Open Access
Add to BibSonomy
Abstract
This study presents a novel physical layer security technique that aims to increase the security level by reducing decryption attempts and improving the resistance to security attacks. To achieve this goal, the proposed approach generates signals that resemble Gaussian noise in both the time and frequency domains. This method utilizes a wavelength-shuffled optical orthogonal frequency division multiplexing (OFDM) scheme, which is combined with the standard blue-excited phosphorus lighting approach. Experimental validation of the proposed system demonstrates a secure data rate of 880 Mb/s in the aggregate, followed by a real-time demonstration showing its practicality. Furthermore, the proposed system generates high-quality white light (with a color rendering index of 83 and correlated color temperature of 5040 K), which makes it suitable for practical illumination applications.
© 2023 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Recently, the field of communication has witnessed remarkable advancements and is expected to revolutionize current e-health, transportation, building and factory automation, and smart agriculture, more often augmented by artificial intelligence (AI). However, the implementation of these services requires substantially greater bandwidth than existing systems currently offer. Such potentials and challenges associated with future wireless systems have been extensively discussed in the literature [1–3].
Although radio frequency (RF) uses a range of frequencies, including sub-6 GHz and millimeter waves, it is challenging to meet the exponential growth of demand for faster data transmission [3]. As a potential candidate to tackle this, optical wireless communication (OWC) is gaining attention for its advantageous features over RF communications, such as immense spectrum, low latency, and high security. Typically, in visible light communication (VLC), data is conveyed by modulating the light within the visible spectrum (400-700 nm). Furthermore, the visible light spectrum does not require any license and has become more widely accessible with the advancement of cost-effective high-power light-emitting diodes (LEDs). Utilizing the existing lighting infrastructure provides simultaneous lighting and data connectivity, as LED lamps are already widely deployed [4,5]. Also, VLC technology is immune to the interferences between RF systems and is considered more cost-effective. VLC is also promising in fields where RF can interfere with sensitive devices, such as hospitals, airplanes, the petrochemical industry, and nuclear power plants [6].
Current communication systems and networks can be more vulnerable to hacking and other security threats in 6 G due to the increased use of wireless communication in various application scenarios and the growing number of connected devices [7]. The modulation format information is an important component of physical layer security, leading to significant security risks when revealed. VLC is generally regarded as having more secure communication channels than RF channels due to the non-penetrating nature of light, resulting in more difficulty for unauthorized parties to eavesdrop or interfere with the VLC systems. However, due to the line-of-sight (LoS) dominant nature of VLC, the eavesdropper can more easily exploit the modulation formats once it is exposed to them. It is especially problematic for the VLC broadcasting channel under an illumination scenario, where multiple users can perceive the same intensity variation [8].
In this research, a method to enhance the security level by producing Gaussian-noise-like signals in the time- and frequency-domain by the proposed wave-length-shuffled optical orthogonal frequency division multiplexing (O-OFDM) is presented. The generation and the characterization of such signal are detailed, followed by the demonstration of the proposed method concealing the modulation format, with a comparative BER analysis between the legitimate user and the eavesdropper. The scheme also uses the standard blue excited phosphorous white-light approach incorporated with additional red, generating a high-quality white light for practical VLC illumination scenarios.
The remainder of this study is structured as follows. Firstly, a review of physical layer security (PLS) schemes, including the potentials and vulnerabilities, is introduced in Section 2. The section also presents previous reviews of various studies to explore the feasibility of PLS in VLC. Section 3 details the proposed wavelength-shuffled O-OFDM system with both a novel PLS mechanism to enhance security and a phosphorous white-light generation to improve white-light quality for practical illumination purposes. In Section 4, the feasibility of the proposed system is demonstrated through a series of experimental analyses with a real-time system demonstration. Finally, discussions and conclusions are presented in Section 5.
2. Physical layer security in VLC
Data is transmitted through the physical layer in wireless communication, but it can be vulnerable to security threats such as eavesdropping, interference, and jamming attacks. To counter these threats, PLS protects wireless communication networks in physical mediums. PLS utilizes the physical layer's properties to make it challenging for attackers to intercept or manipulate the information through wireless channels, aiming to prevent security breaches [9]. PLS can work together with other security measures at the upper layers, including cryptography and authentication, to provide a more comprehensive security solution. Thus, PLS can offer an extra layer of security, making it harder for attackers to interfere with legitimate wireless communications, even if the security measures at the upper layers are compromised.
