CN103916219A - Estimate-and-forward relay transmission method based on unitary space-time modulation - Google Patents

Abstract

The invention discloses an estimate-and-forward relay transmission method based on the unitary space-time modulation. The method comprises the following steps that under the fast fading channel environment of all unknown channel state messages of nodes in the relay network, a source node sends a unitary space-time symbol to the relay nodes for communication; the kth relay node correspondingly processes the received signal, and relay nodes solve conditional expectation for the source sent signal on the premise of the known received relay signal, and the estimated signal, namely the conditional expectation is forwarded to a target node; when the relay nodes forward the estimated signal, uncertain estimated signals are included, and soft information is provided for the decoding of the target node; based on the signal orthogonal decomposition principle, the source signal and a non-relevant noise signal in the received signal of the target node are distinguished, and the power ratio of the source signal to the non-relevant noise signal is the general signal-to-noise ratio of the target node. According to the scheme, when the relay nodes forward the estimated signal with uncertainty, soft information is provided for decoding of the target node, and the system reliability is improved.

Description

A kind of estimation forward relay transmission method of modulation when empty based on the tenth of the twelve Earthly Branches

Technical field

The present invention relates to a kind of estimation forward relay transmission method of modulation when empty based on the tenth of the twelve Earthly Branches.

Background technology

In environment such as high-speed railway communication, air communications, mobile terminal high-speed mobile, channel variation is very fast, very difficult acquisition channel condition information (CSI) accurately, at this moment relevant empty time-code is no longer applicable.When the tenth of the twelve Earthly Branches is empty, modulation (USTM) technology does not all need known CSI at its transmitting terminal and receiving terminal, and under fast fading channel, error performance is still good.But long-distance transmissions or " shadow effect " can cause signal to noise ratio greatly to reduce, when at this moment the tenth of the twelve Earthly Branches is empty, modulation just loses its superiority.Relaying technique can provide energy at each via node, maintains the communications of longer distance.In tradition trunking scheme, amplification forwarding trunking scheme (AF) is amplified noise simultaneously; Decoding forward relay mode (DF) error propagation phenomenon in the time of low signal-to-noise ratio is serious.Therefore, we pay close attention to emerging soft information forward relay technology.This technology utilizes relay reception signal regeneration to become the forward signal containing soft information, reduces the erroneous effects of hard decision to receiving terminal decoding, improves the communication quality of junction network.

Summary of the invention

The deficiency existing for solving prior art, the invention discloses a kind of estimation forward relay transmission method of modulation when empty based on the tenth of the twelve Earthly Branches, be intended to for parallel many junction networks, the theory that provides first its relay function, the power normalization factor and destination node broad sense signal to noise ratio is derived, and provides destination node interpretation method etc.We can be noted as EF by abridging scheme.This scheme forwards and contains probabilistic estimated signal in the time of via node, for destination node decoding provides soft information, improves system reliability.

In many antenna relays network that this patent is paid close attention to, suppose source node configuration M root transmitting antenna, destination node configuration N root reception antenna, K via node all configures K _rroot antenna.For achieving the above object, concrete scheme of the present invention is as follows:

An estimation forward relay transmission method for modulation when empty based on the tenth of the twelve Earthly Branches, comprises the following steps:

Step 1: under the fast fading channels of the equal unknown channel state information of each node, source node transmission space-time symbol at the tenth of the twelve Earthly Branches to via node communicates in junction network;

Step 2: via node is expected to estimate to source transmitted signal solving condition, and this estimated signal is forwarded to destination node under the prerequisite of known relay reception signal;

Step 3: the principle of utilizing signal in orthogonal to decompose, destination node is received to source signal in signal and separate with uncorrelated noise signaling zone, source signal and the ratio of the power of uncorrelated noise signal are the broad sense signal to noise ratio of destination node.

Broad sense signal to noise ratio is for evaluating the reliability of estimating forward relay transmission.Broad sense signal to noise ratio is higher, and system is more stable.

In described step 1, source node transmission space-time symbol at the tenth of the twelve Earthly Branches to via node communicates, at the T × K of k via node _rdimension receiving matrix R _kfor:

$R_k=\sqrt{{\mathrm{\ρ}}_R}{X}_l{H}_{{\mathrm{SR}}_k}+W_k---\left(1\right)$

In formula, that T × M ties up sending metrix, Φ _lsatisfied unitary matrice, wherein, () ^hthe conjugate transpose computing of representing matrix, I _mrepresent the unit matrix of M × M.M × K _rdimension matrix transmission matrix, T × K _rdimension matrix W _kadditive white Gaussian noise matrix, ρ _rrepresent the receiving terminal signal to noise ratio at all via nodes place, when the tenth of the twelve Earthly Branches is empty, the size of constellation is L=2 ^tR, wherein, T is channel coherence time, and R is data transfer rate, and unit is that bits/channel realizes.

