*3.1. Performence of Channel Optical Power Monitoring*

In order to alleviate the influence of the photoelectric conversion characteristics of different PDs and different modulation depths on the accuracy of channel optical power monitoring, we first simulated the curve relationship of the calibration coefficient with the modulation depth in the QPSK/16QAM system. The modulation depth of the QPSK/16QAM system is increased from 2% to 15%. Then, the relationship between the calibration coefficient and the modulation depth is shown in Figure 4.

tion.

*3.1. Performence of Channel Optical Power Monitoring* 

According to Table 3, we find that the optical label performance in QPSK system is

In order to alleviate the influence of the photoelectric conversion characteristics of

different PDs and different modulation depths on the accuracy of channel optical power monitoring, we first simulated the curve relationship of the calibration coefficient with the modulation depth in the QPSK/16QAM system. The modulation depth of the QPSK/16QAM system is increased from 2% to 15%. Then, the relationship between the

calibration coefficient and the modulation depth is shown in Figure 4.

better than that in 16QAM system, and the optical label power basically exhibits a linear relationship as the modulation depth increases. Optical label power and modulation depth *m* are basically linear, and the label information could be recovered accurately. In addition, the SNR of optical label under different *m* is sufficient for error-free demodula-

**Figure 4.** The calibration coefficient versus modulation depth. **Figure 4.** The calibration coefficient versus modulation depth.

Figure 3 shows that as the modulation depth increases, the calibration coefficient of the QPSK system is relatively constant with little fluctuation, but it increases first and then gradually becomes constant as the modulation depth increases in the 16QAM system. The main reasons are as follows: For the constant-module QPSK signals, the beat noise generated by PD detection is relatively small, which means that even at a low modulation depth, the noise only accounts for a small part of the entire optical label power and causes Figure 3 shows that as the modulation depth increases, the calibration coefficient of the QPSK system is relatively constant with little fluctuation, but it increases first and then gradually becomes constant as the modulation depth increases in the 16QAM system. The main reasons are as follows: For the constant-module QPSK signals, the beat noise generated by PD detection is relatively small, which means that even at a low modulation depth, the noise only accounts for a small part of the entire optical label power and causes little effect on the optical label power. As the modulation depth increases, the optical label power also increases almost linearly. Therefore, the change of modulation depth hardly brings about the change of calibration coefficient.

little effect on the optical label power. As the modulation depth increases, the optical label power also increases almost linearly. Therefore, the change of modulation depth hardly brings about the change of calibration coefficient. However, for the 16QAM system with non-constant modulus, larger beat noise will be generated after PD detection. This noise accounts for a ratio of the entire optical label power at low modulation depths, even exceeding that of the digital label signal modulated by the pilot tone. This results in the optical label power being greater than the expected power at the corresponding modulation depth. In the case of monitoring the optical power of the same channel, the calibration factor is small compared with QPSK systems. The PD beat noise is almost constant or slightly increases with increasing modulation depth, but However, for the 16QAM system with non-constant modulus, larger beat noise will be generated after PD detection. This noise accounts for a ratio of the entire optical label power at low modulation depths, even exceeding that of the digital label signal modulated by the pilot tone. This results in the optical label power being greater than the expected power at the corresponding modulation depth. In the case of monitoring the optical power of the same channel, the calibration factor is small compared with QPSK systems. The PD beat noise is almost constant or slightly increases with increasing modulation depth, but the power value of the pilot-modulated useful digital label increases linearly, resulting in a reduced effect of PD beat noise accompanied by an increase in calibration coefficients. This process gradually flattens out as the power of the digital label signal increases. When the modulation depth is increased to 15%, the effect of the beat noise is small, reaching a situation similar to that of the QPSK system, so the calibration coefficients of the two systems are almost the same at this time.

the power value of the pilot-modulated useful digital label increases linearly, resulting in a reduced effect of PD beat noise accompanied by an increase in calibration coefficients. This process gradually flattens out as the power of the digital label signal increases. When the modulation depth is increased to 15%, the effect of the beat noise is small, reaching a situation similar to that of the QPSK system, so the calibration coefficients of the two sys-The maximum absolute errors of optical power monitoring at different modulation depths on 16QAM and QPSK systems after 20 spans fiber transmission are shown in Figure 5. It shows that the scheme of optical power monitoring using spectrum integration is feasible, and the maximum channel power monitoring error is less than 0.6 dB. The QPSK and the 16QAM systems exhibit same performance, and the maximum monitoring error is similar at different label modulation depths. Therefore, the monitoring accuracy of two systems can be considered to be roughly the same.

tems are almost the same at this time.

