Open Access
Add to BibSonomy
Abstract
A pulsed fiber master oscillator power amplifier (MOPA), which is combined with second harmonic generation (SHG) and modulated by directly changing the current of the low-power seed laser, is designed in this paper to overcome the ‘green gap’ of semiconductor lasers and the difficulty of obtaining high-power and wide-bandwidth driving circuits. To decrease the guard slot and increase the data rate of a high-order pulse position modulation (PPM) system, pre-pulse shaping (PPS) is utilized to decrease the fluctuation of pulse power, which is caused by the gain dynamics of multi-order amplification of the MOPA, from 55.6% to 27.5% for 25-ns pulses and from 22.4% to 16.7% for 10-ns pulses, respectively. Moreover, an analog PPM demodulation method is proposed to mitigate the nonlinear effect caused by space charge limitations at dynodes of a photomultiplier tube (PMT) and increase the robustness of the system. In an optical darkroom, a 99-m 64-PPM UWOC transmission, of which the measured link loss is around 13.16 attenuation length (AL), is realized in a water tank with a data rate of 9.14 Mbps. The average received optical power ranges from −60.87 to −52.51 dBm, corresponding to a bit error rate (BER) range of 1.93 × 10−4 to 2.3 × 10−3. To further prove the reliability of the proposed system, we implement a 65-m UWOC experiment with the same data rate at a BER of 3.42 × 10−4 in a 50-m standard swimming pool. The maximum link loss is measured to be 15 AL.
© 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Underwater wireless optical communication (UWOC) has recently attracted increasing attention from both academic and industrial communities [1–3] owing to its predominant advantages of high bandwidth, low time delay and compact structure compared with underwater acoustic communication. For the light source design in an UWOC system, the optimal wavelength varies with water quality, and the green band is suitable for relatively turbid waters. Limited by the ‘green gap’ of semiconductor laser diodes [4], the power of green laser diodes is relatively low compared with blue laser diodes, so they were usually used as an independent channel of wavelength division multiplex to increase data rates or a separate upstream/downstream link [5–8]. Besides, digital signal processing methods, such as the frequency domain equalization, noise prediction or nonlinear equalization, were used to promote the data rate of UWOC system [9,10] based on green laser diodes. The maximum data rate was promoted to several Gbps but the transmission distance was limited. Until now, the largest attenuation distance based on green laser diodes was 14.3 attenuation length (AL), benefited from the outstanding design of optical antenna and the high-sensitivity detector [11].
Other common green lasers such as diode pumping solid state lasers (DPSSLs) and master oscillator power amplifiers (MOPAs) combined with second harmonic generation (SHG) are also used in UWOC systems. In as early as 2008, a MOPA with the output green light power of 8.6 dBm was utilized in a 2-Gbps on-off keying (OOK) UWOC system through a 2-m underwater channel [12]. In 2017, Kong et al. used a DPSSL to achieve a net bit rate of 108.55 Mbps for the 64-quadrature amplitude modulation (QAM) signal over a 2-m underwater channel [13]. In 2020, Yang et al. used a MOPA with an external modulator to achieve a 100-m/100-Mbps transmission in tap water with an attenuation coefficient of 0.73 dB/m. The output 532-nm light power was 1.4 W [14]. Compared with continuous-wave lasers, pulsed lasers see higher SHG efficiency at the same average power, and their high peak powers are more suitable for a long-distance transmission. Pulsed lasers could be combined with high-order pulse position modulation (PPM) to transfer information. An L-PPM symbol consists of L (= 2M, where bit resolution M > 0 is an integer) possible time slots, where one slot is occupied by a pulse and the remaining slots are empty. The position of the pulse in the symbol corresponds to the decimal value of the M-bit input data [15]. The matched PPM format has higher power efficiency and better anti-noise performance for a given bit error rate (BER) compared with continuous-wave modulation formats like OOK or pulse amplitude modulation (PAM) [16–19]. In 2018, Hu et al. used a solid-state laser under 1.5-kHz repeated frequency to achieve a 120-m transmission with a link loss of 136.8 dB [20].