In described step 2, via node forwards and contains probabilistic estimated signal, for destination node decoding provides soft information; Destination node adopts maximum-likelihood decoding criterion to carry out decoding.

In described step 2, via node is processed to received signal, and wherein, k via node is to relay reception signal carry out corresponding relay process, relay function or forward signal are expressed as:

$X_{R_k}=f_{\mathrm{EF}}\left(R_k\right)---\left(2\right)$

In above formula, f _eF(R _k) relay function or forward signal, its expression formula is as follows:

f _EF(R _k)＝β _EF,kE(X|R _k) (3)

${\mathrm{\β}}_{\mathrm{EF},k}=\sqrt{\frac{1}{E\left[{\left|\right|E\left(X\right|R_k)\left|\right|}_F^2\right]}}---\left(4\right)$

Wherein, β _{eF, k}for the power normalization factor, E (X|R _k) be conditional expectation, in formula, symbol || || _frepresent asking F-norm.

Described conditional expectation E (X|R _k) be calculated as follows: supposition source node with equiprobability send the tenth of the twelve Earthly Branches space-time symbol, conditional expectation E (X|R _k) be expressed as:

$\begin{array}{c}E\left(X\right|R_k)=\underset{l=0}{\overset{L-1}{\mathrm{\Σ}}}{X}_l\mathrm{Pr}\left(X_l\right|R_k)\\ =\underset{l=0}{\overset{L-1}{\mathrm{\Σ}}}{X}_l\left(\frac{p\left(R_k\right|X_l)}{{\mathrm{\Σ}}_{l^{\′}=0}^{L-1}p\left(R_k\right|X_{l^{\′}})}\right)\end{array}---\left(5\right)$

Wherein, Pr (X _l| R _k) source space-time symbol at tenth of the twelve Earthly Branches X while representing given relay reception signal _lprobability distribution, p (R _k| X _l) be known source while sending symbol via node receive signal R _kprobability density function.

Described via node receives signal R _kprobability density function can be expressed as:

$p\left(R_k\right|X_l)=\frac{\exp (-{\left|\right|R_k\left|\right|}_F^2+\frac{{\left|\right|X_l^H{R}_k\left|\right|}_F^2}{(T+M/{\mathrm{\ρ}}_R)})}{{\mathrm{\π}}^{\mathrm{TN}}1+{\mathrm{\ρ}}_{R}T/M^{\mathrm{MN}}}---\left(6\right)$

By in formula (6) substitution formula (5), obtain:

$E\left(X\right|R_k)=\frac{{\mathrm{\Σ}}_{l=0}^{L-1}{X}_l\exp \left(\frac{{\left|\right|X_l^H{R}_k\left|\right|}_F^2}{(T+M/{\mathrm{\ρ}}_R)}\right)}{{\mathrm{\Σ}}_{l^{\′}=0}^{L-1}\exp \left(\frac{{\left|\right|X_{l^{\′}}^H{R}_k\left|\right|}_F^2}{(T+M/{\mathrm{\ρ}}_R)}\right)}---\left(7\right)$

Wherein, l' only represents to be different from arbitrary numerical value of l, and l'=0, and 1,2 ..., L-1.

In described step 2, via node is by estimated signal f _eF(R _k) being forwarded to destination node, the gain such as destination node employing receives, destination node T × N dimension receiving matrix Y _eFbe expressed as suc as formula shown in (8):

$Y_{\mathrm{EF}}=\underset{k=1}{\overset{K}{\mathrm{\Σ}}}\sqrt{\frac{{\mathrm{\ρ}}_D}{E{\left|\right|E\left(X\right|R_k)|}_F^2}}E\left(X\right|R_k)H_{R_{k}D}+W_D---\left(8\right)$

In above formula, K _r× N dimension channel coefficient matrix, W _dfor white Gaussian noise matrix, ρ _drepresent the received signal to noise ratio of destination node.

Described destination node adopts maximum-likelihood decoding criterion to carry out decoding, space-time symbol at tenth of the twelve Earthly Branches Φ in gained source after decoding _mLfor:

${\mathrm{\Φ}}_{\mathrm{ML}}=\arg \underset{l}{\max }{\left|\right|Y_{\mathrm{EF}}^H{\mathrm{\Φ}}_l\left|\right|}_F^2---\left(9\right)$

Destination node adopts maximum-likelihood decoding criterion to carry out decoding, and destination node is carried out decoding, reception information.After decoding, just can calculate the error rate, simulate characteristic curve of error code.

