Resource allocation issues in broadband wireless networks with OFDM signaling
Publication or External Link
Wireless broadband technologies are anticipated to flourish in thenext few years, due to the increasing demand for wireless connectivityand the need to support enhanced services and applications in local-or wide-area environments. The primary goal in a communications systemis Quality of service (QoS) provisioning to users, which depends onprocedures that span several communication layers. Although independentconsideration of different layers simplifies system design, it oftenturns out to be insufficient for wireless networks. Cochannelinterference between users that reuse the limited spectrum and theresulting impact of local adaptation actions on overall network performance impose layer interactions in wireless systems. The purposeof this work is to identify and study some of the issues that arisefrom the synergy between the physical and the MAC layer in the contextof multiple access schemes with orthogonal channels.
Using the essential feature of channel orthogonality as a baseline,our approach places emphasis on Orthogonal Frequency Division Multiplexing(OFDM), which is an emerging multiple access and signaling method for futurewireless broadband networks. In OFDM, the broadband spectrum isdivided into orthogonal, narrow-band subcarriers and user symbols aresplit into subsymbols, which are transmitted in parallel over thosevariable-quality subcarriers. OFDM transmission reduces the effectivesymbol transmission rate, simplifies equalization at the receiver andprovides high immunity to inter-symbol interference and delay spread.Furthermore, it defines a framework for flexible adaptation to varyingchannel conditions, by allowing transmission parameter control foreach subcarrier.
We first address the joint problem of channel allocation withsimultaneous adaptation of modulation level and transmission power ina multi-cell OFDM network. We study the impact of those parameters oncochannel interference and channel reuse and present two classes ofcentralized heuristic algorithms to perform the allocation. Next, we focus on a single-cell multi-user system with modulationcontrol and study the problem of subcarrier assignment to userssubject to time resource constraints. We study and compare integral andfractional user assignment, whereby a user is assigned to one subcarrieror can be partially assigned to multiple subcarriers. In addition, weconsider the synergy between link-layer ARQ protocols and physicallayer parameter adaptation. We consider a simple channel monitoringmethod which is based on counting received ACKs and NACKs. For asingle subcarrier, we show that the adaptation policy which maximizeslong-term average throughput per unit time is of threshold type. Wealso expand our policy to the multiple-subcarrier case with similar ordifferent channel qualities.
In the sequel, we study the impact of smart antennas and SpaceDivision Multiple Access (SDMA) on MAC layer channel allocation for a single-cell multi-user system. Our approach encompasses multipleaccess schemes with orthogonal channels, such as OFDM. We first considerthe case of unlimited transceiver resources, where a separate beam canbe formed for each user of a spatially separable cochannel user set ina subcarrier. We present heuristic algorithms to allocate subcarriersto users and adjust down-link beam patterns, transmission powers andrates with the objective to increase total achievable system rate andprovide QoS to users in the form of minimum rate guarantees. Then, we consider the allocation problem forlimited transceiver resources, which arises whenever certainreasons impose limitations on the number of beams that can beformed. We propose meaningful heuristic algorithms to jointly formbeams from corresponding transceivers and assign subcarriers andtransceivers to users, such that the total achievable system rate isincreased.