## Abstract

In this paper, we have proposed and experimentally demonstrated a multiplexed sensing interrogation technique based on a flexibly switchable multi-passband RF filter with a polarization maintaining fiber (PMF) Solc-Sagnac loop. A high-order Solc-Sagnac loop can be used as a spectrum slicer as well as sensing heads, and a multi-passband microwave photonic filter (MPF) can be achieved together with a dispersive medium. Environmental parameter variations will cause a frequency shift of the corresponding passband of the MPF, so by employing only one Sagnac loop, it is possible to monitor several environmental parameters simultaneously. In this article, we have demonstrated and analyzed the performance of the flexibly switchable multi-passband MPF by using a second-order Solc-Sagnac loop. To demonstrate the temperature sensing capabilities of our interrogation system, we have applied temperature changes individually to Sensor Head 1 (${L_{PM{F_1}}} \approx 0.97m$) only, Sensor Head 2 (${L_{PM{F_2}}} \approx 2.97m$) only, and both Sensor Head 1 and 2 in the experiment. By monitoring frequency shift of the MPF’s passbands, the sensitivities for Sensor Head 1 and Sensor Head 2 have been estimated to be -0.275 ± 0.011 MHz/℃ and -0.811 ± 0.013 MHz/℃ respectively, which show a good sensing linearity and stability. By utilizing the longer length of the sensing PMF, higher sensitivity can be achieved. By using Solc-Sagnac loop with higher order, more sensors can be multiplexed.

© 2021 Optical Society of America under the terms of the OSA Open Access Publishing Agreement

## 1. Introduction

Microwave photonic (MWP) has attracted the interest of researchers due to various advantages of processing radio frequency (RF) signals in the optical domain instead of using traditional electric counterparts limited by the so-called electronic bottleneck [1–4]. Microwave photonic filter (MPF) as one of the important techniques in MWP brings supplementary advantages inherent to photonic technology such as compactness, low attenuation, immunity to electromagnetic interference, good tunability and reconfigurability [5]. Generally speaking, MPFs can be implemented in the incoherent regime using a delay-line configuration or in the coherent regime using an optical filter [6]. To avoid a periodic RF spectrum response, a sliced broadband optical source together with dispersive medium has been usually used to implement single passband or multi-passband MPF by continuous sampling instead of discrete sampling [6–10]. In the past few years, MPF technology originally used in communication applications has been investigated for fiber sensing system [11,12]. By introducing microwave modulation into the optical system, the optical detection is synchronized with microwave modulation frequency [13], in which the variation of environmental parameters (such as temperature) can be demodulated via monitoring passbands’ frequency shifts of a single-passband MPF [14]. For practical applications, multiplexed sensing interrogation technique based on multi-passband MPF is one of comprehensive key research field in the world. A sensing multiplexing system based on MWP technique has been proposed and demonstrated to measure variations of physical magnitudes by using the low coherence interferometry (LCI) in [15]. Multiplexing sensing system for fiber Mach-Zehnder interferometer (FMZI) based temperature sensors has also been realized by using FMZIs as the sensing elements as well as the spectrum slicer for a dual-passband MPF [16]. However, in these schemes, multiple fiber-optic interferometers should be used, which will increase the system complexity, cost and instability.

In this paper, we have proposed and experimentally demonstrated a multiplexed sensing interrogation technique based on a flexibly switchable multi-passband MPF by using one high order Solc-Sagnac loop. The Solc-Sagnac loop can be used as a spectrum slicer as well as sensor heads, and together with a dispersive medium a multi-passband MPF can be implemented. Compared with traditional demodulation schemes in the optical domain, the proposed sensing scheme has aperiodic frequency response and can avoid the limitation of the periodic interference spectrum in the optical domain. In our experiment, a second-order Solc-Sagnac loop has been used to demonstrate the temperature sensing capabilities of our interrogation system by changing temperature of Sensor Head 1 (${L_{PM{F_1}}} \approx 0.97m$) only, Sensor Head 2 (${L_{PM{F_2}}} \approx \textrm{2}.97m$) only, and both Sensor Head 1 and 2 individually. By monitoring frequency shift of corresponding passbands, temperature variations can be demodulated with the sensitivities of -0.275 ± 0.011 MHz/℃ and -0.811 ± 0.013 MHz/℃ for Sensor Head 1 and Sensor Head 2 which show a good linearity and stability. For the proposed sensing scheme, longer length of the sensing PMF will result in higher sensitivity, and by using Solc-Sagnac loop with higher order, more sensors can be multiplexed.

