Open Access
Add to BibSonomy
Abstract
A signal-to-noise ratio (SNR) improvement method for microwave photonic (MWP) links enhanced by optical injection locking (OIL) and channelized spectrum stitching (CSS) is investigated and experimentally demonstrated. By exploiting the resonant amplification characteristics of OIL, both optical gain and in-band noise suppression of the input radio frequency signal can be achieved. The injection bandwidth is channelized to further suppress noise during OIL, and the input signal can be well reconstructed by spectrum stitching in the digital domain. Experimental results show that the optimal improvement in SNR of 3.6 dB is achieved for linear frequency modulated signals and at least an additional improvement of 7.2 dB can be obtained by adopting CSS. Other broadband signals for radar and communication are used to further verify the ability to improve SNR. The potential for application scenarios with large operating bandwidth and high optical gain is also demonstrated.
© 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Microwave photonic (MWP) links are core units of applying photonic technology to electronic systems, providing a new dimension to transmit and process microwave signals. Thanks to the advantages of photonic technology such as broadband, agility and reconfigurability [1,2], MWP links have a wide range of applications in digital optical communications, radio frequency over fiber, microwave photonic radar, electronic warfare [3–6], etc. Performance of MWP links is decisive for the overall performance of optoelectronic systems. However, due to the power loss caused by electro-optical interconversion and the quantum noise of photonic devices which is stronger than thermal noise, there is typically 30 dB deterioration in the signal-to-noise ratio (SNR) of radio frequency (RF) signals output from MWP links compared to that of the input signals [7]. Actually, in most application scenarios [8–11], SNR of the broadband RF signals applied to electro-optical modulators in MWP links is usually at a low level, which makes the SNR deterioration introduced by the links more unacceptable.
To overcome this problem and improve the SNR performance, many research methods and investigations are proposed [11–16], which can be categorized into loss compensation and noise suppression. A common solution to compensate for optical power loss is to apply optical amplifiers such as erbium-doped fiber amplifiers (EDFAs) and semiconductor optical amplifiers (SOAs) in the links. However, the amplifiers introduce noise in the optical domain with a theoretical minimum deterioration in optical SNR of 3 dB [12–14]. Another solution is to suppress different kinds of noise in the optical domain, such as phase-sensitive amplification (PSA) for noise of fiber links and balanced detection for common-mode noise. PSA is usually used in optical communication systems, and amplification without optical SNR deterioration can be obtained [11,15]. However, both the complexity and the control difficulty of the system are increased using twice the optical bandwidth in exchange for better optical SNR performance. Balance detection is often used to diminish the common-mode noise from the laser to the modulator, meanwhile it causes shot noise multiplication of the photodiode (PD) [16]. Nevertheless, traditional solutions only alleviate some of the deterioration introduced by the optical domain itself, do not deal with the noise contained in RF signals input to MWP links, and there is still SNR deterioration of RF signals output from the links. In-band noise suppression while maintaining the original input RF signals is a practical requirement for systems based on MWP links. Recently, optical injection locking (OIL) has been used in photonic-assisted measurements and radar systems replacing optical amplifiers to amplify frequency modulated signals [17–19], and data sideband amplifiers based on OIL are also investigated for optical communication systems [20,21]. OIL is a frequency and phase synchronization technique based on the interaction of externally injected photons with photons in the active cavity [22], which has the potential to provide selective amplification. However, the influence of OIL on the SNR performance and in-band noise variation of broadband RF signals has not been paid enough attention for MWP links.