In described step 3, only consider that each node in junction network all configures the situation of single antenna, comprise signal X when empty at source node tenth of the twelve Earthly Branches _ldestination node receive signal Y _eFbe expressed as:

$Y_{\mathrm{EF}}=\frac{E\left(X_l^H{Y}_{\mathrm{EF}}\right)}{E\left(X_l^H{X}_l\right)}(X_l+U_{\mathrm{EF}})---\left(10\right)$

In formula, U _eFrepresent to send symbol X with source _lincoherent error matrix, will in formula be defined as scale factor, so, can obtain,

$U_{\mathrm{EF}}=\frac{Y_{\mathrm{EF}}}{{\mathrm{\α}}_{\mathrm{EF}}}-X_l---\left(11\right)$

Broad sense signal to noise ratio is expressed as:

$\mathrm{GSNR}=\frac{E\left(X_l^H{X}_l\right)}{E\left(U_{\mathrm{EF}}^H{U}_{\mathrm{EF}}\right)}=\frac{T}{E\left({(\frac{Y_{\mathrm{EF}}}{{\mathrm{\α}}_{\mathrm{EF}}}-X_l)}^H(\frac{Y_{\mathrm{EF}}}{{\mathrm{\α}}_{\mathrm{EF}}}-X_l)\right)}---\left(12\right)$

Described scale factor _eFwith irrelevant error matrix U _eFbe expressed as:

${\mathrm{\α}}_{\mathrm{EF}}=\frac{1}{T}\sqrt{{\mathrm{\ρ}}_D}\underset{k=1}{\overset{K}{\mathrm{\Σ}}}\left(\frac{{\mathrm{\β}}_{\mathrm{EF},k}{h}_{R_{k}D}{\mathrm{\Σ}}_{l^{\′}=0}^{L-1}E\left(X_l^H{X}_{l^{\′}}\right)\exp \left(\frac{{\left|\right|X_{l^{\′}}^H{R}_k\left|\right|}_F^2}{(T+M/{\mathrm{\ρ}}_R)}\right)}{{\mathrm{\Σ}}_{l^{\′}=0}^{L-1}\exp \left(\frac{{\left|\right|X_{l^{\′}}^H{R}_k\left|\right|}_F^2}{(T+M/{\mathrm{\ρ}}_R)}\right)}\right)---\left(14\right)$

Wherein, represent the channel fading letter factor from k via node to destination node.Due to the restriction of mathematical theory, in the theoretical derivation of broad sense signal to noise ratio, we only consider that each node all configures the situation of single antenna, so locate no longer a matrix, but a real number.

$U_{\mathrm{EF}}=\frac{\sqrt{{\mathrm{\ρ}}_D}\underset{k=1}{\overset{K}{\mathrm{\Σ}}}\left(\frac{{\mathrm{\β}}_{\mathrm{EF},k}{h}_{R_{k}D}{\mathrm{\Σ}}_{l^{\′}=0}^{L-1}\exp \left(\frac{{\left|\right|X_{l^{\′}}^H{R}_k\left|\right|}_F^2}{(T+M/{\mathrm{\ρ}}_R)}\right)}{{\mathrm{\Σ}}_{l^{\′}=0}^{L-1}\exp \left(\frac{{\left|\right|X_{l^{\′}}^H{R}_k\left|\right|}_F^2}{(T+M/{\mathrm{\ρ}}_R)}\right)}\right)+W_D}{{\mathrm{\α}}_{\mathrm{EF}}}-X_l---\left(15\right)$

By in formula (14) and formula (15) substitution formula (13), obtain the receiving terminal broad sense signal to noise ratio of the EF scheme based on USTM.

Beneficial effect of the present invention:

The present invention is very suitable for many antenna relays system of fast-fading environment.Now, in junction network, source node, each via node and destination node all do not need known channel state information.The present invention provides first the theory of its relay function, the power normalization factor and destination node broad sense signal to noise ratio and derives, and provides destination node interpretation method etc.This scheme forwards and contains probabilistic estimated signal in the time of via node, for destination node decoding provides soft information, improves system reliability.In addition, compared with two kinds of traditional relay forwarding schemes based on USTM, amplification forwarding (AF) forwards (DF) scheme with decoding, and EF scheme stability of the present invention is better.

