Abstract
This article investigates nonorthogonal multiple access (NOMA)-based cloud radio access networks (C-RANs), where edge caching is adopted to cut down the crowdedness of the fronthaul links. We aim to maximize the energy efficiency (EE) by jointly optimizing the power allocation, analog, and digital precoding, which turns out to be an intractable nonconvex optimization problem. To tackle this problem, we first select cluster heads using the selecting cluster-head (SCH) algorithm, where the analog precoding matrix can be resolved by means of maximizing the array gains. Then, the device grouping algorithm is proposed to group devices according to the equivalent channel correlations, and thus, the NOMA devices in the same beam are capable of sharing the same digital precoding vector. Finally, the joint digital precoding design and power allocation algorithm is proposed to decompose the resultant optimization problem into two subproblems and solve them iteratively by applying the Taylor expansion operation and the minimum mean square error (MMSE) detection. Simulation results validate that the proposed NOMA-based C-RANs with a hybrid precoding (HP) scheme can achieve higher spectral efficiency and EE than the traditional orthogonal multiple access (OMA)-based approach and two-stage HP scheme.
Funding Information
  • Key Research and Development Project of Guangdong Province (2019B010156003)
  • Natural Science Foundation of Guangdong Province (2019A1515011607)
  • National Key Research and Development Project (2019YFB1804100)
  • National Natural Science Foundation of China (61971194)
  • Research Fund Program of Guangdong Key Laboratory of Aerospace Communication and Networking Technology (2018B030322004)

This publication has 36 references indexed in Scilit: