ICI消除的技术资源-CSDN文库

需积分: 35 180 浏览量 2015-10-30 14:46:31 上传评论收藏 155KB PDF 举报

标题和描述所指的知识点为ICI（Inter-Carrier Interference，即载波间干扰）消除技术，并且特别关注在移动OFDM（Orthogonal Frequency Division Multiplexing，正交频分复用）系统中的应用，以及具体到DVB-H（Digital Video Broadcasting for Handhelds，便携式数字视频广播）系统上的研究。接下来，我们将详细探讨ICI消除技术及其它相关知识点。我们需要了解OFDM系统的基本概念。OFDM是一种多载波传输技术，其通过将频率宽的信道分割成许多狭窄的子载波来传输数据。OFDM能够有效利用频谱资源，并具有较强的抗多径干扰能力。然而，当载波受到频率选择性衰落的影响时，或者在高速移动环境中，子载波的正交性将被破坏，导致ICI产生。 ICI消除技术的核心目标是在高速移动的场景下，尽可能地消除由多普勒频移引起的ICI，保持子载波的正交性。文章中提到了几种已提出的ICI抑制技术，包括线性最小均方误差（MMSE）均衡器、连续干扰消除技术、基扩展模型（BEM）和时域块滤波等。这些技术通过不同的信号处理方式来补偿信道的快速变化，减少ICI对系统性能的影响。文章中特别提到的混合频域/时域的低复杂度算法，正是利用信道矩阵在频域的带状和对称结构以及在时域的稀疏结构来实现复杂度降低。这种算法能够对高动态移动环境下的OFDM系统中的ICI进行有效抑制。 DVB-H系统是一个典型的移动数字广播系统，它在高速移动的环境中工作，例如汽车和火车中。DVB-H 8K模式特别使用了大量的8192个子载波，这增加了ICI对系统性能影响的挑战。因此，为了提高接收器的可靠性，在高速移动环境中对抗ICI尤为关键。文章强调了在用户终端受限于功率和尺寸的约束条件下，开发低复杂度ICI消除方案的重要性。文章还提到了有限脉冲响应（FIR）MMSE频域均衡器（FEQ）作为ICI消除的一种方案。这种均衡器基于信道矩阵已知的带状结构，旨在以较低的复杂度实现ICI的消除。综合以上内容，我们可知ICI消除技术是无线通信领域中非常重要的一个研究方向，特别是在高速移动通信系统中，如何减少或消除ICI已经成为提升系统性能的关键所在。ICI消除技术的研究不仅包括算法的开发，还包括信号处理、信道估计、滤波器设计等多个层面。未来，随着高速移动通信的快速发展，ICI消除技术将会进一步得到优化和创新，以适应各种复杂多变的无线通信环境。

资源推荐

资源详情

资源评论

IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS, VOL. 7, NO. 7, JULY 2008 1

Coherent and Differential ICI Cancellation for Mobile OFDM with

Application to DVB-H

Sili Lu, Student Member, IEEE and Naofal Al-Dhahir, Fellow, IEEE

Abstract—We develop a reduced-complexity hybrid

frequency/time-domain orthogonal frequency division

multiplexing (OFDM) channel estimation algorithm for

high-mobility scenarios where the channel varies signiﬁcantly

within each OFDM block, resulting in severe intercarrier

interference (ICI). The algorithm exploits the banded and

symmetric structure of the channel matrix in the frequency-

domain and its sparse structure in the time-domain to achieve

signiﬁcant complexity reductions, which we quantify for

the Digital Video Broadcasting-Handheld (DVB-H) system.

Furthermore, we compare coherent and differential mobile

OFDM detection for DVB-H.

Index Terms—ICI mitigation, OFDM, channel estimation,

doppler shift

I. INTRODUCTION

UPPORTING high mobility is a key requirement for

a number of OFDM-based broadband wireless systems,

such as DVB-H[1]. For high-mobility systems, the fast channel

variations within one OFDM block destroy the orthogonal-

ity between subcarriers resulting in ICI proportional to the

Doppler frequency. This problem is even more severe for the

DVB-H 8K mode (which uses a large number o f 8192 densely-

spaced subcarriers) and becomes a signiﬁcant challenge for

receiver reliability at highway and train speeds.

