CN101197603B - Low complexity step-by-step detecting system and method of multi-antenna system based on spherical decoding - Google Patents

Abstract

Translated fromChinese

一种多天线系统基于球形译码的低复杂度分步检测系统及检测方法，其首先根据所述多天线系统接收端接收到的信号估计所述多天线系统发射端所发射的信号的硬判值，并根据星座拆分原理将所述硬判值进行拆分为多个分量，然后根据所得到的多个分量计算所述接收到的信号的相应分信号，并于对所述多个分信号进行校正后计算出多天线系统的外部信息，由此，本发明降低了所述多天线系统球形译码的复杂度。

A low-complexity step-by-step detection system and detection method based on sphere decoding for a multi-antenna system, which first estimates the hard judgment of the signal transmitted by the multi-antenna system transmitting end based on the signal received by the multi-antenna system receiving end value, and split the hard judgment value into multiple components according to the principle of constellation splitting, and then calculate the corresponding sub-signals of the received signal according to the obtained multiple components, and then analyze the multiple components After the signal is corrected, the external information of the multi-antenna system is calculated, thus, the present invention reduces the complexity of spherical decoding of the multi-antenna system.

Description

Translated fromChinese

多天线系统基于球形译码的低复杂度分步检测系统及检测方法Low-complexity step-by-step detection system and detection method based on sphere decoding for multi-antenna system

技术领域technical field

本发明涉及一种多天线系统基于球形译码的低复杂度分步检测系统及检测方法。The invention relates to a low-complexity step-by-step detection system and a detection method based on spherical decoding for a multi-antenna system.

背景技术Background technique

在多天线系统中，迭代检测译码技术可使发送速率达到信道容量值，此时最佳检测器是最大后验概率检测器。但是最大后验概率的复杂度是跟发射/接收天线数和每个调制符号的比特数成指数关系的。现已存在一些次最佳低复杂度的检测结构，如Turbo-分层、迭代树搜索、序列球形译码等技术，但都涉及到软信息的计算，复杂度都很大。In a multi-antenna system, the iterative detection and decoding technique can make the transmission rate reach the channel capacity value, and the optimal detector is the maximum a posteriori probability detector at this time. However, the complexity of the maximum a posteriori probability is exponentially related to the number of transmit/receive antennas and the number of bits per modulation symbol. There are some sub-optimal low-complexity detection structures, such as Turbo-hierarchy, iterative tree search, sequential sphere decoding, etc., but they all involve the calculation of soft information, and the complexity is very large.

再者，现有的检测技术都是性能跟检测复杂度之间的一种折衷。例如，Turbo-分层检测是基于随机分层空时码和迭代接收机结构的，迭代接收机的每次迭代都能快速提高误比特率性能，但不能保证最佳的性能；迭代树搜索检测是使用M算法在一颗结构树中搜索最佳路径，寻找具有最大后验的发送符号向量，在此算法中，当保留路径足够大时可以达到最佳检测的性能；序列球形译码是寻找最接近接收符号的M个相量，把这个集合称为一个候选序列，用这些序列的值产生检测器的输出，当球形译码的初始半径足够大且序列长度M足够长时，序列球形译码可完全达到最佳检测器最大后验概率检测的性能。Furthermore, existing detection techniques are a compromise between performance and detection complexity. For example, Turbo-hierarchical detection is based on random hierarchical space-time codes and an iterative receiver structure, each iteration of the iterative receiver can quickly improve the bit error rate performance, but can not guarantee the best performance; iterative tree search detection It is to use the M algorithm to search for the best path in a structure tree, and to find the transmitted symbol vector with the maximum a posteriori. In this algorithm, when the reserved path is large enough, the best detection performance can be achieved; the sequence sphere decoding is to find The M phasors closest to the received symbols are called a candidate sequence, and the values of these sequences are used to generate the output of the detector. When the initial radius of the spherical decoding is large enough and the sequence length M is long enough, the sequence spherical decoding The code can fully achieve the performance of the maximum a posteriori probability detection of the best detector.

举例来说，对于迭代检测译码的多天线系统，其有N_T个发射天线N_R个接收天线，在发射端，数据流u经过纠错编码后交织到星座映射模块，即每M个编码比特被映射成一个复数QAM信号，再经一个串并转换模块，符号向量$S={[s_1,...,s_{N_T}]}^T$从N_T个发射天线上同时发射出去。而在接收端，接收向量可以由下式复数形式的多天线矩阵模型表示：For example, for a multi-antenna system with iterative detection and decoding, it has_NT transmitting antennas and_NR receiving antennas. At the transmitting end, the data stream u is interleaved to the constellation mapping module after error correction coding, that is, every M codes Bits are mapped into a complex QAM signal, and then through a serial-to-parallel conversion module, the symbol vector$S={[{\mathrm{the<figure-callout\; label="s"\; filenames="H061B9314920070321E000012.png,H061B9314920070321E000021.png"\; state="\{\{state\}\}">s</figure-callout>}}_1,...,{\mathrm{the\; s}}_{N_T}]}^T$ Simultaneously transmit from_NT transmit antennas. At the receiving end, the receiving vector can be represented by a multi-antenna matrix model in complex form as follows:

y＝HS+n                 (1)y=HS+n (1)

