CN102006113A - Parallel MIMO (multiple input multiple output) signal detection method based on zero forcing predetection - Google Patents

Abstract

The invention discloses a parallel MIMO (multiple input multiple output) signal detection method based on zero forcing predetection, which is mainly used for solving the problems that the existing MIMO receiver detection method has high complexity and is not favourable for realizing high-speed processing. The specific steps are as follows: (1) according to a channel transmission matrix, determining the detection sequence of transmission signals; (2) calculating the zero forcing predetection solution of the first-layer detection signals; (3) determining a candidate traversal symbol; (4) eliminating the interference of the fir-layer detection signals; (5) estimating the zero forcing of the second-layer detection signals; (6) according to the zero forcing estimation, removing the interference of the second-layer detection signals, thus obtaining residual vectors; (7) calculating the Euclidean distance of the residual vectors; and (8) outputting a candidate traversal symbol corresponding to parallel branches with the minimum Euclidean distance and zero forcing estimation serving as final detection output. The invention has the advantages that the complexity of the traditional parallel detection method is lowered by combined zero forcing predetection, the performance loss is small and rapid realization is ensured, and the method can be used for the MIMO detection of an LTE (long term evolution) receiver.

Description

Parallel MIMO signal detecting method based on the ZF pre-detection

Technical field

The invention belongs to wireless communication field, relate to a kind of signal detecting method of multiple-input and multiple-output mimo system, can be used for the MIMO receiver in the NGBW communication system.

Background technology

The input of mimo system is a core technology in the NGBW communication system.At present, propose numerous MIMO signal detecting methods, comprised methods such as maximum likelihood ML detection, ZF ZF detection, least mean-square error MMSE detection, spherical SD detection, K-best detection and the PD detection that walks abreast.Wherein, ML detection method best performance, but the computational complexity of ML detection method is too high, is difficult to use in practice; Though and the complexity of linearity test methods such as ZF, MMSE is relatively low, poor-performing; Though method performances such as spherical SD detection, K-best detection, PD detection are near the performance of ML detection method, spherical SD detects and is unfavorable for that parallel processing, processing speed are difficult to improve; The parallel PD detection method that Sanhae Kim proposes can be used for 22 of WiMAX system up-links collects mail and number detects, but this method travels through all constellation symbol to wherein one deck signal demand, and when order of modulation was higher, complexity was still very high; Min Chuin Hoo proposition in patent (07523037) " Reduced complexity detector for multiple-antenna systems " utilizes the lower detection method of certain complexity to reduce the institute's symbol search that might send set earlier, this thought of Reuven utilizing in patent (7720169) " Multiple-input multiple-output (MIMO) detector incorporating efficient signal point search and softinformation refinement " reduces a way of tree-like searching method, proposed k-best detection method, handled but this method equally not too is fit to high-speed parallel at 22 receipts mimo systems.

For NGBW communication system LTE-A standard, typical antenna configurations is 22 and receives, the MIMO detection method needs parallel processing satisfying the system high-speed demand, and said method is because the factor of computational complexity, implementation structure is difficult to satisfy this high-speed requirement.

Summary of the invention

The objective of the invention is to overcome the deficiency of above-mentioned prior art, a kind of parallel (ZF-PD) MIMO signal detecting method based on the ZF pre-detection is proposed, to reduce parallel branch number and the overall complexity that detects, satisfy the requirement of NGBW communication system high speed processing.

The present invention is achieved in that

One. know-why

Define the signal model of 2 reception antennas of 2 transmitting antennas:

R＝HX+W

Wherein, R=[r ₁, r ₂] ^TBe that 2 dimensions receive column vector, r ₁Be the received signal of reception antenna 1, r ₂Be the received signal of reception antenna 2, subscript T represents the transposition computing; X=[x ₁, x ₂] ^TBe that 2 dimensions send signal train vector, x ₁Be the transmission signal of transmitting antenna 1, x ₂Transmission signal for transmitting antenna 2; W=[n ₁, n ₂] ^TBe 2 dimension noise column vectors, n ₁The signal noise of expression reception antenna 1, n ₂The signal noise of expression reception antenna 2;

Be the Channel Transmission matrix, h _IjRepresent the channel fading coefficient of transmitting antenna i to reception antenna j, here, i, j=1,2.

The MIMO input that the present invention relates to is promptly estimated X according to described R and H.