To mitigate ICI, several techniques have been proposed in

the literature including linear minimum mean sq uare error

(MMSE) equalizers [2] [3], successive interference cancella-

tion [4], basis expansion models (BEM) [5] and time-domain

block ﬁltering [6]. Our objective in this paper is to investigate

low-complexity schemes that can suppress ICI effectively

under high-mobility conditions. As a case study, we focus

on the challenging DVB-H application where the number of

subcarriers is very large (2048, 4096, or 8192) [7], hence

it is critical to develop a low-comp lexity ICI cancellation

solution due to the power/size constraints at the user terminal.

One such scheme is the ﬁnite impluse response (FIR) MMSE

frequency-domain equalizer (FEQ) in [8] which is based on

the well-known banded structure of the channel matrix in

the frequency-domain [9][10], and has only few taps per

subcarrier. Nevertheless, all of these approaches are coherent

detection schemes that require reliable channe l estimation at

the receiver wh ich is a cha llenging task at high mobility.

Channel estimation for OFDM can be performed in the

frequency domain or time domain. Conventional frequency-

domain channel estimation algorithms ignore ICI which makes

Manuscript received June 4, 2007; revised September 17, 2007, November

8 2007 and November 25, 2007; accepted December 3, 2007. The associate

editor coordinating the review of this letter and approving it for publication

was Dr. R. Mallik.

This work was supported in part by Semiconductor Research Corporation

under contract No. 2005-HJ-1328 and by a gift from Texas Instruments Inc.

The authors are with University of Texas at Dallas, (e-mail: {sxl059000,

aldhahir}@utdallas.edu)

Digital Object Identiﬁer 10.1109/TWC.2008.070591.

them highly suboptimal under high Doppler. Time-domain

channel estimation algorithms (such as [6]) typically mark

a few rows of the time-domain channel matrix and estimate

them using pilot tones embedded within each OFDM block.

Since the multipath delay spread can be very high for single

frequency network (SFN) DVB systems [1] [7], the number

of unknown channel taps for each marker row can be too

high to estimate with the limited number of available pilots.

Considering a fast-varying channel, Mostoﬁ et al. proposed

in [11] a hybrid frequency/time-domain channel estimation

algorithm based on a linear approximation of the channel time

variations within one OFDM block.

As a low-complexity alternative transceiver structure, we

also investigate and compare the performance of differential

OFDM detection with coherent OFDM detection for the DVB-

H system. Note that differential detection is already used in the

digital audio broadcasting (DAB) standard. Since differential

detection does not require channel estimation, pilot overhead

can be reduced to increase data throughput and channel esti-

mation co mplexity can be eliminated from the user terminal.

Our main contributions in this paper can be summarized as

follows

• We show how to signiﬁcantly reduce the complexity

of the hybrid frequency/time-domain channel estimation

algorithm in [11] by exploiting the channel’s sparse and

banded structure. We quantify these complexity savings

through a detailed complexity analysis for the DVB-H

system.

• We derive a new asymptotic relationship between the sub-

and super-diagonals of the frequency-domain channel

matrix and exploit it to further reduce the complexity

of channel estimation and to design an ICI-mitigating

pilot/data placement scheme.

• We investigate differential OFDM detection in the pres-

ence of ICI and compare its performance with coherent

detection for the DVB-H system.

Notation: We use (·)

to denote the transpose, (·)

∗

the

complex-conjugate, (·)

the complex-conjugate transpose, ·

the ﬂoor operation and (·)

the modulo-N operation. G

i,j

denotes the element in the i-th row and j-th column of matrix

G, where row/column indices begin with zero. The estimated

value of random variable a is denoted by ˆa and I

denotes

the N × N identity matrix. We use the notation of G(:,j) to

denote the j-th column of G.