0.0

0.3

0.6

**absolute error (dB)**

**Channel power monitoring** 

0.9

1.2

1.5

The maximum absolute errors of optical power monitoring at different modulation

*Photonics* **2022**, *9*, x FOR PEER REVIEW 10 of 16

The maximum absolute errors of optical power monitoring at different modulation

0.6

depths on 16QAM and QPSK systems after 20 spans fiber transmission are shown in Figure 5. It shows that the scheme of optical power monitoring using spectrum integration is feasible, and the maximum channel power monitoring error is less than 0.6 dB. The QPSK and the 16QAM systems exhibit same performance, and the maximum monitoring error is similar at different label modulation depths. Therefore, the monitoring accuracy of two

depths on 16QAM and QPSK systems after 20 spans fiber transmission are shown in Figure 5. It shows that the scheme of optical power monitoring using spectrum integration is feasible, and the maximum channel power monitoring error is less than 0.6 dB. The QPSK and the 16QAM systems exhibit same performance, and the maximum monitoring error

systems can be considered to be roughly the same.

systems can be considered to be roughly the same.

0.6

**Figure 5.** The optical power monitoring results of (**a**) 8-channel QPSK, (**b**) 8-channel 16QAM system. **Figure 5.** The optical power monitoring results of (**a**) 8-channel QPSK, (**b**) 8-channel 16QAM system. In addition, taking a QPSK system as an example, the impact of channel number on

In addition, taking a QPSK system as an example, the impact of channel number on the accuracy of channel power monitoring is also investigated. Here, the optical label modulation depth is fixed to 0.10, while the number of monitoring channels is set from 5 to 64. The monitoring error of channel optical power of different wavelength channel In addition, taking a QPSK system as an example, the impact of channel number on the accuracy of channel power monitoring is also investigated. Here, the optical label modulation depth is fixed to 0.10, while the number of monitoring channels is set from 5 to 64. The monitoring error of channel optical power of different wavelength channel numbers after 20-span transmission is shown in Figure 6. the accuracy of channel power monitoring is also investigated. Here, the optical label modulation depth is fixed to 0.10, while the number of monitoring channels is set from 5 to 64. The monitoring error of channel optical power of different wavelength channel numbers after 20-span transmission is shown in Figure 6.

**Number of Channel Figure 6.** The power monitoring error with different channel numbers. **Figure 6.** The power monitoring error with different channel numbers.

