Open Access
Add to BibSonomy
Abstract
We demonstrate an X-band distributed receiving system with 4 remote ends for signal-to-noise ratio (SNR) enhancement. The X-band analog signal received by 4 remote ends is first transmitted to the local end through optical fiber links and is then down-converted with a photonic method for digitization and further coherent combination. Finally, a combined signal with a higher SNR can be obtained. In the proposed system, a frequency-tunable single-tone signal is stably transmitted to the remote end for both down-converting the received signal and for generating a dithered sample clock to eliminate the transmission delay jitter with an unlimited compensation range. Experimentally, X-band binary phase shift keying signals are used for system performance evaluation. After 20 to 25 km transmission, the relative timing drifts between different links are at the order of picoseconds, and a near-theoretical SNR enhancement is achieved. The proposed scheme has a simple remote structure with no need for time synchronization, increasing its signal combining precision, flexibility, and scalability, making it an ideal candidate for long-distance weak signal detection.
© 2023 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
The distributed antenna system has been proven to have incomparable advantages over the centralized single antenna structure in the aspects of detection sensitivity, angular resolution as well as flexibility, and scalability, thanks to its distributed structure and large effective aperture [1,2]. This concept was first employed in satellite communications [3], and now is widely used in deep space exploration [4] and target weak signal detection [5]. For a distributed antenna system, the signals from the same source are first detected by each remote antenna, and they are finally combined in the central station. Generally, the signal-to-noise ratio (SNR) enhancement can be regarded as the performance indicator of the distributed detection system. With the high SNR combined signals, distant targets emitting weak signals can also be detected. The greater the number of antennas and the degree of dispersion, the larger the effective aperture and detection sensitivity will be, provided that the received signals of each link maintain a stable relative delay relationship at the local end [6].
Usually, the distance between the centralized local station and the remote antenna sites is dozens of kilometers, and the optical fiber has been considered the optical transmission media thanks to its low loss, immunity to electromagnetic interference, and other advantages. But the signals transmitted in the optical fiber will experience random timing drift due to environmental variations, decreasing the signal combining performance [7]. Thus, for most distributed antenna systems, the received signals are directly digitized at the remote sites before being transmitted to the local end [8]. In this case, however, a highly stable frequency standard and time synchronization signal are required at every remote end, which is difficult to guarantee considering the practical condition, and the analog-to-digital conversion error will be increased due to the reference voltage drift of the ADC under the influence of ambient temperature changes [9]. Moreover, it also complicates the remote structure and introduces more unstable factors, thereby reducing the deployment flexibility of the remote antennas.
Therefore, directly transmitting the analog signals through the optical fiber to the local end for local digitization and combining has evident advantages. Yet it brings great challenges to the stability of signal transmission, which is essential for the following signal combination. Generally, there are two kinds of stable transmission solutions. One is to actively [10,11] or passively [12,13] adjust the phase of the transmitted signal to eliminate the transmission phase variation. But it can only compensate for the phase variation of a single-frequency signal, which is not applicable for multi-frequency radar signal receiving. The other kind is to compensate for the delay variation of the transmission link with an optical/electrical delay line [14,15] or a wavelength-tunable laser [16]. It ensures that the received signals experience a stable transmission delay, but such systems often have a restricted compensation range, limiting their scope of application.