To decrease the gain dynamics of optical amplifiers caused by PPM signals, several guard slots are usually set between PPM symbols, resulting in the reduction of data rate. Pre-pulse shaping (PPS), which adjusts the seed light amplitude according to pulse intervals of PPM symbols in advance, could reduce the fluctuation of pulse energy caused by the gain dynamics [21]. In this paper, an open loop PPS algorithm is demonstrated to reduce the fluctuations of pulse peak power from 55.6% to 27.5% for 25-ns pulses and from 22.4% to 16.7% for 10-ns pulses, respectively. The order of PPM is 64 and the number of guard slot is reduced to only 1 slot. Besides, the photomultiplier tube (PMT) nonlinear effect, caused by space charge limitations at the dynodes of the high-sensitivity PMT for short optical pulses (∼10 ns) [22,23], would change the width and the rising time of output electrical pulses and further cause symbol errors in a PPM system. An analog demodulation method, which demodulates the symbol by calculating the starting point position of the pulse, is proposed in this paper to solve this problem. In a water tank experiment, a 9.14-Mbps data rate with 64-ary PPM is achieved in a 99-m underwater channel utilizing the proposed demodulation method. The average received optical power ranges from -60.87 to -52.51 dBm, corresponding to a BER range of 1.93 × 10−4 to 2.3 × 10−3. To further prove the validity of PPS and the reliability of the system, a 15-AL underwater experiment with the same data rate is demonstrated in a 50-m standard swimming pool. By utilizing the PPS method, the BER of the UWOC system is reduced. The analog demodulation method ensures that PPM signals could be detected by a high-sensitively PMT and demodulated without bit errors. Together, they significantly improve the performance of the UWOC system. To the best of our knowledge, the UWOC system designed in this paper has transmitted through the maximum attenuation length and achieved the highest data rate under the same average power and PPM order.
The rest of this paper is organized as follows: Section II presents the design of the UWOC system, including the structure of the light source, the principle and the results of PPS. The process of the PPM analog demodulation method is also represented in this section. Section III describes the experimental setup and the results of the proposed UWOC system. Finally, conclusions are drawn in Section IV with some discussions.
2. Design of UWOC system
2.1 Structure of light source
The near-infrared fiber MOPA has the advantages of compact structure, high optical conversion efficiency, and high beam quality. Combined with SHG and a directly-modulated seed laser diode, the MOPA is an ideal optical source for UWOC. In order to reduce the cost of the laser and eliminate the dependence on polarization-maintaining fusion machines, non-polarization-maintaining fibers and devices are used in this paper. The configuration of the pulsed MOPA is shown in Fig. 1.
A pulsed electrical signal was generated by an arbitrary waveform generator (AWG, Tektronix AWG70002A) and amplified by a power amplifier (AMP, Mini-circuit ZHL-6A-S+). Then it drove a distributed feedback laser (DFB, QLD 1061-6430), which had a 600-µW average optical power at 10-ns pulse width and 1-MHz repetition rate. A 4-m Yb3+-doped double-clad fiber (YDCF) with a 10/125-µm core/cladding diameter (YOFC YD1110-A) was utilized in the pre-amplifier. The pumping light from a 2.3-W fiber pigtailed 976-nm laser diode (LD1) was coupled into the YDCF via a (2 + 1) 1 fiber combiner. After the pre-amplification, the average power was amplified to 90 mW. The second-order amplifier consisted of a 2-m YDCF with a 20/125-µm core/cladding diameter (YOFC YD1110-B), which was pumped by a 20-W fiber pigtailed 976-nm laser diode (LD2) via a (2 + 1) 1 fiber combiner. Three isolators were set after the seed laser, the pre-amplifier and the second-order amplifier, respectively. The curve of output fundamental power and the pump power of the second-order amplification is shown in Fig. 2(a).
Fig. 2. (a) The fundamental output power versus launched pump power after the second-order amplification. (b) The maximum and minimum power at different polarization directions. The black line is the total power of fundamental light.
Due to the lack of polarization-maintaining fibers, the polarization changes with the shape of the fiber, causing the instability of SHG efficiency. To control the polarization of the fundamental light, a homemade three-paddle polarization controller (PC) fabricated by a 3D printer was set at the output of the MOPA. For a smaller bending loss, the loop diameter was set to 14 cm. According to the experimental optimization, the loop numbers of the three-paddle PC were set to 1, 2 and 1, respectively. A polarizer was set following the collimator to obtain the maximum power at any polarization by adjusting the PC in a preliminary test to verify the performance of the PC, and the result (the blue line) is shown in Fig. 2(b). It shows that, by adjusting the PC, at least half of the fundamental-light power can locate in an arbitrary polarization at common fiber shapes.
