CN109167648B

CN109167648B - Candidate constellation point set generation method and MIMO spatial multiplexing detection method

Info

Publication number: CN109167648B
Application number: CN201810873601.1A
Authority: CN
Inventors: 朱嘉俊; 程绍刚
Original assignee: Xradiotech Technology Co ltd
Current assignee: Shenzhen Quanzhi online Co.,Ltd.
Priority date: 2018-08-02
Filing date: 2018-08-02
Publication date: 2020-12-25
Anticipated expiration: 2038-08-02
Also published as: CN109167648A

Abstract

The invention provides a method for generating a candidate constellation point set, which comprises the following steps: (1) putting a hard decision result into the candidate constellation point set; (2) finding out a real part which is opposite to the bit of the corresponding position of the real part of the hard decision result and is closest to the real part of the hard decision result, and combining the real part with the imaginary part of the hard decision result to put the real part and the imaginary part into the candidate constellation point set; and/or (3) finding the imaginary part which is opposite to the bit of the position corresponding to the imaginary part of the hard judgment result and is closest to the imaginary part of the hard judgment result, and combining the imaginary part with the real part of the hard judgment result to be put into the candidate constellation point set. The invention forms the candidate constellation point set of each space stream through the hard decision symbol obtained by pre-detection, has simple generation method, can effectively overcome unreliable search range and LLR deficiency, reduces the operation amount and ensures excellent performance.

Description

Candidate constellation point set generation method and MIMO spatial multiplexing detection method

Technical Field

The present invention relates to the field of wireless communication technologies, and in particular, to a method for generating a candidate constellation point set and a method and an apparatus for detecting in a multiple-input multiple-output (MIMO) system.

Background

Nowadays, wireless communication technology, multiple-output (MIMO) multi-antenna transmission technology, is widely used, especially spatial multiplexing (spatial multiplexing) technology. Spatial multiplexing is used to increase the amount of data transmitted in the same time and frequency band by using the spatial dimension, thereby enhancing the throughput of wireless communication. However, due to the multipath channel, the spatial stream data interfere with each other at the receiving end, so it is necessary to extract or recover the signal from the spatial stream interference by using a Detection technique (Detection).

Currently, Detection technologies are roughly classified into Linear Detection (Linear Detection), Non-Linear Detection (Non-Linear Detection), Maximum Likelihood Detection (MLD), Near Maximum Likelihood Detection (Near Maximum Likelihood Detection), and the like. The main representative methods of linear detection are Zero-Forcing (ZF) detection and Minimum mean-squared error (MMSE) detection. The linear detection technology is to shape an inverse matrix similar to a multipath channel matrix, counteract the multipath channel matrix and descramble the spatial stream. The linear detection technology has the advantage of low implementation difficulty, but the performance is poor. The nonlinear detection technique is based on ZF and MMSE, and performs hierarchical Interference Cancellation (spatial Interference Cancellation), which slightly improves performance and complexity. The maximum likelihood detection is a high-performance detection technology, which can enable the system to obtain the best error rate performance, but the complexity is extremely high, and the ergodic search is often difficult to realize or cannot be realized in the actual system due to the Nondeterministic Polynomial (NP) operation complexity. The approximate maximum likelihood detection sacrifices proper performance loss, replaces lower complexity and is beneficial to engineering realization. One of the more widely studied Detection techniques is Sphere decoding Detection (Sphere Detection). However, the approximate maximum likelihood detection technology still generally has the problems of high complexity, serious performance loss, missing or inaccurate log-likelihood ratio llr (likelihood rate) in combination with channel coding, and the like, which cause the degradation of decoding performance.

Disclosure of Invention

Therefore, the invention provides a new simplified maximum likelihood detection technology, which can not only overcome the problem of performance reduction caused by LLR deficiency or inaccuracy, but also effectively reduce the complexity and the time delay, and is suitable for being realized on an ASIC. Most importantly, the performance loss is slight compared to the maximum likelihood detection performance.

In one aspect, the present invention provides a method for generating a candidate constellation point set, including the steps of: (1) putting a hard decision result into the candidate constellation point set; (2) finding out a real part which is opposite to the bit of the corresponding position of the real part of the hard decision result and is closest to the real part of the hard decision result, and combining the real part with the imaginary part of the hard decision result to put the real part and the imaginary part into the candidate constellation point set; and/or (3) finding the imaginary part which is opposite to the bit of the position corresponding to the imaginary part of the hard judgment result and is closest to the imaginary part of the hard judgment result, and combining the imaginary part with the real part of the hard judgment result to be put into the candidate constellation point set.

In addition, in order to increase the constellation points in the candidate constellation point set, the method may further include the steps of: (4) in addition to the constellation points obtained in the steps (1), (2) and/or (3), finding the P constellation points closest to the hard decision result, and putting the P constellation points into the candidate constellation point set; where P < (the real part mapping bit number + the imaginary part mapping bit number) ^ 2.

In addition, the method for adding the constellation points may also be: wherein the set of real part candidate real numbers is generated in step (2); wherein said step (3) produces a set of imaginary candidate real numbers; then the method further comprises the steps of: (4) and combining all elements or part of elements in the real part candidate real number set and all elements or part of elements in the imaginary part candidate real number set into the candidate constellation point set.

In another aspect, the present invention provides a method for generating a summary of a set of candidate real numbers of a constellation diagram, where a fission step is performed on each real number in the constellation diagram: finding a real part/imaginary part which is opposite to the position corresponding to the real number and is closest to the real number; generating a candidate real number set of real numbers for the fission of one real number in the constellation diagram, and summarizing the candidate real number sets generated for all real numbers in the constellation diagram.

Preferably, in each candidate real number set, M candidate real number elements closest to the real number are marked, where M < the number of elements of the real number set.

On the other hand, the invention provides another method for generating a candidate constellation point set, and for each hard decision result, the following steps are executed: (1) putting the hard decision result into the candidate constellation point set; (2) finding two candidate real number sets corresponding to the hard decision real part and the imaginary part from the collection of the candidate real number sets, namely a hard decision real part candidate real number set and a hard decision imaginary part candidate real number set; (3) combining the real part candidate real number set of the hard decision with the imaginary part of the hard decision, and putting the generated constellation point into the candidate constellation point set; and/or combining the real part of the hard decision with the imaginary part candidate real number set of the hard decision, and putting the generated constellation point into the candidate constellation point set.

Preferably, in order to increase the constellation points, all elements or partial elements in the real part candidate real number set and all elements or partial elements in the imaginary part candidate real number set are combined and put into the candidate constellation point set.

Preferably, in each set of candidate real numbers, M elements of the candidate real numbers closest to the real number are marked, where M < the number of elements of the set of real numbers.

Preferably, in order to increase constellation points, (4) putting constellation points obtained by combining M marked candidate elements in the hard-decision real part candidate real number set and M marked candidate elements in the hard-decision imaginary part candidate real number set into the candidate constellation point set.

