THESIS
2017
xi, 50 pages : illustrations ; 30 cm
Abstract
We consider optimizing user throughput for coordinated wireless local area network (WLAN)
under maximum power association, i.e., a user associates with the access point (AP) with
the maximum received power. The APs may be deployed anywhere and user density may
be non-uniform. To improve throughput, one may control the coverage of each AP by tuning
its transmit power to balance the signal strength users receive and interference users
suffer. Furthermore, orthogonal channels should be properly assigned to the APs to reduce
interference.
Because coverage control and channel assignment are mutually dependent, they need
to be jointly optimized. We study the joint problem to maximize proportionally fair user
throughput under maximum power association. Our work is unique, as prior wo...[
Read more ]
We consider optimizing user throughput for coordinated wireless local area network (WLAN)
under maximum power association, i.e., a user associates with the access point (AP) with
the maximum received power. The APs may be deployed anywhere and user density may
be non-uniform. To improve throughput, one may control the coverage of each AP by tuning
its transmit power to balance the signal strength users receive and interference users
suffer. Furthermore, orthogonal channels should be properly assigned to the APs to reduce
interference.
Because coverage control and channel assignment are mutually dependent, they need
to be jointly optimized. We study the joint problem to maximize proportionally fair user
throughput under maximum power association. Our work is unique, as prior works have not
studied the problem of such nature. By modeling maximum power association as Voronoi
partition, we formulate the joint problem and show that it is NP-hard. We then propose an
efficient and simple heuristic termed OPAC (Optimizing Power and Channel) which captures
the user interference in IEEE 802.11 CSMA/CA medium access protocol. Through extensive
simulation, OPAC is shown to outperform state-of-the-art schemes by a wide margin in terms
of throughput (more than 35% improvement in our simulation), fairness and delay.
Post a Comment