II. M

OBILE OFDM SYSTEM MODEL

We consider an OFDM system with N subcarriers where

each OFDM block, denoted by X =[X

...X

N−1

]

,is

converted into time-domain samples x =[x

...x

N−1

]

using

the N -point Inverse Fast Fourier Transform (IFFT) operation

x = F

X where F

is the N-point IFFT matrix. We assume

1536-1276/08$25.00

 2008 IEEE

2 IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS, VOL. 7, NO. 7, JULY 2008

that the cyclic preﬁx (CP) length is equal to or larger than the

channel impulse response (CIR) memory denoted by L. Then,

the received b lock y =[y

...y

N−1

]

after CP removal is

given by

y = Hx + v (1)

where H is an N × N time-domain channel matrix with

elements H

n,l

= h

n,(n−l)

where h

n,l

is the CIR at lag l

for 0 ≤ l ≤ L − 1 and time instant n for 0 ≤ n ≤ N − 1,

and v is the time-d omain noise vector with auto-correlation

matrix σ

. Taking the FFT of (1), we obtain

Y = Fy = FHF

X + Fv = GX + V (2)

where G  FHF

is the frequency-domain channel matrix

and V is the frequency-domain noise vector. For a quasi-static

fading channel, H is a circulant matr ix and hence transform s

into a diagonal G matrix in (2). In this case, the OFDM

subcarriers are decoupled, and hence a one-tap FEQ is optimal.

For example, with a linear Zero-forcing (ZF) FEQ, the data

estimate for the k-th subcarrier

is simply

= G

−1

k,k

For the time-varying channel, however, H is not circulant and

hence G is no longer diagonal. In this case, the input-output

relation for the k-th subcarrier is given by

= G

k,k

N−1



n=0,n=k

k,n

+ V

(3)

The ﬁrst term on the right-hand side of (3) is the desired

signal term while the second term is the ICI term which can be

characterized by the normalized Doppler frequency F

= f

where f

is the Doppler frequency and T is the time duration

of the useful part of one OFDM block.

III. C

OHERENT ICI CANCELLATION

A. ICI Analysis

To better understand the structure of G, we expand the

relation G = FHF

in (2) and get

k,n

L−1



m=0

N−1



r=0

r,m

j2πr(n−k)

−j2πnm

(4)

For a quasi-static channel where h

r,m

= h

s,m

for r =

s, 0 ≤ m ≤ L − 1, and 0 ≤ r, s ≤ N − 1,wehave

k,n





L−1

m=0

avg

−j2πmk

: k = n

0:k = n

(5)

where h

avg



N−1

r=0

r,m

denotes the time-average of the

m-th channel tap and in this case h

avg

= h

r,m

, ∀ 0 ≤ r ≤

N − 1. It is clear from (5) that G is a diagonal matrix with

elements equal to the channel frequency response.

Turning our attention to fast-varying channels, the CIR is

no longer ﬁxed within one OFDM block. In [3] and [11],

a linear model for the time-domain channel variations was

proposed. It is shown in [11] that for Doppler rates normalized

to the subcarrier spacing of up to 20%, the channel time-

variations can be approximated by a piece-wise linear model

with a constant slope over each OFDM block. Let α

denote

the slope of the m-th CIR tap, then the m-th CIR tap at instant

r can be expressed as [11]

r,m

= h

avg

+Δh

r,m

;Δh

r,m

=(r −

N − 1

)α

(6)

where Δh

r,m

denotes the time-variation term at the m-th tap

for 0 ≤ r ≤ N −1, 0 ≤ m ≤ L−1. Here, we used the fact that

the average of each channel tap can be approximated b y the

middle point of the channel tap vector, i.e. h

avg

≈ h

N/2−1,m

as proved in [11]. Combining (4) and (6), we arrive at (7)

shown on top of next page.

Therefore, the time-average term h

avg

corresponds only to

the diagonal elements of G while the time-variation term

Δh

r,m

contributes to the off-diagonal elements giving rise to

ICI.