Two. technical scheme

The present invention is based on the parallel MIMO signal detecting method of ZF pre-detection, comprise the steps:

(1), determine the detection order of transmitting antenna signal according to the Channel Transmission matrix H:

$r=\left\{\begin{array}{cc}1,& \mathrm{if}{\left|h_{22}\right|}^2+{\left|h_{12}\right|}^2>{\left|h_{21}\right|}^2+{\left|h_{11}\right|}^2\\ 2,& \mathrm{else}\end{array}\right.;$

Wherein, h _IjExpression transmitting antenna i is to the channel fading coefficient of reception antenna j, i, j=1,2;

(2) utilize received signal vector R and Channel Transmission matrix H to calculate ground floor detection signal x _mThe ZF pre-detection separate

$x_m^{\%}=\frac{1}{\left|H\right|}(q_{m1}{\mathrm{<figure-callout\; id="1"\; label="r"\; filenames="HDA0000034327140000011.png"\; state="\{\{state\}\}">r</figure-callout>}}_1+q_{m2}{r}_2)$

Wherein, r ₁Be the received signal of first reception antenna, r ₂Be the received signal of second reception antenna, q ₁₁=h ₂₂, q ₁₂=-h ₁₂, q ₂₁=-h ₂₁, q ₂₂=h ₁₁

(3) select to separate with the ZF pre-detection N the constellation symbol that Euclidean distance is nearest

As ground floor detection signal x _mThe candidate of judgement travels through symbol, and k represents the parallel branch index;

(4) to k bar parallel branch, the candidate who utilizes step (3) to provide travels through symbol

From received signal vector R, eliminate ground floor detection signal x _mThe interference that is produced, the output column vector y after the ground floor that is eliminated disturbs _k:

$y_k=R-H_1{\hat{x}}_{m,k}$

Wherein, H ₁=H (:, m) the m column vector of expression Channel Transmission matrix H,

Expression ground floor detection signal x _mThe interference that is produced.

(5), utilize output column vector y to k bar parallel branch _kTo second layer detection signal x _nCarry out ZF and detect estimation, obtain the ZF estimated value

${\hat{x}}_{n,k}=\mathrm{HD}\left(\frac{1}{h_{1\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="HDA0000034327140000011.png,HDA0000034327140000012.png"\; state="\{\{state\}\}">n</figure-callout>}}^2+h_{2n}^2}(h_{1n}^{\&prime;}{y}_{k,1}+h_{2n}^{\&prime;}{y}_{k,2})\right)$

Wherein, HD (.) expression hard decision computing,

H ' _1nExpression h _1nConjugation, h ' _2nExpression h _2nConjugation, y _{K, 1}Expression output column vector y _kFirst element, y _{K, 2}Expression output column vector y _kSecond element;

(6) utilize output column vector y _kEstimate with ZF Second layer detection signal x among the cancellation received signal vector R _nInterference, obtain the remaining vector of k bar parallel branch:

${\mathrm{\&epsiv;}}_k=y_k-H_2{\hat{x}}_{n,k}$

In the formula, H ₂=H (:, n) the n column vector of expression Channel Transmission matrix H,

Expression second layer detection signal x _nThe interference that is produced;

(7) utilize remaining vectorial ε _kCalculate Euclidean distance δ _k=|| ε _k|| ², and from N Euclidean distance { δ _k, k=1,2, L selects minimum δ among the N} _kCorresponding parallel branch

Promptly $\hat{k}=\underset{k}{\arg }({\mathrm{\&delta;}}_{\hat{k}}=\min \{{\mathrm{\&delta;}}_k,k=\mathrm{1,2},L,N\left\}\right);$

(8) output the

The candidate of bar parallel branch correspondence travels through symbol Estimate with ZF

Detection finishes.

The present invention is owing to utilize the ZF of ground floor detection signal to estimate evaluation, from the ground floor detection signal might send and filter out the highest N of reliability the signal and send signal and travel through symbol as the candidate of ground floor detection signal, thereby reduced a way and an overall complexity of parallel detection.

Simulation result shows, it is very little that the present invention reduces the detection performance loss that complexity causes, and satisfies the requirement of NGBW communication system high speed processing and performance of BER.

Description of drawings

Fig. 1 is a main realization flow block diagram of the present invention;

Fig. 2 is a transmission/reception antennas ideograph of the present invention;

Fig. 3 is performance simulation figure of the present invention.

Embodiment

Describe the present invention below in conjunction with accompanying drawing.

With reference to accompanying drawing 1, the specific implementation step of detection method of the present invention is as follows:

Step 1:, determine the detection order m of transmitting antenna signal according to the Channel Transmission matrix H.