In this paper, an SNR improvement method based on OIL for MWP link performance enhancement is investigated and experimentally demonstrated, and channelized spectrum stitching (CSS) is introduced into the method to further improve SNR. A broadband RF signal with low SNR is sufficiently amplified for electro-optical (E/O) conversion, and then a single 1st optical sideband is injected into an SL. Exploiting the resonant amplification characteristics of OIL, the signal in the injected component resonates within the laser cavity causing injection locking to obtain high optical gain. The RF in-band noise modulated at the same spectrum position cannot trigger resonance and therefore is not locked, but can only be gained by the active medium. In total, the broadband RF signal obtains optical gain stronger than that of the in-band noise during OIL, which results in an SNR improvement after optical-electro (O/E) conversion. Due to the nonlinearity in the laser cavity, additional in-band noise from the mixing between the signal and noise appears in the output of OIL. The undesired noise is reduced by channelizing the optical spectrum before injection, which cuts down the bandwidth of noise involved in mixing. By stitching the channels with a digital signal processing (DSP) algorithm, the signal can be well reconstructed, and an additional SNR improvement is obtained compared to using OIL alone. To evaluate the ability of SNR improvement, different formats of RF broadband signals with varying input SNR are used. Factors affecting the SNR performance of OIL have also been investigated. Moreover, the applicability for scenarios with wider operation bandwidth and higher optical gain is demonstrated. In the proof-of-concept experiment, a typical coherent detection structure is deployed. Experimental results show that the optimal improvement in SNR of 3.6 dB is achieved for a group of linear frequency modulated (LFM) signals with 2 GHz bandwidth, accompanied by 29.7 dB optical gain. By adopting CSS, the SNR of LFM signals with 1 GHz bandwidth is further improved by at least 7.2 dB. Due to the improved SNR, the integral sidelobe ratio (ISLR) of nonlinear frequency modulated (NLFM) signals with 2 GHz bandwidth and the error vector magnitude (EVM) of 200 Mbaud quadrature phase-shift keying (QPSK) signals, two SNR-sensitive indicators, are reduced by up to 2.9 dB and 9%, respectively. By tuning the injected laser to approximately follow the input signal with a wider bandwidth, the operation bandwidth of this method is extended by 2 GHz, with a higher optical gain of 8.2 dB.
2. Operation principle and experimental setup
The theory of OIL has been well studied based on the set of Lang-Kobayashi (L-K) equations [23], the essence of which is the forced resonance from the slave laser (SL) under external light injection. For a steady locked state, the injected light can be expressed as:
where Ainj and φinj(t) are the complex field amplitude and the phase of the injected field, respectively. As an extension of single-frequency OIL, such signals can be considered as the sinusoid-like signal, that is, each moment corresponds to a single frequency component. It means that the injected signal with a certain bandwidth such as LFM signal, NLFM signal, and phase-coded signal exhibits stable locking properties in the locking range [22]:Fig 1. shows the schematic diagram of the proposed SNR improvement method. The OIL module not only provide optical gain but also suppress in-band noise for injected optical sideband. It can be flexibly inserted into electro-optical interconversion which is necessary for MWP links. To focus on SNR variation in the electrical domain, which is introduced by OIL, a single-sideband suppressed-carrier (SSB-SC) modulation and coherent detection structure is deployed. It provides pure injection components, and the output is the original input RF signal. Specifically, a continuous-wave (CW) laser is divided into a carrier and an optical local oscillator (LO) with a frequency of fo. An SSB-SC modulator can be realized by combining a Mach-Zehnder modulator (MZM) with an optical bandpass filter (OBPF) or a dual-parallel MZM (QPSK modulator) with a microwave 90-degree hybrid coupler. The modulated light as the ‘master laser (ML)’ can be expressed as:
where Aml is the complex field amplitude of the injected filed and φRF(t) is the phase of the input RF signal. By appropriately adjusting the power of the ML and the free running frequency of the SL, the injected signal falls completely into the locking range to ensure a steady locked state under weak injection. After OIL process, the output frequency of the SL is strictly the same as that of the ML, and the power is constant eliminating the slight amplitude fluctuations from the input signal:SNR of broadband RF signals applied to E/O modulators in MWP links is usually at a low level. To obtain effective E/O conversion, the RF signal is amplified by electrical amplifiers resulting in high-power signal and in-band noise. As shown in Fig. 1, the noise modulated at the same position as the signal in the optical spectrum has a higher power density than that of the noise floor. The output of the SL consists of the injection locked resonant signal and in-band noise reflected by the laser cavity. Although the injected noise traveling through the gain medium within the laser cavity is not locked, it does experience some amplification, which is often observed in optical frequency comb selection [22,25]. Other factors that contribute to the undesired gain of the injected noise will be discussed in the experimental results. In the optical domain, OIL offers differential amplification for the broadband signal and in-band noise which are difficult to be separated from each other in the electrical domain. The output of the SL and the optical LO from the CW laser beat with each other in a photodiode (PD), and the original RF signal with an improved SNR is obtained.
