Open Access
Add to BibSonomy
Abstract
In this paper, a high security chaotic encryption scheme for orthogonal frequency division multiplexing (OFDM) transmission system is proposed by using power division multiplexing (PDM) technology and four-dimensional region joint encryption. The scheme uses PDM to realize simultaneous transmission of multiple user information, which can achieve a good compromise among system capacity, spectral efficiency and user fairness. In addition, bit cycle encryption, constellation rotation disturbance (CRD) and region joint constellation disturbance (RJCD) are used to realize four-dimensional region joint encryption, effectively improving the physical layer security. The masking factor is generated by the mapping of two-level chaotic systems, which can enhance the nonlinear dynamics and improve the sensitivity of encrypted system. A 11.76 Gb/s OFDM signal transmission over 25 km standard single-mode fiber (SSMF) is experimentally demonstrated. At the forward-error correction (FEC) bit error rate (BER) limit -3.8×10−3, the proposed receiver optical power based on quadrature phase shift keying (QPSK) without encryption, QPSK with encryption, variant-8quadrature amplitude modulation (V-8QAM) without encryption and V-8QAM with encryption are about -13.5dBm, -13.6dBm, -12.2dBm, and -12.1dBm. The key space is up to 10128. The results show that this scheme not only improves the security of the system and the ability to resist attackers, but also improves the system capacity and has the potential to serve more users. It has a good application prospect in the future optical network.
© 2023 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Cloud operation, virtual reality, augmented reality, intelligent devices, intelligent transportation, telemedicine, remote control and other applications emerge in an endless stream, which makes the increasing demand for network bandwidth [1,2]. In order to improve the capacity of transmission and spectral efficiency, researchers use multiple dimensions such as wavelength, frequency and polarization [3–5]. Recently, multiplexing technology based on power domain has been introduced into optical communication systems as a new multiplexing dimension [6,7]. The technology distinguishes users by different power levels. At the transmitter end, different transmitting power is allocated to each user according to the given power allocation principle. At the receiver, according to the different power of the received signal, the serial interference cancellation (SIC) receiver is used to recover the signal information of each user in turn [8,9]. Therefore, without increasing available resources, power multiplexing technology can serve more users in a given frequency and time period, and effectively improve spectrum efficiency. However, at the receiver, low-power users need to decode signals of high-power users firstly when realizing SIC extraction, which may easily lead to privacy issues among all legitimate users of the service [10–12].
In addition, with the rapid development of optical transmission system, a large number of private data may be leaked when transmitted on optical transmission system, so in the real information transmission, people pay more and more attention to its privacy and security [13,14]. At present, there are many ways to solve the security problem of optical communication system. For example, L. Gao et al. in reference [15] designed an RFID security authentication protocol based with Rabin encryption algorithm to improve the security of low-cost RFID. However, the upper layer encryption has defects, which cannot realize full data protection. Reference [16] enhanced the security of OFDM-PON by combining the improved DNA encoding encryption with symbol level matrix scrambling, but the introduction of DNA coding technology has further increased the complexity of encryption. Reference [17] adopts the bidirectional scheme of continuously variable quantum key distribution to realize the security of plug and play system. Quantum key distribution can theoretically guarantee that any eavesdropping behavior can be detected, but it still faces the challenge of integration in real-time passive optical networks. Compared with them, chaotic physical layer encryption technologies are generally more efficient and flexible and compatible with mature digital signal processing (DSP) [18]. Among the existing physical layer encryption methods, chaotic system encryption is an effective method to protect signal security and confidentiality due to its randomness and high sensitivity to initial conditions [19–22]. Therefore, how to effectively guarantee the optical information transmission security based on power division multiplexing (PDM) is worth further study.
This paper proposes a high security optical transmission scheme based on PDM four-dimensional region joint encryption. Take the signals of two channels as an example. The input data of channel A is firstly encrypted by bit cycle, and then the constellation rotation disturbance (CRD) is completed. The input data of channel B goes through a bit cycle encryption firstly, and then QAM mapping is completed. The data of two channels are superimposed into one signal by PDM, and then the superimposed signal is encrypted by regional joint constellation disturbance (RJCD). Finally, the encrypted signal is input to the optical transmission channel through orthogonal frequency division multiplexing (OFDM) modulation. The proposed scheme is experimentally verified in a 25 km standard single-mode optical fiber (SSMF) transmission system at the rate of 11.76Gb/s, and the system performance is analyzed. The results show the key space is up to 10128. In addition, at the forward-error correction (FEC) bit error rate (BER) limit -3.8 × 10−3, quadrature phase shift keying (QPSK) without encryption receives optical power of about -13.5dBm, QPSK with encryption receives optical power of about -13.6dBm, and variant-8quadrature amplitude modulation (V-8QAM) without encryption receives optical power of about -12.2dBm, and V-8QAM with encryption receive optical power of about -12.1dBm. The scheme not only guarantees the information security of users with high power ratio, but also can effectively resist illegal attacks and has good robustness.