1.1) determine detection order m by the signal to noise ratio that compares first transmitting antenna and second transmitting antenna, because the signal to noise ratio of transmitting antenna is relevant with the Channel Transmission matrix H, define the signal to noise ratio of first transmitting antenna respectively:

And the signal to noise ratio of second transmitting antenna:

Wherein, matrix G _ZF=H ^-1, subscript-1 representing matrix is inverted G _ZF(j :) representing matrix G _ZFJ capable, j=1,2, || || ²The computing of expression norm, N ₀Expression known noise power,

Represent known transmission signal power;

1.2) signal to noise ratio of first transmitting antenna and second transmitting antenna relatively, if the signal to noise ratio of first transmitting antenna is greater than the signal to noise ratio of second transmitting antenna, then detection order m is 1, on the contrary detection order m is 2, promptly

$m=\left\{\begin{array}{cc}1,& \mathrm{if}{\mathrm{\&gamma;}}_1>{\mathrm{\&gamma;}}_2\\ 2,& \mathrm{else}\end{array}\right.---1)$

1.3) when specifically implementing, formula 1) can simplify, avoid complicated matrix inversion operation, its simplification is the signal to noise ratio γ according to transmitting antenna j _jWith || G _ZF(j :) || ²So the relation of being inversely proportional to is with formula 1) be equivalent to:

$m=\left\{\begin{array}{cc}1,& \mathrm{if}{\left|\right|G_{\mathrm{ZF}}(2,:)\left|\right|}^2>{\left|\right|G_{\mathrm{ZF}}(1,:)\left|\right|}^2\\ 2,& \mathrm{else}\end{array}\right.---2)$

Will

Substitution formula 2), be reduced to:

$m=\left\{\begin{array}{cc}1,& \mathrm{if}{\left|h_{22}\right|}^2+{\left|h_{12}\right|}^2>{\left|h_{21}\right|}^2+{\left|h_{11}\right|}^2\\ 2,& \mathrm{else}\end{array}\right.---3)$

Wherein, h _IjRepresent the channel fading coefficient of transmitting antenna i to reception antenna j, i, j=1,2, as shown in Figure 2, | the determinant of H| representing matrix H;

According to formula 3) can draw: detection order m can compare for passing through in equivalence | h ₂₂| ²+ | h ₁₂| ²With | h ₂₁| ²+ | h ₁₁| ²Size determine.

Step 2: after determining detection order m, utilize received signal vector R and Channel Transmission matrix H to calculate ground floor detection signal x _mThe ZF pre-detection separate

2.1) the ZF pre-detection separates

Provide according to following formula:

$x_m^{\%}=G_{\mathrm{ZF}}(m,:)R---4)$

Wherein, received signal vector R=[r ₁r ₂|| ^T, subscript T representing matrix transposition, r ₁Be the received signal of first reception antenna, r ₂It is the received signal of second reception antenna.

2.2) will

Substitution formula 4) and abbreviation, obtain the ZF pre-detection separate into:

$x_m^{\%}=\frac{1}{\left|H\right|}(q_{m1}{\mathrm{<figure-callout\; id="1"\; label="r"\; filenames="HDA0000034327140000011.png"\; state="\{\{state\}\}">r</figure-callout>}}_1+q_{m2}{r}_2)---5)$

Wherein, q ₁₁=h ₂₂, q ₁₂=-h ₁₂, q ₂₁=-h ₂₁, q ₂₂=h ₁₁

Step 3: select to separate with the ZF pre-detection

Nearest N the constellation symbol of Euclidean distance is as ground floor detection signal x _mThe candidate of judgement travels through symbol.

3.1) calculate each constellation symbol s in the constellation set by following formula _lArrive

Euclidean distance d _{M, l}:

$d_{m,l}={\left|\right|s_l-x_m^{\%}\left|\right|}^2---6)$

Wherein M is given constellation symbol number, and is only relevant with modulation system, l=1, and 2 ..., M;

3.2) M Euclidean distance d of comparison _{M, l}, therefrom select the constellation symbol of N minimum Euclidean distance correspondence, as ground floor detection signal x _mThe candidate travel through symbol, be defined as

K represents the parallel branch index.

Step 4: to k bar parallel branch, the candidate who utilizes step (3) to provide travels through symbol

According to

From received signal vector R, eliminate ground floor detection signal x _mThe interference that is produced, the output column vector y after the ground floor that is eliminated disturbs _k, wherein, H ₁=H (:, m) the m column vector of expression Channel Transmission matrix H,

Expression ground floor detection signal x _mThe interference that is produced.

