KR20090125452A - Ofdm receiving system and method for ici cancellation and equalization for ofdm system - Google Patents

Abstract

PURPOSE: An OFDM receiving system and a method for ICI cancellation and equalization for OFDM system are provided to reduce an error transmission caused by a determination error while reducing computational complexity. CONSTITUTION: In a device, an MMSE-SIC detector(214) uses equalization matrixes to minimize a mean square error in consideration of a decision error. A soft demapper(215) performs symbol demapping to an output of the MMSE-SIC detector. The soft demapper performs the symbol demapping to an output of the MMSE equalizer. A SISO decoder(218) performs SISO-decoding of the soft bit information obtained from the result of the symbol demapping SISO. An ICI remover(221) eliminates ICI by using soft bit information which is processed through SISO decoding. An MMSE equalizer(222) applies equalization matrixes to the signal which the ICI is removed from in consideration of a decision error.

Description

Orthogonal Frequency Division Multiplexing System and Method for ICI Cancellation and Equalization for OFDM System

The present invention relates to orthogonal frequency division multiplexing communication, and more particularly, to an orthogonal frequency division multiplexing system and a method of canceling and equalizing interference between subcarriers in the system.

The demand for high data rate multimedia services is rapidly increasing in the field of wireless digital communication and broadcasting in mobile environments. Orthogonal Frequency Division Multiplexing has attracted attention as one of the powerful systems that meets this need within a multipath fading channel. In particular, most terrestrial digital broadcasting systems such as digital video broadcasting-handheld (DVB-H), forward link only (FLO), integrated services digital broadcasting-terrestrial (ISDB-T), and terrestrial-digital multimedia broadcasting (TDMB) OFDM is adopted.

If the channel response does not change during the OFDM block period in the OFDM system, orthogonality between subcarriers may be maintained. In practice, however, when transmitting high capacity data in the high frequency band, the OFDM system experiences a time and frequency selective channel (or doubly selective channel). In this situation, there is a fatal weakness in which inter-carrier interference (ICI) occurs due to loss of orthogonality. In addition, since the Doppler frequency increases in proportion to the carrier frequency despite movement at the same speed, the reception performance is affected accordingly. Thus, to overcome these problems on time and frequency selective channels, many techniques have been studied to suppress or eliminate ICI. Most of the techniques proposed to date are based on iterative processing between detection and decoding, but they are not able to effectively remove or suppress ICI such as high computational complexity and error propagation due to decision errors in the iteration process. There is a problem.

The present invention provides an orthogonal frequency division multiplexing system and an ICI cancellation and equalization method in the orthogonal frequency division multiplexing system, which can effectively reduce error propagation due to a decision error while reducing computational complexity. There is.

In order to solve the above technical problem, the orthogonal frequency division multiplexing system according to the present invention uses a sequential interference cancellation technique according to a pre-detection criterion for a signal received through a receiving antenna, but minimizes the mean square error in consideration of a determination error. An MMSE-SIC detector that applies an equalization matrix to and a soft demapper, a SISO decoder, an ICI canceler, and an MMSE equalizer to perform iterative ICI removal and equalization, wherein the soft demapper includes the MMSE- Symbol demapping is performed on the output of the SIC detector, symbol demapping is performed on the output of the MMSE equalizer in the subsequent iterations, and the SISO decoder outputs SISO decoded soft bit information obtained as a result of the symbol demapping. The ICI remover removes the ICI using the SISO decoded soft bit information, The MMSE equalizer is characterized by applying an equalization matrix to minimize the mean square error in consideration of the determination error to the signal from which the ICI is removed.

Here, the equalization matrix of the MMSE-SIC detector may be calculated using a decision error covariance matrix.

In addition, the equalization matrix may be obtained according to the following equation.

은 전송 신호의 분산을,

은 잡음의 분산을, ρ는

을, H는 채널 행렬을,

는 크기

의 직 사각행렬을,

는 허미시안(Hermitian) 행렬을,

은 판정 오류 공분산 행렬을, I _N은 크기 N의 단위 행렬을 나타낸다.Where N is the number of subcarriers,

The dispersion of the transmitted signal,

Is the variance of the noise,

Where H is the channel matrix,

The size

Square matrix,

Is a Hermitian matrix,

Denotes a decision error covariance matrix, and I _N denotes an identity matrix of size N.

In addition, the determination error covariance matrix may be defined as follows.

는 복소 공액(complex conjugate)를 나타내고, 조건부 기대값

은, 오류 e_m 및 e_n이 각각 부정확한 판정, 즉

및

로부터 발생됨을 가리키고, 조건부 기대값

은

에 따라 구해지고, 여기서 집합

는 연판정(soft decision) 포인트

및 그것을 둘러싸는 이웃한 성상도 포인트들로 이루어진다.here,

Represents a complex conjugate, the conditional expected value

Means that errors e _m and e _n are each incorrect, i.e.

And

Conditional expected value

silver

Obtained according to, where the set

Is a soft decision point

And neighboring constellation points surrounding it.

In addition, the determination error covariance matrix may be expressed according to the following equation by ignoring non-diagonal elements.

는 대각 행렬을 의미한다. here,

Means diagonal matrix.

로 정의되고, 상기 등화 행렬은 다음 수학식과 같이 표현될 수 이다.Also, Q _err is

Where the equalization matrix can be expressed as

The orthogonal frequency division multiplexing system further includes a channel estimator for estimating channel information, and the MMSE-SIC detector may determine a determination order according to an SINR calculated using the channel information.

In addition, the determination order may be determined according to the following equation.

은 전송 신호의 분산을,

은 잡음의 분산을,

은 p번째 채널 행렬의 m번째 열벡터를 나타낸다.Where N is the number of subcarriers,

The dispersion of the transmitted signal,

Is the dispersion of noise,

Denotes the mth column vector of the pth channel matrix.