Here, we propose and demonstrate a 4-antenna distributed system [17] for the full X-band signal receiving and combining, leading to an evident SNR enhancement. To simplify the remote structure and reduce the system complexity for better combining performance, the received analog signals at the remote ends are directly transmitted to the local end for down-conversion, digitization, and coherent combination. Signals converged on the local end are immune to the large delay variation of the fiber link, which is highly desired for practical applications. Experimentally, we detect X-band binary phase shift keying (BPSK) signals for evaluating the system's performance. The relative timing drift of the sampled digital signals after 20 to 25 km transmission is on the order of picosecond, and the experimental combined SNR enhancement is close to the theoretical value, validating the feasibility of the proposed distributed receiving system. To the best of our knowledge, this is the first experimental demonstration of X-band analog signal transmission and combination for SNR enhancement in a 4-antenna distributed receiving system
2. Principle
As is shown in Fig. 1, the proposed 4-antenna distributed detection system realizing phase-stable receiving and signal combining can be divided into three key functional modules: local oscillator (LO) stabilization and distribution (module I), signal transmission and photonic down-conversion (module II), and phase-stable digital receiving and processing (module III). For each link, firstly, a phase-stable LO from a shared tunable radio frequency (RF) source is transmitted from the local end to the remote end via 20 to 25 km optical fiber. The LO is used for photonic down-converting the remote X-band signal into the receiver bandwidth. Secondly, this remote LO signal and the X-band analog signal together modulate another laser at the remote end and are then transmitted back to the local end. By detecting the beat signal with the LO, the X-band analog signals are down-converted to the intermediate frequency (IF). Thirdly, the IF signal at the local end is sampled in a digital receiver with a dithered sample clock that carries the same transmission delay drift as the IF signal induced by the link, and therefore any delay drift of the signal will be perfectly canceled out, leading to time-stable digitization. Finally, the digitized signals from 4 channels are combined for SNR enhancement. We separately discuss the above 3 modules as follows.
2.1 Frequency-tunable phase-stable LO signal distribution
The phase-stable LO distribution is the foundation of the proposed system because it is used for both down-converting the X-band signal and generating the dithered sample clock for signal digitization. The stability of the LO transmission guarantees the stability of the signal transmission and finally the combination.
The phase-stable LO distribution system is shown in the color yellow in Fig. 1. At the local end, the LO signal is obtained by up-converting a phase-stable RF source with an IF voltage-controlled oscillator (VCO) signal in a single-sideband modulator (SSBM). Therefore, the frequency of the LO signal equals the sum frequency of the RF source and the VCO; the phase of the LO can be precisely adjusted by controlling the driving signal of the VCO. When the LO signal arrives at the remote end,
${\omega _{LO}}$ is the angular frequency of the LO signal; ${\varphi _{vco}}(t )$ is the adjustable phase of the VCO;$\Delta {\tau _p}$ is the variable single-trip transmission delay of the fiber link. Part of the LO signal is reflected at the remote end and transmitted back to the local end. A dual-heterodyne phase-error transfer (DHPT) structure is then used to transfer the phase drift of the LO to an IF signal, whose phase variation is detected and sent to a servo for adjusting the phase of the VCO output so as to control the phase of the transmitted LO [18]. By using a phase-locked loop (PLL), the sum of the round-trip LO phase and the VCO phase is locked to a stable cesium clock, which can be expressed asAs long as the loop is in the locking state, the remote LO signal will be phase-stable. A more detailed deduction can be found in our previous work [19].
In order to cover the full X-band (8∼12 GHz) receiving range with a limited digital receiver bandwidth of merely 1 GHz. Here, a frequency tunable RF source is adopted. By adjusting the frequency of the distributed LO signal according to the frequency of received signals, final down-conversion signals can be always within the 1 GHz receiver bandwidth.
2.2 Photonic down-conversion and signal transmission
Usually, the detected signal such as the radar signal has a limited bandwidth but a higher carrier frequency. Thus, a frequency down-conversion is a requisite. In the proposed system, we down-convert the X-band signal with a photonic solution as shown in the color green in Fig. 1. At the remote end, the X-band received signal and the electrical phase-stable LO signal together modulate a laser with double-sideband carrier-suppression (DSBCS) modulation. The beat of the first-order sidebands of the LO and the X-band signal directly becomes the down-conversion of the X-band signal. The modulated light is then injected into the fiber link through a wavelength division multiplexer (WDM) and is transmitted to the local station. At the local end, the photonic down-conversion signal can be detected by a low-frequency photodetector followed by a low-pass filter.
