ITERATIVE ESTIMATOR AND METHOD OF CHANNEL AND NOISE VARIANCE FOR MULTIPLE INPUT MULTIPLE OUTPUT ORTHOGONAL FREQUENCY DIVISION MULTIPLEXING SYSTEM

Abstract

A reception apparatus including: a detection unit detecting extrinsic information based on an output of a tentative symbol decision unit, an output of a channel estimation unit, and an output of a noise variance estimation unit that are obtained from a previous iteration process; a Cyclic Redundancy Check (CRC) aided channel decoding unit outputting an encoded and interleaved bit, or a posteriori information thereof based on the extrinsic information; the tentative symbol decision unit determining a tentative transmission symbol based on an output of the CRC aided channel decoding unit; the channel estimation unit estimating a channel based on an output of the tentative symbol decision unit or a pilot symbol; and the noise variance estimation unit estimating a noise variance or an initial noise variance based on the output of the tentative symbol decision unit and the output of the channel estimation unit is provided.

Description

TECHNICAL FIELD

The present invention relates to an iterative transmission/reception apparatus and method for improving the performance associated with channel estimation and noise variance estimation and reducing the complexity in a Multiple-Input Multiple-Output (MIMO) orthogonal frequency division multiplexing (OFDM) system.

This work was supported by the IT R&D program of MIC/IITA. [2006-S-001-02, Development of Adaptive Radio Access and Transmission Technologies for 4th Generation Mobile Communications]

BACKGROUND ART

Generally, compared to a scheme of directly estimating a frequency domain channel in a received signal, a Multiple-Input Multiple-Output (MIMO) orthogonal frequency division multiplexing (OFDM) system initially estimates a time domain channel and then finally estimates a frequency domain channel using the estimated time domain channel. The MIMO OFDM system can reduce the noise affect and also can effectively perform interpolation between pilot subcarriers to improve the channel estimation performance. Channel estimation schemes can be generally classified into a disjoint channel estimation scheme and joint channel estimation scheme.

The disjoint channel estimation scheme considers, as interference, other signals except for its own signal to estimate a channel. For example, on the assumption that a number of multi-paths of a time domain channel is P, when the conventional disjoint channel estimation is adopted, P×P matrix inversion is required for each OFDM symbol.

On the assumption that a number of transmit antennas is N_t, the joint channel estimation scheme simultaneously estimates channels corresponding to all the transmit antennas. Therefore,

N_tP×N_tP

matrix inversion is required for each OFDM symbol. Generally, since P>50, it is very difficult to embody the joint channel estimation scheme. In the conventional art, proposed was a scheme of canceling an interference component, caused by another antenna, using interference estimation obtained from an iterative reception process to improve the channel estimation performance. However, since the conventional scheme does not adopt a time domain channel estimation process, the performance may be significantly deteriorated when frequency selectivity or time selectivity is large.

When mixing the conventional schemes, the performance of mixed scheme can be similar to the performance of the joint channel estimation scheme and the complexity thereof may be similar to the complexity of the disjoint channel estimation scheme. However, since the mixed scheme also requires P×P matrix inversion for each OFDM symbol, there still remains a problem that the complexity thereof is very high.

In the case of an initial channel estimation scheme of estimating a channel using a pilot symbol, a channel estimation value in a previous iteration for interference cancellation does not exist and thus the interference cancellation becomes impossible. Specifically, the only solution is to apply the conventional joint channel estimation scheme.

In the case of a conventional noise variance estimation scheme, when per-subcarrier estimation is performed, the performance can be deteriorated due to a noise component.

DISCLOSURE OF INVENTION

Technical Problem

An aspect of the present invention provides a transmission/reception apparatus and method that can adopt a channel estimation scheme with similar performance to a joint channel estimation scheme and with lower complexity due to not requiring a matrix inversion, in a Multiple-Input Multiple-Output (MIMO) orthogonal frequency division multiplexing (OFDM) system.

Another aspect of the present invention also provides a transmission/reception apparatus and method that can perform soft interference cancellation based on a value estimated in a previous OFDM symbol, in an initial channel estimation stage and thereby can improve the channel estimation performance and reduce the complexity.

Another aspect of the present invention also provides a transmission/reception apparatus and method that can improve the performance of interference noise variance estimation based on all the information associated with previous effective subcarriers when estimating noise variance in a MIMO OFDM system.

Technical Solution

According to an aspect of the present invention, there is provided a transmission apparatus including: an encoding unit encoding a transmission packet; an interleaving unit interleaving the encoded transmission packet; a mapping unit mapping the interleaved transmission packet; a pilot insertion unit inserting a pilot symbol and an information symbol into a subcarrier of the mapped transmission packet; and a transmit antenna transmitting the transmission packet with the inserted pilot symbol and information symbol.

According to another aspect of the present invention, there is provided a reception apparatus including: a detection unit detecting extrinsic information based on an output of a tentative symbol decision unit, an output of a channel estimation unit, and an output of a noise variance estimation unit that are obtained from a previous iteration process; a Cyclic Redundancy Check (CRC) aided channel decoding unit outputting an encoded and interleaved bit, or a posteriori information thereof based on the extrinsic information; the tentative symbol decision unit determining a tentative transmission symbol based on an output of the CRC aided channel decoding unit; the channel estimation unit estimating a channel based on an output of the tentative symbol decision unit; and the noise variance estimation unit estimating a noise variance or an initial noise variance based on the output of the tentative symbol decision unit and the output of the channel estimation unit.

