We present a high-order UWB pulses generator based on a microwave photonic filter which provides a set of positive and negative samples by using the slicing of an incoherent optical source and the phase inversion in a Mach-Zehnder modulator. The simple scalability and high reconfigurability of the system permit a better accomplishment of the FCC requirements. Moreover, the proposed scheme permits an easy adaptation to pulse amplitude modulation, bi phase modulation, pulse shape modulation and pulse position modulation. The flexibility of the scheme for being adaptable to multilevel modulation formats permits to increase the transmission bit rate by using hybrid modulation formats.
© 2013 Optical Society of America
Ultra-Wideband (UWB) technology has received a tremendous attention for a wide variety of commercial applications such as wireless communications, sensor networks or radar imaging. UWB is attractive due to characteristics such as its ability to overlay existing band occupants without interfering, potentially high data rates, its high temporal resolution and the capability to penetrate through the obstacles. For wireless communications, the US Federal Communications Commission (FCC) defines the use without license of the spectrum from 3.1 to 10.6 GHz with a restriction in power spectral density of −41.3 dBm/MHz . This fact involves that UWB systems can only cover a short signal transmission distance of less than tens of meters. In this point, the use of optical fiber for UWB signals distributions emerge as a promising solution to increase the coverage area what is known as UWB-over-fiber. Therefore, photonic UWB signals generation is attractive since they can profit from optical components involved in the transmission process .
Many photonic approaches have been reported for generating UWB pulses. We can found schemes to generate classical UWB pulses, monocycle and doublet, based on phase-modulation-to-intensity-modulation (PM-IM) conversion by using a dispersive device  or an optical frequency discriminator implemented by optical filters , polarization-dependent interferometer  and chirped Fiber Bragg Gratings (C-FBG) . Microwave photonic filters have been also proposed for classical UWB signals generation implementing negative taps based on cross-gain modulation in Semiconductor Optical Amplifiers (SOA) , the nonlinear amplification process in a SOA , the combination of the effects of cross-phase modulation and birefringence in a As2S3 rib waveguide  and, also, based on a phase modulator and an asymmetric Mach-Zehnder Interferometer (AMZI) . In order to achieve a more efficient use of the FCC available spectrum, high order pulses have been proposed . In  the generation of third-order pulses is proposed based on four-wave mixing and PM-IM conversion. High-order pulses FCC-compliant has been also experimentally demonstrated using Fiber Bragg Gratings (FBG) with a specific design for optical spectral shaping and frequency-to-time mapping by single mode optical fiber . However, the experimental conditions to obtain the pulses are fixed and reconfigurability is not easy to achieve. In order to increase the flexibility in UWB pulses generation, an N-tap microwave photonic filter has been proposed using phase inversion in electro-optical modulators for obtaining positive and negative taps. Nevertheless, in that case, each coefficient required an additional laser source . Also, a fifth order Gaussian pulses and other linear combination of Gaussian derivatives pulses have been proposed in the literature focused on power spectral efficient pulses  .Apart from an efficient use of the FCC spectral mask, optical UWB pulse generators have a key challenge related to the possibility of pulse codification using different modulation formats such as Pulse Position Modulation (PPM), Bi-Phase Modulation (BPM), Pulse Amplitude Modulation (PAM) and Pulse Shape Modulation (PSM) [1, 11]. Most of the proposals carried out are focused on the implementation of a binary codification of these modulation formats [3–12]. However, as it is well known multilevel modulation formats permit to achieve a more efficient use of the available transmission bandwidth. In the literature, a few numbers of proposals carry out multilevel modulation formats. Recently, a hybrid modulation format has been proposed which permits to increase the transmission bit rate by a factor of about 4 .
In this paper, we propose a high-order UWB pulses generator based an N tap microwave photonic filter fed by different optical signals and featuring positive and negative sample polarity by using phase inversion in a Mach-Zenhder modulator. A high number of taps is achieved by the slicing of a broadband source and they can be controlled independently in terms of amplitude and polarity. Preliminary experimental results of this scheme have been presented in . In this manuscript, we include further experimental measurements and the corresponding theoretical analysis showing a good agreement between experimental and theoretical results obtained by numerical simulations. Moreover, the flexibility of the system allows the adaptation of pulse amplitude modulation (PAM), bi phase modulation (BPM), pulse shape modulation (PSM) and pulse position modulation (PPM). We analyze the potential to increase the transmission bit rate by implementing multilevel modulation formats through hybrid modulation .
