Chapter 1 INTRODUCTION
1.5 Objective and Outlines of the Thesis
In this thesis, four novel RoF systems were proposed for transporting and generating wideband signals at 60 GHz, and the performances were investigated both theoretically and experimentally. The first architecture demonstrates the feasibility of the generation of an RF direct-detection vector signal using optical I/Q up-conversion. The second system demonstrates a short-range RoF system employing a single-electrode Mach-Zehnder modulator (MZM). Both systems transmit single-carrier and multi-carrier signal, and employ several digital signal processing techniques. The third and fourth systems present two simple hybrid access network architectures for generating and transmitting a 60GHz radio frequency (RF) phase-shift keying (PSK) signal with a baseband (BB) on-off keying (OOK) signal simultaneously.
This thesis is organized as the following. Chapter 1 provides the review of wireless systems and the advantages of 60 GHz technology. The challenges of the transmission of the 60 GHz signal are also discussed. Therefore, the 60 GHz RoF systems with digital signal processing attract great interests for future applications. Chapter 2 describes the basic ideas of 60 GHz RoF systems using external modulator. RoF system can be separated into an optical system and a wireless system. Optical system includes the optical transmitter, the optical channel and the optical receiver. The optical double-sideband (DSB) and
double-sideband with carrier suppression (DSB-CS) modulation schemes will be discussed. The properties of 60 GHz components and 60 GHz wireless channel will be discussed. The impairments of 60 GHz RoF systems will also been investigated in Chapter 2. The properties of digital signal modulation formats and digital signal processing will be discussed in Chapter 3. The signal carrier, OFDM, SC-FDE, single-carrier frequency-division multiple access (SC-FDMA), and single-carrier frequency-division multiplexing (SC-FDM) modulation formats will be discussed. The digital signal processing for system impairments and concept of multiple-input multiple-output (MIMO) will also been discussed.
A novel optical I/Q up-conversion RoF system for 60 GHz wireless applications will be proposed in Chapter 4. The advantage of the proposed transmitter is that no electrical mixer is needed to generate RF signals.
Therefore, I/Q data of RF signals are processed at baseband at the transmitter, which is independent of the carrier frequency of the generated RF signal.
Theoretical analysis and experimental demonstration of this system will be performed. The impacts of the I/Q imbalance will also been discussed. In order to achieve multi-standard operation, signal carrier and OFDM signals are utilized in the proposed system. The I/Q imbalance correction and adaptive loading algorithm are used to improve system performance. In Chapter 5, a simple 60GHz RoF system employing a single-electrode MZM are demonstrated. This system uses only one single-electrode MZM with bandwidth less than 35.5 GHz. The impacts of fiber chromatic dispersion and beat-noise on the performance of the RoF system are investigated by theoretical analysis and experimental demonstration. OFDM, SC-FDE and
SC-FDM signal generation and modulation techniques will be developed. 2 x 2 MIMO technologies also investigated to increase the data throughput within the 7 GHz band.
In Chapter 6, two novel multi-service hybrid access network systems for 60GHz wireless and wireline applications using frequency multiplication techniques will be presented. One of architecture uses a single-electrode MZM with frequency doubling technology. The other architecture uses a dual-parallel MZM with frequency quadrupling technology. These two schemes employ a novel pre-coded method that is based on the digital signal processing. The proposed systems does not suffer from RF fading and needs no narrow-band optical filter at the remote node to separate the RF and baseband signals. A frequency multiplication method for RoF link is realized to reduce the bandwidth requirement of the transmitter. Finally, Chapter 7 reviews the main conclusion of the thesis.
Chapter 2
RADIO-OVER-FIBER SYSTEMS USING EXTERNAL MODULATOR
2.1 The Architecture of 60 GHz Radio-over-Fiber Systems
There are two parts in radio-over-fiber (RoF) system: optical system and wireless system [38]. The optical system includes optical transmitter, optical channel, and optical receiver. The wireless system includes wireless transmitter, wireless channel, and wireless receiver. Optical transmitter converts an electrical input signal into the corresponding optical signal and then launches it into the optical fiber serving as a communication channel. Since the optical signal transfer in the optical fiber, the signal would suffer fiber dispersion that would induce frequency-selective fading. This phenomenon would introduce in 2.2.2. The role of an optical receiver is to convert the optical signal back into electrical form using photodiode. The generated electrical signal is at radio frequency and then launches it into the antenna serving as wireless transmitter.
