Open Access
Add to BibSonomy
Abstract
We demonstrate the analog indoor distributed antenna system (DAS) for 5G mobile communications that supports 4 × 4 multiple-input multiple-output (MIMO) configuration. For this, we exploit a pair of intermediate frequency over fiber (IFoF)-based analog optical links, transporting 32 frequency allocation (FA) 5G mobile signals (effectively ~4 GHz bandwidth). The analog optical link manifests its high fidelity in the measured characteristics: small gain variation (< ± 1 dB for the entire transmission bandwidth), low noise (<-136 dBm/Hz), and large dynamic range (spurious free dynamic range of >106 dB∙Hz2/3), subsequently providing superior error vector magnitude (EVM) performance (~2%) for a wide range of ambient temperatures (−20 ~60°C). Consequently, the IFoF-based 4 × 4 MIMO supporting analog indoor DAS is capable of providing record high peak throughput of 5.3 Gb/s for millimeter wave based 5G mobile communication system.
© 2018 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Fifth generation mobile communication (abbreviated 5G) era is coming. It is predicted that 5G network technology would enable 20 Gb/s peak rate per cell and 1,000 times data traffic as compared to the current 4G LTE, readily providing speedier connectivity to not only smart phones but also many other devices such as sensors, robots, cars, etc [1–3]. One of the numerous challenges for realizing 5G network is to address the significantly increasing capacity requirement of densely populated indoor areas, where >80% of whole data traffic would be consumed [2]. However, in 5G network, the millimeter (mm)-wave carried signal is not likely to reach indoor users due to the characteristic of the high frequency electromagnetic waves particularly in case of, e.g., stadiums, tunnels, high-rise buildings constructed with low emissivity windows and concrete walls. In response to this problem, indoor distributed antenna system (DAS) is being more and more important, which is indeed one of the disruptive technologies for realizing 5G network [2,4,5].
Although there could be many different types of indoor DAS, it basically comprises two main units: i) DAS host unit and ii) distribution system. The DAS host unit would include the optical transceivers (TRx) that supply/receive the mobile signal to/from indoor users via the distribution system. The distribution system would comprise the fiber optic cables, splitters, and remote optic units (ROU) that includes optical TRx that receive/transmit the mobile signal from/to the DAS host unit, and antennas placed close to users.
One of the most critical parts of the indoor DAS is the cost/bandwidth efficient analog optical link [4,5]. In the conventional DAS based on digital transmission scheme, the DAS host unit digitizes the analog mobile signal, and transmits the digitized signal to the ROU where the digital signal is converted back to analog signal [6]. However, the digitization procedure demands tremendous signal bandwidth (e.g., ~118 Gb/s eCPRI-formatted data traffic to transport 8 frequency allocation (FA) 100 MHz bandwidth 5G mobile signal with 4 × 4 multiple-input multiple-output (MIMO) configuration), high complexity, and latency increase [7,8]. As an alternative, there are two options available: analog RF over fiber (RFoF) and analog IF over fiber (IFoF), where in both the analog signal from remote radio head (RRH, of mobile front-haul) is transparently transported without digitization process [8–13]. Although the analog IFoF requires more complicated ROU, for electronic frequency up/down conversion to the mmWave/IF frequency, as compared to the analog RFoF, it can be implemented with lower speed (i.e. lower cost) electrical-to-optical (EO) and optical-to-electrical (OE) conversion components, and has better flexibility in the IF carrier frequency allocation [8]. The IFoF-based analog optical link was experimentally demonstrated to transport 8 frequency allocation (FA) 5G mobile signal (~1 GHz analog bandwidth) in conjunction with mmWave based Giga-Korea (GK) 5G prototype that has single-input single-output (SISO) architecture [14,15]. As a further work of [15], we improved the IFoF-based analog optical link so it can transport 16 (2 × 8) FA 5G mobile signals [16].