In another aspect, the present invention further provides a method for detecting MIMO spatial multiplexing, including: step 01, performing pre-detection to obtain a pre-detection symbol

l∈{1,2,...,N_T}，N_TThe number of the spatial streams is sent; step 02, hard decision is carried out on the pre-detected symbols, and the most similar constellation points are found out, namely hard decision results

l∈{1,2,...,N_T}，N_TThe number of the spatial streams is sent; step 03, according to each space stream symbol

Hard decision result of

According to the generation of the candidate constellation point setThe method generates a candidate constellation point set of each spatial stream; step 04, calculating l for each candidate constellation point element according to the candidate constellation point set of each spatial stream₂Norm distance and update the minimum l of "0" and "1" of each bit₂-norm distance; step 05, minimum l of "0" and "1" according to each bit₂Norm distance, calculating the LLR value for each bit.

In another aspect, the present invention further provides a MIMO spatial multiplexing detection apparatus, including: a candidate real number set summary part configured to generate a candidate real number set summary according to the constellation diagram; a pre-detection part configured to perform pre-detection to obtain pre-detection symbol

l∈{1,2,...,N_T}，N_TThe number of the spatial streams is sent; a hard decision part configured to perform hard decision on the pre-detected symbol to find a closest constellation point, i.e., a hard decision result

l∈{1,2,...,N_T}，N_TThe number of the spatial streams is sent; generating a candidate constellation point set portion configured to generate a symbol from each spatial stream

Hard decision result of

Generating a candidate constellation point set of each spatial stream according to the collection part of the candidate real number set; l₂-a norm calculation portion configured to calculate/for each candidate constellation point element from a set of candidate constellation points for each spatial stream₂Norm distance and update the minimum l of "0" and "1" of each bit₂-norm distance; an LLR calculation section configured to calculate a minimum l of "0" and "1" according to each bit₂Norm distance, calculating the LLR value for each bit.

Drawings

For a more complete understanding of the present invention, preferred embodiments thereof are described in detail below with reference to the accompanying drawings.

Fig. 1 is a block diagram of a typical MIMO wireless system using spatial multiplexing.

Fig. 2 is a schematic diagram of a MIMO detection unit according to an embodiment of the present invention.

Fig. 3 is a flow chart of a MIMO detection method corresponding to fig. 2.

Fig. 4-11 show gray mapping constellations using 256QAM modulation; wherein fig. 4-7 show a fission combining (candidate constellation point set) process for a first spatial stream and fig. 8-11 show a fission combining (candidate constellation point set) process for a second spatial stream.

FIG. 12 shows₂-norm distance calculation and LLR calculation procedures.

Fig. 13 shows a set of candidate real numbers for the real part and a set of candidate real numbers for the imaginary part of a first spatial stream generated according to a fission rule.

Fig. 14-17 detail how to generate the set of candidate constellation points for the first spatial stream by fig. 13.

Fig. 18 shows a set of candidate real numbers for the real part and a set of candidate real numbers for the imaginary part of the second spatial stream generated according to the fission rule.

Fig. 19-22 detail how to generate the set of candidate constellation points for the second spatial stream by fig. 18.

Fig. 23 shows a summary of the set of candidate real numbers modulated by 256 QAM.

Fig. 24 shows an example of a set of candidate real numbers based on a hard decision result for a first spatial stream according to an embodiment of the present invention.

Fig. 25 shows an example of a set of candidate real numbers based on a hard decision result for the second spatial stream according to an embodiment of the present invention.

Fig. 26 shows an example of a set of candidate real numbers based on a hard decision result for a third spatial stream according to an embodiment of the present invention.

Fig. 27 is a schematic diagram of a MIMO detection unit according to another embodiment of the present invention.

Fig. 28 is a flow chart illustrating a MIMO detection method corresponding to fig. 27.

Fig. 29 is a diagram showing a summary of the real number candidate sets for BPSK modulation.

Fig. 30 shows a summary of the set of candidate real numbers for the QPSK modulation scheme.

Fig. 31 shows a summary of the candidate real number set for the 16QAM modulation scheme.

Fig. 32 shows a summary of the candidate real number set for the 64QAM modulation scheme.

Detailed Description

Fig. 1 is a block diagram of a typical MIMO wireless system using spatial multiplexing. The transmitter 100 includes a channel coding module 102, a serial-to-parallel conversion module 103, and an orthogonal modulator 104. The channel coding module 102 inputs bits a_i,kThe bit b is coded, and the serial-to-parallel conversion module 103 converts the coded bit b_i,kGrouping to obtain a bit set { c) required by each transmitting antenna_1,i,k}、{c_2,i,k}、…、{c_Nt,i,kA quadrature modulator 104 modulates the set of bits into a modulation symbol s_1,k、s_2,k、…、s_Nt,kThrough an antenna A_t1、A_t2、…、A_tNtThe formed transmitting antenna array is transmitted. The quadrature modulator 104 performs symbol modulation such that the in-phase/quadrature (I/Q) portion of each constellation point is modulated with two carrier signals having a phase of 90 degrees. Examples of quadrature modulation schemes are Quadrature Amplitude Modulation (QAM), Quadrature Phase Shift Keying (QPSK), Quadrature Amplitude Shift Keying (QASK), etc.

Assume that the multipath channel is Flat Fading (Flat Fading), the Noise is White gaussian Noise (Guassian White Noise), and the number of antennas (spatial streams) at the transmitting end is N_TThe number of receiving end antennas is N_RThe channel information is accurately estimated; the received signal model may be represented using the following formula:

Y_k＝H_kS_k+Z_k

wherein, y_i,kA k-th symbol representing an i-th receiving antenna; s_l,kThe kth symbol represented as the l spatial stream; h is_i,l,kRepresenting a channel value of a kth symbol of a ith spatial stream for an ith receiving antenna; z is a radical of_i,kA noise value of a k-th symbol for an i-th receiving antenna; the index k is the time dimension, meaning that the processes are repeated in time. The subscript k in the formula can also be deleted, and the formula after the subscript k is deleted represents a received signal model at a certain moment, so that the accuracy of the formula is not influenced.

At the receiver 101, from the received signal y_1,k、y_2,k、…、y_NR,kThe first step in estimating the input bits is to obtain the log-likelihood ratio { LLR (c) } for each bit in the set of bits for each transmit antenna by MIMO detection unit 105_1,i,k)}、{LLR(c_2,i,k)}、…、{LLR(c_Nt,i,k)}. Then merged by the parallel-to-serial conversion module 106 to obtain { b'_i,kAnd performs Forward Error Correction (FEC) through the channel decoding module 107 to obtain an estimated value of input bits { a'_i,k}. The detection process at the MIMO detection unit 105 determines the complexity and performance of the MIMO system to some extent. The Maximum Likelihood Detection (MLD) is to search all constellation points in a traversal mode, the operation complexity is extremely high, and the MLD is often difficult to realize or cannot be realized in an actual system. The present invention provides a simplified maximum likelihood detection technique.

Fig. 2 is a schematic diagram of a MIMO detection unit according to an embodiment of the present invention. As shown in fig. 2, MIMO detection unit 200 of the present invention includes six parts: a pre-detection part 201, a hard decision part 202, a generation candidate constellation point set part 203, an L2-norm calculation part 204, an LLR calculation part 205, and a constellation map part 206. Please note that the constellation diagram portion 206 is not necessarily included in the MIMO detection unit 200, and the constellation diagram portion 206 may also be located in other elements of the receiving end, such as a channel estimation unit, a phase correction unit, etc., for providing information for MIMO detection, and the present invention does not limit that the constellation diagram portion 206 is located in the MIMO detection unit 200.

