KR20120116165A - Signal detection device and method for multiple input multiple output system using channel coding - Google Patents

Abstract

The present invention provides a signal detection apparatus and method for a multi-input multi-output system using channel coding, and more particularly, in a multi-input multi-output system, signals received through a plurality of transmit antennas are received through a plurality of receive antennas. Relates to an apparatus and a method for detecting

Description

SIGNAL DETECTION DEVICE AND METHOD FOR MULTIPLE INPUT MULTIPLE OUTPUT SYSTEM USING CHANNEL CODING}

The following embodiments are signal detection apparatus and method for a multi-input multiple output system using channel coding, and more particularly, in a multi-input multiple output system, signals transmitted through multiple transmit antennas are received through multiple receive antennas. It relates to a device and a method for receiving and detecting.

In wireless communications, multiple input multiple output systems are not only more efficient in frequency than single input single output systems but also increase channel capacity without increasing bandwidth or transmit power. have. Therefore, multiple input multiple output systems are considered as one of the important technologies in next generation mobile communication.

For example, a multiple input multiple output system is a V-BLAST (vertical-Bell Laboratories layered space-time) system that is relatively simple to design and use.

The V-BLAST system separates data to be transmitted into one or more pieces of data, and transmits each piece of separated data through the separated number of transmit antennas. As such, when using multiple transmit antennas, interference between transmitted signals may occur. Therefore, in a multiple input multiple output system, it may be important to design a detection device to mitigate such interference.

Conventional detection apparatuses have a method of detecting signals according to a maximum likelihood (ML) criterion. In this case, in a system with many antennas or a large signal constellation, the calculation amount may be large.

In order to solve this problem, conventionally, instead of considering all possible combinations of symbol vectors for transmission, the conventional hard decision demodulation method such as SD and QRD-M is extended, and only candidate symbols with high probability are transmitted as candidate lists (candidate). There are studies on sub-optimal detectors obtained by approximating LLR values of sign bits.

One such study is the QRD-M detection algorithm.

This algorithm QR-decomposes a channel matrix and uses a tree structure to descend from the root one by one, leaving nodes with a short node metric length for each layer, and down to the first layer. This is a method of calculating the LLR value of a sign bit using a node with a short node at.

The QRD-M detection algorithm is advantageous in that it is easy to implement in hardware and has a small amount of computation, but a large amount of computation may be required to achieve near ML performance.

According to an embodiment of the present invention, there is provided an MQRD-M-DFE detection method that is improved to reduce the complexity of the QRD-M detection algorithm and uses a partial decision decision equalizer (DFE).

In addition, according to an embodiment of the present invention, there is provided a signal detection method for a multi-input multiple output system using channel coding that can effectively detect a received signal in a multi-input multiple output system using channel coding.

Signal detection method according to an embodiment of the present invention in the multi-input multiple output system using a channel encoding, after receiving a signal transmitted through a plurality of transmit antennas through a plurality of receive antennas, and then detecting the received signal A method comprising: QR-decomposing a channel matrix H of the received signal, calculating an LLR value for each code bit of the received signal based on a tree structure using the QR decomposition, and calculating Determining the LLR value as an input of a channel decoder for channel encoding.

In particular, the step of calculating the LLR for each code bit is to first look at the best branch in the tree structure, and then create a candidate list by adding all branches to the parent node of the node with high reliability in the tree structure. The LLR value for each sign bit is calculated based on the generated candidate list.

According to an embodiment of the present invention, in a multi-input multiple output system, the bit error rate performance may be close to the ML performance and the required amount of calculation may be small. In addition, according to the embodiment of the present invention, since the calculation amount may be constant independent of the channel environment and the noise intensity, it may be easy to implement an actual system.

In particular, an embodiment of the present invention, in the detection algorithm, when making a search through the tree structure, only the best branches from the nodes are made one by one, and the length of the nodes is lowered, compared to other detection methods. The reliability of nodes can be taken into account. Thus, by joining all branches from only the node of the node with high confidence, it is possible to select a more accurate candidate list than before.

In addition, an embodiment of the present invention can reduce unnecessary branch noise by using a smart branch removal technique.

