Mitigation strategy for transmission impairments in millimeterwave radio-over-fiber networks [Invited]

Transcription

1 Vol. 8, No. 2 / February 2009 / JOURNAL OF OPTICAL NETWORKING 201 Mitigation strategy for transmission impairments in millimeterwave radio-over-fiber networks [Invited] Christina Lim, 1, * Ampalavanapillai Nirmalathas, 1,2 Masuduzzaman Bakaul, 2 Ka-Lun Lee, 1 Dalma Novak, 1,3 and Rod Waterhouse 1,3 1 ARC Special Research Centre for Ultra-Broadband Information Networks (CUBIN) 2 National ICT Australia, Victoria Research Laboratory, Department of Electrical and Electronic Engineering, The University of Melbourne, VIC 3010, Australia 3 Pharad LLC, Glen Burnie, Maryland, USA *Corresponding author: c.lim@ee.unimelb.edu.au Received September 30, 2008; revised November 26, 2008; accepted December 4, 2008; published January 27, 2009 Doc. ID Hybrid fiber wireless networks for fixed wireless access operating in the millimeter-wave (mm-wave) frequency region have been actively pursued to provide ultrahigh bandwidth for untethered connectivity. Moving the radio operating frequency into the mm-wave region overcomes the spectral congestion in the lower microwave region and is also capable of providing high-capacity broadband wireless services in a picocellular or microcellular architecture. Optical fiber backhaul provides the broadband interconnectivity between a centralized location and a large number of high-throughput antenna base stations necessary in such an architecture. The transportation of mm-wave wireless signals within the hybrid network is subject to numerous impairments ranging from low conversion efficiency to fiber chromatic dispersion and also to signal degradation due to nonlinearity along the link. One of the major technical challenges in implementing these networks lies in the mitigation of these impairments that the wireless signals experience while traversing the links. In this paper, we present an overview of the different techniques and schemes to overcome some of the impairments for transporting mm-wave signals over optical fibers Optical Society of America OCIS codes: , Introduction Fixed wireless access at millimeter-wave (mm-wave) or sub-millimeter-wave frequencies has the potential of providing users with seamless broadband connectivity [1 3]. With the inherent high-propagation-loss characteristics of radio signals at these frequencies, it is essential to deploy picocellular or microcellular architectures to provide efficient geographical coverage. To accommodate such an architecture, a large number of antenna base stations (BSs) have to be deployed to optimize the antenna coverage. It is therefore essential to simplify the antenna BS by moving all routing, switching, and processing functionalities to the head end, thereby enabling the cost and equipment to be shared among all the base stations. With the coverage zone of each antenna BS being greatly reduced while the throughput is significantly increased, an optical fiber backbone with its inherent low loss and large bandwidth characteristics has proven to be an ideal transport medium. Shown in Fig. 1 is a typical hybrid fiber wireless scenario where an optical head end or central office (CO) acts as the gateway to the optical metropolitan backbone while serving a large number of widely distributed antenna base stations and remote nodes (RNs). A strategy to achieve a centralized control architecture is to optically distribute the radio signals at mm-wave frequencies, thus reducing the complexity of the antenna BS by moving most of the hardware intelligence to the CO. However, the optical distribution of radio signals at mm-wave frequencies is susceptible to a number of impairments that may degrade the overall system performance. These impairments include low optical-to-electrical conversion efficiency, inefficient optical spectral usage, fiber chromatic dispersion, and nonlinearity along the link. It is well established that the mm-wave radio signals are typically weakly modulated onto the optical carrier /09/ /$ Optical Society of America

2 Vol. 8, No. 2 / February 2009 / JOURNAL OF OPTICAL NETWORKING 202 Fig. 1. Hybrid fiber wireless networks. due to the low modulation efficiency at these high radio frequencies. To improve the efficiency, the mm-wave radio signals can be electrically amplified before modulating the optical source. However, this may lead to increased intermodulation distortion (IMD) products at the optical front end, which limits the overall system dynamic range [4], and this condition may further worsen due to other fiber nonlinearities as the radio signals propagate through the fiber links [5]. A number of linearization techniques have been introduced to improve the optical front-end linearity, including optical feed forward [6,7], gain modulation [8], predistortion [9,10], and parallel modulator configurations [11 13]. Alternatively, the link performance can also be improved by increasing the optical power of the mm-wave modulated optical signals via a high-power optical source or optical amplifier. This, on the other hand, may increase the intermodulation distortions at the receiver or even damage the receiver due to too large an optical power incident on the optical detector [14]. A few techniques have been proposed to improve the low mm-wave modulation efficiency, including Brillouin scattering [14,15] and external optical filtering [16 18]. Apart from low modulation efficiency, the optical distribution of mm-wave radio signals is also susceptible to fiber chromatic dispersion that severely limits the transmission distance [19,20]. This shortcoming can be mitigated using an optical singlesideband-with-carrier OSSB+C modulation scheme [21], an optical carrier suppression modulation technique [22 24], optical filtering [25 28], a chirped fiber grating [29], or specialized modulators [30,31]. In this paper, we review the different schemes and strategies to mitigate all these impairments for the realization of an efficient mm-wave fiber wireless network. 2. Optical Impairments in Fiber Wireless Links Figure 2 shows a simple point-to-point mm-wave radio-over-fiber link connecting a remote antenna base station and a central office that quantifies the different signal impairments experienced by the mm-wave wireless signals as they propagate in the Fig. 2. Optical impairments in the mm-wave fiber wireless links.