Fig. 3 is a flowchart illustrating a MIMO detection method corresponding to fig. 2 according to an embodiment of the present invention. As shown in fig. 2 and 3, the method for detecting MIMO symbols according to the embodiment of the present invention includes the following steps.

Step 301: the preliminary detection is performed in the preliminary detection section 201.

In this step, any MIMO detection method may be used to perform pre-detection, for example, by using simple linear detection or nonlinear detection-supported detection techniques to extract or recover symbols in advance. As to which technique is adopted as the pre-detection method, the selection can be made according to the design specification requirements.

Step 302: in the hard decision section 202, hard decision is performed on the pre-detected symbols to find out the closest constellation point, i.e., a hard decision result.

In this step, the hard decision determines the most likely transmitted symbol (e.g., the transmitted symbol with the smallest euclidean distance) according to the result of the pre-detection. Taking BPSK modulation as an example, the transmission set is { -1, +1}, and due to the influence of noise, according to the detection method of step 301, it is possible to obtain that the detection value of the first symbol of the first spatial stream is

At this time, since only { -1, +1} two possible symbols are transmitted, it is considered that the transmitted signal is more likely to be +1, i.e., the hard decision result is + 1.

Step 303: in the candidate constellation point set generating portion 203, according to the hard decision constellation points of each spatial stream symbol, based on the constellation diagram portion 206, a respective required candidate constellation point set is generated.

This step may be considered a fission process. After the hard decision of each spatial stream, a symbol closest to the pre-detection result is found, and then a candidate constellation point set is generated based on the hard decision constellation points according to the fission rule of the present invention. The fission rules of the present invention are as follows:

1. putting the hard decision result of the step 302 into the candidate constellation point set;

2. finding out a real part which is opposite to the bit of the corresponding position of the real part of the hard decision result and is closest to the real part of the hard decision result (the generated set is called a candidate real number set of the real part), and combining the real part with the imaginary part of the hard decision result to put the real part and the candidate constellation point set;

3. and finding out the imaginary part which is opposite to the bit of the position corresponding to the imaginary part of the hard judgment result and is closest to the imaginary part of the hard judgment result (the generated set is called as a candidate real number set of the imaginary part), and combining the imaginary part and the real part of the hard judgment result to be put into the candidate constellation point set.

4. The candidate constellation points are added on the basis of the

rules

1,2 and 3, the number and the rules can be customized, and the rule adopted in the invention is 4 constellation points which are nearest to the hard judgment result and are combined by a real part/imaginary part real number set except the constellation points contained in the

rules

1,2 and 3.

The details of the fission process will be described in detail below.

Furthermore, although the fission rules/processes described above are described in a sequential order, the processes may operate in other orders. In other words, the sequence or order of steps described herein is not necessarily the order in which the steps are performed. The processes and steps described herein may be performed in virtually any order. In addition, some steps may also be performed simultaneously.

For a better understanding of the invention, in the following preferred embodiments, the fission process is

rule

1,2, 3, 4 inclusive. It should be understood, however, that rule 2, rule 3 in the fission rules/procedures described above are relationships of "2 and/or 3," that is, rule 2 may be included without rule 3 in one embodiment; in another embodiment, rule 3 may be included without rule 2; in yet another embodiment, rule 2 and rule 3 may be included. Further, the order of execution of

rules

2 and 3 does not necessarily require that the steps be performed in that order. May be performed in a practical order or may be performed synchronously.

Rule 4 may also be omitted.

Step 304: in l₂A norm calculation section 204 that calculates an L2 norm distance for each candidate constellation point from the set of candidate constellation points for each spatial stream, and updates the minimum L of "0" and "1" for each bit₂The norm distance.

In this step, each spatial stream independently calculates l in parallel or in series according to its own set of candidate constellation points₂The norm distance. The parallel computation means that after the candidate constellation point set is determined, the subsequent computation supports parallel development in implementation and is independent of and independent of influence. Of course, serialization is also possible, depending on whether time-first or resource-first (serialization implies less computational resources and greater delay), but some simplified algorithms can in principle only employ serialization. Here, |₂Norm refers to the square of the euclidean distance.

Step 305: in the LLR calculating section 206, an LLR value of each bit is calculated from the minimum L2-norm distance of "0" and "1" of each bit, and finally output to the channel decoding block after parallel-to-serial conversion.

In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in detail below with reference to specific embodiments. The embodiment takes a two-layer MIMO system and 256QAM as an example, and explains the simplified maximum likelihood detection method provided by the present invention.

In one embodiment, a spatial multiplexing MIMO system includes a 2 x 2 antenna array configuration. Assuming that a multipath channel is Flat Fading (Flat Fading), Noise is White gaussian Noise (Guassian White Noise), a transmitting end is two spatial streams, a receiving end is two antennas, channel information is accurately estimated, and a received signal model can be expressed by the following formula:

Y_k＝H_kS_k+Z_k

wherein, y_i,kA k-th symbol representing an i-th receiving antenna; s_l,kThe kth symbol represented as the l spatial stream; h is_i,l,kRepresenting a channel value of a kth symbol of a ith spatial stream for an ith receiving antenna; z is a radical of_i,kA noise value of a k-th symbol for an i-th receiving antenna; the subscript k is the time dimension.

Pre-detection step 301:

at this step, any MIMO detection method can be used to perform pre-detection, for example, by pre-extracting or restoring symbols using simple linear detection, non-linear detection, or detection techniques supported by the lattice-reduction LLL algorithm. As to which technique is adopted as the pre-detection method, the selection can be made according to the design specification requirements. In this embodiment, zero forcing Detection (ZF Detection) is used for example, and can be expressed by the following formula:

wherein the content of the first and second substances,

the kth detected symbol represented as the l spatial stream; w_kA zero-forcing detection matrix represented as the kth symbol; h_kA MIMO channel matrix representing a kth symbol; [. the]^-1Represents the inverse of a matrix or vector; [. the]^HRepresenting the conjugate transpose of a matrix or vector.

Hypothesis pre-detection

Is-14.5-j12.7; in the second spatial stream, pre-detection is assumed

Is + 12.5-j.0.7.

Hard decision step 302:

and carrying out hard decision on the pre-detected symbols, and finding out the closest constellation point, namely a hard decision result.

In this embodiment, in the first spatial stream, pre-detection is performed

Is-14.5-j12.7; in the second spatial stream, pre-detection

Is +12.5-j · 0.7, then the hard decision result is the symbol of the first spatial stream:

and symbols of the second spatial stream:

wherein the content of the first and second substances,

expressed as the real part of the kth hard decision symbol of the l spatial stream;

represented as the imaginary part of the kth hard decision symbol of the ith spatial stream.