2. Principles
The proposed four-dimensional regional joint encryption scheme based on PDM is shown in Fig. 1. In the figure, two input data are used as an example. First, the input data A is encrypted by the bit cycle. After the binary sequence is encoded by QAM mapping, the chaotic vector rotates the constellation randomly to realize the masking of constellation dimension. Channel A data is preliminarily encrypted. Then, the input data B is encrypted by the bit cycle. The constellation map obtained after QAM mapping and A channel primary encrypted signal are used for PDM superposition. The regional joint constellation disturbance is performed on the signal after superposition. The transmitted data is encrypted. Finally, the encrypted signal modulated by OFDM is sent to the SSMF channel for transmission. The key-driven two-level cascaded chaotic mapping generates four chaotic vectors for bit cycle encryption, constellation rotation disturbance and regional joint constellation disturbance respectively. At the receiver, the inverse RJCD is executed firstly, and then PDM is demodulated by SIC. Finally, the two signals are calculated by the opposite method of the encrypted end to obtain the original data.
2.1 Chaotic sequence generation
In a general chaotic system, the chaotic oscillation is easily transformed into periodic oscillation when the parameters are slightly impacted. Therefore, the two-level cascaded chaotic model is adopted in this scheme, which can enhance the nonlinear dynamics of chaotic system and avoid the degradation of chaotic dynamics caused by the precision problem of digital equipment. At the same time, the encryption sensitivity of the system can be improved in the physical layer encryption. The pseudo-random sequences generated by this model are used as masking factors to realize bit cycle encryption, CRD and RJCD respectively. The expression is as follows:
2.2 Four-dimensional region joint encryption principles
In this scheme, two input signals are taken as an example. QPSK signal is adopted in channel B, and A V-8QAM is adopted in channel A, which corresponds to four different quadrants of QPSK and has four different constellation patterns. The constellation diagram has symmetry in the four quadrants, as shown in Case1-Case 4 of PA in Fig. 3(B). Compared with the traditional rectangular 8QAM, the V-8QAM constellation figure of merit (CFM) has better performance and better anti-noise performance. CFM can be expressed as [23]:
where C represents the given constellation. $d_{\min }^2(C)$ represents the minimum euclidean distance (MED) and P(C) represents the average power of the constellation.The schematic diagram of signal generation at the transmitter is shown in Fig. 3(A) and Fig. 3(B). ${S_A}(t)$ and ${S_B}(t)$ are set to represent channel A and channel B respectively. The transmitter signal $S(t)$ can be expressed as:
where ${P_A}$ and ${P_B}$ indicate the power allocated to the signal of channel A and B respectively. The total power $P = {P_A} + {P_B}$, and the power distribution ratio (PDR) between the information of channels A and B is $\textrm{r} = {P_A}/{P_B}$.In order to ensure the security of the transmitted information, this scheme uses bit cycle encryption, CRD and RJCD to realize four-dimensional region joint encryption. The key is composed of the initial state and control parameters (a, b, c, d, e, f) of the two-level cascaded chaotic mapping. Chaotic sequence (x, y) is used to implement bit cycle encryption. It is worth noting that (x, y) cannot be directly used for bit-cycle encryption. Firstly, x is converted into binary, and the end bit of binary representation is taken as the pseudo-random sequence $\{{{L_i}} \}_{i = 0}^a$. Then two pseudo-random series ${M_i}$, ${N_i}$ are further generated from $\{{{L_i}} \}_{i = 0}^a$, which are expressed as follows:
After the two pseudo-random sequences ${M_i}$ and ${N_i}$ are determined, the channel A encryption process can be expressed as:
The decryption process of channel A can be expressed as:
where $a < < { < _b}x = a > > { > _b}( - x) = \sum\limits_{i = 0}^{b - 1} {({a_i}{2^{(i + x)\bmod b}})}$. Chaotic sequence y is used to encrypt channel B in the same way. z chaotic sequence completes constellation rotation disturbance encryption, and w sequence completes regional joint constellation disturbance encryption.The unencrypted constellation and rotation disturbance encrypted constellation are shown in Fig. 4(A). z-chaotic sequence is used to carry out constellation disturbance encryption for channel A V-8QAM. It is worth mentioning that z chaotic sequence cannot be directly encrypted. z chaotic sequence and encryption process are dealt with as follows:
Fig. 4. Constellation before and after encryption. (A) constellation rotation disturbance encryption; (B) regional joint constellation disturbance encryption.
