KR101093920B1 - Lattice Reduction method-based MIMO receiver under time varying fading environments - Google Patents

Abstract

The present invention achieves the same performance as the conventional Lattice Reduction (LR) based multiple input / output receiver by using the time correlation of fading channels in a multiple input multiple output (MIMO) wireless communication system, and greatly increases the complexity of the receiver. The present invention relates to a lattice reduction method of a multiple input / output receiver that can be reduced. The present invention for this purpose, in the Lattice Reduction (LR: Lattice Reduction) method of a multiple input and output wireless communication system using N transmit antennas and L receive antennas, each of which constitutes a plurality of blocks for storing K channels step; When performing a lattice reduction (LR) at the receiving end of the receiving antenna, LR is not performed for every channel, and a K-th channel, that is, an nK-th (n = 1, 2, 3, ... Performing an LR only on a channel); And a matrix comparison step of comparing the P matrix obtained by performing LR in the nK th channel and the P matrix obtained by performing LR in the previous block (n-1) K th channel.

Description

Lattice Reduction Method of Lattice Reduction-based Multiple Input-Output Receivers in Time-varying Fading Channel Environments {Lattice Reduction method-based MIMO receiver under time varying fading environments}

The present invention relates to a lattice reduction method of a Lattice Reduction (LR) -based Multiple Input Mutiple Output (MIMO) receiver in a time-varying fading channel environment. More specifically, the present invention relates to a multiple input multiple output (MIMO). In the wireless communication system, the Lattice reduction method of a multi-input / output receiver which can reduce the complexity of the receiver while achieving the same performance as a conventional Lattice reduction-based multi-input / output receiver by using the time correlation of fading channels. It is about.

Recently, as high-speed data transmission is required in a wireless communication environment, interest in a MIMO system is increasing. In the spatial multiplexing scheme, in which a transmitting end simultaneously transmits different information to a plurality of transmission antennas, a diversity gain and a transmission speed gain are obtained as the number of transmitting and receiving antennas increases. In this case, since different signals are simultaneously transmitted from a plurality of transmitting antennas, signals received by adding all signals are received at each receiving antenna. Therefore, the receiving end should separate the incoming signal for each antenna. As a technique of detecting a signal for each antenna at the receiving end, there are a maximum likelihood detection (MLD), zero-forcing (ZR), and a minimum mean square error (MMSE).

The MLD method is a technique of measuring a Euclidean distance between transmittable symbol vectors and a received signal and selecting a symbol vector that minimizes the minimum as a received vector. This method is the detection technique that has the most ideal performance and is the basis for performance comparison with other methods. However, as the number of transmitting antennas increases or the modulation order increases, computational complexity increases exponentially. It is also difficult to implement in hardware. To compensate for this drawback, ZF and MMSE methods, which are linear signal detection techniques, have been proposed. This method has a lower computational complexity than the ML method, but there is a serious problem in performance degradation due to noise amplification in the linear detection process.

As a method to solve this problem, a LRAD (Lattice Reduction Aided Detection) method has been proposed. The LRAD method significantly reduces noise amplification by converting the channel matrix to be orthogonal and performing linear detection. In the LRAD method, LR (Lattice Reduction) is used to convert the channel matrix to be almost orthogonal. The Gauss reduction, Minkowski reduction, and Korkine-Zolotareff reduction techniques of the LR method transform the channel matrix to be orthogonal by finding the best basis, but have a disadvantage in that it has a large amount of computation. The LLL (Lenstra-Lenstra-Lovasz) -LR technique, which compensates for these shortcomings, uses the Real LLL-LR technique, Gram-schmidt orthonomalization, which converts a complex channel into a real orthogonal channel by converting the complex channel into a real number. Complex LLL-LR using QR, and LLL-LR with QR using QR.

The LLL-LR technique of transforming the channel matrix to be orthogonal requires a large amount of computation. Therefore, the LLL-LR technique has a disadvantage in that the amount of calculation increases as the number of antennas increases. In addition, in the time-varying fading channel environment, although the matrix P obtained through the LR does not change well, there is a problem in that unnecessary calculation is performed by performing the LR for every channel matrix.