In addition, the equalization matrix of the MMSE-SIC detector may be obtained according to the following equation in the i th step of the sequential interference cancellation technique.

은 전송 신호의 분산을,

은 잡음의 분산을,

은 p번째 채널 행렬의 m번째 열벡터를,

는 허미시안(Hermitian) 행렬을,

및

는 각각 본래 채널 행렬 및 판정 오류 공분산 행렬의 일부 원소들로 재구성한 크기 2q+1(≪N)의 채널 행렬 및 판정 오류 공분산 행렬을, I _2q+1은 크기 2q+1의 단위행렬을 나타낸다.here,

The dispersion of the transmitted signal,

Is the dispersion of noise,

Is the mth column vector of the pth channel matrix,

Is a Hermitian matrix,

And

Denotes a channel matrix of size 2q + 1 (<N) and a decision error covariance matrix reconstructed with some elements of the original channel matrix and the decision error covariance matrix, respectively, and I _{2q + 1} represents a unit matrix of size 2q + 1.

In addition, the determination error covariance matrix may be calculated according to the following equation.

는 대각 행렬을 의미하고, 조건부 기대값

은 here,

Means the diagonal matrix, and the conditional expected value

silver

에 따라 구해지며, 여기서 집합

는 연판정(soft decision) 포인트

및 그것을 둘러싸는 이웃한 성상도 포인트들로 이루어진다.

Is obtained according to

Is a soft decision point

And neighboring constellation points surrounding it.

In addition, the equalization matrix of the MMSE equalizer may be calculated using a decision error covariance matrix. In this case, the determination error covariance matrix may be calculated using a pre-LLR value provided from the SISO decoder.

In addition, the output of the MMSE equalizer for the p-th subcarrier is expressed as

는 상기 ICI 제거기의 출력을,

는 잡음 성분을 나타내며,

이고,

이며, 상기 등화 행렬

는 다음 수학식에 따라 구해질 수 있다.Where N is the number of subcarriers,

Is the output of the ICI eliminator,

Represents the noise component,

ego,

Equalization matrix

Can be obtained according to the following equation.

은 전송 신호의 분산을,

은 잡음의 분산을,

은 p번째 채널 행렬의 m번째 열벡터를 나타내고,

는 본래 채널 행렬의 일부 원소들로 재구성한 크기 2q+1(≪N)의 채널 행렬이며,

는

의 공분산 행렬이고, I _2q+1은 크기 2q+1의 단위행렬을 나타낸다.here,

The dispersion of the transmitted signal,

Is the dispersion of noise,

Represents the mth column vector of the pth channel matrix,

Is a channel matrix of size 2q + 1 (≪N), reconstructed from some elements of the original channel matrix,

Is

Is a covariance matrix of and I _{2q + 1} represents a unit matrix of size 2q + 1.

는 다음 수학식에 따라 계산될 수 있다.Also, the

May be calculated according to the following equation.

는 대각 행렬을 의미한다.here,

Means diagonal matrix.

은 다음 수학식에 따라 계산될 수 있다.Also, the

May be calculated according to the following equation.

는 연판정(soft decision) 포인트

및 그것을 둘러싸는 이웃한 성상도 포인트들로 이루어지고,

는 사전(a priori) LLR 값이다.Where set

Is a soft decision point

And neighboring constellation points surrounding it,

Is a priori LLR value.

In addition, the soft demapper may obtain the LLR value for the p th symbol X _p in the initial iteration according to the following equation.

는 X_p의 i번째 비트이고, u_p는 상기 MMSE-SIC 검출기의 출력으로서

로 표현되고,

이며, 집합

는 연판정(soft decision) 포인트

및 그것을 둘러싸는 이웃한 성상도 포인트들로 이루어지고,

및

는

의 두 서로 배타적인 부분집합으로서 i번째 비트

위치

에 각각 "0" 및 "1"을 가지는 심볼들로 이루어진다.here,

Is the i th bit of X _p and u _p is the output of the MMSE-SIC detector.

Represented by

Is, set

Is a soft decision point

And neighboring constellation points surrounding it,

And

Is

I-bit as two mutually exclusive subsets of

location

Is composed of symbols having "0" and "1" respectively.

In addition, the soft demapper may obtain the LLR value for the p th symbol X _p in the subsequent iteration according to the following equation.

는 X_p의 i번째 비트이고, z_p는 상기 MMSE 등화기의 출력으로서

로 표현되고,

이며, 집합

는 연판정(soft decision) 포인트

및 그것을 둘러싸는 이웃한 성상도 포인트들로 이루어지고,

및

는

의 두 서로 배타적인 부분집합으로서 i번째 비트

위치

에 각각 "0" 및 "1"을 가지는 심볼들로 이루어진다.here,

Is the i th bit of X _p , and z _p is the output of the MMSE equalizer.

Represented by

Is, set

Is a soft decision point

And neighboring constellation points surrounding it,

And

Is

I-bit as two mutually exclusive subsets of

location

Is composed of symbols having "0" and "1" respectively.

In order to solve the above technical problem, according to the present invention, an ICI cancellation and equalization method in an orthogonal frequency division multiplexing system uses a sequential interference cancellation technique based on a pre-detection criterion for a signal received through a receiving antenna, MMSE-SIC detection process that applies equalization matrix to minimize mean squared error in consideration, and symbol demapping process, SISO decoding process, ICI removal process, MMSE equalization process to perform iterative ICI removal and equalization In the symbol demapping process, symbol demapping is performed on an output of the MMSE-SIC detection process at an initial iteration, and symbol demapping is performed on an output of the MMSE equalization process at an iteration, and the SISO decoding process is performed. SISO decodes soft bit information obtained as a result of the symbol demapping, and the ICI removal process is performed by the SISO decoupling. The soft bits of information by using a removing ICI, and the MMSE equalization process is characterized by applying an equalization matrix to minimize the mean square error, taking into account the error in the determination signal obtained by the ICI is removed.

