인지 무선 및 보안을 고려한 협력 NOMA 기반 5G 시스템의 자원 최적화 연구

Abstract

Currently, the deployment of fifth-generation (5G) networks is underway in numerous countries and regions, and it is expected to bring about a new era of innovation and connectivity in the upcoming years. However, to satisfy the increasing demands for high data speed, minimal delay, and pervasive computing in 5G and future wireless communication technologies, there is a critical request for advanced communication system design. In particular, the primary challenge is the efficient assignment of resources, which refers to managing limited quantities such as bandwidth, power, and time in wireless communications. Furthermore, advancements in technologies accompany the progress of wireless communication systems to enable faster data transfer rates, lower latency, and more effective use of network resources. Based on the above observations, this thesis explores and investigates efficient resource allocation strategies for promising 5G and beyond 5G candidate technologies in terms of non-orthogonal multiple access (NOMA), beamforming, and mobile edge computing (MEC). To further enhance the throughput and extend the coverage of the 5G networks, in this thesis, we utilize collaborative communication techniques including relay selection, cognitive radio (CR), and power allocation schemes.

In addition, to address the problem of limited battery life in energy-constrained wireless devices, energy harvesting (EH) techniques from ambient radio frequency (RF) signals have emerged as a promising solution. EH techniques provide a sustainable and eco-friendly manner to power devices, especially in remote locations where traditional power sources may not be available. In this respect, simultaneous wireless information and power transfer (SWIPT) offers another option in EH methods since it enables simultaneously both RF EH and information decoding for the user, utilizing the same RF signal. Moreover, SWIPT eliminates the need for separate power and data transmission systems, which reduces the complexity and cost of the overall system. Motivated by the benefits mentioned above, SWIPT, linear, and non-linear EH models are part of the wireless network designs studied in this dissertation. Despite the potential of cooperative communications in NOMA systems along with CR and SWIPT technologies to meet the requirements of 5G networks, ensuring wireless security remains a significant challenge. In wireless transmission environments, confidential information is particularly at risk of being intercepted by eavesdroppers. Although industry and academia have proposed cryptographic encryption and decryption methods to overcome this issue, these techniques rely heavily on complex decoding/encoding algorithms and encryption key management, which require significant computing power and resource consumption. Furthermore, these complex algorithms can be compromised, as eavesdroppers nowadays have access to high computational power. Therefore, this thesis proposes an alternative approach to conventional security techniques by utilizing physical layer security (PLS) in our systems, which takes advantage of the characteristics of wireless channels. PLS provides security regardless of the computing ability of the communication equipment even if the eavesdropper has strong computing capability. Furthermore, resource allocation optimization is a crucial task in 5G networks, which aims to maximize the utilization of network resources while meeting the quality of service (QoS) requirements of the users. Nevertheless, resource allocation optimization is a complex problem, and solving it can be computationally intensive. Therefore, reducing computational complexity is an important research field in wireless communications that brings several benefits. For instance, simplifying computational complexity ensures data is processed faster, preventing delays and ensuring smooth communication. Moreover, as the number of users in wireless communication systems increases, so does the computational load. By reducing the computational complexity, the system can scale easier without requiring additional resources. Hence, in this thesis, we investigate low-complexity algorithms that can significantly reduce the computational complexity of solutions that lead to resource allocation optimization in 5G networks. In particular, we study metaheuristics techniques that utilize a trade-off between randomization and local search to obtain an approximately optimal solution. In addition, machine learning (ML)-based schemes are considered in this thesis to reduce computational complexity by learning patterns in the data and optimizing the processing accordingly.

Firstly, we study the problem of secure computation efficiency (SCE) in a NOMA MEC system with a nonlinear EH user and a power beacon in the presence of an eavesdropper. To further provide a friendly environment resource allocation design, wireless power transfer is applied. The SCE problem is solved by jointly optimizing the transmission power, the time allocations for energy transfer, the computation time, and the central processing unit (CPU) frequency in the NOMA-enabled MEC system. The problem is non-convex and challenging to solve because of the complexity of the objective function in meeting constraints that ensure the required QoS, such as the minimum value of computed bits, limitations on total energy consumed by users, maximum CPU frequency, and minimum harvested energy and computation offloading times. Therefore, a low-complexity particle swarm optimization (PSO)-based algorithm is proposed to solve this optimization problem. For comparison purposes, time division multiple access and fully offloading baseline schemes are investigated. Simulation results show the superiority of the proposed approach over baseline schemes.

Secondly, we investigate a beamforming design with artificial noise (AN) to improve the security of a multi-user downlink, multiple-input single-output (MISO) NOMA-CR network with SWIPT. To further support power-limited, battery-driven devices, EH users are involved in the proposed network. Specifically, we investigate the optimal AN, power-splitting ratios, and transmission beamforming vectors for secondary users and EH users to minimize the transmission power of the secondary network, subject to the following constraints: a minimum signal-to-interference-plus-noise ratio at the secondary users, minimum harvested energy by secondary users and EH users, maximum power at the secondary transmitter, and maximum permissible interference with licensed users. The proposed solution for the challenging non-convex optimization problem is based on the semidefinite relaxation method. Numerical results show that the proposed scheme outperforms the conventional scheme without AN, the zero-forcing-based scheme, and the space-division multiple-access-based method.

Thirdly, we consider a cooperative non-linear SWIPT-enabled NOMA system with a non-linear EH user. Specifically, we investigate two optimization problems. First, we minimize transmission power, and second, we maximize energy efficiency subject to meeting QoS constraints. Furthermore, we develop the optimal solution based on convex optimization and the exhaustive search (ES) method to validate the results of the proposed PSO-based framework. Afterward, we investigate the performance of five swarm intelligence-based baseline schemes and evaluate an additional low-complexity solution based on the cuckoo search technique. For comparison purposes, we use orthogonal multiple access (OMA), equal power splitting (EPS), and time-fixed (TF) baseline schemes. From the results in terms of total transmission power and energy efficiency, the proposed SWIPT NOMA network outperforms the benchmark schemes, and the proposed PSO-based framework achieves the nearest performance to the optimal scheme with lower complexity than obtained by the comparative swarm intelligence techniques and from convex optimization with the ES method.

Finally, we design a novel artificial intelligence (AI)-based framework for maximizing the secrecy energy efficiency (SEE) in FD cooperative relay underlay CR-NOMA systems with imperfect successive interference cancellation that are exposed to multiple eavesdroppers. First, we formulate the non-convex SEE optimization problem as bi-level optimization, subject to constraints that satisfy the QoS requirements of secondary users. In particular, the outer problem is solved with ensemble learning (EL) to select the optimal relay. Regarding the inner problem, we propose a quantum particle swarm optimization (QPSO)-based technique to optimize power allocation. In addition, for comparison purposes, we describe a cooperative relay CR network with OMA, rate-splitting multiple access, and half-duplex technologies. Moreover, we evaluate comparative schemes based on machine learning algorithms and swarm intelligence baseline schemes. Furthermore, the proposed EL-aided QPSO-based framework achieves performance close to the optimal solutions, with a meaningful reduction in computation complexity.