Fig. 1. Schematic diagram of the SNR improvement method. SSB-SC: single-sideband suppressed-carrier, OIL: optical injection locking, CW: continuous wave, PD: photodiode.
Figure 2 shows a proof-of-concept experimental setup of the proposed SNR improvement method. To show the variation of the RF signal more clearly, a coherent detection structure is deployed. Specifically, a continuous-wave light generated from an ultra-narrow linewidth laser (100 Hz, NKT Koheras BASIK) is divided into two branches by an optical coupler, one serves as the optical LO and the other is to carry the input signal. A broadband RF signal is generated from an arbitrary waveform generator (Keysight M8190A). The signals under test include three formats: LFM, NLFM and QPSK. Firstly, the LFM signal which is widely used in radar systems:
The LFM signal has a period of T ($- T/2 \le t \le T/2$), where fc is the central frequency and B is the bandwidth. Secondly, the NLFM signal with an S-shape frequency function of time [26]:
Fig. 2. Experimental setup of the SNR improvement method. OBPF: optical bandpass filter, CIR: circulator, AWG: arbitrary waveform generator, EC: electrical coupler, PA: power amplifier, CLD: compact laser-diode driver, DSO: digital storage oscilloscope.
3. Experimental results and discussion
3.1 Evaluation of SNR improvement ability
The performance of the proposed method is evaluated by three signal formats including LFM, NLFM and QPSK. Firstly, a group of LFM signals is used to demonstrate the ability to improve SNR. It takes the form of a continuous wave with a period of 100 µs, a center frequency of 7 GHz, and a bandwidth of 2 GHz. By adjusting the variable electrical attenuator in the noise source, the signal to in-band noise power ratio can be varied within a certain range. The SNR range in the LFM signal test is from 4.5 dB to 32.5 dB before E/O modulation (Point A in Fig. 2), measured directly by an electrical signal analyzer (ESA, Agilent N9030A) and corroborated with the calculated results of the waveforms acquired by the DSO. Figure 3 (a), (b) and (c) show the spectrum of the input RF signal at Point A with an SNR of 4.5 dB, 18.3 dB and 32.5 dB, respectively. In order to show the in-band noise clearly, the spectrum is calculated from a single period interception of the acquisition signal. Due to the non-ideal characteristics of the electrical devices, the noise bandwidth is slightly larger than the signal bandwidth, while the redundant part will be removed in DSP.
Fig. 3. Electrical spectrum of the system input: (a), (b), (c) and output: (d), (e), (f). (a) 4.5 dB SNR, (b) 18.5 dB SNR, (c) 32.5 dB SNR, (d) 1.1 dB SNR improvement, (e) 3.6 dB SNR improvement, (f) 0.8 dB SNR improvement.
The CW-laser wavelength is 1549.31 nm and the output power is 13 dBm. The SL is pumped by 79.5 mA at 21 °C controlled by the CLD. The free-running wavelength of the SL is 1549.25 nm with a power of 8.2 dBm. The LFM optical sideband with 2 GHz bandwidth is injected directly into the SL. By properly adjusting the VOA, a steady locked state can be maintained with an injection power as low as -21.5 dBm thus a 29.7 dB optical gain is obtained. Figure 4 shows the optical spectrum before and after OIL. The frequency of the CW-laser, which serves as the optical LO, is set to the normalized center frequency. Thanks to the ultra-narrow linewidth of the CW-laser, the output of OIL and the optical LO can be coherently detected by directly beating at the PD. The system outputs are acquired by the DSO without any electrical amplifier or filter. Figure 3 (d), (e) and (f) show the calculated electrical spectrum. It can be seen from the vertical comparison that in-band noise has been suppressed, and the slight power fluctuation of the original signal is eliminated as well. The SNR improvement of these three inputs is 1.1 dB, 3.6 dB and 0.8 dB, respectively.