Here, the equalization matrix in the MMSE-SIC detection process may be calculated using a decision error covariance matrix.

The ICI removal and equalization method may further include a channel estimating process of estimating channel information, and may determine a determination order according to an SINR calculated using the channel information in the MMSE-SIC detection process.

In addition, the equalization matrix in the MMSE equalization process may be calculated using a decision error covariance matrix. In this case, the determination error covariance matrix may be calculated using a pre-LLR value provided from the SISO decoding process.

In order to solve the above technical problem, according to the present invention, an ICI cancellation and equalization method in an orthogonal frequency division multiplexing system uses a sequential interference cancellation technique based on a pre-detection criterion for a signal received through a receiving antenna, Taking into account the MMSE-SIC detection process applying an equalization matrix to minimize the mean square error, and performing symbol demapping on the output of the MMSE-SIC detection process in the initial iteration, and then performing SISO decoding in the iteration to remove the ICI. The method includes an iterative ICI removal and equalization process that applies an equalization matrix to minimize the mean square error to the obtained signal and performs symbol demapping on the resulting output.

Here, in the ICI removal and equalization process, an equalization matrix for minimizing the mean square error may be applied to the signal from which the ICI is removed in consideration of a determination error.

In order to solve the above technical problem, there is provided a computer-readable recording medium recording a program for executing the ICI removal and equalization method in the orthogonal frequency division multiplexing system according to the present invention described above.

According to the present invention described above, it is possible to effectively reduce error propagation due to decision errors while reducing computational complexity in performing ICI removal and equalization. Therefore, it is possible to obtain an improvement in reception performance or to obtain the same or more reception performance with less computation.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the following description and the accompanying drawings, the substantially identical components are represented by the same reference numerals, and thus redundant description will be omitted. In addition, in the following description of the present invention, if it is determined that a detailed description of a related known function or configuration may unnecessarily obscure the subject matter of the present invention, the detailed description thereof will be omitted.

1 is a schematic block diagram of an OFDM transmission system according to an embodiment of the present invention. In the OFDM transmission system according to the present embodiment, the encoder 110, the bit interleaver 120, the symbol mapper 130, the serial / parallel converter 140, the IFFT unit 150, the CP inserter 160, and the parallel / A serial converter 170, a transmit antenna 180.

The encoder 110 encodes an information bit stream, and a convolutional code, a turbo code, a low density parity code (LDPC), or the like may be used as an encoding scheme. The bit interleaver 120 interleaves the data stream encoded by the encoder in units of bits, and the symbol mapper 130 modulates it in units of symbols. In this case, 64QAM, 16QAM, QPSK, or the like may be used as the modulation scheme. The modulated symbol is converted into a parallel symbol by the serial / parallel converter 140 and an IFFT operation is performed by the IFFT unit 150. The CP inserter 160 inserts a cyclic prefix to remove inter-symbol interference (ISI) in the modulated symbol. The parallel / serial converter 170 converts the parallel symbols into serial symbols, and the OFDM symbols are transmitted over the wireless channel through the transmit antenna 180.

In the present invention, a time and frequency selective channel (or doubly selective channel) is considered as a wireless channel, with a maximum delay T _d , a root mean square delay spread τ _rms , a maximum Doppler The maximum Doppler frequency f _d is considered.

The channel frequency response (CFR) in the time and frequency selective channel is expressed by the following equation.

Here, h _{n, l} is a time domain channel impulse response, and represents a complex random variable for the l (0≤n <L) th path at time n (0≤n <N), and (k) _N denotes the remainder of N modulo k, k denotes the subcarrier index, N denotes the number of subcarriers, which means the size of the FFT or IFFT.

After the IFFT, the signal transmitted at the nth time is given by the following equation.

Here, X _k represents a data symbol of the k th subcarrier in the parking area.

Assuming that a multipath fading channel is composed of L separate paths, a signal received in an OFDM receiving system to be described later may be represented by the following equation.

Here, w _n represents additive white Gaussian noise (AWGN) at the nth time. It can also be assumed that the overall channel impulse response is within the guard interval (0 ≦ (L−1) T _s ≦ N _g T _s ), where T _s and N _g are the sampling interval and the guard interval, respectively. Maximum index of (guard interval).

2 is a schematic block diagram of an OFDM reception system according to an embodiment of the present invention, in which the illustrated OFDM reception system performs ICI removal and equalization and recovers a transmission signal. The OFDM reception system according to the present embodiment includes a reception antenna 200, a serial / parallel converter 211 for converting a received signal sequence into a parallel signal sequence, a CP remover 212 for removing a cyclic prefix, and an FFT operation. FFT unit 213 for performing the operation, MMSE-SIC detector 214 for applying an equalization matrix for minimizing the mean square error by applying a sequential interference cancellation technique to the received signal, iterative ICI cancellation ( soft demapper 215, SISO decoder 218, ICI canceler 221, MMSE equalizer 222, etc., for performing iterative ICI cancellation.

The soft demapper 215 performs symbol demapping on the output of the MMSE-SIC detector 214 in the initial iteration of the iterative ICI removal process, and symbol demapping the output of the MMSE equalizer 222 in subsequent iterations. Do this. Soft bit information from the soft demapper 215 is converted into parallel data by the parallel / serial converter 216, and deinterleaving is performed bit by bit by the deinterleaver 217. The SISO decoder 218 decodes the soft bit information and outputs the soft-input soft-output (SISO). The SISO decoded soft bit information is interleaved by the interleaver 219 and converted into parallel data by the serial / parallel converter 220. The ICI canceler 221 removes the ICI from the received signal using the SISO decoded soft bit information, and the MMSE equalizer 222 applies the equalization matrix to minimize the mean square error to the signal from which the ICI has been removed. Output to the mapper 215. The channel estimator 223 estimates channel information such as a channel frequency response by using a pilot signal, and the estimated channel information is operated by the MMSE-SIC detector 214, the MMSE equalizer 222, and the ICI canceler 221. Is utilized.