3 Vol. 8, No. 2 / February 2009 / JOURNAL OF OPTICAL NETWORKING 203 hybrid link. When the mm-wave wireless signals are received at the remote antenna base station, the signals are first amplified and converted to optical signals. The conversion can be carried out either by direct modulation using an optical source or by external modulation using an external modulator in conjunction with an optical carrier. The mm-wave radio signal modulated onto the optical carrier is typically very weak as a result of low modulation efficiency at these frequencies. The conversion efficiency in this context refers to optical-electrical-optical conversion. In addition, the nonlinear characteristics of the optical intensity modulator limit the amplitude of the electrical modulation to a very narrow window for linear operation. Once the mm-wave radio signals are modulated onto the optical carrier, the optical signal will be transported over the optical fiber link to the central office. The optical distribution of the mm-wave radio signals is subject to the effects of fiber chromatic dispersion that will severely limit the overall transmission distance [19]. In addition, the optical spectral usage for the distribution of the mm-wave radio signal is highly inefficient, considering that the amount of useful information that is being transported 3 Gbits/s is only a fraction of the occupied spectrum 40 GHz. In addition, in a long-reach environment, the optical signal may experience fiber nonlinearities if the optical signal power is required to be amplified in order to overcome the link losses while the amplified optical power is also sufficiently large to trigger the nonlinear fiber effects. Another phenomenon that the signal may experience in a long-reach environment is phase decorrelation between the optical carrier and radio signal, which may introduce an additional power penalty [32,33]. Upon reception at the receiver, the optical signal undergoes an optical-to-electrical conversion using a photodiode. The photodiode is also a nonlinear device and is governed by the square-law process. Thus, the detection process will further introduce distortions into the system. Therefore it is of great importance that these various impairments be mitigated to improve the signal quality and performance of mm-wave hybrid fiber wireless links. 3. Strategies to Overcome Impairments In this section, we review the different mitigation strategies and techniques to overcome the various impairments described in Section 2. 3.A. Optical Fiber Dispersion When the mm-wave radio signals are modulated onto an optical carrier, the modulated signal will be in a double-sideband-with-carrier format where the sidebands are located at the mm-wave signals away and on either side of the optical carrier as illustrated in Fig. 3. When the mm-wave modulated optical signal propagates along a dispersive fiber, the two sidebands will experience different amounts of phase shift relative to the optical carrier. Upon detection at the photodetector, the square-law process generates two beat components at the desired mm-wave frequency. The received RF power of the mm-wave signal varies depending on the relative phase difference Fig. 3. Schematic showing double-sideband (DSB) mm-wave modulated optical signal transport. Inset shows measured and calculated normalized RF power as a function of f mm for L=80 km.

4 Vol. 8, No. 2 / February 2009 / JOURNAL OF OPTICAL NETWORKING 204 between the two beat components. The RF power variation is dependent on the fiber dispersion parameter, the transmission distance, and the mm-wave frequency as governed by the following equation [19]: 2 cld P RF cos f mm 2, 1 where D represents the fiber dispersion parameter in ps/nm/km, c is the velocity of light in a vacuum, L is the fiber transmission length, f mm represents the mm-wave modulating frequency, and f 0 is the optical carrier center frequency. To quantify the severity of the penalty, shown in the inset of Fig. 3 is the calculated normalized received RF power using Eq. (1) as a function of modulating frequencies for a fiber transmission distance of 80 km. Normalized RF power is defined as the ratio of detected RF power at 80 km to the RF power calculated at 0 km of fiber. From the results it can be seen that the RF power varies in a periodic manner with complete power suppression occurring at certain modulating frequencies. 3.A.1. Dispersion Effect Mitigation Techniques Since the fiber-induced dispersion penalties are so severe in direct-detection optically fed mm-wave systems, various techniques have been devised and proposed to overcome the dispersion effects in such systems. Among these techniques are the OSSB +C modulation scheme [21], the optical carrier suppression technique [22 24], external filtering [25 28], chirped fiber gratings [29], fiber nonlinearities [34 36], and phase conjugation [37]. A convenient technique to overcome the fiber dispersion effect is removing one of the optical sidebands in Fig. 3. This can be done via optical filtering using a narrowband notch fiber Bragg grating where the reflective band coincides with the unwanted sideband [25]. Although this technique is simple to implement, the limited flexibility makes the implementation difficult to accommodate for modifying the mm-wave frequency. Another technique to generate OSSB+C is via cancellation of the unwanted optical sideband within an external optical modulator. This can be done using a dualelectrode Mach Zehnder modulator (DEMZM) biased at quadrature and with the RF signal applied to both electrodes with a 90 phase shift between the two electrodes [21] as illustrated in Fig 4. The interaction between the RF modulation and the optical signals results in the suppression of one of the odd-harmonics modulation sidebands. Also shown in Fig. 4 is the measured optical spectrum at the output of the DEMZM for a modulating frequency of GHz clearly showing the OSSB+C signal format and also the measured normalized RF power for OSSB+C as a function of modulating frequency for L=80 km. The results indicate that the OSSB+C format is able to overcome the RF power fading due to fiber chromatic dispersion. Both the opti- f 0 Fig. 4. OSSB+C signal generation using DEMZM. Inset shows the measured optical spectrum at the output of the DEMZM and the measured normalized RF power for OSSB+C for L =80 km.

