Scholarlink

<p>Wireless Body Area Network (WBAN) grabs the attention of researchers as it is a needy technology for e-healthcare. Due to its limited protocol support, providing Quality of Service (QoS) is a challenging task. Despite the previous research, a critical research gap persists in superframe time slot allocation to avoid collisions among the Bio Sensor Nodes (BSNs). To address this, a robust Superframe Contention Slot Scheduling (SCSS) algorithm is proposed to identify potential BSNs to occupy slots under the Exclusive Access Period (EAP) slots of the superframe. It incorporates a Markov Decision Process (MDP) policy function to select these BSNs using several runtime operational parameters. The deep Reinforcement Learning (DRL) algorithm implements the MDP process to get the optimized rewards. The proposed algorithm adopts sleep schedules and energy harvesting techniques to avoid the dead nodes problem. A Two-state Markov Chain (TMC) model is adopted to theoretically analyze the expected throughput and loss performance. The comprehensive simulation results show a 41.8% improvement in throughput, a 42.62% reduction in packet loss, a 6.9% less energy consumption, and enhanced overall network performance compared to existing protocols.</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

Superframe contention slot scheduling (SCSS): deep reinforcement learning-based time slot allocation for wireless body area network

  • 期刊论文

  • 发表日期：

    2025年2月18日

查看全文
  • Vamsi Kiran Mekathoti,
  • B. Nithya

摘要

Wireless Body Area Network (WBAN) grabs the attention of researchers as it is a needy technology for e-healthcare. Due to its limited protocol support, providing Quality of Service (QoS) is a challenging task. Despite the previous research, a critical research gap persists in superframe time slot allocation to avoid collisions among the Bio Sensor Nodes (BSNs). To address this, a robust Superframe Contention Slot Scheduling (SCSS) algorithm is proposed to identify potential BSNs to occupy slots under the Exclusive Access Period (EAP) slots of the superframe. It incorporates a Markov Decision Process (MDP) policy function to select these BSNs using several runtime operational parameters. The deep Reinforcement Learning (DRL) algorithm implements the MDP process to get the optimized rewards. The proposed algorithm adopts sleep schedules and energy harvesting techniques to avoid the dead nodes problem. A Two-state Markov Chain (TMC) model is adopted to theoretically analyze the expected throughput and loss performance. The comprehensive simulation results show a 41.8% improvement in throughput, a 42.62% reduction in packet loss, a 6.9% less energy consumption, and enhanced overall network performance compared to existing protocols.