**Figure 6.** The power monitoring error with different channel numbers. It can be seen form Figure 6 that when the number of channels is increased from 5 to 64, the monitoring error also increases gradually. The main reason is that as the number of wavelength channels increases, the launch power increases accordingly, and the fiber Kerr nonlinear effect increases, which deteriorates the signal quality of the optical label. In addition, as the number of wavelength channels increases, the spectrum range increases; thus, the inter-channel crosstalk caused by the SRS effect is more serious, and consequently, it will also lead to deterioration of accuracy of optical power monitoring. Moreover, we can see that the monitoring errors fluctuate in a relatively small range. This is because in each simulation with different channel numbers, the optical label sequence is randomly generated, and the random change of the label sequence characteristics It can be seen form Figure 6 that when the number of channels is increased from 5 to 64, the monitoring error also increases gradually. The main reason is that as the number of wavelength channels increases, the launch power increases accordingly, and the fiber Kerr nonlinear effect increases, which deteriorates the signal quality of the optical label. In addition, as the number of wavelength channels increases, the spectrum range increases; thus, the inter-channel crosstalk caused by the SRS effect is more serious, and consequently, it will also lead to deterioration of accuracy of optical power monitoring. Moreover, we can see that the monitoring errors fluctuate in a relatively small range. This is because in each simulation with different channel numbers, the optical label sequence is randomly generated, and the random change of the label sequence characteristics causes the monitoring results to fluctuate slightly. In general, although there is a rising It can be seen form Figure 6 that when the number of channels is increased from 5 to 64, the monitoring error also increases gradually. The main reason is that as the number of wavelength channels increases, the launch power increases accordingly, and the fiber Kerr nonlinear effect increases, which deteriorates the signal quality of the optical label. In addition, as the number of wavelength channels increases, the spectrum range increases; thus, the inter-channel crosstalk caused by the SRS effect is more serious, and consequently, it will also lead to deterioration of accuracy of optical power monitoring. Moreover, we can see that the monitoring errors fluctuate in a relatively small range. This is because in each simulation with different channel numbers, the optical label sequence is randomly generated, and the random change of the label sequence characteristics causes the monitoring results to fluctuate slightly. In general, although there is a rising trend in the error of channel power monitoring, the maximum monitoring error after 20-span transmission does not exceed 0.9 dB, so it is believed that our proposed scheme is applicable to WDM optical networks.

trend in the error of channel power monitoring, the maximum monitoring error after 20-

causes the monitoring results to fluctuate slightly. In general, although there is a rising trend in the error of channel power monitoring, the maximum monitoring error after 20-

### *3.2. Performence of OSNR Estimation* In order to validate the performance of OSNR estimation practically, the NF value of

span transmission does not exceed 0.9 dB, so it is believed that our proposed scheme is

*Photonics* **2022**, *9*, x FOR PEER REVIEW 11 of 16

applicable to WDM optical networks.

*3.2. Performence of OSNR Estimation* 

In order to validate the performance of OSNR estimation practically, the NF value of EDFA in each optical channel is set to 5 dB, typically. Meanwhile, the length of each span is identically set to 100 km in 20-span transmission. The input power and output power of EDFAs are monitored based on optical labels, and the OSNR in each channel is calculated by using the proposed method. Then, 16 G Baud PM-16QAM and PM-QPSK 8, 32 and 64 channels WDM transmission simulations are carried out. The result of OSNR estimation as well as the estimate error after 20-span transmission are shown in Figure 7. EDFA in each optical channel is set to 5 dB, typically. Meanwhile, the length of each span is identically set to 100 km in 20-span transmission. The input power and output power of EDFAs are monitored based on optical labels, and the OSNR in each channel is calculated by using the proposed method. Then, 16 G Baud PM-16QAM and PM-QPSK 8, 32 and 64 channels WDM transmission simulations are carried out. The result of OSNR estimation as well as the estimate error after 20-span transmission are shown in Figure 7.

**Figure 7.** Result of OSNR estimation and corresponding error under 100 km equal-length span in (**a**) 8-channel QPSK, (**b**) 8-channel 16QAM, (**c**) 32-channel QPSK, (**d**) 32-channel 16QAM systems, (**e**) 64-channel QPSK systems, and (**f**) 64-channel 16QAM system. **Figure 7.** Result of OSNR estimation and corresponding error under 100 km equal-length span in (**a**) 8-channel QPSK, (**b**) 8-channel 16QAM, (**c**) 32-channel QPSK, (**d**) 32-channel 16QAM systems, (**e**) 64-channel QPSK systems, and (**f**) 64-channel 16QAM system.

could be estimated accurately in each span both in 8/32/64-channels QPSK and 16QAM