## 2. Operation principle and experimental setup

The schematic diagram of the second-order Solc-Sagnac interferometer is shown as Fig. 1, which comprises a 3dB fiber optic coupler (OC), two sections of PMFs, two polarization controllers (PCs), and three sections of standard single-mode ﬁber (SMF). Two PCs and two pieces of PMFs are connected with two output ports of the OC. The incident light enters the interferometer from port 1 (the input of the OC), and is split into two counter-propagating beams by the 3dB coupler, which are coupled at the same coupler again after propagating along the loop. Port 1 is also the reflection port of the interferometer, and the output from port 2 (another input of the OC) is the transmitted light.

The transmission characteristics can be analyzed by Jones Matrix [17,18]. For the second-order Solc-Sagnac interferometer, the reflection and the transmission can be expressed as [17]:

_{1}, PMF

_{2}, and can be simply given as:

One can see from Eq. (2) that there are four period components of transmission spectrum corresponding to the phase terms ${\phi _\textrm{1}},({\phi _\textrm{1}} - {\phi _\textrm{2}}),{\phi _\textrm{2}},({\phi _\textrm{1}}\textrm{ + }{\phi _\textrm{2}})$ caused by the birefringence, when the four phase differences for different wavelengths are equal to 2π, the corresponding wave-length spacings can be calculated as:

It can be observed that $\Delta {\lambda _1}$ and $\Delta {\lambda _3}$ are only related to ${L_{PM{F_1}}}$ and ${L_{PM{F_2}}}$ respectively, while $\Delta {\lambda _2}$ and $\Delta {\lambda _4}$ are affected by ${L_{PM{F_1}}}$ and ${L_{PM{F_2}}}$ collectively.

The second-order Solc-Sagnac loop described above can be taken as the sensing element as well as a spectrum slicer for a multi-passband MPF. Figure 2 portrays the schematic diagram of the proposed sensing interrogation system based on the multi-passband MPF with second-order Solc-Sagnac loop. The light from the broadband optical source (BOS) is sent into the port 1 of the second-order Solc-Sagnac loop (shown in the dashed box) after an isolator (ISO). Two section of PMFs are used as the Sensing Head 1 and 2, and Sensor Head 1 is placed in the thermostatic water bath for temperature control while Sensor Head 2 is put in a temperature controller . The transmission light from port 2 of the Solc-Sagnac loop is launched into a 90:10 optical coupler (OC), which splits it into two beams and 10% of the output is monitored by an optical spectrum analyzer (OSA), while the other 90% of light is modulated by an electro-optic modulator (EOM, JDSU 20 GHz). Then the modulated light is sent into a dispersion compensator fiber (DCF). After amplified by an erbium-doped optical fiber amplifier (EDFA), the signal is recovered by a photodiode (PD) and measured by the vector network analyzer (VNA).

The overall frequency response of the MPF can be expressed as [19]:

*D*and

*L*are the dispersion coefficient and the length of the DCF, respectively. If the length of PMFs is constant, the central frequency of the passband will be mainly affected by

*B*. The birefringence change of the PMF caused by temperature variations can be expressed as: where ${B_0}$ is a constant which is determined by the characteristic of the PMF and $\Delta T$ means the temperature change of PMF. When the environmental parameter such as temperature is changed, the change of birefringence ($\Delta B$) and an elongation ($\Delta {L_{PMF}}$) of PMF can be introduced. Since $\Delta {L_{PMF}} < < {L_{PMF}}$ [17], the thermal expansion effect can be ignored and the thermo-optic effect is considered [21,22]. Since PMF

_{1}and PMF

_{2}are the same kind of PMF, the relationship between the central frequency of the four passbands and temperature variations can be expressed as:

According to Eq. (6)-(8), one can see that by keeping the dispersion and length of DCF a constant, the central frequencies of the MPF’s passbands can be changed by varying the wavelength spacing of the interference patterns of the Solc-Sagnac loop, which is affected by the birefringence changes induced by the temperature variations. So, when the environmental parameters such as external temperature changes, the birefringence coefficient *B* of the PMF changes, and then the central frequencies of MPF’s passbands shift. Supposing $\frac{{{B_0}}}{{DL\lambda _0^2}} = M$ and $\frac{1}{{DL\lambda _0^2}}\frac{{dB}}{{dT}} = N$, Eq. (8) can be rewritten as:

When the length of two section of PMFs remain unchanged, the central frequency of the passbands is only related to temperature. From Eq. (9), one can see that ${f_2}$ and ${f_4}$ are associated with ${L_{PM{F_1}}}$ and ${L_{PM{F_\textrm{2}}}}$ collectively, while ${f_\textrm{1}}$ is only determined by ${L_{PM{F_1}}}$ and ${f_3}$ depends on ${L_{PM{F_2}}}$. Therefore, we can trace the temperature change of two section of PMFs by monitoring the frequency shift of ${f_\textrm{1}}$ and ${f_\textrm{3}}$, separately. The coefficient before ΔT, $\frac{{{L_{PMF}}}}{{DL\lambda _0^2}} \cdot \frac{{dB}}{{dT}}$ represents the relationship, i.e. the sensitivity. If the length of PMF is 3 m and the coefficient is $\frac{{dB}}{{dT}} ={-} \textrm{4}\textrm{.123} \times {10^{ - 7}}$[23], one can calculate that the sensitivity is -0.774 MHz/℃ when the dispersion of the DCF is -665.3 ps/nm (DL = 665.3 ps/nm) and λ_{0}≈1550 nm. If higher order Solc-Sagnac loop used, more sensors can be multiplexed, in this way, multiplexed sensing interrogation can be realized by the multi-passband MPF by using only one Solc-Sagnac loop. What’s more, one can see that the coefficient before $\Delta T$ determines the sensitivity of each passband, and by using PMFs with different lengths the sensitivity can be tailorable.

## 3. Experimental results and discussions

#### 3.1 Flexibly switchable multi-passband MPF

In our experiment, firstly we have tested the frequency response characteristics of the proposed the multi-passband MPF based on the second-order Solc-Sagnac loop and dispersive medium. The quasi-Gaussian spectral output from a BOS is sliced by second-order Solc-Sagnac loop, which consists of two PCs and two PMFs. In the experiment, we have selected two combinations of PMFs with different lengths (i.e. ${L_{PM{F_1}}} \approx 0.97m,{L_{PM{F_\textrm{2}}}} \approx 1.94m$ and ${L_{PM{F_1}}} \approx 0.97m,{L_{PM{F_\textrm{2}}}} \approx 2.97m$).

When choosing two PMFs with lengths 0.97 *m* and 1.94 *m*, and by adjusting the angle of two PCs (${\theta _2}\textrm{ = }\frac{\pi }{\textrm{4}},{\theta _1} + {\theta _3}\textrm{ = }\frac{\pi }{\textrm{4}}$), there are three passbands with different central frequencies of ∼0.223GHz, ∼0.447 GHz and ∼0.672 GHz, whose 3dB bandwidths are ∼0.041 GHz, ∼0.050 GHz and ∼0.068 GHz, respectively. Since the length of one piece of PMFs is twice of the other, we can obtain the central frequency ${f_1} \approx {f_2}$ according to the Eq. (6), and only three passbands can be obtained. By carefully adjusting two PCs in the loop, the MPF can be switched between three passbands, two passbands and one passband states, whose frequency responses are shown in Fig. 3.