Generating a candidate constellation point set step 303:

as mentioned above, this step may be considered a fission process. After the above hard decision step, based on the constellation point of the hard decision result, a candidate constellation point set is generated from the constellation diagram according to the fission rule of the present invention, so the fission rule is also called a candidate constellation point set rule. The fission rule is as follows:

1. putting the hard decision result of the step 302 into a candidate constellation point set;

4. The candidate constellation points are added on the basis of the

rules

1,2 and 3, the number and the rules can be customized, and the rules adopted in the invention are 4 constellation points which are closest to the hard judgment result and are combined by real part/imaginary part real number sets except the constellation points contained in the

rules

1,2 and 3. Rule 4, which is used by way of example in the present invention, can be implemented as follows: and two real numbers which are closest to the real part of the hard decision result in the candidate real number set of the real part are used as the real part, two real numbers which are closest to the imaginary part of the hard decision result in the candidate real number set of the imaginary part are used as the imaginary part, and the two real numbers are arranged and combined to form 4 constellation points.

The fission rule of the present invention is explained in detail by specific examples next.

In this embodiment, there are two spatial streams, using 256QAM modulation. Fig. 4-11 show gray mapping constellations, where for 256QAM modulation, 4 bits for each of the real and imaginary parts, that is, 4-bit gray codes for both the real and imaginary parts, one constellation point is represented by an 8-bit gray code, the first 4-bit gray code represents the real part, the last 4-bit gray code represents the imaginary part, the horizontal axis is the real part I, and the vertical axis is the imaginary part Q.

Fig. 4-7 show a fission combining (candidate constellation point set) process for a first spatial stream. Fig. 8-11 show the fission combining (candidate constellation point set) process for the second spatial stream.

As described above, the hard decision result for the first spatial stream is

And according to the rule 1, putting the hard decision result of-15-j.13 into the candidate constellation point set to obtain { -15-j.13 }. As shown in fig. 4, the constellation points-15-j · 13 of the hard decision result are indicated by bold line boxes in fig. 4.

According to rule 2, the real part of the hard decision result is-15, -15 corresponds to 0000 (gray code), and the real part bit is found to be xxx1, while the real part closest to-15 is-13; find bit xx1x, while the real part nearest-15 is-11; finding the real bit as x1xx, while the real bit nearest to-15 is-7; find the real bit to be 1xxx while the real nearest to-15 is + 1. The four values and the imaginary part are combined to obtain four constellation points { -13-j.13, -11-j.13, -7-j.13, + 1-j.13 }. As shown in fig. 5, four constellation points obtained according to rule 2 are indicated by thin line boxes.

Where the set of real numbers { -13, -11, -7, +1} generated by rule 2 is referred to as the set of candidate real numbers for the real parts. And (4) placing two real numbers-13 and-11 which are closest to the real part-15 of the hard decision result at the forefront of the candidate real number set, namely the first two candidate real number elements in the candidate real number set. The role of this set of real part candidates will be described in detail below.

According to rule 3, the imaginary part of the hard decision result is-13, the gray code corresponding to-13 is 0001, the imaginary part bit is found to be xxx0, and the imaginary part closest to-13 is found to be-15; find the imaginary bit as xx1x, while the imaginary bit closest to-13 is-11; find the imaginary bit as x1xx, while the imaginary part closest to-13 is-7; the imaginary bit is found to be 1xxx while the imaginary nearest to-13 is found to be + 1. The four values and the real part are combined to obtain the other four constellation points which are { -15-j.15, -15-j.11, -15-j.7, -15+ j.1 }. As shown in fig. 6, four constellation points obtained according to rule 3 are indicated by a two-line box.

The set of real numbers { -15, -11, -7, +1} generated by rule 3 is referred to as the set of candidate real numbers for the imaginary part. And (4) placing two real numbers-15 and-11 closest to the imaginary part-13 of the hard decision result at the forefront of the candidate real number set, namely the first two candidate real number elements in the candidate real number set. The role of this set of candidate real numbers for the imaginary part will be described in detail below.

According to rule 4, knowing the hard decision result-15-j.13, four constellation points { -13-j.15, -13-j.11, -11-j.15, -11-j.11 } closest to the hard decision result-15-j.13 are added, usually 4 constellation points diagonally opposite to the hard decision constellation point. As shown in fig. 7, four constellation points obtained according to rule 4 are indicated by dashed boxes.

Thus, a candidate constellation point set xi of the first spatial stream is obtained₁I.e. by

ξ₁＝{-15-j·13 -13-j·13 -11-j·13 -7-j·13 +1-j·13 -15-j·15 -15-j·11 -15-j·7 -15+j·1 -13-j·15 -13-j·11 -11-j·15 -11-j·11}

The fission rules of the present invention are also applied on the second spatial stream hard decision result, as described below.

As described above, the hard decision result of the second spatial stream is

And according to the rule 1, putting the hard decision result of + 13-j.1 into the candidate constellation point set to obtain { + 13-j.1 }. As shown in fig. 8, the constellation point +13-j · 1 of the hard decision result is indicated by a bold-line box in fig. 8.

According to rule 2, the real part of the hard decision result is +13, 1001 (gray code) corresponds to +13, and the real part bit is found to be xxx0, while the real part nearest to +13 is found to be + 15; find bit xx1x, while the real part nearest +13 is + 11; finding the real part bit as x1xx while the real part nearest to +13 is + 7; the real bit is found to be 0xxx while the real bit nearest to +13 is found to be-1. The four values and the imaginary part are combined to obtain four constellation points { + 15-j.1, + 11-j.1, + 7-j.1, -1-j.1 }. As shown in fig. 9, four constellation points obtained according to rule 2 are indicated by thin line boxes.

Wherein the set of real numbers { +15, +11, +7, -1} generated by rule 2 is referred to as a set of candidate real numbers for real parts. As described above, the two real numbers +15, +11 closest to the real part +13 of the hard decision result are placed at the forefront of the candidate real number set, i.e., the first two candidate real number elements in the candidate real number set. The role of this set of real part candidates will be described in detail below.

According to rule 3, the imaginary part of the hard decision result is-1, the gray code corresponding to-1 is 0100, the bit of the imaginary part is xxx1, and the imaginary part closest to-1 is-3; find the imaginary bit as xx1x, while the imaginary part closest to-1 is-5; finding the imaginary bit as x0xx with the imaginary part closest to-1 as-9; the imaginary bit is found to be 1xxx while the imaginary nearest to-1 is found to be + 1. The four values are combined with the real part to obtain the other four constellation points { + 13-j.3, + 13-j.5, + 13-j.9, +13+ j.1 }. As shown in fig. 10, four constellation points obtained according to rule 3 are indicated by a two-line box.

Wherein the set of real numbers { -3, +1, -5, -9} generated by rule 3 is referred to as the set of candidate real numbers for the imaginary part. As described above, the two real numbers-3, +1 closest to the imaginary part-1 of the hard decision result are placed at the forefront of the candidate real number set, i.e., the first two candidate real number elements in the candidate real number set. The role of this set of candidate real numbers for the imaginary part will be described in detail below.

According to rule 4, knowing the hard decision result +13-j · 1, four constellation points { +11-j · 3, +11+ j · 1, +15-j · 3, +15+ j · 1} closest to the hard decision result +13-j · 1 are added, usually 4 constellation points diagonally opposite to the hard decision constellation point. As shown in fig. 11, four constellation points obtained according to rule 4 are indicated by dashed boxes.