According to still another aspect of the present invention, there is provided a transmission method including: encoding a transmission packet; interleaving the encoded transmission packet; mapping the interleaved transmission packet; inserting a pilot symbol and an information symbol into a subcarrier of the mapped transmission packet; and transmitting the transmission packet with the inserted pilot symbol and information symbol.

According to yet another aspect of the present invention, there is provided a reception method including: detecting extrinsic information based on a tentative symbol decision, a channel estimation, and a noise variance estimation that are obtained from a previous iteration process; outputting an encoded and interleaved bit or a posteriori information thereof based on the extrinsic information; determining a tentative transmission symbol based on the outputting; estimating a channel based on a tentative symbol decision; and estimating a noise variance or an initial noise variance based on the tentative symbol decision and the channel estimation.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a block diagram illustrating a configuration of a transmission apparatus in a Multiple-Input Multiple-Output (MIMO) orthogonal frequency division multiplexing (OFDM) system according to an embodiment of the present invention;

FIG. 2 is a block diagram illustrating a configuration of a reception apparatus in a MIMO OFDM system according to an embodiment of the present invention;

FIG. 3 is a block diagram illustrating a configuration of a Cyclic Redundancy Check (CRC) aided channel decoding unit shown in FIG. 2;

FIG. 4 is a block diagram illustrating a configuration of a tentative symbol decision unit shown in FIG. 2;

FIG. 5 is a block diagram illustrating a configuration of a channel estimation unit shown in FIG. 2;

FIG. 6 is a block diagram illustrating a configuration of a channel estimation unit using a pilot symbol;

FIG. 7 is a block diagram illustrating a configuration of a noise variance estimation unit shown in FIG. 2; and

FIG. 8 is a block diagram illustrating a configuration of an initial noise variance estimation unit.

MODE FOR THE INVENTION

Reference will now be made in detail to embodiments of the present invention, examples of which are illustrated in the accompanying drawings, wherein like reference numerals refer to the like elements throughout. The embodiments are described below in order to explain the present invention by referring to the figures.

Hereinafter, an iterative transmission/reception apparatus and method for a Multiple-Input Multiple-Output (MIMO) orthogonal frequency division multiplexing (OFDM) system will be described in detail with reference to the accompanying drawings. For example, the present invention will be described based on the MIMO OFDM system with N_t transmit antennas and N_r receive antennas. Here, a total number of subcarriers is N in which N_c subcarriers are used to substantially transmit data and N-N_c subcarriers are not used. Among the N_c effective subcarriers, N_p subcarriers are allocated to a pilot symbol and remaining subcarriers are allocated to an information symbol. N_t pilot symbols transmitted from all the transmit antennas are overlappingly transmitted to a single pilot subcarrier.

FIG. 1 is a block diagram illustrating an iterative transmission apparatus 100 for a MIMO OFDM system according to an embodiment of the present invention.

Referring to FIG. 1, the iterative transmission apparatus 100 for the MIMO OFDM system includes encoding units 110-1, . . . , 110-n, interleaving units 120-1, . . . , 120-n, Gray mapping units 130-1, . . . , 130-n, pilot insertion units 140-1, . . . , 140-n, and transmit antennas 150-1, . . . , 150-n. Specifically, the transmission apparatus 100 is in the horizontal encoding structure and transmits a different bit stream or packet for each of the transmit antennas 150-1, . . . , 150-n.

The encoding units 110-1, . . . , 110-n may receive and encode information bit streams b to output encoded bit streams c, respectively. Also, the encoding units 110-1, . . . , 110-n may receive and encode a transmission packet to output encoded transmission packets. Specifically, the information bit streams b may be converted to the encoded bit streams c via the encoding units 110-1, . . . , 110-n, respectively. The transmission packets may be converted to the encoded transmission packets via the encoding units 110-1, . . . , 110-n, respectively.

The independency between coded bits should be secured for appropriate operations of the transmission apparatus 100. For this, the interleaving units 120-1, . . . , 120-n may receive and interleave the encoded bit streams c to output interleaved bit streams x, respectively. Specifically, the encoded bit streams c may be converted to the interleaved bit streams x via the interleaving units 120-1, . . . , 120-n, respectively. The interleaving units 120-1, . . . , 120-n may receive and interleave the encoded transmission packets to output interleaved transmission packets.

The Gray mapping units 130-1, . . . , 130-n may receive and perform Gray mapping for the interleaved bit streams x to output information symbols s, respectively. Specifically, an Mc number of interleaved bit streams x may be converted to the information symbols s via the Gray mapping units 130-1, . . . , 130-n, respectively. The Gray mapping units 130-1, . . . , 130-n may receive and perform Gray mapping for the interleaved transmission packets to output mapped transmission packets.

The pilot insertion units 140-1, . . . , 140-n may insert N_p pilot symbols into subcarriers

p₁, p₂, . . . , p_N,

and insert

N_c−N_p

information symbols into the subcarriers with the inserted pilot symbols to generate a packet consisting of N_c symbols. For example, subscript[k] in b[k] denotes a bit stream to be transmitted via a k^th transmit antenna 150-k.

S_n,j

is defined as a symbol that is transmitted to a j^th transmit antenna 150-j via an n^th subcarrier.

y_n,i

is defined as a signal that is received at an i^th receive antenna via the n^th subcarrier. In this case, in the MIMO OFDM system with N_c effective subcarriers, received signal vector

y_n=[y_n,1 y_n,2 . . . y_n,Nτ]^T

of the n^th subcarrier may be expressed as,

y _n =H _n s _n +v _n , n=1, 2, . . . , N_c,  [Equation 1]

where

s_n=[s_n,1 s_n,2 . . . s_n,Ni]^T

denotes a transmission symbol vector,

v_n=[v_n,1 v_n,2 . . . v_n,Nr]^T

denotes a noise signal vector,

CN(0,R_n,vv)

denotes a Gaussian probability density function, and

R_n,vv=E{v_n^Hv_n}=s_n²I_Nr.