Therefore, the LR technique is required to show the same performance as the conventional LR technique in a time-varying fading channel environment and to reduce the computational complexity.

The present invention has been invented in consideration of the above situation, and has a low complexity and a conventional LRAD (Lattice Reduction-Aided Detection) method using a time correlation of a fading channel in a multiple input multiple output (MIMO) wireless communication system. The purpose of the present invention is to provide a Lattice Reduction (LR) method of a multiple input / output receiver having the same performance.

인 것을 특징으로 한다. In order to achieve the above object, a Lattice Reduction (LR) method of a multiple input / output receiver according to the present invention is a Lattice Reduction (LR) of a multiple input / output wireless communication system using N transmit antennas and L receive antennas. A method comprising: storing K channels and received symbols at a receiving end; When performing a lattice reduction (LR) at the receiving end of the receiving antenna, LR is not performed in every channel, and LR is performed only in the Kth channel, that is, the nKth channel (n = 1, 2, 3, ...). Performing an LR step; A matrix comparison step of comparing a P matrix obtained by performing LR on the nK th channel and a P matrix obtained by performing LR on a previous block (n-1) K th channel, wherein P is N × in the P matrix. Is of size N, all components are integers, and the determinant

It is characterized by that.

Preferably, the P matrix of the nK th channel and the P matrix of the (n-1) K th channel are compared, and when the two matrices are not the same, LR in the remaining K-1 channel frames without performing LR. Performing P to obtain a P matrix (sequential mode); Comparing the P matrix of the nK th channel and the P matrix of the (n-1) K th channel, when the two matrices are the same, the remaining K-1 P matrices without performing LR are P matrixes of the nK th channel. It is set as the same matrix as (block mode).

Preferably, when performing LR in the remaining K-1 channel frames, LR is performed using the P matrix obtained in the previous channel frame as an initial condition.

Preferably, when performing LR in the nK-th channel, LR is performed by setting the P matrix of the nK-th channel matrix in the (n-1) K-th channel as an initial condition and n = 1. In this case, LR is performed by setting the P matrix as a unit matrix having an N × N matrix.

Preferably, after obtaining the P matrix of the K channels in the above manner, after estimating the K received symbol vectors through the conventional Lattice Reduction-Aided Detection (LRAD), the method is repeated to estimate the received symbol vectors. (N = n + 1).

Preferably, when storing the K channels, K, which minimizes the calculation amount, is determined based on the calculation amount of lattice reduction and the channel capacity of the receiver since K, which minimizes the calculation amount, depends on the Doppler frequency.

Preferably, the K may vary depending on the Doppler frequency or may be set to a fixed K.

Application of the lattice reduction method of the multi-input / output receiver according to the present invention to an LR-based MIMO receiver in a time-varying fading channel environment can provide an advantage of greatly reducing the complexity of the receiver while obtaining the same performance as a conventional LR-based MIMO receiver.

1 is a flowchart of an embodiment of a lattice reduction method of a multiple input / output receiver according to the present invention.
2 is an exemplary diagram of a complex LLL-LR technique used to describe a lattice reduction method of a multiple input / output receiver according to the present invention.
3 is an exemplary diagram for explaining a conventional lattice reduction-aided detection technique.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings. In describing the present invention, if it is determined that a detailed description of a related function or configuration is unnecessary to the gist of the present invention, the detailed description thereof will be omitted.

First, the theoretical background of the present invention is as follows. That is, in the actual MIMO environment, the channel environment changes with time, but the channel environment does not change independently, but has a characteristic that changes slowly and has a correlation with the channel of the previous time. In this time-varying fading channel environment, it is confirmed that the matrix P obtained through the LR technique is the same or similar to the matrix P obtained in the previous channel due to the time correlation characteristics of the channel. Accordingly, the present invention can provide an LR technique using such a characteristic to reduce the computational amount of the computational LR technique.