Thus, a candidate constellation point set xi of the second space stream is obtained₂I.e. by

Finally, a candidate constellation point set xi of the first spatial stream is obtained₁And a second set ξ of spatial stream candidate constellation points₂。

It should be noted that the candidate constellation point set is generated according to the inverse rule of each bit corresponding to the hard decision result. For example, if the first bit of the hard decision result is "0", the first candidate constellation point is the constellation point whose first bit is "1" and is closest to the hard decision constellation point. If the hard decision second bit is "1", then the second candidate constellation point is the second bit is "0" and closest to the hard decision constellation point. By analogy, if the current modulation order is 256QAM (8 bits map one symbol), and rules 1,2, and 3 are generated according to the constellation diagram, each spatial stream will generate 1+8 candidate constellation points, and then a plurality of candidate constellation points are added (rule 4). The constellation mapping principle used in this example is a Gray code (Gray code) constellation point mapping principle, and 4 constellation points of diagonal hard decision constellation points are also added to the candidate constellation point set. Therefore, each spatial stream will yield 13 candidate constellation points for a total of 26 candidate constellation points.

₂l-norm distance calculation step 304:

according to respective candidate constellation point sets, each space flow carries out l in parallel₂Norm distance calculation and updating the minimum l of "0" and "1" of each bit₂The norm distance.

Substituting into the nth candidate constellation point set element of the first spatial stream

The method of step 301 re-estimates the pre-detection value corresponding to the second spatial stream

Namely, it is

Wherein the content of the first and second substances,

according to the pre-detected value

The method re-estimates the hard decision value corresponding to the second spatial stream in step 302

Then, l is calculated using the following formula₂-norm：

Wherein the content of the first and second substances,

an nth candidate constellation point element of a kth symbol of the first spatial stream;

to substitute for

Hard decision values of a kth symbol of the second spatial stream of (1); l is_1,n,kSubstituting the calculated l for the nth candidate constellation point element of the kth symbol of the first spatial stream₂-norm。

Simultaneously, the candidate constellation point set elements of the second space flow are substituted in parallel

Re-estimating the corresponding value of the first spatial stream by the method described in

steps

301 and 302

Then, l is calculated using the following formula₂-norm：

Wherein the content of the first and second substances,

an nth candidate constellation point element of a kth symbol of the second spatial stream;

to substitute for

A corresponding value of a kth symbol of the first spatial stream of (a); l is_2,n,kSubstituting the calculated l for the nth candidate constellation point element of the kth symbol of the second spatial stream₂-norm。

The above-described 256QAM modulation example, ξ, for a 2 × 2 dual-antenna spatial multiplexing MIMO system₁And xi₂There are 26 candidate constellation point elements in total, i.e. 26 l will be output₂Norm value to comparator 1201 as shown in fig. 12.

The above is for l of two spatial streams₂Norm calculation method, for N, analogously_TL of a spatial stream₂Norm calculation, which can be summarized as follows:

for each candidate constellation point set element of each spatial stream l

Substituting the following formula:

the above formula is regarded as that the number of the transmission space streams is N_T-1 MIMO system, pre-detecting other spatial streams by the method of step 301:

wherein the content of the first and second substances,

according to the pre-detected value

Re-estimating the hard decision values corresponding to other spatial streams via step 302

p∈{1,2,...,N_TAnd p ≠ l; then, l is calculated using the following formula₂-norm：

In this step, the transmitted symbols of other spatial streams are obtained by adopting a pre-detection and hard decision mode, and the computation amount is greatly reduced compared with the transmitted symbols of other spatial streams obtained by a search detection method.

LLR calculation step 305:

l calculated in step 304₂The norm value is compared and updated to the register of bit position "0" or "1" corresponding to the corresponding candidate constellation point element.

Minimum l of "0" and "1" according to each bit₂The norm distance, calculating the LLR value of each bit, and finally outputting to a Channel Decoder (Channel Decoder) module.

Still as an example of 256QAM modulation for a 2 x 2 dual antenna spatial multiplexing MIMO system as described above, there are a total of 16 bits and their corresponding two spatial streams in a slotThere is a "0" minimum distance and a "1" minimum distance per bit for the LLR values of (1). Therefore, there are 32 minimum distances. The LLR is calculated as the "0" minimum distance minus the "1" minimum distance for the corresponding bit. The whole₂The norm distance calculation and LLR calculation process is shown in fig. 12.

The simplified maximum likelihood detection method of the present invention is described in detail above by using an example of a two-layer MIMO system and a modulation scheme of 256 QAM. In describing step 303 of generating a set of candidate constellation points, constellation diagrams 4-11 are used for explanation.

Except that the constellation diagrams (such as fig. 4-11) are used to describe the candidate constellation point set xi of the first and second spatial streams₁、ξ₂May also use the fission map approach as shown in fig. 13-23 to describe the candidate constellation point set ξ₁、ξ₂The generation process of (1).

Fig. 13 shows a set of candidate real numbers for the real part and a set of candidate real numbers for the imaginary part of the first spatial stream generated according to the fission rule described above. Wherein the first fission diagram I (also called umbrella diagram) represents the set of real number candidates { -13, -11, -7, +1} for real part-15, and the second fission diagram Q (umbrella diagram) represents the set of real number candidates { -15, -11, -7, +1} for imaginary part-13. The umbrella top of the fission diagram I represents the real part-15 of the hard decision result, the umbrella ribs represent fission, and the umbrella bottom/umbrella edge represents the fission result { -13, -11, -7, +1} of the hard decision result. The first two gray heavy ribs particularly indicate the first two real number candidate elements of the set of real number candidates, which are used for the application of rule 4 described above, and will be described in detail below. Similarly, the umbrella top of the fission diagram Q represents the imaginary part-13 of the hard decision result, the umbrella ribs represent the fission, and the umbrella bottom/umbrella edge represents the fission result { -15, -11, -7, +1} of the hard decision result. The first two gray heavy ribs then indicate in particular the first two real number candidate elements of the set of real number candidates, which are to be used for the application of rule 4 above, as will be described in more detail below.

How to generate the candidate constellation point set ξ of the first spatial stream by fragmenting the graph 13 is described in detail below in connection with fig. 14-17₁. Note that, in making the candidate constellation point combination,the first umbrella I in fig. 14-17 is the real part and the second umbrella Q is the imaginary part. Wherein the black bold line represents the combining process of the real and imaginary parts.

As shown in fig. 14, 1 candidate constellation point is generated according to rule 1, i.e., hard decision result { -15-j · 13 }.

As shown in fig. 15, the combination is performed according to rule 2, i.e. the set of real candidates for the real part of hard decision is combined with the imaginary part of hard decision to generate 4 constellation points { -13-j · 13, -11-j · 13, -7-j · 13, +1-j · 13 }.

As shown in fig. 16, the combination is performed according to rule 3, i.e. the hard decision real part and the hard decision imaginary part are combined to generate 4 candidate constellation points { -15-j · 15, -15-j · 11, -15-j · 7, -15+ j · 1 }.

As shown in fig. 17, the combination is performed according to rule 4, that is, the first two elements of the real number candidate set of hard decision real parts and the first two elements of the real number candidate set of hard decision imaginary parts are combined pairwise with each other to generate 4 constellation point candidates { -13-j · 15, -13-j · 11, -11-j · 15, -11-j · 11 }.