Figure 7 shows that the scheme of OSNR estimation is feasible because that OSNR

Figure 7 shows that the scheme of OSNR estimation is feasible because that OSNR could be estimated accurately in each span both in 8/32/64-channels QPSK and 16QAM systems. It shows that the OSNR monitoring error after 20 span transmissions is less than 0.45 dB in both 8-channel PM-QPSK/16QAM systems, while it is about 0.7 dB in 32-channel PM-QPSK/16QAM systems. When the number of transmission channels increases to 64, the OSNR monitoring error after 20 span transmissions is less than 1 dB both in 64-channel PM-QPSK/16QAM systems. The main monitoring error occurs because of the error that occurred during the previous channel power monitoring. With the increase of transmission spans, the ASE noise induced by EDFA continues to accumulate due to multiple cascade amplification in the transmission link, resulting in an increase of channel power monitoring and thereby causing larger OSNR estimation error. In addition, the nonlinear effects and inter-channel cross talk also become more serious as the number of channels increases, which leads to a larger error in OSNR monitoring. Furthermore, other factors such as chromatic dispersion, polarization dependent impairments and so on will also affect the accuracy of power monitoring, which leads to a situation where the longer the transmission distance is, the greater the monitoring error is. 0.45 dB in both 8-channel PM-QPSK/16QAM systems, while it is about 0.7 dB in 32-channel PM-QPSK/16QAM systems. When the number of transmission channels increases to 64, the OSNR monitoring error after 20 span transmissions is less than 1 dB both in 64 channel PM-QPSK/16QAM systems. The main monitoring error occurs because of the error that occurred during the previous channel power monitoring. With the increase of transmission spans, the ASE noise induced by EDFA continues to accumulate due to multiple cascade amplification in the transmission link, resulting in an increase of channel power monitoring and thereby causing larger OSNR estimation error. In addition, the nonlinear effects and inter-channel cross talk also become more serious as the number of channels increases, which leads to a larger error in OSNR monitoring. Furthermore, other factors such as chromatic dispersion, polarization dependent impairments and so on will also affect the accuracy of power monitoring, which leads to a situation where the longer the transmission distance is, the greater the monitoring error is. Meanwhile, to make the simulation more realistic and reliable, the span lengths of

Meanwhile, to make the simulation more realistic and reliable, the span lengths of the 20-spans transmissions in the 8-channel system are set to 100, 60, 80, 40, 100, 50, 70, 20, 100, 60, 30, 70, 50, 80, 20, 100, 40, 100, 30 and 80 km, respectively. Thus, it is convincing to evaluate practically the OSNR estimation and corresponding errors of the target channel 1 (NF = 4) and channel 3 (NF = 5). Figure 8 shows that the maximum OSNR monitoring error is about 0.3 dB after 20-span transmission in both PM-QPSK/16QAM WDM systems. The scheme still works in the case of different span lengths, and the OSNR estimation error also tends to increase for the same reason. However, due to that the total transmission length here is reduced, resulting in a reduction of channel power monitoring, and it further makes the OSNR estimation error smaller than the results in Figure 7a,b. the 20-spans transmissions in the 8-channel system are set to 100, 60, 80, 40, 100, 50, 70, 20, 100, 60, 30, 70, 50, 80, 20, 100, 40, 100, 30 and 80 km, respectively. Thus, it is convincing to evaluate practically the OSNR estimation and corresponding errors of the target channel 1(NF = 4) and channel 3(NF = 5). Figure 8 shows that the maximum OSNR monitoring error is about 0.3 dB after 20-span transmission in both PM-QPSK/16QAM WDM systems. The scheme still works in the case of different span lengths, and the OSNR estimation error also tends to increase for the same reason. However, due to that the total transmission length here is reduced, resulting in a reduction of channel power monitoring, and it further makes the OSNR estimation error smaller than the results in Figure 7a,b.

**Figure 8.** Result of OSNR estimation and corresponding error under [20–100 km] unequal-length span in (**a**,**b**) QPSK and (**c**,**d**) 16QAM systems. **Figure 8.** Result of OSNR estimation and corresponding error under [20–100 km] unequal-length span in (**a**,**b**) QPSK and (**c**,**d**) 16QAM systems.

optical IQ modulator, and arbitrary waveform generator (AWG) available, only single-

To further verify the actual performance of the proposed scheme, the transmission

**4. Experimental Setup and Performance Analysis** 

### **4. Experimental Setup and Performance Analysis** *Photonics* **2022**, *9*, x FOR PEER REVIEW 13 of 16 random bit sequences of service signal are generated and mapped into PM-QPSK/16QAM