이다. Referring to FIG. 1, in a lattice reduction method of a multiple input / output receiver according to the present invention, storing K channels and a reception symbol vector in a multiple input / output wireless communication system using N transmit antennas and L receive antennas (S100). ); When performing a lattice reduction (LR) at the receiving end of the receiving antenna, LR is not performed for every channel, and LR is performed only in the K-th channel, that is, the nK-th channel (n = 1, 2, 3, ...). LR performing step (S200) to perform; And a matrix comparison step S300 for comparing the P matrix obtained by performing LR in the nK th channel and the P matrix obtained by performing LR in the (n-1) K th channel. In addition, in the case where the P matrix of the nK th channel and the P matrix of the (n-1) th channel are the same, the block mode in which the P matrix of the K-1 channel frame is the same as the P matrix of the nK th channel without performing LR. Step S400; If the P matrix of the n-th channel is not the same as the P-matrix of the (n-1) K-th channel, the sequential mode step of performing LR to obtain a P-matrix of K-1 channel frames without performing LR (S500) ; After calculating P matrix of K channels, performing LRAD, and repeating the above scheme (n = n + 1) to estimate a received symbol vector (S600). Herein, in the P matrix, P has a size of N × N, all components are integers, and a determinant is

to be.

The operation of the lattice reduction method of the multiple input / output receiver according to the present invention configured as described above will be described with reference to FIGS. 1 to 3.

For example, in a multiple input / output wireless communication system composed of N transmit antennas and L receive antennas, a channel, a transmission signal, and a reception signal are represented by Equation 1 below.

[Equation 1]

In Equation 1, s _j denotes a transmission signal transmitted from a j th transmission antenna, y _i denotes a reception signal through an i th reception antenna, and h _ji denotes between a j th transmission antenna and an i th reception antenna. The channel gain of, n _i denotes noise added to the i-th receiving antenna.

In the MIMO system, the LRAD technique can significantly reduce the noise amplification that occurs by using a virtual channel in which the channel matrix is almost orthogonal through the LR technique. In the LR technique, H ′, which is substantially orthogonal to the channel H, may be represented by Equation 2 below.

[Equation 2]

이다. 채널 행렬의 열벡터를 거의 직교를 이루는 새로운 기저 벡터로 변환시키는 LR 방식은 다양하다. 본 발명에서는 도 2에 나타낸 컴플렉스(complex) LLL-LR 기법을 적용하였다. In Equation 2, P has a size of N × N, all components are integers, and a determinant is

to be. There are a variety of LR methods for converting a column vector of a channel matrix into a new basis vector that is almost orthogonal. In the present invention, the complex LLL-LR technique shown in FIG. 2 is applied.

In the signal detection process using the LR, the received symbol vector may be represented by Equation 3 below.

&Quot; (3) "

을 곱하여

를 구할 수 있다.

를 구한 후 양자화(quantized)한 후 P를 곱하여 송신 심벌 벡터를 검출한다. Lattice Reduction-Aided detection using LR is shown in FIG. 3. First, in the received symbol vector y as shown in [Equation 4]

Multiply by

Can be obtained.

After quantizing and quantizing and multiply by P to detect the transmission symbol vector.

&Quot; (4) "

In the time-varying fading channel environment, P does not change significantly. In the present invention, the LR is performed only in the K-th channel by storing K channels, which is a block, instead of performing LR every channel frame. In this case, when the remaining K-1 channels, i, are (n-1) K <i <nK, the channel H _i and the transmission signal y are stored without performing LR. only if i is to be an integer multiple of K (i = nK) after performing the LR made in P _nK matrix and the old block of nK second channel frame (n-1) P of the K-th channel frame _{(n-1) K} Compare In the method of the present invention, LR is performed by dividing into a block mode and a sequential mode according to a result of comparing two P matrices.

First, the block mode is operated only when P _nK and P _{(n-1) K} are compared. If the two matrices P _nK and P _{(n-1) K} are the same, the channel does not change quickly, so the probability of P _i ( (n-1) K <i <n ) of the remaining K-1 channels is equal to P _nK This is very high. In this case, in the K-1 channel frame that does not perform LR, the conventional adaptive LR method is used by using the same matrix as P _nK without obtaining the integer matrix P _i ( (n-1) K <i <nK ) through LR. Reduce unnecessary computation.