For UWOC transmitters, the watertight packaging and the security of equipment are crucial. Reducing the device footprint and using passive devices underwater are necessary. In this paper, an over 2-m passive DCF was utilized between the MOPA and the frequency doubling module (FDM) to transmit the fundamental light through the air-water interface. In the MOPA, different laser modes come from the same seed laser and they are coherent. For a 20/125-µm DCF, there only exist Lp01 and Lp11 modes for the 1064-nm laser light at the same time. So, when the PC was adjusted, the phase difference between the two linear polarization modes changed and it caused the movement of the energy center after the collimator. The complex relationship between the laser beam pattern and polarization brings challenges to the FDM structure [24], resulting in a reduction of the focusing efficiency and thus the SHG efficiency. The structure of the FDM is shown in Fig. 3. The fundamental light was collimated by a fiber collimator with a working distance of 20 mm. Then the light was expanded by a beam expander and focused by a plano-convex lens. Both focal lengths of the focusing and the collimating plano-convex lenses in the system were 75 mm. A 3 × 3 × 7 mm3 potassium titanyl phosphate (KTP) crystal cut for type-II phase matching at 1064 nm was used for second harmonic generation. Here, a single plano-convex lens was used as the collimator to reduce the volume of the FDM. The divergence angle of 532-nm light was 1.16 mrad. The thermoelectric cooler (TEC) was made by a semiconductor chilling plate. A harmonic separator was used to separate the green light from the fundamental light. Finally, considering the performance of the detector and the output power of the MOPA in this paper, which would be demonstrated in the next sections, the pulse width of the driving current was set to 10 ns, and the order of PPM was set to 64 with one guard interval. By adjusting the lens group and the polarization controller, the highest average power of the green laser was 600 mW (27.8 dBm) under a 4.83-W fundamental power.
2.2 Pre-pulse shaping
For PPM format, the randomness of the pulse position causes the pulse interval randomness, leading to the gain dynamics of the optical amplifier. The resulting pulse fluctuations will reduce the signal-to-noise ratio of UWOC. In a directly modulated communication system, the waveform of the injection current could be designed properly to solve this problem. The intensity relationship between the input seeding light and the output light can be expressed by Eq. (1) and Eq. (2) [25] which derivate from the Frantz-Nodvik-equations [26]:
For designing the waveform of the injected current, the relationship between the gain G0 and the pulse interval is necessary. The active ions density in the excited state n2 increases linearly with the time t (or pulse intervals) from the laser rate equations [25] when the spontaneous emission and the change of the pump intensity are ignored. Therefore, according to the above equations, Δ0 and G0 vary linearly with the pulse interval. To demonstrate the relationship directly, the parameters with 1-W signal average power, 10-W pump power and 10-ns pulses in [27] were used for simulations. The pulse interval ranged from 0 to 126 times the pulse width when 64-ary PPM was utilized. To prevent gain saturation and ensure enough amplification, n2 was set to half n0 after the previous pulse and 1/10 n0 was equally divided into 127 parts to simulate the situation of different pulse intervals, as shown in Fig. 4(a) and (b). The signal power changed from 0.4 W to 1.3 W with the pump power of 10 W. Based on Eq. (2), the normalized intensity of pulse peak value at different pulse intervals and different signal powers are shown in Fig. 4(c). To get a constant output power at each pulse interval, which is shown as the black solid line, signal power ranges from 1.3 to 0.4 W with corresponding pulse interval of 0 to 126. The signal power at the pulse interval 63 is 0.7 W. As the pulse interval increases, the corresponding signal power decreases slowly.
Fig. 4. (a) the densities of active ions in the excited state n2 at different pulse intervals (b) the diagram of pulses of different pulse intervals and (c) the output normalized peak intensity at different signal powers and different pulse intervals.