In this regard, we have chosen two PMFs with the length of 0.97 *m* and 2.97 *m*. In this case, the central frequency ${f_1}$ is no longer equal to ${f_2}$, and four-passband MPF with the central frequencies of ∼0.223 GHz, ∼0.459 GHz, ∼0.685 GHz, ∼0.905 GHz has been realized, whose 3dB bandwidths are ∼0.042 GHz, ∼0.049 GHz, ∼0.066 GHz, ∼0.071 GHz, respectively. By carefully adjusting the PCs, the MPF can be switched between one passband, three passbands, four passbands and no passband states, whose frequency responses are shown in Fig. 4. Especially, the passbands ${f_1}$ and ${f_3}$ are corresponding to the length of PMF_{1} and PMF_{2}, respectively, and by tracing the frequency shift of the two passbands, environmental variations on PMF_{1} and PMF_{2} can be monitored separately. The design principle for the sensing interrogation system is that other passbands will not overlap with the passbands that have one-to-one relationship with the sensor heads (such as ${f_\textrm{1}}$ and ${L_{PM{F_1}}},{f_\textrm{3}}$ and ${L_{PM{F_2}}}$ respectively), which can be realized by carefully designing the length of PMFs. For higher order Solc-Sagnac loop, the filtering characteristics of the Solc-Sagnac interferometer become more flexible and the choices of the MPF passband state also become more diverse by adjusting the PCs and designing the length of PMFs. In this case, to ensure non-overlapping passbands, several aspects should be paid attention to, for example estimating the maximum temperature measurement range, tailoring the lengths of PMFs and adjusting the appropriate MPF state including the passbands with a one-to-one relationship with the sensor heads. In this case, sensing multiplexing can be realized by using only one Sagnac loop. And by employing Solc-Sagnac loop with higher order, more sensors can be multiplexed.

#### 3.2 Temperature sensing multiplexing based on the multi-passband MPF

In our experiment, the MPF with the PMF length of 0.97 *m* and 2.97 *m* in the second-order Solc-Sagnac loop is employed to implement the temperature sensing multiplexing. The PMFs ${L_{PM{F_1}}} \approx 0.97m$ and ${L_{PM{F_2}}} \approx 2.97m$ are used as Sensor Head 1 and 2, respectively. Firstly, The Sensor Head 1, PMF_{1} with the length of 0.97 *m* is put in a temperature controller whose temperature is controlled by the temperature control module (TCM) with temperature accuracy ±0.01 ℃. For Sensor Head 2, PMF_{2} with the length of 2.97 *m* is placed at room temperature (∼24℃). Figure 5(a) plots the frequency response of the MPF when the temperature of Sensor Head 1 is varied from 20°C to 70°C, while the temperature of Sensor Head 2 remains unchanged. The inset of Fig. 5 (a) shows the details of passband 1 to 4.

As can be observed in Fig. 5 (a), the MPF’s passbands of the central frequency located at ${f_1}$ and ${f_4}$ shift to the lower frequency when increased temperature is only applied to Sensor Head 1. The passband of the central frequency at ${f_2}$ shifts to the higher frequency and the frequency of passband at ${f_3}$ remains the same as the temperature of Sensor Head 2 keeps unchanged. Figure 5 (b) depicts the relationship between central frequency shifts of four passbands and temperature variations, which exhibits a good linearity. The sensitivities are estimated to be -0.275 ± 0.011 MHz/℃, 0.315 ± 0.016 MHz/℃, ∼0 MHz/℃, and -0.275 ± 0.030 MHz/℃ for the corresponding passbands at ${f_1},{f_{2,}}{f_3},{f_4}$, respectively, whose values are all calculated to three decimal places.