Step 5:, utilize output column vector y to k bar parallel branch _kTo second layer detection signal x _nCarry out ZF and detect estimation.

5.1) according to formula 8) calculating ZF detected value

$x_{n,k}^{\%}=\frac{{\mathrm{<figure-callout\; id="2"\; label="H"\; filenames="HDA0000034327140000011.png,HDA0000034327140000012.png"\; state="\{\{state\}\}">H</figure-callout>}}_2^{\&prime;}}{{\left|\right|{\mathrm{<figure-callout\; id="2"\; label="H"\; filenames="HDA0000034327140000011.png,HDA0000034327140000012.png"\; state="\{\{state\}\}">H</figure-callout>}}_2\left|\right|}^2}{y}_k---8)$

Wherein, H ₂=H (:, n) the n column vector of expression Channel Transmission matrix H, H ' ₂Expression H ₂Conjugate transpose,

Behind the abbreviation, formula 8) be equivalent to:

$x_{n,k}^{\%}=\frac{1}{{\mathrm{<figure-callout\; id="1"\; label="h"\; filenames="HDA0000034327140000011.png"\; state="\{\{state\}\}">h</figure-callout>}}_{1,\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="HDA0000034327140000011.png,HDA0000034327140000012.png"\; state="\{\{state\}\}">n</figure-callout>}}^2+{\mathrm{<figure-callout\; id="2"\; label="h"\; filenames="HDA0000034327140000011.png,HDA0000034327140000012.png"\; state="\{\{state\}\}">h</figure-callout>}}_{2,n}^2}({\mathrm{<figure-callout\; id="1"\; label="h"\; filenames="HDA0000034327140000011.png"\; state="\{\{state\}\}">h</figure-callout>}}_{1,n}^{\&prime;}{y}_{k,1}+{\mathrm{<figure-callout\; id="2"\; label="h"\; filenames="HDA0000034327140000011.png,HDA0000034327140000012.png"\; state="\{\{state\}\}">h</figure-callout>}}_{2,n}^{\&prime;}{y}_{k,2})---9)$

Wherein, y _{K, 1}Expression output column vector y _kFirst element, y _{K, 2}Expression output column vector y _kSecond element, h ' _1nExpression h _1nConjugation, h ' _2nExpression h _2nConjugation;

5.2) the ZF detected value

By the hard decision computing, obtain the ZF estimated value

Wherein, HD (.) expression hard decision computing.

Step 6: utilize output column vector y _kEstimate with ZF Second layer detection signal x among the cancellation received signal vector R _nInterference, obtain the remaining vectorial ε of k bar parallel branch _k:

${\mathrm{\&epsiv;}}_k=y_k-H_2{\hat{x}}_{n,k}---10)$

In the formula,

Expression second layer detection signal x _nThe interference that is produced.

Step 7: utilize the vectorial ε of the resulting remnants of step 6 _kCalculate its Euclidean distance: δ _k=|| ε _k|| ²And from N Euclidean distance { δ _k, k=1,2, L selects minimum δ among the N} _kThe index of corresponding parallel branch

That is:

$\hat{k}=\underset{k}{\arg }({\mathrm{\&delta;}}_{\hat{k}}=\min \{{\mathrm{\&delta;}}_k,k=\mathrm{1,2},L,N)---11)$

Step 8: output the

The candidate of bar parallel branch correspondence travels through symbol Estimate with ZF

If during m=1, then 2 dimension transmission signal X are estimated as

If during m=2, then 2 dimension transmission signal X are estimated as

Detection finishes.

Effect of the present invention can further specify by some emulation.

Simulated conditions: system uses 22 mimo systems of receiving, and channel adopts the Rayleigh fast fading channel, and modulation system is chosen as 16-QAM (M=16) and 64-QAM (M=64), wherein, to 16-QAM, the candidate travels through symbol numbers N and is chosen as 6,8 respectively, 10 3 kinds of values, to 64QAM, the candidate travels through symbol numbers N and is chosen as 8,12 respectively, 16,32 four kinds of values.

Emulation content and result:

Carry out emulation relatively with the performance that ZF-PD method of the present invention and traditional PD method change with signal to noise ratio bit error rate under above-mentioned simulated conditions, simulation result as shown in Figure 3.