For the wireless transmitter, the electrical amplifier amplified radio frequency signal and then broadcast the signal into the air by using antenna.
The signal transfers the energy using electro-magnetic waves without wires. At wireless receiver side, the radio frequency signal down-converted to lower frequency and demodulated the signal. In this chapter, we will do an introduction about the radio-over-fiber systems using external Mach-Zehnder Modulator (MZM), and investigate its impairments of theoretically and experimentally in 2.3.
2.2 Optical System
2.2.1 Optical Transmitter
The traditional optical transmitter concludes RF driving signal, optical source, and optical modulator. The RF driving signal with high spectral efficiency modulation is required to provide higher data-rate transmission because the bandwidth of a wireless channel is limited. Therefore, high order quadrature amplitude modulation (QAM) is a good candidate [41]. The corresponding system requires I/Q mixer to up-convert the in-phase (I) and quadrature phase (Q) signals.
Presently, most RoF systems are using laser as optical source. The advantages of laser are compact size, high efficiency, good reliability small emissive area compatible with fiber core dimensions, and possibility of direct modulation at relatively high frequency.
For the optical modulator, direct modulation and external modulation are two modulations of generated optical signal. When the bandwidth of direct modulation signal is above 10 GHz, the frequency chirp imposed on signal becomes large enough. Hence, it is difficult to apply direct modulation to generate microwave/mm-wave signal. However, the bandwidth of signal generated by external modulator can exceed 10 GHz easily. Presently, most RoF systems are using external modulation with MZM or electro-absorption modulator (EAM) [18-22]. The most commonly used MZM are based on LiNbO3 (lithium niobate) technology. According to number of electrode, there are two types of LiNbO3 device: dual-drive Mach-Zehnder modulator (DD-MZM) and single-drive Mach-Zehnder modulator (SD-MZM) [42, 43].
The SD-MZM has two arms and an electrode. The optical phase in each arm can be controlled by changing the voltage applied on the electrode. When the lightwaves are in phase, the modulator is in “on” state. On the other hand, when the lightwaves are in opposite phase, the modulator is in “off ” state, and the lightwave cannot propagate by waveguide for output.
The modulator is used for converting electrical signal into optical form.
Because the external dual-parallel modulator was composed of MZMs, we select MZM as modulator to build the architecture of optical transmitter. Figure 2-1 shows the schematic model of the MZM. The incident optical filed Ein is divided into two optical fields by optical coupler. The two optical fields send into two isolated paths, which are called upper arm and lower arm. The applied electrical signal modifies the velocities of the optical fields by the Pockels-Effect [43]. The velocity variation corresponds to the phase modulation of the optical field. The output optical filed for upper arm is
E E ∙ √a ∙ e∆ (2-1) Where a is the power splitting ratio of power coupler. ∆φ is the optical carrier phase difference that is induced by driving voltage for upper arm. The power splitting ratio is defined as the ratio of the transmission power to the coupling power. A coupler always has non-ideal power splitting ratios due to manufacturing limitations, e.g. a 1 2⁄ . Therefore, the output optical filed for lower arm is
E E ∙ √1 a ∙ e∆φ (2-2)
∆φ is the optical carrier phase difference that is induced by driving voltage for lower arm. The output optical filed for MZM is
E E ∙ √a ∙ √b ∙ e∆φ √1 a ∙ √1 b ∙ e∆φ (2-3)
