Physical layer security performance of satellite networks

Yahia, Ben Olfa
Süreli Yayın başlığı
Süreli Yayın ISSN
Cilt Başlığı
Graduate School
A vision of a three-layer Vertical Heterogeneous Network (VHetNet) is being discussed in the state-of-the-art Sixth Generation (6G) network design. This concept is in line with the 3rd Generation Partnership Project (3GPP)'s non-terrestrial network (NTN) activities. A satellite (space) network, an aerial network, and a terrestrial network represent the three levels. The space layer includes geostationary Earth orbit (GEO) satellites and non-GEO satellites including low Earth orbit (LEO) and medium Earth orbit (MEO) satellites. This layer will provide orbit or space Internet services in applications including space travel as well as wireless coverage for unserved and underserved areas not served by terrestrial networks by densely deploying LEO, MEO, and GEO constellations. By intensively utilizing flying base stations (BS), including unmanned aerial vehicles (UAV), and floating BSs, such as high altitude platform systems (HAPS), the aerial layer would enable a more flexible and better quality of service for critical events or in inaccessible areas. For most human activities, the terrestrial layer will continue to be the primary method for delivering wireless coverage. Satellite communication (SatCom) systems are typically radio frequency (RF)-based, with multiple frequency bands employed for different applications. However, RF channels suffer from spectrum congestion, license issues, interference with other frequency bands, and security threats. To address these issues, free-space optical (FSO) communication has lately been recommended for use in SatCom links because of its capacity to provide exceptionally large bandwidth, unlicensed spectrum, better security, low interference, ease of deployment, and other features over its RF counterpart. Nevertheless, similar to RF channels, FSO communication presents several drawbacks related to atmospheric effects, including atmospheric turbulence, scattering, and attenuation. Furthermore, FSO communication suffers from beam wander, beam divergence, and pointing issues. Different techniques have been proposed in the current literature to reduce these effects and enhance the overall performance including the incorporation of the RF link with the FSO to benefit from their complementary characteristics. This combination includes two possible architectures: mixed RF-FSO and hybrid RF/FSO communications. In mixed RF-FSO communication, a dual-hop network is considered, in which an RF channel is employed at one hop and an FSO channel is used at the other. For hybrid RF/FSO communication, RF and FSO channels are used in parallel. These combinations may reap all the advantages of both RF and FSO communications while minimizing weather-related issues. On the one hand, hybrid RF/FSO communication has received much attention in the recent literature. On the other hand, for satellite networks, there is a substantial gap in hybrid RF/FSO communication. In addition, there are a few works about mixed RF-FSO networks in NTNs while sending the same information to a group of users on a particular multicast group address, known as multicast services. Due to the broadcasting nature of the wireless channel, security is the most critical issue and challenging concern in NTNs. Unlike conventional methods, which often handle security at the application layer, physical layer security (PLS) guarantees information-theoretic security by extracting the variations in the physical properties of channels such that a degraded signal at an eavesdropper is always assured and therefore the original message can be hardly obtained regardless of how the signal is processed at the illegitimate receiver. PLS has been investigated as a good alternative to provide a solid form of security. Many studies have been conducted up to date to determine the fundamental performance limits of PLS under various wiretap channel models. It's worth noting that none of the available studies have investigated PLs in NTN systems. This Ph.D. research has been motivated by these key challenges involving future networks. In this thesis, our work sheds light on the PLS performance in NTNs. Therefore, understanding the outage and error performance of the physical layer for future networks from the communication point of view is of great importance to studying the security aspect. In the first part of the thesis, we study the hybrid RF/FSO transmission and mixed RF-FSO approaches for SatCom motivated by the complementary characteristics of RF and FSO communication. In the first proposed model, we assume a single-hop SatCom system where the satellite selects RF or FSO links based on weather conditions collected from sensors and employed for context awareness. We obtain the outage probability (OP) expressions