For the sake of simplicity, two 1/4 circles were used as the amplitude of signal power at different pulse intervals and they are shown in Fig. 5(a) and Fig. 6(a) for different pulse widths. A preliminary experiment using the MOPA designed above was demonstrated to verify the compensation effect of PPS. An avalanche photon diode (APD, Menlo Systems APD 210) with a frequency range of 1 to 1600 MHz was utilized to detect the output fundamental light and the waveform was captured by a mixed-signal oscilloscope (MSO, MSO71254C). The histograms of the peak value distribution with and without PPS for 25-ns pulses are shown in Fig. 5(b). The maximum fluctuation of modulated optical pulses was decreased from 55.6% to 27.5%, as shown in Fig. 5(c) and (d). Besides, the fluctuation of 10-ns fundamental light pulses was decreased from 22.4% to 16.7%, as shown in Fig. 6, and the corresponding parameters of the transmitter were used in the following UWOC experiments. It is obvious that PPS has a better effect for wider pulses and pulse intervals.
Fig. 5. For 25-ns pulses: (a) The pre-pulse shaping amplitude at different pulse intervals. (b) The distribution of pulse amplitude with PPS (w/ PPS) and without PPS (w/o PPS). The inset zooms in the distribution when the amplitude is under 0.85. The normalized peak value of pulses for PPM signals (c) without PPS (w/o PPS) and (d) with PPS (w/ PPS).
Fig. 6. For 10-ns pulses: (a) The pre-pulse shaping amplitude at different pulse intervals. (b) The distribution of pulse amplitude with PPS (w/ PPS) and without PPS (w/o PPS). The normalized peak value of pulses for PPM signals (c) without PPS (w/o PPS) and (d) with PPS (w/ PPS).
2.3 PPM analog demodulation method
For short optical pulses, the PMT nonlinear effect caused by space charge limitations at dynodes increases the width of output electrical pulses far beyond the width of optical pulses, leading to demodulation problems for PPM signals. An AMP and a laser diode (NDB7875), both with a 3-dB bandwidth far beyond 100 MHz, are utilized to measure the frequency response of the PMT and the result is shown in Fig. 7. For 10-ns optical pulses, the width of output electrical pulses is proportional to the intensity of the received optical power. The situations can be divided into charge-saturated and intensity-saturated situations, as shown in Fig. 8. (a), (b) and (c), (d), respectively. The neighboring situations, like (a) and (b), might occur at the same average received power due to the shot noise of PMT and the energy fluctuation of the optical source. For Fig. 8(c) and (d), the pulse covers the next slot or next symbol.
Fig. 8. The received waveforms (Rw) from the PMT and the transmitted waveform (Tw) at (a) low-power situation, (b) charge-saturated situation, (c) intensity-saturated situation and (d) intensity-saturated situation when two adjacent pulses are overlapped.
In this paper, an analog demodulation method, in which the starting points of pulses are used to judge time slots, was proposed to solve the PMT nonlinear effect. Depending on the pulse width of the training sequence, demodulation could be divided into single-pulse and overlapped-pulse cases. The demodulation method can be described as the following steps which are shown in Fig. 9.
Fig. 9. The diagrammatic sketch of the demodulation method for (a) relatively low-power pulses, (b) non-overlapped intensity-saturated pulses and (c) overlapped intensity-saturated pulses.
For the single-pulse case, there exists a more than 5-ns rising edge in a symbol duration. For output electrical pulses with the peak value lower than 0.085 V, the point i and point ii, located in the 80% and the 20% of peak value, are obtained to calculate the slope of the rising edge and estimate the position of the starting point iii, as shown in Fig. 9(a). Limited by the sampling rate of the MSO, the points near the 20% and the 80% of peak value are usually selected. For intensity-saturated pulses, the point iv and the point v with amplitude close to 0.085 and 0.017 are used to calculate the location of the starting point vi, as shown in Fig. 9(b). For all these cases, finding a long-rising edge and calculating its slope are necessary. Some overlapped pulses also satisfy this condition at relatively low optical intensity. If there is no long rising edge in a frame, the pulse width and the ending point of pulses are needed to estimate the starting point position. The pulse width can be estimated from the training sequence, which is made of several equidistant pulses at the head of PPM signals. And the ending point x in Fig. 9(c) can be obtained by finding two points of the falling edge and calculating the slope. Finally, by subtracting the pulse width from the point x position, the point vii position is obtained as the time slot of the later frame. Through circulating the starting point position of each pulse, the training sequence could be identified and used for frame synchronization. With the mentioned method, the BER at sufficient signal noise ratio could be zero in the back to back experiment.