2. Principle of operation
The experimental scheme of the system proposed for generating high-order UWB pulses is shown in Fig. 1. It corresponds to an N-tap microwave photonic filter whose spectral properties have been considered in . The principle of the proposed system is based on the use of two electro-optic modulators (EOM1 and EOM2) biased in regions with opposite slopes and modulated by the same electrical pulse using a RF splitter.
Different optical subcarriers are obtained by the slicing of a broadband source using a structure based on arrayed waveguide gratings (AWG). After the first AWG, the optical broadband spectra is sliced, each optical output channel goes through an optical attenuator followed by an optical switch, which state selects the EOM employed for every optical channel. All the channels modulated by a given EOM are multiplexed by the second concatenated AWG just before being launched into it. As can be observed, different number of AWGs and EOMs are used in this structure which can lead to a complicated implementation. Nevertheless, the advances in integrated photonic circuits design permit to achieve low-cost, -size, -weight solutions with reduced power consumption .
As it is well known, the use of incoherent optical sources introduces larger noise than using coherent sources  which could influence the spectral power efficiency of the UWB pulses after long-distance transmission over fiber. However, this limitation is overcome through proper averaging of the output waveforms . In this way, incoherent optical sources can be exploited to increase considerably the number of optical taps as reported in  and . Therefore, the proposed approach shows more capacity in terms of optical taps numbers compared to previous solutions based on a set of laser sources as reported in .
According to the optical response of the AWGs, the spectral density of each optical subcarrier is given by a gaussian profile:Fig. 1 shows the optical spectrum of the concatenated channel which has 0.31 nm bandwidth for the channel centered at 1546.92 nm.
The output of AWG1 and AWG2 is launched into the electro-optical modulators EOM1 and EOM2 by means of an optical switch. Each EOM are biased to operate with opposite slope and they are fed with an electrical signal coming from an electrical pulse generator. The bias voltage applied to each EOM is 1.2 and 6.3 V since the pi-voltage Vπ of the EOMs is around 5 V. The amplitude of the driving electrical pulse is 2 V to operate in a quasi linear region. In this way, the output pulse from the EOM is proportional to the electrical pulse since the amplitude of the driving pulse and the bias point do not have any effect on it. The modulated signals coming from both EOMs are coupled and launched into dispersive element that introduces a different delay between each channel according to its central wavelength. Finally, the optical signal is photodetected.
Following  and , in this case we find that the electrical transfer function HRF(Ω) of the system which represents an equivalent microwave photonic filter, is given by the following expression:Fig. 1.
The second term in Eq. (2) corresponds to the time spreading due to the finite linewidth of the AWGs channels. This term introduces a low passband effect in the system, which can be neglected when the condition β2Lδω/2σo<<1 is satisfied which involves the dispersion (β2) and propagation distance (L), AWG channels optical bandwidth (δω) and pulse width (σo).
The performance of the proposed generator is analyzed by selecting the EOM1 or EOM2 when only a channel centered at 1546.92 nm is employed. Each EOM is fed with an electrical signal coming from an electrical pulse generator with a fixed pattern of one “1” and sixtythree “0” (total 64 bits) and 12.5 Gb/s bit rate. The electrical pulse width is σo = 80 ps. Figure 2 shows the normalized optical pulses corresponding to either EOM1 or EOM2 after propagating through the SMF link. A positive and negative polarity pulse is obtained depending on the bias point of the modulator. We use a 5.43 km standard SMF-28 fiber link with a total dispersion of 93 ps/nm and the channel has a bandwidth of 0.31 nm at the output of AWG1 and AWG2 satisfying adequately the previous condition (β2Lδω/2σo<<1). In principle, the UWB generated pulses could be degraded if the dispersion were not controlled. Indeed, the optical delay between optical taps which determines the FSR of the equivalent microwave photonic filter depends on the dispersion (β2L) and the wavelength separation between channels as shown previously. However, this limitation is overcome provided compensating dispersion modules are included in scenarios with different fibre lengths.