其中$y={[y_1,...,y_{N_R}]}^T$是接收到的复符号向量；H是已知的N_R×N_T维理想信道矩阵，其每个元素都是统计独立的零均值单位方差的复高斯随机变量；向量$n=[n_1,...,n_{N_R}]$是零均值方差为σ²的复高斯噪声。对于序列球形译码检测，其是基于最大后验概率检测算法的，而在最大后验概率检测中，设每个发射符号s_i，i＝1，...，N_T是通过M比特x_i^m，m＝1，...，M的映射得到的，x_i^m，m＝1，...，M的取值为正负1，则相应于发射符号向量s的每个编码比特x_k，k＝1，...，N_TM的外部对数似然比可以表示为：in$\mathrm{the\; y}={[{\mathrm{the<figure-callout\; label="y"\; filenames="H061B9314920070321E000012.png,H061B9314920070321E000021.png"\; state="\{\{state\}\}">y</figure-callout>}}_1,...,{\mathrm{the\; y}}_{N_R}]}^T$ is the received complex symbol vector; H is the known N_R ×N_T dimensional ideal channel matrix, each element of which is a statistically independent complex Gaussian random variable with zero mean and unit variance; the vector$\mathrm{no}=[{\mathrm{no}}_1,...,{\mathrm{no}}_{N_R}]$ is complex Gaussian noise with zero mean and variance^σ2 . For sequence sphere decoding detection, it is based on the maximum a posteriori probability detection algorithm, and in the maximum a posteriori probability detection, it is assumed that each transmitted symbol s_i , i=1,..., N_T is passed through M bits x_i^m , m=1,..., obtained from the mapping of M, x_i^m , m=1,..., the value of M is plus orminus 1, which corresponds to each coded bit of the transmitted symbol vector s The external log-likelihood ratio of x_k , k=1, ..., N_T M can be expressed as:

${\mathrm{<span><span>L</span>L</span>}}_{\mathrm{<span><span>E</span>E.</span>}}<span><span>(</span>(</span>{\mathrm{<span><span>x</span>x</span>}}_{\mathrm{<span><span>k</span>k</span>}}<span><span>|</span>|</span>\mathrm{<span><span>y</span>the\; y</span>}<span><span>)</span>)</span><span><span>\&ap;</span>\&ap;</span>\frac{\mathrm{<span><span>1</span>1</span>}}{\mathrm{<span><span>2</span>2</span>}}\underset{\mathrm{<span><span>X</span>x</span>}<span><span>\&Element;</span>\&Element;</span>{\mathrm{<span><span>X</span>x</span>}}_{\mathrm{<span><span>k</span>k</span>}<span><span>,</span>,</span><span><span>+</span>+</span>\mathrm{<span><span>1</span>1</span>}}}{\mathrm{<span><span>max</span>max</span>}}<span><span>\{</span>\{</span>\frac{<span><span>-</span>-</span>\mathrm{<span><span>1</span>1</span>}}{{\mathrm{<span><span>\&sigma;</span>\&sigma;</span>}}^{\mathrm{<span><span>2</span>2</span>}}}{<span><span>|</span>|</span><span><span>|</span>|</span>\mathrm{<span><span>y</span>the\; y</span>}<span><span>-</span>-</span>\mathrm{<span><span>HS</span>HS</span>}<span><span>|</span>|</span><span><span>|</span>|</span>}^{\mathrm{<span><span>2</span>2</span>}}<span><span>+</span>+</span>{\mathrm{<span><span>X</span>x</span>}}^{\mathrm{<span><span>T</span>T</span>}}{\mathrm{<span><span>L</span>L</span>}}_{\mathrm{<span><span>A</span>A</span>}}<span><span>\}</span>\}</span>$

$<span><span>-</span>-</span>\frac{\mathrm{<span><span>1</span>1</span>}}{\mathrm{<span><span>2</span>2</span>}}\underset{\mathrm{<span><span>X</span>x</span>}<span><span>\&Element;</span>\&Element;</span>{\mathrm{<span><span>X</span>x</span>}}_{\mathrm{<span><span>k</span>k</span>}<span><span>,</span>,</span><span><span>-</span>-</span>\mathrm{<span><span>1</span>1</span>}}}{\mathrm{<span><span>max</span>max</span>}}<span><span>\{</span>\{</span>\frac{<span><span>-</span>-</span>\mathrm{<span><span>1</span>1</span>}}{{\mathrm{<span><span>\&sigma;</span>\&sigma;</span>}}^{\mathrm{<span><span>2</span>2</span>}}}{<span><span>|</span>|</span><span><span>|</span>|</span>\mathrm{<span><span>y</span>the\; y</span>}<span><span>-</span>-</span>\mathrm{<span><span>HS</span>HS</span>}<span><span>|</span>|</span><span><span>|</span>|</span>}^{\mathrm{<span><span>2</span>2</span>}}<span><span>+</span>+</span>{\mathrm{<span><span>X</span>x</span>}}^{\mathrm{<span><span>T</span>T</span>}}{\mathrm{<span><span>L</span>L</span>}}_{\mathrm{<span><span>A</span>A</span>}}<span><span>\}</span>\}</span><span><span>-</span>-</span>{\mathrm{<span><span>L</span>L</span>}}_{\mathrm{<span><span>A</span>A</span>}}<span><span>(</span>(</span>{\mathrm{<span><span>x</span>x</span>}}_{\mathrm{<span><span>k</span>k</span>}}<span><span>)</span>)</span><span><span>-</span>-</span><span><span>-</span>-</span><span><span>-</span>-</span><span><span>(</span>(</span>\mathrm{<span><span>2</span>2</span>}<span><span>)</span>)</span>$