1 is a diagram schematically illustrating a configuration of a multi-input multi-output system to which a signal detection method for a multi-input multi-output system using channel coding according to an embodiment of the present invention can be applied.
FIG. 2 is a diagram for describing an algorithm capable of operating in the multi-input multiple output detector 111 shown in FIG. 1.
3 is a diagram illustrating a signal detection method for a multi-input multiple output system using channel coding according to an embodiment of the present invention.
4 illustrates an example of a signal constellation according to a signal detection method for a multi-input multi-output system using channel coding according to an embodiment of the present invention.
5A to 5C are diagrams illustrating signal constellations according to a signal detection method for a multi-input multiple output system using channel coding according to another embodiment of the present invention.
6 is a view for explaining the operation of the signal detection method for a multiple input multiple output system using channel coding according to another embodiment of the present invention.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

The multi-input multi-channel system to which the signal detection method according to an embodiment of the present invention can be applied can provide a transmission rate close to channel capacity with high reliability by using channel coding.

In other words, the multi-input multi-output system can reduce the bit error increase due to attenuation, interference, noise, etc. that a signal may experience in a wireless channel through channel coding. In general, when decoding a channel coded bit stream, it is possible to use a soft decision value rather than a hard decision value of coded bits. An example describes a signal detection method for calculating the LLR value for soft decision.

Therefore, in a multi-input multiple-output system, if one wants to accurately calculate the log-likelihood ratio (LLR) value, which is the soft decision value of the sign bit, based on the existing ML, the signal detector may consider all possible combinations of symbol vectors transmitted from the antenna. There is a need. However, in this case, the amount of calculation may increase exponentially in proportion to the number of transmitting antennas and the size of the signal constellation.

The present invention proposes an effectively improved MQRD-M-DFE detector to reduce the complexity of the QRD-M detector.

First, the MQRD-M (modified QRD-M) detector examines the tree structure in the same way as the conventional QRD-M detector, which calculates the LLR value of the sign bit with all nodes in the first layer. Therefore, the MQRD-M detector performs better than the QRD-M detector because it has more nodes than the QRD-M detector and calculates the LLR value.However, the wrong information may be used to calculate the LLR value. As a result, performance may be degraded.

To solve this, the MQRD-M detector uses DFE. MQRD-M-DFE detectors using DFE do not discard nodes that are cut from layer to layer, but use the nodes to obtain DFE solutions and connect them downstairs. Accordingly, the MQRD-M-DFE detector can obtain more nodes in the first layer than the QRD-M detector, and thus can calculate a more accurate LLR.

In addition, the MQRD-M-DFE detector can determine the reliability of nodes by descending through the tree structure, making the nodes of the best branches from each node, and by measuring the length of the nodes. The MQRD-M-DFE detector selects a node having a short node length as a node having high reliability, and generates a candidate list for LLR calculation by connecting all branches only at the parent node of the node having high reliability.

1 is a diagram schematically illustrating a configuration of a multi-input multi-output system to which a signal detection method for a multi-input multi-output system using channel encoding according to an embodiment of the present invention can be applied.

개의 송신 안테나,

개의 수신 안테나를 사용하고, 송신기(100) 및 수신기(110)를 포함한다.Referring to Figure 1, the multi-input multiple output system

Transmit antennas,

Two receive antennas, and include a transmitter 100 and a receiver 110.

The transmitter 100 passes the binary data bits u provided from the binary information source 101 to the channel encoder 102, obtains the code bit stream c through the channel encoder 102, and obtains the obtained code bit stream c. Signal to the constellation / antenna counter 103.

비트씩 분리하고, 각각의 송신 안테나마다 동일한 직교진폭변조(quadrature amplitude modulation: QAM) 별자리 집합 Q를 기초로 상기 분리된 스트림을 신호 별자리 심볼로 대응시킨다. 그리고, 신호 별자리/안테나 대응기(103)는 상기 신호 별자리 심볼로 대응된 각 스트림들을

개의 송신 안테나를 통해 수신기(110)로 송신한다.The signal constellation / antenna corresponder 103 generates a code bit stream c.

Bits are separated and the separated streams are mapped to signal constellation symbols based on the same quadrature amplitude modulation (QAM) constellation set Q for each transmit antenna. In addition, the signal constellation / antenna corresponder 103 selects each stream corresponding to the signal constellation symbol.

Transmits to the receiver 110 through two transmit antennas.

