Open Access
Add to BibSonomyAbstract
We summarize several different approaches for the realization of large capacity (>100Gb/s) fiber wireless integration system, including optical polarization-division-multiplexing (PDM) combined with multiple-input multiple-output (MIMO) reception, advanced multi-level modulation, optical multi-carrier modulation, electrical multi-carrier modulation, antenna polarization multiplexing and multi-band multiplexing. These approaches can effectively reduce the signal baud rate as well as the required bandwidth for optical and electrical devices. We also investigate the problems, such as wireless multi-path effect due to different wireless transmission distance, existing in the large capacity fiber wireless integration system. We demonstrate these problems can be effectively solved based on advanced digital-signal-processing (DSP) algorithms including classic constant modulus algorithm (CMA). Moreover, based on the combination of these approaches as well as advanced DSP algorithms, we have successfully demonstrated a 400G fiber wireless integration system, which creates a capacity record of wireless delivery and ushers in a new era of ultra-high bit rate (>400Gb/s) optical wireless integration communications at mm-wave frequencies.
© 2013 Optical Society of America
1. Introduction
It is well known that fourth-generation (4G) wireless systems, with the potential data rate of 100Mb/s and beyond, promote a substantial increase in throughput over existing second-generation (2G) and third-generation (3G) cellular systems. Furthermore, wireless devices with the bit rate of 1Gb/s and beyond have been commercially available using a variety of techniques. Adopting high-frequency millimeter-wave (mm-wave) frequency bands as well as high-speed wireless devices, some wireless systems can realize 10-Gb/s and even 40-Gb/s signal delivery over short wireless distances. Such high-rate wireless delivery can truly be described as comparable to or perhaps faster than fiber. Meanwhile, recently, smart mobile and fixed terminals have been equipped with very high definition (HD) camera such as 8k video (compared with current 1k). For example, some advanced companies, such as Nokia and Samsung, have introduced 8k and 4k super HD (SHD) video camera in smart phones that require transmission speed for uncompressed SHD video images of 60Gb/s and 30Gb/s, respectively [1]. Evidently, it's not desirable to require such thin and light-weight mobile terminals to install heavy high definition multimedia interfaces (HDMIs) or fiber cables.
On the other hand, due to wider bandwidths and higher frequencies, wireless delivery in mm-wave frequency bands is expected to provide multi-gigabit mobile data transmission, and has been intensively studied in the research community [2–41]. Furthermore, 100-Gb/s and beyond wireless mm-wave signal can be achieved based on photonic mm-wave technique, in which the ultra-high-speed optical baseband signal is first generated by externally modulating a continuous-wavelength (CW) lightwave, and then beat with another CW lightwave at different wavelength to generate the ultra-high-speed wireless mm-wave signal. The photonic mm-wave technique thus further promotes the seamless integration of wireless and fiber-optics networks.
However, the photonic mm-wave generation based on the fiber wireless integration system, just as shown in Fig. 1(a), is different from that based on traditional radio-over-fiber (RoF) system just as shown in Fig. 1(b) [42]. For the latter, the optical mm-wave or microwave wave (at lower frequency) is generated at the optical line terminator (OLT) or central office and thus the optical carrier and the optical baseband signal are simultaneously transmitted over the fiber [43], while for the former, the optical baseband signal is up-converted at the remote station and only the optical baseband signal is transmitted over the fiber. Compared to the traditional RoF system, the fiber wireless integration system will suffer less from linear fiber transmission impairments, such as chromatic dispersion (CD) and polarization mode dispersion (PMD), and nonlinear fiber transmission impairments at the cost of a more complex optical up-converter. Thus, the fiber wireless integration system can realize longer-haul fiber transmission.
Fig. 1 Photonic mm-wave generation based on (a) fiber wireless integration system and (b) traditional RoF system. OLT: optical line terminator.
Based on photonic mm-wave technique, the required bandwidth for optical and electrical devices in the large capacity (>100Gb/s) fiber wireless integration system can be reduced by reducing the signal baud rate, for which, just as shown in Fig. 2, several different approaches can be adopted. The first approach is optical polarization division multiplexing (PDM) together with multiple-input multiple-output (MIMO) reception, which can effectively double the transmission bit rate [19–26]. The second approach is advanced multi-level modulation, such as 16-ary quadrature amplitude modulation (16QAM) and 64QAM, but higher receiver sensitivity is required [26]. The third approach is optical multi-carrier modulation, such as optical orthogonal frequency division multiplexing (OFDM) and Nyquist wavelength division multiplexing (WDM), which also gives the possibility for optical sub-carrier optimization [27–30]. The fourth approach is electrical multi-carrier modulation, such as electrical OFDM, which is robust to fiber CD and PMD [31–37]. The fifth approach is antenna polarization multiplexing, but more optical and electrical devices are required [38]. The last but not the least approach is multi-band multiplexing [39–41]. For example, the ultra-high-speed wireless signal can be delivered at 40-, 60-, 80-, and 100-GHz mm-wave frequencies at the same time. A series of problems, such as wireless multi-path effect due to different wireless transmission distance, nonlinearity, component filtering and so on, exist in the large capacity fiber wireless integration system. We have demonstrated these problems can be effectively solved based on advanced digital-signal-processing (DSP) algorithms including classic constant modulus algorithm (CMA).
Fig. 2 Different approaches for the realization of large capacity (>100Gb/s) fiber wireless integration system.
In this paper, we summarize several different approaches for the realization of large capacity (>100Gb/s) fiber wireless integration system, including optical PDM combined with MIMO reception, advanced multi-level modulation, optical multi-carrier modulation, electrical multi-carrier modulation, antenna polarization multiplexing and multi-band multiplexing. These approaches can effectively reduce the signal baud rate as well as the required bandwidth for optical and electrical devices. We also investigate the problems, such as multi-path effect due to different wireless transmission distance, existing in the large capacity fiber wireless integration system. We demonstrate these problems can be effectively solved based on advanced DSP algorithms including classic CMA. Moreover, based on the combination of these approaches as well as advanced DSP algorithms, we have successfully demonstrated a 400G fiber wireless integration system, which creates a capacity record of wireless delivery and ushers in a new era of ultra-high bit rate (>400Gb/s) optical wireless integration communications at mm-wave frequencies.
2. Approaches for the realization of large capacity (>100Gb/s) fiber wireless integration system
As mentioned in the introduction, several approaches, such as optical PDM combined with MIMO reception, advanced multi-level modulation, optical multi-carrier modulation, electrical multi-carrier modulation, antenna polarization multiplexing and multi-band multiplexing, can be adopted to reduce the signal baud rate and the required bandwidth for optical and electrical devices, and thus promote the realization of large capacity (>100Gb/s) fiber wireless integration system. The following part of this section is the detailed introduction of these approaches as well as the corresponding experimental demonstrations.
2.1 Optical PDM combined with MIMO reception
Figure 3 shows the schematic diagram for the fiber wireless integration system adopting optical PDM combined with MIMO reception, including optical baseband transmitter to generate optically-modulated PDM baseband signal, optical heterodyne up-converter to up-convert the optical PDM baseband signal into the mm-wave frequency band, and wireless mm-wave receiver to down-convert the received wireless mm-wave signal into the baseband.
Fig. 3 Schematic diagram for fiber wireless integration system adopting optical PDM combined with MIMO reception. (a) 2 × 2 MIMO wireless link without interference. (b) 2 × 2 MIMO wireless link with large interference. Opt. Mod.: optical modulator, Pol. Mux: polarization multiplexer, Pow. Div.: power divider.
At the optical baseband transmitter, the CW lightwave from a laser is modulated by an optical external modulator and then polarization-multiplexed by a polarization multiplexer to generate optical PDM signal. The optical external modulator is driven by the transmitter electrical data. After fiber-optic transmission, the optical PDM signal is received by the optical heterodyne up-converter. At the optical heterodyne up-converter, there is a laser functioned as the local oscillator (LO), an optical 180° hybrid, two fast-response photo detectors (PDs) and two transmitter horn antennas (HAs). Here, the LO is used as the carrier-frequency generating source. The frequency spacing between the LO and the laser at the optical baseband transmitter is located in the mm-wave frequency band in order to generate the mm-wave central carrier frequency for the up-converted wireless signal. The optical 180° hybrid includes two polarization beam splitters (PBSs) and two optical couplers (OCs), and is used to implement polarization diversity of the received optical signal and the LO in optical domain before heterodyne beating. Next, two fast-response PDs, functioned as two photo-mixers, directly up-convert the X- and Y-polarization components of the optical PDM signal into the mm-wave frequency band, respectively. It’s worth noting that, in Fig. 3, X- or Y-polarization component of the PDM signal after polarization diversity does not mean that only X- or Y-polarization signal exists at each output port of PBSs. In fact, each output port contains both X- and Y-polarization signals. In this paper, we define one output port of each PBS as X-polarization component and the other as Y-polarization for simplification. Then, the X- and Y-polarization up-converted components, at the same time, are independently sent into free space by two transmitter HAs, and then received by two corresponding receiver HAs, which makes up a 2 × 2 MIMO wireless link based on microwave polarization multiplexing. At the wireless mm-wave receiver, there is a two-stage down conversion. In the first-stage analog conversion, the X- and Y-polarization components are respectively down-converted to a lower intermediate frequency (IF) in analog domain based on balanced mixer and sinusoidal radio-frequency (RF) signal, and then sent into a digital storage oscilloscope (OSC) to implement analog-to-digital conversion. Finally, IF down conversion and data recovery is realized with DSP in digital domain.