b is the power splitting ratio of second power coupler in MZM. The power splitting ratio of two couplers of a balanced MZM is 0.5. The output optical filed of the balanced MZM is given by
E ∙ E ∙ e∆φ e∆φ (2-4) E E ∙ cos ∆ ∙ exp j∆φ ∆φ (2-5)
∆φ is the combined phase shift and can be calculated as the difference between
∆φ and ∆φ (∆φ ∆φ ∆φ ). For single electro x-cut MZM, the driving voltage for lower arm is the opposite of driving voltage for upper arm (∆φ ∆φ ) [44]. Therefore, the output optical field for single electro x-cut MZM can be simplify as
E E ∙ cos ∆ (2-6)
∆φ ≜ ∙ π (2-7) The v is called half-wave voltage that can induce π combined phase shift if the half-wave voltage is applied to the driving voltage [45]. If driving voltage equals to v , the modulator output power has its minimum value. The optical field E0 of the laser can be expressed by
E E cos t (2-8) Where E0 and denote the amplitude and angular frequency of input optical field, respectively. After add time component, the optical field will become as
E E ∙ cos ∆ ∙ cos t (2-9) The loss of MZM is neglected in previous equations. consisting of an electrical sinusoidal signal and a dc biased voltage can be written as,
cos t (2-10)
where is the dc biased voltage, and are the amplitude and the angular frequency of the electrical driving signal, respectively. The optical carrier phase difference induced by is given by
∆ ∙ ∙ π (2-11)
Equation (2-10) can be written as:
E E ∙ cos
π ∙ π ∙ cos ω E ∙ cos b m ∙ cos t ∙ cos ω E ∙ cos ω t ∙ cos b ∙ cos m ∙ cos ω t sin b ∙ sin m ∙ cos ω t (2-12) where ≜ π is a constant phase shift that is induced by the dc biased voltage, and ≜ π is the MZM modulation index (MI). The time dependent terms in Eq. (2-12) can be substituted with cos x cos θ and sin x cos θ . The cos x cos θ and sin x cos θ terms can be expanded applying Bessel functions. The expansion results can be summarized as follows [46]
cos x cos θ J x 2 1 J x cos 2nθ
∞
sin x cos θ 2 1 J x cos 2n 1 θ
∞
(2-13) Expanding Equation (2-12) using Bessel functions, as detailed in Equation (2-13). The optical field at the output of the MZM can be written as:
E E ∙ cos ∙
cos ∙ 2 ∙ 1 ∙ m ∙ cos 2
∞
sin ∙ 2 ∙ 1 ∙ ∙ cos 2 1
∞
(2-14) where is the Bessel function of the first kind of order n. the optical field of the mm-wave signal can be written as
E E ∙ cos ∙ ∙ cos
E ∙ cos ∙ ∙ cos
∞
2 π
E ∙ sin ∙ ∙ cos
∞
2 1
(2-15) When the MZM is biased at the maximum transmission point, the bias voltage is set at 0, and cosb = 1 and sinb = 0. Consequently, the optical field of the mm-wave driving signal can be written as
E E ∙ ∙ cos E ∙ ∙ cos
∞
2
(2-16) The amplitudes of the generated optical sidebands are proportional to those of the corresponding Bessel functions associated with the phase modulation index . With the amplitude of the electrical driving signal equal to , the is π 2⁄ . Due to the properties of Bessel function of the first kind, the value of would increase as order n decrease when 0
and 1. As shown in Fig. 2-2, , , , and are
0.5668, 0.2497, 0.069, and 0.014, respectively. Therefore, the optical sidebands with the Bessel function of first kind of high order term can be ignored, and Eq.
(2-16) can be further simplified to
E E ∙ ∙ cos E ∙ ∙ cos 2 (2-17)
When the MZM is biased at the middle point, the bias voltage is set at V , and cos b √ and sin b √ . Consequently, the optical field of the mm-wave signal using double sideband (DSB) can be written as
E √ ∙ E ∙ ∙ cos ω t
√ ∙ E ∙ ∙ cos (2-18) When the MZM is biased at the null point, the bias voltage is set at
V , and cos 0 and sin 1. Consequently, the optical field of the mm-wave signal using DSB with carrier suppression (DSBCS) modulation can be written as
E E ∙ ∙ cos (2-19) The generated optical spectrums of DSB and DSBCS signal are shown in Fig. 2-3 (a) and (b), respectively.