는 송신 안테나마다 심볼로 대응하는 비트 수를 나타낸다. 그리고, 수신기(110)의 j번째 수신 안테나에서 받은 복소 신호를

라 하면,

개의 수신 안테나로 수신된 신호 벡터는 로 나타낼 수 있다. 여기서, 위첨자 T는 벡터 전치를 나타낸다.here,

Denotes the number of bits corresponding to a symbol for each transmit antenna. Then, the complex signal received from the j-th reception antenna of the receiver 110

Say,

The signal vector received by the two receiving antennas may be represented by. Here, the superscript T represents the vector transpose.

개의 수신 안테나를 통해 수신된 신호를 다중입력 다중출력 검파기(111)로 전달한다.On the other hand, the receiver 110

The signal received through the two receiving antennas is transmitted to the multi-input multiple output detector 111.

The multiple input multiple output detector 111 calculates L, which is an LLR value for each code bit stream c, from the received signal vector y, and transfers the calculated L to the channel decoder 112.

를 출력한다. The channel decoder 112 decodes the binary data bits based on L, which is an LLR value for each code bit stream c.

.

Accordingly, the receiver 110 may estimate the bit stream transmitted from the transmitter 100 through the output decoding value.

To describe the operation of the multi-input multi-output detector 111 in detail, the multi-input multi-output detector 111 may define the received signal vector y in one symbol interval as a discrete time baseband model as shown in Equation 1 below. have.

[Equation 1]

는 구성원의 평균이 0이고, 분산이 1이며 독립이고 분포가 같은 복소 정규 확률변수인

채널 행렬을 나타낸다.

는 송신 심볼 벡터를 나타내고,

는 평균이 0이고 분산이

이며 독립이고 분포가 같은 복소 정규 확률변수들이 이루는 벡터를 나타낸다. 여기서, 수신기(110)는 이러한 채널 행렬들을 미리 알고 있는 것으로 가정한다.here,

Is a complex normal random variable with a mean of 0, a variance of 1, and independent and equal distributions

Represents a channel matrix.

Denotes a transmit symbol vector,

Is 0 and the variance is

Denotes a vector of complex normal random variables that are independent and have the same distribution. Here, it is assumed that the receiver 110 knows these channel matrices in advance.

The multiple input multiple output detector 111 calculates a Log-Likelihood Ratio (LLR) value of a sign bit using the received signal vector y. In particular, when the multi-input multi-output detector 111 calculates the LLR based on the existing ML, the LLR may be calculated as in Equation 2 below.

&Quot; (2) "

이고,

인 경우,

는 i번째 심볼

로 대응되는 b번째 비트를 나타낸다. 또한

는 송신 안테나 심볼 벡터 s가 가질 수 있는 모든 심볼 벡터들의 집합 S에서,

(a=0,1)인 벡터들의 집합을 나타낸다. here,

ego,

Quot;

Is the i symbol

Indicates the b-th bit corresponding to. Also

Is a set S of all symbol vectors that the transmit antenna symbol vector s may have,

It represents a set of vectors with (a = 0,1).

Since the calculation of the LLR through Equation 2 may be somewhat complicated, the multi-input multiple output detector 111 may use max-log approximation, as shown in Equation 3 below.

&Quot; (3) "

을 계산하기 위해서는 QR 분해를 이용한 나무 얼개 구조를 많이 이용한다.here,

In order to calculate, we use a lot of tree structure using QR decomposition.

인 m x m 단위 행렬(unitary matrix) 이고,

는 m x m 위쪽 삼각 행렬(upper triangular matrix)이 된다. 따라서, 수학식 1의 양변에

를 곱하면, 아래의 수학식 4를 얻을 수 있다.That is, if we decompose the channel matrix into H = QR, Q is

Mxm unitary matrix,

Is an m-by-m upper triangular matrix. Therefore, on both sides of Equation 1

By multiplying, the following Equation 4 can be obtained.

&Quot; (4) "

는

이고,

는 잡음 성분

이다. here,

The

ego,

Is noise component

to be.

는 수학식 1의 잡음 성분 v와 통계적인 특성이 같을 수 있으므로,

가 될 수 있다. 따라서, 다중입력 다중출력 검파기(111)는

대신 위쪽 삼각 행렬 R을 이용하여,

를 나무 얼개로 계산할 수 있다.On the other hand, since Q is a unit matrix, the noise component of Equation 4

Since the noise component v of Equation 1 may have the same statistical characteristics,

. Therefore, the multi-input multi-output detector 111

Instead, using the upper triangular matrix R,

Can be calculated as a tree trunk.