5 Vol. 8, No. 2 / February 2009 / JOURNAL OF OPTICAL NETWORKING 205 cal filtering scheme and OSSB+C techniques suffer a 6 db electrical loss since half of the optical sideband power is removed compared with the optical double-sideband case. Another approach to generating OSSB+C modulation is via two electroabsorption modulators (EAMs) [38] where the optical signal is first split into two paths with a 90 phase shift. These signals are input into the two EAMs, which are driven by the same electrical modulating signal with one delayed by a quarter of the signal period. The optical carrier suppression scheme is another effective method to combat dispersion effects in mm-wave fiber wireless links [22 24]. By biasing a single-electrode Mach Zehdner modulator (MZM) at the minimum transmission point of the transfer function, the optical carrier will be suppressed and a double-sideband-suppressedcarrier optical signal will be generated. Such an implementation requires only half the desired modulating frequency to drive the MZM. The mixing of the two optical carriers in a high-speed photodetector generates a single beat component at twice the drive frequency, which is not affected by dispersion-induced RF penalties. Despite the simple and elegant approach, this technique requires a large RF drive power to obtain a desirable modulation depth since the modulator is driven in the nonlinear region. In addition to the external modulation techniques, tapered linearly chirped fiber gratings have also been used to counter dispersion in mm-wave fiber links [29]. A specially designed tapered linearly chirped fiber grating is inserted at the output of the fiber via an optical coupler to overcome the dispersion, and the reflected light is then directed to an optical photodetector for detection [29]. Other proposed methods for mitigation of dispersion effects include phase conjugation [37], chirped EAM [30,31], and fiber nonlinearity [34 36]. The latter technique uses the nonlinear property of the fiber to combat dispersion in mm-wave fiber links. The factor limiting the nonlinearity technique is stimulated Brillouin scattering (SBS), which requires the optical power to be carefully tailored to ensure that it is just sufficient to overcome the effects of dispersion while ensuring that the SBS is at a negligible level. Another effective scheme for reducing fiber dispersion effects in mm-wave fiber radio systems is to transmit the radio signals over fiber at a lower intermediate frequency (IF) rather than at mm-wave frequencies with remote upconversion performed at the base station [39]. Recently the possibility of digitization of the radio signals using a bandpass sampling technique has also been shown [40]. 3.A.2. Other Dispersion-Induced Penalties Subsection 3.A.1 described the different strategies to reduce dispersion effects that cause degradation in the RF power of the recovered RF signals. Although the optical suppression technique described in the previous subsection is not affected by dispersion-induced RF power penalty, it has been shown that data imposed on the two sidebands experience bit walk-off and are still subject to a dispersion penalty [41,42]. The effect of dispersion on the data is not as severe as the intensity-modulated mm-wave signal; however, the penalty still limits the distance of data transportation over fiber [41,42]. To cope with this effect, a delay-line filter can be used to separate one of the two optical modes, and information is imposed on this mode via a separate modulator, before the modulated sideband is combined with the unmodulated one [22]. However, such an implementation increases the complexity of the technique and imposes stringent requirements for the filter designs to match the frequency separation and the coherence time of the optical sideband [43]. For the OSSB+C modulation scheme, data is only imposed on the sideband. In this scenario, the OSSB+C technique is not limited by bit walk-off [44]. However, each finite linewidth optical mode in the OSSB+C signal experiences different propagation delays due to dispersion as it propagates along the fiber. Consequently this results in a state of partial phase decorrelation or complete phase correlation upon detection at the photodetector [32,33]. The amount of phase decorrelation increases the phase noise of the remotely generated RF signals, which in turn affects the system performance. In general, phase-shift-keyed (PSK) modulation with higher-order modulation formats are more susceptible to the induced phase noise. It has been demonstrated that for long-reach transmission where optical amplifiers are required, the maximum transmission distance for the transmission of an OSSB+C signal at 28 GHz is limited by phase noise to 90 km [45]. Here the phase noise contributions are from the phase

6 Vol. 8, No. 2 / February 2009 / JOURNAL OF OPTICAL NETWORKING 206 decorrelation between the optical carrier and the sideband and also from the amplified spontaneous emission (ASE) noise from the amplifier. Dispersion-induced phase noise can be compensated by introducing a path imbalance for the two optical signals generated at the transmitter before they are recombined for transmission [32]. The influence of the path imbalance has been successfully quantified in an optical heterodyne system with injection locking [46]. It has been shown that the path imbalance is able to counteract the dispersion-induced phase noise; however, this requires careful design [46]. Essentially the introduction of the path imbalance offers a predistortion feature in the dual-frequency generator that is able to eliminate the dispersion-induced phase noise for a fixed fiber length and mm-wave frequency with carefully tailored delays. Despite the different schemes, these techniques have their unique advantages and disadvantages, which are briefly summarized in Table 1. 3.B. Optical Spectral Efficiency In addition to helping mitigate RF power penalties, the OSSB+C modulation scheme improves the optical spectral usage by at least 50%. Nevertheless the transportation of OSSB+C modulated radio signals at mm-wave frequencies still leads to the inefficient use of the optical spectrum especially in a wavelength-division-multiplexed (WDM) environment where the actual information bandwidth of the radio signals modulated onto an optical WDM channel (typically at 50 or 100 GHz spacing) is 2 GHz. Therefore the optical transport of mm-wave modulated optical signals leads to inefficient use of optical bandwidth. Recently there have been a number of proposed schemes to improve the optical spectral usage for the transportation of optically modulated mm-wave signals [47 49]. These techniques are based on interleaving multiple mm-wave optical signals, making use of the unused spectral band between the optical carrier and sideband in a OSSB+C signal (as illustrated in Fig. 5) or between the two sidebands in an optical carrier suppression technique. In principle, wavelength interleaving is able to enhance the overall capacity within the standard 1550 nm erbium-doped fiber amplifier (EDFA) gain window by 3 times [50] for a 37.5 GHz fiber wireless link. The successful implementation of wavelength interleaving depends largely on effective multiplexing and demultiplexing (MUX/DEMUX) schemes. There have been a number of different MUX/DEMUX architectures demonstrated to support the wavelength-interleaving scheme using fiber Bragg gratings (FBGs) [51,52] and arrayed waveguide gratings (AWGs) [53 55]. Shown in Figs. 6(a) and 6(b) are the different schemes based on AWGs. Figure 6(a) shows an interface based on a 2N+2 2N+2 AWG in conjunction with optical circulators and isolators that is capable of simultaneous multiplexing and demultiplexing of N wavelength-interleaved mm-wave optical signals [53]. This scheme was demonstrated using a 12.5 GHz 8-port AWG for the multiplexing and demultiplexing of 37.5 GHz optical signals. Here the incoming N interleaved channels from the downlink direction (DL) are split by a3dbcoupler before entering the AWG via port A 1 and port A 4 The input ports A 1 and A 4 were cho- Table 1. Advantages and Disadvantages of Different Modulation Schemes Optical Millimeter-Wave Modulation Schemes Advantages Disadvantages Intensity modulation (double-sideband with carrier format) Optical single-sideband with carrier OSSB+C format Optical carrier suppression technique Simple to generate Overcomes dispersioninduced RF power fading Minimal bit walk-off impact Overcomes dispersioninduced RF power fading Uses half the required LO frequency Performance limited by dispersion-induced RF power penalty Low receiver sensitivity Performance may be limited by phase decorrelation in a longreach environment Suffers from bit walk-off for longer transmission distance Requires large RF drive power

