Energy-Efficient Short Packet Communications for Uplink NOMA-Based Massive MTC Networks

Abstract

The 5th generation (5G) mobile networks and beyond need to support massive machine-type communications (MTC) devices with limited available radio resources. In this paper, we study the power-domain non-orthogonal multiple access (NOMA) technology to support energy-efficient massive MTC networks, where MTC devices exchange information using sporadic and low-rate short packets. We investigate the subchannel allocation and power control policy to maximize the achievable effective energy efficiency (EE) for uplink NOMA-based massive MTC networks, taking into account of short-packet communication characteristics. We model the subchannel allocation problem as a multi-agent Markov decision process and propose an efficient Q-learning algorithm to solve it. Furthermore, we obtain the optimal transmission power policy by approximating the achievable effective rate of uplink NOMA-based short packet communications. Compared with the existing OFDMA scheme, simulations validate that the proposed scheme can improve the achievable effective EE of massive MTC networks with 5.93%.

A NOMA-Based Q-Learning Random Access Method for Machine Type Communications

“…The use of reinforcement learning has great potential in MTC networks [8], [10]- [12], [14], [15], specially the widely adopted Q-Learning algorithm, because it is model-free and can be implemented in a distributed fashion. By modeling the RA in an MTC network as a Markov Decision Process (MDP) allows us to use Q-Learning.…”

“…Whilst the MTC devices do select the best BS in this scheme, thus efficiently organizing RA within a cell, it does not deal with the growth in density of mMTC networks and therefore overload is still a problem. In [15], Non-Orthogonal Multiple Access (NOMA) [16] and Q-Learning are utilized in order to maximize energy efficiency in short packet communications. The method in [15] makes use of Q-Learning for pairing devices in sub channels.…”

“…In [15], Non-Orthogonal Multiple Access (NOMA) [16] and Q-Learning are utilized in order to maximize energy efficiency in short packet communications. The method in [15] makes use of Q-Learning for pairing devices in sub channels. In order to maximize energy efficiency, they propose a power allocation scheme, finding the optimal transmit power for each device.…”

“…In [13] machine learning techniques are not used. Finally, compared to [15], we are looking to improve throughput rather than energy efficiency, although our power control scheme prevents the excessive use of power.…”

HARQ Assisted Short-Packet Communications for Cooperative Networks Over Nakagami-m Fading Channels

“…Short-packet communications are widely existing in multiple IoT applications and considering the finite packet-length coding is realistic for real communication systems [14]- [16]. Chen et al [17] proposed a wireless-powered IoT network with finite packet-length coding, where the effectivethroughput (ET) and effective-amount-of-information are adopted as performance metrics and maximized by optimizing the delay and packet error rate.…”