를 생성하여, LLR 값을 아래의 수학식 5와 같이 추정할 수 있다.Here, in order to accurately obtain the result of Equation 3, a large amount of computation may be necessary because the set S of all possible antenna symbol vectors must be considered. Accordingly, the multiple input multiple output detector 111 has a small number of candidate lists.

The LLR value can be estimated by Equation 5 below.

[Equation 5]

이다.here,

to be.

이고, 한 마디에서 아래층으로 이을 수 있는 가지 수가 |Q|인 나무 얼개 구조를 고려할 수 있다.In order to obtain Equation 5 as a tree structure,

We can consider a tree structure with | Q | which is the number of branches that can go from one word downstairs.

째 층을 말한다. 그리고, 이 나무 얼개 구조의 (k+1)번째 층과 k번째 층 사이에 있는 가지(branch)는 송신 신호 벡터 s의 k번째 원소

를 나타내고, 상기 나무 얼개 구조의 마디(node)는 해당 마디와 뿌리를 잇는 가지들이 이루는 벡터를 나타낸다. 또한, k번째 층에 있는 l번째 마디를

차원 심볼 벡터의

로 나타낸다면, k는

이고, l은

이고,

는 상기 나무 얼개 구조의 뿌리를 나타낸다.Here, the root is the highest layer

Say the second layer. The branch between the (k + 1) th and kth layers of this tree structure is the kth element of the transmission signal vector s.

The node of the tree structure represents a vector of branches connecting the node and the root. Also, the lth node on the kth layer

Dimensional symbol of vector

Where k is

And l is

ego,

Represents the root of the tree structure.

과 그 부모 마디 사이에 있는 가지 길이는 branch metric로서, 아래의 수학식 6과 같이 산출할 수 있고, 마디

의 마디 길이는 node metric로서, 아래의 수학식 7과 같이 산출할 수 있다.Where

The length of the branch between and its parent node is a branch metric, which can be calculated as shown in Equation 6 below.

Node length is a node metric, and can be calculated as shown in Equation 7 below.

&Quot; (6) "

[Equation 7]

=2이고, 가지 수가 |Q|=4인 나무 얼개 구조를 보여준다.2 is a view for explaining an algorithm that can operate in the multi-input multiple output detector 111 shown in FIG.

= 2 and tree branch structure with | Q | = 4.

Referring to Fig. 2,? Represents the root of the tree structure, solid line represents the branches of the tree structure, and? Represents the nodes of the tree structure. The numbers next to each solid line indicate the length of each branch, and the numbers next to each node indicate the length of each node. In this case, the length of a node may be equal to the sum of the roots and the lengths of the branches connecting the node.

In the structure below, it is assumed that a single root exists and that there is no connection between the branches, and each tree is connected from the root to the structure of the tree structure.

In addition, the multi-input multi-output detector 111 adopts a breadth-first search method according to an embodiment of the present invention, and descends one by one from the root through the following steps. Look at the structure of the structure.

3 is a view for explaining a signal detection method for a multi-input multi-output system using channel coding according to an embodiment of the present invention, in which the multi-input multi-output detector 111 looks at the tree structure and Shows the process of calculating the LLR value for.

,

를 입력받는다. 그리고, 다중입력 다중출력 검파기(111)는 상기 입력을 기초로, H를 QR 분해하고,

를 계산한다. 여기서,

인 경우,

는

인 정수이고,

인 경우,

는

및

에 만족하는 정수이다.In step 310, the multi-input multiple output detector 111, through the receiving antenna, H, y,

,

Get input. The multi-input multi-output detector 111 performs QR decomposition on the basis of the input,

. here,

Quot;

The

Is an integer

Quot;

The

And

An integer satisfying.

층에 새로 생긴 마디들의 마디 길이를 계산한다. 여기서, i는

-1로 설정한다.In step 320, the multi-input multi-output detector 111 connects all the branches that can be rooted from the root,

Calculate the node length of new nodes on the floor. Where i is

Set to -1.