Recently, a great deal of research has focused on the VLC-based PLS. This technique capitalizes on the inherent properties of light to further enhance security, primarily using its high directionality, narrow beam width, and non-penetrating characteristic. Moreover, since light waves are less likely to interfere with each other in typical VLC device configurations, the VLC channel is less vulnerable to deep fading leading to severe attenuation and phase shifts that eavesdroppers can maliciously exploit. Additionally, VLC-based PLS shows potential for securing wireless communication where RF waves are unavailable or strictly regulated [10].
VLC-based PLS can use optical beam-forming or beam-steering [11,12]. These involves using multiple LEDs to create a locally confined beam of light that illuminates a specific area. For instance, a group of LEDs located above the legitimate user are modulated with the desired signal, while the other groups of LEDs are just given DC biases that generate the same amount of light without information. Thus, it becomes difficult for eavesdroppers to recognize and tap the signal from the wireless channel. On the other hand, the optical beam-steering technique directs a narrow beam of light carrying information toward the intended receiver, further complicating the interception by eavesdroppers.
However, there is a limitation of this method that it requires precise alignment between the transmitter and receiver. Otherwise, the signal may not reach the receiver correctly and could be lost or intercepted by a nearby eavesdropper. Moreover, the use of multiple independently controlled LED groups or localizing, tracking, and steering systems are necessary, which increases the complexity and cost of the system.
Another technique employed in VLC-based PLS is the use of physical layer encryption by coding schemes [13]. This technique generally utilizes secret keys for the encryption and decryption of transmitted data. The transmitter and receiver generate and keep these keys confidential from eavesdroppers. The information is transmitted typically using intensity modulation in such a way that it remains invisible to eavesdroppers, providing an additional measure of security to data transmission. However, this method can be vulnerable if the modulation format is exposed to eavesdroppers, often in VLC systems with a high signal-to-noise ratio (SNR).
The leakage of modulation format information is critical as it conveys essential contents of the transmitted signal, including frame, synchronization, and other phase and timing characteristics. The security level can be compromised even if the eavesdropper cannot access the encryption keys used to secure the data. The modulated signal can be captured and demodulated using advanced signal processing or machine learning techniques, leading to potential security breaches [14,15]. Even if the secret message is not extracted after the demodulation, the eavesdropper can still launch various security attacks by knowing the modulation format. The modulation format determines how the data is spread across the frequency spectrum that can be exploited for the eavesdropper to intercept and decode the signal. Also, the modulation format can reveal information about the encryption type being used, leading to a higher chance for the attacker to break the encryption and read the data. When it comes to tracking the movement of the communication device, the information about the modulation format can be used as well, possibly revealing the device and user locations to the eavesdropper. Therefore, it is desirable to make the existence of the communication channel unnoticed by the eavesdropper initially. Also, it is crucial to make the impact of such attacks insignificant by hiding the modulation format when applying VLC-based PLS. The following sections describe the proposed solution to deal with these.
3. Wavelength-shuffled optical OFDM
Orthogonal frequency division multiplexing (OFDM) is a widely used digital modulation technique in contemporary wireless communication. It employs a technique of subcarrier modulation by dividing a high-speed data stream into multiple lower-speed subcarriers, with each subcarrier carrying a unique part of the data. By modulating each subcarrier with a different phase and amplitude, the subcarriers become orthogonal to one another, thus avoiding any frequency domain overlap. This method results in high spectral efficiency and reduced interference [16]. Furthermore, OFDM's flexibility in the subcarrier selection can render a better SNR in comparison to pulse-based modulations. Optical OFDMs generally are employed in an intensity modulation direct modulation (IM/DD) channel and require a positive and real signal. Optical OFDM systems are broadly categorized into two groups, namely dc-biased optical OFDM (DCO-OFDM) and asymmetrically clipped optical OFDM (ACO-OFDM). DCO-OFDM is a spectral-efficient modulation that utilizes all subcarriers, while ACO-OFDM is a power-efficient modulation that uses only odd-numbered carriers by employing its antisymmetric time domain structure [17]. DCO-OFDM is combined with dc-bias to meet the positive signal value requirement, and ACO-OFDM is used with negative clipping. These two schemes result in an amplitude envelope that follows a Gaussian distribution [18].
The Gaussian distribution in the OFDM time domain signal ensures that the signal is seen as noise, thus making it difficult for eavesdroppers to demodulate it. Nonetheless, when the eavesdropper exhaustively finds the synchronization point and analyzes the signal in the frequency domain, the security level can still be undermined due to the exposure of critical information, such as the type of applied OFDM, modulation and coding levels, and sub-carrier and power allocation methods. This study investigates a DCO-OFDM-based PLS scheme, with which the received signal can be seen as noise in both time and frequency domains, consequently enhancing the level of security.