The signal down-conversion should be performed before transmission because it can effectively lower the requirements for transmission stability. Compared with the traditional down conversion in the electrical field with a mixer and a bunch of electrical amplifiers at the remote end, the equivalent photonic down-conversion is compatible with the optical transmission and results in a much simpler remote structure. Meanwhile, it has better harmonics suppression because the electrical harmonics from signal amplification and mixing are greatly reduced. Therefore, the adopted photonic down-conversion is more suitable for the proposed system.
2.3 Local time-stable digitization and coherent combination
The remote phase-stable LO signal ensures stable photonic down-conversion, but when the signal is transmitted to the local end, it will still experience the delay drift induced by the fiber link. Note that the time synchronization error will greatly lower the coherence performance, this random delay drift should be carefully eliminated [6]. Instead of using electrical or optical delay lines for the delay variation compensation, which always has a limited compensation range and speed, we generate a dithered sample clock for delay jitter cancellation in the digitization process.
The principle of the time-stable digitization is to generate a sample clock that experiences the same delay variation as the transmitted signal, enabling the recovery of the original signal after the analog-to-digital conversion. It should be noted that although the transmission delay of different frequency components within a broadband signal are different due to chromatic dispersion of the fiber, the delay variation of each frequency component can be considered the same, making this method feasible for a broadband signal transmission [20]. From Eq. (2), it yields that the phase of the VCO is conjugated with the phase variation of the LO induced by the link. Thus, by phase-conjugating the VCO and up-converting it with the RF signal, we regenerate the LO signal carrying the time delay information of the link. Then it is divided into the frequency of the sample clock, which can be expressed as
$n$ is the ratio of the LO frequency to the sample clock frequency of the analog to digital converter (ADC). After the analogy-to-digital conversion using this regenerated dithered sample clock, the delay drift of the transmitted signals will be naturally eliminated. Therefore, the system can stably receive signals within the receiver bandwidth. Since the dithered sample clock is generated using the VCO signal whose phase is the integral of its frequency variation over time, the dithered clock can introduce an unlimited delay variation within the compensation bandwidth of the phase-locked loop. Thus, this method can eliminate unlimited delay variation of the transmitted signals.After time-stable digitization, the received signals from 4 remote ends are coherently combined with a well-designed combination algorithm. The delay difference among 4 signals is calibrated using the second correlation assisted by the wavelet transformation [21]. The phase difference is adjusted with the SUMPLE algorithm [22]. Then the received signals are combined with full alignment.
Compared with the traditional distributed antenna system with remote digitization and electrical down-conversion, the proposed system with local digitization and photonic down-conversion results in a simpler remote structure where the time synchronization signal, the ADC, the electrical mixer are all omitted. Meanwhile, the time-stable digital receiver with the dithered sample clock can withstand unlimited delay variation of the fiber link, which cannot be easily achieved with electrical or optical delay lines. Thus, the proposed receiving system has remarkable practicability and scalability. Ideally, the SNR of the combined signal will be increased by 6 dB with the current 4 remote ends, dramatically boosting the weak signal detection capability.
3. Experiment and result
The diagram of the experimental setup containing 4 identical links is shown in Fig. 2. As described in the principle part, the system can be divided into 3 function modules.
Fig. 2. Experimental setup. AOFS: acousto-optic frequency-shifter; Conj: phase conjugate circuit; DPMZM: dual-parallel Mach-Zehnder modulator; EF: electric filter; LF: loop filter; MZM: Mach-Zehnder modulator; M: mixer; OPMC: optical polarization-maintaining coupler; PBS: polarization beam splitter; PA: power amplifier; PD: photodetector; PFD: digital phase and frequency discriminator; PNA: phase noise analyzer; PRFM: partial reflective Faraday mirror; SSBM: single-sideband modulator; SMF: single-mode fiber; TIA: time interval analyzer; VCO: voltage-controlled oscillator; 2: frequency doubler.