To further verify the actual performance of the proposed scheme, the transmission experiment is carried out. Figure 9 shows the experimental setup of optical-label based OPM over PM-QPSK and PM-16QAM systems. Since there is only one set of laser source, optical IQ modulator, and arbitrary waveform generator (AWG) available, only single-channel optical fiber transmission and monitoring verification is carried out in this experiment. In addition, limited by the storage depth and sampling rate of the 8190A instrument (up to 12 GSa/s), the service signal rate in the offline experiment is set to 4 G Baud, and the modulation depth of the optical tag to the service signal is set to 0.10. Firstly, random bit sequences of service signal are generated and mapped into PM-QPSK/16QAM symbols. Then, we modulate the 2 Mbit/s ID information [1, 1, 1, 0, 0, 0, 1, 1, 1, 0, 0, 0, 1, 1, 1] onto a 40 MHz pilot tone and load the pilot-tone-carried labels onto the service signals by offline DSP. Next, the discrete signals after the matched filter are sent into AWG (M8190A) working at 400 MSa/s. After 80 km standard single-mode fiber (SSMF) transmission and EDFA amplification, an adjustable optical attenuator is used to change the optical power of the channel within the PD's acceptable working range of [−20–0 dBm]. A spectrometer (AQ6370) is used to observe the optical signal spectrum, while a 200 MHz-bandwidth PD is used to detect the optical labels. Then the detected 16QAM and QPSK signals are sampled by a real-time oscilloscope with 400 MSa/s for offline processing. channel optical fiber transmission and monitoring verification is carried out in this experiment. In addition, limited by the storage depth and sampling rate of the 8190A instrument (up to 12 GSa/s), the service signal rate in the offline experiment is set to 4 G Baud, and the modulation depth of the optical tag to the service signal is set to 0.10. Firstly, random bit sequences of service signal are generated and mapped into PM-QPSK/16QAM symbols. Then, we modulate the 2 Mbit/s ID information [1, 1, 1, 0, 0, 0, 1, 1, 1, 0, 0, 0, 1, 1, 1] onto a 40 MHz pilot tone and load the pilot-tone-carried labels onto the service signals by offline DSP. Next, the discrete signals after the matched filter are sent into AWG (M8190A) working at 400 MSa/s. After 80 km standard single-mode fiber (SSMF) transmission and EDFA amplification, an adjustable optical attenuator is used to change the optical power of the channel within the PD's acceptable working range of [−20–0 dBm]. A spectrometer (AQ6370) is used to observe the optical signal spectrum, while a 200 MHzbandwidth PD is used to detect the optical labels. Then the detected 16QAM and QPSK signals are sampled by a real-time oscilloscope with 400 MSa/s for offline processing. symbols. Then, we modulate the 2 Mbit/s ID information [1, 1, 1, 0, 0, 0, 1, 1, 1, 0, 0, 0, 1, 1, 1] onto a 40 MHz pilot tone and load the pilot-tone-carried labels onto the service signals by offline DSP. Next, the discrete signals after the matched filter are sent into AWG (M8190A) working at 400 MSa/s. After 80 km standard single-mode fiber (SSMF) transmission and EDFA amplification, an adjustable optical attenuator is used to change the optical power of the channel within the PD's acceptable working range of [−20–0 dBm]. A spectrometer (AQ6370) is used to observe the optical signal spectrum, while a 200 MHzbandwidth PD is used to detect the optical labels. Then the detected 16QAM and QPSK signals are sampled by a real-time oscilloscope with 400 MSa/s for offline processing. **Double Polarization IQ Modulator CW Laszer 193.1THz Service Data and Optical EDFA & SSMF 1:99 Coupler &Attenuator**

*Photonics* **2022**, *9*, x FOR PEER REVIEW 13 of 16

channel optical fiber transmission and monitoring verification is carried out in this experiment. In addition, limited by the storage depth and sampling rate of the 8190A instrument (up to 12 GSa/s), the service signal rate in the offline experiment is set to 4 G Baud, and the modulation depth of the optical tag to the service signal is set to 0.10. Firstly,