Thus, a set ξ of constellation point candidates for the first spatial stream is generated₁。

Similarly, fig. 18 shows a set of candidate real numbers for the real part and a set of candidate real numbers for the imaginary part of the second spatial stream generated according to the fission rule described above. Wherein the first fission diagram I (also called umbrella diagram) represents the set of real number candidates { +15, +11, +7, -1} for real part +13, and the second fission diagram Q (umbrella diagram) represents the set of real number candidates { -3, -5, -9, +1} for imaginary part-1. The umbrella top of the fission diagram I represents a real part of a hard decision result +13, the umbrella ribs represent fission, and the umbrella bottom/umbrella edge represents a fission result { +15, +11, +7, -1} of the hard decision result. The first two gray ribs particularly indicate the first two real number candidate elements of the set of real number candidates, which are used for the application of rule 4 above. Similarly, the umbrella top of the fission diagram Q represents the imaginary part-1 of the hard decision result, the umbrella ribs represent the fission, and the umbrella bottom/umbrella edge represents the fission result { -3, +1, -5, -9} of the hard decision result. The first two gray ribs indicate in particular the first two real number candidate elements of the set of real number candidates, which are used for the application of rule 4 above.

How to generate the candidate constellation point set ξ of the second spatial stream by cracking the graph 18 is described in detail below in conjunction with fig. 19-22₂. Note that, when candidate constellation point combination is performed, the first umbrella I in fig. 19 to 22 is a real part, and the second umbrella Q is an imaginary part. Wherein the black bold line represents the combining process of the real and imaginary parts.

As shown in fig. 19, according to rule 1, i.e. the hard decision result, 1 candidate constellation point { +13-j · 1} is generated.

As shown in fig. 20, the combination is performed according to rule 2, i.e. the set of real candidate numbers of the real hard decision part is combined with the imaginary hard decision part to generate 4 constellation point candidates { +11-j · 1, +15-j · 1, +7-j · 1, -1-j · 1 }.

As shown in fig. 21, the combination is performed according to rule 3, that is, the hard decision real part and the hard decision imaginary part are combined to generate 4 candidate constellation points { +13-j · 3, +13+ j · 1, +13-j · 5, +13-j · 9 }.

As shown in fig. 22, the combination is performed according to rule 4, that is, the first two elements of the set of real candidate numbers of the real hard decision part and the first two elements of the set of real candidate numbers of the imaginary hard decision part are combined pairwise with each other to generate 4 candidate constellation points { +11-j · 3, +11+ j · 1, +15-j · 3, +15+ j · 1 }.

Thus, a set ξ of constellation point candidates for the second spatial stream is generated₂。

In summary, the process of generating a set of candidate constellation points by cracking/umbrella can be summarized as: the top layer of the fission/umbrella-shaped graph is a hard judgment result, the bottom layer is a fission result (also called a candidate real number set), a real part hard judgment result and an imaginary part hard judgment result are combined (rule 1), the candidate real number set and the imaginary part hard judgment result of the real part hard judgment result are combined (rule 2), the candidate real number set and the real part hard judgment result of the imaginary part hard judgment result are combined (rule 3), and the fission results connected by gray bold lines are subjected to internal permutation and combination (rule 4).

The inventors have further found that if the first half of the

fission rules

2, 3 are performed for a real number (whether it is the real or imaginary part): the fission results are the same if the real/imaginary parts closest to the real/imaginary parts of the hard decision result (the resulting set is called a candidate real set of real/imaginary parts) are found with opposite bits from the corresponding positions of the real/imaginary parts of the hard decision result.

For example in constellation diagram 4, if the first half of fission rule 2 is performed for one real number-11 (0011) (i.e., assuming-11 is the hard decision result real part), the fission result (i.e., the real part candidate real number set) { -13(0001), -9(0010), -7(0110), +1(1100) } may be obtained. If the first half of fission rule 3 is performed for real number-11 (0011) (i.e., assuming-11 is the hard decision result imaginary part), then the same fission result (i.e., the imaginary candidate real number set) { -13 (0001)), -9(0010), -7(0110), +1 (1100)) } results.

As such, the first half of

rule

2 or 3 may be performed over all real numbers in constellation diagram 4: and finding the real part or the imaginary part which is opposite to the position corresponding to the real number and is nearest to the real number. And summarizing the candidate real number sets generated by respectively fissioning all real numbers in the constellation map to obtain a candidate real number set summary.

Since the real part and the imaginary part of 256QAM are independent and the bit mapping rules are the same, there are 16 possibilities for the real part and the imaginary part, and the candidate real number sets corresponding to the 16 possibilities can all be represented according to the same principle, such as the 256QAM candidate real number sets shown in fig. 23 are summarized, and there are 16 fission/umbrella maps in total.

Fig. 23 is a set of all possible candidate real numbers of the constellation diagram 4, where each independent fission diagram (i.e., each umbrella diagram) represents a set of candidate real numbers of real numbers. In other words, the top of all fission maps/umbelliform maps in fig. 23 can represent the information of the constellation, and the other information is the embodiment of the fission rule of the present invention.

Specifically, 256QAM carries 8 bits per symbol, with 4 bits in the real part and 4 bits in the imaginary part, so that there are 16 possibilities in the real part, 16 possibilities in the imaginary part, and the real and imaginary parts are independent, meaning that all possibilities are 16 × 16 — 256. The real and imaginary parts are just two latitudes, only values need to be concerned in the fission process, so that fig. 23 has only 16 subgraphs, which is applicable to both the real and imaginary parts.

When the method is applied, an umbrella-shaped graph (candidate real number set) corresponding to a real part and an imaginary part can be found in the graph 23 directly according to a hard decision result of each spatial stream, and the umbrella-shaped graph is combined according to a rule to obtain a candidate constellation point set, so that a constellation representation method is separated.

For example, if the hard decision result for the first spatial stream is +5-j · 7, then the fission maps corresponding to real part +5 and imaginary part-7 are found directly in fig. 23, constituting the hard decision result based candidate real set instance for the first spatial stream as shown in fig. 24. If the hard decision result of the second spatial stream is-9 + j · 9, then the fission maps corresponding to the real part-9 and the imaginary part-9 are directly found in fig. 23, and a candidate real number set example based on the hard decision result of the second spatial stream as shown in fig. 25 is constructed. If the hard decision result of the third spatial stream is-1 + j · 3, then the fission maps corresponding to the real part-1 and the imaginary part +3 are directly found in fig. 23, and a candidate real number set example based on the hard decision result of the third spatial stream is formed as shown in fig. 26. The candidate constellation point sets for each spatial stream are then generated according to the combination rules described above (e.g., fig. 14-17).

Therefore, after the constellation bit mapping relationship is determined, the graph/table (as shown in fig. 23) determined according to the first half of the

rule

2 or 3 is fixed, and when the constellation bit mapping relationship is implemented, the graph/table (table) is directly looked up to obtain a candidate real number set, and then the candidate real number set is combined to obtain a candidate constellation point set without calculating and finding again in the constellation according to the rule.

As described above, fig. 23 can also be expressed in the form of a table, and fig. 23 is converted to the table shown in table 1 below (only the cleavage map of the first row of fig. 23 is shown).