A MIMO channel of a frequency domain of the n^th subcarrier may be defined as given by,

$\begin{array}{cc}{H}_n=\left[\begin{array}{cccc}{h}_{n,11}& h_{n,12}& \dots & h_{n,1N_t}\\ h_{n,21}& h_{n,22}& \dots & h_{n,2N_t}\\ ⋮& ⋮& \ddots & ⋮\\ h_{n,N_r1}& h_{n,N_r2}& \dots & h_{n,N_rN_t}\end{array}\right],& \left[\mathrm{Equation}2\right]\end{array}$

where

h_n,ij

denotes a channel coefficient of the frequency domain of the n^th subcarrier between the j^th transmit antenna and the i^th receive antenna.

FIG. 2 is a block diagram illustrating a configuration of a reception apparatus 200 in a MIMO OFDM system according to an embodiment of the present invention.

Referring to FIG. 2, the reception apparatus 200 includes a minimum mean squared error with soft cancellation (MMSE-SC) detection unit 210, a Cyclic Redundancy Check (CRC) aided channel decoding unit 220, a tentative symbol decision unit 230, a channel estimation unit 240, and a noise variance estimation unit 250.

The MMSE-SC detection unit 210 may output bit extrinsic LLR

L_Ed^(q+1)

of a per-antenna interleaved bit stream to the CRC aided channel decoding unit 220.

Hereinafter, a configuration and operation of the CRC aided channel decoding unit 220 will be further described in detail with reference to FIG. 3.

FIG. 3 is a block diagram illustrating a configuration of the CRC aided channel decoding unit 220 shown in FIG. 2.

Referring to FIG. 3, the CRC aided channel decoding unit 220 includes a deinterleaving units 310-1, . . . , 310-n, channel decoding units 320-1, . . . , 320-n, CRC checking units 330-1, . . . , 330-n, encoding units 340-1, . . . , 340-n, and interleaving units 350-1, . . . , 350-n.

The CRC aided channel decoding unit 220 may receive bit extrinsic LLR

L_Ed^(q+1)

from the MMSE-SC detection unit 210 to generate a priori LLR

L_Ac^(q+1)

of the channel decoding units 320-1, . . . , 320-n via the deinterleaving units 310-1, . . . , 310-n.

The channel decoding units 320-1, . . . , 320-n may calculate transmission information bit extrinsic LLR

L_Ec^(q+1)

using the a priori LLR

L_Ac^(q+1)

to output the calculated transmission information bit extrinsic LLR

L_Ec^(q+1)

to the CRC checking units 330-1, . . . , 330-n.

The CRC checking units 330-1, . . . , 330-n may perform CRC check for the transmission information bit extrinsic LLR

L_Ec^(q+1)

to determine whether an error has occurred while receiving a packet. Specifically, when it is determined no error has occurred while receiving the packet, the CRC checking units 330-1, . . . , 330-n may set Rx_check to “1” and close an upper switch to thereby encode corresponding received information bits via the encoding units 340-1, . . . , 340-n and output the encoded corresponding received information bits to the interleaving units 350-1, . . . , 350-n, respectively. For example, when it is determined an error has occurred, the CRC checking units 330-1, . . . , 330-t may set Rx_check to “0” and close a lower switch to generate a posteriori LLR

L_Pc^(q+1),

output from the channel decoding units 320-1, . . . , 320-n, as a priori LLR

L_Ad^(q+1)

of the tentative symbol decision unit 230 via the interleaving units 350-1, . . . , 350-n. The upper switch may be switched so that an output of the CRC checking units 330-1, . . . , 330-n may be transferred to the encoding units 340-1, . . . , 340-n according to the reception error check signal Rx_check. The lower switch may be switched so that an output of the channel decoding units 320-1, . . . , 320-n may be transferred to the interleaving units 350-1, . . . , 350-n according to the reception error check signal Rx_check.

The channel decoding units 320-1, . . . , 320-n may output a posteriori LLR

L_Pc^(q+1)

of a coded bit, which becomes a priori LLR

L_Ad^(q+1)

of the tentative symbol decision unit 230 via the interleaving units 350-1, . . . , 350-n.

The tentative symbol decision unit 230 may determine tentative symbol

S ^(q+1)

based on an output of the CRC aided channel decoding unit 220. Specifically, the tentative symbol decision unit 230 may tentatively determine the transmission symbol based on the output of the CRC aided channel decoding unit 220. Hereinafter, a configuration and operation of the tentative symbol decision unit 230 will be described in detail with reference to FIG. 4.

FIG. 4 is a block diagram illustrating a configuration of the tentative symbol decision unit 230 shown in FIG. 2.

Referring to FIG. 4, the tentative symbol decision unit 230 includes a decision selection unit 410, a pilot replacement unit 420, a hard decision unit 430, and a soft decision unit 440.

The decision selection unit 410 may select a tentative decision scheme based on Rx_check and a subcarrier index n. For example, the decision selection unit 410 may select any one decision scheme from three tentative symbol decision schemes based on Rx_check and the subcarrier index n.

When a transmission symbol is determined as a pilot symbol as a result of checking the subcarrier index n, the pilot replacement unit 420 may replace tentative symbol decision

S _n,k ^(q+1)

of

S_n,k

with pilot symbol

S_n,k^p.

Specifically,

S _n,k ^(q+1)=S_n,k^p.