For the phase-stable LO distribution module, Laser1 is split into two parts. One part is modulated as a reference light in MZM1 by a tunable RF signal with DSBCS modulation. The RF signal is generated from a microwave source synchronized with a 10 MHz cesium clock. The modulated reference light is then divided into 4 parts and is sent to 4 links separately. The other part is first split into 4 parts and directly sent into 4 links separately. Each part is then amplitude modulated as the LO light in MZM2. The LO signal is generated by up-converting the RF signal with a 10 MHz VCO signal in SSBM1. Then the optically carried LO is sent to the S-polarized port of the polarization beam splitter (PBS) and is injected into a wavelength division multiplexer (WDM) and 20 to 25 km fiber consecutively from the combining port of the PBS. At the remote end, part of the LO light is reflected by a partial reflective Faraday mirror (PRFM) with a 90-degree polarization rotation. While the backward light is transmitted back to the local end, it experiences an orthogonal polarization evolution. Therefore, the polarization state of the returned LO light at the combining port of the PBS is fixed [23]. Thus all the returned LO light passes through the PMC in the P-polarized port, avoiding power variation for the following coherent detection. A 40 MHz acousto-optic frequency shifter (AOFS) is used to shift the optical frequency so that the beating IF signals are frequency distinguishable. Inserting the AOFS at the remote end before the WDM can also prevent the effect of Rayleigh backscattering and ensure the symmetry of the forward and backward link. Then, the returned LO light goes through the optical polarization-maintaining coupler (OPMC) and beats the reference light in a photodetector (PD1). The IF signal with twice the phase of the returned LO signal is obtained by using a DHPT structure [18]. After mixed with the frequency-doubled VCO, the IF signal and the 10 MHz reference signal are sent to the phase frequency detector (PFD), generating the error signal to drive the VCO signal. Finally, the phase variation of the LO signal is compensated once the loop is locked.
For the photonic down-conversion and transmission module. the other part of the phase-stable LO light signal at the remote end is transmitted through the PRFM and detected by PD3, generating the electrical LO signal for the photonic down-conversion as well as performance evaluation. A 20 Mbps X-band BPSK signal is generated by an arbitrary waveform generator (AWG) to simulate the practical X-band radar signal whose bandwidth is usually at the order of tens of megahertz. Then the LO signal and X-band signal are sent to different ports of a dual-parallel Mach-Zehnder modulator (DPMZM) to DSBCS modulate the laser2 with another wavelength. Then it is injected into the fiber link through a WDM and transmitted to the local end. The transmitted signal is separated from the returned LO signal by the local WDM and detected by PD2. The X-band signal is down-converted by beating its sideband with the corresponding sideband of the LO signal. After passing through a followed low-pass filter, only the down-conversion signal remains.
The digital receiver at the local end is used to digitize the transmitted IF signals with the generated dithered sample clock. To generate the dithered sample clock, the VCO signal is phase-conjugated and up-converted to the LO frequency by the RF signal in SSBM2. Because the ADC (Teledyne SP Devices ADQ14) uses a 1 GHz sample clock and 2 GHz sample frequency, the up-converted VCO signal is then frequency divided to 1 GHz. The down-converted transmitted signal is then sampled by the ADC with the dithered clock in each link. Finally, 4 digitized signals are further combined in a computer for full performance evaluation.