Figure 5 shows the improvement in SNR of system outputs compared with that of input RF signals when the input SNR is varied from 4.5 dB to 32.5 dB in 1 dB steps. Each point is the mean of the calculated results from four independent measurements. The blue solid line in Fig. 5 (a) keeps an optical gain of 29.7 dB corresponding with Fig. 4. The optimal result using OIL alone in this experiment is 3.6 dB improvement when the input SNR is 18.3 dB. It can be seen that this method realized signal enhancement over a certain range of input SNR variations.
Fig. 4. Optical spectrum before and after OIL. Bule solid line: injected light measured at Point B, Yellow dashed line: locked result measured at Port 3 of CIR, Orange solid line: locked result and optical LO measured at Point C.
Fig. 5. SNR improvement curve. (a) Using only the proposed OIL based method. Blue solid line: locked by one SL, Orange solid line: cascaded locked by two SLs under different injection conditions. (b) Without OIL while the same power of optical sideband.
As the input SNR decreases, the competition for carriers between the signal and the in-band noise in the SL cavity becomes more intense with a constant injection power, which is directly shown in the power distribution between locked resonance and medium gain. However, when it increases above the optimal point, the performance begins to diminish as well, due to the shot noise of PD. Figure 5 (b) shows the change in SNR of the system output removing the OIL module, with the optical sideband power remaining -21.5 dBm. The power of the optical sideband is quite smaller than that of the optical LO, and the output power of the SL remains constant as well. The power passed into the PD is nearly unchanged, thus the power of shot noise is approximately a constant value that becomes more significant as the input SNR increases:
Besides the gain medium, another major source of the noise gain during OIL process is the nonlinearity within the laser cavity. Figure 6 shows the electrical spectrogram of the signal with 18.3 dB input SNR over 4 periods. Figure 6 (a) and (b) are from the input RF signal at Point A in Fig. 2 and the system output, respectively. Although the total background for in-band noise is suppressed using OIL, there is parallelogram-shaped additional noise superimposed over the entire period. We think that it comes from the four-wavelength mixing (FWM) effect of nonlinear properties inside the laser cavity. As shown in Fig. 6 (c), the in-band noise is always present intact in the cavity when the signal sweeps across the bandwidth. The FWM occurs between the noise on both sides and the signal resulting new noise in the signal frequency band, which deteriorates the performance of SNR improvement. It can be alleviated by reducing the injection noise during OIL process, as will be shown later.
Fig. 6. Electrical spectrogram of the system input, output and schematic diagram of laser nonlinearity. (a) 18.3 dB input SNR, (b) 3.6 dB SNR improvement, (c) undesired additional noise due to the FWM between LFM signals and in-band noise.
Another two kinds of broadband signals for radar and communication are used to further verify the ability to improve SNR. ISLR and EVM are SNR-sensitive indicators of NLFM signals and QPSK signals, respectively. The NLFM signal used in this experiment with low PSLR and low ISLR, is from [26]. An S-shaped frequency function of time is obtained by editing the phase based on Fourier series. The group of NLFM-CW signals has a period of 100 µs, a center frequency of 7 GHz, a bandwidth of 2 GHz, and SNR ranges from 5.1 dB to 29.7 dB in 1 dB steps. The electrical spectrogram of the system input with 13.9 dB SNR is shown in Fig. 7 (a), the frequency modulation is flat in the middle of the period. The settings of the CW-laser and the SL are the same as the 1st OIL for the LFM signal test (Fig. 5). The optical gain is 27.7 dB with an injection power of -19.5 dBm.
Fig. 7. NLFM signal enhancement with 13.9 dB input SNR. (a) Electrical spectrogram of the system input, (b) Electrical spectrum of the system input and output, (c) Autocorrelation of the system input and output, Bule solid line: original RF input, Orange solid line: system output.
Table 1. shows the changes of SNR and ISLR compared with the original inputs. Each sample is calculated from two independent measurements. The output SNR is improved over the entire test range. Figure 7 (b) shows the electrical spectrum of the optimal SNR improvement (3.5 dB under 13.9 dB input).
Table 1. SNR and ISLR improvement of the NLFM signal
ISLR characterizes the proportion of the main lobe energy in the entire autocorrelation result, and the lower the value, the better performance. The results show that the ISLR decreases correspondingly with the increase of SNR, and the optimal value is -2.9 dB. The autocorrelation of the optimal sample is shown in Fig. 7 (c), the impulse response width is maintained with a slightly enhancement of PSLR. For NLFM signals, OIL can improve the SNR while maintaining other properties.
The QPSK signal under test has a phase-coded period of 1 ms. In order to shrink the main lobe for better locked performance while keeping the amplitude constant, the sampling rate of the AWG is set to 10 GSa/s playing a 200 MBaud QPSK signal, which ensures the smooth filtering for phase jitter between adjacent codes. The signal has a carrier frequency of 7 GHz, and the modulated result is played by the AWG directly. Both the system input and output are QPSK signals containing the carrier. To examine the SNR changes of RF signals after OIL process, decoding is based on DSP. The input SNR ranges from 5.7 dB to 30.7 dB in 1 dB steps. Due to the component conditions, the operation band of the electrical BPF before E/O conversion (at Point A) is still covered 6 GHz to 8 GHz, which makes unnecessary noise injected and affects the performance of OIL. The injection power of the optical sideband is -20 dBm and a 28 dB optical gain is obtained. Figure 8 shows the decoded results of input RF signals and system outputs, each point is calculated from 20 consecutive measurements.
EVM, which is closely related to SNR, shows the difference between the actual waveform and the ideal encoded signal. As we can see that the EVM after OIL process is improved when the input SNR is between 5.7 dB and 18.7 dB, after which the EVM is deteriorated. Unlike the previous results for frequency modulated signals, the EVM improvement reaches up to 9% at the worst of the original input SNR. From the constellation diagrams inserted in Fig. 8, it can be found that the suppression effect of OIL on amplitude noise is significant over a wide range of input SNR, even at the inflection point of EVM (18.7 dB), which is consistent with the previous results. The difference mainly comes from phase dispersion as shown in the constellation diagrams. The fiber link introduces additional phase noise, and both the high-power white noise and the instability of the SL driver are translated into phase noise during OIL:
Fig. 8. EVM curve of 200 MBaud QPSK signals of original RF inputs: Blue solid line, and system outputs: Orange solid line. Insets show constellation diagrams for both conditions with 5.7 dB, 14.7 dB, 18.7 dB and 30.7 dB input SNR, respectively.
3.2 Further improvement of SNR introduced by CSS
To deal with the additional in-band noise from the laser nonlinearity, a CSS-assisted method is proposed to reduce the noise in the injected components and reconstruct the signal by DSP. Figure 9 shows the schematic diagram of the improved method. A fiber Bragg grating (FBG) replaces the OBPF in the original scheme shown in Fig. 2, and the link after the MZM forms an independent channel, of which the number depends on that of FBGs. An ultra-narrowband FBG (TeraXion, TFN-1550-U100), as shown in Fig. 10 (a), is used to cut the optical sideband into slices. The 3dB-bandwidth of the transmission port is 150 MHz. Therefore, the injected in-band noise is reduced in both power and bandwidth. By stitching the acquisition of each channel, a reconstructed signal with improved SNR benefitting from OIL and CSS is obtained.