In this paper, we present the IFoF-based analog indoor DAS supporting 4 × 4 MIMO configuration that is capable of delivering 32 (4 × 8) FA 5G mobile signal simultaneously. For this, we exploit a pair of the IFoF-based analog optical links to double the transmission capacity (from 16 FA to 32 FA). In addition, we spatially multiplex the multiple 5G signals through the 4 × 4 MIMO (where each antenna delivers 8 FA signal) to increase system capacity from 8 FA to 32 FA that has not been presented before in the field of 5G indoor DAS network. The use of the IFoF-based analog optical link let the system be efficient by simplifying the function blocks (e.g. removal of the digitization and framing/deframing units) and by using small signal bandwidth (i.e. removal of high-speed electronic components). We provide a full description of the indoor DAS architecture and analog optical TRxs which are key components of the analog optical transmission technology. We also introduce the technical details required for tackling the challenges that occur when implementing the 4 × 4 MIMO supporting DAS with the IFoF-based analog optical link employed. Subsequently, we characterize the analog optical link by measuring fundamental features: frequency response (S21), return losses (S11 and S22), noise performance, nonlinearity, and temperature dependent transmission performances. We also investigate the crosstalk- and fiber transmission-induced performance degradation. Finally, we demonstrate the indoor DAS that enables to provide real-time UHD video streaming services with multi-Giga bit data throughput and ~ms order latency.
2. IFoF-based analog indoor DAS supporting 4 × 4 MIMO
Figure 1 shows the indoor DAS that supports 4 × 4 MIMO (i.e. 8 antennas) with using a pair of the IFoF-based analog optical links. 5G baseband unit (BBU) generates the 5G mobile signal on IF-carriers, clock, and TDD synchronization signals. The main host unit (MHU), placed at the DAS host unit of e.g. a building, comprises: i) MHU IF unit (MIFU), ii) MHU analog TRx (MAT), iii) MHU digital unit (MDU), and iv) MHU control processing unit (MCPU). The IF based 5G mobile signal from the BBU is band-pass filtered and adjusted in its power by the MIFU, and is delivered to the MATs for fiber optic transmission. The MDU receives and digitizes: i) the frequency synchronization clock signal (100 MHz in our system) and ii) the TDD-synchronization signal from the BBU, and iii) the control/monitoring (C/M) signal from the MCPU, and then transmits those to the remote optic units (ROU) via the coarse wavelength division (de)multiplexers (CWDM) of MAT. The MDU employs 1330 nm for uplink (UL) and 1270 nm for downlink (DL), respectively. The data rate of the MDU is 100 Mb/s where the control/monitoring signal is transmitted by 15.36 Mb/s rate. Here we utilized two MATs to build 4 × 4 antenna system, where each ROU providing 2 × 2 MIMO is remotely connected to each MAT with 1-km long single-mode fiber (SMF). The received IF 5G signals at the ROU analog TRx (RAT) is sent to the ROU mmWave unit (RmmU) where the IF based 5G signal is electrically up-converted to the air-interface frequency (i.e. 28 GHz) for wireless transmission towards the 5G terminal, separated by ~2 m from the ROU. For this, the transferred 100 MHz clock signal is extracted at the ROU digital unit (RDU) and sent to the RmmU to frequency/phase-lock the 28 GHz voltage controlled oscillator (VCO). The TDD signal is also sent to the RmmU for triggering the electrical switches for TDD-operation of ULs and DLs, see the inset-i of Fig. 1. Inset-ii to -v show the photographs of the developed 5G BBU prototype, MHU, ROU, and Antennas and 5G-terminal, respectively.
