Adaptive Scheduling and Power Control for Multi-Objective Optimization in IEEE 802.15.6 Based Personalized Wireless Body Area Networks
Beom-Su Kim, Babar Shah, Ki‐Il Kim
Abstract
Multi-objective optimization (MOO) has been a topic of intense interest in providing flexible trade-offs between conflicting optimization criteria in wireless body area networks (WBANs). To solve diverse multi-objective optimization problems (MOPs), conventional resource management schemes have dealt with the classic issues of WBANs, such as traffic heterogeneity, emergency response, and body shadowing. However, existing approaches have difficulty achieving MOO because, despite the personalization of WBANs, they still miss the new constraints or considerations derived from user-specific characteristics. To address this problem, in this paper, we propose an adaptive scheduling and power control scheme for MOO in personalized WBANs. Specifically, we investigate the existing scheduling and power control schemes for solving MOPs in WBANs, clarify their limitations, and present two feasible solutions: priority-based adaptive scheduling and deep reinforcement learning (DRL) power control. By integrating these two mechanisms in compliance with the IEEE 802.15.6 standard, we can jointly improve the optimization criteria, that is, differentiated quality of service (QoS), transmission reliability, and energy efficiency. Through comprehensive simulations, we captured the performance variations under realistic WBAN deployment scenarios and verified that the proposed scheme can achieve a higher throughput and packet delivery ratio, lower power consumption ratio, and shorter delay compared with a conventional approach.