A variance calculation unit 421 may calculate variance of the tentative symbol decision

S _n,k ^(q+1).

The variance of the tentative symbol decision

S _n,k ^(q+1)

is

$\mathrm{var}\left\{s_{n,k}^{\left(q+1\right)}\right\}=0.$

Also, when the transmission symbol is determined as an information symbol as the result of checking the subcarrier index n and Rx_check is “1”, the hard decision unit 430 may perform Gray mapping for an interleaved bit, received without causing any error, via a Gray mapping unit 431 to generate tentative symbol decision

S ^(q+1).

Specifically,

S ^(q+1)=S.

Also, the hard decision unit 430 calculates variance of the errorlessly received interleaved bit via a variance calculation unit 432. The variance of

S _n,k ^(q+1)

is

$\mathrm{var}\left\{s_{n,k}^{\left(q+1\right)}\right\}=0.$

Also, when the transmission symbol is determined as the information symbol as the result of checking the subcarrier index n and Rx_check is “0”, the soft decision unit 440 may calculate a bit probability and a symbol probability via a priori probability calculation unit 441. The bit probability may be calculated using a posteriori LLR

L_Pc^(q+1)(x)

of channel decoding, as given by,

$\begin{array}{cc}{p}_{A_d}^{\left(q+1\right)}\left(x_{n,k,m}\right)=\frac{{}^{\pm L_{p_c}^{\left(q+1\right)}\left(x_{n,k,m}\right)}}{1+{}^{\pm L_{p_c}^{\left(q+1\right)}\left(x_{n,k,m}\right)}}.& \left[\mathrm{Equation}3\right]\end{array}$

The symbol probability may be calculated as given by,

$\begin{array}{cc}{p}_{A_d}^{\left(q+1\right)}\left(s_{n,k}\right)=\prod _{m=1}^{M_c}p_{A_d}^{\left(q+1\right)}\left(x_{n,k,m}\right).& \left[\mathrm{Equation}4\right]\end{array}$

A soft decision calculation unit 742 may determine

S ^(q+1)

and perform soft decision, as given by,

$\begin{array}{cc}{\stackrel{\_}{s}}_{n,k}^{\left(q+1\right)}\left(n\right)=\left\{s_{n,k}^{\left(q+1\right)}\right\}=\sum _{s_k\in }s_{n,k}p_{A_d}^{\left(q+1\right)}\left(s_{n,k}\right),& \left[\mathrm{Equation}5\right]\end{array}$

where

S_n,j

is a total number of

2^Mo

sets of

S_n,j.

A variance calculation unit 443 may calculate symbol variance as given by,

$\begin{array}{cc}\mathrm{var}\left\{S_{n,k}^{\left(q+1\right)}\right\}=\sum _{s_{n,k}\in S}{S_{n,k}^{\left(q+1\right)}}^2p_{A_d}^{\left(q+1\right)}\left(S_{n,k}\right)-{\left({\stackrel{\_}{S}}_{n,k}^{\left(q+1\right)}\right)}^2& \left[\mathrm{Equation}6\right]\end{array}$

FIG. 5 is a block diagram illustrating a configuration of the channel estimation unit 240 shown in FIG. 2.

Referring to FIG. 5, the channel estimation unit 240 includes soft interference cancellation units 510-1, . . . , 510-n, initial frequency domain channel estimation units 520-1, . . . , 520-n, time domain channel estimation units 530-1, . . . , 530-n, and final frequency domain channel estimation units 540-1, . . . , 540-n.

When a received signal vector for channel estimation is defined as

y_(i)=[y_1,i y_2,i . . . y_No,i]^T,

it may be represented as,

$\begin{array}{cc}{y}_{\left(i\right)}=S_kh_{\mathrm{ik}}+\sum _{j=1,j\ne k}^{N_t}S_jh_{\mathrm{ij}}+v_{\left(i\right)}& \left[\mathrm{Equation}7\right]\end{array}$

where

h_ij=[h_1,ij h_2,ij . . . h_No,ij]^T

S _k=diag(s_1,k, s_2,k, . . . , s_No,k)

v_(i)=[v_1,i v_2,i . . . v_No,i]^T.

When

g_p,ij

is defined as a channel impulse response (CIR) of the time domain with respect to a p^th multi-path channel between a j^th transmit antenna and a j^th receive antenna, the channel relationship between the frequency domain and the time domain may be represented as,

$\begin{array}{cc}{h}_{n,\mathrm{ij}}=\sum _{p=0}^{P-1}g_{p,\mathrm{ij}}{}^{-\mathrm{j2\pi }\mathrm{pn}/N_o}& \left[\mathrm{Equation}8\right]\end{array}$

where

P

denotes a number of multi-paths. When it is defined that

g_kj=[g_1,kj g_2,kj . . . g_P,kj]^T,

the relationship shown in the following Equation 9 may be established,

h_kj=F_cg_kj,  [Equation 9]

where

F_cε

denotes a discrete Fourier transform (DFT) matrix for the CIR and

[F_c]_n,p=e^{−j2π(n−1)(p−1)/N}.

Generally, columns of

F_c

are not orthogonal to each other and are given by,

F_c^†F_c=I_p,  [Equation 10]

where

F_c⁺

denotes a pseudo-inverse matrix of

F_c.

Equation 9 and Equation 10 may be induced to make the relationship as given by,

g_kj=F_c^†h_kj.  [Equation 11]

Accordingly, Equation 7 may be represented as,

$\begin{array}{cc}{y}_{\left(i\right)}=S_kF_c^{+}g_{\mathrm{ik}}+\sum _{j=1,j\ne k}^{N_t}S_jF_c^{+}g_{\mathrm{ij}}+v_{\left(i\right)}& \left[\mathrm{Equation}12\right]\end{array}$

A channel estimation scheme may be generally classified into a disjoint channel estimation scheme and a joint channel estimation scheme.