Then, we change the temperature of Sensor Head 2, while keeping the temperature of Sensor Head 1 a constant. The measured frequency response is shown as Fig. 6(a), and one can see that the MPF’s passbands of the central frequency at ${f_2},{f_3}$ and ${f_4}$ shift to the lower frequency when temperature of Sensor Head 2 increases from 20℃ to 70℃, while the frequency passband at ${f_1}$ keeps the same. The sensitivities for the corresponding passbands at ${f_1},{f_{2,}}{f_3},{f_4}$ are evaluated to be ∼0MHz/℃, -0.853 ± 0.057MHz/℃, -0.811 ± 0.013MHz/℃, -0.755 ± 0.020 MHz/℃ in Fig. 6(b), respectively. Therefore, it can be concluded that the cross-sensitivity of sensor heads is very small and can basically be ignored in our experiment. For both cases, the sensitivities for *f*_{2} and *f*_{4} are not the same with *f*_{1} or *f*_{3}, due to temperature fluctuations of room temperature environment at the spare sensor heads as well as errors in the data process and fitting. For sensing applications, only passbands corresponding to the PMFs (*f*_{1} and *f*_{3} in our experiment) need to be traced.

Figure 7 depicts the relationship between the central frequency shifts of two passband at ${f_1}$ and ${f_3}$ with temperature variations on Sensor Head 1 and Sensor Head 2, separately, which shows good agreement with the theoretical analysis above. One can see that sensor head with longer length enjoys higher sensitivity, as we have used the same kind PMF. As we can see that passbands at ${f_1}$ and ${f_3}$ are corresponding to only Sensor Head 1and only Sensor Head 2, respectively, by tracing the frequency shifts of these two passbands, multiplexing sensing scheme with two sensor heads can be realized. What’s more, by using Solc-Sagnac loop with higher order, more sensor heads can be multiplexed.

The interrogation for both PMFs used as temperature sensors has also been carried out, and Sensor Head 2 is put into a temperature controller, while Sensor Head 1 is immersed all in a thermostatic water bath to control the temperature. The measured frequency responses when the temperature of two sensor heads are set at different values is shown as Fig. 8(a). One can see from Fig. 8(a) that the central frequency located at ${f_\textrm{1}}$ shifts to the higher frequency with the falling temperature at Sensor Head 1 and the passband at ${f_\textrm{3}}$ shifts to the lower frequency when the temperature of Sensor Head 2 is increasing. The sensitivities for the corresponding passbands at ${f_1},{f_\textrm{3}}$ are evaluated to be -0.270 ± 0.010 MHz/℃ and -0.804 ±0.021 MHz/℃ in Fig. 8(b), respectively, which shows good agreement with the sensitivities when these sensor heads perform sensing individually. Also, it can be extended to a higher number of multiplexed sensors by using higher order Solc-Sagnac loop. In this way, we can use only one Sagnac loop to implement multiplexed sensing without cross talk.

In our experiment, the bandwidth of BOS is measured to be ∼39 nm and the loss of the Solc-Sagnac loop filter is measured to be ∼6.99 dB. Here we have used DCF with dispersion coefficient about -665 ps/nm at 1550 nm to act as the dispersive medium. And the MPF we proposed for temperature interrogation works in the low-frequency domain (<1 GHz), where the influence of high-order dispersion is very small and can be ignored [19]. What’s more, low-frequency devices also make the sensor system cost effective and beneficial for practical applications. In our experimental demonstration, we have used a VNA to trace the passbands’ peak variation, which is relatively expensive, however, in practical applications, a scanning signal generator and an RF dynamometer can be employed to further reduce the cost.

Figure 9 gives the optical spectrum of the second-order Solc-Sagnac interferometer with the PMF lengths of 0.97 *m* and 2.97 *m* and the corresponding frequency response of the MPF. One can see from Fig. 9(a) that, it has the limitation of periodic spectrum in optical domain and temperature measurement range is greatly limited. What’s more, the sliced optical spectrum has four different envelopes, and it is not easy to track the shifts when the temperature changes. Multiplexed sensing interrogation technique based on our proposed MPF shown in Fig. 9(b) can avoid these problems and expand the range of temperature measurement range as it converted the periodic interference pattern shifts in optical domain into the non-periodic passbands shift in electronic domain. In our experiment, when two Sensor Heads undergo temperature changes at the same time, its temperature measurement range (which is decided by the frequency spacing between passbands as well as the sensitivity.) is estimated to be ∼278°C, as the sensitivities for the passband of ${f_1}$ and ${f_3}$ are -0.275 MHz/℃ and -0.811 MHz/℃, respectively. When the temperature interrogation is in the state of the three passbands shown in Fig. 4(c) or (d), its temperature measurement range is estimated to be ∼858°C (the sensitivity for the passband of ${f_1}$ is -0.275 MHz/℃) or ∼569°C (the sensitivity for the passband of ${f_3}$ is -0.811 MHz/℃). When only one sensor experiences temperature changes, it can be adjusted to the state in Fig. 4(a) to have an unlimited measurement range. There are some limitations for the interrogation scheme with signal modulation and dispersion, which requires sufficient dispersion coefficient and stable performance of the dispersion module. It makes the system maintenance requirements more stringent. With the development of integrated microwave photonic technology, the problems will be solved in the future.