Figure 1 illustrates the algorithmic process of the proposed PLS scheme combined with the DCO-OFDM system. Two sources are employed in the system, and the encrypted signal is received by two receivers through an optical wireless channel. The procedure commences with the generation of the OFDM signal. The data streams are modulated using quadrature amplitude modulation (QAM), and the symbol-mapped data of $X{{\; }_{ch,k}}$ is transformed to the time-domain OFDM signal using the process in equation (1)–(2) [17,18]. The complex sequences turn into real signals by performing the Hermitian symmetry operation. The fast Fourier transform (FFT) size is represented by N, and ‘$ch,k$‘ represents the source and symbol index, respectively.
These generated OFDM signals are encrypted by the shared key between Alice (information sender) and Bob (legitimate user) before the transmission. In the proposed method, the key to encrypt the OFDM signal in the nth time sequence, $ke{y_n}$, is generated randomly and is composed of binary digits of ‘0’ or ‘1’. The OFDM signal ${x_{ch,n}}$ is then encrypted to the ${\textrm{y}_{ch,n}}$ through the block of ‘Encryption’ in Fig. 1 and as shown in equation (3)–(4).
In the time domain, if the $ke{y_n}$ is ‘1’, the generated OFDM signals of $x{{\; }_{1,n}}$ and $x{{\; }_{2,n}}$ are the mapped to ${\textrm{y}_{1,n}}$, and ${\textrm{y}_{2,n}}$, respectively. The OFDM sequences are not changed at the point where the key is the ‘1’. However, with the $ke{y_n}$ of ‘0’, sequences are shuffled. Two encrypted signals are summed and transmitted after the electrical-to-optical (E/O) conversion leading to ${s_{1,n}}$ and ${s_{2,n}}$, respectively. The combined signal goes through the optical wireless channel, h. The received signals by two receivers are combined with the additive white Gaussian noise (AWGN), z, as shown in the equations (5)–(6).
The digital signal processing (DSP) module separates the two channels and equates the signal strength after the optical-to-electrical (O/E) conversion. However, the noise power can be different since the channel can be asymmetrical. Equations (7)–(8) show the received signals through the optical wireless channel.
The two estimated signals are then decrypted in the way shown in equations (9)–(10). At the time sequence where the key is ‘1’, the received sequences are unchanged, while the sequences are reshuffled with the key of ‘0’.
Before the decryption, although the two channels are affected by different noise contributions due to the optical wireless channel and the E/O-O/E conversion efficiency, the noise components are mixed after the decryption process, leading to the same impact on both ${\hat{x}_{1,n}}$ and ${\hat{x}_{2,n}}$. This is experimentally verified in the following section.
4. Experimental verification with high-quality lighting
Figure 2 depicts the experimental setup used in this study. Two wavelength channels are shuffled according to the same procedures explained above. Initially, the OFDM signals generated by the PC are converted to electrical signals through an arbitrary waveform generator (Keysight, 81150A). These signals are applied to the blue and red laser diodes (Thorlabs, PL450B, and L670VH1). After the E/O conversion by laser diodes, the two colors are combined to form white light through the phosphor plate [19]. The diffused white light is then sent through the optical wireless channel. The light signals are received by two receivers (Hamamatsu C12702) equipped with two color filters. The distance between the transmitters and receivers is set to 1.2 meters. These two received signals, after O-E conversion, are captured by an oscilloscope (DSO 6054A), sent to the PC, and then processed by MATLAB for decryption and OFDM demodulation.
First and foremost, using the encrypted signal, the generation of Gaussian-noise-like signals in both the time and frequency domain is confirmed. Figure 3 shows the histogram of the received DCO-OFDM signal of each channel in the time domain. The normalized Gaussian noise distribution (reference distribution) is depicted by the dark grey lines in Fig. 4(a) and (b). The histograms agree with the reference lines, implying that the signal and the noise may be indistinguishable from the receiver's perspective. The scatter plots in the frequency domain (as shown in the insets) suggest that the signal also has complex Gaussian distribution in the frequency domain. Consequently, obtaining the applied modulation information and the subsequent secret message is challenging without knowing how the colors are shuffled.
Fig. 3. The histogram of received signal and Gaussian distribution. (a) Blue channel (ch1), (b) red channel (ch2).
Fig. 4. The comparison of 16-QAM communication performance by the key existence in the (a) blue channel (ch1) and (b) red channel (ch2).