Hereinafter, the operation of the above-described OFDM reception system will be described in more detail.

After the cyclic prefix is removed in the CP remover 212, the frequency domain signal is obtained by performing the FFT of y _n in Equation 3 in the FFT unit 213 as shown in the following equation.

Here, N _k represents a noise component of the k th subcarrier in the frequency domain. In the above equation, α _k and β _k represent multiplicative distortion and ICI in the kth desired subcarrier, respectively. These may each be expressed as the following equations.

As a result, the signal received in the k-th subcarrier in the frequency domain is represented by the following equation.

이다.here,

to be.

When the multipath channel is time-variant during the block period, Equation 4 may be expressed in a vector form as in the following equation.

이고,

이며, H는 다음 수학식과 같다.here,

ego,

H is represented by the following equation.

이고,

는 전치(transpose)를 나타낸다. 상기 수학식들에서, ICI는 채널 주파수 응답들(CFRs)

및 k번째 부반송파에서 요망되지 않는 데이터 심볼들

으로 이루어짐을 알 수 있다. 상기 수학식 8에서 전송 데이터 벡터 X를 복원하기 위해서는 채널 추정이 수행되어야 한다. 본 실시예에서는 OFDM 수신 시스템 측에서 채널 추정기(223)을 통하여 채널 행렬 H를 알고 있는 것으로 가정하기로 한다.here,

ego,

Represents a transpose. In the above equations, ICI is the channel frequency responses (CFRs).

And undesired data symbols on the kth subcarrier

It can be seen that. In Equation 8, channel estimation must be performed to reconstruct the transmission data vector X. In the present embodiment, it is assumed that the channel matrix H is known through the channel estimator 223 at the OFDM receiving system.

라 하고, 단순화를 위해, 후술할 선검출 기준(pre-detection criterion)에 따라 판정 순서(decision ordering)가

과 같이 형성된 것으로 가정하자. 그리고

,

,

로 정의하자. 논문 [H. Wang, X. Chen, S. Zhou, and M. Zhoa, and Y. Yao, ""Low-Complexity ICI cancellation in frequency domain for OFDM systems in time-varying multipath channels,"" IEICE Trans. Commun., vol. E89-B, no. 3, pp. 1020??-1023, Mar. 2006.]에 의하면, 순차적 간섭 제거의 (i-1)번째 스텝에서 선검출된 심볼 벡터

이, i번째 스텝에서의 수신 벡터 신호로부터 제거된다면, 수정되는 수신 벡터

는 다음 수학식과 같이 주어진다.symbols detected at the kth subcarrier

For the sake of simplicity, decision ordering is based on a pre-detection criterion described below.

Assume that it is formed as follows. And

,

,

Let's define Thesis [H. Wang, X. Chen, S. Zhou, and M. Zhoa, and Y. Yao, "" Low-Complexity ICI cancellation in frequency domain for OFDM systems in time-varying multipath channels, "IEICE Trans. Commun., Vol. E89-B, no. 3, pp. 1020 ??-1023, Mar. 2006.], symbol vector pre-detected in step (i-1) of sequential interference cancellation.

Is removed from the received vector signal at the i th step,

Is given by the following equation.

인 경우에만 달성된다. 일반적으로 이전 스텝의 판정들이 반드시 정확하지는 않으므로 오류 전파(error propagation)로 인한 성능 저하가 발생할 수 있다. However, this development is only possible if the judgments of the previous step are correct, i.e.

Is achieved only if In general, since the determinations of the previous step are not necessarily accurate, performance degradation due to error propagation may occur.

Therefore, in the present embodiment, the MMSE-SIC detector 214 uses a sequential interference cancellation technique, but considers a determination error, and applies an equalization matrix to the received signal to minimize the mean square error. The operation of this MMSE-SIC detector 214 is described in more detail below.

은 i번째 스텝에서 간섭들로 간주된다. 이러한 상황에서 오류 전파가 불가피해지고, 판정 오류의 존재로 인해 심볼 검출을 신뢰할 수가 없게 된다. Remaining undetected symbols in detecting X _p in the p-th desired subcarrier

Is considered interference in the i th step. In such a situation, error propagation becomes inevitable, and the presence of a decision error makes symbol detection unreliable.

Therefore, in the present embodiment, Equation 10 is modified as follows.

이고,

로 정의된다. 순차적 간섭 제거 과정의 진행에서 일어나는 오류 전파를 피하기 위해서, 본 실시예에 따른 MMSE-SIC 검출기(214)는, 상기 수학식 11에 기초한 판정 오류를 고려한다.Where a determination error

ego,

Is defined as In order to avoid error propagation occurring in the sequential interference cancellation process, the MMSE-SIC detector 214 according to the present embodiment considers a determination error based on Equation 11 above.

G equalizer matrix to minimize the mean square error (MSE) defined by the following equation,

G is described in J. Proakis, Digital communications, Prentice-Hall, 3rd Edition, 1995.] can be obtained by the following equation from the orthogonality principle.

Therefore, G is obtained as follows.

는 허미시안(Hermitian) 행렬을 의미하고, 공분산(covariance) 행렬이

및

로 정의된다. here,

Is a Hermitian matrix, and the covariance matrix is

And

Is defined as

Since all transmission signals are independent and identically distributed (iid) with an average of 0, and the transmission signal and noise component are uncorrelated, G is obtained as follows.

는 크기

의 직사각행렬이고,

은 전송 신호의 분산을,

은 잡음의 분산을, ρ는

을, I _N은 크기 N의 단위 행렬을 나타내며,

이고

라는 사실이 이용된다.here,

The size

Is a rectangular matrix of,

The dispersion of the transmitted signal,

Is the variance of the noise,

Where I _N represents an identity matrix of size N,

ego

Is used.