We have also carried out the stability test in the experiment, and the results are shown in Fig. 10, from which one can see that ${f_\textrm{1}},{f_\textrm{2}},{f_\textrm{3}}$ remain unchanged within 60 minutes, and the frequency fluctuations are 0 MHz, and the frequency instability at ${f_\textrm{4}}$ is about ±1 MHz, which will have little influence since we have used passbands at ${f_1}$ and ${f_3}$ (which are only related to Sensors Head 1-${L_{PM{F_1}}}$, Sensors Head 2-${L_{PM{F_2}}}$, respectively) for measurement. The SNR of the passbands at ${f_1},{f_{2,}}{f_3},{f_4}$ are greater than 14 dB, and power fluctuation is less than ±0.92 dB in our experiment. The small fluctuations in power will have little influence, since frequency demodulation technique is used in our sensing scheme.

As for the resolution, when the intermediate frequency (IF) band of the VNA is set to 100 Hz, the temperature resolutions by using passbands at ${f_1}$ and ${f_3}$ are estimated to be ∼3.6 × 10^{−4}℃,∼1.2 × 10^{−4}℃, as the sensitivities for passbands at ${f_1}$ and ${f_3}$ are -0.275 MHz/℃ and -0.811 MHz/℃, respectively. The sensing resolution of measurement wavelength shift in optical domain by an OSA can be estimated to 5.5 × 10^{−2}℃, if the sensitivity is -0.1822nm/°C [14]and the resolution restriction of our OSA is ∼0.01 nm (almost the best resolution of a commercial OSA we can get). To demonstrate the performance of this work more clearly, the comparison of the sensing performance between our scheme and other recently reported schemes is listed in Table 1. Compared with other temperature sensing scheme, our scheme effectively improves the sensing performance. i.e. the sensitivity and resolution. This method provides a low-cost and simple structure for temperature multiplexed sensing by using only one fiber interferometer.

## 4. Conclusions

In this work, a sensing interrogation technique based on a flexibly switchable multi-passband RF filter with a Solc-Sagnac loop has been proposed and experimentally demonstrated. By using MPF technology, the environmental variation induced interference pattern shifts in optical domain can be converted to the frequency shift of MPF’s passband in electronic domain, which facilitates the measurement. By adopting high-order Solc-Sagnac loop, multiplexing sensing interrogation system can be implemented by using only one Sagnac loop. In our experiment, a second-order Solc-Sagnac loop is employed and the sensitivities of -0.275 ± 0.011MHz/℃ and -0.811 ±0.013 MHz/℃ can be obtained by tracing the passbands at ${f_\textrm{1}}$ and ${f_\textrm{3}}$ which are corresponding to Sensor Head 1 (${L_{PM{F_1}}} \approx 0.97m$) and Sensor Head 2 (${L_{PM{F_2}}} \approx 2.97m$) respectively, when only Sensor Head 1 and only Sensor Head 2 experience temperature variations. Results show good linearity and longer sensing fiber gives higher sensing responsivity. If higher order Solc-Sagnac loop is used, more sensors can be simultaneously multiplexed. In summary, our proposed sensing interrogation technique shows good adjustability and stability, the ease of multiplexing and high sensitivity, resolution, which is suitable for application scenarios that simultaneously monitor and detect temperature changes at multiple points such as monitoring heat-prone parts in high-voltage power distribution devices [27–29] and detecting insulation materials of related components [30] in power systems.

