A_WLAN_UplinkBeamforming_Technique.pdf资源-CSDN文库

需积分: 5 115 浏览量 2024-09-08 18:01:32 上传评论收藏 2MB PDF 举报

资源推荐

资源详情

资源评论

11042 IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY, VOL. 69, NO. 10, OCTOBER 2020

A WLAN Uplink Collision-Resolving Scheme

Using Multi-User Beamforming Technique

Yi-Hsiu Lee , Student Member, IEEE, and Wen-Rong Wu , Member, IEEE

Abstract—To effectively improve the throughput of the IEEE

contention-based wireless local area network (WLAN) system,

many uplink medium access control layer (MAC) protocols based

on multi-user multiple-input multiple-output (MU-MIMO) tech-

nology have been proposed. However, a common problem of these

approaches is that many physical layer (PHY) operations, e.g.,

synchronization, are not considered. As a result, existing methods

may ﬁnd difﬁculties in real-world implementation. In this article,

we propose a simple yet practical uplink scheme for systems with

multiple antennas. The main idea is to conduct multi-user beam-

forming (MU-BF) at an access point such that it can resolve collided

signals from multiple competitive users; the proposed scheme can

effectively reduce the upstream collision probability and improve

system throughput. The other distinct advantage of the proposed

scheme is that the MAC protocol of the original IEEE 802.11

standard, i.e., distributed coordination function (DCF), can remain

unchanged. The mathematical analysis of system performance is

also presented and veriﬁed. Simulations show that the proposed

scheme signiﬁcantly outperforms conventional WLAN systems.

Index Terms—IEEE 802.11, uplink wireless local area network

(WLAN), distributed coordination function (DCF), multi-packet

reception (MPR), multi-user beamforming (MU-BF).

I. INTRODUCTION

EGINNING in 1997, the 802.11 family of the IEEE stan-

dards association began to develop a series of compre-

hensive wireless local area network (WLAN) standards [1]. In

the past few years, to meet trafﬁc demands, IEEE 802.11 has

continuously made breakthroughs in physical layer (PHY) and

medium access control layer (MAC) technologies. For example,

IEEE 802.11n [2] and 802.11ac [3] both adopt multi-antenna

systems and successfully implement single-user multiple-input

multiple-output (SU-MIMO) and downlink multi-user multiple-

input multiple-output (MU-MIMO) transmission, respectively.

The IEEE 802.11 working group has developed a widely used

MAC mechanism, i.e., distributed coordination function (DCF),

employing the carrier sense multiple access with collision avoid-

ance (CSMA/CA) strategy, for resource contention between

stations (STAs) [4]. As mentioned, downlink MU-MIMO has

Manuscript received September 4, 2019; revised December 30, 2019, April

13, 2020, and June 10, 2020; accepted June 18, 2020. Date of publication July 3,

2020; date of current version October 22, 2020. This work was supported in part

by the Ministry of Science and Technology, Taiwan, under Grants 108-2218-E-

009-027 and 109-2218-E-009-003. The review of this article was coordinated

byDr.Y.Ma.(Corresponding author: Yi-Hsiu Lee.)

The authors are with the Department of Electrical and Computer Engi-

neering, National Chiao Tung University, Hsinchu 30010, Taiwan (e-mail:

cmmint0160225.cm01g@nctu.edu.tw; wrwu@faculty.nctu.edu.tw).

Digital Object Identiﬁer 10.1109/TVT.2020.3007085

been used in IEEE 802.11ac. It uses the space division multiple

access (SDMA) technology that allows access point (AP) to

simultaneously transmit spatially separated downlink streams

to different STAs. However, uplink MU-MIMO is still not

available due to some technical challenges [5]. For supporting

simultaneous uplink multi-user transmissions, many researchers

have tried to solve the multi-packet reception (MPR) problem.

The objective is to design new protocols f or multi-channel access

[6], [7]. In this regard, coordinated [8]–[13] and uncoordinated

[14]–[19] uplink MU-MIMO based MAC proposals are then

proposed, respectively. The ideal uplink throughput of the MPR-

enabled system is ﬁrst discussed in [20]. Various mathematical

frameworks are then developed to analyze the performance of

MU-MIMO based MPR system with different considerations

[10], [12], [14]–[16]. Other related works can be seen in [21],

[22], and [23].

As well known, there are two kinds of multi-antenna transmis-

sion technology, referred to as MIMO (i.e., spatial multiplexing)

and beamforming (BF). All of the above uplink MPR protocols

adopt the former to receive multiple data streams simultane-

ously. Many of them redesign the frame ﬁeld, which may require

AP and STAs to exchange information. Although these MPR

protocols can effectively improve overall network throughput,

implementation issues in PHY layer are rarely discussed. Note

that the original frame format designed for PHY synchroniza-

tion including packet detection, timing offset estimation, and

frequency offset estimation do not consider MPR. In MPR, these

synchronization operations become difﬁcult, if not impossible,

to conduct with current WLAN PHY speciﬁcations. Since mul-

tiple antennas have been used in downlink MU-MIMO, we then

wonder if there are other approaches that can take advantage of

multiple antennas enhancing uplink throughput and at the same

time without changing the current PHY and MAC protocols.