3. Experimental setup and results
3.1 Experiments in the water tank
The experimental setup of the UWOC system in the water tank is illustrated in Fig. 10. The experiment was conducted in an optical darkroom to decrease the influence of background light. On the transmitter side, random binary sequences were generated and modulated by MATLAB in an offline personal computer. The parameters of the signal are shown in Table 1. Then the signal was loaded to an AWG, and the output amplitude was adjusted by an AMP and then drove the transmitter (TX), as shown in Fig. 10. After the FDM, the modulated 532-nm light was collimated and transmitted in the water tank. Six mirrors were set in a 7-m water tank to extend the optical path length to 99 m. On the receiver (Rx) side, a paper tube was used as a spatial filter (SF) with black tape pasted on the inner wall, which was used to reduce the stray light from pool walls and mirrors. The optical signal was detected by a PMT and the output electrical waveform was captured by the MSO. Except for the signal and the dark-current pulses, no obvious pulse was observed in the waveform, proving the availability of the SF. The attenuation coefficient of the tank water was calculated to be 0.5776 dB/m, as shown in Fig. 11.
Fig. 10. Experimental setup of the UWOC system in the water tank. Insets: i is the main part of MOPA, ii are the pump lasers and iii is the FDM.
Table 1. Parameters in the experimental system
Through adjusting the second-order amplification pump power, different green light output power was achieved in the water tank experiment and the results are recorded in Fig. 12. Threshold method, peak value method and analog method are used to demodulate the 9.14-Mbps PPM signals. Threshold method means after the symbol synchronization, the first value reaching the threshold value in the symbol duration represents the symbol value. Peak value method is that the peak value of the symbol duration represents the symbol value. At the pump powers of 1.69 W and 5.38 W, the powers of green light were about 11.76 dBm and 21.14 dBm and the corresponding BERs of analog method were 1.93 × 10−4 and 2.30 × 10−3, respectively. Compared with the threshold method and peak value method, the analog method decreases the BER at each received optical power. Limited by the volume of the water tank, the communication distance could not be further extended, and PPS was not utilized because of little gain effect with low optical gain in the MOPA. Besides, the output power was raised to 23.4 dBm (the stable maximum power was 27.8 dBm) to measure the link loss, and the received power was about −49.22 dBm. So, the total attenuation was calculated to be 72.63 dB/13.16 attenuation length (AL), which was coincident with the attenuation coefficient of the tank water.
Fig. 12. The BER performance versus the pump power of the second-stage amplification using different demodulation methods. AN: analog method. TH: threshold method. PE: peak value method.
3.2 Experiments in the swimming pool
In order to verify the effectiveness of PPS and the reliability of the transmitter, the same equipment and signal parameters were used in a standard 50-m swimming pool experiment. The experimental setup is shown in Fig. 13. Due to the limited length of the passive DCF, the shape of the pipe, which supported and protected the passive DCF, was difficult to adjust. With a 10-W second-order pump power, the maximum power of the green light was only 21 dBm. The attenuation of the water in the swimming pool was about 1.0029 dB/m according to Fig. 14.
With the larger attenuation coefficient and smaller output power, the average received power dropped and the spot area decreased, which increased the difficulty of alignment. Therefore, in the swimming pool experiment, a 65-m communication distance was achieved with a data rate of 9.14 Mbps. A box plot of BER performance corresponding to the 50-m and 65-m communication in the swimming pool and the optical back-to-back experiment in an optical darkroom is shown in Fig. 15. For the 50-m communication, the received power was saturated when the receiver was aligned with the transmitter, so the horizontal distance between the light spot and the receiver was set to 1 m. To further prove the effectiveness of PPS, a back-to-back experiment in an optical darkroom was conducted to reduce the environmental impact. From Fig. 15, we can conclude that PPS could decrease the median and the fluctuation of BERs.
Fig. 15. The BER performances with PPS (w/ PPS) and without PPS (w/o PPS) in an optical darkroom, at 50-m distance with 1-m non-alignment and 65-m distance in the swimming pool.
In the swimming pool experiments, the received power of the PMT was lower than the detection range of the power meter (Thorlabs PM100D). Therefore, the link loss and the received power were calculated to be about 78.05 dB and −57.05 dBm, respectively.