이 다음과 같이 정의될 수 있다.And a decision error covariance matrix of size i

This can be defined as:

는 복소 공액(complex conjugate)를 나타내고, 조건부 기대값

이, 오류 em 및 en이 각각 부정확한 판정, 즉

및

로부터 발생됨을 가리키기 위해 사용된다. 대각 원소(diagonal element)들

를 구하는 방법은 후술하기로 한다. m≠n인 비대각 원소들

는, 가우시안 프로세스처럼 판정 오류를 근사화함으로써 성상도(constellations)의 대칭성으로 인해 무시될 수 있다. 따라서 이하에서는 계산의 복잡도를 줄이고자

의 비대각 원소들은 무시하기로 한다. 그러면,

는 다음 수학식과 같이 나타낼 수 있다.here,

Represents a complex conjugate, the conditional expected value

This implies that the errors em and en are each incorrect, i.e.

And

Used to indicate that it originates from. Diagonal elements

How to obtain will be described later. non-diagonal elements with m ≠ n

Can be ignored due to the symmetry of constellations by approximating the decision error, like a Gaussian process. Therefore, to reduce the complexity of the calculation below

The non-diagonal elements of are ignored. then,

Can be expressed as the following equation.

는 대각 행렬을 의미한다. Q _err 를 다음과 같이 정의하면,here,

Means diagonal matrix. If you define Q _err as

Equation 15 may be expressed as the following equation.

As described above, the use of the equalization matrix in consideration of the decision error in the MMSE-SIC detector 214 can compensate for the decision error that occurs during the iterative operation of sequential interference cancellation.

In the time and frequency selective channel, the determinant according to Equation 8 can be simplified. Since most of the energy of most ICI components is concentrated in several neighboring subcarriers of the desired data symbol, elements that do not significantly affect Y _k in Equation 9 can be ignored. This can be expressed as follows.

Where 2q represents the number of major ICI terms. In other words, the ICI components come from several nearby subcarriers. In this respect, it is possible to reduce the size except for the main region with respect to Equation (9). If an approximation according to Equation 20 is applied to Equation 9, it may be expressed as the following Equation.

3 shows the structure of such a channel matrix.

와 같이 MMSE 등화 벡터의 크기를 줄일 수 있다. 이 경우 상기 수학식 8에 따른 수신 벡터는 다음과 같이 다시 표현될 수 있다.Similarly, use only 2q + 1 (<N) taps to avoid large matrix inverse transformations.

As can be seen, the size of the MMSE equalization vector can be reduced. In this case, the reception vector according to Equation 8 may be represented as follows.

는 다음과 같고,here,

Is

이며,

은 p번째 채널 행렬

의 m번째 열벡터이다. 이와 같이, 본래 채널 행렬 H의 일부 원소들로 재구성한 크기 2q+1(≪N)의 채널 행렬을 사용함으로써 행렬 연산의 복잡도를 감소시킬 수 있다. 마찬가지로, 판정 오류 공분산 행렬 Q _err 역시 크기 2q+1(≪N)의 행렬로 재구성되어야 할 것이다.

,

Is the p-th channel matrix

M-th column vector of. As such, the complexity of the matrix operation can be reduced by using a channel matrix of size 2q + 1 (&quot; N) reconstructed with some elements of the original channel matrix H. Similarly, the decision error covariance matrix Q _{err will} also have to be reconstructed into a matrix of size 2q + 1 (&quot; N).

4 is a flow diagram illustrating one embodiment of the operation of the MMSE-SIC detector 214 based on the above.

First, in step 410, a pre-detection SINR of all symbols of N points is calculated. Pre-detection criterion can be expressed as the following equation.

In step 420, a symbol having the largest SINR among the undetected symbols is selected according to the above equation. For simplicity, assume that the desired symbol is X _p at the i th step.

In step 430, an equalization matrix is obtained using the decision error covariance matrix. If it is the i th step, the MMSE coefficient may be calculated according to the following equation.

은 전송 신호의 분산을,

은 잡음의 분산을, I _2q+1은 크기 2q+1의 단위행렬을 나타내며, 크기 2q+1의 판정 오류 공분산 행렬(decision error covariance matrix)은 다음 수학식에 따라 계산될 수 있다.here,

The dispersion of the transmitted signal,

_Denotes a variance of noise, I _{2q + 1} denotes a unit matrix of size 2q + 1, and a decision error covariance matrix of size 2q + 1 may be calculated according to the following equation.

를 구하고, 경판정(hard decision)을 이용하여

를 구한다. 이때,

는 다음 수학식과 같이 표현된다.In step 440, output to the soft demapper to be described later

Is obtained, and using hard decision

Obtain At this time,

Is expressed as the following equation.

이고,

이다.here,

ego,

to be.

In step 450, for the next step, the interference is removed to regenerate the received vector as shown in the following equation.

The above steps 420 to 450 are repeated until all data is detected (step 460).

It is known that by applying soft decoding a significant improvement in system performance is obtained. Some assumptions will be made about the output of the MMSE-SIC detector 214, and the soft demapper 215 will derive an optimal soft bit metric that takes into account detection errors.

를 정의하자. 여기서

을 모든 성분이 0인 (2q+1)ㅧ1 열벡터이다. 그리고

를 다음과 같이 정의한다.

Let's define it. here

Is a (2q + 1) ㅧ 1 column vector with all components equal to zero. And

Define as

이다. 그리고 상기 수학식 27에 관하여

및

로 대신하면, MMSE-SIC 검출기(214)의 출력은 다음과 같이 수정될 수 있다.here,

to be. And with respect to Equation 27

And

Instead, the output of MMSE-SIC detector 214 can be modified as follows.

및

이다. here,

And

to be.