其中，M是调制阶数，M＝2为四相相移键控(QPSK)调制、M＝4为16正交幅度(QAM)调制、M＝6为64-QAM调制，公式(2)采用了最大对数近似准则，L_A是已知先验概率，且X_k，+1＝{X|x_k＝+1}，X_k，-1＝{X|x_k＝-1}。为使搜索空间降低到一个适度的水平，序列球形译码模块只需找到最接近于最大似然估计值的N_candi个候选项，即可兼顾检测的准确性与检测速度。根据现有文献公开的经验值，当MN_T≤8时，可做最大似然检测。当8＜MN_T≤32时N_candi＝512、32＜MN_T≤48时N_candi＝1024。然而序列球形译码方法在高调制阶数系统中应用时其复杂度是不可接受的，仅当发射符号向量s是二进制形式的，软硬球形译码在基于 $2H^T\tilde{y}={\mathrm{\&sigma;}}^2{L}_A$的变换后可以用下式来计算外部信息：Among them, M is the modulation order, M=2 is quadrature phase shift keying (QPSK) modulation, M=4 is 16-quadrature amplitude (QAM) modulation, M=6 is 64-QAM modulation, formula (2) uses The criterion of maximum logarithm approximation is established,_LA is the known prior probability, and X_k,+1 ={X|x_k =+1}, X_k,-1 ={X|x_k =-1}. In order to reduce the search space to a moderate level, the sequence sphere decoding module only needs to find the N_candi candidates closest to the maximum likelihood estimation value, which can take into account the detection accuracy and detection speed. According to the empirical values published in the existing literature, when MN_T ≤ 8, maximum likelihood detection can be performed. N_candi =512 when 8<MN_T ≤32, and N_candi =1024 when 32<MN_T ≤48. However, the complexity of the sequence sphere decoding method is unacceptable when it is applied in a high modulation order system. Only when the transmitted symbol vector s is in binary form, the soft and hard sphere decoding is based on$2h^T\widetilde{\mathrm{the\; y}}={\mathrm{\&sigma;}}^2{L}_A$ After the transformation of can use the following formula to calculate the external information:

${\mathrm{<span><span>L</span>L</span>}}_{\mathrm{<span><span>E</span>E.</span>}}<span><span>(</span>(</span>{\mathrm{<span><span>x</span>x</span>}}_{\mathrm{<span><span>k</span>k</span>}}<span><span>|</span>|</span>\mathrm{<span><span>y</span>the\; y</span>}<span><span>)</span>)</span><span><span>\&ap;</span>\&ap;</span>\frac{<span><span>-</span>-</span>{\stackrel{<span><span>^</span>^</span>}{\mathrm{<span><span>x</span>x</span>}}}_{\mathrm{<span><span>k</span>k</span>}<span><span>,</span>,</span>\mathrm{<span><span>map</span><figure-callout\; label="map"\; filenames="H061B9314920070321E000012.png"\; state="\{\{state\}\}">map</figure-callout></span>}}}{\mathrm{<span><span>2</span>2</span>}{\mathrm{<span><span>\&sigma;</span>\&sigma;</span>}}^{\mathrm{<span><span>2</span>2</span>}}}{<span><span>|</span>|</span><span><span>|</span>|</span>\mathrm{<span><span>y</span>the\; y</span>}<span><span>+</span>+</span>\stackrel{<span><span>~</span>~</span>}{\mathrm{<span><span>y</span>the\; y</span>}}<span><span>-</span>-</span>\mathrm{<span><span>H</span>h</span>}{\stackrel{<span><span>^</span>^</span>}{\mathrm{<span><span>X</span>x</span>}}}_{\mathrm{<span><span>map</span>map</span>}}<span><span>|</span>|</span><span><span>|</span>|</span>}^{\mathrm{<span><span>2</span>2</span>}}$

$<span><span>+</span>+</span>\frac{{\stackrel{<span><span>^</span>^</span>}{\mathrm{<span><span>x</span>x</span>}}}_{\mathrm{<span><span>k</span>k</span>}<span><span>,</span>,</span>\mathrm{<span><span>map</span><figure-callout\; label="map"\; filenames="H061B9314920070321E000012.png"\; state="\{\{state\}\}">map</figure-callout></span>}}}{\mathrm{<span><span>2</span>2</span>}{\mathrm{<span><span>\&sigma;</span>\&sigma;</span>}}^{\mathrm{<span><span>2</span>2</span>}}}\underset{\mathrm{<span><span>X</span>x</span>}<span><span>\&Element;</span>\&Element;</span>{\mathrm{<span><span>X</span>x</span>}}_{\mathrm{<span><span>k</span>k</span>}<span><span>,</span>,</span><span><span>-</span>-</span>{\stackrel{<span><span>^</span>^</span>}{\mathrm{<span><span>x</span>x</span>}}}_{\mathrm{<span><span>k</span>k</span>}<span><span>,</span>,</span>\mathrm{<span><span>map</span>map</span>}}}}{\mathrm{<span><span>min</span>min</span>}}{<span><span>|</span>|</span><span><span>|</span>|</span>\mathrm{<span><span>y</span>the\; y</span>}<span><span>+</span>+</span>\stackrel{<span><span>~</span>~</span>}{\mathrm{<span><span>y</span>the\; y</span>}}<span><span>-</span>-</span>\mathrm{<span><span>HX</span>HX</span>}<span><span>|</span>|</span><span><span>|</span>|</span>}^{\mathrm{<span><span>2</span>2</span>}}<span><span>-</span>-</span>{\mathrm{<span><span>L</span>L</span>}}_{\mathrm{<span><span>A</span>A</span>}}<span><span>(</span>(</span>{\mathrm{<span><span>x</span>x</span>}}_{\mathrm{<span><span>k</span>k</span>}}<span><span>)</span>)</span><span><span>-</span>-</span><span><span>-</span>-</span><span><span>-</span>-</span><span><span>(</span>(</span>\mathrm{<span><span>3</span>3</span>}<span><span>)</span>)</span>$