Fig. 9. Schematic diagram of the channelized spectrum stitching-assisted method. The RF source and laser driver are omitted. FBG: fiber Bragg grating, DSP: digital signal processing.
Fig. 10. Optical spectrum of the FBG and output waveform of channels. (a) Transmission port used in the experiment (viewed by APEX, AP2041B), Inset shows the zoom-in spectrum of the passband, (b) Ch.1 to Ch.7 from left to right, Upper: without OIL, Downer: using OIL.
In this experiment, the signal under test is the LFM-CW signal with a period of 100 µs, a bandwidth of 1 GHz and an SNR ranging from 2.9 dB to 22.2 dB in 3 dB steps. Due to the hardware constraint, the time-division test method is used to simulate a seven-channel structure by tuning the FBG. The electrical BPF before the DSO in Fig. 7 is replaced by DSP according to the operation band of the FBG in each channel. The injection power of each slice is about -23.6 dBm, and the free-running power of the SL pumped by 90 mA at 21 °C is 9 dBm. The 32.6 dB optical gain is obtained due to a narrow bandwidth for injection. The CW-laser is tuned to 1549.31 nm with a power of 13 dBm.
The waveforms output from each channel without OIL or using OIL are shown in Fig. 10 (b). Since the temperature-controlled tuning of the FBG is non-uniformly stepped, the envelope of the waveform in each channel without OIL is not consistent. The valuable components of signal in each channel can be extracted by simply setting a power threshold. The reconstructed signal is stitched using time-frequency filtering, which has the ability to suppress noise on its own. However, the stitching cannot be achieved due to the high fading of power at the edges of each channel without OIL. Figure 11 shows the result of the improved method. Figure 11. (a) has the input SNR of 14.2 dB, the reconstructed LFM signal has a 13.6 dB peak sidelobe ratio (PSLR) with 11.2 dB SNR improvement in total (3.6 dB for OIL alone). Figure 11 (a-iii) and (a-iv) show the time filtering part of channelized OIL and signal stitching. The total result is shown in Fig. 11 (b), each point is calculated from two independent measurements. Compared with using only OIL, the CSS-assisted method can further provide at least 7.2 dB SNR improvement over the entire test range of SNR.
Fig. 11. Result of the channelized spectrum stitching-assisted method. (a) Reconstructed signal with 14.2 dB input SNR: (i) Electrical spectrum, (ii) Autocorrelation, (iii) Spectrogram of Ch.6 without BPF, (iv) Spectrogram of stitched signal with threshold filter in the time domain, (b) Additional SNR improvement compared with OIL alone.
3.3 Extension of operation bandwidth
Moreover, in order to meet the adaptation of this method to wider operation band and higher optical gain in different application scenarios, the SL is tuned approximately matched with the period and the bandwidth of the injected sideband. As shown in Fig. 12, the 2nd channel of the AWG provides a sweep function, which has the synchronization period of the signal under test, to the CLD. By adjusting the function power, the pump current of the SL is modulated so that the free-running frequency and the injected sideband can cover the similar range. Thanks to the locking bandwidth, the function does not require fine control.
Fig. 12. Tuning lock for LFM signals with a 4 GHz bandwidth (a) Optical spectrum before and after OIL, Yellow dashed line: free-running without function, Blue: injection power of -21.4dBm, Light bule: locked for the blue line without function, Red: injection power of -29.6 dBm, Purple: locked for the red line with function, (b) Electrical spectrogram of (i) free-running with function, (ii) unlocked output, (iii) locked result with function.
The signal under test is positive and negative frequency modulated with 1 ms period for both sides, 4 GHz bandwidth and 7 GHz center frequency. The free-running power of the SL without function is 6.2 dBm. As shown in Fig. 12 (a), the maximum optical gain without function is 27.6 dB, 2.1 dB lower than that of 2 GHz LFM signal. By applying the function to the CLD, a steady locked state with the minimum injection power of -29.6 dBm is achieved, 8.2 dB higher optical gain and 2 GHz wider bandwidth is obtained. Figure 12 (b) shows the electrical spectrogram of system output in different conditions.