7 Vol. 8, No. 2 / February 2009 / JOURNAL OF OPTICAL NETWORKING 207 Fig. 5. Schematic showing wavelength interleaving technique. sen in such a way that the optical carriers C D1,C D2,...,C DN and the corresponding sidebands S D1,S D2,...,S DN are demultiplexed together and exit the AWG via the odd-numbered output ports B 1,B 3,..., B 2N 1. The optical circulators used in the setup are to isolate the downlink and uplink (UL) signals. In the UL direction, the multiple OSSB+C channels enter the AWG via the oddnumbered ports B 1,B 3,..., B 2N 1 as shown in Fig. 6(a). Due to the reciprocal and cyclic characteristics of the AWG, the UL carriers and their corresponding sidebands S U1,S U2,...,S UN combine at ports A 4 and A 1, respectively. The composite UL carriers C U1,C U2,...,C CN at port A 4 are looped back to the AWG via port B 2 which routes the optical carriers to the odd-numbered A 3,A 5,...,A 2N 1, starting from A 3. To establish the multiplexing and interleaving for the UL channels, the distributed UL carriers C U1,C U2,...,C CN are again looped back to the AWG via the evennumbered ports B 4,B 6,...,B 2N, starting from port B 4, and the resulting outcome combines all the UL optical carriers and exits the AWG via port A 1. This establishes the multiplexing functionality. Toda et al. [54] have proposed another demultiplexing scheme that is also based on an AWG. The demultiplexer is based on a 2 N AWG and a high-finesse Fabry Perot etalon as shown in Fig. 6(b) [54]. The free spectral range of the etalon is tuned to match the channel spacing of the mm-wave optical channels. The etalon has the function of separating the optical carriers from the sidebands of the mm-wave modulated optical signals. All the optical carriers enter one port of the AWG while all the sidebands enter the other. The AWG routes the optical carriers and the corresponding sidebands to the same output port, thus accomplishing the demultiplexing functionality. The multiplexing functionality can be accomplished using the same technique but in the reverse order with the optical circulator operating in the reverse direction [54]. 3.C. Improving Optical Modulation Depth for Transporting Millimeter-Wave Signals It is well known that mm-wave radio signals are typically weakly modulated onto the optical carrier due to the narrow linear region of intensity modulators. Consequently Fig. 6. (a) MUX/DEMUX using 2N+2 2N+2 AWG and (b) DEMUX using 2 N AWG for wavelength-interleaved channels.

8 Vol. 8, No. 2 / February 2009 / JOURNAL OF OPTICAL NETWORKING 208 the power of the optically modulated mm-wave sideband can be more than 20 db below that of the optical carrier for an OSSB+C signal. To improve the link performance, the optical power of the signals can be increased by using a high-power optical source or an optical amplifier; however, this may lead to increased IMD at the receiver or even damage of the receiver due to too large an optical power incident upon the optical detector [14]. A few techniques have been proposed to improve the modulation efficiency of these signals including Brillouin scattering [14,15], external optical filtering [16,17], and optical attenuation [18]. Shown in Fig. 7(a) is a technique to improve the modulation efficiency of OSSB+C signals using an optical filtering scheme. Here a narrowband external fiber Bragg grating is used to remove a portion of the optical carrier, leaving only a fraction of the optical carrier power to be detected at the receiver [56]. In this particular investigation, a number of FBGs with 3 db reflection bandwidths of 2.7 GHz and reflectivity ranging from 3 db (50%) to 30 db (99.9%) were used to quantify the optical link performance as a function of modulation efficiency. Figure 7(b) shows the optical spectrum of an OSSB+C signal carrying 155 Mbits/s data at 35 GHz before and after the FBG with 95% reflectivity that clearly indicated that the optical carrier was suppressed by 14 db. The corresponding bit-error-rate (BER) curves are shown in Fig. 7(b) with a 4.25 db improvement in the sensitivity at a BER of 10 9 with the carrierto-sideband ratio (CSR) decreased by 14 db [57]. Therefore there exists an optimum receiver sensitivity for OSSB+C, which corresponds to an optical carrier-to-sidebandratio of 0 db. 3.D. Base-Station Technologies To support full-duplex operation in a mm-wave fiber wireless network, the optical interfaces within the antenna base station must include optical sources that can be modulated by the mm-wave uplink radio signals. In addition, optical sources with narrow linewidth at well-specified wavelengths are required at the base stations to minimize phase noise degradation. However, this scenario is not an attractive option for uplink signal transmission because ultrastable, low-cost, narrow-linewidth optical sources are difficult to realize. Therefore there is a significant advantage to completely removing the need for an optical source in the antenna base station. This moves the wavelength assignment and source monitoring functionalities to the central office, which relaxes the stringent requirements of the antenna base stations. Fig. 7. (a) Technique to improve modulation efficiency using a FBG. (b) Measured optical spectra and BER using a 95% reflectivity FBG.