3. Experimental capabilities of the system
3.1 High-order Impulse Radio UWB pulses generation
In order to evaluate the capabilities of the system, different waveforms, corresponding to the derivatives of a Gaussian pulse , have been obtained experimentally as shown Fig. 3. In this case, we can observe that high order pulses become an interesting solution since they are more efficient to fulfill the FCC mask in terms of power spectral density .
The delay of each optical sample is a key point in the system design for generating a given waveform. From the Eq. (2), the transfer function of the equivalent microwave photonic filter is defined by the delay suffered for each optical sample (τn), As it is known, for Finite Impulse Response (FIR) filters, the Free Spectral Range (FSR) is given by the difference delay between consecutive samples (Δτ = τn - τn-1) :
In our proposal, the output RF signal is given by the input pulse and the corresponding electrical transfer function. The ideal scenario is that the electrical transfer function determines completely the waveform of the output pulse. Therefore, the electrical bandwidth of the driving pulse should be higher than the frequency operation range where the pulse spectra is located. In this way, the condition 1/σo >> FSR/2 implies driving pulse does not limit the output waveform which is conformed entirely by the microwave photonic filtering. Therefore, the width σo of the driving pulse should be lower than the optical delay 2Δτ. In our approach, the electrical bandwidth of the pulse is around 1/σo ~12.5 GHz and the FSR/2 is close to 6.8 GHz.
Firstly, the power of two channels is adjusted with identical amplitude as can be seen in Fig. 3(a). The generated pulse and the corresponding electrical spectrum are shown in Fig. 3(b) and Fig. 3(c), respectively, which corresponds to a monocycle waveform. The next implemented waveform is a doublet pulse which involves the use of three channels with a given optical power [Fig. 3(d)] and switches state to give the following equivalent coefficients [0.5,-1, 0.5]. Figures 3(e) and 3(f) show the doublet pulse shape and its electrical power spectrum, respectively. As can be observed from the electrical power spectra, the monocycle and doublet pulses do not fit the FCC mask, especially around 0.96-1.61 GHz band. This fact involves the need of reducing the signal power spectral density for fulfilling the FCC mask. In order to show the flexibility of this structure to increase the order of the pulses we generate a third and fourth order UWB pulses which are known as triplet and quadruplet. In the first case, four channels are used adjusting the variable optical attenuators [Fig. 3(g)] and the state of the switches to an equivalent filter of [-0.3, 1,-1, 0.3]. The waveform obtained and its electrical power spectrum is plotted in Figs. 3(h) and 3(i), respectively. As can be observed, a nearly fulfillment of the spectral requirements is achieved. Finally, a quadruplet pulse has been implemented using five channels [Fig. 3(j)] and a configuration of the system for obtaining the coefficients [-0.1, 0.6,-1, 0.6, −0.1]. The pulse shape resulting is shown in Fig. 3(k) and the corresponding electrical power spectrum in Fig. 3(l). Note as in this case the frequency band in 0.96-1.61 GHz is free from signal content fulfilling the FCC spectral requirements only restricted by noise. For the different waveforms, the transfer function of the equivalent microwave photonic filter has been included in the spectra graphs. As can be observed, the spectrum of the different pulses is shaped according to this transfer function. Therefore, we have experimentally demonstrated the flexibility of our system for generating high-order UWB waveforms. This fact permits to fit the pulses to the FCC mask avoiding the need of reducing the signal power in order to achieve an efficient solution in terms of power spectral density. As can be observed from Fig. 3, the central frequency of the generated waveforms is around 5.5 GHz. A displacement from the central frequency of the FCC mask is due to the spectral response of the electrical pulse. The signal bandwidth changes from 7.5 to 5.9 GHz for the monocycle and quadruplet, respectively. Regarding to the spectral power efficiency, a numerical estimation has been realized from Fig. 3 according to reference . As observed from each signal spectrum, with higher order waveforms the best fulfillment of the FCC mask is achieved. Therefore, the spectral efficiency is improved for high-order pulses improving the performance of UWB technology. We find a power efficiency around 0.2, 2.1, 33.0 and 40% for monocycle, doublet, triplet and quadruplet, respectively.