**200MHz**

**400MSa/s Oscilloscope**

**Offline DSP** 

**Figure 9.** Diagram of experimental platform. **Figure 9.** Diagram of experimental platform. shown in Figure 10. It shows that the transmitted DPSK digital optical labels, i.e., ID in-

The waveform of the optical label after differential demodulation in offline DSP is shown in Figure 10. It shows that the transmitted DPSK digital optical labels, i.e., ID information of a wavelength channel, could be accurately recovered without error by using a low-bandwidth PD and low-sample-rate ADC with the help of DSP processing. Furthermore, we can see that the signal quality of labels in QPSK system is better than that of 16QAM. This is due to the fact that after PD reception, the amount of beat noise of QPSK constant-module signals is smaller than that of 16QAM, thereby making the QPSK system The waveform of the optical label after differential demodulation in offline DSP is shown in Figure 10. It shows that the transmitted DPSK digital optical labels, i.e., ID information of a wavelength channel, could be accurately recovered without error by using a low-bandwidth PD and low-sample-rate ADC with the help of DSP processing. Furthermore, we can see that the signal quality of labels in QPSK system is better than that of 16QAM. This is due to the fact that after PD reception, the amount of beat noise of QPSK constant-module signals is smaller than that of 16QAM, thereby making the QPSK system a better optical label signal-to-noise ratio when demodulating the labels. formation of a wavelength channel, could be accurately recovered without error by using a low-bandwidth PD and low-sample-rate ADC with the help of DSP processing. Furthermore, we can see that the signal quality of labels in QPSK system is better than that of 16QAM. This is due to the fact that after PD reception, the amount of beat noise of QPSK constant-module signals is smaller than that of 16QAM, thereby making the QPSK system a better optical label signal-to-noise ratio when demodulating the labels.

The result of power monitoring in experiments is shown in Figure 11. The result **Figure 10.** The demodulated waveform of optical labels in (**a**) QPSK and (**b**) 16QAM systems. **Figure 10.** The demodulated waveform of optical labels in (**a**) QPSK and (**b**) 16QAM systems.

shows that the optical label power can be calculated more accurately, and then, the accurate channel optical power monitoring result can be obtained. The power monitoring error of the QPSK system is similar to that of the 16QAM system, which are both less than 0.3

The result of power monitoring in experiments is shown in Figure 11. The result

shows that the optical label power can be calculated more accurately, and then, the accu-

of the QPSK system is similar to that of the 16QAM system, which are both less than 0.3 dB under 80 km fiber transmission. It is important to point out that the multi-span transmission cannot be fully experimented with, resulting in better channel power monitoring

The result of power monitoring in experiments is shown in Figure 11. The result shows that the optical label power can be calculated more accurately, and then, the accurate channel optical power monitoring result can be obtained. The power monitoring error of the QPSK system is similar to that of the 16QAM system, which are both less than 0.3 dB under 80 km fiber transmission. It is important to point out that the multi-span transmission cannot be fully experimented with, resulting in better channel power monitoring performance than that of our simulation. Moreover, although the error of power monitoring fluctuates irregularly within a range, the accurate channel optical power monitoring performance is sufficient to make it a practical OPM solution. *Photonics* **2022**, *9*, x FOR PEER REVIEW 14 of 16 performance than that of our simulation. Moreover, although the error of power monitoring fluctuates irregularly within a range, the accurate channel optical power monitoring performance is sufficient to make it a practical OPM solution.

**Figure 11.** Results of channel power monitoring and monitoring error (**a**) in QPSK system, and (**b**) **Figure 11.** Results of channel power monitoring and monitoring error (**a**) in QPSK system, and (**b**) in 16QAM system.

### **5. Conclusions**

read and agreed to the published version of the manuscript.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

in 16QAM system.

**5. Conclusions** 

(2022RC09).

going study.