Fig. 1 IFoF-based analog indoor distributed antenna system (DAS) supporting 4 × 4 MIMO configuration. (Inset-i: temporal waveforms of TDD-synchronization signal, downlink signal, and uplink signal, Inset-ii: 5G BBU prototype, Inset-iii: MHU, Inset-iv: ROU, Inset-v: Antennas and 5G terminal)
As mentioned earlier, we utilize a pair of IFoF-based analog optical links (i.e. using two MATs and two RATs) to build 4 × 4 MIMO system. The detailed configurations of the MAT and RAT, which are the key components of the IFoF link, are illustrated in Fig. 2(a) and 2(b), respectively. Each TRx module includes a pair of transmitters (Tx) and a pair of receivers (Rx), forming four paths (two uplinks: UL#1, UL#2, and two downlinks: DL#1, and DL#2). Four different wavelengths are assigned for the four paths: 1510 nm for UL#0, 1530 nm for UL#1, 1550 nm for DL#0, and 1570 nm for DL#1, respectively, with 20 nm guard-band, see Fig. 3(a). All Rxs are identical in its configuration, having a positive intrinsic negative-photodiode (PIN-PD), two cascaded RF amplifiers that give ~40 dB gain (at room temperature), and an output RF power control block (dashed line box in Fig. 2, which has a variable RF attenuator together with an RF amplifier, being able to provide −18.5 ~ + 13 dB gain) that aids automatic link gain control enabled by the MCPU and the RCPU. The output RF power control blocks are necessary for two main reasons: i) to prevent the system from being damaged by high incident RF power, and ii) to keep the constant link (RF) gain (e.g. 0 dB in our system), regardless of any environmental changes that could induce the link gain variation. MAT-Tx comprises a directly modulated uncooled distributed feedback laser diode (DFB LD) only. On the other hand, RAT-Tx has an input RF power control block (consisting of an RF amplifier with a variable RF attenuator, being able to provide −18.5 ~ + 13 dB gain) prior to the uncooled DFB LD, so we are able to adjust the RF power of the 5G mobile signal from user equipments (5G terminal in our set-up) via the RmmU that may be lower than −10 dBm which is the power requirement of our mmWave based 5G prototype [15,16]. All Txs have automatic output power control function so the uncooled DFB-LDs keep the constant output power: + 6dBm for our system, at which the transmitted signal has the optimum noise and nonlinearity characteristics that are to be discussed later. Four analog optical signals (UL#1, UL#2, DL#1, and DL#2) are combined with the digital signals (see the optical spectra in Fig. 3(a)) from the MDU and the RDU, by using a CWDM.
Fig. 2 Analog optical transceiver (TRx) modules for the IFoF-based indoor distributed antenna system (DAS) application: (a) MHU Analog TRx (MAT), (b) ROU Analog TRx (RAT).
Fig. 3 (a) Measured optical spectra of analog and digital signals of the IFoF-based analog indoor DAS and (b) measured electrical signal spectrum and error vector magnitude (EVM) of each FA of the input 5G mobile signal.
To characterize the analog optical link, we make a connection between the MAT and the RAT with use of a variable optical attenuator (VOA). In the experiment, we utilize an arbitrary waveform generator (AWG, M8190A Keysight) to generate 8 FA 5G mobile signal occupying from 1.7 to 2.7 GHz, see the black line in Fig. 3(b), where each FA is 125 MHz 64-QAM OFDM signal. We intentionally use this frequency range to avoid any influences of 2nd order nonlinear distortions (e.g. 2nd order harmonic distortion (HD2) and 2nd order intermodulation distortion (IMD2)) as those significantly degrade the signal quality according to our previous studies [18]. The total RF power of the 8FA 5G mobile signal to each Tx is −10 dBm that is, again, the RF power requirement of the mm-wave based 5G prototype [15,16]. For the same reason, the output RF power of the Rx also needs to be −10 dBm, requiring 0 dB RF link gain that is maintained by the output RF power control function block (black dashed line boxes in Fig. 2) in Rxs during operation. The error vector magnitude (EVM) of the each FA of the input 5G mobile signal from the AWG to Tx was measured to be <2%, see red circles in Fig. 3(b).
3. Experimental results and discussion
The analog optical link is required to have high fidelity so the temporal waveform of the IF mobile signal is well preserved after transmission. There are three main factors of the link that are related to the signal distortion: i) gain flatness in frequency response, ii) noise induced by the E-O-E conversion, including thermal, shot, and relative intensity noise (RIN), and iii) nonlinear distortions mainly caused by the RF amplifiers and the optoelectronic components (i.e. LD and PD) [17]. Firstly, we characterized the gain flatness by measuring the frequency response (S21) of our analog optical link in optical back-to-back (B-t-B) configuration, shown in Fig. 4. For each link, the received optical power was set to be 0 dBm and the link RF gain was fixed at 0 dB. The S21 graph, Fig. 4(a), shows that the RF gain deviation over the entire IF mobile signal bandwidth (1.7 ~2.7 GHz) is less than ± 1 dB for all links (i.e UL#1, UL#2, DL#1, and DL#2). This is due to the fact that we utilized a set of RF, photonic devices, and impedance matching by which flat frequency response can be achieved. It leads the power deviation of 8FA after the optical B-t-B transmission ( = output RF power – input RF power for each FA) to be <1 dB, which is depicted in Fig. 4(b). We also measured the input/output return losses (S11 and S22) illustrated in Fig. 5. For the IF mobile signal bandwidth (1.7 GHz ~2.7 GHz), both S11 and S22 were measured to be around/less than −14 dB that corresponds to the voltage standing wave ratio (VSWR) of 1.5:1.