其中${\hat{X}}_{\mathrm{map}}=\underset{X\&Element;X}{\arg \min }{\left|\right|y+\tilde{y}-\mathrm{HX}\left|\right|}^2$可以由硬球形译码模块计算而得，软硬球形译码只需产生MN_T+1个候选项来精确地计算后验比特概率值。in${\hat{x}}_{\mathrm{map}}=\underset{x\&Element;x}{\arg \min }{\left|\right|\mathrm{the\; y}+\widetilde{\mathrm{the\; y}}-\mathrm{HX}\left|\right|}^2$ It can be calculated by the hard sphere decoding module, and the soft and hard sphere decoding only needs to generate MN_T + 1 candidates to accurately calculate the posterior bit probability value.

由上可知，若要到达或接近最佳检测器的性能，通常检测器的复杂度都非常大。而当发射符号是二进制的，采用软硬球形译码方法(soft to hard sphere decoder)，其性能完全等同于序列球形译码而其计算复杂度有明显下降，在高信噪比下的平均复杂度是O(M⁴)(其中M是数据块长度)。由于软硬球形译码只能在二元幅移键控(2ASK)和四元相移键控(4PSK，也即QPSK)调制系统中使用，因此，如何降低现有高阶调制系统的译码复杂度问题实已成为本领域技术人员亟待解决的技术课题。It can be seen from the above that if the performance of the optimal detector is to be reached or approached, the complexity of the detector is usually very large. And when the transmitted symbol is binary, using the soft to hard sphere decoder, its performance is completely equivalent to the sequence sphere decoding and its computational complexity is significantly reduced, and the average complexity under high SNR The degree is O(M⁴ ) (where M is the data block length). Since soft and hard spherical decoding can only be used in binary amplitude shift keying (2ASK) and quaternary phase shift keying (4PSK, ie QPSK) modulation systems, how to reduce the decoding cost of existing high-order modulation systems? The complexity problem has actually become a technical subject to be solved urgently by those skilled in the art.

发明内容Contents of the invention

本发明的目的在于提供一种多天线系统基于球形译码的低复杂度分步检测系统及检测方法，以降低所述多天线系统球形译码的复杂度。The object of the present invention is to provide a low-complexity step-by-step detection system and detection method based on sphere decoding for a multi-antenna system, so as to reduce the complexity of sphere decoding for the multi-antenna system.

为了达到上述目的，本发明提供一种多天线系统基于球形译码的低复杂度分步检测系统，其包括：用于根据所述多天线系统接收端接收到的信号估计所述多天线系统发射端所发射的信号的硬判值，并根据星座拆分原理将所述硬判值拆分为多个分量的硬判检测及星座拆分模块、分别用于根据各自对应的所述硬判检测及星座拆分模块所获得的分量计算所述接收到的信号相应分信号的多个软硬球形译码模块、用于根据所述多个软硬球形译码模块所得到的各分信号及预先获得的先验信息计算相应的外部信息的外部信息计算模块。In order to achieve the above object, the present invention provides a low-complexity step-by-step detection system based on sphere decoding for a multi-antenna system, which includes: used for estimating the transmission The hard judgment value of the signal transmitted by the end, and according to the principle of constellation splitting, the hard judgment value is split into a plurality of components of hard judgment detection and constellation splitting modules, which are respectively used for the hard judgment detection according to the respective corresponding and the components obtained by the constellation splitting module to calculate a plurality of soft and hard spherical decoding modules of the corresponding sub-signals of the received signal, and are used to obtain each sub-signal according to the plurality of soft and hard spherical decoding modules and pre-set The obtained prior information is calculated by an external information calculation module corresponding to the external information.

其中，所述多个软硬球形译码模块包括用于根据所述星座拆分原理校正所述各分信号的调整单元。Wherein, the plurality of soft and hard spherical decoding modules include an adjustment unit for correcting the sub-signals according to the constellation splitting principle.

进一步，本发明还提供一种多天线系统基于球形译码的低复杂度分步检测方法，其包括步骤：1)根据所述多天线系统接收端接收到的信号估计所述多天线系统发射端所发射的信号的硬判值，并根据星座拆分原理将所述硬判值进行拆分为多个分量；2)根据所获得的多个分量计算所述接收到的信号的相应分信号；3)根据所获得的各分信号及预先获得的先验信息计算相应的外部信息。Further, the present invention also provides a low-complexity step-by-step detection method for a multi-antenna system based on sphere decoding, which includes the steps of: 1) estimating the transmit end of the multi-antenna system according to the signal received by the receive end of the multi-antenna system The hard judgment value of the transmitted signal, and split the hard judgment value into multiple components according to the constellation splitting principle; 2) calculate the corresponding sub-signal of the received signal according to the obtained multiple components; 3) Calculate the corresponding external information according to the obtained sub-signals and the prior information obtained in advance.