가 복소 가우시안 분포를 가진다고 말할 수 있다.

내의 각 항들은 서로 독립적이기 때문에,

의 분산은 다음 수학식과 같이 간단히 계산될 수 있다.[HV Poor and S. Verdüu, "" Probability of error in MMSE multiuser detection, "IEEE Trans. Inf. Theory, vol. 43, no. 3, pp. 858 ??-871, May 1997.], it is known that the error probability of the MMSE detector can be easily approximated under the assumption that the distribution of residual interference-plus-noise is Gaussian. have. Similarly,

Can be said to have a complex Gaussian distribution.

Since each term within is independent of each other,

The variance of can be simply calculated as

및

를 사용하여 어떻게 상기 수학식 17의

을 구하는지 살펴본다. MMSE-SIC 검출기(214)의 다음 스텝 (i+1)을 위하여, 조건부 기대값

이 다음 수학식에 따라 구해질 수 있다.Now, the equations 30 and 31

And

How to use equation 17 above

See if you get. Conditional expected value for the next step (i + 1) of the MMSE-SIC detector 214

This can be obtained according to the following equation.

는 연판정(soft decision) 포인트

및 그것을 둘러싸는 이웃한 성상도 포인트들로 이루어진다. 조건부 확률

는 가우시안 근사(Gaussian approximation)를 이용하여 계산될 수 있다. Where set

Is a soft decision point

And neighboring constellation points surrounding it. Conditional probability

Can be calculated using a Gaussian approximation.

를 가우시안 랜덤 변수로 가정함으로써 계산될 수 있다.

의 조건부 확률 분포 함수(conditional probability density function)는 다음과 같이 주어진다.In the soft demapper 215, the log likehood ratio (LLR) for the soft bit information is

Can be calculated by assuming that is a Gaussian random variable.

The conditional probability density function of is given by

를 성상도 집합

의 하나의 성분이라 하자.

는 X_p의 i번째 비트이고,

의 두 서로 배타적인 부분집합

및

가 i번째 비트

위치

에 각각 "0" 및 "1"을 가지는 심볼들로 이루어진다고 정의하자. 또한 일반적으로, 전송되는 심볼들은 동일한 발생 확률을 가진다고 가정하자. 그러면, 소프트 디매퍼(215)에서 MMSE-SIC 검출기(214)의 출력으로부터 구해지는 소프트 정보인

의 사후(a posteriori) LLR은 다음과 같이 정의될 수 있다. 여기서,

는

의 p번째 비트를 나타낸다.

Constellation set

Let's call it one component.

Is the i th bit of X _p ,

Two mutually exclusive subsets of

And

Is the i bit

location

Let's define that it consists of symbols with "0" and "1" respectively. Also, in general, assume that transmitted symbols have the same probability of occurrence. Then, the soft demapper 215 is soft information obtained from the output of the MMSE-SIC detector 214.

A posteriori LLR may be defined as follows. here,

Is

Represents the p-th bit of the.

Thesis [F. Tosato and P. Bisaglia, "" Simplified soft-output demapper for binary interleaved COFDM with application to HIPERLAN / 2, "" in Proc. IEEE Int. Conf. Commun., Pp. 664 ??-668, Sep. Applying the max-log rule presented in 2002. and applying a little approximation, Equation 34 can be obtained as the following equation.

.

.

Now, iterative ICI cancellation and MMSE equalization performed in the OFDM receiving system shown in FIG. 2 will be described. As shown in FIG. 2, channel information obtained from the channel estimator 223 may be utilized to further improve reception performance by repetitive ICI removal. Since the output of SISO decoder 218 is not yet reliable at an early stage, in this embodiment MMSE equalizer 222 considers a decision error to mitigate error propagation. As a result, a priori LLR provided from SISO decoder 218 is used to correctly reproduce the ICI.

에 대한 사전(a priori) LLR이 상기 수학식 22의

로부터 ICI를 제거하기 위해 사용된다.

를 사전 LLR

로부터 변환되어 ICI 제거기(221)로 입력되는 소프트 정보라 하자.In order to effectively detect the desired symbol X _p in the p th desired subcarrier, the surrounding symbols

A priori LLR for

Used to remove ICI from

Advance LLR

It is assumed that the soft information is converted from and inputted into the ICI eliminator 221.

는 ICI를 제거함으로써 다음과 같이 수정된다.In the ICI remover 221, the received signal

Is modified as follows by removing the ICI:

,

및

이다.here,

,

And

to be.

는 전송 심볼 X_p의 추정치를 얻고 잔여 간섭 및 잡음을 줄이기 위해 상기 수학식 36의 수정된 수신 벡터에 적용되기 위해 크기 1ㅧ(2q+1)로 형성되는 것이 바람직하다. 이때 MMSE 등화기(222)의 출력은 다음과 같이 표현된다. Equalization Matrix of MMSE Equalizer 222

Is preferably formed with a magnitude of 1 ms (2q + 1) to be applied to the modified received vector of equation (36) to obtain an estimate of the transmission symbol X _p and reduce residual interference and noise. At this time, the output of the MMSE equalizer 222 is expressed as follows.

Here, it is necessary to minimize the variance of the estimation error defined as follows.

Similarly, from Equation 14, the equalization matrix of the MMSE equalizer 222 can be obtained as follows.

는

의 공분산 행렬로서 다음과 같다.here,

Is

The covariance matrix of is given by

이 다음과 같이 계산될 수 있다.If SISO decoder 218 is applied,

This can be calculated as follows.

이고,

는 SISO 디코더(218)로부터 제공되는, i번째 부반송파에 대한 사전 LLR이다. 인터리버(219)로 인해, 모든 비트들이 독립적임이 명백하다. 따라서 판정 오류 확률은 다음 수학식에 따라 구해질 수 있다.here,

ego,

Is the prior LLR for the i th subcarrier, provided from the SISO decoder 218. Due to interleaver 219, it is clear that all bits are independent. Therefore, the determination error probability can be obtained according to the following equation.