## Funding

National Natural Science Foundation of China (61975167).

## 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, “On the cascade of incoherent discrete-time microwave photonic filters,” J. Lightwave Technol. **24**(7), 2564–2578 (2006). [CrossRef]

**2. **J. Capmany and D. Novak, “Microwave photonics combines two worlds,” Nat. Photonics **1**(6), 319–330 (2007). [CrossRef]

**3. **J. Yao, “Microwave Photonics,” J. Lightwave Technol. **27**(3), 314–335 (2009). [CrossRef]

**4. **X. Zou, W. Bai, W. Chen, P. Li, B. Lu, G. Yu, W. Pan, B. Luo, L. Yan, and L. Shao, “Microwave Photonics for Featured Applications in High-Speed Railways: Communications, Detection, and Sensing,” J. Lightwave Technol. **36**(19), 4337–4346 (2018). [CrossRef]

**5. **J. Capmany, B. Ortega, and D. Pastor, “A tutorial on microwave photonic filters,” J. Lightwave Technol. **24**(1), 201–229 (2006). [CrossRef]

**6. **Z. Jiao and J. Yao, “Passband-Switchable and Frequency-Tunable Dual-Passband Microwave Photonic Filter,” J. Lightwave Technol. **38**(19), 5333–5338 (2020). [CrossRef]

**7. **J. Ge and M. P. Fok, “Passband switchable microwave photonic multiband filter,” Sci. Rep. **5**(1), 15882 (2015). [CrossRef]

**8. **J. Ge, A. James, A. K. Mathews, and M. P. Fok, “Simultaneous 12-passband microwave photonic multiband filter with reconfigurable passband frequency,” 2016 Optical Fiber Communications Conference and Exhibition (OFC), W1G.2 (2016).

**9. **T. X. H. Huang, X. Yi, and R. A. Minasian, “Single passband microwave photonic filter using continuous-time impulse response,” Opt. Express **19**(7), 6231–6242 (2011). [CrossRef]

**10. **Y. Jiang, P. P. Shum, P. Zu, J. Zhou, G. Bai, J. Xu, Z. Zhou, H. Li, and S. Wang, “A Selectable Multiband Bandpass Microwave Photonic Filter,” IEEE Photonics J. **5**(3), 5500509 (2013). [CrossRef]

**11. **J. Hervas, A. L. Ricchiuti, W. Li, N. H. Zhu, C. R. Fernandez-Pousa, S. Sales, M. Li, and J. Capmany, “Microwave Photonics for Optical Sensors,” IEEE J. Sel. Top. Quantum Electron. **23**(2), 327–339 (2017). [CrossRef]

**12. **D. Xiao, L. Shao, C. Wang, W. Lin, F. Yu, G. Wang, T. Ye, W. Wang, and M. I. Vai, “Optical sensor network interrogation system based on nonuniform microwave photonic filters,” Opt. Express **29**(2), 2564–2576 (2021). [CrossRef]

**13. **L. Hua, Y. Song, B. Cheng, W. Zhu, Q. Zhang, and H. Xiao, “Coherence-length-gated distributed optical fiber sensing based on microwave-photonic interferometry,” Opt. Express **25**(25), 31362–31376 (2017). [CrossRef]

**14. **X. Chen, S. Yang, J. Yun, K. Wang, H. Fu, and N. Chen, “A Sensing Interrogation System for Sagnac Interferometer With Polarization Maintaining Fiber Utilizing Microwave Photonic Filtering Technique,” IEEE Sens. J. **20**(3), 1202–1207 (2020). [CrossRef]

**15. **J. Benitez, M. Bolea, and J. Mora, “Demonstration of multiplexed sensor system combining low coherence interferometry and microwave photonics,” Opt. Express **25**(11), 12182–12187 (2017). [CrossRef]

