministry of Science and Technology 104-2221-E-150-044-. Author Contributions: Chih-Ta Yen performed the simulations and algorithm analysis, and contributed to a main part of manuscript writing. Wen-Bin Chen contributed in conceiving and the corresponding data analysis. All authors contributed to the writing of the paper. Conflicts of Interest: The authors declare no conflict of interest.

Sensors 2016, 16, 1440

16 of 16

References 1. 2. 3. 4. 5. 6. 7. 8.

9. 10. 11.

12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24.

Wu, J.S.; Wu, J.; Tsao, H.W. A radio-over-fiber network for microcellular system application. IEEE Trans. Veh. Technol. 1998, 47, 84–94. [CrossRef] Al-Raweshidy, H.; Komaki, S. Radio over Fiber Technologies for Mobile Communication Networks; Artech House: Boston, MA, USA, 2002. Kavehrad, M.; Zaccarin, D. Optical Code-Division-Multiplexed Systems Based on Spectral Encoding of Noncoherent Sources. IEEE J. Lightwave Technol. 1995, 13, 534–545. [CrossRef] Huang, J.F.; Hsu, D.Z. Fiber-grating-based optical CDMA spectral coding with nearly orthogonal M-sequence codes. IEEE Photonics Technol. Lett. 2000, 12, 1252–1254. [CrossRef] Yen, C.T.; Huang, J.F.; Chang, Y.T.; Chen, B.H. Polarization diversity scheme on spectral polarization coding optical code-division multiple-access network. Opt. Eng. 2010, 49, 125005-1–125005-9. Huang, J.F.; Yen, C.T. Phase noise suppression in multilevel optical code-division multiple-access network coding system with embedded orthogonal polarizations. Opt. Eng. 2006, 45, 065005-1–065005-8. [CrossRef] Ng, E.K.H.; Weichenberg, G.E.; Sargent, E.H. Dispersion in Multiwavelength Optical CDMA Systems: Impact and Remedies. IEEE Trans. Commun. 2002, 50, 1811–1816. [CrossRef] Cai, J.X.; Feng, K.M.; Willner, A.E.; Grubsky, V.; Starodubov, D.S.; Feinberg, J. Simultaneous tunable dispersion compensation of many WDM channels using a sampled nonlinearly chirped fiber Bragg grating. IEEE Photonics Technol. Lett. 1999, 11, 1455–1457. [CrossRef] Shirasaki, M. Chromatic dispersion compensator using virtually imaged phased array. IEEE Photonics Technol. Lett. 1997, 9, 1598–1600. [CrossRef] Madsen, C.K.; Lenz, G.; Bruce, A.J.; Cappuzzo, M.A.; Gomez, L.T.; Scotti, R.E. Integrated all-pass filters for tunable dispersion and dispersion slope compensation. IEEE Photonics Technol. Lett. 1999, 11, 1623–1625. [CrossRef] Doerr, C.R.; Stulz, L.W.; Chandrasekhar, S.; Pafchek, R. Colorless tunable dispersion compensator with 400-ps/nm range integrated with a tunable noise filter. IEEE Photonics Technol. Lett. 2003, 15, 1258–1260. [CrossRef] Vengsarkar, A.M.; Reed, W.A. Dispersion-compensating single-mode fibers: Efficient design for first- and second- order compensation. Opt. Lett. 1993, 18, 924–926. [CrossRef] [PubMed] Yen, C.T. Integrated dispersion slope equalizer of AWG-based optical CDMA for radio-over-fiber transmissions. Photonic Netw. Commun. 2010, 19, 311–319. [CrossRef] Ravelo, B. Investigation on microwave negative group delay circuit. Electromagnetics 2011, 31, 537–549. [CrossRef] Ravelo, B. Distributed NGD active circuit for RF-microwave communication. Int. J. Electron. Commun. (AEÜ) 2014, 68, 282–290. [CrossRef] Bingham, J.A.C. ADSL, VDSL, and Multicarrier Modulation, 1st ed.; John Wiley & Sons, Inc.: New York, NY, USA, 2000. ETSI. Radio Broadcasting System, Digital Audio Broadcasting (DVB) to Mobile, Portable and Fixed Receiver; ETS 300 401; European Telecommunication Standards Institute: Sophia Antipolis, France, 1995. Hoeg, W.; Lauterbach, T. Digital Audio Broadcasting: Principle and Applications; John Wiley & Sons. Inc.: New York, NY, USA, 2000. Reid, N.P.; Seide, R. 802.11(Wi-Fi) Networking Handbook, 1st ed.; McGraw Hill Education-Europe: New York, NY, USA, 2002. IEEE Standards Department. IEEE 802.11 Draft Standard for Wireless LAN Medium Access Control (MAC) and Physical Layer (PHY) Specification; P802.11 D6.1; IEEE Standards Association: Piscataway, NJ, USA, 1997. Shieh, W.; Athaudage, C. Coherent optical orthogonal frequency division multiplexing. Electron. Lett. 2006, 42, 587–589. [CrossRef] International Telecommunication Union. Characteristics of a Single-Mode Optical Fiber and Cable. Available online: https://www.itu.int/rec/T-REC-G.652/en (accessed on 12 May 2012). Thorlabs. SMF-28-J9 Spec Sheet. Available online: https://www.thorlabs.com/thorproduct.cfm?partnumber= SMF-28-J9 (accessed on 31 May 2000). Thorlabs. DCF-38 Spec Sheet. Available online: https://www.thorlabs.com/thorproduct.cfm?partnumber= DCF38 (accessed on 23 February 2012). © 2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC-BY) license (http://creativecommons.org/licenses/by/4.0/).