Figure 4(B) shows the constellation diagram before and after disturbance encryption of the regional joint constellation. w chaotic sequence is used to encrypt the regional joint constellation disturbance of the constellation map after PDM. Firstly, w chaotic sequence is preprocessed. The specific operations are as follows:
We used ${w_i}$ chaotic sequence to generate pseudo-encrypted constellation points $({w_i},{w_{i + 1}})$. Assuming that the position of the constellation point is L (I, Q), and the distance between the pseudo-encrypted point and the constellation point is:
The masked constellation angle is:
The position of the constellation point after perturbation is:
The principle of signal recovery at the receiver is shown in Fig. 5. We equalize the received data and decrypt the joint perturbation using the key. Then we demodulated PDM by SIC. The basic principle of SIC is to use step by step interference elimination strategy, as shown in Fig. 5(B). In the received signal, multiple users are determined one by one, and the multiple access interference (MAI) caused by the user signal is subtracted when a user is determined. In this scheme, because the power allocated to the B channel signal is greater than that of the A channel signal, the B channel signal after SIC can be directly decoded and received to the QPSK constellation signal. Therefore, channel estimation is performed first to obtain channel response H1. Then the signal balance processing, decryption, demodulation operation. At last, the original signal Data B can be obtained by bit cycle decryption. The A channel signal is a low-power signal that is first decoded for the B channel signal and then subtracted from the received signal. Therefore, the re-modulation of Data B is first completed, that is, the signal obtained after the first decryption at the receiver is re-modulated and multiplied by the channel response H1. The new modulated signal is then subtracted from the signal on the receiver to obtain the low power level user signal Data A. Finally, after channel estimation, channel balancing, decryption and demodulation, bit cycle decryption is carried out to recover the original signal Data A. It is worth noting that the encryption scheme is implemented at the digital end, which is flexible and compatible with other advanced modulation formats. Most of the digital modulation methods used in non-orthogonal multiple access technology can be compatible with this encryption scheme.
Fig. 5. (A) Schematic diagram of signal recovery at the receiver; (B) Schematic diagram of SIC principle.
3. Experiment setup and results
The experimental setup of the proposed scheme is shown in Fig. 6. The encrypted signal is generated on the optical line terminal (OLT) through off-line DSP. The receiver is equipped with a normal receiving optical network units (ONUs) with a security key, while the illegal ONUs without a security key can only obtain the information by brute force. At the transmitter, the signal output by DSP is converted into an analog radio frequency (RF) signal at a speed of 10 GSa/s by an arbitrary waveform generator (AWG,TekAWG70002A) with a sampling rate of 50 GSa/s. The number of subcarriers is set to 256, the number of IFFT points to 1024, and the protection interval to 1/16. We calculated the effective information transmission rate: $10GSa/s \times 5bits/symbol \times 256 \div 1024 \times 1 \div (1 + 1/16) = 11.76Gb/s$. After passing through an electrical amplifier (EA), the signal is then loaded through a mach-zehnder modulator (MZM) to light with a wavelength of 1550 nm and a power of 12dBm generated by a laser with a linewidth of less than 100 KHz. The two OFDM signals allocate power PA and PB according to PA < PB. Finally, the optical signal is sent to the standard single-mode fiber transmission channel of 25 km for transmission. Erbium-doped fiber amplifiers (EDFA) are used to amplify optical signals. At the receiver, the received optical power is adjusted by a variable optical attenuator (VOA). Photodiode (PD) converts the optical signal into electrical signal, and then through the mixed signal oscilloscope (MSO, TekMSO73304DX) with sampling rate of 50 GSa/s to complete the analog-to-digital conversion. Finally, the original data is recovered by offline DSP.
Fig. 6. Experimental setup (AWG: arbitrary waveform generator; EA: electrical amplifier; MZM: Mach-Zehnder modulator; EDFA: erbium-doped fiber amplifier PS: power splitter; VOA: variable optical attenuator; PD: photodiode; MSO: mixed signal oscilloscope; MCF: multicore fiber).