In this article, we propose such a scheme based on multi-user

BF (MU-BF) technology. The idea is based on the fact that

receive BF conducted in AP can form multiple beams pointing

to different directions, and multiple data streams coming from

different directions (corresponding to different STAs) can be

received simultaneously. In our scheme, AP divides its service

area into multiple spatial regions (SRs) according to signal

angle-of-arrival (AoA). It i s also equipped with multiple single-

input-single-output (SISO) receivers. As a result, even if a STA

transmits a signal at each SR simultaneously, uplink multi-user

signals will not collide. Each SISO receiver can process the

received signal as that in the conventional WLAN system. This

will effectively reduce the probability of transmission collision

See https://www.ieee.org/publications/rights/index.html for more information.

Authorized licensed use limited to: Hong Kong Baptist University. Downloaded on September 07,2024 at 13:22:23 UTC from IEEE Xplore. Restrictions apply.

LEE AND WU: WLAN UPLINK COLLISION-RESOLVING SCHEME USING MULTI-USER BEAMFORMING TECHNIQUE 11043

Fig. 1. Uniform linear array.

and enhance system throughput. This is particularly true when

the number of STAs is large. On top of that, the biggest advantage

of the proposed scheme is that the original IEEE 802.11 DCF can

be reserved, and PHY processing for each SR remains the same

as the original WLAN system. No extra operations are required

except those for BF. To the best of our knowledge, this work is the

ﬁrst of the kind exploiting MU-BF in the WLAN uplink system.

We also analyze the theoretical performance of the proposed

scheme, and the validity of the theoretical results is veriﬁed by

simulation. Finally, numerical results show that the proposed

scheme can effectively enhance the system performance.

The remainder of the article is organized as follows: Section II

reviews the basic principle of BF technology. Section III de-

scribes the proposed MU-BF scheme for WLAN uplink. With a

mathematical model, Section IV presents performance analysis

for the proposed MU-BF based WLAN system. Section V

reports simulation r esults verifying the accuracy of our analysis

and demonstrating the effectiveness of the proposed scheme.

Finally, Section VI draws conclusions.

II. I

NTRODUCTION TO BEAMFORMING

Consider a uniform linear array (ULA) shown in Fig. 1. In

the ﬁgure, φ is the AoA, and d is the antenna spacing which is

usually set as the half wavelength of the carrier.

A common operation with an array antenna is receive BF.

Receive BF is obtained by a linear combination of the signals

received from antennas. With proper weights, a spatial ﬁlter,

called beamformer, can be formed. The purpose is to enhance

the received signal strength in the desired AoA direction and

attenuate others. Fig. 2 shows the receive beamformer for a

communication system in which transmitter has one antenna

and the receiver has N antennas. Note that the BF technique can

be used at the transmitter also (i.e., transmit BF).

From Fig. 1, we see that signal arrives each antenna element

will have different delays, and the delay is a function of AoA.

The delay in turn will change the phase of the received signal

at each antenna. Without loss of generality, let the phase shift

experienced in the leftmost antenna be zero. Then, we can write

the phase response of the antenna array with a vector shown

below:



1,e

−j

2πd

sin φ

,...,e

−j

2π(N −1)d

sin φ



(1)

where λ is the wavelength of the carrier. The vector a

is referred

to as the steering vector. Let w

=[ω

,ω

,...,ω

]

Fig. 2. Receive BF in a wireless communication system.

be the BF vector consisting of the weights of the receive

beamformer, and y

(t)=[y

(t),y

(t),...,y

(t)]

x(t) be the received signal vector. Here, x(t) is the signal

from the transmitter. Then, the output of the beamformer can be

written as

y(t)=





(t). (2)

It is apparent that when w

= a

, the output can have the

maximum signal-to-noise ratio for a signal received from the

direction of φ. This is the basic principle of BF. Also, note that

multiple beams can be formed at the same time if multiple w

’s

are used.

III. P

ROPOSED MU-BF BASED WLAN SYSTEM

We consider a MU-BF based WLAN system composed of

an AP, equipped with N antennas, and n single-antenna STAs,

competing for the uplink channel. AP is assumed to be located

at the center of a basic service set, and STAs are distributed

around it. For simplicity, let N = 2

, where c is a positive integer.

In general, N represents the maximum number of orthogonal

beams that the AP can form. In other words, the service area

of the AP can thus be partitioned into N independent SRs by

BF. In addition, the channel is assumed to be ideal line-of-sight

(LoS). For an SR, at most one STA is allowed to transmit at one

time. Otherwise, the uplink transmission collision will occur.