For the 2 × 2 MIMO wireless link shown in Fig. 3(a), almost no wireless MIMO interference exists because of good directionality of transmitter and receiver HAs. However, for the 2 × 2 MIMO wireless link shown in Fig. 3(b), each receiver HA can receive wireless power from two transmitter HAs, and thus there exists large wireless MIMO interference. The wireless MIMO interference can be suppressed by frequency domain equalization proposed by [44]. However, considering the optical polarization de-multiplexing, a uniform method for the fiber wireless integration system would be better. For PDM signal, the fiber link and the 2 × 2 MIMO wireless link can be both considered based on a 2 × 2 MIMO model. Thus, the equalization of the wired and wireless integrated 2 × 2 MIMO channel can be uniformly realized by classic CMA based on DSP. The overall transfer function for the fiber-wireless 2 × 2 MIMO transmission can be expressed as
Where denotes the received PDM signal after both fiber and wireless transmission, while and denote the transmitted PDM signal and the noise, respectively. For Eq. (1), the Jones matrix includes the channel response of both the fiber link and the 2x2 MIMO wireless link, that is,Thus, in order to recover the transmitted PDM signal , the task is to estimate the total Jones matrix and obtain its inverse matrix. Just as in the digital coherent communication, four butterfly-configured adaptive digital equalizers based on CMA can be used to de-multiplex the received PDM signal.Correspondingly, we experimentally demonstrated a fiber wireless integration system adopting optical PDM combined with MIMO reception, which can deliver 108-Gb/s PDM quadrature-phase-shift-keying (PDM-QPSK) signal through 80-km single-mode fiber-28 (SMF-28) and 1-m 2 × 2 MIMO wireless link at 100GHz [24]. The X- and Y-polarization components of the PDM-QPSK optical baseband signal are simultaneously up-converted to 100-GHz wireless carrier by optical polarization-diversity heterodyne beating, and then delivered over a 2 × 2 MIMO wireless link. At the wireless receiver, two-stage down conversion is firstly done in analog domain based on balanced mixer and sinusoidal RF signal, and then in digital domain based on DSP. The classic CMA equalization based on DSP is used to realize polarization de-multiplexing. The bit-error ratio (BER) for the 108-Gb/s PDM-QPSK signal is under the pre-forward-error-correction (pre-FEC) threshold of 3.8 × 10−3 [45] after both 80-km SMF-28 transmission and 1-m MIMO wireless delivery at 100GHz.
Figure 4 shows the experimental setup for the seamless integration of 108-Gb/s PDM-QPSK 80-km SMF-28 transmission and 1-m MIMO wireless delivery at 100GHz. At the optical baseband transmitter, there is an external cavity laser (ECL) at 1558.51nm with linewidth less than 100kHz and output power of 14.5dBm. The CW lightwave from the ECL is modulated by an in-phase/quadrature (I/Q) modulator. The I/Q modulator is driven by a 27-Gbaud electrical binary signal, which, with a pseudo-random binary sequence (PRBS) length of 215-1, is generated from a pulse pattern generator (PPG). For optical QPSK modulation, the two parallel Mach-Zehnder modulators (MZMs) in the I/Q modulator are both biased at the null point and driven at the full swing to achieve zero-chirp 0- and π-phase modulation. The phase difference between the upper and lower branches of the I/Q modulator is controlled at π/2. The polarization multiplexing is realized by a polarization multiplexer, comprising a polarization-maintaining OC to split the signal into two branches, an optical delay line (DL) to provide a 150-symbol delay, an optical attenuator to balance the power of two branches and a polarization beam combiner (PBC) to recombine the signal. The generated signal is launched into 80-km SMF-28, which has 18-dB average loss and 17-ps/km/nm CD at 1550nm without optical dispersion compensation. An erbium-doped fiber amplifier (EDFA) is used to compensate for the fiber loss. The launched power (after EDFA) is 0dBm.
Fig. 4 Experimental setup. (a) X-polarization optical spectrum (0.01-nm resolution) after polarization-diversity splitting. (b) Electrical spectrum after first-stage analog down conversion. (c) Received QPSK constellation.
At the optical heterodyne up-converter, an ECL with linewidth less than 100kHz is used as LO at 1557.71nm, which has 100-GHz frequency offset (FO) relative to the received optical signal. Two PBSs and two OCs are used to implement polarization diversity of the received optical signal and LO in optical domain before heterodyne beating. Two single-ended PDs, each with 75-GHz 3-dB bandwidth and 7.5-dBm input power, directly up-convert the X- and Y-polarization components of the optical PDM-QPSK signal into W-band wireless signals, respectively. The X- and Y-polarization up-converted components carried by 100-GHz wireless carrier independently pass through two 100-GHz narrowband electrical amplifiers (EAs) with 32-dB gain, and then, are simultaneously sent into a 2 × 2 MIMO wireless air link. Each pair of transmitter and receiver HAs have a 0.5~1.5-m wireless distance, the X- and Y-polarization wireless links are parallel and two transmitter (receiver) HAs have a 40-cm distance. Each HA has a 25-dBi gain. Almost no wireless MIMO interference exists because of good directionality of transmitter and receiver HAs. Inset (a) in Fig. 4 shows the X-polarization optical spectrum after polarization-diversity splitting. The frequency spacing and power difference between the LO and the received optical signal is 100GHz and 20dB, respectively.
Two-stage down conversion is first implemented at the W-band wireless receiver. A 12-GHz sinusoidal RF signal firstly passes through an active frequency doubler ( × 2) and an EA in serial, and is then split into two branches by a power divider. Next, each branch passes through a passive frequency tripler ( × 3) and an EA. As a result of this cascaded frequency doubling, an equivalent 72-GHz RF signal is provided for the corresponding balanced mixer. Therefore, the X- and Y-polarization components with 28-GHz IF are obtained after first-stage down conversion, as shown in inset (b) of Fig. 4. Each band-pass low-noise amplifier (LNA) after the mixer is centered on 100GHz and has a 5-dB noise figure. The analog-to-digital conversion is realized in the digital storage OSC with 120-GSa/s sampling rate and 45-GHz electrical bandwidth. The detailed DSP is given as follows. Firstly, the clock is extracted using the “square and filter” method, and then the digital signal is re-sampled at twice of the baud rate based on the recovered clock. Secondly, the received signals are down-converted to the baseband by multiplying synchronous cosine and sine functions, which are generated from a digital LO for down conversion [46]. Thirdly, a T/2-spaced time-domain finite impulse response (FIR) filter is used for CD compensation, where the filter coefficients are calculated from the known fiber CD transfer function using the frequency-domain truncation method. Fourthly, two complex-valued, 13-tap, T/2-spaced adaptive FIR filters, based on the classic CMA, are used to retrieve the modulus of the PDM-QPSK signal and realize polarization de-multiplexing. The subsequent step is carrier recovery, which includes FO estimation and carrier phase estimation. The former is based on fast Fourier transform (FFT) method while the latter based on fourth-power Viterbi-Viterbi algorithm. Finally, differential decoding is used to eliminate the π/2 phase ambiguity before BER counting. Inset (c) in Fig. 4 shows the X-polarization received constellations at the BER of 1.2 × 10−3.
2.2 Advanced multi-level modulation
It’s well known that compared to QPSK modulation, 16QAM modulation can make each symbol carry more bits and thus further reduce the baud rate. We experimentally demonstrated a fiber wireless integration system adopting optical PDM combined with MIMO reception, which can deliver 14-Gbaud (112-Gb/s) PDM-16QAM signal through 400-km SMF-28 and 0.5-m 2 × 2 MIMO wireless link at 35GHz [26]. The 14-Gbaud PDM-16QAM optical baseband signal after 400-km SMF-28 transmission is first up-converted to 35-GHz mm-wave frequency and then delivered over 0.5-m 2 × 2 MIMO wireless link. At the wireless receiver, the mm-wave down conversion, CD compensation, polarization de-multiplexing and carrier recovery are all realized in the digital domain after analog-to-digital conversion. The BER is less than the pre-FEC threshold of 3.8 × 10−3 after both 0.5-m MIMO wireless delivery at 35GHz and 400-km SMF-28 transmission.