**16. **H. Chen, S. Zhang, H. Fu, B. Zhou, and N. Chen, “Sensing interrogation technique for fiber-optic interferometer type of sensors based on a single-passband RF filter,” Opt. Express **24**(3), 2765–2773 (2016). [CrossRef]

**17. **Y. Liu, B. Liu, X. Feng, W. Zhang, G. Zhou, S. Yuan, G. Kai, and X. Dong, “High-birefringence fiber loop mirrors and their applications as sensors,” Appl. Opt. **44**(12), 2382–2390 (2005). [CrossRef]

**18. **S. Yang, W. Dai, and H. Fu, “Multi-passband Microwave Photonic Filter with Flexible Passband,” 2019 Photonics & Electromagnetics Research Symposium - Fall (PIERS - Fall), 3329–3333 (2019)

**19. **J. Mora, B. Ortega, A. Diez, J. L. Cruz, M. V. Andres, J. Capmany, and D. Pastor, “Photonic microwave tunable single-bandpass filter based on a Mach-Zehnder interferometer,” J. Lightwave Technol. **24**(7), 2500–2509 (2006). [CrossRef]

**20. **H. Fu, K. Zhu, H. Ou, and S. He, “A tunable single-passband microwave photonic filter with positive and negative taps using a fiber Mach–Zehnder interferometer and phase modulation,” Opt. Laser Technol. **42**(1), 81–84 (2010). [CrossRef]

**21. **X. Cai, J. Luo, H. Fu, Y. Bu, and N. Chen, “Temperature measurement using a multi-wavelength fiber ring laser based on a hybrid gain medium and Sagnac interferometer,” Opt. Express **28**(26), 39933–39943 (2020). [CrossRef]

**22. **J. Shi, Y. Wang, D. Xu, H. Zhang, G. Su, L. Duan, C. Yan, D. Yan, S. Fu, and J. Yao, “Temperature Sensor Based on Fiber Ring Laser With Sagnac Loop,” IEEE Photonics Technol. Lett. **28**(7), 794–797 (2016). [CrossRef]

**23. **Z. Li, X. S. Yao, X. Chen, H. Chen, Z. Meng, and T. Liu, “Complete Characterization of Polarization-Maintaining Fibers Using Distributed Polarization Analysis,” J. Lightwave Technol. **33**(2), 372–380 (2015). [CrossRef]

**24. **O. Frazao, J. L. Santos, and J. M. Baptista, “Strain and Temperature Discrimination Using Concatenated High-Birefringence Fiber Loop Mirrors,” IEEE Photonics Technol. Lett. **19**(16), 1260–1262 (2007). [CrossRef]

**25. **N. Wu, M. Xia, Y. Wu, S. Li, R. Qi, Y. Huang, and L. Xia, “Microwave photonics interrogation for multiplexing fiber Fabry-Perot sensors,” Opt. Express **29**(11), 16652–16664 (2021). [CrossRef]

**26. **D. Leandro, M. Bravo, and M. Lopez-Amo, “High resolution polarization-independent high-birefringence fiber loop mirror sensor,” Opt. Express **23**(24), 30985–30990 (2015). [CrossRef]

**27. **N. K. Mukhopadhyay, B. K. Dutta, and H. S. Kushwaha, “On-line fatigue–creep monitoring system for high-temperature components of power plants,” Int. J. Fatigue **23**(6), 549–560 (2001). [CrossRef]

**28. **S. Nandi, H. A. Toliyat, and X. Li, “Condition Monitoring and Fault Diagnosis of Electrical Motors—A Review,” IEEE Trans. Energy Convers. **20**(4), 719–729 (2005). [CrossRef]

**29. **G. Stone, “Advancements during the Past Quarter Century in On-line Monitoring of Motor and Generator Winding Insulation,” IEEE Trans. Dielectr. Electr. Insul. **9**(5), 746–751 (2002). [CrossRef]

**30. **Y. Han and Y. H. Song, “Condition monitoring techniques for electrical equipment-a literature survey,” IEEE Trans. Power Deliv. **18**(1), 4–13 (2003). [CrossRef]