As seen from Figure 3, to 16QAM and two kinds of modulation systems of 64QAM, the candidate travels through symbol numbers N and is respectively 8 and at 12 o'clock, and the present invention compares with traditional PD method, its performance of BER loss can be ignored, and implementation complexity of the present invention only is about 1/2 of a traditional PD method.

Claims (4)

1. the parallel MIMO signal detecting method based on the ZF pre-detection comprises the steps:

(1), determine the detection order m of transmitting antenna signal according to the Channel Transmission matrix H:

$m=\left\{\begin{array}{cc}1,& \mathrm{if}{\left|h_{22}\right|}^2+{\left|h_{12}\right|}^2>{\left|h_{21}\right|}^2+{\left|h_{11}\right|}^2\\ 2,& \mathrm{else}\end{array}\right.;$

Wherein, h _IjExpression transmitting antenna i is to the channel fading coefficient of reception antenna j, i, j=1,2;

(2) utilize received signal vector R and Channel Transmission matrix H to calculate ground floor detection signal x _mThe ZF pre-detection separate

$x_m^{\%}=\frac{1}{\left|H\right|}(q_{m1}{r}_1+q_{m2}{r}_2)$

Wherein, r ₁Be the received signal of first reception antenna, r ₂Be the received signal of second reception antenna, q ₁₁=h ₂₂, q ₁₂=-h ₁₂, q ₂₁=-h ₂₁, q ₂₂=h ₁₁

(3) select to separate with the ZF pre-detection

N the constellation symbol that Euclidean distance is nearest

As ground floor detection signal x _mThe candidate of judgement travels through symbol, and k represents the parallel branch index;

(4) to k bar parallel branch, the candidate who utilizes step (3) to provide travels through symbol

From received signal vector R, eliminate ground floor detection signal x _mThe interference that is produced, the output column vector y after the ground floor that is eliminated disturbs _k

(5), utilize output column vector y to k bar parallel branch _kTo second layer detection signal x _nCarry out ZF and detect estimation, obtain the ZF estimated value

(6) utilize output column vector y _kEstimate with ZF

Second layer detection signal x among the cancellation received signal vector R _nInterference, obtain the remaining vector of k bar parallel branch;

(7) utilize remaining vectorial ε _kCalculate its Euclidean distance δ _k=|| ε _k|| ², and from N Euclidean distance { δ _k, k=1,2, L selects minimum δ among the N} _kCorresponding parallel branch Promptly $\hat{k}=\underset{k}{\arg }({\mathrm{\&delta;}}_{\hat{k}}=\min \{{\mathrm{\&delta;}}_k,k=\mathrm{1,2},L,N\left\}\right);$

(8) output the

The candidate of bar parallel branch correspondence travels through symbol

Estimate with ZF Detection finishes.

2. MIMO signal detecting method according to claim 1, wherein the described candidate who utilizes step (3) to provide of step (4) travels through symbol

From received signal vector R, eliminate ground floor detection signal x _mThe interference that is produced is to be undertaken by following formula:

$y_k=R-H_1{\hat{x}}_{m,k}$

Wherein, y _kThe output column vector after ground floor disturbs, H are eliminated in expression ₁=H (:, m) the m column vector of expression Channel Transmission matrix H,

Expression ground floor detection signal x _mThe interference that is produced.

3. MIMO signal detecting method according to claim 1, the wherein described utilization output of step (5) column vector y _kTo second layer detection signal x _nCarrying out ZF and detect estimation, is to be undertaken by following formula:

${\hat{x}}_{n,k}=\mathrm{HD}\left(\frac{1}{h_{1n}^2+h_{2n}^2}(h_{1n}^{\&prime;}{y}_{k,1}+h_{2n}^{\&prime;}{y}_{k,2})\right)$

Wherein,

Be the ZF estimated value, the computing of HD (.) expression hard decision,

H ' _1nExpression h _1nConjugation, h ' _2nExpression h _2nConjugation, y _{K, 1}Expression output column vector y _kFirst element, y _{K, 2}Expression output column vector y _kSecond element.

4. MIMO signal detecting method according to claim 1, the wherein described utilization output of step (6) column vector y _kEstimate with ZF Second layer detection signal x among the cancellation received signal vector R _nInterference, be to be undertaken by following formula:

${\mathrm{\&epsiv;}}_k=y_k-H_2{\hat{x}}_{n,k}$

In the formula, ε _kBe the remaining vector of k bar parallel branch, H ₂=H (:, n) the n column vector of expression Channel Transmission matrix H,

Expression second layer detection signal x _nThe interference that is produced.

Images