BAC-NOMA for Secondary Transmission

Introducing backscatter (BAC) devices into a legacy non-orthogonal multiple access (NOMA) network greatly improves spectrum efficiency, which provides a promising solution for the combination of internet of things (IoT) and wireless networks. Deep learning (DL) as an emerging optimization tool gradually attracts people’s interest in wireless communication area. In this letter, a BAC-NOMA network is investigated, where a sum-rate maximization problem is formulated and the closed-form solution of backscattering coefficient is derived. The original problem is transformed and solved by a semi-definite relaxation (SDR) based algorithm and a learning based algorithm. The simulation results show that both algorithms have their own advantages and disadvantages and should be chosen wisely according to actual situations.

I. Introduction

Ultra-massive machine type communications (umMTC) as a key technique of the envisioned sixth-generation (6G) communication system is the focus of research [1]. The main feature of umMTC is that massive low-power devices, e.g., energy-constrained Internet of Things (IoT) sensors, are connected. It is challenging to serve a huge number of devices simultaneously due to the spectrum constraint.