Scholarlink

<p>With the rapid development of 5G and Internet of Things (IoT) technologies, massive machine-type communication (mMTC) faces critical challenges in random access channel (RACH) congestion and energy inefficiency. This paper proposes an uplink random access optimization framework called Non-Orthogonal Multiple Access with Short Packet Communication via Multi-Agent Q-Learning (NOMA-SPC-MAQL), specifically designed for high-density short-packet communication scenarios in mMTC to enhance both energy efficiency and access performance. The proposed scheme employs NOMA-based Slotted ALOHA (NOMA-SA) to enable multi-device time-frequency resource sharing, utilizes Multi-Agent Q-learning (MAQL) to model the system as a Multi-Agent Markov Decision Process (MA- MDP) for optimal strategy learning, and incorporates Access Class Barring (ACB) to dynamically regulate access load in dense deployments. A tailored objective function minimizes total energy consumption while ensuring access success probability. Simulation results confirm that our method significantly reduces energy consumption while maintaining required access success rates under dense deployment conditions, demonstrating its advantages in both energy efficiency and reliability.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Energy-efficient uplink random access for mMTC using Multi-agent Q-learning and ACB-enhanced NOMA

  • 期刊论文

  • 发表日期：

    2026年1月7日

查看全文
  • Xu Zhao,
  • Zeyu Wu,
  • Xiang Wei,
  • Jie Ding

摘要

With the rapid development of 5G and Internet of Things (IoT) technologies, massive machine-type communication (mMTC) faces critical challenges in random access channel (RACH) congestion and energy inefficiency. This paper proposes an uplink random access optimization framework called Non-Orthogonal Multiple Access with Short Packet Communication via Multi-Agent Q-Learning (NOMA-SPC-MAQL), specifically designed for high-density short-packet communication scenarios in mMTC to enhance both energy efficiency and access performance. The proposed scheme employs NOMA-based Slotted ALOHA (NOMA-SA) to enable multi-device time-frequency resource sharing, utilizes Multi-Agent Q-learning (MAQL) to model the system as a Multi-Agent Markov Decision Process (MA- MDP) for optimal strategy learning, and incorporates Access Class Barring (ACB) to dynamically regulate access load in dense deployments. A tailored objective function minimizes total energy consumption while ensuring access success probability. Simulation results confirm that our method significantly reduces energy consumption while maintaining required access success rates under dense deployment conditions, demonstrating its advantages in both energy efficiency and reliability.