4. Conclusion and discussion
In this paper, we have designed a transmitter based on the MOPA structure and optimized it through PPS. The maximum average power of green light is about 27.8 dBm, and the divergence angle is 1.16 mrad. The FDM can be placed underwater separately from the MOPA to reduce the difficulty of underwater packaging and protect the electrical equipment. By using PPS, the fluctuation is decreased from 55.6% to 27.5% for 25-ns pulses and from 22.4% to 16.7% for 10-ns pulses. In addition, an analog PPM demodulation method has been designed based on the characteristics of the PMT to solve the problem of limited bandwidth and saturation. The average received power ranges from −60.87 dBm to −52.51 dBm with a corresponding BER ranging from 1.93 × 10−4 to 2.30 × 10−3 in a 99-m water tank experiment. The maximum link loss is measured to be −72.63 dB. A 65-m swimming pool experiment is also achieved with a corresponding BER of 3.42 × 10−4 at the same data rate. The maximum link loss is measured to be −78.95 dB.
There is still much room for improvement in our work. The main limitation to be overcome is the output power of green light, which is limited by the power and structure of the homemade TEC. A higher power from the transmitter (more than 30 dBm) could be achieved by increasing the pump power of the second-order amplifier.
Funding
National Natural Science Foundation of China (61971378); Strategic Priority Research Program of the Chinese Academy of Sciences (XDA22030208); Zhoushan-Zhejiang University Joint Research Project (2019C81081).
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. M. A. Khalighi and M. Uysal, “Survey on free space optical communication: a communication theory perspective,” IEEE Commun. Surv. Tutorials 16(4), 2231–2258 (2014). [CrossRef]
2. Z. Zeng, S. Fu, H. Zhang, Y. Dong, and J. Cheng, “A survey of underwater optical wireless communications,” IEEE Comm. Surv. and Tutor. 19(1), 204–238 (2017). [CrossRef]
3. M. Z. Chowdhury, M. K. Hasan, M. Shahjalal, M. T. Hossan, and Y. M. Jang, “Optical wireless hybrid networks: trends, opportunities, challenges, and research directions,” IEEE Commun. Surv. Tutorials 22(2), 930–966 (2020). [CrossRef]
4. T. Langer, A. Kruse, F. A. Ketzer, A. Schwiegel, L. Hoffmann, H. Jonen, H. Bremers, U. Rossow, and A. Hangleiter, “Origin of the “green gap”: Increasing nonradiative recombination in indium-rich GaInN/GaN quantum well structures,” Phys. Status Solidi C 8(7–8), 2170–2172 (2011). [CrossRef]
5. M. Kong, W. Lv, T. Ali, R. Sarwar, C. Yu, Y. Qiu, F. Qu, Z. Xu, J. Han, and J. Xu, “10-m 9.51-Gb/s RGB laser diodes-based WDM underwater wireless optical communication,” Opt. Express 25(17), 20829–20834 (2017). [CrossRef]
6. W. S. Tsai, H. H. Lu, H. W. Wu, Y. C. Huang, and S. E. Tsai, “500 Gb/s PAM4 FSO-UWOC convergent system with a R/G/B five-wavelength polarization-multiplexing scheme,” IEEE Access 8(99), 16913–16921 (2020). [CrossRef]
7. X. Liu, S. Yi, X. Zhou, S. Zhang, Z. Fang, Z.-J. Qiu, L. Hu, C. Cong, L. Zheng, R. Liu, and P. Tian, “Laser-based white-light source for high-speed underwater wireless optical communication and high-efficiency underwater solid-state lighting,” Opt. Express 26(15), 19259–19274 (2018). [CrossRef]
8. X. Yang, Z. Tong, Y. Dai, X. Chen, H. Zhang, H. Zou, and J. Xu, “100 m full-duplex underwater wireless optical communication based on blue and green lasers and high sensitivity detectors,” Opt. Commun. 498, 127261 (2021). [CrossRef]
9. X. Chen, W. C. Lyu, Z. J. Zhang, J. Zhao, and J. Xu, “56-m/3.31-Gbps underwater wireless optical communication employing Nyquist single carrier frequency domain equalization with noise prediction,” Opt. Express 28(16), 23784–23795 (2020). [CrossRef]
10. J. M. Wang and H. Chun, “100 m/500 Mbps underwater optical wireless communication using an NRZ-OOK modulated 520 nm laser diode,” Opt. Express 27(9), 12171–12181 (2019). [CrossRef]
11. Z. Chen, X. Tang, C. Sun, Z. Li, W. Shi, H. Wang, L. Zhang, and A. Zhang, “Experimental demonstration of over 14 AL underwater wireless optical communication,” IEEE Photonics Technol. Lett. 33(4), 173–176 (2021). [CrossRef]
12. F. Hanson and S. Radic, “High bandwidth underwater optical communication,” Appl. Opt. 47(2), 277–283 (2008). [CrossRef]
13. J. Xu, M. Kong, A. Lin, Y. Song, J. Han, Z. Xu, B. Wu, S. Gao, and N. Deng, “Directly modulated green-light diode-pumped solid-state laser for underwater wireless optical communication,” Opt. Lett. 42(9), 1664–1667 (2017). [CrossRef]
14. Y. Yang, L. Fan, F. He, Y. Song, Z. Duan, Y. Zhu, and B. Li, “Long-distance underwater optical wireless communication with PPLN wavelength conversion,” in Proc. SPIE 11717, 24th National Laser Conference & Fifteenth National Conference on Laser Technology and Optoelectronics, 117172J (2020).