For the test of the communication performance using the proposed method, the bit-error rate (BER) is measured by increasing the date rate with the 16-QAM DCO-OFDM scheme. The performances of both channel 1 and channel 2 are tested as shown in Fig. 4(a) and (b), respectively. In the figures, the BERs obtained from Bob and those from Eve are compared, also with the constellation scatter plots at data rates of 200 Mb/s and 440 Mb/s. With the distinctive 16-QAM constellations from Bob and Gaussian-noise-like constellations from Eve, it is seen that the proposed method is proven to work adequately. It is also found that, as expected, the performance of channel 1 and channel 2 are similar, due to the noise components being equally mixed in the decryption process. Each channel shows the 440 Mb/s, satisfying a BER under the forward-error-correction (FEC) code threshold of 3.8 × ${10^{ - 3}}$. In aggregate, a data rate of 880 Mb/s from both channels is achieved.
Figure 5 presents a spectrum analysis of the proposed method using a spectrometer (Thorlabs, CCS 200). The resulting spectrum distribution is displayed in Fig. 5(a), which exhibits two peaks approximately at 450 nm and 660 nm from the two laser diode sources and the broad yellow spectrum from the excited phosphor plate. The relative power between the blue and red channels in the proposed method is determined by monitoring the properties of white light in the CIE color space shown in Fig. 5(b).
Fig. 5. The spectrum analysis: (a) relative spectrum of the generated white-light, (b) CIE color space.
For most lighting applications, the spectral distribution of the generated white light is essential in accurately reproducing the natural colors. The measured diagram shows such color quality, which is generally represented by the color rendering index (CRI) and the correlated color temperature (CCT). The measured CRI and CCT values are 8 and 5040 K (neutral white), respectively. A CRI of more than 80 is generally considered to render a high-quality color in most indoor and commercial lighting applications.
Next, a real-time demonstration showing the feasibility of the proposed security-enhancing scheme is presented. Figure 6 shows the setup with universal software radio peripheral (USRP) configured and controlled by LabVIEW software. The USRP used in this demonstration is Ni-USRP 2901, a tunable RF (70MHz-6 GHz) transceiver with full-duplex capability. In LabVIEW, two DCO-OFDM blocks are generated and encrypted. The signal is applied to the USRP with a carrier frequency of 80 MHz. A relatively low I/Q rate of 1 MHz is set to overcome the limitation on the latency with these real-time peripherals. Each electrical signal from channel 1 and 2 are combined with DC bias and is adjusted to designated levels for the blue and the red channels shown in the previous demonstration. Through the same optical system and wireless channel shown in Fig. 2, the received signals are fed to the USRP receiver port, followed by the decryption and OFDM demodulation in LabVIEW.
Figure 7 shows the signal distribution without the secure key (Eve) and with the key (Bob). Figure 7(a) and (c) present real-time counts in time domain signal distribution, and (b) and (d) show the real-time 16-QAM scatter plots in the frequency domain, without and with the key, respectively. The results show that the applied OFDM modulation format can be hidden in both the time and frequency domain to Eve. Hence, this real-time system can provide protection against the attacks such as the attempt to decrypt the secret message, the generation of signal-like interference for jamming, and the replay attack for unauthorized access to the secure system.
Fig. 7. Signal distribution with and without the secure key: (a) time-domain real-time counts without secure key, (b) frequency-domain real-time scatter plot without secure key, (c) time-domain real-time counts with secure key, and (d) frequency-domain real-time scatter plot with secure key.
5. Conclusion
In this study, a wavelength-shuffled optical orthogonal frequency division multiplexing method was presented to enhance security by producing signals that resemble Gaussian noise in both the time and frequency domains. The nature of Gaussian distribution which is the innate characteristic of DCO-OFDM in the time domain and the color-shuffling that breaks the orthogonality among subcarriers leading to complex Gaussian distribution in the frequency domain were utilized. The proposed system was experimentally verified with the standard blue-excited phosphorus lighting technique with the inclusion of red color to generate high-quality white light (83 and 5040 K) that is suitable for practical illumination applications. The secure communication performances were evaluated, and each channel showed 440 Mb/s (880 Mb/s in aggregate). Also, a real-time system evaluating the practicality of the proposed system was successfully demonstrated. Future work includes analytic investigations of the proposed method for dealing with various attacks, such as jamming and replay attacks, with more analytic security performance indicators, such as secrecy capacity analysis under different channel characteristics and coding schemes.
Acknowledgments
This work was supported by Organic Materials Laboratory, Samsung Advanced Institute of Technology (SAIT).