Fig. 4 Measured (a) frequency response (S21) and (b) power deviation of transmitted 8 frequency allocation (FA) of IF-based mobile signals in optical back-to-back configuration.
Fig. 5 Measured (a) input (S11) and (b) output (S22) return losses of the analog optical link for IFoF based indoor DAS in optical back-to-back configuration.
Figure 6 shows the noise and nonlinear distortions, i.e. 2nd order intermodulation distortions (IMD2, blue triangles) and the 3rd order intermodulation distortions (IMD3, red circles), compared with the signal output power (black squares) of the analog optical link as a function of the input RF power to the Tx in optical B-t-B configuration. We omit the results of UL#2 and DL#2 as those are identical to UL#1 and DL#1 in the configurations (see Fig. 2), thus having the similar noise and nonlinear characteristics with UL#1 and DL#1. The IMD2 lies over the frequency range of 3.4~5.4 GHz that is out of the IF mobile signal bandwidth (1.7~2.7 GHz), thus IMD3 only needs to be considered [18]. As the ULs (Fig. 2(b)) have an additional RF amplifier (at the RF power control block in RAT-Tx) when compared to the DLs (Fig. 2(a)), the ULs have lower 3rd order input interception point (IIP3) (14 dBm) than that of the DLs (26 dBm). For the same reason, however, the ULs have relatively lower noise floor (−146 dBm/Hz) than that of the DLs (−136 dBm/Hz) as the electrical amplification (by the input RF power control block) prior to the EO conversion (by the DFB LD) at the RAT-Tx leads the UL to have relatively lower E-O-E conversion noise than the DL [17]. The measured noise level (<-136 dBm/Hz) indicates that the SNR would be >36 dB, considering −10 dBm input RF power and 1 GHz total signal bandwidth (8FA). Consequently, the spurious free dynamic range (SFDR) of UL (108 dB∙Hz2/3) comes to be slightly lower than the SFDR of DL (106 dB∙Hz2/3).
Fig. 6 Measured noise and nonlinear characteristics of (a) uplink#1 (UL#1) and (b) downlink#1 (DL#1). (IMD: Intermodulation Distortion, IIP3: 3rd-order Input Intercept Point)
The EVM of the transmitted IF-based 5G mobile signal includes all the influences of the measured i) gain flatness, ii) noise, and iii) nonlinear distortions of the link (Fig. 4-6). We measured the EVM values as a function of optical modulation index (OMI) with optical B-t-B configuration for various ambient temperatures of TRx’s (−20 ~60 °C) to see the temperature dependent transmission performances, Fig. 7. To examine the effect of temperature changes on the TRx separately, we placed either the MAT or the RAT in the chamber during the measurement: UL#1 (MAT-Rx under temperature control, Fig. 7(a)), UL#2 (RAT-Tx under temperature control, Fig. 7(b)), DL#1 (MAT-Tx under temperature control, Fig. 7(c)), DL#2 (RAT-Rx under temperature control, Fig. 7(d)). In the measurement, we controlled the OMI by varying the RF power (of the IF-based 5G signal) to Tx from the AWG with using a variable RF attenuator. At Rx, a vector signal analyzer (VSA, FSW Rohde & Schwarz) was used to capture the received IF 5G signal for the following MATLAB-based EVM analysis. In Fig. 7, we show the EVM of FA#1 and #8 only although the EVM of other FAs (#2 ~#7) were all measured to be at the comparable level. For the OMI lower than 0.1, the noise comes to have dominant influence on the EVM performance, while the nonlinear distortions become the key factor of determining the EVM for the OMI larger than 0.1 [19]. For the OMI around 0.1, the EVM performance of transmitted IF-based 5G signals are around 2%. Also, for the wide range of OMI (0.01~0.25), the EVM performance meets the 3GPP-defined mobile signal EVM requirement (<9%) [20]. At the −10 dBm input RF power to the Tx, the OMI of the ULs and DLs are 0.09 and 0.03 where the EVM degradation induced by the temperature variation (from −20 to 60 °C) was only <0.4%. This is mainly due to the fact that the modulation characteristic and noise/nonlinear properties of the uncooled DFB-LD remain unchanged thanks to the automatic output power control function (i.e. actively monitoring and controlling the DFB-LD that is enabled by the MCPU/RCPU in Fig. 1) regardless of the ambient temperature changes that could be occurred by not only the diurnal/annual temperature ranges but also by the system /hardware configurations.