9 Vol. 8, No. 2 / February 2009 / JOURNAL OF OPTICAL NETWORKING 209 The first demonstration of a source-free antenna base station was by British Telecom s passive pico cell concept [58] where an EAM was used as both the detector and modulator by careful choice of the biasing condition [58]. In this technique, the EAM was optimized independently for two different optical signals to be used as the downlink and uplink carriers. This scheme was further improved and optimized for 60 GHz transmission where the EAM was termed as an electroabsorption transceiver (EAT) [59,60]. A more convenient method to establish a source-free base station is by providing the uplink optical carrier remotely from the central office [61,62]. In this approach, the central office will generate the downlink signals and also provide optical carriers at different wavelengths for the uplink transmission. Another source-free scheme that has been proposed and demonstrated is called the wavelength reuse technique where a portion of the downlink carrier is extracted and reused for uplink transmission [63]. Shown in Fig. 8 is the schematic of the wavelength reuse technique for OSSB+C modulated signals where an optical carrier recovery interface is located within the base station. The optical carrier recovery interface consisting of a 3-port circulator and a narrowband FBG with 50% reflectivity is shown in the interface. As can be seen from Fig. 8, the incoming downlink OSSB+C signal enters the optical carrier recovery interface via port 1 of the circulator and where 50% of the optical carrier power is reflected by a FBG with a center wavelength at the optical carrier, which is located at the output of port 2. The remaining 50% of the carrier and the corresponding sideband feed a photodetector and the detected downlink signal enters the base station downlink RF interface for wireless transmission. The reflected optical carrier exits the optical interface via port 3, where it will be reused as the uplink optical carrier. Knowledge of the operating wavelength is sufficient for the design of the FBG, making it more flexible in terms of frequency assignment at the base station air interface. To demonstrate how to develop a simplified BS architecture with improved performance, a full-duplex demonstration as shown in Fig. 9 was implemented in [57]. The transmission was over 20 km of single-mode fiber (SMF) for both downlink and uplink transmission for OSSB+C signals at 35 GHz with each carrying 155 Mbits/ s binaryphase-shift-keyed (BPSK) data. The carrier-to-sideband ratio of the OSSB+C downlink signal was 20 db and was amplified with an EDFA to 5 mw. The signal after 20 km of fiber was filtered using a 98.5% FBG in conjunction with a 3-port circulator. The extracted optical carrier was remodulated with the upstream 155 Mbits/ s BPSK wireless data at 35 GHz in an OSSB+C format before being transmitted over another 20 km of fiber to the central office. The remaining optical carrier and the corresponding sideband were detected using a photodetector in the downlink direction. Figure 10 shows the optical spectra of the downlink signal before [Fig. 10(a)] and after [Fig. Fig. 8. Wavelength reuse scheme for OSSB+C signals in the base station.

10 Vol. 8, No. 2 / February 2009 / JOURNAL OF OPTICAL NETWORKING 210 Fig. 9. Full-duplex transmission for 35 GHz signal in OSSB+C format over 20 km of fiber with improved modulation efficiency and incorporating the wavelength reuse scheme in the BS. 10(b)] the FBG and the uplink signal before [Fig. 10(c)] and after [Fig. 10(d)] remodulation. The downlink carrier-to-sideband ratio was reduced from 20 to 2 db, and errorfree reception for both uplinks and downlinks was achieved. Shown in Fig. 10(e) are the measured bit error rates of the full-duplex transmission over 20 km of fiber. This demonstration illustrates the combined techniques of improving the modulation efficiency and reusing the optical carrier to achieve a source-free base station. 3.E. Front-End Nonlinearity In a wireless access network with a multicarrier environment, linearity plays an important role in the improvement of the system dynamic range. It has been shown that the nonlinearity of the optical front end in a fiber distributed wireless network limits the overall system dynamic range [4] and this condition worsens due to other fiber nonlinearities as the radio signals propagate through the fiber link [5]. The issue of linearity in fiber wireless links has been widely investigated. A number of linearization techniques have been demonstrated to combat IMD products and improve the dynamic range of optical analog links, including optical feed forward [6,7], gain modulation [8], predistortion [9,10], and parallel modulator configurations [11 13]. Fig. 10. Measured optical spectra of downlink signals (a) before FBG, (b) after FBG and uplink signals, (c) before modulation, and (d) after modulation. (e) Measured BER curves for downlink and uplink after 20 km of SMF.