In order to verify the sensitivity of the key, we analyze the changes of the received BER corresponding to different parameters when the initial value of the system changes very small. As shown in Fig. 7, the abscissa represents the degree of refinement of parameters changed by initial value, and the ordinate is the corresponding BER. For x and y sequences, when some parameter of the key changes above E-9; For z and w sequences, when a parameter of the key changes above E-16, the BER increases sharply, leading to information decryption failure. The meaning of E-9 specifically refers to: suppose the value of x is 1, when it is slightly changed to 1 + 10-9, the BER will be significantly increased, and the receiver cannot decrypt normally, which proves the high sensitivity of the encryption system. The sensitivity of z and w sequences generated by system (2) is higher than that of x and y sequences generated by system (1), because the initial value of the input of system (2) is the output sequence of system (1). The key space can be expressed as (a, b, c, d, e, f, x, y, z, w), which can realize the key space of $[{{{(1{0^{11}})}^3} \times {{({{10}^{15}})}^3} \times {{({{10}^9})}^2} \times {{({{10}^{16}})}^2}} ]= {10^{128}}$. Therefore, the two-level cascaded chaotic sequence has high sensitivity and can effectively resist brute force cracking. Compared with the encryption algorithm of a single chaotic system with the same dimension, the two-level chaotic system will increase the key space, and the cascaded encryption will further improve the sensitivity of the encryption system. The dynamic factor is introduced into the cascade chaotic encryption algorithm, which can make the attractor switch in different regions, so that the key can be updated and iterated constantly.
This scheme uses PDM technology. In order to achieve the best performance and get the optimal power distribution scheme, power distribution is particularly important. Therefore, we analyze the effect of different PDRs on the user's BER performance. When the received optical power is fixed at -12dBm, the BER of the two users without encryption is shown in Fig. 8, and the BER curve of the two users with encryption is shown in Fig. 9. As can be seen from the figure, the change trend of BER curves in the two cases is similar. With the increase of PDR, the power allocated to high power ratio channel B is also increasing, and its BER is decreasing. However, the whole signal at the receiver is demodulated as high-power signal QPSK. On this basis, the low-power signal V-8QAM is modulated using SIC. Once errors occur in the QPSK signal, these errors are introduced into the demodulation of the V-8QAM. So, for low power ratio channel A, when PDR is less than 9 dB, the BER of low power ratio signal-A decreases with the increase of PDR. When PDR is greater than 9 dB, the BER starts to increase. Therefore, in order to ensure the signal reception quality of the two channels, the PDR with the lowest BER of V-8QAM signal is the optimal PDR of the system.
We compare the BER of different received optical power. Setting the PDR at 9, channel A uses the low-power V-8QAM signal, and channel B uses a high-power QPSK signal. Figure 10 shows the BER curves of the system under different receiving power levels. As can be seen from the figure, the BER of high-power signal is significantly lower than that of low-power signal. The proposed receiver optical power based on QPSK without encryption, QPSK with encryption, V-8QAM without encryption and V-8QAM with encryption are about -13.5dBm, -13.6dBm, -12.2dBm, and -12.1dBm under the FEC BER limit of 3.8 × 10−3, respectively. V-8QAM with encryption has 0.1 dB sensitivity gain. Compared with the unencrypted system, the performance of the system is almost unaffected in the encrypted case. So the encryption algorithm introduced in this scheme can guarantee the communication quality and performance of the system. Therefore, the scheme can not only effectively resist illegal attacks, but also has good robustness.
Under normal reception conditions, ONUs uses the same key as the transmitter to correctly decrypt the encrypted message, while the receiver without the correct key can only obtain the message by illegal means. Fig. 11 illustrates the BER curves of V-8QAM-OFDM transmission system under legal and illegal conditions. The legal condition includes 25 km SSMF transmission system with encryption, 25 km SSMF transmission system without encryption, and back-to-back (BTB) transmission. In the case of illegal ONUs, the BER of the signal is greater than 0.3. In the case of legal ONUs, the BER in different scenes decreases with the increase of receiving power. When the received optical power is greater than -13dBm, the BER of BTB transmission is less than 10−2. When the received optical power is greater than -12dBm, the BER of the transmission under the three legal conditions is less than 10−3. As can be seen from the figure, the BER of V-8QAM-OFDM without encryption is very close to that of V-8QAM-OFDM with encryption, so the encryption scheme does not add complexity to the system. To sum up, the scheme proposed in this paper has good recovery performance of the original data in the SSMF transmission system experiment.