Figure 5 shows the experimental setup for the seamless integration of 112-Gb/s PDM-16QAM 400-km SMF-28 transmission and 0.5-m MIMO wireless delivery at 35GHz. At the optical baseband transmitter, the CW lightwave at 1549.8nm is generated from an ECL, and then modulated by an I/Q modulator. The 14-Gbaud electrical four-level signal with a PRBS length of 215-1 is generated from arbitrary waveform generator (AWG) to drive the I/Q modulator. It is worth noting that before I/Q modulation, the two middle levels of the 14-Gbaud electrical four-level signal are pre-compensated by multiplying 0.89 to overcome the nonlinear characteristics of the two parallel MZMs in the I/Q modulator. The subsequent polarization multiplexing is realized by a polarization multiplexer identical to that adopted in Fig. 4. The generated PDM-16QAM optical baseband signal is launched into the straight line consisting of 5 spans of 80-km SMF-28. An EDFA is used to compensate for the loss of each span. The total launched power (after EDFA) into each span is 1dBm.
Fig. 5 Experimental setup. (a) Parallel HA array. (b) Orthogonal HA array. (c) Detailed DSP. (d) X-polarization optical spectrum after polarization-diversity splitting. (e) Electrical spectrum after analog-to-digital conversion. Recovered constellations in the case of (f) BTB (OSNR = 28dB) and (g) 400-km SMF-28 transmission (OSNR = 28dB).
The optical heterodyne up-converter is quite similar to that adopted in Fig. 4, except that the FO between the LO and the received optical signal is 35GHz and two balanced PDs (BPDs), instead of two single-ended PDs, are used for photonic up-conversion. The generated two 35-GHz mm-wave signals are independently power-amplified by two EAs with 10-GHz 3-dB bandwidth and then broadcasted by two transmitter HAs with 25-dBi gain. Insets (a) and (b) in Fig. 5 show the parallel and orthogonal HA arrays, respectively. Each pair of transmitter and receiver HAs have a 0.5-m wireless distance for both parallel and orthogonal conditions. The crosstalk is not large because of good directionality of transmitter and receiver HAs. But it should be fine even if the crosstalk is large because we use blind equalization based on CMA to realize polarization de-multiplexing in the subsequent DSP process.
Then, the analog-to-digital conversion is realized in the digital storage OSC with 120-GSa/s sampling rate and 45-GHz electrical bandwidth. Inset (c) in Fig. 5 shows the detailed DSP after analog-to-digital conversion. The first three steps of the DSP process, i.e., time recovery, IF down conversion and CD compensation, are identical to those adopted in Fig. 4. Then, two complex-valued, 13-tap, T/2-spaced adaptive FIR filters are used to retrieve the modulus of the 16QAM signal. The two adaptive FIR filters are based on the classic CMA followed by three-stage CMA, to realize multi-modulus recovery and polarization de-multiplexing [47, 48]. The carrier recovery, including residual FO estimation and carrier phase estimation, is performed in the subsequent step. The fourth power is used to estimate the FO between the LO and the received optical signal. The phase recovery is obtained by feed-forward and Least-Mean-Square (LMS) algorithms. Finally, differential decoding is used to eliminate the π/2 phase ambiguity before BER counting. Inset (d) in Fig. 5 gives the X-polarization optical spectrum after polarization-diversity splitting, where the optical signal-to-noise ratio (OSNR) difference of the LO and the received optical signal is larger than 20dB. Inset (e) in Fig. 5 shows the electrical spectrum after analog-to-digital conversion. It can be seen that the central frequency is 35GHz. An obvious frequency dip appears at the frequency of 45GHz, where is the end of analog-to-digital converter (ADC) bandwidth.
Insets (f) and (g) in Fig. 5 give the recovered constellations in the case of back-to-back (BTB) (OSNR = 28dB) and 400-km SMF-28 transmission (OSNR = 28dB), respectively. BTB denotes no fiber transmission. Because of the adoption of conventional digital coherent demodulation algorithms (particularly CMA), the multiplexing effect caused by both fiber and 2 × 2 MIMO wireless links can be mostly compensated, and thus fiber nonlinearity is the key factor that influences the transmission performance. Although the 112-Gb/s PDM-16QAM signal is quite sensitive to fiber nonlinearity, compared to the BTB case, only 3.5-dB OSNR penalty is caused by 400-km SMF-28 transmission [26]. Considering the very small penalty after 400-km SMF-28 transmission, it clearly demonstrates the feasibility of this 2 × 2 MIMO wireless transmission system.
2.3 Optical multi-carrier modulation
Compared to optical single-carrier modulation, optical multi-carrier modulation can also effectively reduce the transmission baud rate. Figure 6 shows the schematic diagram of the proposed multi-carrier optical heterodyne up-converter and wireless mm-wave receiver with joint-channel DSP. Take a three-channel dense WDM (DWDM) signal (ch1, ch2 and ch3) as an example. The wireless mm-wave carrier frequency is equal to the frequency spacing between LO and ch2. The optical heterodyne up-conversion and wireless mm-wave reception for the three-channel DWDM signal is quite similar to that for the single-carrier signal. Moreover, after analog-to-digital conversion, the generated digital signal with full channel information can be processed by joint-channel DSP, which will be introduced in detail in the following part.
Fig. 6 The schematic diagram of the proposed multi-carrier optical heterodyne up-converter and wireless mm-wave receiver with joint-channel DSP.
Correspondingly, based on the optical multi-carrier modulation technique, we experimentally demonstrated a fiber wireless integration system that delivers 120-Gb/s multi-channel PDM-QPSK signal through 80-km SMF-28 and 2-m 2 × 2 MIMO wireless link at 92GHz [30]. In this experiment, the 3 × 40-Gb/s three-channel PDM-QPSK signal has 12.5-GHz neighboring frequency spacing. Each receiver HA can receive signals from two transmitter HAs, and thus there exists wireless MIMO interference in the 2 × 2 MIMO wireless link. At the wireless receiver, the classic CMA equalization based on DSP is used to realize polarization de-multiplexing and suppress the wireless MIMO interference. We also experimentally demonstrated that more CMA taps are required for this system with the additional 2x2 MIMO wireless link.
Figure 7 shows the experimental setup for the seamless integration of 3 × 40-Gb/s multi-channel PDM-QPSK signal transmission through 80-km SMF-28 and 2-m 2 × 2 MIMO wireless link at 92GHz. At the optical baseband transmitter, an ECL at 1554.43nm is used with linewidth less than 100kHz and output power of 14.5dBm. For optical QPSK modulation, the I/Q modulator is driven by a 10-Gbaud electrical binary signal with a PRBS length of 210-1. After I/Q modulation, the 20-Gb/s optical QPSK signal is split into two branches. One branch is polarization-multiplexed to generate even-channel PDM-QPSK signal (channel 2), while the other is injected into a MZM. The MZM is driven by a 12.5-GHz RF signal and biased at null point, in order to implement optical carrier suppression (OCS) [49]. After OCS, the two sidebands of 25-GHz channel spacing are polarization-multiplexed to generate two odd-channel PDM-QPSK signals (channel 1 and channel 3). Both the I/Q modulator and the polarization multiplexer are identical to those adopted in Fig. 4. The odd and even channels are combined together by 12.5/25-GHz arrayed waveguide grating before fiber transmission. The generated 120-Gb/s multi-channel signal is launched into 80-km SMF-28. An EDFA is used to compensate for the fiber loss.
Fig. 7 Experimental setup. (a) Optical spectrum (0.02-nm resolution) after arrayed waveguide grating. (b) Optical spectrum (0.05-nm resolution) after polarization diversity splitting. (c) Electrical spectrum of the three-channel signal with 20-GHz IF. (d) Detailed DSP. (e) QPSK constellation with 13 taps. (f) QPSK constellation with 25 taps.
The optical heterodyne up-converter is quite similar to that adopted in Fig. 4, except that the FO between the LO and the received optical signal is 92GHz. The generated 3 × 40-Gb/s three-channel PDM-QPSK wireless signal at 92GHz is delivered by a 2 × 2 MIMO wireless link. Each pair of transmitter and receiver HAs have a 2-m wireless distance, while the distance between two transmitter HAs and two receiver HAs are 5cm and 10cm, respectively. Each HA has a 25-dBi gain. Each receiver HA can receive signals from two transmitter HAs, and thus there exists wireless MIMO interference in the 2 × 2 MIMO wireless link. Insets (a) and (b) in Fig. 7 show the optical spectra after 12.5/25-GHz arrayed waveguide grating and polarization-diversity splitting, respectively. The frequency spacing and power difference between Channel 2 and LO is 92GHz and 20dB, respectively.