In a second item of Equation 12, the disjoint channel estimation scheme considers, as interference, signals transmitted to other transmit antennas, and estimates time domain channel

g_ik

and then finally estimates frequency domain

h_kj

using Equation 9.

The joint channel estimation scheme simultaneously estimates

g_ij,j=1,2,L,N_t

of all the antennas. Therefore,

N_tP×N_tP

matrix inversion is required for each OFDM symbol.

Since an iterative receiver uses not only a pilot symbol for channel estimation but also a tentative decision value of an information symbol obtained from a previous iteration process, it is possible to improve the channel estimation performance. A representative example of iterative channel estimation is expectation and maximization (EM) channel estimation.

Generally, when a disjoint EM channel estimation scheme is used, it may be expressed as,

ĥ _ik ^(q+1) =F _c F _c ⁺ [F _c ⁺(S_k^(q+1))^HS_k^(q+1)F_c+R_vv^(q)(R_hh^(q+1))⁻¹]⁻¹(S_k^(q+1))^Hy_(i),  [Equation 13]

where

S _k ^(q+1)

denotes an estimate of

S_k

obtained by the tentative symbol decision unit 230 for each current iteration process,

R_hh^(q+1)

denotes an estimate of

R_hh=E{h_ik^Hh_ik}

obtained by the channel estimation unit 240 for a previous iteration process, and

R_vv^(q)

denotes an estimate of

R_vv=σ_n²I_Nc

obtained by the noise variance estimation unit 250 for the previous iteration process.

Since the disjoint EM channel estimation scheme considers, as interference, signals transmitted to other antennas, the channel estimation performance may be deteriorated. Therefore, the present invention proposes a method that can perform soft cancellation for an interference component caused by another antenna, based on interference estimation, to thereby improve the channel estimation performance. When mixing the disjoint channel estimation scheme and the joint channel estimation scheme, the channel estimation as shown in the following Equation 14 may be enabled,

ĥ _ik ^(q+1) =F _c F _c ⁺ [F _c ⁺(S_k^(q+1))^HS_k^(q+1)F_c+R_vv^(q)(R_hh^(q+1))⁻¹]⁻¹(S_k^(q+1))^Hy_(i)^(q+1).  [Equation 14]

The difference between Equation 13 and Equation 14 lies in that the received signal vector

y_(i)

is changed to

y _(i) ^(q+1)

as given by,

$\begin{array}{cc}{\stackrel{\_}{y}}_{\left(i\right)}^{\left(q+1\right)}=y_{\left(i\right)}-\sum _{j=1,j\ne k}^{N_t}{\stackrel{\_}{S}}_j^{\left(q+1\right)}{\hat{h}}_{\mathrm{ij}}^{\left(q\right)}& \left[\mathrm{Equation}15\right]\end{array}$

In comparison to the joint channel estimation scheme, the scheme as shown in Equation 14 initially performs soft interference cancellation and then performs distribution estimation. Therefore, it is possible to reduce a computational amount while maintaining the joint estimation performance.

However, since the above scheme shown in Equation 14 requires P×P matrix inversion for each OFDM symbol, the complexity may be very large.

In order to reduce the computational amount, the present invention proposes a channel estimation scheme as given by,

ĥ _ik ^(q+1) =F _c F _c ⁺(S_k^(q+1)(S_k^(q+1)^H)⁻¹(S_k^(q+1))^Hy_(i)^(q+1).  [Equation 16]

Here,

F_c⁺

is a fixed value at all times. Therefore, it is possible to avoid the matrix inversion by pre-calculating and storing the fixed value. Since the proposed channel estimation scheme does not require the P×P matrix inversion for each OFDM symbol, it is possible to significantly reduce the complexity.

In order to cancel soft interference in the received symbol, the soft interference cancellation units 510-1, . . . , 510-n may be embodied as given by,

$\begin{array}{cc}{\stackrel{\_}{y}}_{\left(i\right)}^{\left(q+1\right)}=y_{\left(i\right)}-\sum _{j=1,j\ne k}^{N_t}{\stackrel{\_}{S}}_j^{\left(q+1\right)}{\hat{h}}_{\mathrm{ij}}^{\left(q\right)}& \left[\mathrm{Equation}17\right]\end{array}$

According to an aspect of the present invention, when it is determined no error has occurred while receiving a packet, the channel estimation unit 240 may use an errorlessly received symbol and thus it is possible to improve the channel estimation performance.

In order to estimate an initial frequency domain channel for a received symbol, the initial frequency domain channel estimation units 520-1, . . . , 520-n may be embodied as given by,

$\begin{array}{cc}{\stackrel{⋓}{h}}_{\mathrm{ik}}^{\left(q+1\right)}={\left({{\stackrel{\_}{S}}_k^{\left(q+1\right)}\left({\stackrel{\_}{S}}_k^{\left(q+1\right)}\right)}^H\right)}^{-1}{\left({\stackrel{\_}{S}}_k^{\left(q+1\right)}\right)}^H{\stackrel{\_}{y}}_{\left(i\right)}^{\left(q+1\right)}& \left[\mathrm{Equation}18\right]\end{array}$

In order to estimate a time domain channel for the received symbol, the time domain channel estimation units 530-1, . . . , 530-n may be embodied as given by,

ĝ_ik^(q+1)=F_c⁺{hacek over (h)}_ik^(q+1).  [Equation 19]

In order to estimate a final frequency domain channel for the received symbol, the final frequency domain channel estimation units 540-1, . . . , 540-n may be embodied as given by,

ĥ_ik^(q+1)=F_cĝ_ik^(q+1).  [Equation 20]

An iterative EM channel estimation scheme may perform soft interference cancellation to thereby reduce a computational amount. However, in the case of an initial iteration, that is, in the case q=0,

ĥ_ij⁽⁰⁾

does not exist and thus the soft interference cancellation using Equation 17 may be impossible. Therefore, the present invention proposes a soft interference cancellation scheme for a pilot symbol using channel estimation obtained from a previous OFDM symbol.