개의 마디들의 가지들 중 가장 좋은 가지를 선택하고, 상기 선택된 가지에 대한 마디 길이를 산출한다.In step 330,

The best branch among the branches of the nodes is selected, and the node length for the selected branch is calculated.

개의 마디를 선택한다. 그리고, 다중입력 다중출력 검파기(111)는 상기

개의 마디들 중 가장 좋은 가지(best branch)를 아래의 수학식 8과 같이 산출할 수 있다.In other words, the multi-input multiple output detector 111 has a shorter node length among the nodes in the (i + 1) th layer.

Select nodes And, the multiple input multiple output detector 111 is the

The best branch of the nodes can be calculated as Equation 8 below.

[Equation 8]

는

마디들 중 하나를 나타내고,

가 된다. 그리고, D(a)는 신호 별자리 집합 Q에서 a에 가까운 심볼을 이어주는 심볼 역대응 함수를 나타낸다. Here, the best branch refers to the branch having the shortest length among the branches that can be connected in one node.

The

Represents one of the nodes,

. In addition, D (a) represents a symbol inverse corresponding function that connects a symbol close to a in the signal constellation set Q.

를 잇고, 상기 이음에 의한 새로운 마디인

를 생성하고, 상기 생성된 마디에 대한 마디 길이를 아래의 수학식 9와 같이 산출할 수 있다.Then, the multi-input multi-output detector 111 is the next best branch and

Joining, the new node by the joint

And generate a node length for the generated node as shown in Equation 9 below.

&Quot; (9) "

개의 마디에 대응하는 형제 마디(sibling node)를 잇는다. 여기서, 형제 마디는 부모 마디가 같은 마디들을 말한다. In step 340, the multi-input multiple output detector 111 is

Connect sibling nodes corresponding to nodes. Here, brother nodes refer to nodes whose parents are the same.

개의 마디들을 선택한다. 다중입력 다중출력 검파기(111)는 상기 선택된

개의 마디들과, 상기

개의 마디들 각각의 형제 마디들을 모두 잇고, 상기 이음에 의해 생성되는 새로운 마디

의 마디 길이를 산출할 수 있다. 상기 새로운 마디

의 마디 길이는 아래의 수학식 10과 같이 산출될 수 있다.Therefore, the multi-input multi-output detector 111 has a shorter node length among the nodes generated in step 330.

Select nodes Multiple input multiple output detector 111 is selected

Nodes and above

A new node created by joining all of the sibling nodes of each of the two nodes

The node length of can be calculated. Said new node

The node length may be calculated as in Equation 10 below.

&Quot; (10) "

는

의 부모 마디를 나타낸다.here,

The

Indicates the parent node of.

In operation 350, the multi-input multi-output detector 111 determines the end of the algorithm according to whether i is 1.

As a result of the determination, if i is greater than 1, the multi-input multiple output detector 111 assumes that the termination is not the end, sets i = i-1, and returns to step 2.

As a result of the determination, when i is 1, the multi-input multi-output detector 111 determines that the algorithm is finished. In operation 360, the multi-input multi-output detector 111 selects the nodes in the first layer as the candidate list L, and obtains the LLR value of the sign bit using Equation 5.

개만 이으면서도, 상기 한 마디에서 모든 가지를 잇는 경우와 동일한 비트 오류율의 성능을 달성할 수 있다.Further, the multi-input multi-output detector 111 may use a smart branch elimination technique when connecting sibling nodes in the second layer (i = 2) in operation 340 to further reduce the amount of calculation. When the multi-input multiple output detector 111 uses the smart branch elimination scheme, the multi-input multi-output detector 111 branches from a word in the second layer to the first layer.

Even with only dogs, it is possible to achieve the same bit error rate performance as when connecting all branches in one word.

In addition, the signal constellation / antenna counter 103 of the transmitter 100 may use quadrature phase shift keying (QPSK) according to an embodiment of the present invention. In this case, the signal constellation may be expressed as shown in FIG. 4.

Referring to FIG. 4, the branches 10, 00, 11, and 01 of the second layer are shown.

First, it can be seen that the transformed received signal 410 and the signal constellation point closest to the corrected signal in the first layer are 10 420. In addition, it can be seen that 10 (420) is the same as the best branch that can be reached in a word in the second layer by Equation (8).