는

의 i번째 비트를 나타낸다. 상기 수학식 34에 따른 LLR의 정 의로부터,

는 다음과 같이 주어진다. here,

Is

Represents the i-th bit of. From the definition of LLR according to Equation 34,

Is given by

을 구함으로써 판정 오류 공분산 행렬을 효과적으로 구할 수 있다.As described above, the determination error probability is obtained from the prior LLR, and the expected value is used using this determination error probability.

By determining, the decision error covariance matrix can be effectively obtained.

을 계산한 후, 상기 수학식 39에서

가

와 함께 얻어질 수 있다. MMSE 등화기(222)

를 상기 수학식 36의

에 적용하면, X_p의 편향 추정(biased estimate)을 다음과 같이 얻을 수 있다. From Equation 41

After calculating the following equation (39)

end

Can be obtained with MMSE Equalizer (222)

Of Equation 36

When applied to, the biased estimate of X _p can be obtained as

이고,

이다.here,

ego,

to be.

로부터, 소프트 비트 정보로서 X_p에 대한 사후 LLR 값이 소프트 디매퍼(215)에서 다음 수학식과 같이 계산될 수 있다. In iterations after the beginning of the iterative ICI removal process, the output of 222 of the MMSE equalizer

From, the post LLR value for X _p as soft bit information can be calculated in the soft demapper 215 as follows.

According to the present invention described above, by using MMSE detection using a sequential interference cancellation technique, it is possible to improve reception performance at an initial stage. In this case, error propagation can be minimized by applying an equalization matrix in consideration of a decision error. In addition, error propagation can be further reduced by applying an equalization matrix in consideration of decision errors in repetitive ICI elimination and equalization processes in subsequent stages. In addition, when the determination error covariance matrix is obtained during iterative ICI removal and equalization, it can be effectively obtained by using a prior LLR value.

의 동작이 요구된다. 보통, 후검출 기준(post-detection criterion)에 기초하여 순차적 간섭 제거 기법이 적용된 MMSE 검출기의 경우 복잡도

의 동작이 필요하다. 상술한 본 발명에 의하면, 초기 반복 또는 그 이후의 반복에서 상기 수학식 25 또는 상기 수학식 39에 따른 등화 행렬을 계산하는 데에는 복잡도

의 동작이 요구되므로, 복잡도가 상당히 경감된다. 또한 크기 2q+1(≪N)의 채널 행렬 및 공분산 행렬을 사용함으로써 모든 반복 과정에서

의 행렬 역변환이 요구되므로, 연산량이 감소된다. As is already known, an MMSE detector using a total of N taps has a complexity

Operation is required. Usually, the complexity of MMSE detectors with sequential interference cancellation based on post-detection criterion

Operation is required. According to the present invention described above, the complexity of calculating the equalization matrix according to Equation (25) or (39) in the initial or subsequent iterations is complicated.

Operation is required, so the complexity is considerably reduced. In addition, by using a channel matrix and covariance matrix of size 2q + 1 (

Since the matrix inverse of is required, the amount of computation is reduced.

As a result, according to the present invention, it is possible to effectively reduce error propagation due to determination errors while significantly reducing computational complexity, thus leading to an improvement in reception performance.

Meanwhile, the above-described embodiments of the present invention can be written as a program that can be executed in a computer, and can be implemented in a general-purpose digital computer that operates the program using a computer-readable recording medium. The computer-readable recording medium may be a magnetic storage medium (for example, a ROM, a floppy disk, a hard disk, etc.), an optical reading medium (for example, a CD-ROM, DVD, etc.) and a carrier wave (for example, the Internet). Storage medium).

So far I looked at the center of the preferred embodiment for the present invention. Those skilled in the art will appreciate that the present invention can be implemented in a modified form without departing from the essential features of the present invention. Therefore, the disclosed embodiments should be considered in descriptive sense only and not for purposes of limitation. The scope of the present invention is shown in the claims rather than the foregoing description, and all differences within the scope will be construed as being included in the present invention.

1 is a schematic block diagram of an OFDM transmission system according to an embodiment of the present invention.

2 is a schematic block diagram of an OFDM reception system according to an embodiment of the present invention.

3 is a schematic diagram showing the structure of an approximated channel matrix according to an embodiment of the present invention.

4 is a flowchart illustrating an operation of an MMSE-SIC detector according to an embodiment of the present invention.

Claims (25)

An MMSE-SIC detector using sequential interference cancellation based on pre-detection criteria for signals received through the receiving antenna and applying an equalization matrix to minimize the mean square error in consideration of decision errors; Includes a soft demapper, a SISO decoder, an ICI remover, an MMSE equalizer, to perform repetitive ICI removal and equalization, The soft demapper performs symbol demapping on the output of the MMSE-SIC detector at an initial iteration, and symbol demapping on the output of the MMSE equalizer at an iteration afterwards. The SISO decoder outputs SISO decoded soft bit information obtained as a result of the symbol demapping, The ICI remover removes ICI using the SISO decoded soft bit information, And the MMSE equalizer applies an equalization matrix to minimize the mean squared error in consideration of a decision error to the signal from which the ICI has been removed. The method of claim 1, An equalization matrix of the MMSE-SIC detector is calculated using a decision error covariance matrix. The method of claim 2, The equalization matrix is orthogonal frequency division multiplexing system according to the following equation.

Where N is the number of subcarriers,

The dispersion of the transmitted signal,

Is the variance of the noise,

Where H is the channel matrix,

The size

The rectangular matrix of,

Is a Hermitian matrix,

Denotes a decision error covariance matrix, and I _N denotes an identity matrix of size N. The method of claim 3, And the determination error covariance matrix is defined as in the following equation.

here,

Represents a complex conjugate, the conditional expected value

Means that errors e _m and e _n are each incorrect, i.e.