4. Conclusion
In this paper, a high robustness, high spectral efficiency and high security encryption scheme is proposed by combining PDM technology with four-dimensional region joint encryption, which is applied to OFDM transmission system. The PDM technology is adopted to accomplish the multiplexing of the overlapping signals of spectrum under different power ratios, serving more users on the same resources, and effectively improving the system capacity and user access capability. Using the four-dimensional region joint encryption technology and OFDM modulation, the security of legitimate users to eavesdroppers and the privacy of all legitimate users are guaranteed without affecting the performance of BER of legitimate users. The scheme is verified by experiments in a 25 km SSMF system. The experimental results show that the proposed scheme has high sensitivity. The key space is up to 10128, and the scheme has high anti-hacker ability and good key leakage performance. In addition, at the FEC BER limit -3.8 × 10−3, the proposed receiver optical power based on V-8QAM with encryption is -12.1dBm, compared to V-8QAM without encryption, it has 0.1 dB sensitivity gain. Therefore, this scheme can guarantee the communication performance of the system even after the encryption algorithm is introduced. In conclusion, this scheme can not only improve the system capacity and user access capability, but also ensure the security of the transmitted information. It is a candidate scheme for future large-capacity and high-security optical transmission systems.
Funding
National Key Research and Development Program of China (2018YFB1800901); National Natural Science Foundation of China (62225503, 61835005, 62205151, 62171227, 61935005); Jiangsu Provincial Key Research and Development Program (BE2022079, BE2022055-2); The Natural Science Foundation of the Jiangsu Higher Education Institutions of China (22KJB510031); The Startup Foundation for Introducing Talent of NUIST.
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. Cisco, “Cisco visual networking index: Global mobile data traffic forecast update, 2016–2021 White Paper,” San Jose, CA, USA, 2017. [Online]. Available: https://s3.amazonaws.com/media.mediapost.com/uploads/CiscoForecast. pdf.
2. S. Le, T. Drenski, A. Hills, M. King, K. Kim, Y. Matsui, and T. Sizer, “100Gbps DMT ASIC for Hybrid LTE-5 G Mobile Fronthaul Networks,” J. Lightwave Technol. 39(3), 801–812 (2021). [CrossRef]
3. K. Igarashi, T. Tsuritani, and I. Morita, “Polybinary Shaping for Highly-Spectral-Efficient Super-Nyquist WDM QAM Signals,” J. Lightwave Technol. 34(8), 1724–1731 (2016). [CrossRef]
4. X. Fang, Y. Xu, Z. Chen, and F. Zhang, “Frequency-Domain Channel Estimation for Polarization-Division-Multiplexed CO-OFDM/OQAM Systems,” J. Lightwave Technol. 33(13), 2743–2750 (2015). [CrossRef]
5. Q. Wu, Y. Zhu, Q. Zhuge, and W. Hu, “Dual-Carrier-Assisted Phase Retrieval for Polarization Division Multiplexing,” J. Lightwave Technol. 40(22), 7297–7306 (2022). [CrossRef]
6. V. Dixit and A. Kumar, “Error analysis of L-PPM modulated MIMO based multi-user NOMA-VLC system with perfect and imperfect SIC,” Appl. Opt. 61(4), 858–867 (2022). [CrossRef]
7. F. Lu, M. Xu, L. Cheng, J. Wang, and G. Chang, “Power-Division Non-Orthogonal Multiple Access (NOMA) in Flexible Optical Access With Synchronized Downlink/Asynchronous Uplink,” J. Lightwave Technol. 35(19), 4145–4152 (2017). [CrossRef]
8. J. Shi, J. He, K. Wu, and J. Ma, “Enhanced Performance of Asynchronous Multi-Cell VLC System Using OQAM/OFDM-NOMA,” J. Lightwave Technol. 37(20), 5212–5220 (2019). [CrossRef]