11 Vol. 8, No. 2 / February 2009 / JOURNAL OF OPTICAL NETWORKING 211 Fig. 11. Optical components generated at the output of DEMZM under a two-tone test. The optical feed-forward technique has been shown to be effective in suppressing third-order IMD and also in reducing laser relative intensity noise (RIN) over a wide bandwidth [64]. This technique consists of two sections with one determining the error and the other canceling the error [6,7]. Therefore it relies on the careful tuning of the amplitudes and phases of the signals to ensure perfect cancellation. Another technique to combat IMD is predistortion. The predistortion technique has also been explored for fiber wireless applications [9,10]. In general the predistortion technique requires a predistorter at the source that combats IMD by generating frequency components of the same amplitude but opposite in phase. To date there has been limited work undertaken to cancel IMD for OSSB+C modulated optical signals in mm-wave fiber radio applications. Recently a linearization technique has been proposed based on the removal of the dominant sources of IMD in the optical domain [65]. To quantify the IMD contributions from the different optical components, a DEMZM is driven by a two-tone signal at frequency 1 and 2. Due to the nonlinear characteristics of the DEMZM, other optical components are also generated apart from the desired two-tone signal as shown in Fig. 11. When these optical components are detected using a photodetector, other RF components are also generated, and out of these, only the third-order IMD that falls within the band is of interest. Via this quantification, it was identified that the optical components at c and c contribute most to the third-order IMD in the RF domain [65]. Therefore, to improve the carrier-to-interference ratio performance of OSSB+C modulated signals, these optical components were removed based on a scheme shown in Fig. 12. The optical carrier is first split into two paths, with one driving the DEMZM to generate the OSSB+C signal. A narrowband filter is used at the output of the DEMZM to remove the optical carrier, together with any components that lie within the vicinity of the carrier, i.e., c and c The output of the filter is then recombined with the clean optical carrier from the bottom arm before photodetection. Here the amplitude and the phase of the clean optical carrier are adjusted to optimize the components at f 1 and f 2. The proposed scheme does not completely remove all the third-order IMD in the RF domain but rather a large portion of it as quantified in [65]. Figures 13(a) and 13(b) show the measured RF spectra of a threetone test (27.5, , and GHz) without and with the linearization technique Fig. 12. Schematic of technique to reduce IMD for OSSB+C signals.

12 Vol. 8, No. 2 / February 2009 / JOURNAL OF OPTICAL NETWORKING 212 Fig. 13. Measured RF spectra (a) without linearization and (b) with linearization for a threetone test. as a proof-of-concept experiment [65]. It is evident from Fig. 13(b) that, when the linearization scheme was applied, the third-order IMD was suppressed with a maximum suppression of 9 db achieved [65]. 4. Conclusions This paper has provided a comprehensive summary and review of the research carried out in the area of mm-wave fiber wireless systems, in particular looking at the techniques and strategies to overcome the various optical impairments that exist within the link. The main focus of this research has been directed at the realization of highly efficient and optimized fiber wireless links. The introduction of OSSB+C and optical carrier suppression techniques to overcome issues associated with fiber chromatic dispersion, wavelength interleaving to improve the optical spectral usage, and the wavelength reuse technique to simplify the BS architecture into the mm-wave fiber wireless network have all further enhanced the overall link performance and efficiency. Various strategies have been discussed to improve the link linearity and also improve the modulation efficiency in such hybrid transmission systems. Acknowledgment The authors thank Marik Attygalle of the Defence Science and Technology Organisation (DSTO), Australia, for part of the work published in this manuscript. References 1. A. Seeds and K. Williams, Microwave photonics, J. Lightwave Technol. 24, (2006). 2. J. Capmany and D. Novak, Microwave photonics combines two worlds, Nat. Photonics 1, (2007). 3. A. Nirmalathas, C. Lim, D. Novak, and R. B. Waterhouse, Progress in millimeter-wave fiber-radio access networks, in Millimeter Waves in Communication Systems, Innovative Technology Series Information Systems and Networks, M. Ney, ed. (Hermes Penton Science Ltd., 2003), pp T. Kurniawan, A. Nirmalathas, C. Lim, D. Novak, and R. Waterhouse, Performance analysis of optimized millimeter-wave fiber radio links, IEEE Trans. Microwave Theory Tech. 54, (2006). 5. W. H. Chen and W. I. Way, Multichannel single-sideband SCM/DWDM transmission systems, J. Lightwave Technol. 7, (2004). 6. S.-H. Park and Y.-W. Choi, Significant suppression of the third intermodulation distortion in transmission system with optical feedback forward linearized transmitter, IEEE Photon. Technol. Lett. 17, (2005). 7. T. Ismail, C.-P. Liu, J. E. Mitchell, and A. J. Seeds, High-dynamic-range wireless-over-fiber link using feedforward linearization, J. Lightwave Technol. 25, (2007). 8. S. H. Lee, J. M. Kang, Y. Y. Won, H. C. Kwon, and S. K. Han, Linearization of RoF optical source by using light-injected gain modulation, in International Topical Meeting on Microwave Photonics (IEEE, 2005), pp L. Roselli, V. Borgioni, F. Zepparelli, F. Ambrosi, M. Comez, P. Faccin, and A. Cassini, Analog laser predistortion for multiservice radio-over-fiber systems, J. Lightwave Technol. 21, (2003). 10. A. R. Shah and B. Jalali, Adaptive equalisation for broadband predistortion linearisation of optical transmitters, IEE Proc.: Optoelectron. 152, (2005). 11. A. Djupsjobacka, A linearization concept for integrated-optic modulators, IEEE Photon. Technol. Lett. 4, (1992). 12. H. Skeie and R. Johnson, Linearization of electro-optic modulators by a cascade coupling of phase modulating electrodes, Proc. SPIE 1583, (1991).