TABLE 1 tabular representation of FIG. 23

In summary, the overall fission rule can be considered to be divided into a fission part (first half of rules 2, 3) and a combination part (second half of

rules

2, 3 and rule 4). The fission is performed on all real numbers in the constellation diagram, and the set of candidate real numbers of all real numbers is summarized (fig. 23/table 1) and stored. The collection of candidate real numbers can be calculated at the time of hardware initialization and stored in a register of a correlation unit, and the receiving of the receiver is not really started; in the present invention, it is referred to as "offline calculation". Then, for each hard decision result, a table look-up manner can be used to directly find a corresponding candidate real number set (see fig. 13 and 18) from the collection of candidate real number sets, and then the combination is performed according to the combination part (see fig. 15-17 and 20-22) to obtain a candidate constellation point set.

A MIMO detection unit and a MIMO detection method flow according to another embodiment of the present invention are described below with reference to fig. 27 and 28. Fig. 27 is a schematic diagram of a MIMO detection unit according to another embodiment of the present invention. Fig. 28 is a flow chart illustrating a MIMO detection method corresponding to fig. 27. Compared with the above-described embodiments of fig. 2 and 3, the MIMO detection unit and MIMO detection method of this embodiment may have different flows.

In this embodiment, MIMO detection unit 2700 of the present invention includes seven parts: the system comprises a pre-detection part 201, a hard decision part 202, a candidate constellation point set generation part 2703, an L2-norm calculation part 204, an LLR calculation part 205, a constellation diagram part 206 and a candidate real number set summary part 2707. Compared with fig. 2, fig. 27 has one more candidate real number set summary portion 2707 between constellation diagram 206 and generation candidate constellation point set portion 2703; and the candidate constellation point set generation portion 2703 is also different from the candidate constellation point set generation portion 203 in fig. 2, which will be described below. The rest is the same and the same reference numerals are used.

As described above, the constellation diagram portion 206 is not necessarily included in the MIMO detection unit 2700, and the constellation diagram portion 206 may also be located in other components of the receiving end, such as a channel estimation unit, a phase correction unit, and the like, for providing information for MIMO detection.

Fig. 28 differs from fig. 3 in the difference between step 2803 in fig. 28 and step 303 in fig. 3.

In step 301 in fig. 3, the generate candidate constellation point set portion 203 finds a candidate constellation point set from the constellation diagram 206 according to the fission rule based on the hard decision result. This fission process involves calculations and is performed once for each hard decision result.

In step 2803 in fig. 28, the candidate constellation point set generation part 2703 finds out a corresponding candidate real number set from the candidate real number set summarization part 2707 based on the hard decision result, and combines the candidate constellation point set according to the combination part rule. This combining process does not involve calculations and is performed only once for each hard decision result. The candidate real set summary portion 2707 is computed from the constellation diagram 206, computing only 16 real candidate sets (for 256QAM), and this computation is one-time, and the computation results (i.e., the candidate real set summary) are stored for each hard decision result at a later time.

Thus, in the embodiments of fig. 2 and 3, the constellation is stored, with a computational effort for each detection process. In the embodiment of fig. 27 and 28, the summary of the candidate real number set is stored, each detection process only involves table lookup and combination, and no calculation amount is needed, and the calculation amount for obtaining the summary of the candidate real number set is almost negligible (one-time calculation) compared with a large number of detection processes.

Similarly, other modulation schemes can be summarized as the real candidate set, as shown in fig. 29-32. Fig. 29 is a diagram showing a summary of a set of candidate real numbers for BPSK modulation; fig. 30 is a diagram showing a summary of a set of candidate real numbers for QPSK modulation; fig. 31 shows a summary of a set of candidate real numbers for a 16QAM modulation scheme; fig. 32 shows a summary of the candidate real number set for the 64QAM modulation scheme.

Therefore, according to the above description, the candidate set corresponding to each hard decision result in each modulation scheme (BPSK/QPSK/16QAM/64QAM/256QAM) can be simply found from the graphs (fig. 29, fig. 30, fig. 31, fig. 32, fig. 23).

The number of candidate constellation points generated by the hard decision result of each spatial stream is:

min (total number of constellation points, 1 (rule 1) + number of real part mapping bits (rule 2) + number of imaginary part mapping bits (rule 3) +4 (rule 4)).

Where min represents taking a smaller value because all constellation points have been covered after applying

rules

1,2, 3, and 4 for modulation schemes with fewer constellation points.

For example, for BPSK, min (2, 1+1+0+4) is 2, for QPSK, min (4, 1+1+1+4) is 4, for 16QAM, min (16, 1+2+2+4) is 9, for 64QAM, min (64, 1+3+3+4) is 11, and for 256QAM, min (256, 1+4+4+4) is 13.

The total candidate constellation point set is the sum of the number of candidate constellation points of each spatial stream, so for the present invention, adding one spatial stream increases the calculation at most 13 times, the algorithm complexity is linear growth relation (increasing n times) with the number of spatial streams, while most other algorithms are exponential growth relation (increasing n times).

The technical solution of the present invention is described by taking gray mapping as an example, but the present invention is not limited thereto, and the fission rule of the present invention is also applicable to non-gray mapping.

It should be noted that rule 4 in the present example is only a preferred embodiment, that is, the constellation points diagonally opposite to the hard decision result are put into the candidate constellation point set. The invention is not so limited. Rule 4 can be customized by adding the number of constellation points to the candidate constellation point set, for example, more than 4 constellation points can be added, for example, the real part candidate real number set and the imaginary part candidate real number set are combined with each other, but not limited to the combination of two real numbers closest to the real part/imaginary part of the hard decision result, that is, the thick gray connecting line is omitted in the fission diagram/umbrella diagram, and the thick gray connecting line is replaced by the thin connecting line. Or the combination of three real numbers closest to the real/imaginary parts of the hard decision result, etc. The constellation points may not be increased according to the combination of the real part candidate real number set and the imaginary part candidate real number set, for example, P points closest to the hard decision result may be increased, where P ═ is (real part mapping bit number + imaginary part mapping bit number) ^ 2; (ii) a Rule 4 may also be omitted.

The invention obtains the hard decision symbol shape through the pre-detectionAnd each space stream candidate set is formed, the generation method is simple, unreliable search range and LLR (log likelihood ratio) loss can be effectively overcome, and the calculation amount is reduced to ensure excellent performance. And candidate elements are decided without referring to the case of other spatial streams. Each spatial stream can be independently and parallelly calculated according to own candidate set₂Norm distance, independent and independent of influence. The method is beneficial to reducing complexity and delay, and is convenient for ASIC realization.

The invention changes the constellation diagram into the candidate real number set for summarizing, so that the generation of the candidate constellation point set is simple and convenient, the detection device has small time delay, low complexity and excellent performance. As the number of MIMO antennas increases, the complexity only grows approximately linearly.

Reference throughout this specification to "one embodiment" or "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of the present invention. The appearances of the phrase "in one embodiment" in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments.

It should be understood that the elements shown in the fig. may be implemented in various forms of hardware, software or combinations thereof. These elements may be implemented in a combination of hardware and software on one or more appropriately programmed general-purpose devices, which may include a processor, memory and input/output interfaces.

This specification describes the principles of the present invention. It will thus be appreciated that those skilled in the art will be able to devise various arrangements that, although not explicitly described or shown herein, embody the principles of the invention and are included within its spirit and scope.