Figure 2-1 The principle diagram of the optical mm-wave generation using MZM.
0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 -0.25
0.00 0.25 0.50 0.75 1.00
Modulation Index
J1(m) J2(m) J3(m) J4(m)
Figure 2-2 The magnitude of Bessel functions versus different RF modulation
index.
Figure 2-3 The generated optical spectrum: (a) DSB signal; (b) DSBCS signal.
2.2.2 Optical Channel
Communication channel concludes fiber, optical amplifier, etc.. Presently, most RoF systems are using standard single-mode fiber or dispersion compensated fiber (DCF) as the transmission medium. When the optical signal transmits in optical fiber, dispersion will be happened. DCF is use to
compensate dispersion. The transmission distance of any fiber-optic communication system is eventually limited by fiber losses. For long-haul systems, the loss limitation has traditionally been overcome using regenerator witch the optical signal is first converted into an electric current and then regenerated using a transmitter. Such regenerators become quite complex and expensive for WDM lightwave systems. An alternative approach to loss management makes use of optical amplifiers, which amplify the optical signal directly without requiring its conversion to the electric domain. Presently, most RoF systems are using erbium-doped fiber amplifier (EDFA). An optical band-pass filter (OBPF) is necessary to filter out the amplified spontaneous emission (ASE) noise. The model of communication channel is shown in Fig.
2-4.
When optical RF signals are transmitted over a standard single-mode fiber with dispersion, a phase shift to each optical sideband relative to optical carrier is induced. The propagation constant of the dispersion fiber can be expressed as [47]
⋯ (2-20)
where is the derivative of the propagation constant evaluated at . The effect of high order fiber dispersion at 1550-nm band is neglected. For carrier tones with central frequency at , the eq.
(2-20) can be expressed as
≅ (2-21)
∙ D (2-22)
where c is light speed in free space and D is the chromatic dispersion parameter.
For a standard single-mode fiber, D is 17-ps/(nm.km) [43]. The fiber loss is ignored. Therefore, after transmission over a standard single-mode fiber of length z, the optical field for the DSB modulation scheme can be written as
E √ E ∙ cos
cos ∓ (2-23)
The optical field for the DSBCS modulation scheme can be written as
E E ∙ cos ∓ (2-24)
Figure 2-4 The model of optical channel in a RoF system. (OBPF: optical
bandpass filter; EDFA: erbium doped fiber amplifier.)
2.2.3 Optical Receiver
Optical receiver usually consists of the photodiode and the trans-impedance amplifier (TIA). The function of photodiode is to convert optical signal to electrical current. The function of TIA is to convert current to output voltage. The generated electrical signal is proportional to the square of the optical filed.
∙ (2-25)
where R is the responsivity of photodiode [47]. The optical field for the DSB modulation scheme will generate the electrical signal after photodiode. The photocurrent for the DSB modulation scheme without optical fiber transmission at different frequency can be expressed as
∙ ∙ 2 ∙
∙ ∙ ∙ ∙
∙ ∙ ∙ (2-26) For the RoF application, the frequency of desired mm-wave signal is ω . The reason is that the photocurrent at ω is proportional to driving signal v without significant distortion for small modulation index. Therefore, the DSB modulation scheme is one of RoF schemes for vector signal generation. In order to investigate the impairment of fiber dispersion, the he photocurrent with optical fiber transmission at ω can be expressed as
∙ ∙ ∙ ∙ ∙ cos (2-27) Due to fiber dispersion effect, the RF fading issue would be observed. The RF signal power is related to cos . Therefore, the RF fading issue would become serious when the magnitude of frequency become large.
As shown in the Fig. 2-5, when the frequency increases, the RF power will drop off rapidly. For 60-GHz applications, the frequency is fixed at 60.5 GHz. The first deep appears following 1-km fiber transmission.