At the wireless receiver, the first stage of the two-stage down conversion is identical to that adopted in Fig. 4. The three-channel PDM-QPSK signal centered on around 20GHz is obtained after first-stage analog down conversion. Then, the analog-to-digital conversion is realized in the digital storage OSC with 120-GSa/s sampling rate and 45-GHz electrical bandwidth. The electrical spectrum of the three-channel signal with 20-GHz IF is shown in inset (c) of Fig. 7. We can see that, due to the low-pass frequency response of wireless and electrical devices, channel 1 and channel 2 have a lower power compared to channel 3. Inset (d) in Fig. 7 shows the detailed DSP after analog-to-digital conversion. Firstly, the received three-channel PDM-QPSK signal is de-multiplexed in the electrical domain, which is realized by a digital fifth-order Bessel filter. Then, each sub-channel is down-converted to the baseband by multiplying synchronous cosine and sine functions with a different frequency (32.5GHz, 20GHz and 7.5GHz for channel 1, channel 2 and channel 3, respectively), which are generated from a digital LO for down conversion. The subsequent CD compensation, CMA equalization, carrier recovery, differential decoding and BER counting is quite similar to that adopted in Fig. 4. Here, two complex-valued, 25-tap, T/2-spaced adaptive FIR filters, based on the classic CMA, are used to retrieve the modulus of the PDM-QPSK signal, to realize polarization de-multiplexing and wireless interference suppression [25]. Insets (e) and (f) in Fig. 7 show the constellations with CMA-tap number of 13 and 25, respectively. The constellation for the 25-tap CMA length is much clearer than that for the 13-tap CMA length. To our knowledge, the CMA tap for PDM-QPSK in most coherent systems is around 13, which means that more taps are required for this system with the additional 2 × 2 MIMO wireless link.
2.4 Electrical multi-carrier modulation
As one kind of electrical multi-carrier modulation, OFDM has been widely used in the optical networks due to high resistance to fiber CD and PMD as well as high spectral efficiency (SE). OFDM can also be used in the fiber wireless integration system to reduce the transmission baud rate as well as improve performance and SE [31–37].
Correspondingly, we experimentally demonstrated a fiber wireless integration system adopting optical PDM combined with MIMO reception, which can deliver 30.67-Gb/s PDM OFDM signal through 40-km SMF-28 and 2-m 2 × 2 MIMO wireless link at 100GHz [37]. In this experiment, there also exists wireless MIMO interference in the 2 × 2 MIMO wireless link. At the wireless receiver, de-multiplexing is realized by channel estimation based on a pair of time-interleaved training sequences (TSs). The BER for the 30.67-Gb/s PDM OFDM signal is less than the pre-FEC threshold of 3.8 × 10−3 when the OSNR is larger than 17dB after both 40-km SMF-28 transmission and 2-m wireless delivery at 100GHz.
The experimental setup for the 30.67-Gb/s PDM OFDM optical baseband transmitter is quite similar to that shown in Fig. 4, except that the I/Q modulator is driven by an electrical baseband OFDM signal. The electrical OFDM signal is generated by AWG and the sampling rate is 11.5Gsa/s. For electrical OFDM modulation as shown in Fig. 8(a), the inverse FFT (IFFT) size is 256. Among the 256 subcarriers, 192 subcarriers are allocated for data transmission with 4QAM, 8 subcarriers are used as pilots for phase estimation, the first subcarrier is set to zero for DC-bias and the rest 55 null subcarriers at the edge are reserved for oversampling. After IFFT, cyclic prefix (CP), which is 1/8 of IFFT size, is added in the OFDM symbol. Two types of TSs are added in the front of the data stream. The first type includes only one TS used for time and frequency synchronization, while the other is comprised by one TS surrounded by two null symbols in order to construct a pair of time-interleaved TSs after polarization multiplexing and used for channel estimation. As shown in Fig. 8(b), Y-polarization is delayed by exactly one symbol compared to X-polarization after polarization multiplexer, which can construct a pair of time-interleaved TSs for de-multiplexing. The total bit rate is 30.67Gb/s (11.5 × 192/288 × 2 × 2Gb/s = 30.67Gb/s) after the polarization multiplexer. The generated signal is launched into 40-km SMF-28. The launched power (after EDFA) is 0dBm.
Fig. 8 (a) Electrical OFDM modulation. (b) Delay between two polarizations. (c) Detailed DSP after analog-to-digital conversion.
The operation of the optical heterodyne up-converter is identical to that adopted in Fig. 4, that is, the received 30.67-Gb/s PDM OFDM optical baseband signal is up-converted to the 30.67-Gb/s PDM OFDM wireless signal at 100GHz. For the 2 × 2 MIMO wireless link, each pair of transmitter and receiver HAs have a 2-m wireless distance, while the distance between two transmitter HAs and receiver HAs are 8cm and 10cm, respectively. Wireless interference exists in the 2 × 2 MIMO wireless link. At the wireless receiver, the first stage of the two-stage down conversion is also identical to that adopted in Fig. 4. Figure 8(c) shows the detailed DSP after analog-to-digital conversion. Firstly, the 30.67-Gb/s PDM OFDM wireless signal with 28-GHz IF is down-converted to baseband with the RF pilot, which is DC component of the signal injected into I/Q modulator. Secondly, time synchronization is realized by the conjugate symmetric OFDM symbol in time domain placed in the front of the frame as the first type of TS at the transmitter. Thirdly, channel estimation for the MIMO channel is implemented by a pair of TSs set as the second type of TS at the transmitter in two polarizations, and then, de-multiplexing can be realized in order to minimize crosstalk between two branches. Fourthly, phase noise cancellation in two branches is implemented with the pilots inserted in each OFDM symbol, and after the decision and de-mapper procedure, the feed-back algorithm is applied to improve the accuracy of the phase noise estimation. The final step is BER counting. It’s noted that intra-symbol frequency-domain averaging (ISFA) algorithm is applied to improve the accuracy of channel estimation. The subcarrier number used for ISFA is 13.
2.5 Antenna polarization multiplexing
It is well known that there exist two antenna polarization states, that is, horizontal-polarization (H-polarization) state and vertical-polarization (V-polarization) state. Figure 9 shows the schematic diagram of the fiber wireless integration system adopting antenna polarization multiplexing. Take a two-channel DWDM signal (ch1 and ch2) as an example. At the optical heterodyne up-converter, an additional wavelength selective switch (WSS) is adopted to de-multiplex the two-channel DWDM signal. Then, ch1 and ch2 are independently heterodyne up-converted at the same time. The heterodyne up-conversion of ch1 or ch2 is quite similar to that adopted in Fig. 4. The MIMO wireless link includes four transmitter HAs and four receiver HAs. The upper two transmitter HAs and two receiver HAs are all horizontally polarized, and thus form an H-polarization HA array to deliver ch1. Meanwhile, the lower two transmitter HAs and two receiver HAs are all vertically polarized, and thus form a V-polarization HA array to deliver ch2. After MIMO wireless delivery, the received wireless two-channel signal is sent into digital storage OSC, and offline DSP is implemented after analog-to-digital conversion. The signal baud rate and performance requirements for optical and wireless devices can be reduced by adopting antenna polarization multiplexing, but meanwhile double antennas and devices are required. The adoption of antenna polarization multiplexing can also increase wireless transmission capacity at the cost of stricter requirements for V-polarization.
Fig. 9 Schematic diagram of fiber wireless integration system adopting antenna polarization multiplexing.
Correspondingly, we experimentally demonstrated the 2 × 56-Gb/s two-channel PDM-QPSK signal delivery based on antenna polarization multiplexing over 80-km SMF-28 and 2-m Q-band (33-50GHz) wireless link [38]. At the wireless receiver, the classic CMA equalization based on DSP can realize polarization de-multiplexing and remove the crosstalk at the same antenna polarization. For 2-m wireless delivery, the BER of each channel after 80-km SMF-28 transmission can be under 3.8 × 10−3, while the BER without fiber transmission under 1 × 10−5. We also experimentally demonstrated the isolation is only about 19dB when the V-polarization deviation approaches to 100, which will affect the wireless delivery for high-speed (>50Gb/s) signal.
Figure 10(a) shows the experimental setup for the optical mm-wave generator for the 2 × 56-Gb/s two-channel PDM-QPSK wireless signal at Q-band. At the optical baseband transmitter, there are two ECLs with linewidth less than 100kHz. In the upper path, the CW lightwave from ECL1 at 1554.43nm is modulated by an I/Q modulator. The I/Q modulator is driven by a 14-Gbaud electrical binary signal, which, with a PRBS length of 215-1, is generated from PPG. The optical PDM-QPSK modulation is realized by the I/Q modulator and the subsequent polarization multiplexer, which is identical to that adopted in Fig. 4. In the lower path, the same operation is implemented except that the CW lightwave from ECL2 is at 1554.83nm. Next, by the combination of an OC, the generated 2 × 56-Gb/s two-channel PDM-QPSK optical baseband signal with 50-GHz channel spacing is amplified by an EDFA, and then launched into 80-km SMF-28. The second EDFA is used to compensate for the fiber loss. The total optical power after the first and second EDFAs is 4dBm and 10dBm, respectively. In the following part of this section, ch1 and ch2 are used to denote the two channels at 1554.43nm and 1554.83nm, respectively.