Fig. 7 Measured error vector magnitude (EVM) as a function of optical modulation index (OMI) for various temperatures (−20~60 °C) of the analog optical transceivers (TRx): (a) uplink#1 (UL#1), (b) uplink#2 (UL#2), (c) downlink#1 (DL#1), and (d) downlink#2 (DL#2).
Integrated with the 5G network system, the analog optical link operates based in the time division duplexing (TDD), in which the ULs and DLs do not work at the same time, to be more specific, ULs are off while DLs are on, and vice versa. Thus, during the TDD-based operation there is no interaction between the ULs and DLs. However, there can be crosstalk between UL#1 and UL#2, and between DL#1 and DL#2 caused by: i) electro-magnetic interference (EMI) at the TRx printed circuit boards (PCB) and ii) optical interference at the CWDM filter of the MATs and RATs. To minimize the EMI, we spatially separate the Txs and Rxs inside the MAT/RAT module. Also, the optical interference level at the CWDM filter was measured to be <-50 dB in optical power due to the large guard band (i.e. 20 nm) between channels. To investigate the effects of crosstalk, we measure the frequency response between the UL#1 and UL#2 (or between the DL#1 and DL#2) with a network analyzer (N5225A, Keysight). Also, two non-identical IF signals are applied to the Txs of the UL#1 and UL#2 (or DL#1 and DL#2) for the EVM measurement. The crosstalk was measured in optical B-t-B to be <-40 dB over the entire IF based 5G signal bandwidth (1.7 ~2.7 GHz) for both the ULs (Fig. 8(a)) and the DLs (Fig. 8(b)). For all cases, the measured EVM degradation due to the crosstalk was less than 0.1%.
Fig. 8 Measured crosstalk (a) between the uplink#1 (UL#1) and the uplink#2 (UL#2), and (b) between the downlink#1 (DL#1) and the downlink #2 (DL#2).
The reach of the indoor DAS, i.e. the distance between the MHU and the ROU, is supposed to be shorter than 5 km, thus it has negligible loss by SMF (e.g. < 1 dB). In case the system is traditionally configured in the point-to-multipoint transmission architecture, however, there could be significant loss caused by optical power splitters (e.g. ~10 dB loss by 1 × 8 splitting and excess loss). Here, we performed the EVM measurement as a function of the received optical power for the optical B-t-B (filled black rectangles: FA#1, filled red circles: FA#4, and filled blue triangles: FA#8) and for 5km SMF transmission (hollow black rectangles: FA#1, hollow red circles: FA#4, and hollow blue triangles: FA#8), which are shown in Fig. 9. We utilize output RF power control blocks of the Rxs in order to maintain the constant RF link gain (i.e. 0 dB) while the received optical power is changing. The available lowest received optical power to keep the 0 dB RF link gain (i.e. −10 dBm output RF power) was −11 dBm and −5.5 dBm for ULs and DLs, respectively. As the received optical power decreases, the E-O-E conversion induced noise increases, leading to the EVM degradation, still lower than 5% though for the minimum received optical power. Considering the + 6 dBm Tx output optical power, the optical link budget are 17 dB and 11.5 dB, indicating >32 and >8 split-ratio, showing its good practicability in terms of spatial coverage. Also, the EVM degradation by the 5km SMF transmission (as compared to optical B-t-B) was <0.4% for all cases.