其中，所述步骤1)中采用硬判检测方法估计所述硬判值，所述硬判检测方法包括ZF算法或M算法，当所述多天线系统为16正交幅度(QAM)调制的系统时，设所述硬判值为S，根据星座拆分原理其可被拆分为分量S₁及S₂，其中， 

当所述接收到的信号为y，其对应的分信号分别为y₁、y₂，满足： 

其中，H是理想信道矩阵，所述步骤2)还包括一根据所述星座拆分原理校正所述各分信号的校正步骤，即根据所述星座拆分原理将所述分信号y₁校正为y₁/4。Wherein, in the step 1), a hard decision detection method is used to estimate the hard decision value, and the hard decision detection method includes the ZF algorithm or the M algorithm, when the multi-antenna system is a system of 16 quadrature amplitude (QAM) modulation , the hard judgment value is set to S, which can be split into components S₁ and S₂ according to the principle of constellation splitting, where,

When the received signal is y, its corresponding sub-signals are respectively y₁ and y₂ , satisfying:

Wherein, H is an ideal channel matrix, and the step 2) also includes a correction step of correcting the sub-signals according to the constellation splitting principle, that is, correcting the sub-signals_y1 according to the constellation splitting principle as y_1/4 .

综上所述，本发明的多天线系统基于球形译码的低复杂度分步检测系统及检测方法，根据星座拆分原理将高阶调制系统拆分为多个QPSK系统，可降低所述多天线系统球形译码的复杂度。In summary, the multi-antenna system of the present invention is based on a low-complexity step-by-step detection system and detection method based on spherical decoding, and the high-order modulation system is split into multiple QPSK systems according to the principle of constellation splitting, which can reduce the number of QPSK systems described above. Complexity of sphere decoding for antenna systems.

附图说明Description of drawings

图1为16-QAM星座点格雷映射结构示意图。FIG. 1 is a schematic diagram of a 16-QAM constellation point Gray mapping structure.

图2为本发明多天线系统基于球形译码的低复杂度分步检测系统结构示意图。FIG. 2 is a schematic structural diagram of a low-complexity step-by-step detection system based on sphere decoding for a multi-antenna system according to the present invention.

图3为四发四收多天线16-QAM调制系统分步球形译码检测误比特率性能示意图。Fig. 3 is a schematic diagram of the bit error rate detection performance of the four-transmission and four-reception multi-antenna 16-QAM modulation system step-by-step spherical decoding.

图4为基于实数乘法的复杂度比较示意图。Fig. 4 is a schematic diagram of complexity comparison based on real number multiplication.

具体实施方式Detailed ways

请参见图1，本发明的多天线系统基于球形译码的低复杂度分步检测方法，其包括以下步骤：Please refer to Fig. 1, the multi-antenna system of the present invention is based on the low-complexity step-by-step detection method of sphere decoding, and it comprises the following steps:

1)根据所述多天线系统接收端接收到的信号估计所述多天线系统发射端所发射的信号的硬判值，并根据星座拆分原理将所述硬判值进行拆分为多个分量，通常采用硬判检测方法估计所述硬判值，其中，所述硬判检测方法包括ZF算法及M算法等，现有星座拆分原理已经揭露一个16-QAM星座点可以被拆分成两个QPSK子星座点，因此，对于16正交幅度(QAM)调制的多天线系统，设定所述硬判值为S，根据现有星座拆分原理将其拆分为s₁，s₂，三者满足下式：1) Estimate the hard judgment value of the signal transmitted by the multi-antenna system transmitting end according to the signal received by the multi-antenna system receiving end, and split the hard judgment value into multiple components according to the principle of constellation splitting , the hard-judgment detection method is usually used to estimate the hard-judgment value, wherein the hard-judgment detection method includes the ZF algorithm and the M algorithm, etc., and the existing constellation splitting principle has revealed that a 16-QAM constellation point can be split into two QPSK sub-constellation points, therefore, for a multi-antenna system with 16 quadrature amplitude (QAM) modulation, set the hard decision value S, and split it into s₁ , s₂ according to the existing constellation splitting principle, The three satisfy the following formula:

$\mathrm{<span><span>S</span>S</span>}<span><span>=</span>=</span><span><span>(</span>(</span>\mathrm{<span><span>2</span>2</span>}<span><span>/</span>/</span>\sqrt{\mathrm{<span><span>5</span>5</span>}}<span><span>)</span>)</span>{\mathrm{<span><span>S</span><figure-callout\; label="S"\; filenames="H061B9314920070321E000012.png,H061B9314920070321E000021.png"\; state="\{\{state\}\}">S</figure-callout></span>}}_{\mathrm{<span><span>1</span>1</span>}}<span><span>+</span>+</span><span><span>(</span>(</span>\mathrm{<span><span>1</span>1</span>}<span><span>/</span>/</span>\sqrt{\mathrm{<span><span>5</span>5</span>}}<span><span>)</span>)</span>{\mathrm{<span><span>S</span>S</span>}}_{\mathrm{<span><span>2</span>2</span>}}<span><span>-</span>-</span><span><span>-</span>-</span><span><span>-</span>-</span><span><span>(</span>(</span>\mathrm{<span><span>4</span>4</span>}<span><span>)</span>)</span>$

