Sector Throughput Using Frequency-and-Time Domain Channel-Dependent Packet

VIP免费

3.0 朱铭铭 2024-09-20 7 4 204.15KB 5 页 100积分

侵权投诉

Scheduling with Channel Prediction in OFDMA Downlink Packet Radio Access

Yoshiaki Ofuji, Akihito Morimoto, Hiroyuki Atarashi, and Mamoru Sawahashi

IP Radio Network Development Department, NTT DoCoMo, Inc.

3-5 Hikari-no-oka Yokosuka-shi, Kanagawa-Ken, 239-8536 Japan

Abstract-This paper presents the throughput performance using frequency-and-

time domain channel-dependent packet scheduling and adaptive modulation

and coding (AMC) that exploit channel prediction in OFDMA packet radio

access based on system-level simulations with a multi-cell configuration. The

simulation results show that, for the full-queue traffic model without considering

the traffic types, by employing the channel prediction, the sector throughput

performance using 32 chunks (chunk bandwidth of 288 kHz) is improved by

approximately 15% compared to the case without using channel prediction for

the GSM Typical Urban channel model with the maximum Doppler frequency

of 55.5 Hz and a 10-MHz channel bandwidth. Furthermore, assuming the same

channel conditions, when the real-time(RT) traffic and non-real-time (NRT)

traffic models are taken into account, the user throughput performance of the

NRT users is improved by approximately 10 to 15% by employing channel

prediction compared to the case without using channel prediction. Therefore,

we show that the channel prediction is effective in improving the throughput

performance based on system-level evaluation, even when mixed- traffic models

are taken into account.

I. INTRODUCTION

Investigation on the long-term evolution of the UMTS Terrestrial Radio

Access (UTRA) and the UMTS Terrestrial Radio Access Network

(UTRAN) was initiated as a study item, i.e., Evolved UTRA and UTRAN,

in the 3rd Generation Partnership Project (3GPP) [1]. In this study item,

the UTRA radio interface mainly targets a channel bandwidth wider than

5 MHz. Furthermore, OFDM-based radio access is considered to be

promising as a downlink radio access scheme, since it is inherently robust

against time dispersion in a multipath fading channel. In the Evolved UTRA

and UTRAN, one of the most important requirements is to achieve high

spectrum efficiency. To achieve this, the application of packet scheduling

associated with the multi-user diversity effect is essential, in which multiple

users share the total channel bandwidth and the base station (BS) controls

the radio resource assignment to these users. Since an OFDM signal

comprises many sub-carriers in the frequency domain, the channel

bandwidth can be segmented into multiple frequency blocks, i.e., frequency

chunks. Then, the optimum frequency chunk allocation based on the channel

quality of each chunk (hereafter frequency-and-time domain channel-

dependent packet scheduling) [2],[3] is very beneficial to improving sector

throughput thanks to multi-user diversity in a frequency-selective

(multipath) fading channel. When frequency-and-time domain channel-

dependent packet scheduling is applied, each chunk is assigned to nearly

the best user who has the best channel quality represented as the received

signal-to-interference plus noise power ratio (SINR). Thus, packet decoding

error hardly occurs when the channel variation in the time domain is almost

constant. However, when the channel conditions vary more frequently

than the control loop delay of the packet scheduling, this large channel

variation causes decoding errors in the received packets. In such a situation

with time-varying channel fluctuations, channel prediction of each chunk

based on the channel quality information fed-back from the receiver can

mitigate the packet decoding errors [4].

Therefore, this paper presents the sector throughput performance using

frequency-and-time domain channel-dependent packet scheduling and

adaptive modulation and coding (AMC) that exploit channel prediction in

OFDMA packet radio access based on system-level simulations with a

multi-cell configuration. The combined effect of channel-dependent

scheduling with channel prediction, AMC, and hybrid ARQ was presented

in [4] assuming TDMA radio access with a single carrier. In [4], the potential

benefit of channel prediction in time-domain channel-dependent scheduling

was proven in a single-cell model assuming channel prediction error

following a lognormal distribution. Thus, in the current paper, we investigate

the effect of channel prediction in the frequency-and-time domain channel-

dependent scheduling in OFDMA radio access using more realistic models.