4. Conclusion
In this paper, an OIL based method for SNR performance enhancement of MWP links is investigated and experimentally demonstrated. It is suitable for functional modules and applications in compact scenarios such as photonic-assisted radar transceiver and integrated MWP radar. The SNR of RF signals output from the system adopting the method is improved compared with that of the input. The in-band noise of input RF signals is suppressed during OIL process due to the resonant amplification characteristics which provides different optical gain for the signal and in-band noise. The performance of this method is evaluated by three signal formats including LFM, NLFM and QPSK. The improvements of 3.6 dB SNR, 2.9 dB ISLR, and 9% EVM are achieved, respectively. The undesired amplification of in-band noise is investigated as well. A CSS-assisted method is proposed to reduce the noise gain from laser nonlinearity and further provide at least 7.2 dB SNR improvement compared with using OIL alone. Besides, a higher optical gain and a wider operation band for more application scenarios is demonstrated by tuning the SL.
Funding
National Key Research and Development Program of China (2019YFB2203300, 2019YFB2203302); National Natural Science Foundation of China (62104232); Key Research Program of Frontier Science, Chinese Academy of Sciences (ZDBS-LY-JSC016).
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. J. Capmany and D. Novak, “Microwave photonics combines two worlds,” Nat. Photonics 1(6), 319–330 (2007). [CrossRef]
2. J. Yao, “Microwave photonics,” J. Lightwave Technol. 27(3), 314–335 (2009). [CrossRef]
3. R. Waterhouse and D. Novak, “Realizing 5G: Microwave photonics for 5G mobile wireless systems,” IEEE Microw. Mag. 16(8), 84–92 (2015). [CrossRef]
4. K. Xu, R. Wang, Y. Dai, F. Yin, J. Li, Y. Ji, and J. Lin, “Microwave photonics: radio-over-fiber links, systems, and applications,” Photonics Res. 2(4), B54–B63 (2014). [CrossRef]
5. P. Ghelfi, F. Laghezza, F. Scotti, G. Serafino, A. Capria, S. Pinna, D. Onori, C. Porzi, M. Scaffardi, A. Malacarne, V. Vercesi, E. Lazzeri, F. Berizzi, and A. Bogoni, “A fully photonics-based coherent radar system,” Nature 507(7492), 341–345 (2014). [CrossRef]
6. M. E. Manka, “Microwave photonics for electronic warfare applications,” in 2008 International Topical Meeting on Microwave Photonics jointly held with the 2008 Asia-Pacific Microwave Photonics Conference (IEEE, 2008), pp. 275–278.