And

Conditional expected value

silver

Obtained according to, where the set

Is a soft decision point

And neighboring constellation points surrounding it. The method of claim 4, wherein And the determination error covariance matrix is represented according to the following equation by ignoring non-diagonal elements.

here,

Means diagonal matrix. The method of claim 5, Q _err is

And the equalization matrix is expressed by the following equation.

The method of claim 1, Further comprising a channel estimator for estimating channel information, And the MMSE-SIC detector determines the order of determination according to the SINR calculated using the channel information. The method of claim 7, wherein Orthogonal frequency division multiplexing system according to the following equation.

Where N is the number of subcarriers,

The dispersion of the transmitted signal,

Is the dispersion of noise,

Denotes the mth column vector of the pth channel matrix. The method of claim 2, And an equalization matrix of the MMSE-SIC detector is obtained according to the following equation in an i th step of the sequential interference cancellation scheme.

here,

The dispersion of the transmitted signal,

Is the dispersion of noise,

Is the mth column vector of the pth channel matrix,

Is a Hermitian matrix,

And

Denotes a channel matrix of size 2q + 1 (<N) and a decision error covariance matrix reconstructed with some elements of the original channel matrix and the decision error covariance matrix, respectively, and I _{2q + 1} represents a unit matrix of size 2q + 1. The method of claim 9, And the determination error covariance matrix is calculated according to the following equation.

here,

Means the diagonal matrix, and the conditional expected value

silver

Is obtained according to

Is a soft decision point

And neighboring constellation points surrounding it. The method of claim 1, And an equalization matrix of the MMSE equalizer is calculated using a decision error covariance matrix. The method of claim 11, And the determination error covariance matrix is calculated using a pre-LLR value provided from the SISO decoder. The method of claim 11, The output of the MMSE equalizer for the p th subcarrier is expressed as

Where N is the number of subcarriers,

Is the output of the ICI eliminator,

Represents the noise component,

ego,

, The equalization matrix

Orthogonal frequency division multiplexing system characterized in that is obtained according to the following equation.

here,

The dispersion of the transmitted signal,

Is the dispersion of noise,

Represents the mth column vector of the pth channel matrix,

Is a channel matrix of size 2q + 1 (≪N), reconstructed from some elements of the original channel matrix,

Is

Is a covariance matrix of and I _{2q + 1} represents a unit matrix of size 2q + 1. The method of claim 13, remind

Orthogonal frequency division multiplexing system characterized in that is calculated according to the following equation.

here,

Means diagonal matrix. The method of claim 14, remind

Orthogonal frequency division multiplexing system characterized in that is calculated according to the following equation.

Where set

Is a soft decision point

And neighboring constellation points surrounding it,

Is a priori LLR value. The method of claim 2, The soft demapper obtains an LLR value for a p th symbol X _p at an initial iteration according to the following equation.

here,

Is the i th bit of X _p and u _p is the output of the MMSE-SIC detector.

Represented by

Is, set

Is a soft decision point

And neighboring constellation points surrounding it,

And

Is

I-bit as two mutually exclusive subsets of

location

Is composed of symbols having "0" and "1" respectively. The method of claim 11, The soft demapper, in subsequent iterations, obtains the LLR value for the p-th symbol X _p according to the following equation.

here,

Is the i th bit of X _p , and z _p is the output of the MMSE equalizer.

Represented by

Is, set

Is a soft decision point

And neighboring constellation points surrounding it,

And

Is

I-bit as two mutually exclusive subsets of

location

Is composed of symbols having "0" and "1" respectively. In the method of ICI cancellation and equalization in orthogonal frequency division multiplexing system, MMSE-SIC detection process for applying the equalization matrix to minimize the mean squared error by using sequential interference cancellation technique according to pre-detection criteria for the signal received through the receiving antenna, Including symbol demapping, SISO decoding, ICI removal, MMSE equalization, to perform repetitive ICI removal and equalization, In the symbol demapping process, symbol demapping is performed on an output of the MMSE-SIC detection process at an initial iteration, and symbol demapping is performed on an output of the MMSE equalization process at an iteration. The SISO decoding process SISO decodes soft bit information obtained as a result of the symbol demapping, The ICI removal process removes ICI using the SISO decoded soft bit information, In the MMSE equalization process, the ICI removal and equalization method is characterized by applying an equalization matrix to minimize the mean squared error in consideration of a decision error on the signal from which the ICI has been removed. The method of claim 18 The equalization matrix in the MMSE-SIC detection process is calculated using a decision error covariance matrix. The method of claim 18, A channel estimation process of estimating channel information, The MMSE-SIC detection process determines the decision order according to the SINR calculated using the channel information. The method of claim 18, The equalization matrix in the MMSE equalization process is calculated using a decision error covariance matrix. The method of claim 21, And the decision error covariance matrix is calculated using a pre-LLR value provided from the SISO decoding process. A computer-readable recording medium having recorded thereon a program for executing an ICI removal and equalization method in an orthogonal frequency division multiplexing system according to any one of claims 18 to 22. In the method of ICI cancellation and equalization in orthogonal frequency division multiplexing system, MMSE-SIC detection process for applying the equalization matrix to minimize the mean squared error by using sequential interference cancellation technique according to pre-detection criteria for the signal received through the receiving antenna, In the initial iteration, symbol demapping is performed on the output of the MMSE-SIC detection process, and in subsequent iterations, an equalization matrix is applied to minimize the mean square error to the SISO-decoded ICI-removed signal and the resulting output ICI removal and equalization method comprising an iterative ICI removal and equalization process that performs symbol demapping. The method of claim 24, In the ICI removal and equalization process, the ICI removal and equalization method, characterized in that to apply the equalization matrix to minimize the mean square error in consideration of the decision error to the signal from which the ICI is removed.

Images