Moreover, all statements herein reciting principles, aspects, and embodiments of the invention, as well as specific examples thereof, are intended to encompass both structural and functional equivalents thereof. Additionally, it is intended that such equivalents include both currently known equivalents as well as equivalents developed in the future, such as any elements capable of performing the same function, regardless of structure.

Thus, for example, it will be appreciated by those skilled in the art that the block diagrams presented herein represent conceptual views of systems and devices embodying the principles of the invention.

In the claims hereof, any element expressed as a means for performing a specified function is intended to encompass any way of performing that function including, for example, a) a combination of circuit elements that performs that function or b) software in any form, including, therefore, firmware, microcode or the like, combined with appropriate circuitry for executing that software to perform the function. The invention as defined by such claims resides in the fact that the functionalities provided by the various recited means are combined and brought together in the manner which the claims call for. Thus, any means that can provide those functionalities are deemed equivalent to those shown herein.

It should be understood that the above preferred embodiments are only for illustrating the technical solutions of the present invention, and not for limiting the same, and those skilled in the art can modify the technical solutions described in the above preferred embodiments or substitute some technical features thereof; and all such modifications and alterations are intended to fall within the scope of the appended claims.

Claims

1. A method for generating a candidate constellation point set, comprising the steps of:

(1) putting a hard decision result into the candidate constellation point set;

(2) finding out a real part which is opposite to the bit of the corresponding position of the real part of the hard decision result and is closest to the real part of the hard decision result, and combining the real part with the imaginary part of the hard decision result to put the real part and the imaginary part into the candidate constellation point set;

and/or (3) finding the imaginary part which is opposite to the bit of the position corresponding to the imaginary part of the hard judgment result and is closest to the imaginary part of the hard judgment result, and combining the imaginary part with the real part of the hard judgment result to be put into the candidate constellation point set.

2. The method of claim 1, further comprising:

(4) in addition to the constellation points obtained in the steps (1), (2) and/or (3), finding the P constellation points closest to the hard decision result, and putting the P constellation points into the candidate constellation point set; where P < (the real part mapping bit number + the imaginary part mapping bit number) ^ 2.

3. The method of claim 2, wherein P < ═ real part/imaginary part mapping bits number.

4. The method of claim 1, wherein:

wherein in the step (2), "the real part which is opposite to the bit of the corresponding position of the real part of the hard decision result and is nearest to the real part of the hard decision result" is found, and a real part candidate real part set is generated;

wherein in the step (3), "finding the imaginary part which is opposite to the bit of the position corresponding to the imaginary part of the hard decision result and is nearest to the imaginary part of the hard decision result", generating an imaginary part candidate real number set;

further comprising the steps of:

(4) and combining all elements in the real part candidate real number set and all elements in the imaginary part candidate real number set and putting the combined elements into the candidate constellation point set.

5. The method of claim 1, wherein:

further comprising the steps of:

(4) combining part elements in the real part candidate real number set and part elements in the imaginary part candidate real number set to obtain R constellation points closest to the hard judgment result, and putting the R constellation points in the candidate constellation point set;

wherein R < number of elements of the real part candidate real number set x number of elements of the imaginary part candidate real number set.

6. A method for detecting MIMO spatial multiplexing, comprising:

step 01: performing pre-detection to obtain pre-detected symbol

Where l is in the range of {1,2_T}，N_TThe number of the spatial streams is sent;

step 02: hard decision is carried out on the pre-detected symbols to find out the most similar constellation points, namely the hard decision result

step 03: according to each space stream symbol

Hard decision result of

The method according to any of claims 1-5, generating a set of candidate constellation points for each spatial stream;

step 04: according to the candidate constellation point set of each space stream, calculating l for each candidate constellation point element₂Norm distance and update the minimum l of "0" and "1" of each bit₂-norm distance;

step 05: minimum l of "0" and "1" according to each bit₂Norm distance, calculating the LLR value for each bit.

7. The method according to claim 6, wherein in step 04, each spatial stream independently calculates l in parallel or in series according to its own candidate constellation point set₂The norm distance.

8. Method according to claim 6 or 7, characterized in that in step 04 l₂-the calculation of norm distance comprises:

for each candidate constellation point set element of each spatial stream l

Substituting the following formula:

wherein, y_i,kA k-th symbol representing an i-th receiving antenna; s_l,kThe kth symbol represented as the l spatial stream;

an nth candidate constellation point element representing a kth symbol of the ith spatial stream; h is_i,l,kRepresenting a channel value of a kth symbol of a ith spatial stream for an ith receiving antenna; z is a radical of_i,kA noise value of a k-th symbol for an i-th receiving antenna; subscript k is the time dimension, and the number of transmit side antennas is N_TThe number of receiving end antennas is N_R；

The above formula is regarded as that the number of the transmission space streams is N_T-1, re-estimating hard decision values corresponding to other spatial streams by said steps 01 and 02

Wherein

To substitute for

Hard decision values of the kth symbol of the p-th spatial stream of (1); l is_l,n,kSubstituting the calculated l for the nth candidate constellation point element of the kth symbol of the l space stream₂-norm。

9. The method of claim 6, wherein said step 05 comprises: for the l₂Comparing the norm distance values, updating the norm distance values to a register of a bit position '0' or '1' corresponding to the corresponding candidate constellation point element, subtracting the '1' minimum distance of the corresponding bit from the '0' minimum distance of the corresponding bit, and calculating the LLR value of each bit.

10. An apparatus for detecting MIMO spatial multiplexing, comprising:

a pre-detection part configured to perform pre-detection to obtain pre-detection symbol

a hard decision part configured to perform hard decision on the pre-detected symbol to find a closest constellation point, i.e., a hard decision result

generating a candidate constellation point set portion configured to generate a symbol from each spatial stream

Hard decision result of

Any of claims 1 to 5The method of claim, generating a set of candidate constellation points for each spatial stream;

l₂-a norm calculation portion configured to calculate/for each candidate constellation point element from a set of candidate constellation points for each spatial stream₂Norm distance and update the minimum l of "0" and "1" of each bit₂-norm distance;

an LLR calculation section configured to calculate a minimum l of "0" and "1" according to each bit₂Norm distance, calculating the LLR value for each bit.

11. The apparatus of claim 10, wherein said/, is₂-the calculating step of the norm calculating section includes:

for each candidate constellation point set element of each spatial stream l

Substituting the following formula:

wherein, y_i,kA k-th symbol representing an i-th receiving antenna; s_l,kThe kth symbol represented as the l spatial stream; h is_i,l,kRepresenting a channel value of a kth symbol of a ith spatial stream for an ith receiving antenna;

a kth symbol representing an nth candidate constellation point of the ith spatial stream; z is a radical of_i,kA noise value of a k-th symbol for an i-th receiving antenna; subscript k is the time dimension, and the number of transmit side antennas is N_TThe number of receiving end antennas is N_R；

Wherein

To substitute for

12. The apparatus of claim 10, wherein the LLR calculating section comprises a comparator for comparing the/and a register₂-norm calculated by the calculating part₂-norm value and updated into said register of bit positions "0" or "1" corresponding to the respective candidate constellation point element.