When

S_pn,j^p

is defined as a pilot symbol of a j^th transmit antenna that is transmitted via a P_n subcarrier and

y_(i)^p=[y_p1,i y_p2,i . . . y_pNp,i]^T,

a received signal for a pilot subcarrier may be expressed as,

$\begin{array}{cc}{y}_{\left(i\right)}^p=S_k^ph_{\mathrm{ik}}^p+\sum _{j=1,j\ne k}^{N_t}S_j^ph_{\mathrm{ij}}^p+v_{\left(i\right)}^p& \left[\mathrm{Equation}21\right]\end{array}$

where

h_ij^p=[h_p1,ij h_p2,ij . . . h_pNp,ij]^T

S_k^p=diag(s_p1,k^p, s_p2,k^p, . . . , s_pNp^p,k)

v_(i)^p=[v_p1,k v_p2,k . . . v_pNp,k]^T.

When

F_Pε

is defined as a pilot DFT matrix for CIR and

[F_p]_n,p=e^{−j2π(n−1)(p−1)/N},

the relationship as shown in the following Equation 22 may be accomplished,

h_kj^p=F_pg_kj.  [Equation 22]

Generally, columns of

F_p

are not orthogonal to each other and are given by,

F_p^†F_p=I_p  [Equation 23]

where

F_p⁺

denotes a pseudo-inverse matrix of

F_p.

Therefore, it is possible to induce the relationship as given by,

g_kj=F_p⁺h_kj^p.  [Equation 24]

Finally, Equation 21 may be represented as,

$\begin{array}{cc}{y}_{\left(i\right)}^p=S_k^pF_p^{+}g_{\mathrm{ik}}+\sum _{j=1,j\ne k}^{N_t}S_j^pF_p^{+}g_{\mathrm{ij}}+v_{\left(i\right)}^p& \left[\mathrm{Equation}25\right]\end{array}$

The present invention may cancel soft interference in the pilot symbol using the channel estimation obtained from the previous OFDM symbol, as given by,

$\begin{array}{cc}{\stackrel{\_}{y}}_{\left(i\right)}^p=y_{\left(i\right)}^p-\sum _{j=1,j\ne k}^{N_t}S_j^p{\hat{h}}_{\mathrm{ij}}^{\mathrm{pre}}& \left[\mathrm{Equation}26\right]\end{array}$

where

$h_{\mathrm{ij}}^{\mathrm{pre}}={\left[\begin{array}{cccc}{h}_{p_1,\mathrm{ij}}^{\mathrm{pre}}& h_{p_2,\mathrm{ij}}^{\mathrm{pre}}& \dots & h_{p_{N_p},\mathrm{ij}}^{\mathrm{pre}}\end{array}\right]}^T$

and

h_pn,ij^pre

denotes a channel estimate of

h_pn,ij

that is finally obtained from the previous OFDM symbol. A final value may be obtained from a preamble OFDM symbol.

Accordingly, the initial channel estimation using the pilot symbol may be represented as,

ĥ _ik ⁽⁰⁾ =F _e(F_p^H(S_k^p)^HS_k^pF_p)⁻¹F_p^H(S_k^p)^Hy_(i)^p.  [Equation 27]

Here,

(F_p^H(S_k^p)^HS_k^pF_p)⁻¹

is a fixed value at all times. Therefore, it is possible to avoid the matrix inversion by pre-calculating and storing the fixed value.

FIG. 6 is a block diagram illustrating a configuration of a channel estimation unit 600 using a pilot symbol.

Referring to FIG. 6, in order to cancel interference in the pilot symbol, soft pilot interference cancellation units 610-1, . . . , 610-n may be embodied as given by,

$\begin{array}{cc}{\stackrel{\_}{y}}_{\left(i\right)}^p=y_{\left(i\right)}^p-\sum _{j=1,j\ne k}^{N_t}S_j^p{\hat{h}}_{\mathrm{ij}}^{\mathrm{pre}}.& \left[\mathrm{Equation}28\right]\end{array}$

In order to estimate an initial frequency domain channel with respect to the pilot symbol, initial pilot-based frequency domain channel estimation units 620-1, . . . , 620-n may be embodied as given by,

{hacek over (h)} _ik ⁽⁰⁾=(S_k^p)^Hy_(i)^p.  [Equation 29]

In order to estimate a time domain channel of the pilot symbol, initial pilot-based time domain channel estimation units 630-1, . . . , 630-n may be embodied as given by,

ĝ _ik ⁽⁰⁾=(F_p^H(S_k^p)^HS_k^pF_p)⁻¹F_p^H{hacek over (h)}_ik⁽⁰⁾.  [Equation 30]

In order to estimate a final frequency domain channel of the pilot symbol, final pilot-based frequency domain channel estimation units 640-1, . . . , 640-n may be embodied as given by,

ĥ_ik⁽⁰⁾=F_ey_(i)^pĝ_ik⁽⁰⁾.  [Equation 31]

FIG. 7 is a block diagram illustrating a configuration of the noise variance estimation unit 250 shown in FIG. 2.