The stability of the LO distribution determines the phase stability of the down-converted signals and the phase adjustment accuracy of the dithered clock, both of which affect the stability of the signal digitization and the further combining. Here, we set the RF source to 9.99 GHz so that the frequency of the LO signal is 10 GHz. Figure 3 (a) shows the spectra of the closed-loop feedback signal before entering the PFD measured by a real-time spectrum analyzer (Tektronix RSA 5103A). It indicates the in-loop stability of the LO signal. Compared with the free-running state, distinct noise suppression can be observed at the low-frequency range, proving its better long-term in-loop stability. To further evaluate the stability of the remote LO signal, its residual phase noise is measured by a phase noise analyzer (Symmetricom 5125A). Limited by the frequency range of the phase noise analyzer, we down-convert the remote LO signal to 10 MHz by mixing it with the same 9.99 GHz RF source. The residual phase noise of the remote LO is shown in Fig. 3 (b). After phase locking, the phase noise at the low offset frequency has been evidently suppressed. The residual phase noise decreases to -62 dBc/Hz at 0.01 Hz, which is reduced by over 45 dB compared with the free-running condition. Around the loop bandwidth of 1 kHz, the jitter peaking is observed, which has been optimized but still slightly rises the phase noise level [24]. However, the root mean square (RMS) integrated timing jitter is merely 33 fs, and the integrated phase jitter is 0.0021 rad. It indicates superior long-term stability of the LO transmission. Note that we only demonstrate the frequency stability of the remote LO signal for link1, yet other links have similar performance.
Fig. 3. Performance of the phase-locked loop. (a) the power spectra of the compensated in-loop signal of link1; (b) the residual phase noise of the remote LO signal of link1.
To intuitively evaluate the LO stability in the time domain, the timing drifts of the remote distributed LO signals of 4 links are measured by a time interval analyzer (TIA, GuideTech GT668) instead of the phase noise analyzer. The time intervals between the 10 MHz down-converted LO signal and the cesium clock are measured by comparing their zero-crossing point pairs. Since the measured down-converted 10 MHz signal has the same phase value as the 10 GHz LO signal, the timing drift is amplified by a thousand times, leading to a measurement precision at a femtosecond level. As shown in Fig. 4, when lack of phase compensation, the timing drift is as large as 1030 ps, which is dependent on the environmental variations. The RMS timing jitter of link1 to link4 is reduced to 80 fs, 85 fs, 116 fs, and 105 fs, respectively within 3700 s once the loop is phase-locked. Note that this measured RMS jitter is more accurate than the integrated timing jitter from the phase noise. As a result, the phase jitter of the down-converted signal induced by the remote LO signal is less than 0.0073 rad, which can be reasonably ignored. Meanwhile, the phase adjustment accuracy of the dithered clock is also well guaranteed at the same time.
After verifying the LO stability, we evaluate the transmission stability of the X-band signal because the relative delay stability of different links is essential for the following coherent combination. We generate an X-band signal with a carrier frequency of 10.4 GHz and a baud rate of 20 Mbps at the remote end by the AWG to simulate the practical radar signal. The LO signal is still 10 GHz. Thus, the down-converted IF signals have a 400 MHz carrier frequency, which is within the ADC bandwidth. We obtain the relative delay drift of each two links by calculating their cross-correlation in the frequency domain. As shown in Fig. 5(a), the slope of the fitted curve represents the relative delay value of the two links. Figure 5(b) shows the relative delay drift between the received radar signals from every two links. If just using a phase-stable clock, however, the relative timing drift of the received signals between link1 and link2 increases dramatically to 1050 ps and has a clear turning point caused by the abrupt environmental change. When using the dithered sample clock, the RMS relative timing drift is 6.7 ps between link1 and link2. It is 6.1 ps between link3 and link4. Compared with a stabler LO signal, the stability degradation of the transmission delay is mainly caused by the synchronization error of different ADC boards. But note that after the signal down-conversion to 400 MHz, this time drift induced by the link is equivalent to just 0.017 rad phase variation. The proposed system can still effectively eliminate the timing drift of the transmitted signals and guarantee the following signal combination.
Fig. 5. (a) The relative delay calculation from the slope of the fitting curve; (b) the relative timing drift between every two links. L1 to L4 represents link1 to link4.