Brief description of the drawings

Fig. 1 is the system model of many relayings parallel network;

Fig. 2 is the communication block diagram of the many relayings parallel network based on USTM;

Fig. 3 is the relay process block diagram of the EF scheme based on USTM that proposes of the present invention;

Fig. 4 is carried EF and AF, DF scheme bit error rate simulation curve;

Fig. 5 is carried EF scheme and AF, DF scheme broad sense signal to noise ratio simulation curve;

When Fig. 6 is number of antennas variation, the EF that carries of institute and AF, DF scheme bit error rate curve comparison diagram;

Fig. 7 is while changing coherence time, the EF that carries of institute and AF, DF scheme bit error rate curve comparison diagram.

Embodiment:

Below in conjunction with accompanying drawing, the present invention is described in detail:

The system model of many relayings parallel system as shown in Figure 1.Many relay systems are by source node S, destination node D and K parallel via node R ₁, R ₂..., R _kform.Suppose source node configuration M root transmitting antenna, destination node configuration N root reception antenna, each via node all configures K _rroot antenna.When the tenth of the twelve Earthly Branches is empty, the generation of constellation adopts the Systematic Method based on DFT.

Taking k via node as example, the communication block diagram of the many antenna relays system based on USTM as shown in Figure 2.

First stage, source node transmission space-time symbol at the tenth of the twelve Earthly Branches to each via node communicates, at the T × K of k via node _rdimension receiving matrix R _kcan be written as:

$R_k=\sqrt{{\mathrm{\ρ}}_R}{X}_l{H}_{{\mathrm{SR}}_k}+W_k---\left(1\right)$

In formula, that T × M ties up sending metrix, M × K _rdimension matrix transmission matrix, T × K _rdimension matrix W _kit is additive white Gaussian noise matrix.In addition ρ, _rrepresent the receiving terminal signal to noise ratio at all via nodes place.When the tenth of the twelve Earthly Branches is empty, the size of constellation is L=2 ^tR, wherein, R is data transfer rate, unit is that bits/channel realizes.

Second stage, k via node processed to received signal accordingly, and this can be expressed as:

$X_{R_k}=f\left(R_k\right)---\left(2\right)$

In above formula, f () represents corresponding relay function.T × N dimension receiving matrix Y can be expressed as

$Y=\underset{k=1}{\overset{K}{\mathrm{\Σ}}}\sqrt{{\mathrm{\ρ}}_D}f\left(R_k\right)H_{R_{k}D}+W_D---\left(3\right)$

In above formula, T × K _rdimension matrix f (R _k) be relay function, K _r× N dimension channel coefficient matrix, W _dfor white Gaussian noise matrix, ρ _drepresent the received signal to noise ratio of destination node.

The present invention proposes the estimation forward relay scheme based on USTM, can note the scheme into EF by abridging.The soft information processing block diagram that via node carries out as shown in Figure 3.Estimated signal to source transmitted signal when via node forwards known reception signal, the relay function of this scheme is:

f _EF(R _k)＝β _EF,kE(X|R _k) (4)

${\mathrm{\β}}_{\mathrm{EF},k}=\sqrt{\frac{1}{E\left[{\left|\right|E\left(X\right|R_k)\left|\right|}_F^2\right]}}---\left(5\right)$

Wherein, β _{eF, k}for the power normalization factor.

Conditional expectation E (X|R in formula (4) _k) be calculated as follows.Suppose that source node sends space-time symbol at the tenth of the twelve Earthly Branches, conditional expectation E (X|R so with equiprobability _k) can be expressed as:

$\begin{array}{c}E\left(X\right|R_k)=\underset{l=0}{\overset{L-1}{\mathrm{\Σ}}}{X}_l\mathrm{Pr}\left(X_l\right|R_k)\\ =\underset{l=0}{\overset{L-1}{\mathrm{\Σ}}}{X}_l\left(\frac{p\left(R_k\right|X_l)}{{\mathrm{\Σ}}_{l^{\′}=0}^{L-1}p\left(R_k\right|X_{l^{\′}})}\right)\end{array}---\left(6\right)$

For given space-time symbol X at the tenth of the twelve Earthly Branches _l, via node receives signal R _kprobability density function can be expressed as

$p\left(R_k\right|X_l)=\frac{\exp (-{\left|\right|R_k\left|\right|}_F^2+\frac{{\left|\right|X_l^H{R}_k\left|\right|}_F^2}{(T+M/{\mathrm{\ρ}}_R)})}{{\mathrm{\π}}^{\mathrm{TN}}1+{\mathrm{\ρ}}_{R}T/M^{\mathrm{MN}}}---\left(7\right)$

By in formula (7) substitution formula (6), can obtain

$E\left(X\right|R_k)=\frac{{\mathrm{\Σ}}_{l=0}^{L-1}{X}_l\exp \left(\frac{{\left|\right|X_l^H{R}_k\left|\right|}_F^2}{(T+M/{\mathrm{\ρ}}_R)}\right)}{{\mathrm{\Σ}}_{l^{\′}=0}^{L-1}\exp \left(\frac{{\left|\right|X_{l^{\′}}^H{R}_k\left|\right|}_F^2}{(T+M/{\mathrm{\ρ}}_R)}\right)}---\left(8\right)$

(4) and (5) are updated in (3), and destination node receives signal and can be expressed as suc as formula shown in (9).