9. F. Lu, M. Xu, L. Cheng, J. Wang, J. Zhang, and G. Chang, “Non-Orthogonal Multiple Access With Successive Interference Cancellation in Millimeter-Wave Radio-Over-Fiber Systems,” J. Lightwave Technol. 34(17), 4179–4186 (2016). [CrossRef]
10. Q. Zhong, J. Ren, B. Liu, Y. Li, Z. Guo, Y. Mao, X. Wu, R. Ullah, Y. Wu, L. Zhao, and T. Sun, “High-security UFMC optical transmission system of seven-core fiber based on updating the 3D discrete chaotic model,” Opt. Lett. 47(9), 2254–2257 (2022). [CrossRef]
11. Y. Wan, B. Liu, J. Ren, R. Ullah, Y. Mao, S. Zhu, S. Chen, X. Wu, F. Wang, T. Sun, Y. Wu, and L. Zhao, “Performance-Enhanced Optical Non-Orthogonal Multiple Access Enabled by Orthogonal Chirp Division Multiplexing,” J. Lightwave Technol. 40(16), 5440–5449 (2022). [CrossRef]
12. Y. Yang, C. Chen, W. Zhang, X. Deng, P. Du, H. Yang, W. Zhong, and L. Chen, “Secure and private NOMA VLC using OFDM with two-level chaotic encryption,” Opt. Express 26(26), 34031–34042 (2018). [CrossRef]
13. Z. Wang, Y. Xiao, S. Wang, Y. Yan, B. Wang, Y. Chen, Z. Zhou, J. He, and L. Yang, “Probabilistic shaping based constellation encryption for physical layer security in OFDM RoF system,” Opt. Express 29(12), 17890–17901 (2021). [CrossRef]
14. M. Furdek, C. Natalino, F. Lipp, D. Hock, A. Giglio, and M. Schiano, “Machine Learning for Optical Network Security Monitoring: A Practical Perspective,” J. Lightwave Technol. 38(11), 1 (2020). [CrossRef]
15. L. Gao, L. Zhang, and M. Ma, “Low Cost RFID Security Protocol Based on Rabin Symmetric Encryption Algorithm,” Wirel. Pers. Commun. 96(1), 683–696 (2017). [CrossRef]
16. X. Song, B. Liu, H. Zhang, R. Ullah, Y. Mao, J. Ren, S. Chen, J. Zhang, J. Zhao, S. Han, X. Liu, D. Zhao, and X. Xin, “Security-enhanced OFDM-PON with two-level coordinated encryption strategy at the bit-level and symbol-level,” Opt. Express 28(23), 35061–35073 (2020). [CrossRef]
17. R. Goncharov, D. Kirichenko, I. Vorontsova, I. Filipov, Y. Adam, B. Pervushin, B. Nasedkin, E. Samsonov, and V. Egorov, “Security of plug-and-play continuous-variable quantum key distribution,” J. Opt. Technol. 89(7), 430–433 (2022). [CrossRef]
18. Y. Chen, J. Chen, M. Zhang, W. Li, D. Liu, and M. Tang, “High-security constellation shaped self-homodyne coherent system with 4-D joint encryption,” Opt. Express 31(2), 3153–3167 (2023). [CrossRef]
19. N. Jiang, A. Zhao, S. Liu, C. Xue, B. Wang, and K. Qiu, “Generation of broadband chaos with perfect time delay signature suppression by using self-phase-modulated feedback and a microsphere resonator,” Opt. Lett. 43(21), 5359–5362 (2018). [CrossRef]
20. Y. Xiao, Z. Wang, J. Cao, R. Deng, Y. Liu, J. He, and L. Chen, “Time–Frequency Domain Encryption With SLM Scheme for Physical-Layer Security in an OFDM-PON System,” J. Opt. Commun. Netw. 10(1), 46–51 (2018). [CrossRef]
21. Y. Xiao, Y. Chen, and Z. Wang, “Secure transmission of W-band millimeter-wave based on CNN and dynamic resource allocation,” Opt. Lett. 46(22), 5583–5586 (2021). [CrossRef]
22. Q. Zhong, B. Liu, J. Ren, Y. Li, Z. Guo, Y. X. Wu, Y. Ma, Y. Wu, L. Zhao, T. Sun, and R. Ullah, “Self-propagated chaotic dynamically enhanced optical physical layer encryption communication system based on bidirectional long short-term memory neural network,” Opt. Express 30(20), 36379–36393 (2022). [CrossRef]
23. J. Ren, B. Liu, X. Xu, L. Zhang, Y. Mao, X. Wu, Y. Zhang, L. Jiang, and X. Xin, “A probabilistically shaped star-CAP-16/32 modulation based on constellation design with honeycomb-like decision regions,” Opt. Express 27(3), 2732–2746 (2019). [CrossRef]