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. W. Akhtar, S. A. Hassan, R. Ghaffar, H. Jung, S. Garg, and M. S. Hossain, “The shift to 6 G communications: vision and requirements,” Hum. Cent. Comput. Inf. Sci. 10(1), 53 (2020). [CrossRef]
2. H. Tataria, M. Shafi, A. F. Molisch, M. Dohler, H. Sjöland, and F. Tufvesson, “6 G wireless systems: Vision, requirements, challenges, insights, and opportunities,” Proc. IEEE 109(7), 1166–1199 (2021). [CrossRef]
3. A. Minotto, P. A. Haigh, Ł. G. Łukasiewicz, E. Lunedei, D. T. Gryko, I. Darwazeh, and F. Cacialli, “Visible light communication with efficient far-red/near-infrared polymer light-emitting diodes,” Light: Sci. Appl. 9(1), 70 (2020). [CrossRef]
4. N. Chi, Y. Zhou, Y. Wei, and F. Hu, “Visible light communication in 6G: Advances, challenges, and prospects,” IEEE Veh. Technol. Mag. 15(4), 93–102 (2020). [CrossRef]
5. S. U. Rehman, S. Ullah, P. H. J. Chong, S. Yongchareon, and D. Komosny, “Visible light communication: a system perspective—overview and challenges,” Sensors 19(5), 1153 (2019). [CrossRef]
6. S. Idris, A. M. Aibinu, G. Koyunlu, and J. Sanusi, “A survey of modulation schemes in visible light communications,” In 2019 3rd International Conference on Trends in Electronics and Informatics (ICOEI) (pp. 1–7). IEEE (2019).
7. G. Blinowski, “Security of visible light communication systems—a survey,” Phys. Commun. 34, 246–260 (2019). [CrossRef]
8. Y. Liang, H. V. Poor, and S. Shamai (Shitz), “Physical layer security in broadcast networks,” Security Comm. Networks 2(3), 227–238 (2009). [CrossRef]
9. Y.-S. Shiu, S. Y. Chang, H.-C. Wu, S. C.-H. Huang, and H.-H. Chen, “Physical layer security in wireless networks: a tutorial,” IEEE Wireless Commun. 18(2), 66–74 (2011). [CrossRef]
10. M. A. Arfaoui, M. D. Soltani, I. Tavakkolnia, A. Ghrayeb, M. Safari, C. M. Assi, and H. Haas, “Physical Layer Security for Visible Light Communication Systems: A Survey,” IEEE Commun. Surv. Tutorials 22(3), 1887–1908 (2020). [CrossRef]
11. M. A. Arfaoui, Z. Rezki, A. Ghrayeb, and M. S. Alouini, “On the input distribution and optimal beam-forming for the MISO VLC wiretap channel,” 2016 IEEE Global Conference on Signal and Information Processing (GlobalSIP), Washington, DC, USA, pp. 970–974 (2016).
12. S. Cho, G. Chen, and J. P. Coon, “Enhancing Security in VLC Systems Through Beam-forming,” GLOBECOM 2020 - 2020 IEEE Global Communications Conference, Taipei, Taiwan, pp. 1–6, (2020).
13. Z. Che, J. Fang, Z. L. Jiang, X. Yu, G. Xi, and Z. Chen, “A physical-layer secure coding schcme for visible light communication based on polar codes,” 2017 Conference on Lasers and Electro-Optics Pacific Rim (CLEO-PR), Singapore, pp. 1–2, (2017).
14. X. -C. Chen and Y. -J. Chen, “A Machine Learning Based Attack in UAV Communication Networks,” 2019 IEEE 90th Vehicular Technology Conference (VTC2019-Fall), Honolulu, HI, USA, pp. 1–2, (2019).
15. I. Alouani, “Breaking (and Fixing) Channel-based Cryptographic Key Generation: A Machine Learning Approach,” 2022 25th Euromicro Conference on Digital System Design (DSD),Maspalomas, Spain, pp. 383–390 (2022).
16. I. B. Djordjevic and B. Vasic, “Orthogonal frequency division multiplexing for high-speed optical transmission,” Opt. Express 14(9), 3767–3775 (2006). [CrossRef]
17. S. D. Dissanayake and J. Armstrong, “Comparison of ACO-OFDM, DCO-OFDM and ADO-OFDM in IM/DD Systems,” J. Lightwave Technol. 31(7), 1063–1072 (2013). [CrossRef]
18. L. Chen, B. Krongold, and J. Evans, “Performance Analysis for Optical OFDM Transmission in Short-Range IM/DD Systems,” J. Lightwave Technol. 30(7), 974–983 (2012). [CrossRef]
19. L. Issaoui, S. Cho, and H. Chun, “High CRI RGB Laser Lighting With 11-Gb/s WDM Link Using Off-the-Shelf Phosphor Plate,” IEEE Photon. Technol. Lett. 34(2), 97–100 (2022). [CrossRef]