$Y_{\mathrm{EF}}=\underset{k=1}{\overset{K}{\mathrm{\Σ}}}\sqrt{\frac{{\mathrm{\ρ}}_D}{E{\left|\right|E\left(X\right|R_k)|}_F^2}}E\left(X\right|R_k)H_{R_{k}D}+W_D---\left(9\right)$

Destination node adopts maximum-likelihood decoding criterion to carry out decoding, as follows:

${\mathrm{\Φ}}_{\mathrm{ML}}=\arg \underset{l}{\max }{\left|\right|Y_{\mathrm{EF}}^H{\mathrm{\Φ}}_l\left|\right|}_F^2---\left(10\right)$

The definition of broad sense signal to noise ratio during according to employing BPSK modulation, source signal and uncorrelated noise signaling zone that we need to receive destination node in signal separate, and then calculate broad sense signal to noise ratio by the ratio of both power.Therefore signal X while, comprising source node sky at the tenth of the twelve Earthly Branches _ldestination node receive signal Y _eFcan be expressed as follows formula.Hereby indicate: the present invention only considers that each node in junction network all configures the situation of single antenna.

$Y_{\mathrm{EF}}=\frac{E\left(X_l^H{Y}_{\mathrm{EF}}\right)}{E\left(X_l^H{X}_l\right)}(X_l+U_{\mathrm{EF}})---\left(11\right)$

In formula, U _eFrepresent to send symbol X with source _lincoherent error matrix.Will in formula be defined as scale factor.So, can obtain,

$U_{\mathrm{EF}}=\frac{Y_{\mathrm{EF}}}{{\mathrm{\α}}_{\mathrm{EF}}}-X_l---\left(12\right)$

For many relay systems of considering herein, easily obtain therefore,, in the parallel many relay systems based on USTM, broad sense signal to noise ratio can be expressed as

$\mathrm{GSNR}=\frac{E\left(X_l^H{X}_l\right)}{E\left(U_{\mathrm{EF}}^H{U}_{\mathrm{EF}}\right)}=\frac{T}{E\left({(\frac{Y_{\mathrm{EF}}}{{\mathrm{\α}}_{\mathrm{EF}}}-X_l)}^H(\frac{Y_{\mathrm{EF}}}{{\mathrm{\α}}_{\mathrm{EF}}}-X_l)\right)}---\left(13\right)$

According to the difference of selected relay forwarding mode, the present invention provides the analytical expression of corresponding destination node broad sense signal to noise ratio.

For the EF scheme based on USTM, scale factor _eFwith irrelevant error matrix U _eFcan be expressed as:

${\mathrm{\α}}_{\mathrm{EF}}=\frac{1}{T}\sqrt{{\mathrm{\ρ}}_D}\underset{k=1}{\overset{K}{\mathrm{\Σ}}}\left(\frac{{\mathrm{\β}}_{\mathrm{EF},k}{h}_{R_{k}D}{\mathrm{\Σ}}_{l^{\′}=0}^{L-1}E\left(X_l^H{X}_{l^{\′}}\right)\exp \left(\frac{{\left|\right|X_{l^{\′}}^H{R}_k\left|\right|}_F^2}{(T+M/{\mathrm{\ρ}}_R)}\right)}{{\mathrm{\Σ}}_{l^{\′}=0}^{L-1}\exp \left(\frac{{\left|\right|X_{l^{\′}}^H{R}_k\left|\right|}_F^2}{(T+M/{\mathrm{\ρ}}_R)}\right)}\right)---\left(16\right)$

$U_{\mathrm{EF}}=\frac{\sqrt{{\mathrm{\ρ}}_D}\underset{k=1}{\overset{K}{\mathrm{\Σ}}}\left(\frac{{\mathrm{\β}}_{\mathrm{EF},k}{h}_{R_{k}D}{\mathrm{\Σ}}_{l^{\′}=0}^{L-1}\exp \left(\frac{{\left|\right|X_{l^{\′}}^H{R}_k\left|\right|}_F^2}{(T+M/{\mathrm{\ρ}}_R)}\right)}{{\mathrm{\Σ}}_{l^{\′}=0}^{L-1}\exp \left(\frac{{\left|\right|X_{l^{\′}}^H{R}_k\left|\right|}_F^2}{(T+M/{\mathrm{\ρ}}_R)}\right)}\right)+W_D}{{\mathrm{\α}}_{\mathrm{EF}}}-X_l---\left(17\right)$

By in formula (16) and formula (17) substitution formula (13), can obtain the receiving terminal broad sense signal to noise ratio of the EF scheme based on USTM.

