Doctor of Philosophy
School of Electrical, Computer and Telecommunications Engineering
Nowadays, wireless Access Points (APs) operating in Wireless Local Area Networks (WLAN) or Wi-Fi networks are densely deployed to satisfy user demands for high data rates. In addition, future Wi-Fi networks will be required to support devices, in terms of data communications and energy delivery, that are operating as part of an Internet of Things (IoT) eco-system. However, the increasing energy consump- tion and interference of Wi-Fi networks have become key concerns to operators and researchers. To address these concerns, a promising approach is to equip APs with Energy Harvesting (EH) capabilities. Another approach is to exploit channel bond- ing or transmit power control to increase the capacity of a Wi-Fi network. Channel bonding allows an AP to form a wider channel that offers higher data rates. It also helps increase spectrum efficiency. Lastly, transmit power control allows an AP to either increase the received signal strength at users or reduce the interference to neighboring APs.
The aforementioned approaches, however, lead to a number of problems. First, a bonded or wider channel may cause an AP to be more susceptible to interfer- ence from neighboring APs because of the use of overlapping channels. Second, as spectrum resource is limited, APs must be assigned an appropriate number of chan- nels corresponding to varying traffic demands over time. Otherwise, they may be under-provisioned, resulting in delays and poor user satisfaction, or over-provisioned, resulting in low spectrum efficiency. Third, a high transmit power level may lead to higher interference at APs sharing the same channel and also causes an AP to have higher energy expenditure. This is especially important if an AP is powered by a renewable energy source, because such an AP will experience energy outages frequently in the future if it fails to manage the use of energy, resulting in service suspension. On the other hand, a low transmit power level may result in low data rates or energy harvesting rates at IoT devices. Therefore, it is necessary to de- velop adaptive resource allocation solutions that assign one or more channel(s) and a transmit power level to APs according to their traffic load or channel state. Unfor- tunately, most resource allocation solutions to date fail to consider traffic variation and random Channel State Information (CSI). Moreover, they assume CSI is fixed or is known by APs in advance.
Luo, Yizhou, Learning Algorithms for Resource Allocation in Wireless Local Area Networks, Doctor of Philosophy thesis, School of Electrical, Computer and Telecommunications Engineering, University of Wollongong, 2021. https://ro.uow.edu.au/theses1/1072
Unless otherwise indicated, the views expressed in this thesis are those of the author and do not necessarily represent the views of the University of Wollongong.