For evaluating the signal combining performance, the received signals from 4 links are combined in a computer with a well-designed algorithm. The received signals from link1 and the corresponding combined signal are compared both in the time and frequency domain as shown in Fig. 6. For ease of comparison, we normalize the time-domain amplitude and frequency-domain power of the signals. Figure 6(a) shows the overlapped BPSK signals in the time domain after being folded 1000 times. The waveform of the combined signal is evidently cleaner with fewer amplitude variations, showing significant signal quality improvement. Figure 6(b) shows the normalized power spectra of the combined signal and received signal from link1. We take the noise from 800 MHz to 1000 MHz as the average noise power in the entire receiving band, and the calculated SNR of the combined signal is 18.38 dB, which is 5.8 dB higher than that from link1 of the 12.58 dB. The SNR enhancement is clearly observed both in the time and frequency domains.
Fig. 6. The received signal from link1 and the combined signal in (a) the time domain and (b) the frequency domain.
We continuously record the SNRs of all the received signals and the combined signals for 5000 s as shown in Fig. 7. Theoretically, for 4 signals with the same SNR, their combined SNR will be increased by 6 dB. However, in reality, the distances of each antenna from the target are different, resulting in different SNRs of the signals received by each antenna. Therefore, we make the signals from different links have different SNRs to simulate the real situation. And this will result in a different combined SNR gain. For a fair comparison, we calculated the theoretical combined SNR of the received signals from the 4 links which is indicated by the black dashed line. And the SNR of the actual experimental combined signal is increased by about 2.7 dB over the best link and is about 18.4 dB, which is very close to the theoretical value of about 18.9 dB. The 0.5 dB penalty is due to the correlation reduction caused by the noise and the relative phase jitter between the received signals. It finally validates the feasibility of the proposed distributed receiving system. With the SNR enhancement, the detection distance and detection sensitivity will be both improved.
Finally, we demonstrate the receiving capability of the proposed system with a 1-GHz receiver bandwidth. Previously we have presented the receiving stability of BPSK signals with several hundred Mbps in [25], and now we use a 200 Mbps BPSK signal with different carrier frequencies in X-band to verify the frequency tunability of the receiving system. For a 200 Mbps BPSK signal, the sidelobes will cover the whole 1 GHz bandwidth, which will interfere with the observation of the SNR. Thus, we apply a raised cosine filter on the generated BPSK signal to remove its sidelobes. By changing the frequency of the RF signal so as to change the frequency of the LO signal, the generated BPSK signals with different carrier frequencies in X-band can all be down-converted into the digital receiver bandwidth, provided that the frequency division ratio is changed accordingly to maintain 1-GHz dithered sample clock. Figure 8 shows the normalized power spectra of the combined signal of the received 200 Mbps BPSK signals with the carrier frequency of 8.5 GHz, 9.5 GHz, 10.5 GHz, and 11.5 GHz, respectively. The X-band signals are down-converted and received successively by changing the LO signal to the frequency of 8 GHz, 9 GHz, 10 GHz, and 11 GHz. Besides, Fig. 8(d) also shows the SNR increase of the combined signal compared with 4 separate links. 5.21 dB, 5.68 dB, 5.64 dB, and 5.58 dB gain increases are observed respectively. Actually, the proposed system can not only cover the whole X-band but also be further extended to other bands by tuning the RF frequency correspondingly. And the bandwidth of the signal is only limited by the bandwidth of the ADC.
Fig. 8. The spectra of combined signals and received signals with different carrier frequencies: (a) 8.5 GHz; (b) 9.5 GHz; (c) 10.5 GHz; (d) 11.5 GHz.
Finally, we list a comparison table of typical photonic receiving systems as shown in Table 1. We compare multiple crucial indicators including the number of links, transmission distance, link stability, compensation range of the delay variation, types of the transmitted signals and the SNR improvement after signal combination. Among these different systems, the proposed local digitization method shows high transmission stability and superior delay compensation capability. It can support long distance broadband transmission with a relatively simple remote structure. And to the best of our knowledge, this is the first experimental demonstration of X-band analog signal transmission and combination for SNR enhancement in a 4-antenna distributed receiving system.