In this article, the channel access strategy adopted by STAs is

the traditional IEEE 802.11 DCF protocol. Moreover, we assume

that the signal of a STA transmitted over the wireless channel can

be sensed by any other STA in the WLAN. Prior to transmissions,

no speciﬁc information (e.g., which SR a STA is located) is

required to be exchanged between STAs and AP.

Fig. 3 shows the proposed MU-BF based AP architecture.

First, there are N SRs, and BF is conducted for each SR. Then,

the signal received from each SR is processed by a conventional

SISO WLAN transceiver. For convenience, the receiver for SR i

is denoted as RX i where i = 1, 2,...,N. As mentioned, beam-

formed receive signal can be obtained by combining weighted

received signals from antennas. Since there are N beams to be

formed, N BF vectors are needed. From (2), N outputs can be

Authorized licensed use limited to: Hong Kong Baptist University. Downloaded on September 07,2024 at 13:22:23 UTC from IEEE Xplore. Restrictions apply.

11044 IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY, VOL. 69, NO. 10, OCTOBER 2020

Fig. 3. Receiver architecture in MU-BF based AP.

obtained as

(t)=





(t),i= 1, 2,...,N, (3)

where y

(t) is the output of the ith SISO receiver and w

[ω

,ω

,...,ω

]

is the corresponding BF vector. By

doing so, AP can concurrently receive signals from N SRs, and

each of the beamformed receive signals can be viewed as the

signal received by a conventional AP. That is, with MU-BF,

N virtual APs are operating at the same time. Note that the

same set of beamforming vectors can be used for concurrent

transmissions also. Let the transmitter for SR

be denoted as TX

i where i = 1, 2,...,N. In uplink, AP will send an acknowl-

edgement (ACK) packet back to a STA whenever a transmission

is successful. For MU-BF, there will be a beam overlapped

(BO) area between the edge of two neighboring SRs. Any two

concurrent STAs transmit in the area will cause non-negligible

mutual interference and affect uplink transmissions. This effect

is not considered in our previous work [24]. In this article, we

will speciﬁcally take this effect into account.

As noted above, most uplink distributed MU-MIMO protocols

do not consider PHY operations, such as packet detection and

symbol timing, resulting in implementation difﬁculties. For ex-

ample, in order to achieve MPR, the MU-MIMO based WLAN

system requires that all the uplink packets must be operated

in a synchronized manner. That is, all packets must arrive at

the AP at the same time. This is challenging in real-world

WLAN environments. Even this can be achieved, AP cannot

conduct synchronization for the uplink packets received from

different STAs since the preambles of the packets, designed for

synchronization, are all the same, and they are severely interfered

Fig. 4. Beam pattern of MU-BF based WLAN system (N = 4, ULA).

with each other. Therefore, multiple packets will collide and

cannot be decoded successfully. The proposed uplink scheme

can effectively avoid these problems.

As shown in (3), N beams have t o be formed, and the problem

now is how can we determine the BF vectors? A simple way is to

use orthogonal beamformers, meaning that the BF vectors are

orthogonal to each other while pointing to desired directions.

Let W =[w

, w

,...,w

] be a BF matrix. A frequently

used BF matrix is the discrete Fourier transform (DFT) matrix.

Fig. 4 shows the beam pattern of an ULA with a four-point DFT

matrix (N = 4). As we can see, there are two beams for each

SR. This is because, for the same φ, the ULA cannot distinguish

between the signals coming from its front and back.

With MU-BF, STAs in different SRs can ignore contention

and transmit signals without causing collision. However, in our

scheme, while a STA transmits, all other STAs will sense the

transmitted signal and treat the channel as busy. This is because,

for any other STA, there is no way to distinguish that a sensed

signal is transmitted from its SR or other SRs. In other words,

a transmitting STA will prohibit all other STAs from competing

for the channel. Even so, our proposed architecture provides

multiple spatial channels through BF, and forms multiple SRs.

Each SR can be treated as a collision domain, thereby signiﬁ-

cantly reducing the collision probability of simultaneous trans-

missions. For this reason, system performance will be effectively

improved. Furthermore, it is interesting to note that in some

cases, the system performance of the proposed MU-BF based

WLAN system can even outperform MU-MIMO based WLAN

system; we will elaborate this with an example below. To further

understand the characteristic of the proposed MU-BF scheme,

we use an example shown in Fig. 5 for explanation. In this ﬁgure,

the beam patterns in Fig. 4 are redrawn for ease of illustration;

two beams in each SR are merged into one. Note that in Fig. 4,

the size of each SR may not be the same, especially for SR1.

While this is not a constraint for our analysis, it will signiﬁcantly

complicate the result. For simplicity, we will treat the size of all

Authorized licensed use limited to: Hong Kong Baptist University. Downloaded on September 07,2024 at 13:22:23 UTC from IEEE Xplore. Restrictions apply.

剩余12页未读，继续阅读

评论收藏

内容反馈