by considering various weather situations to assess the performance of the proposed network. In addition, asymptotic analysis is carried out to determine the diversity order. We also consider the effects of non-zero boresight pointing errors and show how aperture averaging can significantly mitigate the effects of misalignment and atmospheric turbulence. The results reveal that the suggested technique outperforms dual-mode traditional hybrid RF/FSO communication in terms of OP while providing some power gain. Then, we present a novel downlink dual-hop SatCom model using an intermediate HAPS node. FSO communication is adopted between the LEO satellite and the HAPS node, whereas a hybrid FSO/RF transmission mechanism is considered between the HAPS node and the ground station (GS). The satellite chooses the HAPS node with the best signal-to-noise ratio in the initial phase of transmission. The signal is decoded and forwarded to the GS in the second phase by the designated HAPS. To examine the proposed system's performance, OP expressions for exponentiated Weibull and shadowed-Rician fading models are obtained, by considering atmospheric turbulence, scattering, stratospheric attenuation, path loss, and pointing errors. Aside from that, asymptotic analysis and diversity gain are obtained. The effect of the aperture averaging technique, wind speed, and temperature are also explored. Finally, the results show that using a HAPS node increases system performance and that the suggested model outperforms all other models currently in use. Thereafter, we investigate the performance of a multiple-hop mixed RF-FSO communication-based decode-and-forward protocol for multicast networks. So far, delivering real-time applications to a large number of users at the same time has been seen as a promising technique to deal with high data traffic demands. Thus, we present two realistic use-cases. In the first model, we propose a mixed RF/FSO/RF communication strategy assisted by a HAPS node, in which a terrestrial node communicates with a cluster GSs via two HAPS systems. In the second model, we consider that due to substantial attenuation produced by wide propagation distances, we consider that line of sight (LOS) connectivity between the two HAPS systems is unavailable. As a result, we propose a mixed RF/FSO/FSO/RF communication system supported by an LEO satellite. Closed-form expressions of OP and bit error rate are obtained for the given scenarios. In addition, diversity gains are derived to show the asymptotic behavior of the proposed models. For both scenarios, ergodic capacity and energy efficiency (EE) are also presented. Finally, simulation results are presented in order to verify the theoretical derivations. The obtained results demonstrate that the satellite-assisted mixed RF/FSO/FSO/RF model outperforms the HAPS-assisted mixed RF/FSO/RF model in terms of OP, whereas the HAPS-assisted mixed RF/FSO/RF scenario achieves better EE. From the security perspective, we propose different wiretap models and analyze the PLS performance. At first, for a HAPS-aided SatCom system, we study the secrecy performance of RF eavesdropping. In the proposed architecture, FSO communication is used between HAPS and the LEO satellite, while RF communication is used between HAPS and the GS as LOS communication cannot be established. Closed-form of secrecy outage probability (SOP) and probability of positive secrecy capacity (PPSC) expressions are derived to interpret the overall secrecy performance of the proposed approach. We also consider the impact of the pointing error and shadowing severity parameters. In what follows, we introduce optical eavesdropping in NTNs for different practical scenarios. In this context, we consider that a HAPS eavesdropper node is seeking to collect sensitive data from an LEO satellite, and a UAV eavesdropper is attempting to intercept confidential data from a HAPS node. We obtain closed-form SOP and PPSC expressions and validate them with Monte Carlo (MC) simulations to evaluate the overall performance of both scenarios. Secondly, a satellite eavesdropping strategy is presented, in which an attacker spacecraft trying to intercept optical communications established between an LEO satellite and a HAPS node. We assume satellite-to-HAPS (downlink) and HAPS-to-satellite (uplink) optical communications, in which the eavesdropper spacecraft can capture either the transmitted or received signal. The average secrecy capacity and SOP expressions are obtained and validated with MC simulations to quantify the secrecy performance. We also investigate secrecy throughput performance. Finally, according to our findings, turbulence-induced fading has a considerable impact on the secrecy performance of FSO communication.
Thesis(Ph.D.) -- Istanbul Technical University, Graduate School, 2022
Anahtar kelimeler
satellite networks, uydu ağları