But, in tradition decoding pass-through mode (DF), because via node need to carry out hard decision to relay reception signal, so the forward signal of regeneration may be the specific function of source space-time symbol at the tenth of the twelve Earthly Branches hardly, the expression formula that obtain broad sense signal to noise ratio is very difficult.But we can obtain its broad sense signal-to-noise performance performance by MATLAB emulation.

In conjunction with invention Fig. 3, the EF scheme based on USTM proposed by the invention and traditional AF, DF scheme are carried out to MATLAB emulation.When emulation, average statistical is truly difficult to try to achieve, and we utilize time average to replace average statistical to carry out design conditions expectation, and the content during the time average of expectation can be accorded with by desired operation is determined divided by the size of considered source symbolic blocks.If there is no specified otherwise, all properties simulation result in the present invention all obtains in following situation.As shown in Figure 2, the number of antennas of source node, each via node and destination node configuration is respectively K _r=M=1, N=1, be T=2 coherence time, relaying number K=1, and data transfer rate is the realization of R=1 bits/channel.Can be obtained when coherence time is when T=2 by search completely, the u value sequence that can generate optimal constellation is u=[1,3]; When coherence time is when T=5, the u value sequence that can generate optimal constellation is u=[1,3,6,7,16].

Utilize above-mentioned thought, according to formula (4)～(10) and, we can obtain the simulation curve of the bit error rate of carried EF and traditional AF, DF scheme, as shown in Figure 4.Can find out, for adopting in the parallel double bounce junction network of USTM, compared with AF, DF scheme, when bit error rate is 10 ^-2time, the EF scheme of carrying can obtain respectively the snr gain of about 1.5dB and 2dB.According to formula (16)～(17), we can obtain the simulation curve of EF scheme broad sense signal-to-noise performance, as shown in Figure 5.Can find out, compared with other two schemes, the EF scheme of carrying can obtain higher broad sense signal to noise ratio in destination node.

In addition, for the parallel many relay systems based on USTM, the present invention still carries out emulation on factors such as number of antennas and channel coherence times to the impact of systematic function, respectively as shown in Figure 6 and Figure 7.Can be known by Fig. 6, M=2, in the situation of N=2, with M=1, the situation of N=1 is compared, when bit error rate is 10 ^-3time, put forward EF scheme and can obtain the snr gain of about 12dB.Fig. 7 shows that working as bit error rate is 10 ^-2time, when channel T=5 coherence time, compared with the situation of T=2, put forward EF scheme and can reach the snr gain of about 4dB.

Claims (10)

1. when empty based on the tenth of the twelve Earthly Branches, an estimation forward relay transmission method for modulation, is characterized in that, comprises the following steps:

Step 1: under the fast fading channels of the equal unknown channel state information of each node, source node transmission space-time symbol at the tenth of the twelve Earthly Branches to via node communicates in junction network;

Step 2: via node is expected to estimate to source transmitted signal solving condition, and this estimated signal is forwarded to destination node under the prerequisite of known relay reception signal;

Step 3: the principle of utilizing signal in orthogonal to decompose, destination node is received to source signal in signal and separate with uncorrelated noise signaling zone, source signal and the ratio of the power of uncorrelated noise signal are the broad sense signal to noise ratio of destination node.

2. the estimation forward relay transmission method of a kind of modulation when empty based on the tenth of the twelve Earthly Branches as claimed in claim 1, is characterized in that, in described step 1, source node sends tenth of the twelve Earthly Branches space-time symbol and communicates to via node, at the T × K of k via node _bdimension receiving matrix R _kfor:

$R_k=\sqrt{{\mathrm{\ρ}}_R}{X}_l{H}_{{\mathrm{SR}}_k}+W_k---\left(1\right)$

In formula, that T × M ties up sending metrix, Φ _lsatisfied unitary matrice, wherein, () ^hthe conjugate transpose computing of representing matrix, I _mrepresent the unit matrix of M × M, M × K _rdimension matrix transmission matrix, T × K _rdimension matrix W _kadditive white Gaussian noise matrix, ρ _rrepresent the receiving terminal signal to noise ratio at all via nodes place, when the tenth of the twelve Earthly Branches is empty, the size of constellation is L=2 ^tR, wherein, T is channel coherence time, and R is data transfer rate, and unit is that bits/channel realizes.