A second sequential mode, the probability P _nK and P _(n-1) If _K is not equal to the remaining K-1 channel frame _{P i ((n-1)} K <i <nK) is not the same as compared to the P _nK Since LR is performed, P _i ( (n-1) K <i <nK) is obtained. In this case, when LR is executed, P _i ( (n-1) K <i <nK) is set as the unit matrix (I _N ), and LR is not performed but LR is performed with P _{i - 1} obtained from the previous channel as an initial condition. do.

The operation procedure of an embodiment of the present invention is as follows.

1) Initialize to n = 1.

2) H _i when (n-1) K <i <nK And transmit signal y.

3) LR is performed in the nK th channel frame. When performing LR, set P _i = P _{(n-1) K} (P _i = I _N when n = 1) and perform LR.

4)

IF P _nK = P _{(n-1) K} : block mode

FOR i = (n-1) K + 1,... , nK-1Do

SET P _i = P _nK

COMPUTE

END FOR

ELSE: sequential mode

FOR i = (n-1) K + 1,... , nK-1Do

_{{Perform LR for H i with initial} setting of P i = P i -1 in the LR iteration}

END FOR

를 이용하여 LRAD를 수행한다. 5)

Perform LRAD using.

6) Repeat the process of n = n + 1, 2) ~ 5).

In the present invention, the value of K, which minimizes the amount of calculation, depends on the Doppler frequency of the time-varying fading channel. The value of K can be determined according to the Doppler frequency through experiments, and a fixed value of K can be used instead of using a different K value according to the Doppler frequency.

As described above, the embodiments of the present invention have been described, but these are merely exemplary, and those skilled in the art may understand that other embodiments corresponding to various modifications and equivalents are possible therefrom. will be. Therefore, the true technical protection scope of the present invention will be defined by the technical spirit of the appended claims.

Claims (7)

In a Lattice Reduction (LR) method of a multiple input / output wireless communication system using N transmit antennas and L receive antennas,
Storing K channels and a received symbol vector;
When performing a lattice reduction (LR) at the receiving end of the receiving antenna, LR is not performed for every channel, and LR is performed only in the K-th channel, that is, the nK-th channel (n = 1, 2, 3, ...). Performing an LR step;
A matrix comparison step of comparing a P matrix obtained by performing LR on the nK th channel and a P matrix obtained by performing LR on a previous block (n-1) K th channel;
Comparing the P matrix of the nK th channel and the P matrix of the (n-1) K th channel, when the two matrices are not the same, LR is performed in the remaining K-1 channel frames that do not perform LR and P is performed. Sequential mode step of obtaining a matrix; And
Comparing the P matrix of the nK th channel and the P matrix of the (n-1) K th channel, when the two matrices are the same, the remaining K-1 P matrices without performing LR are P matrixes of the nK th channel. Lattice reduction method of a multiple input-output wireless communication system comprising a block mode step to be equal to. The method of claim 1,
In the P matrix, P has a size of N × N, all components are integers, and the determinant is

Lattice reduction method of a multiple input-output wireless communication system, characterized in that. The method of claim 2,
Lattice reduction method of a multiple input-output wireless communication system, characterized in that when performing the LR in the remaining K-1 channel frame, LR is performed using the P matrix obtained in the previous channel frame as an initial condition. The method of claim 2,
When LR is performed in the nK th channel, LR is performed by setting the P matrix of the nK th channel matrix in the (n−1) K th channel as an initial condition. A Lattice reduction method of a multiple input / output wireless communication system, characterized in that LR is performed by setting a matrix as a unit matrix having an N × N matrix. The method of claim 1,
After obtaining the P matrix of the K channels, the K received symbol vectors are estimated through the conventional Lattice Reduction-Aided Detection (LRAD), and the steps are repeated (n = n + 1) to estimate the received symbol vectors. Lattice reduction method of a multiple input-output wireless communication system, characterized in that. The method of claim 1,
When storing the K channels, the minimum K is calculated according to the Doppler frequency, so the Lattice reduction method of a multiple input / output wireless communication system is determined based on the calculation amount of the lattice reduction and the channel capacity of the receiver. The method of claim 6,
The K is a lattice reduction method of a multiple input and output wireless communication system, characterized in that it can vary depending on Doppler frequency or use a fixed K.

Images