Fig. 10 Experimental setup. (a) Optical mm-wave generator. (b) Optical spectra (0.02-nm resolution) after polarization-diversity splitting. (c) Q-band HA array. (d) X-polarization constellation after 80-km SMF-28 transmission. (e) X-polarization BTB constellation.
At the optical up-converter, another two ECLs (ECL3 and ECL4) with linewidth less than 100kHz are used as LOs. ELC3 an ECL4 have a 37.5-GHz FO relative to ch1 and ch2, respectively. After passing through a 1 × 2 programmable WSS, ch1 and ch2 are independently heterodyne up-converted at the same time. The heterodyne up-conversion of ch1 or ch2 is quite similar to that adopted in Fig. 4, except that two single-ended PDs, each with 60-GHz 3-dB bandwidth and 7.5-dBm input power, directly up-convert the PDM-QPSK optical signal into the wireless signal at 37.5GHz. Figure 10(b) shows BTB optical spectra (0.02-nm resolution) after polarization diversity corresponding to ch2. Figure 10(c) shows the Q-band HA array including four transmitter HAs (TX1-TX4) and four receiver HAs (TX1-TX4). TX1, TX2, RX1 and RX2 are vertically polarized, while TX3, TX4, RX3 and RX4 are horizontally polarized. The V- and H-polarization HA arrays are used to deliver ch1 and ch2, respectively. The isolation between H- and V-polarization HA arrays is 33dB, but there exists large crosstalk at the same antenna polarization due to the deliberate HA arrangement. All receiver HAs can get the same power from each transmitter HA. Each HA has 25-dBi gain and frequency range of 33~50GHz, and is connected with a 60-GHz EA with 30-dB gain and 24-dBm saturation output power. After 2-m wireless delivery, the received two-channel PDM-QPSK wireless signal at 37.5GHz is first amplified by four parallel 60-GHz EAs and then directly sent to the digital storage OSC with 120-GSa/s sampling rate and 45-GHz electrical bandwidth. The subsequent DSP is quite similar to that adopted in Fig. 4. It is worth noting that, here, CMA equalization with 53 T/2 taps is used to realize polarization de-multiplexing and remove the crosstalk at the same antenna polarization [25]. Figure 10(d) shows the X-polarization constellation after 80-km SMF-28 transmission and 2-m wireless delivery at the BER of 4 × 10−4 and 24-dB OSNR. Figure 10(e) shows the X-polarization BTB constellation after 2-m wireless delivery when no crosstalk exists and the BER is 1 × 10−5.
2.6 Multi-band multiplexing
Similar to antenna polarization multiplexing, the adoption of multiple frequency bands can also effectively reduce the signal baud rate and performance requirements for optical and wireless devices, but at the cost of more antennas and devices. We experimentally demonstrate an optical wireless integration system simultaneously delivering 2 × 112-Gb/s two-channel PDM-16QAM wireless signal at 37.5GHz and 2 × 108-Gb/s two-channel PDM-QPSK wireless signal at 100GHz, adopting two mm-wave frequency bands, two orthogonal antenna polarizations, MIMO, photonic mm-wave generation and advanced DSP [41]. In the case of no fiber transmission, the BERs for both the 112-Gb/s PDM-16QAM signal after 1.5-m wireless delivery at 37.5GHz and the 108-Gb/s PDM-QPSK signal after 0.7-m wireless delivery at 100GHz are under the pre-FEC threshold of 3.8 × 10−3. To our knowledge, this is the first demonstration of a 400G optical wireless integration system in mm-wave frequency bands and also a capacity record of wireless delivery.
Figure 11(a) shows the experimental setup for the optical mm-wave generator for 2 × 112-Gb/s two-channel PDM-16QAM wireless signal at Q-band and 2 × 108-Gb/s two-channel PDM-QPSK wireless signal at W-band. At the optical baseband transmitter, there are four ECLs with linewidth less than 100kHz and output power of 14.5dBm. For optical PDM-16QAM modulation, the two CW lightwaves from ECL1 at 1554.83nm and ECL2 at 1554.43nm are first combined together by an OC, and then modulated by an I/Q modulator. The I/Q modulator is driven by a 14-Gbaud electrical four-level signal, which, with a PRBS length of 215-1, is generated from AWG. The operation of the I/Q modulator and the subsequent polarization multiplexer is identical to that adopted in Fig. 4. Thus, the 2 × 112-Gb/s two-channel PDM-16QAM optical baseband signal is generated with 50-GHz channel spacing. For optical PDM-QPSK modulation, similarly, the two CW lightwaves from ECL3 at 1553.82nm and ECL4 at 1553.22nm are combined together by an OC, then, modulated by an I/Q modulator and polarization-multiplexed by a polarization multiplexer. What is different from optical PDM-16QAM modulation is that the I/Q modulator is driven by a 27-Gbaud electrical binary signal, which, with a PRBS length of 215-1, is generated from PPG. Thus, the 2 × 108-Gb/s two-channel PDM-QPSK optical baseband signal is generated with 75-GHz channel spacing. The generated 2 × 112-Gb/s two-channel PDM-16QAM and 2 × 108-Gb/s two-channel PDM-QPSK optical signals are combined together by an OC, amplified by an EDFA, and then launched into 80-km SMF-28. The second EDFA is used to compensate for the fiber loss. The total optical power after the first and second EDFAs is 15dBm and 18dBm, respectively. Figures 11(b) and 11(c) show the optical spectra (0.1-nm resolution) after the first and second EDFA, respectively. In the following part of this section, ch1, ch2, ch3 and ch4 are used to denote the four channels at 1554.83nm, 1554.43nm, 1553.82nm and 1553.22nm, respectively.
Fig. 11 Experimental setup. (a) Optical mm-wave generator. (b)-(c) Optical spectra (0.1-nm resolution) after EDFA. (d)-(g) Optical spectra (0.02-nm resolution) after polarization diversity. (h) Q-band and W-band HA system. (i) X-polarization QPSK constellation. (j) X-polarization 16-QAM constellation.
At the optical up-converter, another four ECLs with linewidth less than 100kHz are used as LOs. ELC5 an ECL6 respectively have 37.5-GHz FO relative to ch1 and ch2, while ELC7 and ECL8 have 100-GHz FO relative to ch3 and ch4. After passing through a 1 × 4 programmable WSS, ch1-ch4 are independently heterodyne up-converted at the same time. Here, 5-m SMF is used to de-correlate ch1 and ch2 or ch3 and ch4. The heterodyne up-conversion of ch1 or ch2 is quite similar to that adopted in Fig. 5, except that two BPDs directly up-convert the PDM-16QAM optical signal into the PDM-16QAM wireless signal at 37.5GHz. Each BPD has a 10-dBm input power. The heterodyne up-conversion of ch3 or ch4 is quite similar to that adopted in Fig. 4, except that two PDs, each with 90-GHz 3-dB bandwidth and 7.5-dBm input power, directly convert the PDM-QPSK optical signal into the PDM-QPSK wireless signal at 100GHz. Figures 11(d)-11(g) show the optical spectra (0.02-nm resolution) after polarization diversity corresponding to ch1-ch4, respectively.
Figure 11(h) shows the HA system including a Q-band HA array and a W-band HA array. Each HA array includes four transmitter HAs and four receiver HAs. The two pairs of transmitter and receiver HAs in the middle of the W-band HA array are horizontally polarized, while the other two pairs vertically polarized. The 2 × 108-Gb/s two-channel PDM-QPSK wireless signal at 100GHz is delivered by the W-band HA array, with ch3 and ch4 corresponding to the H- and V-polarization HA array, respectively. For the W-band HA array, each receiver HA can only get the wireless power from corresponding transmitter HA as shown by the black dashed lines, which is because the distance between the HAs is much larger with respect to a wavelength of only 3 mm for the 100 GHz wireless mm-wave signal. Thus, there hardly exists crosstalk at the same polarization (H or V) for the W-band HA array. Moreover, the crosstalk, even if exists, can be compensated by a long-tap CMA [25]. On the other hand, the two pairs of transmitter and receiver HAs in the middle of the Q-band HA array are vertically polarized, while the other two pairs horizontally polarized. The 2 × 112-Gb/s two-channel PDM-16QAM wireless signal at 37.5GHz is delivered by the Q-band HA array, with ch1 and ch2 corresponding to the H- and V-polarization HA array, respectively. For the Q-band array, RX2 or RX3 can get the same wireless power from TX2 and TX3, while RX1 or RX4 can get the same wireless power from TX1 and TX4, as shown by the blue dashed lines. Thus, there exists large crosstalk at the same polarization (H or V) for the Q-band HA array. The wireless distance is 0.7m for the wireless transmission links at W-band, while 1.5m at Q-band. Each HA has a 25-dBi gain. Each Q-band HA is connected with a 60-GHz EA with 30-dB gain and 24-dBm saturation output power, while each W-band transmitter HA connected with a 100-GHz EA with 30-dB gain and 10-dBm saturation output power. The 3-dB beam width at the input of each receiver HA is about 10° × 10°. The transmitted and received RF power is about 9dBm and −10dBm, respectively. The isolation between H- and V-polarization HA array is 33dB. Thus, the crosstalk between H- and V-polarization signals can be ignored and other arrangements apart from H-V-V-H and V-H-H-V are also feasible for the HA system.