其中 

且S，S₁，S₂都是归一化形式，由图1所示的16-QAM星座点格雷映射结构示意图可知，图1中的大方框代表了S₁星座图的大小，小方框代表了S₂星座图的大小，S₁与16-QAM星座点的前两个比特相关，S₂与16-QAM星座点的后两个比特相关。此外，对于64QAM调制的多天线系统，也可通过相应的拆分公式将64QAM星座点拆分为三个QPSK星座点，在此不再赘述。in

And S, S₁ , and S₂ are all normalized forms. From the schematic diagram of the gray mapping structure of 16-QAM constellation points shown in Figure 1, it can be known that the large box in Figure 1 represents the size of the S₁ constellation, and the small box Represents the size of the S₂ constellation diagram, S₁ is related to the first two bits of the 16-QAM constellation point, and S₂ is related to the last two bits of the 16-QAM constellation point. In addition, for a multi-antenna system modulated by 64QAM, the 64QAM constellation point can also be split into three QPSK constellation points through a corresponding split formula, which will not be repeated here.

2)根据所获得的多个分量计算所述接收到的信号的相应分信号，设定所述多天线系统接收端所接收到的信号为y，其对应的分信号分别为y₁、y₂，应满足： 

其中，H是理想信道矩阵，在所述多天线系统中，y₁、y₂分别通过两个软硬球形译码算法即可得到。由图1可知，S₁星座图的大小是_S2星座图的大小子模块2的两倍，为抵消不同星座尺寸的影响，需要对分信号进行校正，由于星座图尺寸的两倍相当于能量大了4倍，因此，在本实施方式中，将得到的分信号y₁校正为y₁/4。2) Calculate the corresponding sub-signals of the received signal according to the obtained multiple components, set the signal received by the receiving end of the multi-antenna system as y, and the corresponding sub-signals are y₁ and y₂ respectively , should satisfy:

Wherein, H is an ideal channel matrix, and in the multi-antenna system, y₁ and y₂ can be obtained through two soft and hard spherical decoding algorithms respectively. It can be seen from Figure 1 that the size of the S₁ constellation is twice the size of thesub-module 2 of the_S 2 constellation. In order to offset the influence of different constellation sizes, the divided signal needs to be corrected, because twice the size of the constellation is equivalent to the energy is 4 times larger, therefore, in this embodiment, the obtained partial signal y₁ is corrected to y₁ /4.

3)根据所获得的各分信号及预先所获得的先验信息计算各分信号对应的外部信息，通常，将所述各分信号减去所对应的先验信息即得到各外部信息，所述外部信息输入至所述多天线系统的信道解码模块以供其进行相应的解码，其中所述先验信息由所述信道解码模块所提供，其为本领域技术人员所知悉，在此不再赘述。3) Calculate the external information corresponding to each sub-signal according to the obtained sub-signals and the prior information obtained in advance, usually, subtract the corresponding a priori information from each sub-signal to obtain each external information, the said The external information is input to the channel decoding module of the multi-antenna system for corresponding decoding, wherein the prior information is provided by the channel decoding module, which is known to those skilled in the art and will not be repeated here .

请参见图2，本发明的多天线系统基于球形译码的低复杂度分步检测系统包括：一硬判检测及星座拆分模块、多个软硬球形译码模块、以及一外部信息计算模块。Please refer to Fig. 2, the low-complexity step-by-step detection system based on sphere decoding of the multi-antenna system of the present invention includes: a hard judgment detection and constellation splitting module, a plurality of soft and hard sphere decoding modules, and an external information calculation module .

所述硬判检测及星座拆分模块用于根据所述多天线系统接收端接收到的信号估计所述多天线系统发射端所发射的信号的硬判值，并根据星座拆分原理将所述硬判值进行拆分为多个分量，例如，对于16正交幅度(QAM)调制的多天线系统，根据星座拆分原理可将其拆分为两个QPSK子星座点，对于64QAM调制的多天线系统，也可通过相应的拆分公式将64QAM星座点拆分为三个QPSK星座点。The hard judgment detection and constellation splitting module is used to estimate the hard judgment value of the signal transmitted by the multi-antenna system transmitter according to the signal received by the multi-antenna system receiving end, and divide the constellation according to the constellation splitting principle. The hard decision value is split into multiple components. For example, for a multi-antenna system modulated by 16 quadrature amplitude (QAM), it can be split into two QPSK sub-constellation points according to the principle of constellation splitting. For a multi-antenna system modulated by 64QAM The antenna system can also split the 64QAM constellation point into three QPSK constellation points through a corresponding split formula.