15. Z. Ghassemlooy, W. Popoola, and S. Rajbhandari, Optical Wireless Communications: System and Channel Modelling With MATLAB (Taylor & Francis, 2012).
16. T. Y. Elganimi, “Performance comparison between OOK, PPM and pam modulation schemes for free space optical (FSO) communication systems: analytical study,” International Journal of Computer Applications. 79(11), (2013).
17. Q. Tang, L. Yang, G. B. Giannakis, and T. Qin, “Battery power efficiency of PPM and FSK in wireless sensor networks,” IEEE Trans. Wireless Commun. 6(4), 1308–1319 (2007). [CrossRef]
18. F. Qu, L. Yang, and A. Swami, “Battery power efficiency of PPM and OOK in wireless sensor networks,” in 2007 IEEE International Conference on Acoustics, Speech and Signal Processing-ICASSP'07. 3 III-525-III-528 (2007).
19. G. A. Mahdiraji and E. Zahedi, “Comparison of selected digital modulation schemes (OOK, PPM and DPIM) for wireless optical communications,” in 4th Student Conference on Research and Development (IEEE, 2006), pp. 5–10.
20. S. Hu, L. Mi, T. Zhou, and W. Chen, “35.88 attenuation lengths and 3.32 bits/photon underwater optical wireless communication based on photon-counting receiver with 256-PPM,” Opt. Express 26(17), 21685–21699 (2018). [CrossRef]
21. D. Engin, W. Lu, F. Kimpel, Y. Chen, M. Akbulut, and S. Gupta, “Fiber-MOPA based multi-kW uplink laser beacons for deep-space communication links,” Free-Space Laser Communication Technologies XXIII. 7923, 79230D (2011). [CrossRef]
22. M. Babicz, V. Bellini, M. Bonesini, T. Cervi, A. Falcone, A. Menegolli, and A. Zani, “Linearity and saturation properties of Hamamatsu R5912-MOD photomultiplier tube for the ICARUS T600 light detection system,” Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment 936, 554–555 (2019). [CrossRef]
23. D. H. Hartman, “Pulse mode saturation properties of photomultiplier tubes,” Rev. Sci. Instrum. 49(8), 1130–1133 (1978). [CrossRef]
24. G. Feng, S. Zhou, and C. Gao, Laser Mode Field and Beam Quality Characterization (National Defense Industry2016), pp. 197–198.
25. D. N. Schimpf, C. Ruchert, D. Nodop, J. Limpert, A. Tünnermann, and F. Salin, “Compensation of pulsedistortion in saturated laser amplifiers,” Opt. Express 16(22), 17637–17646 (2008). [CrossRef]
26. L. M. Frantz and J. S. Nodvik, “Theory of Pulse Propagation in a Laser Amplifier,” J. Appl. Phys. 34(8), 2346–2349 (1963). [CrossRef]
27. I. Kelson and A. Hardy, “Optimization of strongly pumped fiber lasers,” J. Lightwave Technol. 17(5), 891–897 (1999). [CrossRef]