Fig. 9 Measured error vector magnitude (EVM) as a function of the received optical power for optical back-to-back and 5-km single mode fiber (SMF) transmission: (a) uplink#1 (UL#1), (b) uplink#2 (UL#2), (c) downlink#1 (DL#1), and (d) downlink#2 (DL#2).
Finally, we demonstrate the analog indoor DAS system that supports 4 × 4 MIMO configuration with using a pair of the IFoF-based analog optical links, depicted in Fig. 1. We use the 1 GHz (8 × 125 MHz) bandwidth 64-QAM OFDM signal that is generated by the 5G BBU. Each 125 MHz OFDM signal has the symbol length (FFT size) of 1024 and the cyclic prefix (CP) length of 128. The number of subcarriers is 624 with 180 kHz carrier spacing. One of the important factors determining the 5G mobile communication system performance is the phase noise of the mmWave carrier (28 GHz) that is supposed to be , where is the phase noise of the reference clock (i.e. 100 MHz in our system), is the additive phase noise of the transport link (i.e. MHU-to-ROU), and N is the frequency multiplication factor of the clock to the mm-wave (280 = 28GHz/100MHz in our system). Thus the phase noise added by the transport link needs to be minimized. The black curve in Fig. 10 indicates the measured phase noise of the clock signal at the clock board of the 5G BBU (i.e. ). The blue curve in Fig. 10 shows the phase noise of the received clock signal measured at the ROU (i.e. ) where the difference between two phase noise curves is the added phase noise by the MHU-ROU link that includes the MDU, MAT, 1km-SMF, RAT, and RDU. In our previous study, we experimentally confirmed that the phase noise of the received clock (100 MHz) at the ROU is required be below −100 dBc/Hz at 10 kHz offset frequency where our proposed system clearly meets the requirement [15].
Fig. 10 Measured phase noise of the 100 MHz clock signal at the 5G baseband unit (BBU, black curve) and at the remote mm-wave unit (RmmU, blue curve).
Table 1 shows the measured EVM values and constellations of the first FA (FA#1) of the 5G signal received by the four antennas (#1 ~#4) of the 5G-terminal. The measured EVM values at all antennas were lower than 7% that satisfies the 3GPP-defined EVM requirement (i.e. 9%) [20]. Antenna#3 shows the worst EVM performance (i.e. 6.9%) that is attributed to the frequency up-conversion performance of the used RmmU. The EVM degradation by the IFoF-based 1-km analog optical link (as compared to the electrical B-t-B measurement) was less than 0.5% for all antennas. In addition, the FA-dependent EVM degradation was measured to be less than 1%.
Table 1. Error vector magnitude (EVM) and constellations at the 4 antennas of the 5G terminal of the IFoF-based analog indoor DAS supporting 4 × 4 MIMO configuration.
We performed the real-time 5G demonstration through the IFoF-based analog indoor DAS system, providing 4K UHD video streaming service from the BBU to the 5G-terminal. During the service, the 4 × 4 MIMO system maintained real-time downlink data-rate over 5 Gb/s as seen in Fig. 11(a). Also, the measured peak data-rate reached 5.3 Gb/s, which is, to the best of our knowledge, the highest data-rate attained in real-time 5G services provided by the analog indoor DAS technologies. The user-plane roundtrip latency (i.e. ping roundtrip time) was measured to be <5 ms, that is, the one way latency is <2.5 ms. This is greatly shorter than the currently reported conventional digital DAS for the 4G LTE that has hundreds of ms response speed [6]. The short latency is enabled by i) making the transmission time interval (TTI) as short as possible (200 µs in our system) so as to reduce the time required for the signal processing and scheduling, and ii) eliminating the time taken for converting and mapping the 5G signal between the digital and analog formats.
