One of the key applications for the Internet of Things (IoT) is the eHealth service that targets sustaining patient health information in digital environments, such as the Internet cloud with the help of advanced communication technologies. In eHealth systems, wireless networks, such as wireless local area networks (WLAN), wireless body sensor networks (WBSN), and wireless medical sensor networks (WMSNs), are prominent technologies for early diagnosis and effective cures. The next generation of these wireless networks for IoT-based eHealth services is expected to confront densely deployed sensor environments and radically new applications. To satisfy the diverse requirements of such dense IoT-based eHealth systems, WLANs will have to face the challenge of assisting medium access control (MAC) layer channel access in intelligent adaptive learning and decision-making. Machine learning (ML) offers services as a promising machine intelligence tool for wireless-enabled IoT devices. It is anticipated that upcoming IoT-based eHealth systems will independently access the most desired channel resources with the assistance of sophisticated wireless channel condition inference. Therefore, in this study, we briefly review the fundamental models of ML and discuss their employment in the persuasive applications of IoT-based systems. Furthermore, we propose Q-learning (QL) that is one of the reinforcement learning (RL) paradigms as the future ML paradigm for MAC layer channel access in next-generation dense WLANs for IoT-based eHealth systems. Our goal is to contribute to refining the motivation, problem formulation, and methodology of powerful ML algorithms for MAC layer channel access in the framework of future dense WLANs. This paper also presents a case study of next-generation WLAN IEEE 802.11ax that utilizes the QL algorithm for intelligent MAC layer channel access. The proposed QL-based algorithm optimizes the performance of WLAN, especially for densely deployed devices environment.
Read full abstract