RESOURCE ALLOCATION FOR NON-ORTHOGONAL MULTIPLE ACCESS BASED MOBILE EDGE COMPUTING NETWORKS

  • Haodong Li

Student thesis: Phd

Abstract

The computational capacity and the battery life of mobile devices are two critical aspects that limit the service quality of mobile applications with growing complexity. Therefore, driven by the envisioned sixth-generation (6G) network, the mobile edge computing (MEC) technology has been recognised as a promising technology to enhance the computational capability of mobile devices. Conventional mobile cloud computing (MCC) has a centralised structure, which cannot satisfy the ultra-low latency demand in 6G networks. Therefore, MEC is proposed to shorten the physical distance between the user and the powerful server by integrating the server at the edge of the base station (BS). Furthermore, non-orthogonal multiple access (NOMA) can be integrated with MEC to improve the service from the communication aspect, which enables more than one user to be allocated with the same radio resource. Thus, NOMA has become a promising technology to improve spectral efficiency. This thesis studies the combination of NOMA with the MEC network from four different aspects. First, the hybrid multiple access is considered in the MEC network, in which NOMA and OMA are both adopted for offloading. Secondly, the hybrid successive interference cancellation (SIC) scheme is proposed by formulating the optimisation problem with respect to the SIC decoding order and dynamically changing the SIC decoding order to improve the performance. Following that, a physical layer security (PLS) scheme is proposed by optimising the power allocation, task assignment, and SIC decoding to combat the overhearing of an eavesdropper. Fourthly, the backscatter communication (BackCom)-assisted MEC protocol is studied, in which the BackCom-equipped devices can offload by utilising a legacy signal to excite the circuits of the BackCom devices. Moreover, both offloading energy and latency minimisation are studied in this thesis by studying the convexity and providing efficient algorithms. The simulation results reveal that the proposed NOMA-MEC scheme can outperform the OMA-MEC scheme in terms of energy consumption and latency, which improves the overall offloading performance.
Date of Award1 Aug 2023
Original languageEnglish
Awarding Institution
  • The University of Manchester
SupervisorKa Chun So (Supervisor) & Zhiguo Ding (Supervisor)

Keywords

  • 5G
  • Backscatter communication
  • Physical layer security (PLS)
  • 6G
  • Wireless communication
  • Non-orthogonal multiple access (NOMA)
  • Reinforcement learning
  • Resource allocation
  • Convex optimisation
  • Mobile edge computing (MEC)

Cite this

'