13 Vol. 8, No. 2 / February 2009 / JOURNAL OF OPTICAL NETWORKING S. Korotky and R. DeRidder, Dual parallel modulation schemes for low-distortion analog optical transmission, IEEE J. Sel. Areas Commun. 8, (1990). 14. K. J. Williams and R. D. Esman, Stimulated Brillouin scattering for improvement of microwave fibre-optic link efficiency, Electron. Lett. 30, (1994). 15. S. Tonda-Goldstein, D. Dolfi, J.-P. Huignard, G. Charlet, and J. Chazelas, Stimulated brillouin scattering for microwave signal modulation depth increase in optical links, Electron. Lett. 36, (2000). 16. M. J. LaGasse, W. Charczenko, M. C. Hamilton, and S. Thaniyavarn, Optical carrier filtering for high dynamic range fibre optic links, Electron. Lett. 30, (1994). 17. R. D. Esman and K. J. Williams, Wideband efficiency improvement of fiber optic systems by carrier substraction, IEEE Photon. Technol. Lett. 7, (1995). 18. H. Toda, T. Yamashita, T. Kuri, and K. Kitayama, 25-GHz channel spacing DWDM multiplexing using an arrayed waveguide grating for 60-GHz band radio-on-fiber systems, in International Topical Meeting on Microwave Photonics (IEEE, 2003), pp G. J. Meslener, Chromatic dispersion induced distortion of modulated monochromatic light employing direct detection, IEEE J. Quantum Electron. QE-20, (1984). 20. H. Schmuck, Comparison of optical millimeter-wave system concepts with regard to chromatic dispersion, Electron. Lett. 31, (1995). 21. G. H. Smith, D. Novak, and Z. Ahmed, Technique for optical SSB generation to overcome dispersion penalties in fibre-radio systems, Electron. Lett. 33, (1997). 22. R. A. Griffin, P. M. Lane, and J. J. O Reilly, Dispersion-tolerant subcarrier data modulation of optical millimetre-wave signals, Electron. Lett. 32, (1996). 23. J. M. Fuster, J. Marti, and J. L. Corral, Chromatic dispersion effects in electro-optical upconverted millimetre-wave fibre optic links, Electron. Lett. 33, (1997). 24. Z. Jia, J. Yu, and G.-K. Chang, A full-duplex radio-over-fiber system based on optical carrier suppression and reuse, IEEE Photon. Technol. Lett. 18, (2006). 25. J. Park, W. V. Sorin, and K. Y. Lau, Elimination of the fibre chromatic dispersion penalty on 1550 nm millimetre-wave optical transmission, Electron. Lett. 33, (1997). 26. K. Yonenaga and N. Takachio, A fiber chromatic dispersion compensation technique with an optical SSB transmission in optical homodyne detection systems, IEEE Photon. Technol. Lett. 5, (1993). 27. E. Vourch, D. Le Berre, and D. Herve, Lightwave single sideband wavelength self-tunable filter using InP:Fe crystal for fiber-wireless systems, IEEE Photon. Technol. Lett. 14, (2002). 28. J. Capmany, D. Pastor, P. Munoz, S. Sales, B. Ortega, and A. Martinez, WDM-SSB generation and dispersion mitigation in radio over fiber systems with improved performance using an AWG multiplexer with flat top resonances, in International Topical Meeting on Microwave Photonics (IEEE, 2003), pp J. Marti, J. M. Fuster, and R. I. Laming, Experimental reduction of chromatic dispersion effects in lightwave microwave/millimetre-wave transmissions using tapered linearly chirped fibre gratings, Electron. Lett. 33, (1997). 30. V. Polo, J. Marti, F. Ramos, and D. Moodie, Mitigation of chromatic dispersion effects employing electroabsorption modulator-based transmitters, IEEE Photon. Technol. Lett. 7, (1999). 31. J. A. J. Fells, M. A. Gibbon, I. H. White, G. H. B. Thompson, R. V. Penty, C. J. Armistead, E. M. Kimber, D. J. Moule, and E. J. Thrush, Transmission beyond the dispersion limit using a negative chirp electroabsorption modulator, Electron. Lett. 30, (1994). 32. U. Gliese, S. Norskov, and T. N. Nielsen, Chromatic dispersion in fiber-optic microwave and millimeter-wave links, IEEE Trans. Microwave Theory Tech. 44, (1996). 33. R. Hofstetter, H. Schmuck, and R. Heidemann, Dispersion effects in optical millimeterwave systems using self-heterodyne method for transport and generation, IEEE Trans. Microwave Theory Tech. 43, (1995). 34. F. Ramos, J. Marti, V. Polo, and J. M. Fuster, On the use of fiber-induced self-phase modulation to reduce chromatic dispersion effects in microwave/millimeter-wave optical systems, IEEE Photon. Technol. Lett. 10, (1998). 35. F. Ramos, J. Marti, and V. Polo, Compensation of chromatic dispersion effects in microwave/millimeter-wave optical systems using four-wave-mixing induced in dispersionshifted fibers, IEEE Photon. Technol. Lett. 11, (1999). 36. P. C. Won, W. Zhang, and J. A. R. Williams, Self-phase modulation dependent dispersion mitigation in high power SSB and DSB dispersion compensated modulated radio-overfiber links, IEEE MTT-S International Microwave Symposium Digest (IEEE, 2006), pp J. Marti and F. Ramos, Compensation for dispersion-induced nonlinear distortion in subcarrier systems using optical-phase conjugation, Electron. Lett. 33, (1997). 38. E. Vergnol, D. Tanguy, J. F. Cadiou, A. Carenco, and E. Penard, Multi-carrier and m-qam modulation based on integrated single side band lighwave source, in Optical Fiber Communication Conference (Optical Society of America, 1999), pp G. H. Smith and D. Novak, Broadband millimeter-wave fiber-radio network incorporating remote up/downconversion, in IEEE MTT-S International Microwave Symposium Digest (IEEE, 1998), pp P. Gamage, A. Nirmalathas, C. Lim, D. Novak, and R. Waterhouse, Experimental demonstration of digitized RF transport over optical fiber links, in International Topical Meeting on Microwave Photonics, Jointly Held with the 2008 Asia-Pacific Microwave Photonics Conference. MWP/APMP 2008 (IEEE, 2008), pp J. Ma, J. Yu, C. Yu, X. Xin, J. Zeng, and L. Chen, Fiber dispersion influence on