In this paper, a low-cost and highly efficient optical labels enabled channel power monitoring and OSNR estimation of WDM system has been proposed and demonstrated. DPSK digital optical labels modulated on pilot tones with frequency of tens of MHz were applied as identity indicators and performance predictors tied up to each wavelength channel. The system can not only monitor the performance of all wavelength channels simultaneously using low-cost PD detection, but it is also able to reliably deliver any specific surveillance information to further sense the working state of WDM channels. In addition, in our proposed DSP processing in OPM modules, the channel optical power is monitored using the method of spectral integration, and the OSNR is further estimated by calculating the ASE noise accumulation form all amplifiers. The simulation results of 8, 32 and 64 channels WDM systems under 20 spans transmission show that the maximum monitoring error of channel optical power and the estimation error of OSNR are both less than 1 dB in 16 G Baud PM-QPSK and PM-16QAM systems, respectively. Furthermore, an offline experiment platform was constructed by using a PD with 300 MHz bandwidth and an ADC with 600 MSa/s sample rate, and the results show that the DPSK digital labels In this paper, a low-cost and highly efficient optical labels enabled channel power monitoring and OSNR estimation of WDM system has been proposed and demonstrated. DPSK digital optical labels modulated on pilot tones with frequency of tens of MHz were applied as identity indicators and performance predictors tied up to each wavelength channel. The system can not only monitor the performance of all wavelength channels simultaneously using low-cost PD detection, but it is also able to reliably deliver any specific surveillance information to further sense the working state of WDM channels. In addition, in our proposed DSP processing in OPM modules, the channel optical power is monitored using the method of spectral integration, and the OSNR is further estimated by calculating the ASE noise accumulation form all amplifiers. The simulation results of 8, 32 and 64 channels WDM systems under 20 spans transmission show that the maximum monitoring error of channel optical power and the estimation error of OSNR are both less than 1 dB in 16 G Baud PM-QPSK and PM-16QAM systems, respectively. Furthermore, an offline experiment platform was constructed by using a PD with 300 MHz bandwidth and an ADC with 600 MSa/s sample rate, and the results show that the DPSK digital labels could be accurately recovered with the help of the proposed DSP processing, while the monitoring error of channel optical power is less than 0.3 dB. Therefore, the advantages of low cost and high efficiency make our scheme more pragmatic and more robust and therefore easier to implement and more practical for WDM system applications.

could be accurately recovered with the help of the proposed DSP processing, while the monitoring error of channel optical power is less than 0.3 dB. Therefore, the advantages of low cost and high efficiency make our scheme more pragmatic and more robust and therefore easier to implement and more practical for WDM system applications. **Author Contributions:** Conceptualization, T.Y. and X.C.; formal analysis, T.Y., K.L. and Z.L.; investigation, T.Y. and S.S.; data curation, T.Y. and X.C.; writing—original draft preparation, T.Y., K.L. and X.W.; writing—review and editing, T.Y., L.W. and X.C.; visualization, T.Y. and K.L.; supervision, T.Y., K.L., Z.L., X.W., S.S., L.W. and X.C.; project administration, T.Y. and X.C. All authors have read and agreed to the published version of the manuscript.

**Author Contributions:** Conceptualization, T.Y. and X.C.; formal analysis, T.Y., K.L. and Z.L.; investigation, T.Y. and S.S.; data curation, T.Y. and X.C.; writing—original draft preparation, T.Y., K.L. and X.W.; writing—review and editing, T.Y., L.W. and X.C.; visualization, T.Y. and K.L.; supervi-**Funding:** This work is partly supported by National Natural Science Foundation of China (62001045), Beijing Municipal Natural Science Foundation (4214059), Fund of State Key Laboratory of IPOC (BUPT) (IPOC2021ZT17), and Fundamental Research Funds for the Central Universities (2022RC09).

(62001045), Beijing Municipal Natural Science Foundation (4214059), Fund of State Key Laboratory of IPOC (BUPT) (IPOC2021ZT17), and Fundamental Research Funds for the Central Universities

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author. The data are not publicly available due to the data also forms part of an on-

sion, T.Y., K.L., Z.L., X.W., S.S., L.W. and X.C.; project administration, T.Y. and X.C. All authors have

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author. The data are not publicly available due to the data also forms part of an ongoing study.

**Acknowledgments:** The authors express their appreciation to reviewers for their valuable suggestions.

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**