Certainly, our experimental results correspond to the 4th derivative of Gaussian pulse. We cannot achieve higher order pulses due to the limitation of feasibilities in terms of attenuation and optical switching. In order to increase the power efficiency by means of higher order derivatives, as proposed in previous references , our proposal only requires the addition of more channels in the AWG structure. In such scheme, the power of each channel is adjusted independently, according to the amplitude of the corresponding derivative.
Finally, we can observe that UWB pulses of Fig. 3 show a relatively large noise due to the use of incoherent optical sources related to the high signal-to-noise ratio. This limitation can be overcome through a proper average of the output waveforms as previously mentioned .
3.2 Multilevel modulation formats
As pointed out in the introduction, most of UWB generation proposals in the literature are focused on implementing binary modulation formats as PAM, BPM, PSM and PPM [3–12]. In our approach, the flexibility of the system allows us to extend the UWB signals generator proposed to a multilevel version of these modulation formats.
In the proposed architecture, PAM with different levels can be implemented easily by means of variable optical attenuators which provide the amplitude of each optical tap. Using this scheme, BPM can be easily achieved by two different methods. Pulse polarity inversion can be achieved by changing the modulator bias in an opposite linear slope region. Also, BPM can be implemented by setting an opposite state for each optical switch in order to invert the input modulator for each optical tap. Furthermore, the control of optical tap amplitude by means of the attenuators and the states of the optical switches can be exploited for alternatively selecting the UWB pulse shapes. For example, 4-levels PSM have been carried out experimentally in Fig. 4(a). The possibility of generating high order pulses allows us to achieve 4 different waveforms for PSM format. On the other hand, the relationship between the channels wavelength and waveforms delay by the fiber dispersion permits to control the relative position of pulses in the time domain. In this way, depending on the selected channels we can control the delay of pulses as it is shown in Fig. 4(b) where a monocycle is moved a multiple of a base delay (TC) from a reference position.
As shown, the flexibility of the scheme for being adaptable to the multilevel modulation formats is interesting. However, the presented scheme can be extended for high-speed modulation considering a novel and attractive modulation format as hybrid modulation format  by combining PAM, BPM, PSM and PPM.
We have demonstrated an UWB pulse generator based on an N tap microwave photonic filter based on slicing an incoherent optical source and featuring positive and negative samples by means of polarity inversion in an electro-optical modulator. High-order UWB pulses have been experimentally obtained using different number of positive and negative taps in order to obtain a better fulfillment of the FCC mask. We have demonstrated a more efficient photonic solution in terms of power spectral density than conventional monocycle and doublet UWB pulses. The use of incoherent optical sources in combination with AWGs significantly increases the flexibility of the scheme and makes it suitable for multilevel modulation formats. Furthermore, the AWG-based UWB signal generator can be applied for not only multilevel modulation formats but also a hybrid modulation format, which make the transmission bit rate increase to several times the original.
The authors wish to acknowledge the financial support given by the national project TEC TEC2011-26642 (NEWTON) funded by the Ministerio de Economía y Competitividad, the project FEDER UPVOV10-3E-492 and Research Excellency Award Program GVA PROMETEO 2013/012 Next Generation Microwave Photonics Technologies funded by the Generalitat Valenciana.