14 Vol. 8, No. 2 / February 2009 / JOURNAL OF OPTICAL NETWORKING 214 transmission of the optical millimeter-waves generated using LN-MZM intensity modulation, J. Lightwave Technol. 25, (2007). 42. J. Yu, Z. Jia, L. Yi, Y. Su, G.-K. Chang, and T. Wang, Optical millimeter-wave generation or up-conversion using external modulator, IEEE Photon. Technol. Lett. 18, (2006). 43. L. T. Nichols and R. D. Esman, Single sideband modulation techniques and applications, in Optical Fiber Communication Conference (Optical Society of America, 1999), pp C. Lim, K. L. Lee, A. Nirmalathas, D. Novak, and R. Waterhouse, Impact of chromatic dispersion on 60 GHz radio-over-fiber transmission, 21st Annual Meeting of the IEEE Lasers and Electro-Optics Society (IEEE, 2008), pp C. Lim, A. Nirmalathas, D. Novak, and R. Waterhouse, Impact of ASE on phase noise in LMDS incorporating optical fibre backbones, in International Topical Meeting on Microwave Photonics (IEEE, 2000), pp C. G. Schaffer, F. H. Lubeck, R. P. Braun, G. Grosskopf, and F. Schmidt, Compensation of fiber dispersion in an optical mm-wave system in the 60 GHz-band, in IEEE MTT-S International Microwave Symposium Digest (IEEE, 1998), pp C. G. Schaffer, M. Sauer, K. Kojucharow, and H. Kaluzni, Increasing the channel number in WDM mm-wave systems by spectral overlap, in International Topical Meeting on Microwave Photonics (IEEE, 2000), pp C. Lim, A. Nirmalathas, D. Novak, R. S. Tucker, and R. B. Waterhouse, Technique for increasing optical spectral efficiency in millimeter-wave WDM fiber-radio, Electron. Lett. 37, (2001). 49. H. Toda, T. Yamashita, K.-I. Kitayama, and T. Kuri, A DWDM mm-wave fiber-radio system by optical frequency interleaving for high spectral efficiency, in International Topical Meeting on Microwave Photonics 2001 (IEEE, 2002), pp C. Lim, A. Nirmalathas, D. Novak, and R. B. Waterhouse, Capacity analysis for WDM fiber-radio backbones with star-tree and ring architecture incorporating wavelength interleaving, J. Lightwave Technol. 21, (2003). 51. C. Marra, A. Nirmalathas, D. Novak, C. Lim, L. Reekie, J. Besley, and N. Baker, Wavelength interleaved OADMs incorporating optimized multiple phase-shifted FBGs for fiber-radio systems, J. Lightwave Technol. 21, (2003). 52. M. Bakaul, A. Nirmalathas, and C. Lim, Multifunctional WDM optical interface for millimeter-wave fiber-radio antenna base station, J. Lightwave Technol. 23, (2005). 53. M. Bakaul, A. Nirmalathas, C. Lim, D. Novak, and R. B. Waterhouse, Simultaneous multiplexing and demultiplexing of wavelength-interleaved channels in DWDM millimeterwave fiber-radio networks J. Lightwave Technol. 24, (2006). 54. H. Toda, T. Yamashita, T. Kuri, and K.-I. Kitayama, Demultiplexing using an arrayedwaveguide grating for frequency-interleaved DWDM millimeter-wave radio-on-fiber systems, J. Lightwave Technol. 21, (2003). 55. P. A. Gamage, A. Nirmalathas, C. Lim, M. Bakaul, D. Novak, and R. Waterhouse, Efficient transmission scheme for AWG-based DWDM millimeter-wave fiber-radio systems, IEEE Photon. Technol. Lett. 19, (2007). 56. M. Attygalle, C. Lim, G. J. Pendock, A. Nirmalathas, and G. Edvell, Transmission improvement in fiber wireless links using fiber Bragg gratings, IEEE Photon. Technol. Lett. 17, (2005). 57. M. Attygalle, C. Lim, and A. Nirmalathas, Extending optical transmission distance in fiber wireless links using passive filtering in conjunction with optimized modulation, J. Lightwave Technol. 24, (2006). 58. D. Wake, D. Johansson, and D. G. Moodie, Passive pico-cell new in wireless network infrastructure, Electron. Lett. 33, (1997). 59. A. Stohr, K. Kitayama, and D. Jager, Full-duplex fiber-optic RF subcarrier transmission using a dual-function modulator/photodetector, IEEE Trans. Microwave Theory Tech. 47, (1999). 60. K. Kitayama, A. Stohr, T. Kuri, R. Heinzelmann, D. Jager, and Y. Takahashi, An approach to single optical component antenna base stations for broad-band millimeter-wave fiberradio access systems, IEEE Trans. Microwave Theory Tech. 48, (2000). 61. L. Chen, H. Wen, and S. Wen, A radio-over-fiber system with a novel scheme for millimeter-wave generation and wavelength reuse for up-link connection, IEEE Photon. Technol. Lett. 18, (2006). 62. J. Yu, Z. Jia, T. Wang, and G.-K. Chang, A novel radio-over-fiber configuration using optical phase modulator to generate an optical mm-wave and centralized lightwave for uplink connection, IEEE Photon. Technol. Lett. 19, (2007). 63. A. Nirmalathas, D. Novak, C. Lim, and R. B. Waterhouse, Wavelength re-use in the WDM optical interface of a millimeter-wave fiber wireless antenna base-station, IEEE Trans. Microwave Theory Tech. 49, (2001). 64. L. S. Fock and R. S. Tucker, Simultaneous reduction of intensity noise and distortion in semiconductor lasers by feed-forward compensation, Electron. Lett. 27, (1991). 65. C. Lim, A. Nirmalathas, K. L. Lee, D. Novak, and R. Waterhouse, Intermodulation distortion improvement for fiber-radio applications incorporating OSSB+C modulation in an optical integrated access environment, J. Lightwave Technol. 25, (2007).