Referring to FIG. 7, the noise variance estimation unit 250 includes per-subcarrier noise variance estimation units 710-1, . . . , 710-n and a final noise variance estimation unit 720. The noise variance estimation unit 250 may estimate

{circumflex over (R)}_vv^(q+1)

based on output

S ^(q+1)

of the tentative symbol decision unit 230, output

Ĥ^(q+1)

of the channel estimation unit 240, and the received signal Y. An estimate of

R_vv

in a

(q+1)^th

iteration may be given by,

$\begin{array}{cc}{\hat{R}}_{\mathrm{vv}}^{\left(q+1\right)}=\frac{1}{N}\sum _{n=1}^N\left(y_n-{\hat{H}}_n^{\left(q+1\right)}{\stackrel{\_}{s}}_n^{\left(q+1\right)}\right){\left(y_n-{\hat{H}}_n^{\left(q+1\right)}{\stackrel{\_}{s}}_n^{\left(q+1\right)}\right)}^H.& \left[\mathrm{Equation}32\right]\end{array}$

The per-subcarrier noise variance estimation units 710-1, . . . , 710-n may estimate each per-subcarrier noise variance as given by,

{circumflex over (R)} _n,vv ^(q+1)=(y_n−Ĥ_n^(q+1)s_n^(q+1))(y_n−Ĥ_n^(q+1)s_n^(q+1))^H.  [Equation 33]

Next, the final noise variance estimation unit 720 may average the variance to estimate the final noise variance, as given by,

$\begin{array}{cc}{\hat{R}}_w^{\left(q+1\right)}=\frac{1}{N}\sum _{n=1}^N{\hat{R}}_{n,\mathrm{vv}}^{\left(q+1\right)}& \left[\mathrm{Equation}34\right]\end{array}$

FIG. 8 is a block diagram illustrating a configuration of an initial noise variance estimation unit 800.

Referring to FIG. 8, the initial noise variance estimation unit 800 obtains an initial value of a noise variance matrix using a pilot symbol. An initial noise variance matrix may be estimated as given by,

$\begin{array}{cc}{\hat{R}}_{\mathrm{vv}}^{\left(0\right)}=\frac{1}{N_p}\sum _{n=1}^{N_p}\left(y_{\mathrm{pn}}-{\hat{H}}_{\mathrm{pn}}^{\left(0\right)}s_{\mathrm{pn}}^p\right){\left(y_{\mathrm{pn}}-{\hat{H}}_{\mathrm{pn}}^{\left(0\right)}s_{\mathrm{pn}}^p\right)}^H& \left[\mathrm{Equation}35\right]\end{array}$

Initial per-subcarrier noise variance estimation units 810-1, . . . , 810-n may estimate each per-subcarrier noise variance of the pilot symbol as given by,

{circumflex over (R)} _n,vv ⁽⁰⁾=(y_pn−Ĥ_pn⁽⁰⁾s_pn^p)(y_pn−Ĥ_pn⁽⁰⁾s_pn^p)^H.  [Equation 36]

An initial final noise variance estimation unit 820 may estimate the final noise variance of the pilot symbol by averaging the per-subcarrier noise variance of the pilot symbol, as given by,

$\begin{array}{cc}{\hat{R}}_w^{\left(0\right)}=\frac{1}{N_p}\sum _{p_n=1}^{N_p}{\hat{R}}_{n,\mathrm{vv}}^{\left(0\right)}.& \left[\mathrm{Equation}37\right]\end{array}$

As described above, according to the present invention, it is possible to provide a channel estimation scheme with similar performance to a joint channel estimation scheme and with lower complexity due to not requiring a matrix inversion, in a MIMO OFDM system.

Also, according to the present invention, it is possible to perform soft interference cancellation based on a value, estimated in a previous OFDM symbol, in an initial channel estimation stage and thereby improve the channel estimation performance and reduce the complexity.

Also, according to the present invention, it is possible to improve the performance of interference noise variance estimation based on all the information associated with previous effective subcarriers when estimating noise variance in a MIMO OFDM system.

Although a few embodiments of the present invention have been shown and described, the present invention is not limited to the described embodiments. Instead, it would be appreciated by those skilled in the art that changes may be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. A transmission apparatus comprising: an encoding unit encoding a transmission packet;an interleaving unit interleaving the encoded transmission packet;a mapping unit mapping the interleaved transmission packet;a pilot insertion unit inserting a pilot symbol and an information symbol into a subcarrier of the mapped transmission packet; anda transmit antenna transmitting the transmission packet with the inserted pilot symbol and information symbol.

2. A reception apparatus comprising: a detection unit detecting extrinsic information based on an output of a tentative symbol decision unit, an output of a channel estimation unit, and an output of a noise variance estimation unit that are obtained from a previous iteration process;a Cyclic Redundancy Check (CRC) aided channel decoding unit outputting an encoded and interleaved bit, or a posteriori information thereof based on the extrinsic information;the tentative symbol decision unit determining a tentative transmission symbol based on an output of the CRC aided channel decoding unit;the channel estimation unit estimating a channel based on an output of the tentative symbol decision unit or a pilot symbol; andthe noise variance estimation unit estimating a noise variance or an initial noise variance based on the output of the tentative symbol decision unit and the output of the channel estimation unit.