에 가까운 3개의 점만 이용하고, 두번째 층에서는 한 마디에 3개의 가지를 잇도록 구현될 수 있다.
On the other hand, another branch 01 (430) is farther than the other branches 00 (440), 11 (450) from the corrected signal. However, since the bits representing the other branch 01 430 partially overlap with the bits representing the other branches 00 440 and 11 (450), the multi-input multiple output detector 111 has the another branch. LLR values of sign bits may be calculated using only the bits representing the 00 440 and the 11 450. Therefore, the multi-input multi-output detector 111 is the fixed signal in the signal constellation according to an embodiment of the present invention.

Only three points close to can be used, and the second layer can be implemented to connect three branches in a word.

5A to 5C illustrate signal constellations according to a signal detection method for a multi-input multi-output system using channel coding according to another embodiment of the present invention, and the signal constellation / antenna counter 103 of the transmitter 100 is illustrated. ) Shows the case of using 16QAM. That is, FIGS. 5A to 5C are divided into three cases according to the positions of the signal constellation points closest to the fixed signal.

FIG. 5A is the case where the nearest signal constellation point is at the corner of the signal constellation (510), FIG. 5B is the case where the closest signal constellation point is at an edge other than the corner of the signal constellation (520), and FIG. 5C is the The closest signal constellation point is centered in the signal constellation (530).

First, it can be seen that the positions of other close signal constellations (black points) differ according to the positions of the closest signal constellation points (gray points). In addition, since the multi-input multiple output detector 111 can know information about all bits only by the closest signal constellation point and the other close signal constellations, there will be no problem in calculating the LLR value of the sign bit. That is, in the case of 16QAM, the multi-input multi-output detector can obtain the same LLR value as in the case of using all signal points by using only five signal points.

Similarly, in the case of 64QAM, seven signal points may be used in the signal constellation according to the position of the nearest signal constellation point. That is, the multi-input multi-output detector 111 may calculate an LLR value using at least one or more other close signals in addition to the signal closest to the fixed signal.

이고, QPSK를 사용하는 시스템에서, 이고,

인 경우의 검파 과정을 설명한다.6 is a diagram illustrating an operation of a signal detection method for a multiple input multiple output system using channel encoding according to another embodiment of the present invention. The wooden structure shown in Figure 6 is

In a system using QPSK,

The detection process in case of is explained.

In particular, referring to FIG. 6, the numbers written in each node indicate the order in which each node is made. However, these orders may be different when using the smart branch removal technique.

A general method for examining the amount of calculation of the multi-input multiple output detector 111 is to count the multiplication amount. In the following, the multiplication amount used in the detection method of the multi-input multiple output detector 111 according to an embodiment of the present invention is calculated.

번의 곱셈이 사용될 수 있고,

를 산출하는 데에는 약

번의 곱셈이 사용될 수 있다.First, the step multiplication of the multi-input multiple output detector 111 is described.

Times multiplication can be used,

To calculate

Times multiplication can be used.

In step 2, 4 | Q | multiplications can be used to connect all possible branches from the root and to calculate the node length of the new nodes.

개의 마디들에서 가장 좋은 가지를 얻는데에는 수학식 8의 경우,

번의 곱셈이 사용될 수 있고, 새로운 마디의 마디 길이를 산출하는 데에는 수학식 9의 경우,

번의 곱셈이 사용될 수 있다.In step 3,

In order to get the best branch from the nodes of Equation 8,

The multiplication of times can be used, and in order to calculate the node length of a new node,

Times multiplication can be used.

개의 마디에 대응하는 형제 마디들을 잇고 상기 형제 마디들의 마디 길이를 계산하는 데에는, 수학식 10의 경우,

번의 곱셈이 사용될 수 있다. 만약, 단계 4가 스마트 가지 제거 기법을 이용한다면, 사용되는 곱셈의 수는

가 될 수 있다.In step 4,

In order to connect the node nodes corresponding to the four nodes and calculate the node lengths of the brother nodes, in Equation 10,

Times multiplication can be used. If step 4 uses the smart branch elimination technique, the number of multiplications used is

.

Subsequently, the process of calculating the LLR value of the sign bit in step 5 assumes that the same number of multiplication numbers are used in most detectors. Therefore, the detection method using the smart branch removal technique may use the following multiplication amount considering the preprocessing step.