For the received 100-GHz PDM-QPSK wireless signal (corresponding to ch3 or ch4), analog down conversion is firstly implemented at the wireless receiver, which is identical to that adopted in Fig. 4. The 28-GHz IF signals after first-stage analog down conversion are amplified by two 40-GHz EAs, and then sent to the digital storage OSC with 120-GSa/s sampling rate and 45-GHz electrical bandwidth. The subsequent digital DSP, similar to that adopted in Fig. 4, includes IF down conversion, CD compensation, CMA equalization, carrier recovery, differential decoding and BER counting. Here, CMA equalization with 53 T/2 taps is used to realize polarization de-multiplexing and remove the crosstalk due to wireless delivery [25]. Analog down conversion is unnecessary for the received 37.5-GHz PDM-16QAM wireless signal (corresponding to ch1 or ch2), which is amplified by 60-GHz EAs and then directly sent to the real-time OSC with 120-GSa/s sampling rate and 45-GHz electrical bandwidth. The subsequent DSP, similar to that adopted in Fig. 5, includes IF down conversion, CD compensation, cascaded multi-modulus algorithm (CMMA) equalization, carrier recovery, differential decoding and BER counting. Here, CMMA equalization is used to realize multi-modulus recovery and polarization de-multiplexing. Figure 11(i) shows the X-polarization constellation for the 108-Gb/s PDM-QPSK signal after 80-km SMF-28 transmission and 0.7-m wireless delivery when the OSNR is 36dB and the CMA-tap number is 53. Figure 11(j) shows the X-polarization constellation for the 112-Gb/s PDM-16QAM signal after 80-km SMF-28 transmission and 1.5-m wireless delivery at the BER of 9 × 10−4 and 33-dB OSNR.
3. Problems existing in the large capacity fiber wireless integration system and corresponding solutions
3.1 Wireless multi-path effect due to different wireless transmission distance
Due to the wireless multi-path effect, wireless interference exists in the MIMO wireless transmission. We experimentally investigate the MIMO wireless interference in a 100-GHz optical wireless integration system, which can deliver 50-Gb/s PDM-QPSK signal over 80-km SMF-28 and a 2 × 2 MIMO wireless link [25]. For the parallel MIMO wireless link, each receiver HA can only get the wireless power from corresponding transmitter HA, and thus there is no wireless interference. However, for the cross-over ones, receiver HA can get the wireless power from two transmitter HAs, and thus there exists wireless interference. Polarization de-multiplexing is realized by CMA based on DSP at the wireless receiver. Compared to the parallel case, about 2-dB OSNR penalty at the BER of 3.8 × 10−3 is caused by the wireless interference for the cross-over cases if similar CMA taps are employed. The increase of tap length can reduce the wireless interference and improve the BER performance. More taps should be adopted when two pairs of transmitter and receiver HAs have different wireless distance.
The experimental setup for the 100-GHz 50-Gb/s PDM-QPSK optical wireless integration system is quite similar to that shown in Fig. 4, except that the I/Q modulator at the optical baseband transmitter is driven by a 12.5-Gbaud electrical binary signal with a PRBS length of 215-1, and the 2 × 2 MIMO wireless link has a deliberate HA arrangement in order to investigate the MIMO wireless interference. Figure 12 shows one parallel and three cross-over MIMO wireless links by fixing the locations of two receiver HAs and adjusting those of two transmitter HAs. There exists an 80-km SMF-28 transmission with 2-dBm launched power into fiber (after EDFA), and two receiver HAs have a 10-cm wireless distance. The 3-dB beam width at the input of the receiver HA is 40° × 40°. For Case 1 shown in Fig. 12(a), each pair of transmitter and receiver HAs have a 0.6-m wireless distance and two transmitter (receiver) HAs have a 10-cm wireless distance. There is no wireless interference due to its high directionality. For Case 2 shown in Fig. 12(b), two transmitter HAs are both moved to the location that has the same wireless distance from two receiver HAs, and each pair of transmitter and receiver HAs have a 0.6-m wireless distance horizontally. Thus, each receiver HA can get the same wireless power from two transmitter HAs. For Case 3 shown in Fig. 12(c), transmitter HA1 is fixed, transmitter HA2 is moved to the location that has the same wireless distance from two receiver HAs, and each pair of transmitter and receiver HAs have a 0.6-m wireless distance horizontally. Thus, receiver HA2 can only get the wireless power from transmitter HA2, whereas receiver HA1 can get the same wireless power from two transmitter HAs. For Case 4 shown in Fig. 12(d), two transmitter HAs are both moved to the locations that have the same wireless distance from two receiver HAs, and the horizontal wireless distance is 0.6 m between transmitter HA2 and receiver HA2 while 0.8 m between transmitter HA1 and receiver HA1. Thus, each receiver HA can also get the same wireless power from two transmitter HAs. The block effect of transmitter HA2 is removed by locating transmitter HA2 a little lower than transmitter HA1, that is, two transmitter HAs have different height on the optical table. Because receiver HA can get the wireless power from two transmitter HAs, there exists wireless interference for these cross-over cases.
Figure 12(e) gives BER versus OSNR for Cases 1 to 4. The CMA tap number is 19 for Cases 1 to 3, while 27 and 35 for Case 4 because the BER performance is very poor if only 19 taps are used. The reason more taps are required for Case 4 is that two pairs of transmitter and receiver HAs have different wireless distance, which is equivalent to a large differential group delay (DGD) effect existing in the transmission fiber. The wireless interference in Case 4 can be almost removed completely when the tap number is 35. Compared to Case 1, only about 2-dB OSNR penalty is caused at the BER of 3.8 × 10−3 by the wireless interference for Cases 2 to 4 if similar CMA taps are employed, which shows the uniform equalization of the wired and wireless integrated 2 × 2 MIMO channel can be well realized by classic CMA.
Relative to Case 1, the optimal additional tap number in Case 4 can be calculated as follows
Where n is the medium index ( = 1 in the air), l is the wireless distance difference between two pairs of transmitter and receiver HAs, c is the optical speed in vacuum, and b is the baud rate. The optimal tap number for Case 4 will be about 19 + 16 = 35 if that for Case 1 is 19. When the adopted tap number is 35 for Case 4, we can get the BER curve similar to Case 1 as shown in Fig. 12(e). Thus, the increase of tap length can overcome the wireless interference and improve the BER performance. More CMA taps should be adopted when two pairs of transmitter and receiver HAs have different wireless distance, and the interference can be almost removed completely if the tap number is large enough.
3.2 Advance algorithms based on DSP
In the large capacity fiber wireless integration system, the signal quality degradation due to the limited bandwidth of optical and electrical components can be pre-compensated by high-speed digital-to-analog converter (DAC) based on advanced DSP [50]. Furthermore, if the bandwidth of optical and electrical components is narrow enough, the QPSK signal will be converted into 9QAM-like one, and thus the digital post filter combined with 1-bit maximum likelihood sequence estimation (MLSE) [51, 52] can be adopted to improve the signal quality. On the other hand, the high peak-to-average power ratio (PAPR) characteristic of OFDM modulation will also degrade the performance of the fiber wireless integration system [53], for which discrete Fourier transform (DFT) [54–57] can be introduced to reduce PAPR and thus improve the system performance. Many related investigations are still being carried out now.
4. Conclusion
We summarize several different approaches for the realization of large capacity (>100Gb/s) fiber wireless integration system, including optical PDM combined with MIMO reception, advanced multi-level modulation, optical multi-carrier modulation, electrical multi-carrier modulation, antenna polarization multiplexing and multi-band multiplexing. These approaches can effectively reduce the signal baud rate as well as the required bandwidth for optical and electrical devices. We also investigate the problems, such as multi-path effect due to different wireless transmission distance, existing in the large capacity fiber wireless integration system. We demonstrate these problems can be effectively solved based on advanced DSP algorithms including classic CMA. Moreover, based on the combination of these approaches as well as advanced DSP algorithms, we have successfully demonstrated a 400G fiber wireless integration system, which creates a capacity record of wireless delivery and ushers in a new era of ultra-high bit rate (>400Gb/s) optical wireless integration communications at mm-wave frequencies. We believe, in the near future, 1Tb/s and beyond fiber wireless integration system can be realized either by adopting more optical/electrical carriers, more frequency bands, higher-level modulation and so on, or by introducing more emerging techniques, such as orbital angular momentum (OAM).