Fig. 11 Real-time demonstration of 5G mobile service with the analog IFoF-based indoor distributed antenna system (DAS): (a) measured real-time and peak data rate, (b) measured latency, and (c) snapshot of 4K UHD video streaming
Table 2 compares our IFoF system to other analog RFoF and IFoF technologies [10–13]. [10] and [11] demonstrated the RFoF scheme for its application in (25 km and 15.5 km reaching) mobile front-haul based on 60 GHz and 14 GHz carrier frequencies, respectively. Ref. [10] employs optical heterodyne technique to perform photonic frequency up-conversion that allows for simple implementation of the remote radio head (RRH). Ref. [11] presents the compact four channel analog optical transceivers (TRx) integrated with an arrayed waveguide grating (AWG) using silica-based planar lightwave circuit (PLC) technology. For both, the demonstrated analog signal bandwidth was limited to hundreds of MHz. Ref. [12,13] show the capacity increase in the IFoF technology using digital signal processing (DSP) and parallel intensity/phase modulators, respectively, for both in order to improve the transmission performances, enabling several GHz analog signal delivery over tens of km. They aim to show the maximum capacity of a single IFoF link with the EVM approaching 8%, where additional EVM degradation (e.g. by up-conversion to air-interface frequency) is not allowed. Thus, there is a trade-off between the available maximum capacity of an analog RFoF/IFoF link and the EVM performance. For our system, to secure a large room (in terms of the EVM value) for other function blocks than the IFoF link, such as frequency up/down conversion (between IF and mmWave) unit, we tried to reduce the EVM value of the IFoF link as low as possible at the expense of the analog signal bandwidth per wavelength. Instead, we compactly integrated multi-wavelength TRx to interface with the DAS and 5G system so it can provide up-to 5.3 Gb/s real-time throughput and <5 ms user-plane roundtrip latency. This is in line with our purpose of this work that is to build a not only large capacity but also deployable 5G DAS network that is highly robust and reliable in long-term basis.
Table 2. Comparison of various analog RFoF/IFoF technologies
4. Summary
We demonstrated the IFoF-based analog indoor DAS that provides 5G mobile services via 4 × 4 MIMO configuration with using a pair of analog optical links. Each analog optical link was possible to transmit 16 frequency allocation (FA) IF based 5G mobile signal simultaneously where the error vector magnitude (EVM) of the delivered 5G signal meets 3GPP limit (9%), reaching 2% for wide temperature variation range: from −20 to 60 °C. Consequently, we were able to obtain the record high peak rate of 5.3 Gb/s for 5G mobile communication services provided by the 4 × 4 MIMO supporting IFoF-based analog indoor DAS.
Funding
'The Cross-Ministry Giga KOREA Project' grant funded by the Korea government Ministry of Science, ICT and Future Planning (MSIT) (No. GK18N0300, Development of Indoor DAS Technology, based on IFoF for 5G Mobile Communication).
References
1. R. Waterhouse and D. Novak, “Realizing 5G,” IEEE Microw. Mag. 16(8), 84–92 (2015). [CrossRef]
2. F. Boccardi, R. W. Heath, A. Lozano, T. L. Marzetta, and P. Popovski, “Five Disruptive Technology Directions for 5G,” IEEE Commun. Mag. 52(2), 74–80 (2014). [CrossRef]
3. D. Pliatsios, P. Sarigiannidis, S. Goudos, and G. K. Karagiannidis, “Realizing 5G vision through Cloud RAN: technologies, challenges, and trends,” EURASIP J. Wirel. Commun. Netw. 2018(1), 136 (2018). [CrossRef]
4. M. Crisp, R. V. Penty, and I. H. White, “Wideband Radio over Fiber Distributed Antenna Systems for Energy Efficient In-Building Wireless Communications,” in Proceedings of the 71st IEEE Vehicular Technology Conference, 11389883 (2010).
5. J. Beas, G. Castanon, I. Aldaya, A. Aragon-Zavala, and G. Campuzano, “Millimeter-Wave Frequency Radio over Fiber Systems: A Survey,” IEEE Commun. Surv. Tutorials 15(4), 1593–1619 (2013). [CrossRef]
6. J. Wang, S. Zhaob, X. Hu, E. Guan, Z. Ding, and Y. Yaob, “DDAS-DRoF Based a New In-Building Signal Coverage System Supporting Multiple Services,” in Proceedings of the 8th International Congress of Information and Communication Technology, 558–563 (2017).