3. the estimation forward relay transmission method of a kind of modulation when empty based on the tenth of the twelve Earthly Branches as claimed in claim 1, is characterized in that, in described step 2, via node is processed to received signal, and wherein, k via node is to relay reception signal carry out corresponding relay process, relay function or forward signal f _eF(R _k) be expressed as:

$X_{R_k}=f_{\mathrm{EF}}\left(R_k\right)---\left(2\right)$

In above formula, relay function or forward signal f _eF(R _k) expression formula as follows:

f _EF(R _k)＝β _EF,kE(X|R _k) (3)

${\mathrm{\β}}_{\mathrm{EF},k}=\sqrt{\frac{1}{E\left[{\left|\right|E\left(X\right|R_k)\left|\right|}_F^2\right]}}---\left(4\right)$

Wherein, β _{eF, k}for the power normalization factor, E (X|R _k) be conditional expectation.

4. the estimation forward relay transmission method of a kind of modulation when empty based on the tenth of the twelve Earthly Branches as claimed in claim 3, is characterized in that, supposes that source node sends space-time symbol at the tenth of the twelve Earthly Branches, conditional expectation E (X|R with equiprobability _k) be expressed as:

$\begin{array}{c}E\left(X\right|R_k)=\underset{l=0}{\overset{L-1}{\mathrm{\Σ}}}{X}_l\mathrm{Pr}\left(X_l\right|R_k)\\ =\underset{l=0}{\overset{L-1}{\mathrm{\Σ}}}{X}_l\left(\frac{p\left(R_k\right|X_l)}{{\mathrm{\Σ}}_{l^{\′}=0}^{L-1}p\left(R_k\right|X_{l^{\′}})}\right)\end{array}---\left(5\right)$

Wherein, X _lfor the space-time symbol at the tenth of the twelve Earthly Branches that source node sends, Pr (X _l| R _k) source space-time symbol at tenth of the twelve Earthly Branches X while representing given relay reception signal _lprobability distribution, p (R _k| X _l) be via node reception signal R _kconditional probability density function.

5. the estimation forward relay transmission method of a kind of modulation when empty based on the tenth of the twelve Earthly Branches as claimed in claim 4, is characterized in that, described via node receives signal R _kconditional probability density function p (R _k| X _l) can be expressed as:

$p\left(R_k\right|X_l)=\frac{\exp (-{\left|\right|R_k\left|\right|}_F^2+\frac{{\left|\right|X_l^H{R}_k\left|\right|}_F^2}{(T+M/{\mathrm{\ρ}}_R)})}{{\mathrm{\π}}^{\mathrm{TN}}1+{\mathrm{\ρ}}_{R}T/M^{\mathrm{MN}}}---\left(6\right)$

By in formula (6) substitution formula (5), obtain:

$E\left(X\right|R_k)=\frac{{\mathrm{\Σ}}_{l=0}^{L-1}{X}_l\exp \left(\frac{{\left|\right|X_l^H{R}_k\left|\right|}_F^2}{(T+M/{\mathrm{\ρ}}_R)}\right)}{{\mathrm{\Σ}}_{l^{\′}=0}^{L-1}\exp \left(\frac{{\left|\right|X_{l^{\′}}^H{R}_k\left|\right|}_F^2}{(T+M/{\mathrm{\ρ}}_R)}\right)}---\left(7\right)$

Wherein, l' only represents to be different from arbitrary numerical value of l, and l'0, and 1,2 ..., L-1.

6. the estimation forward relay transmission method of a kind of modulation when empty based on the tenth of the twelve Earthly Branches as claimed in claim 1, is characterized in that, in described step 2, via node is by estimated signal f _eF(R _k) being forwarded to destination node, the gain such as destination node employing receives, destination node T × N dimension receiving matrix Y _eFbe expressed as suc as formula shown in (8):

$Y_{\mathrm{EF}}=\underset{k=1}{\overset{K}{\mathrm{\Σ}}}\sqrt{\frac{{\mathrm{\ρ}}_D}{E{\left|\right|E\left(X\right|R_k)|}_F^2}}E\left(X\right|R_k)H_{R_{k}D}+W_D---\left(8\right)$

In above formula, K _r× N dimension channel coefficient matrix, W _dfor white Gaussian noise matrix, ρ _drepresent the received signal to noise ratio of destination node.