References and links
1. G. R. Aiello and G. D. Rogerson, “Ultra-wideband wireless systems,” IEEE Microw. Mag. 4(2), 36–47 (2003). [CrossRef]
2. J. Yao, “Photonics for ultrawideband communications,” IEEE Commun. Mag. 41, 66–74 (2003).
3. F. Zeng and J. Yao, “An approach to ultrawideband pulse generation and distribution over optical fiber,” IEEE Photon. Technol. Lett. 18(7), 823–825 (2006). [CrossRef]
5. Y. Dai and J. Yao, “High-chip-count UWB biphase coding for multiuser UWB-over-fiber system,” J. Lightwave Technol. 27(11), 1448–1453 (2009). [CrossRef]
6. Y. Dai and J. Yao, “Optical generation of binary phase-coded direct-sequence UWB signals using a multichannel chirped fiber Bragg grating,” J. Lightwave Technol. 26(15), 2513–2520 (2008). [CrossRef]
7. Q. Wang, F. Zeng, S. Blais, and J. Yao, “Optical ultrawideband monocycle pulse generation based on cross-gain modulation in a semiconductor optical amplifier,” Opt. Lett. 31(21), 3083–3085 (2006). [CrossRef] [PubMed]
8. Y. Yu, J. Dong, X. Li, and X. Zhang, “UWB monocycle generation and Bi-phase modulation based on Mach-Zehnder modulator and semiconductor optical amplifier,” IEEE Photonics J. 4(2), 327–339 (2012). [CrossRef]
9. K. Tan, D. Marpaung, R. Pant, F. Gao, E. Li, J. Wang, D. Y. Choi, S. Madden, B. Luther-Davies, J. Sun, and B. J. Eggleton, “Photonic-chip-based all-optical ultra-wideband pulse generation via XPM and birefringence in a chalcogenide waveguide,” Opt. Express 21(2), 2003–2011 (2013). [CrossRef] [PubMed]
10. S. Pan and J. Yao, “UWB-over-fiber communications: modulation and transmission,” J. Lightwave Technol. 28(16), 2445–2455 (2010). [CrossRef]
11. M. Bolea, J. Mora, B. Ortega, and J. Capmany, “Optical UWB pulse generator using an N tap microwave photonic filter and phase inversion adaptable to different pulse modulation formats,” Opt. Express 17(7), 5023–5032 (2009). [CrossRef] [PubMed]
12. W. Li, L. X. Wang, W. Hofmann, N. H. Zhu, and D. Bimberg, “Generation of ultra-wideband triplet pulses based on four-wave mixing and phase-to-intensity modulation conversion,” Opt. Express 20(18), 20222–20227 (2012). [CrossRef] [PubMed]
13. M. Abtahi and L. A. Rusch, “Arbitrary UWB waveform generator supporting OOK, PPM and PSK modulation formats,” in Proceedings of IEEE Conference on Microwave Photonics (Institute of Electrical and Electronics Engineers, Montreal, 2010), pp. 294–297. [CrossRef]
14. P. Li, H. Chen, M. Chen, and S. Xie, “Gigabit/s Photonic Generation, Modulation, and Transmission for a Reconfigurable Impulse Radio UWB Over Fiber System,” IEEE Photonics Journal 4(3), 805–816 (2012). [CrossRef]
15. P. Li, H. Chen, X. Wang, H. Yu, M. Chen, and S. Xie, “Photonic Generation and Transmission of 2-Gbit/s Power-Efficient IR-UWB Signals Employing an Electro-Optic Phase Modulator,” IEEE Photon. Technol. Lett. 25(2), 144–146 (2013). [CrossRef]
16. K. Tan, J. Shao, J. Sun, and J. Wang, “Photonic ultra-wideband pulse generation, hybrid modulation and dispersion-compensation-free transmission in multi-access communication systems,” Opt. Express 20(2), 1184–1201 (2012). [CrossRef] [PubMed]
17. M. Bolea, J. Mora, B. Ortega, and J. Mora, “High-order UWB pulse generation based on a microwave photonic filter using incoherent optical sources,” in Proceedings of IEEE Conference on Microwave Photonics (Institute of Electrical and Electronics Engineers, Singapore, 2011), pp. 462–465. [CrossRef]
18. J. Capmany, J. Mora, B. Ortega, and D. Pastor, “Microwave photonic filters using low-cost sources featuring tunability, reconfigurability and negative coefficients,” Opt. Express 13(5), 1412–1417 (2005). [CrossRef] [PubMed]
19. L. A. Coldren, S. C. Nicholes, L. Johansson, S. Ristic, R. S. Guzzon, E. J. Norberg, and U. Krishnamachari, “High performance InP-based photonic ICs – A tutorial,” J. Lightwave Technol. 29(4), 554–570 (2011). [CrossRef]