7. eCPRI Specification V1.1, “Commom Public Radio Interface: eCPRI Interface Specification,” Jan. 2018.
8. S. Cho, H. Park, H. Chung, K. Doo, S. Lee, and J. Lee, “Cost-effective Next Generation Mobile Fronthaul Architecture with Multi-IF Carrier Transmission Scheme,” in Optical Fiber Communications Conference, paper Tu2B (2014). [CrossRef]
9. L. Xie, J. W. Man, B. J. Wang, Y. Liu, X. Wang, H. Q. Yuan, L. J. Zhao, H. L. Zhu, N. H. Zhu, and W. Wang, “24-GHz directly modulated DFB laser modules for analog applications,” IEEE Photonics Technol. Lett. 24(5), 407–409 (2012). [CrossRef]
10. C. Browning, E. P. Martin, A. Farhang, and L. P. Barry, “60 GHz 5G Radio-Over-Fiber Using UF-OFDM With Optical Heterodyning,” IEEE Photonics Technol. Lett. 29(23), 2059–2062 (2017). [CrossRef]
11. J. Liu, Y. Ye, L. Deng, L. Liu, Z. Li, F. Liu, Y. Zhou, J. Xia, and D. Liu, “Integrated four-channel directly modulated O-band optical transceiver for radio over fiber application,” Opt. Express 26(17), 21490–21500 (2018). [CrossRef] [PubMed]
12. X. Liu, H. Zeng, N. Chand, and F. Effenberger, “Efficient Mobile Fronthaul via DSP-Based Channel Aggregation,” J. Lightwave Technol. 34(6), 1556–1564 (2016). [CrossRef]
13. S. Ishimura, B. G. Kim, K. Tanaka, K. Nishimura, H. Kim, Y. C. Chung, and M. Suzuki, “Broadband IF-Over-Fiber Transmission With Parallel IM/PM Transmitter Overcoming Dispersion-Induced RF Power Fading for High-Capacity Mobile Fronthaul Links,” IEEE Photonics J. 10(1), 7900609 (2018). [CrossRef]
14. S. Song, K. Chang, C. Yoon, and J. Chung, “Special Issue on 5G Communications and Experimental Trials with Heterogeneous and Agile Mobile networks,” ETRI J. 40(1), 7–9 (2018). [CrossRef]
15. M. Sung, S. Cho, J. Kim, J. Lee, J. Lee, and H. Chung, “Demonstration of IFoF-Based Mobile Fronthaul in 5G Prototype With 28-GHz Millimeter wave,” J. Lightwave Technol. 36(2), 601–609 (2018). [CrossRef]
16. J. Kim, M. Sung, E. Kim, S. Cho, and J. Lee, “Experimental demonstration of analog IFoF-based seamless fiber-wireless interface for 5G indoor DAS supporting 8 FA and 2×2 MIMO configuration,” in Optical Fiber Communications Conference (2018), paper W4B.4. [CrossRef]
17. X. Xie, K. Li, Q. Zhou, A. Beling, and J. C. Campbell, “High-Gain, Low-Noise-Figure, and High-Linearity Analog Photonic Link Based on a High-Performance Photodetector,” J. Lightwave Technol. 32(21), 4187–4192 (2014). [CrossRef]
18. M. Sung, S. H. Cho, H. S. Chung, S. M. Kim, and J. H. Lee, “Investigation of transmission performance in multi-IFoF based mobile fronthaul with dispersion-induced intermixing noise mitigation,” Opt. Express 25(8), 9346–9357 (2017). [CrossRef] [PubMed]
19. C. Han, S. Cho, M. Sung, H. Chung, and J. Lee, “Clipping Distortion Suppression of Directly Modulated Multi-IF-over-Fiber Mobile Fronthaul Links Using Shunt Diode Predistorter,” ETRI J. 38(2), 227–234 (2016). [CrossRef]
20. 3GPP TS 36.141 v. 10.1.0, “LTE; Evolved Universal Terrestrial Radio Access (E-UTRA); Base Station (BS) conformance testing,” Technical Specification Group Radio Access Network, Rel. 10, Jan. 2011.