It is a well-known fact that G has most of its energy

concentrated along its main diagonal and hence can be ap-

proximated as a banded matrix (see e.g. [9][10] and the

references therein), which has non-zero elements only along

(2D +1)main diagonals and two D × D triangular matrices

in the top-right and bottom-left corners. This banded structure

implies that most of the ICI term energy in (3) results from

the adjacent subcarriers. Based on this banded structure, we

proposed in [8] an FIR FEQ designed using the MMSE

criterion (hence the name FIR-MMSE FEQ) which is based

only on the following (2D +1)× (4D +1)submatrix of G.

⎡

⎢

⎣

(m−D)

,(m−2D)

... G

(m−D)

,(m+2D)

m,(m−2D)

... G

m,(m+2D)

(m+D)

,(m−2D)

... G

(m+D)

,(m+2D)

⎤

⎥

⎦

(8)

The Q-tap (where Q =2D +1) FIR-MMSE FEQ is given

by [8]

= g

+ σ

)

−1

(9)

where g

= G

(:, 2D +1) is the middle column of G

Hence, the Q-tap FEQ output for the m-th subcarrier is

where Y

=[Y

(m−D)

...Y

(m+D)

]

To compute the coefﬁcients for this FIR-MMSE FEQ, we

need to estimate the Q (typically Q =2D +1=3) diagonals

of G. The hybrid frequency/time-domain channel estimation

algorithm of the next subsection provides a computationally-

efﬁcient solution for this problem.

Starting from (7), we show in the Appendix that for large

N, the elements of the sub-diagonals of G are approximately

the negatives of the corresponding elements of the super-

diagonals, i.e. G

i+1,i

≈−G

i−1,i

. We will utilize this property

to simplify the channel estimation algorithm in Subsection

III-B and develop an efﬁcient ICI-mitigating pilot/data place-

ment scheme in Subsection III-C.

B. Reduced-Complexity Hybrid Channel Estimation Algo-

rithm

To estimate the Q main diagonals of G, we revisit

the hybrid frequency/time-domain channel estimation algo-

rithm in [11]. Let h

avg

=[h

avg

...h

avg

L−1

] and α =

剩余6页未读，继续阅读

评论收藏

内容反馈

dyq8888

粉丝: 0
资源: 1

ICI消除的技术

回声消除技术

快速时变衰落信道上的新ICI自消除映射方案

OFDM系统中ICI自消除算法研究

基于SLM和PTS的子载波间干扰消除技术

水体中高氯酸盐消除技术研究 (2010年)

论文研究-引入加权系数消除ICI的二次快时变信道估计算法.pdf

消除PCB沉银层的技术和方法

ofdm-simulink.rar_LTE 无线干扰_LTE干扰消除_ofdm simulink _ofdm 跳频_扩频 抗干扰

基于干扰消除的OFDM / OQAM系统先导设计

比奈奎斯特信令快的干扰消除技术

回声消除技术_一个开源程序

视频人声消除技术.docx

VoIP技术之回声消除技术

高移动环境下利用无线信号空域特征实现OFDM系统的ICI消除

一种新的基于导频簇方案的时变OFDM系统双重ICI消除算法

PCB技术中的消除PCB沉银层的技术和方法

OFDM技术抗ISI和ICI的性能分析

【教练技术】消除身边的“流行病”.docx

ofdm系统的载波频偏干扰消除程序

无循环前缀OFDM无线传输的关键技术_刘喜庆.caj

CDMA 协作检测信号检测内容。很全的。论文

OFDM移动通信技术原理与应用

matlab-ofdm-.rar_ICI_LTE 无线干扰_OFDM抗干扰_ofdm 跳频_跳频OFDM

数字电视解调芯片关键技术分析.pdf

ICI reduction for OFDM

MIMO-OFDM无线通信技术及MATLAB实现pdf（包括matlab代码）无缺少的

LASIP_Image_Restoration_DemoBox_v120_restoration_lpa-ici_局部多项式_自

最新资源

ofdm-simulink.rar_LTE 无线干扰_LTE干扰消除_ofdm simulink _ofdm 跳频_扩频抗干扰