7. the estimation forward relay transmission method of a kind of modulation when empty based on the tenth of the twelve Earthly Branches as claimed in claim 1, is characterized in that, in described step 2, via node forwards and contains probabilistic estimated signal, for destination node decoding provides soft information; Destination node adopts maximum-likelihood decoding criterion to carry out decoding.

8. the estimation forward relay transmission method of a kind of modulation when empty based on the tenth of the twelve Earthly Branches as claimed in claim 7, is characterized in that, described destination node adopts maximum-likelihood decoding criterion to carry out decoding, space-time symbol at tenth of the twelve Earthly Branches Φ in gained source after decoding _mLfor:

${\mathrm{\Φ}}_{\mathrm{ML}}=\arg \underset{l}{\max }{\left|\right|Y_{\mathrm{EF}}^H{\mathrm{\Φ}}_l\left|\right|}_F^2---\left(9\right)$

Destination node adopts maximum-likelihood decoding criterion to carry out decoding, and destination node is carried out decoding, reception information.

9. the estimation forward relay transmission method of a kind of modulation when empty based on the tenth of the twelve Earthly Branches as claimed in claim 1, is characterized in that, in described step 3, only considers that each node in junction network all configures the situation of single antenna, comprises signal X when empty at source node tenth of the twelve Earthly Branches _ldestination node receive signal Y _eFbe expressed as:

$Y_{\mathrm{EF}}=\frac{E\left(X_l^H{Y}_{\mathrm{EF}}\right)}{E\left(X_l^H{X}_l\right)}(X_l+U_{\mathrm{EF}})---\left(10\right)$

In formula, U _eFrepresent to send symbol X with source _lincoherent error matrix, will in formula be defined as scale factor, so, can obtain,

$U_{\mathrm{EF}}=\frac{Y_{\mathrm{EF}}}{{\mathrm{\α}}_{\mathrm{EF}}}-X_l---\left(11\right)$

Broad sense signal to noise ratio is expressed as:

$\mathrm{GSNR}=\frac{E\left(X_l^H{X}_l\right)}{E\left(U_{\mathrm{EF}}^H{U}_{\mathrm{EF}}\right)}=\frac{T}{E\left({(\frac{Y_{\mathrm{EF}}}{{\mathrm{\α}}_{\mathrm{EF}}}-X_l)}^H(\frac{Y_{\mathrm{EF}}}{{\mathrm{\α}}_{\mathrm{EF}}}-X_l)\right)}---\left(12\right).$

10. the estimation forward relay transmission method of a kind of modulation when empty based on the tenth of the twelve Earthly Branches as claimed in claim 9, is characterized in that described scale factor _eFwith irrelevant error matrix U _eFbe expressed as:

${\mathrm{\α}}_{\mathrm{EF}}=\frac{1}{T}\sqrt{{\mathrm{\ρ}}_D}\underset{k=1}{\overset{K}{\mathrm{\Σ}}}\left(\frac{{\mathrm{\β}}_{\mathrm{EF},k}{h}_{R_{k}D}{\mathrm{\Σ}}_{l^{\′}=0}^{L-1}E\left(X_l^H{X}_{l^{\′}}\right)\exp \left(\frac{{\left|\right|X_{l^{\′}}^H{R}_k\left|\right|}_F^2}{(T+M/{\mathrm{\ρ}}_R)}\right)}{{\mathrm{\Σ}}_{l^{\′}=0}^{L-1}\exp \left(\frac{{\left|\right|X_{l^{\′}}^H{R}_k\left|\right|}_F^2}{(T+M/{\mathrm{\ρ}}_R)}\right)}\right)---\left(14\right)$

Wherein, represent the channel fading letter factor from k via node to destination node;

$U_{\mathrm{EF}}=\frac{\sqrt{{\mathrm{\ρ}}_D}\underset{k=1}{\overset{K}{\mathrm{\Σ}}}\left(\frac{{\mathrm{\β}}_{\mathrm{EF},k}{h}_{R_{k}D}{\mathrm{\Σ}}_{l^{\′}=0}^{L-1}\exp \left(\frac{{\left|\right|X_{l^{\′}}^H{R}_k\left|\right|}_F^2}{(T+M/{\mathrm{\ρ}}_R)}\right)}{{\mathrm{\Σ}}_{l^{\′}=0}^{L-1}\exp \left(\frac{{\left|\right|X_{l^{\′}}^H{R}_k\left|\right|}_F^2}{(T+M/{\mathrm{\ρ}}_R)}\right)}\right)+W_D}{{\mathrm{\α}}_{\mathrm{EF}}}-X_l---\left(15\right)$

By in formula (14) and formula (15) substitution formula (13), obtain the receiving terminal broad sense signal to noise ratio of the EF scheme based on USTM.