7. V. J. Urick, K. J. Williams, and J. D. McKinney, Fundamentals of microwave photonics (John Wiley & Sons, 2015).
8. R. Li, W. Li, M. Ding, Z. Wen, Y. Li, L. Zhou, S. Yu, T. Xing, B. Gao, Y. Luan, Y. Zhu, P. Guo, Y. Tian, and X. Liang, “Demonstration of a microwave photonic synthetic aperture radar based on photonic-assisted signal generation and stretch processing,” Opt. Express 25(13), 14334 (2017). [CrossRef]
9. X. Ye, F. Zhang, Y. Yang, and S. Pan, “Photonics-based radar with balanced I/Q de-chirping for interference-suppressed high-resolution detection and imaging,” Photonics Res. 7(3), 265–272 (2019). [CrossRef]
10. J. Dong, F. Zhang, Z. Jiao, Q. Sun, and W. Li, “Microwave photonic radar with a fiber-distributed antenna array for three-dimensional imaging,” Opt. Express 28(13), 19113–19125 (2020). [CrossRef]
11. R. Kakarla, J. Schröder, and P. A. Andrekson, “One photon-per-bit receiver using near-noiseless phase-sensitive amplification,” Light: Sci. Appl. 9(1), 153–157 (2020). [CrossRef]
12. D. M. Baney, P. Gallion, and R. S. Tucker, “Theory and measurement techniques for the noise figure of optical amplifiers,” Opt. Fiber Technol. 6(2), 122–154 (2000). [CrossRef]
13. D. R. Zimmerman and L. H. Spiekman, “Amplifiers for the masses: EDFA, EDWA, and SOA amplets for metro and access applications,” J. Lightwave Technol. 22(1), 63–70 (2004). [CrossRef]
14. L. Galdino, D. Semrau, M. Ionescu, A. Edwards, W. Pelouch, S. Desbruslais, J. James, E. Sillekens, D. Lavery, S. Barnes, R. I. Killey, and P. Bayvel, “Study on the impact of nonlinearity and noise on the performance of high-capacity broadband hybrid Raman-EDFA amplified system,” J. Lightwave Technol. 37(21), 5507–5515 (2019). [CrossRef]
15. S. L. I. Olsson, H. Eliasson, E. Astra, M. Karlsson, and P. A. Andrekson, “Long-haul optical transmission link using low-noise phase-sensitive amplifiers,” Nat. Commun. 9(1), 2513 (2018). [CrossRef]
16. Y. Painchaud, M. Poulin, M. Morin, and M. Têtu, “Performance of balanced detection in a coherent receiver,” Opt. Express 17(5), 3659–3672 (2009). [CrossRef]
17. B. Wang, X. Fan, S. Wang, J. Du, and Z. He, “Millimeter-resolution long-range OFDR using ultra-linearly 100 GHz-swept optical source realized by injection-locking technique and cascaded FWM process,” Opt. Express 25(4), 3514–3524 (2017). [CrossRef]
18. Y. Dong, Z. Zhu, X. Tian, L. Qiu, and D. Ba, “Frequency-modulated continuous-wave LIDAR and 3D imaging by using linear frequency modulation based on injection locking,” J. Lightwave Technol. 39(8), 2275–2280 (2021). [CrossRef]
19. J. Li, X. Xue, J. Zhang, Z. Xue, B. Yang, S. Li, and X. Zheng, “Ultrahigh-factor frequency multiplication based on optical sideband injection locking for broadband linear frequency modulated microwave generation,” Opt. Express 29(25), 40748–40758 (2021). [CrossRef]
20. Y. Hung and S. Hwang, “Photonic microwave amplification for radio-over-fiber links using period-one nonlinear dynamics of semiconductor lasers,” Opt. Lett. 38(17), 3355–3358 (2013). [CrossRef]
21. A. Fragkos, A. Bogris, D. Syvridis, and R. Phelan, “Amplitude Noise Limiting Amplifier for Phase Encoded Signals Using Injection Locking in Semiconductor Lasers,” J. Lightwave Technol. 30(5), 764–771 (2012). [CrossRef]
22. Z. Liu and R. Slavík, “Optical injection locking: From principle to applications,” J. Lightwave Technol. 38(1), 43–59 (2020). [CrossRef]
23. R. Lang and K. Kobayashi, “External optical feedback effects on semiconductor injection laser properties,” IEEE J. Quantum Electron. 16(3), 347–355 (1980). [CrossRef]
24. E. K. Lau, L. J. Wong, and M. C. Wu, “Enhanced modulation characteristics of optical injection-locked lasers: A tutorial,” IEEE J. Sel. Top. Quantum Electron. 15(3), 618–633 (2009). [CrossRef]
25. D. S. Wu, R. Slavík, G. Marra, and D. J. Richardson, “Direct selection and amplification of individual narrowly spaced optical comb modes via injection locking: design and characterization,” J. Lightwave Technol. 31(14), 2287–2295 (2013). [CrossRef]
26. Q. Xie, H. Zeng, Z. Mo, and W. Li, “A Two-Step Optimization Framework for Low Sidelobe NLFM Waveform Using Fourier Series,” IEEE Geosci. Remote Sensing Lett. 19, 1–5 (2022). [CrossRef]