3. The reception apparatus of claim 2, wherein the CRC aided channel decoding unit comprises: a deinterleaving unit receiving a bit extrinsic log-likelihood ratio (LLR) from the detection unit and deinterleaving the received bit extrinsic LLR to generate a priori LLR of a channel decoding unit;the channel decoding unit decoding the priori LLR to output a transmission information bit extrinsic LLR and a posteriori LLR of an encoded bit;a CRC checking unit checking the transmission information bit extrinsic LLR to determine whether an error has occurred while receiving a packet;an encoding unit encoding a received information bit corresponding to an output of the CRC checking unit; andan interleaving unit interleaving the encoded bit or the posteriori LLR thereof to generate a priori LLR of the tentative symbol decision unit.

4. The reception apparatus of claim 2, wherein the tentative symbol decision unit comprises: a decision selection unit selecting any one scheme from tentative symbol decision schemes;a pilot replacement unit replacing a tentative symbol decision with a pilot symbol when the transmission symbol is the pilot symbol;a hard decision unit mapping a received encoded bit corresponding to an output of the CRC checking unit to determine a tentative symbol, when it is determined no error has occurred while receiving a packet; anda soft decision unit determining the tentative symbol with respect to the transmission symbol based on a posteriori LLR of the interleaved bit corresponding to an output of the channel decoding unit, when it is determined the error has occurred.

5. The reception apparatus of claim 2, wherein the channel estimation unit comprises: a soft interference cancellation unit canceling an interference component based on a channel estimation obtained from the previous iteration process and a tentative symbol decision;an initial frequency domain channel estimation unit estimating an initial frequency domain channel based on a received signal and the tentative symbol decision;a time domain channel estimation unit estimating a time domain channel based on the estimated initial frequency domain channel; anda final frequency domain channel estimation unit estimating a final frequency domain channel based on the estimated time domain channel.

6. The reception apparatus of claim 2, wherein the channel estimation unit comprises: a soft pilot interference cancellation unit canceling a pilot interference component based on a pilot symbol and a channel estimation that is obtained from a previous orthogonal frequency division multiplexing (OFDM) symbol;an initial pilot-based frequency domain channel estimation unit estimating an initial pilot-based frequency domain channel based on a received pilot signal and the pilot symbol;a pilot-based time domain channel estimation unit estimating a pilot-based time domain channel based on the estimated initial pilot-based frequency domain channel; anda final pilot-based frequency domain channel estimation unit estimating a final pilot-based frequency domain channel based on the estimated pilot-based time domain channel.

7. The reception apparatus of claim 2, wherein the noise variance estimation unit comprises: a per-subcarrier noise variance estimation unit estimating a per-subcarrier noise variance based on a received signal, a tentative symbol decision, and a channel estimation obtained from the previous iteration process; anda final-noise variance estimation unit averaging the estimated per-subcarrier noise variance to estimate a final noise variance.

8. The reception apparatus of claim 2, wherein the noise variance estimation unit comprises: a initial per-subcarrier noise variance estimation unit estimating a pilot per-subcarrier noise variance based on a received pilot signal, a pilot symbol, and a channel estimation obtained from the previous iteration process; andan initial final noise variance estimation unit averaging the estimated pilot per-subcarrier noise variance to estimate a final pilot noise variance.

9. A transmission method comprising: encoding a transmission packet;interleaving the encoded transmission packet;mapping the interleaved transmission packet;inserting a pilot symbol and an information symbol into a subcarrier of the mapped transmission packet; andtransmitting the transmission packet with the inserted pilot symbol and information symbol.

10. A reception method comprising: detecting extrinsic information based on a tentative symbol decision, a channel estimation, and a noise variance estimation that are obtained from a previous iteration process;outputting an encoded and interleaved bit or a posteriori information thereof based on the extrinsic information;determining a tentative transmission symbol based on the outputting;estimating a channel based on a tentative symbol decision or a pilot symbol; andestimating a noise variance or an initial noise variance based on the tentative symbol decision and the channel estimation.

11. The method of claim 10, wherein the determining of the tentative transmission symbol comprises: selecting any one scheme from tentative symbol decision schemes;replacing the tentative symbol decision with a pilot symbol when the transmission symbol is the pilot symbol;mapping a received encoded bit to determine a tentative symbol, when it is determined no error has occurred while receiving a packet; anddetermining the tentative symbol with respect to the transmission symbol based on a posteriori LLR of the interleaved bit, when it is determined the error has occurred.

12. The method of claim 10, wherein the estimating of the channel comprises: canceling an interference component based on the channel estimation obtained from the previous iteration process and the tentative symbol decision;estimating an initial frequency domain channel based on a received signal and the tentative symbol decision;estimating a time domain channel based on the estimated initial frequency domain channel; andestimating a final frequency domain channel based on the estimated time domain channel.

13. The method of claim 10, wherein the estimating of the channel comprises: canceling a pilot interference component based on a pilot symbol and a channel estimation that is obtained from a previous OFDM symbol;estimating an initial pilot-based frequency domain channel based on a received pilot signal and the pilot symbol;estimating a pilot-based time domain channel based on the estimated initial pilot-based frequency domain channel; andestimating a final pilot-based frequency domain channel based on the estimated pilot-based time domain channel.

14. The method of claim 10, wherein the estimating of the noise variance comprises: estimating a per-subcarrier noise variance based on a received signal, the tentative symbol decision, and the channel estimation obtained from the previous iteration process; andaveraging the estimated per-subcarrier noise variance to estimate a final noise variance.

15. The method of claim 10, wherein the estimating of the noise variance comprises: estimating an initial per-subcarrier noise variance based on a received pilot symbol, a pilot symbol, and the channel estimation obtained from the previous iteration process; andaveraging the estimated initial per-subcarrier noise variance to estimate an initial final pilot noise variance.