所述多个软硬球形译码模块分别用于根据各自对应的所述硬判检测及星座拆分模块所获得的分量计算所述接收到的信号相应分信号，在本实施方式中，所述多个软硬球形译码模块即第一软硬球形译码模块及第二软硬球形译码模块，当所述硬判值被拆分为两个分量S₁和S₂，第一软硬球形译码模块根据分量S₂计算所述接收到的信号相应分信号y₁，第二软硬球形译码模块根据分量S₁计算所述接收到的信号相应分信号y₂，根据星座拆分原理可知，S₁星座图的大小是S₂星座图的大小子模块2的两倍，为抵消不同星座尺寸的影响，需要对分信号进行校正，由于星座图尺寸的两倍相当于能量大了4倍，因此，在本实施方式中，所述第一软硬球形译码模块还设有调整单元，用于将所述分信号y₁校正为y₁/4。需注意的是，软硬球形译码模块的个数并非以本实施方式为限，本领域技术人员可根据实际的需要予以设定，再有，调整单元的设定也可根据实际的需要予以设定，例如，对于64QAM调制的多天线系统，相应的设有三个软硬球形译码模块用于分别计算各相应的分信号。The plurality of soft and hard spherical decoding modules are respectively used to calculate the corresponding sub-signals of the received signal according to the components obtained by the corresponding hard-judgment detection and constellation splitting modules. In this embodiment, the A plurality of soft and hard spherical decoding modules are the first soft and hard spherical decoding module and the second soft and hard spherical decoding module. When the hard judgment value is split into two components S₁ and S₂ , the first soft and hard spherical decoding module The spherical decoding module calculates the corresponding sub-signal y₁ of the received signal according to the component S₂ , and the second soft and hard spherical decoding module calculates the corresponding sub-signal y₂ of the received signal according to the component S₁ , and splits according to the constellation The principle shows that the size of the S₁ constellation is twice the size of thesub-module 2 of the S₂ constellation. In order to offset the influence of different constellation sizes, the divided signal needs to be corrected, because twice the size of the constellation is equivalent to alarger energy 4 times, therefore, in this embodiment, the first soft and hard spherical decoding module is further provided with an adjustment unit for correcting the sub-signal y₁ to y₁ /4. It should be noted that the number of soft and hard spherical decoding modules is not limited to this embodiment, and those skilled in the art can set them according to actual needs. Moreover, the setting of the adjustment unit can also be set according to actual needs. It is assumed that, for example, for a multi-antenna system modulated by 64QAM, three soft and hard spherical decoding modules are correspondingly provided to calculate respective sub-signals.

所述外部信息计算模块用于根据所述多个软硬球形译码模块所得到的各分信号及预先获得的先验信息计算相应的外部信息，例如，将所述分信号y₁减去第一二位的先验信息即得到第一二位的外部信息，将所述分信号y₂减去第三四位的先验信息即得到第三四位的外部信息，由此即可得到每比特的外部信息。The external information calculation module is used to calculate the corresponding external information according to the sub-signals obtained by the multiple soft and hard spherical decoding modules and the prior information obtained in advance, for example, subtracting the sub-signal_y1 from the sub-signal The prior information of one or two digits is to obtain the external information of the first two digits. Subtracting the prior information of the third and fourth digits from the sub-signal_y2 obtains the external information of the third and fourth digits, and thus each bits of external information.

以下为对对本发明进行仿真的仿真结果，系统仿真参数如下表一所示，图3为本发明在平坦衰落信道下及不同信噪比条件下的误比特率性能，现有的序列球形译码的性能作为一个参考也在图中示出，其仿真参数也如下表所示：The following is the simulation result of the simulation of the present invention. The system simulation parameters are shown in Table 1 below. Fig. 3 shows the bit error rate performance of the present invention under flat fading channels and different SNR conditions. The performance of is also shown in the figure as a reference, and its simulation parameters are also shown in the table below:

表一仿真条件设定Table 1 Simulation condition setting

  天线配置Antenna configuration  4×44×4  调试方式Debugging method  16QAM16QAM  信道 channel  平坦衰落瑞利信道flat fading Rayleigh channel  每个数据块的比特数The number of bits per data block  40044004  每个信噪比条件下的仿真帧数目The number of simulation frames under each SNR condition  25002500  信道编码channel coding  1/2Turbo(5，7)1/2 Turbo(5, 7)  Turbo迭代译码内迭代次数Number of iterations in Turboiterative decoding  8 8  球形译码初始搜索半径Initial search radius for spherical decoding  1e41e4  迭代接收机的迭代次数The number of iterations of theiterative receiver  44

初始球形译码搜索半径设定为一个足够大的值以保证QPSK子模块能找到合适的点来计算外部信息，因为满足前述式(3)的候选项是考虑了先验信息通过硬球形译码模块搜索得到的，所以需要一个足够大的半径来包含在考虑了先验以后所能搜到的点。The initial spherical decoding search radius is set to a sufficiently large value to ensure that the QPSK sub-module can find a suitable point to calculate the external information, because the candidates satisfying the aforementioned formula (3) consider the prior information through hard spherical decoding It is obtained by module search, so a radius large enough to contain the points that can be searched after considering the prior is required.

在误比特率为10^-4时方案硬球形译码跟分步球形译码的级联时(SD/MSD)的性能较之LSD(序列球形译码)下降了2dB，同时也给出了方案最小均方误差跟分步球形译码的级联时(MMSE/MSD)的性能，其又下降了1dB。When the bit error rate is 10^-4 , the performance of the concatenation of hard sphere decoding and step sphere decoding (SD/MSD) is 2dB lower than that of LSD (sequential sphere decoding), and the scheme is also given When the minimum mean square error is concatenated with stepwise sphere decoding (MMSE/MSD), it drops another 1dB.