First, we investigate the effect of channel prediction in system-level

simulations using a three-sectored seven-cell configuration considering

distance-dependent path loss, shadowing variation, and instantaneous

multipath Rayleigh fading. Then, the influence of actual variations in inter-

cell interference is considered. Second, we actually estimate future channel

variation from the extrapolation of the M-th order polynomial, which is

generated by curve fitting employing the least square method (LSM). Third,

we apply packet scheduling criteria considering the delay requirement of

each traffic flow, the type of packet whether initially transmitted packet or

retransmitted packet, and the channel quality of each accessing user at

each chunk[5]. Note that we use the notation of channel-dependent

scheduling in the paper to emphasize the criterion of channel quality for

each accessing user at each chunk. Fourth, we consider the mixed traffic

conditions containing real-time (RT) traffic users and non-real-time (NRT)

traffic users following a VoIP traffic model and Web browsing traffic model,

respectively. The rest of the paper is organized as follows. We present the

channel prediction method in the frequency-and-time domain channel-

dependent scheduling and AMC in Section II. Next, Section III describes

the packed scheduling method assumed in the paper. Then, after describing

the simulation configuration in Section IV, Section V presents the simulation

results.

II. CHANNEL PREDICTION METHOD IN CHANNEL-DEPENDENT PACKET

SCHEDULING

Figure 1 illustrates the principle of channel prediction in with frequency-

and-time domain channel-dependent packet scheduling OFDMA packet

radio access. Packet scheduling has a control loop delay, which is defined

as the delay from the timing of the SINR measurement for the packet

scheduling to the timing of the reception of the corresponding scheduled

packet. Therefore, we apply the channel prediction method in order to

compensate for the estimation error in the measured received SINR caused

by the control loop delay of the packet scheduling. In the channel prediction,

we generate the M-th order polynomial by fitting the past instantaneous

received SINR samples at every transmission time interval (TTI) employing

the LSM. More specifically, the channel prediction method is expressed

by using the following equation. First, in order to generate the curve to fit

the instantaneous received SINR variation of the past measurements, we

assume the curve of the M-th order polynomials, )(

ˆn

tu , which is given by

∑

nmn tatu

)(

ˆ. (1)

Figure 1. Channel prediction with frequency-and-time domain channel-

dependent scheduling.

Chunk 1

Chunk 2

Chunk 3

Chunk L

Frequency

Time

SINR

Measured SINR

Predicted SINR

TTI length Control loop delay of scheduling

Chunk 1

Chunk 2

Chunk 3

Chunk L

Frequency

Time

SINR

Measured SINR

Predicted SINR

TTI length Control loop delay of scheduling

1589

文档加载中……请稍候！
如果长时间未打开，您也可以点击刷新试试。

下载文档到电脑，查找使用更方便

100 积分 4人已下载

立即下载 VIP免费下载

摘要：

SectorThroughputUsingFrequency-and-TimeDomainChannel-DependentPack\etSchedulingwithChannelPredictioninOFDMADownlinkPacketRadioAccessYoshiakiOfuji,AkihitoMorimoto,HiroyukiAtarashi,andMamoruSawahashi\IPRadioNetworkDevelopmentDepartment,NTTDoCoMo,Inc.3-5Hikari-no-okaYokosuka-shi,Kanagawa-Ken,239-8536Ja...

展开>> 收起<<

Sector Throughput Using Frequency-and-Time Domain Channel-Dependent Packet.pdf

共5页,预览1页

还剩页未读，继续阅读

Sector Throughput Using Frequency-and-Time Domain Channel-Dependent Packet

相关推荐

开通VIP享超值会员特权

作者详情

相关内容

推荐作者

热门标签

举报选择: