Slot Allocation in Wireless Body Area Networks (WBANs) using Q-learning Approaches

Wireless Body Area Networks (WBANs) are vital in modern healthcare, allowing for real-time monitoring via wearable and implanted devices. However, efficient Medium Access Control (MAC) remains a critical challenge, particularly in achieving high throughput, low packet delay, reliable packet delivery ratio (PDR), and energy efficiency under varying network conditions. This study addresses these challenges by proposing a Q-learning-based MAC protocol (QL-MAC) for slot allocation in WBANs. It dynamically utilizes reinforcement learning (RL) to optimize slot allocation, improving network performance across key metrics. Compared to other MAC protocols, such as adaptive MAC (ADT-MAC), dynamic multi-threshold MAC (DMTMMAC), traffic aware MAC (TA-MAC), multi-converge cast MAC (McMAC), and IEEE 802.15.6 MAC. QL-MAC outperformed them. The results show that QL-MAC achieves faster throughput, fewer packet delays, and better packet delivery ratio (PDR) at different network densities. Furthermore, it consumes less energy, making it an excellent choice for energy-constrained WBAN environments. These findings support the use of reinforcement learning to manage the special problems of WBANs and ensure dependable and energy-efficient communication for healthcare applications.