Acknowledgments
Authors would like to thank Ze Dong, Junwen Zhang and Fan Li from ZTE USA for their help in finishing a part of experiments, and Prof. Gee-Kung Chang from Georgia Institute of Technology for his encouragement. This work was partially supported by NNSF of China (61250018) and NKTR&DP of China (2012BAH18B00).
References and links
1. M. Fitzsimmons, “Sharp dazzles with 8K TV prototype, VP says production about 4 years off” (CES 2013). http://www.techradar.com/us/news/television/hdtv/sharp-dazzles-with-8k-tv-prototype-vp-says-production-about-4-years-off-1124330.
2. C. H. Chang, P. C. Peng, H. H. Lu, C. L. Shih, and H. W. Chen, “Simplified radio-over-fiber transport systems with a low-cost multiband light source,” Opt. Lett. 35(23), 4021–4023 (2010). [CrossRef] [PubMed]
3. J. Yu, G. K. Chang, Z. Jia, A. Chowdhury, M. F. Huang, H. C. Chien, Y. T. Hsueh, W. Jian, C. Liu, and Z. Dong, “Cost-effective optical millimeter technologies and field demonstrations for very high throughput wireless-over-fiber access systems,” J. Lightwave Technol. 28(16), 2376–2397 (2010). [CrossRef]
4. X. Pang, A. Caballero, A. Dogadaev, V. Arlunno, R. Borkowski, J. S. Pedersen, L. Deng, F. Karinou, F. Roubeau, D. Zibar, X. Yu, and I. T. Monroy, “100 Gbit/s hybrid optical fiber-wireless link in the W-band (75-110 GHz),” Opt. Express 19(25), 24944–24949 (2011). [CrossRef] [PubMed]
5. A. Kanno, K. Inagaki, I. Morohashi, T. Sakamoto, T. Kuri, I. Hosako, T. Kawanishi, Y. Yoshida, and K. I. Ki-tayama, “40 Gb/s W-band (75-110 GHZ) 16-QAM radio-over-fiber signal generation and its wireless transmission,” Proc. ECOC2011, Geneva, Switzerland, We.10.P1.112.
6. C. W. Chow, F. M. Kuo, J. W. Shi, C. H. Yeh, Y. F. Wu, C. H. Wang, Y. T. Li, and C. L. Pan, “100 GHz ultra-wideband (UWB) fiber-to-the-antenna (FTTA) system for in-building and in-home networks,” Opt. Express 18(2), 473–478 (2010). [CrossRef] [PubMed]
7. D. Zibar, R. Sambaraju, A. Caballero, J. Herrera, U. Westergren, A. Walber, J. B. Jensen, J. Martí, and I. T. Monroy, “High-capacity wireless signal generation and demodulation in 75- to 110-GHz band employing all-optical OFDM,” Photon. Technol. Lett. 23(12), 810–812 (2011). [CrossRef]
8. D. Zibar, C. Antonio, X. Yu, X. Pang, A. K. Dogadaev, and I. T. Monroy, “Hybrid optical fibre-wireless links at the 75-110 GHz band supporting 100 Gbps transmission capacities,” Proc. MWP/APMP 2011, Singapore, 445–449(2011). [CrossRef]
9. C. Ye, L. Zhang, M. Zhu, J. Yu, S. He, and G. K. Chang, “A bidirectional 60-GHz wireless-over-fiber transport system with centralized local oscillator service delivered to mobile terminals and base stations,” IEEE Photon. Technol. Lett. 24(22), 1984–1987 (2012). [CrossRef]
10. N. Cvijetic, D. Qian, J. Yu, Y. K. K. Huang, and T. Wang, “Polarization-multiplexed optical wireless transmission with coherent detection,” J. Lightwave Technol. 28(8), 1218–1227 (2010). [CrossRef]
11. N. Ghazisaidi and M. Maier, “Fiber-wireless (FiWi) access networks: Challenges and opportunities,” IEEE Netw. 25(1), 36–42 (2011). [CrossRef]
12. Z. Jia, J. Yu, A. Agarwal, H. C. Chien, A. Chowdhury, G. Ellinas, J. L. Jackel, and G. K. K. Chang, “Photonic generation and processing technologies for converged ultra-high throughput fiber-wireless systems,” Proc. OFC/NFOEC2010, San Diego, California, OTuF6. [CrossRef]
13. A. Lebedev, T. T. Pham, M. Beltrán, X. Yu, A. Ukhanova, L. Deng, N. G. González, R. Llorente, I. T. Monroy, and S. Forchhammer, “Optimization of high-definition video coding and hybrid fiber-wireless transmission in the 60 GHz band,” Proc. ECOC2011, Geneva, Switzerland, We.10.P1.97.
14. A. K. Dogadaev and I. T. Monroy, “Challenges and capacity analysis of 100 Gbps optical fibre wireless links in 75-110 GHz band,” IEEE Photonics Conference, Arlington, VA, 268–269 (2011). [CrossRef]
15. A. H. M. R. Islam, M. Bakaul, A. T. Nirmalathas, and G. E. Town, “Simplified generation, transport, and data recovery of millimeter-wave signal in a full-duplex bidirectional fiber-wireless system,” Photon. Technol. Lett. 24(16), 1428–1430 (2012). [CrossRef]
16. S. Z. Pinter and X. N. Fernando, “Estimation and equalization of fiber-wireless uplink for multiuser CDMA 4G networks,” IEEE Trans. Commun. 58(6), 1803–1813 (2010). [CrossRef]
17. C. Lim, A. T. Nirmalathas, M. Bakaul, P. A. Gamage, K.-L. L. Lee, Y. M. Yang, D. Novak, and R. B. Waterhouse, “Fiber-wireless networks and subsystem technologies,” J. Lightwave Technol. 28(4), 390–405 (2010). [CrossRef]
18. X. Pang, A. Caballero, A. K. Dogadaev, V. Arlunno, L. Deng, R. Borkowski, J. S. Pedersen, D. Zibar, X. Yu, and I. T. Monroy, “25 Gbit/s QPSK Hybrid Fiber-Wireless Transmission in the W-Band (75-110 GHz) With Remote Antenna Unit for In-Building Wireless Networks,” Photon. J. 4(3), 691–698 (2012). [CrossRef]
19. C. T. Lin, A. Ng’oma, W. Y. Lee, C. C. Wei, C. Y. Wang, T. H. Lu, J. Chen, W. J. Jiang, and C. H. Ho, “2 × 2 MIMO radio-over-fiber system at 60 GHz employing frequency domain equalization,” Opt. Express 20(1), 562–567 (2012). [CrossRef] [PubMed]
20. Y. Zhao, X. Pang, L. Deng, M. B. Othman, X. Yu, X. Zheng, H. Y. Zhang, and I. T. Monroy, “Experimental demonstration of 5-Gb/s polarization-multiplexed fiber-wireless MIMO systems,” Proc. MWP/APMP 2011, Singapore, 13–16 (2011). [CrossRef]
21. A. Kanno, T. Kuri, I. Hosako, T. Kawanishi, Y. Yasumura, Y. Yoshida, and K. Kitayama, “Optical and radio seamless MIMO transmission with 20-Gbaud QPSK,” Proc. ECOC2012, Amsterdam, The Netherlands, We.3.B.2. [CrossRef]
22. L. Tao, Z. Dong, J. Yu, N. Chi, J. Zhang, X. Li, Y. Shao, and G. K. Chang, “Experimental demonstration of 48-Gb/s PDM-QPSK radio-over-fiber system over 40-GHz mm-wave MIMO wireless transmission,” Photon. Technol. Lett. 24(24), 2276–2279 (2012). [CrossRef]
23. X. Li, J. Yu, Z. Dong, Z. Cao, N. Chi, J. Zhang, Y. Shao, and L. Tao, “Seamless integration of 57.2-Gb/s signal wireline transmission and 100-GHz wireless delivery,” Opt. Express 20(22), 24364–24369 (2012). [CrossRef] [PubMed]
24. X. Li, Z. Dong, J. Yu, N. Chi, Y. Shao, and G. K. Chang, “Fiber wireless transmission system of 108-Gb/s data over 80-km fiber and 2×2 MIMO wireless links at 100GHz W-Band frequency,” Opt. Lett. 37(24), 5106–5108 (2012). [CrossRef] [PubMed]
25. X. Li, J. Yu, Z. Dong, J. Zhang, N. Chi, and J. Yu, “Investigation of interference in multiple-input multiple-output wireless transmission at W band for an optical wireless integration system,” Opt. Lett. 38(5), 742–744 (2013). [CrossRef] [PubMed]
26. Z. Dong, J. Yu, X. Li, G. K. Chang, and Z. Cao, “Integration of 112-Gb/s PDM-16QAM wireline and wireless data delivery in millimeter wave RoF system,” Proc. OFC2013, Anaheim, California, OM3D.2. [CrossRef]
27. T. Kuri, H. Toda, J. J. V. Olmos, and K. I. Kitayama, “Fiber-wireless DWDM networks and radio-over-fiber technologies,” Proc. MWP 2009, Valencia, 1–4 (2009).