[Equation 12]

As a result, the detection method of the multi-input multiple output detector 111 according to an embodiment of the present invention may take a smaller amount of calculation than the ML detector. In addition, when the multi-input multiple output detector 111 uses the smart branch elimination technique, the calculation amount may be smaller than that of the (M) QRD-M detector and the (M) QRD-M-DFE detector.

In summary, the detection apparatus of the signal detection method for a multi-input multiple output system using null coding according to an embodiment of the present invention performs QR decomposition on the channel matrix H of the received signal, and uses a tree decomposition structure using the QR decomposition. The LLR value is calculated for each code bit of the received signal, and the calculated LLR value is determined as an input of a channel decoder for the channel encoding.

In particular, in order to calculate the LLR for each sign bit, the detector first looks at the best branch in the tree structure first, and then adds all the branches to the parent node of the node with high reliability in the tree structure. A candidate list is generated and an LLR value for each sign bit is calculated based on the generated candidate list.

층에 새로 생긴 마디들의 마디 길이를 계산하고, 상기 계산 결과를 기초로 마디 길이가 짧은 복수의 마디들을 선택하고, 상기 선택된 마디들 각각에서 가지를 이어 새로운 마디들을 생성한 후 상기 생성된 마디들 중 마디 길이가 짧은 다른 복수의 마디들을 선택하고, 상기 선택된 다른 복수의 마디들 각각에 대응하는 형제 마디들을 잇는다. 상기 검파 장치는 이러한 과정을 반복하여 이어진 가지들을 기초로, 상기 후보 리스트를 생성하여 각 부호 비트에 대한 LLR 값을 계산한다.In more detail, the detection device follows all branches that may be at the root of the tree structure.

Calculate the node lengths of the new nodes in the layer, select a plurality of nodes with short node lengths based on the calculation result, generate new nodes by branching from each of the selected nodes, and then, among the generated nodes. A plurality of nodes having a shorter node length are selected, and sibling nodes corresponding to each of the selected plurality of other nodes are connected. The detection apparatus calculates the LLR value for each code bit by generating the candidate list on the basis of the following branches by repeating this process.

층에 새로 생긴 마디들의 마디 길이를 계산시, 수학식 7을 사용할 수 있다. 또한, 상기 검파 장치는

마디에서 가지를 하나씩 이을 때 수학식 8을 사용할 수 있다.Here, the detection device

Equation (7) may be used to calculate the node lengths of nodes newly formed in the layer. In addition, the detection device is

Equation 8 can be used when joining branches one by one.

마디에서 생성된 새로운 마디의 마디 길이는 수학식 9를 이용하여 산출할 수 있다. 또한, 상기 검파 장치는

마디의 형제 마디들의 길이는 수학식 10을 이용하여 산출할 수 있다.The detection device is

The node length of the new node generated from the node can be calculated using Equation 9. In addition, the detection device is

The length of sibling nodes of a node can be calculated using Equation 10.

The detection device may apply a smart removal technique, or may obtain a fixed signal by using Equation (11).

The methods according to embodiments of the present invention may be implemented in the form of program instructions that can be executed through various computer means and recorded in a computer-readable medium. The computer readable medium may include program instructions, data files, data structures, etc. alone or in combination. The program instructions recorded on the medium may be those specially designed and constructed for the present invention or may be available to those skilled in the art of computer software.

As described above, the present invention has been described by way of limited embodiments and drawings, but the present invention is not limited to the above embodiments, and those skilled in the art to which the present invention pertains various modifications and variations from such descriptions. This is possible.

Therefore, the scope of the present invention should not be limited to the described embodiments, but should be determined by the equivalents of the claims, as well as the claims.

Claims (1)

In a multi-input multiple output system using channel coding, after receiving a signal transmitted through a plurality of transmit antennas through a plurality of receive antennas, in the method for detecting the received signal,
QR decomposing the channel matrix H of the received signal;
Calculating an LLR value for each code bit of the received signal based on the tree structure using the QR decomposition; And
Determining the calculated LLR value as an input of a channel decoder for channel encoding,
Computing the LLR for each sign bit
First of all, the best branches in the tree structure are first examined, and then a candidate list is generated after all the branches of the parent node of the node with high reliability in the tree structure, and the respective code bits are generated based on the generated candidate list. To calculate the LLR value for
Signal Detection Method for Multiple Input Multiple Output System Using Channel Coding.

Images