从图4可得SD/MSD跟MMSE/MSD的复杂度相当，但是从性能上考虑，SD/MSD更加优越。It can be seen from Figure 4 that the complexity of SD/MSD is comparable to that of MMSE/MSD, but in terms of performance, SD/MSD is superior.

进一步地，对复杂度进行分析如下：Further, the complexity analysis is as follows:

在上述仿真条件下，对于每个接收向量，序列球形译码模块需要搜索512最好的候选项来保证其性能基本等同于最大后验概率检测器的性能。然而，本发明所提出的方法把检测分成两步实现，只需要2×9个候选项来计算每个比特的后验概率。因为对于每个QPSK子模块，需要MN_T+1＝2×4+1＝9个候选项来得到每个比特的软信息。Under the above simulation conditions, for each received vector, the sequential sphere decoding module needs to search 512 for the best candidate to ensure that its performance is basically equal to that of the maximum a posteriori probability detector. However, the method proposed by the present invention divides the detection into two steps and only needs 2×9 candidates to calculate the posterior probability of each bit. Because for each QPSK sub-module, MN_T +1=2×4+1=9 candidates are required to obtain the soft information of each bit.

图4给出了本发明的方法与序列球形译码检测的的基于实数乘法的复杂度比较，前者的复杂度是后者的0.6％。Fig. 4 shows the complexity comparison between the method of the present invention and sequence sphere decoding detection based on real number multiplication, the complexity of the former is 0.6% of the latter.

在迭代检测的第二到第四次迭代中，子模块使用了同第一次迭代一样的拆分星座点。尽管最小均方误差检测比硬球形译码的检测简单得多，但这个硬判的过程只进行一次，在整体上SD/MSD跟MMSE/MSD的复杂度基本相同。表二给出了具体的实数乘法复杂度比较。In the second to fourth iterations of iterative detection, the submodules used the same split constellation points as in the first iteration. Although the minimum mean square error detection is much simpler than the detection of hard spherical decoding, this hard judgment process is only performed once, and the overall complexity of SD/MSD and MMSE/MSD is basically the same. Table 2 gives the specific real number multiplication complexity comparison.

表二迭代过程的实数乘法复杂度比较Table 2 Comparison of real number multiplication complexity of iterative process

Claims (10)

1. a multiaerial system is characterized in that comprising based on the low complexity step-by-step detection system of globular decoding:

Declare firmly and detect and constellation fractionation module, be used for estimating value of declaring firmly of the signal that described multiaerial system transmitting terminal is launched, and described value of declaring firmly is split as a plurality of components according to constellation fractionation principle according to the signal that described multiaerial system receiving terminal receives;

A plurality of soft or hard globular decoding modules are respectively applied for according to each self-corresponding described declare firmly detection and the corresponding sub-signal of the constellation fractionation component that module the obtained described signal that receives of calculating;

The external information computing module is used for calculating corresponding external information according to the prior information of resulting each sub-signal of described a plurality of soft or hard globular decoding modules and acquisition in advance.

2. multiaerial system as claimed in claim 1 is characterized in that based on the low complexity step-by-step detection system of globular decoding: described a plurality of soft or hard globular decoding modules comprise and are used for splitting the adjustment unit that principle is proofreaied and correct described each sub-signal according to described constellation.

3. a multiaerial system is characterized in that may further comprise the steps based on the low complexity step-by-step detection method of globular decoding:

1) signal that receives according to described multiaerial system receiving terminal is estimated value of declaring firmly of the signal that described multiaerial system transmitting terminal is launched, and according to constellation fractionation principle described value of declaring firmly is split as a plurality of components;

2) calculate the corresponding sub-signal of the described signal that receives according to a plurality of components that obtained;

3) calculate corresponding external information according to the prior information of each sub-signal that is obtained and acquisition in advance.

4. multiaerial system as claimed in claim 3 is characterized in that based on the low complexity step-by-step detection method of globular decoding: adopt in the described step 1) and declare the described value of declaring firmly of detection method estimation firmly.

5. multiaerial system as claimed in claim 4 is characterized in that based on the low complexity step-by-step detection method of globular decoding: describedly declare detection method firmly and comprise the ZF algorithm.

6. multiaerial system as claimed in claim 4 is characterized in that based on the low complexity step-by-step detection method of globular decoding: describedly declare detection method firmly and comprise the M algorithm.

7. multiaerial system as claimed in claim 3 is based on the low complexity step-by-step detection method of globular decoding, it is characterized in that: when described multiaerial system is the system of 16 quadrature amplitudes (QAM) modulation, if described value of declaring firmly is S, it can be split as component S to split principle according to constellation₁And S₂, wherein,

8. multiaerial system as claimed in claim 7 is characterized in that based on the low complexity step-by-step detection method of globular decoding: the described signal that receives is y, and its corresponding sub-signal is respectively y₁, y₂, satisfy:

Wherein, H is the ideal communication channel matrix.

9. multiaerial system as claimed in claim 3 is characterized in that based on the low complexity step-by-step detection method of globular decoding: described step 2) also comprise an aligning step according to described each sub-signal of described constellation fractionation principle correction.

10. multiaerial system as claimed in claim 8 is characterized in that based on the low complexity step-by-step detection method of globular decoding: split principle with described sub-signal y according to described constellation₁Proofread and correct and be y₁/ 4.

Images