4. Conclusion
We have demonstrated an X-band distributed 4-antenna system with local digitization and combination for SNR enhancement. The local digitization and photonic down-conversion result in a simple remote structure and excellent receiving performance. The transmission delay variation is eliminated using a dithered sample clock without compensation range limitation. After combining the received signals from 4 remote ends, the SNR enhancement is close to the theoretical value, which can boost the detection distance and sensitivity. It also validates the feasibility and the superiority of the proposed distributed receiving system. Our future work includes link performance equalization and further performance evaluation with real targets and antennas.
Funding
National Natural Science Foundation of China (61827807, 61901039).
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. K. M. Cuomo, S. D. Coutts, J. C. McHarg, N. B. Pulsone, and F. C. Robey, “Wideband aperture coherence processing for next generation radar (NexGen),” MIT Lincoln Laboratory. Tech. Rep. ESC-TR-2004–087 (2004) [CrossRef]
2. F. Zhang, B. Gao, and S. Pan, “Photonics-based MIMO radar with high-resolution and fast detection capability,” Opt. Express 26(13), 17529 (2018). [CrossRef]
3. J. Eberle, “An adaptively phased, four-element array of thirty-foot parabolic reflectors for passive (Echo) communication systems,” IEEE Trans. Antennas Propag. 12(2), 169–176 (1964). [CrossRef]
4. M. Tarenghi, “The Atacama Large Millimeter/Submillimeter Array: overview & status,” Astrophys. Space Sci. 313(1–3), 1–7 (2008). [CrossRef]
5. J. Chen, F. Wang, J. Zhou, and C. Shi, “Information divergence-based low probability of intercept waveform detection for multi-antenna intercept receivers,” J. eng. 2019(20), 7034–7038 (2019). [CrossRef]
6. H. Wang, S. Li, X. Xue, X. Xiao, and X. Zheng, “Distributed coherent microwave photonic radar with a high-precision fiber-optic time and frequency network,” Opt. Express 28(21), 31241 (2020). [CrossRef]
7. A. H. Hartog, A. J. Conduit, and D. N. Payne, “Variation of pulse delay with stress and temperature in jacketed and unjacketed optical fibres,” Opt. Quant Electron 11(3), 265–273 (1979). [CrossRef]
8. M. Zhang, Y. Ji, Y. Zhang, Y. Wu, H. Xu, and W. Xu, “Remote Radar Based on Chaos Generation and Radio Over Fiber,” IEEE Photonics J. 6(5), 1–12 (2014). [CrossRef]
9. N. Saputra, M. A. P. Pertijs, K. A. A. Makinwa, and J. H. Huijsing, “12-bit accurate voltage-sensing ADC with curvature-corrected dynamic reference,” Electron. Lett. 46(6), 397 (2010). [CrossRef]
10. M. Kumagai, M. Fujieda, S. Nagano, and M. Hosokawa, “Stable radio frequency transfer in 114 km urban optical fiber link,” Opt. Lett. 34(19), 2949 (2009). [CrossRef]
11. S. Wang, D. Sun, Y. Dong, W. Xie, H. Shi, L. Yi, and W. Hu, “Distribution of high-stability 10 GHz local oscillator over 100 km optical fiber with accurate phase-correction system,” Opt. Lett. 39(4), 888 (2014). [CrossRef]
12. J. Wei, F. Zhang, Y. Zhou, D. Ben, and S. Pan, “Stable fiber delivery of radio-frequency signal based on passive phase correction,” Opt. Lett. 39(11), 3360 (2014). [CrossRef]
13. Z. Li, L. Yan, B. Luo, and W. Pan, “Phase Fluctuation Cancellation for Uplink Radar Arrays Based on Passive Frequency Mixing,” IEEE Photonics J. 10(2), 1–7 (2018). [CrossRef]
14. O. Lopez, A. Amy-Klein, C. Daussy, C. Chardonnet, F. Narbonneau, M. Lours, and G. Santarelli, “86-km optical link with a resolution of 2 × 10-18 for RF frequency transfer,” Eur. Phys. J. D 48(1), 35–41 (2008). [CrossRef]
15. P. Krehlik, L. Sliwczynski, L. Buczek, and J. Kolodziej, “Fiber-Optic UTC(k) Timescale Distribution With Automated Link Delay Cancelation,” IEEE Trans. Ultrason., Ferroelect. 66(1), 163–169 (2019). [CrossRef]
16. A. Zhang, Y. Dai, F. Yin, T. Ren, K. Xu, J. Li, and G. Tang, “Phase stabilized downlink transmission for wideband radio frequency signal via optical fiber link,” Opt. Express 22(18), 21560 (2014). [CrossRef]
17. K. Wang, W. Wei, P. Wang, D. Wang, W. Xie, and Y. Dong, “4-antenna Distributed Receiving System for Broadband Signal Transmission and Combination,” 2022 Optical Fiber Communications Conference and Exhibition (OFC) (pp. 1–3).IEEE, (2022).
