(An) energy-conscious access point system to increase the energy efficiency of iot devices for medi-care iot services in wi-fi networks

Abstract

Wi-Fi networks are the most common and globally used wireless access communication technology. In addition, Wi-Fi network infrastructure can be used easily anywhere, anytime in the world, and guarantees high data transmission speed. In the trend of increasing wireless network traffic, Wi-Fi protocol technology has been developed to optimize and increase bandwidth, transmission distance, and transmission speed. However, recently, research to reduce the energy consumption of mobile terminals (MT) has been actively conducted, focusing on protocols such as 802.11ax. In particular, with the high demand and spread of the Internet of Things (IoT) over Wi-Fi networks, research to reduce the energy consumption of IoT devices, such as increasing the energy efficiency of battery-typed IoT devices, has been actively conducted recently as a significant task to be solved. Among these studies, IoT data transmission with bio-information is performed with high reliability to provide medicare IoT services in an indoor environment such as a hospital, and efforts to increase the energy of IoT devices are essential. In a situation where Wi-Fi wireless network traffic increases, the cells covered by the Wi-Fi network's access point (AP) are also becoming smaller and more densely populated. Due to the dense AP, mobile terminals (MTs), including IoT devices, exist within the coverage of multiple overlapped cells. In this situation, the MT generally uses a method of accessing the AP that has received the strongest signal. However, in this legacy method, a phenomenon in which MTs are concentrated in a specific AP brings contentions between MTs, and transmission collisions occur in many cases. In addition, low bandwidth utilization in the non-crowded APs reduces overall network performance. This reduces energy efficiency and causes a long delay time due to the retransmissions of the MT that include IoT devices. Therefore, it is necessary to solve the problem of selecting the optimal AP considering IoT devices' energy efficiency and latency in a multi-coverage Wi-Fi network environment. In order to solve this problem, the development of a more intelligent AP system was proposed to increase the energy efficiency of IoT devices for reliable IoT services. The proposed energy-conscious AP (hereafter referred to as eAP) implements various functions in software to reduce the energy consumption of IoT devices. First, the environment of TCP protocol-based transmission in which ACKs are exchanged for reliable data transmission was considered. In addition, this dissertation tried to propose an eAP system operating in 802.11n or 802.11ax Wi-Fi, which are the latest WLAN protocols. The existing AP is a simple router-like hardware that acts as a relay to send and receive data between IoT devices or MTs and servers. The proposed eAP is an intelligent AP that control the transmission period, adaptive transmitting power, and optimal AP selection of IoT devices in consideration of various layers of information. The proposed eAP analyzes the characteristics of IoT devices to adjust the transmission period, generates and delivers prompt TCP ACKs, and increases the energy efficiency of IoT devices through caching and retransmission functions. In addition, by integrating multiple IoT data traffic and delivering it to the server in a burst-type packet, the number of network traffic between the AP and the server was reduced to improve network performance. The proposed eAP system selects an optimal AP that can increase IoT devices' energy efficiency and latency efficiency in a multi-coverage environment. The eAP calculates the optimal transmitting power and delivers this to the IoT devices for IoT data transmission through network load balancing. The eAP delivers the control message to the IoT device and controls to select the optimal AP that significantly reduces the possibility of collision during IoT data transmission through network load balancing. In addition, according to the selection of the optimal AP, the IoT device is controlled to transmit the IoT data with the optimal transmitting power in consideration of the distance between the IoT device and the selected AP. In order to solve the problem of selecting the optimal AP considering the energy efficiency and latency of IoT devices, reinforcement learning, a branch of machine learning, was used to solve the problem. For this, we propose a new energy and latency reinforcement learning model (Energy and Latency Reinforcement Learning, hereafter referred to as EL-RL) based on the Proximal Policy Optimization (PPO) algorithm, which has fast learning and processing times and excellent performance in reinforcement learning. Through the proposed EL-RL model, it is possible to select the optimal AP for a given environment by considering the energy and latency of IoT devices in real-time. The energy consumption of the IoT device was mathematically analyzed through the proposed eAP system. The IoT performance was analyzed by constructing a simulation environment based on the parameter values of the actual Wi-Fi device based on 802.11ax. To analyze the energy performance of this IoT device, the energy consumption of the IoT device according to the transmission period, the energy consumption of the IoT device according to the change of the DTIM (Delivery of Traffic Indication Map) value, and the set target SINR (Signal to Interference plus Noise Ratio) value change. Also, we calculate the probability of transmission collision and analyze the energy consumption of the IoT device according to the collision probability change through mathematical calculations and simulations. In addition, the expected life expectancy of the battery-type IoT device was calculated, and the delay time of the IoT device was reduced by using the proposed eAP. Through such a proposed eAP system, it is possible to increase the energy efficiency of MTs, including IoT devices, improve service quality by lowering latency, and significantly improve the expected life expectancy of IoT devices. In addition, the eAP system that is implemented by software is expected to provide a much higher quality of medicare IoT service than before through simple software installation on the existing AP system.