28. M. Luizink, C. A. M. Steenbergen, Koonen, and M. J. T. Antonius, “Capacity allocation using WDM in fiber-wireless access networks,” Proc. OFC/IOOC 1999, San Diego, California, 61–63 (1999).
29. A. T. Nirmalathas, C. Lim, D. Novak, D. Castleford, R. B. Waterhouse, and G. H. Smith, “Millimeter-wave fiber-wireless access systems incorporating wavelength division multiplexing,” Microwave Conference 2000, 625–629 (2000). [CrossRef]
30. J. Zhang, J. Yu, N. Chi, Z. Dong, X. Li, and G.-K. Chang, “Multichannel 120-Gb/s data transmission over 2x2 MIMO fiber-wireless link at W-band,” Photon. Technol. Lett. 25(8), 780–783 (2013). [CrossRef]
31. L. Deng, M. Beltrán, X. Pang, X. Zhang, V. Arlunno, Y. Zhao, A. Caballero, A. K. Dogadaev, X. Yu, R. Llorente, D. Liu, and I. T. Monroy, “Fiber wireless transmission of 8.3-Gb/s/ch QPSK-OFDM signals in 75-110-GHz band,” Photon. Technol. Lett. 24(5), 383–385 (2012). [CrossRef]
32. Z. Cao, L. Zou, L. Chen, and J. Yu, “Impairment mitigation for a 60 GHz OFDM radio-over-fiber system through an adaptive modulation technique,” J. Opt. Commun. Netw. 3(9), 758–766 (2011). [CrossRef]
33. J. Yu, J. Hu, D. Qian, Z. Jia, G. K. K. Chang, and T. Wang, “16 Gbit/s super broadband OFDM-radio-over-fibre system,” Electron. Lett. 44(6), 450–451 (2008). [CrossRef]
34. Z. Cao, J. Yu, M. Xia, Q. Tang, Y. Gao, W. Wang, and L. Chen, “Reduction of intersubcarrier interference and frequency-selective fading in OFDM-ROF systems,” J. Lightwave Technol. 28(16), 2423–2429 (2010). [CrossRef]
35. C. Liu, H. C. Chien, Z. Gao, W. Jian, A. Chowdhury, J. Yu, and G. K. K. Chang, “Multi-band 16QAM-OFDM vector signal delivery over 60-GHz DSB-SC optical millimeter-wave through LO enhancement,” Proc. OFC/NFOEC2011, Los Angeles, California, OTHJ2. [CrossRef]
36. L. Tao, J. Yu, Q. Yang, Y. Shao, J. Zhang, and N. Chi, “A novel transform domain processing based channel estimation method for OFDM radio-over-fiber systems,” Opt. Express 21(6), 7478–7487 (2013). [CrossRef] [PubMed]
37. F. Li, Z. Cao, X. Li, Z. Dong, and L. Chen, “Fiber-wireless transmission system of PDM-MIMO-OFDM at 100 GHz frequency,” J. Lightwave Technol. 31(14), 2394–2399 (2013). [CrossRef]
38. X. Li, J. Yu, J. Zhang, Z. Dong, and N. Chi, “Doubling transmission capacity in optical wireless system by antenna horizontal- and vertical-polarization multiplexing,” Opt. Lett. 38(12), 2125–2127 (2013). [CrossRef] [PubMed]
39. A. Chowdhury, H. C. Chien, S. H. Fan, J. Yu, and G. K. K. Chang, “Multi-band transport technologies for in-building host-neutral wireless over fiber access systems,” J. Lightwave Technol. 28(16), 2406–2415 (2010). [CrossRef]
40. Y. T. Hsueh, Z. Jia, H. C. Chien, A. Chowdhury, J. Yu, and G. K. K. Chang, “Multiband 60-GHz wireless over fiber access system with high dispersion tolerance using frequency tripling technique,” J. Lightwave Technol. 29(8), 1105–1111 (2011). [CrossRef]
41. X. Li, J. Yu, J. Zhang, Z. Dong, F. Li, and N. Chi, “A 400G optical wireless integration delivery system,” Opt. Express 21(16), 18812–18819 (2013). [CrossRef] [PubMed]
42. J. Yu, Z. Jia, and G. K. Chang, “Optical millimeter wave generation or up-conversion using external modulator,” IEEE Photon. Technol. Lett. 18(1), 265–267 (2006). [CrossRef]
43. 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(16), 1726–1728 (2006). [CrossRef]
44. C. T. Lin, A. Ng’oma, W. Y. Lee, C. C. Wei, C. Y. Wang, T. H. Lu, J. Chen, W. J. Jiang, and C. H. Ho, “2 × 2 MIMO radio-over-fiber system at 60 GHz employing frequency domain equalization,” Opt. Express 20(1), 562–567 (2012). [CrossRef] [PubMed]
45. ITU-T Recommendation G.975.1, “Forward error correction for high bit-rate DWDM submarine system,” 2004.
46. J. Zhang, Z. Dong, J. Yu, N. Chi, L. Tao, X. Li, and Y. Shao, “Simplified coherent receiver with heterodyne detection of eight-channel 50 Gb/s PDM-QPSK WDM signal after 1040 km SMF-28 transmission,” Opt. Lett. 37(19), 4050–4052 (2012). [CrossRef] [PubMed]
47. Z. Dong, X. Li, J. Yu, and N. Chi, “6×128-Gb/s Nyquist-WDM PDM-16QAM generation and transmission over 1200-km SMF-28 with SE of 7.47b/s/Hz,” J. Lightwave Technol. 30(24), 4000–4005 (2012). [CrossRef]
48. Z. Dong, X. Li, J. Yu, and J. Yu, “Generation and transmission of 8 × 112-Gb/s WDM PDM-16QAM on a 25-GHz grid with simplified heterodyne detection,” Opt. Express 21(2), 1773–1778 (2013). [CrossRef] [PubMed]
49. C. Liu, H. C. Chien, S. H. Fan, J. Yu, and G. K. Chang, “Enhanced vector signal transmission over double-sideband carrier-suppressed optical millimeter-waves through a small LO feedthrough,” Photon. Technol. Lett. 24(3), 173–175 (2012). [CrossRef]
50. X. Li, J. Yu, Z. Dong, J. Zhang, L. Tao, N. Chi, and Y. Shao, “Performance improvement by pre-equalization in W-band (75-110GHz) RoF system,” Proc. OFC2013, Anaheim, California, OW1D.3. [CrossRef]
51. X. Li, J. Yu, N. Chi, Z. Dong, J. Zhang, and J. Yu, “The reduction of the LO number for heterodyne coherent detection,” Opt. Express 20(28), 29613–29619 (2012). [CrossRef] [PubMed]
52. X. Li, Z. Dong, J. Yu, J. Yu, and N. Chi, “Heterodyne coherent detection of WDM PDM-QPSK signals with spectral efficiency of 4b/s/Hz,” Opt. Express 21(7), 8808–8814 (2013). [CrossRef] [PubMed]
53. F. Li, J. Yu, Z. Cao, J. Xiao, H. Chen, and L. Chen, “Reducing the peak-to-average power ratio with companding transform coding in 60 GHz OFDM-ROF systems,” J. Opt. Commun. Netw. 4(3), 202–209 (2012). [CrossRef]
54. L. Tao, J. Yu, Y. Fang, J. Zhang, Y. Shao, and N. Chi, “Analysis of noise spread in optical DFT-S OFDM systems,” J. Lightwave Technol. 30(20), 3219–3225 (2012). [CrossRef]
55. Q. Yang, Z. He, Z. Yang, S. Yu, X. Yi, and W. Shieh, “Coherent optical DFT-spread OFDM transmission using orthogonal band multiplexing,” Opt. Express 20(3), 2379–2385 (2012). [CrossRef] [PubMed]
56. Y. Tang, W. Shieh, and B. S. Krongold, “DFT-Spread OFDM for fiber nonlinearity mitigation,” Photon. Technol. Lett. 22(16), 1250–1252 (2010). [CrossRef]
57. L. Tao, J. Yu, Q. Yang, M. Luo, Z. He, Y. Shao, J. Zhang, and N. Chi, “Spectrally efficient localized carrier distribution scheme for multiple-user DFT-S OFDM RoF- PON wireless access systems,” Opt. Express 20(28), 29665–29672 (2012). [CrossRef] [PubMed]