18. D. Sun, Y. Dong, L. Yi, S. Wang, H. Shi, Z. Xia, W. Xie, and W. Hu, “Photonic generation of millimeter and terahertz waves with high phase stability,” Opt. Lett. 39(6), 1493–1496 (2014). [CrossRef]
19. X. Wang, Z. Liu, S. Wang, D. Sun, Y. Dong, and W. Hu, “Photonic radio-frequency dissemination via optical fiber with high-phase stability,” Opt. Lett. 40(11), 2618–2621 (2015). [CrossRef]
20. M. J. Hamp, J. Wright, M. Hubbard, and B. Brimacombe, “Investigation into the temperature dependence of chromatic dispersion in optical fiber,” IEEE Photonics Technol. Lett. 14(11), 1524–1526 (2002). [CrossRef]
21. L. Yang, W. Wei, Z. Liu, X. Wang, N. Deng, and Y. Dong, “Time delay estimation based on the second correlation algorithm and wavelet transformation,” in 2018 Asia Communications and Photonics Conference (ACP) (pp. 1–3). IEEE, (2018).
22. D. H. Rogstad, “The sumple algorithm for aligning arrays of receiving radio antennas: coherence achieved with less hardware and lower combining loss,” IPN Progress Report 42(162), 1–29 (2005).
23. N. C. Pistoni and M. Martinelli, “Polarization noise suppression in retracing optical fiber circuits,” Opt. Lett. 16(10), 711–713 (1991). [CrossRef]
24. A. Hajimiri, “Noise in phase-locked loops,” in 2001 Southwest Symposium on Mixed-Signal Design (Cat. No.01EX475) (pp. 1–6). IEEE, (2001).
25. K. Wang, W. Wei, P. Whang, C. Ma, W. Xie, and Y. Dong, “Phase-stable broadband remote receiving system based on a dithered sample clock,” in 2022 International Topical Meeting on Microwave Photonics (MWP) (to be published).
26. Z. Chen, T. Zhou, X. Zhong, J. Liu, M. Wang, and J. Ye, “Stable Downlinks for Wideband Radio Frequencies in Distributed Noncooperative System,” J. Lightwave Technol. 36(19), 4514–4518 (2018). [CrossRef]
27. L. Lembo, S. Maresca, G. Serafino, F. Scotti, F. Amato, P. Ghelfi, and A. Bogoni, “In-Field Demonstration of a Photonic Coherent MIMO Distributed Radar Network,” in 2019 IEEE Radar Conference (RadarConf) (pp. 1–6). IEEE, (2019).
28. A. Ben-Amram, Y. Stern, Y. London, Y. Antman, and A. Zadok, “Stable closed-loop fiber-optic delay of arbitrary radio-frequency waveforms,” Opt. Express 23(22), 28244–28257 (2015). [CrossRef]