The photocurrent for the DSBCS modulation scheme without optical fiber transmission at different frequency can be expressed as
∙ ∙
∙ ∙ ∙ (2-28) The frequency of desired RoF signal is 2 . DSBCS modulation scheme has effective in the millimeter-wave range with excellent spectral efficiency, a low bandwidth requirement for electrical components. However, the photocurrent at 2 is proportional to the square of driving signal. Therefore, DSBCS schemes can only support on-off keying (OOK) format, and none can transmit vector modulation formats, such as phase-shift keying (PSK), QAM, or OFDM signals, which are of utmost importance for wireless applications. In order to investigate the impairment of fiber dispersion, the photocurrent with optical fiber transmission at 2 is the same with the optical signal without optical fiber transmission. The reason is that the desired RoF signal comes from one copy of the optical signal at photodiode. In order to combine with wireless system, the electrical bandpass filter will be used to reject out-of-band signals in wireless transmitter.
0 2 4 6 8 10
Figure 2-5 Simulated RF power of the generated mm-wave signal versus standard single-mode fiber length.
2.3 Wireless System
2.3.1 Wireless Transmitter
Thanks to the RoF technology, the wireless transmitter does not need electrical mixer and local oscillator to up-convert signal to wanted carrier frequency. Therefore, the wireless transmitter only consists of electrical bandpass filter, electrical amplifier and, antennas, as shown in Fig. 2-6. Since the high transmit power is necessary to overcome the higher path loss at 60 GHz, FCC allocated the 60 GHz license-free band in US from 57 to 64 GHz with a maximum equivalent isotropically radiated power (EIRP) of 40 dBm average and 43 dBm peak [12]. However, the output power for 60 GHz amplifier is typically limited to 10 dBm because the implementation of efficient power amplifiers at 60 GHz is challenging. Thanks for the huge antenna gain can be used to increase EIRP. The huge antenna gain at 60 GHz has significantly boosted the allowable EIRP limits.
Figure 2-6 The model of wireless transmitter of RoF system.
2.3.2 Wireless channel
In order to improve the quality of 60GHz communications, understand the characteristics of 60GHz wireless channel is very important. The path loss is an important parameter for the wireless application. The Friis free space propagation formula could be express as
P P G G (2-29) where PR and PT are the transmitted and received power, respectively. λ is the signal wavelength, d is the transmission dictation between transmitter antenna and receiver antenna, GT and GR are transmit and receiver antenna gain, respectively. From this equation, the higher carrier frequency signals have lower wavelength and high path loss. Compared with signal propagation in 2.4 GHz, the path loss in 60 GHz is 20-30dB attenuation.
The other phenomenon of wireless communication is multipath fading.
The wireless signals at low frequency suffer serious frequency-selective fading because of the scattering effect which comes from objects that are roughness compared to the wavelength. Things change in 60 GHz signal, the signal wavelength is much smaller than low frequency signal. The reflection effect which comes from objects that are smooth compared to the wavelength.
Therefore, the 60-GHz signal does not suffer too much scattering effect, but reflection effect. Because of the reflection loss for the 60-GHz signal is about 10dB, the main propagation phenomenon for 60-GHz signal consists line-of-sight (LoS), first, and second order reflections [12].
2.3.3 Wireless Receiver
The wireless signal is received by receiver antenna. After receiver antenna,
RF bandpass filter is used to reject out-of-band signals. In order to keep the SNR of signal as high as possible, the in-band RF signals are then amplified by a low-noise amplifier (LNA). After the LNA, there have three methods for the RF front-end architecture [12].
In first method, the RF signal split into two signals, two RF signals are down-converted to baseband with two mixers and the quadrature local oscillators. Two baseband signals are transferred to digital signals by using analog to digital converters, as shown in Fig. 2-7 (a). Since this direct down-convert system is very challenging in 60 GHz circuit, the system has
In first method, the RF signal split into two signals, two RF signals are down-converted to baseband with two mixers and the quadrature local oscillators. Two baseband signals are transferred to digital signals by using analog to digital converters, as shown in Fig. 2-7 (a). Since this direct down-convert system is very challenging in 60 GHz circuit, the system has