CN102882570A - Optimum transceiving combined processing method for communication among equipment in mobile communication network - Google Patents

Abstract

The invention discloses an optimum transceiving combined processing method for communication among equipment in the mobile communication network. The optimum transceiving combined processing method includes that firstly, a base station in a time-division duplex communication system can equivalently acquire downlink channel information to UE 3 (user equipment 3) and downlink interference channel information to a D2D (device to device) receiving users by estimating an uplink channel; UE 1 can acquire transmission channel information to UE2 and interference channel information to the UE3 by the similar channel estimation method, and feed the same back to the base station; secondly, the base station computes precoding matrixes and receiving matrixes from the base station to the UE3 and the UE1 as well as from the base station to the UE2 as the minimized mean square error principle or maximized channel capacity principle according to the acquired four kinds of channel information; and finally, the base station sends the precoding matrixes and the receiving matrixes which are computed to notify the UE1, the UE2 and the UE3 via control channels.

Description

The optimum transmit-receive combination processing method of communication between devices under the mobile communications network

Technical field

The present invention relates to the optimum transmit-receive combination processing method of communication between devices under a kind of mobile communications network, belong to the transmit-receive combination treatment design field in many antenna interference channel.

Background technology

Along with the further evolution development of LTE-A, merged the Novel movable communication technologys such as heterogeneous network, multi-point cooperative transmission (CoMP) to existing mobile network and caused widely concern in recent years.Device-to-Device (D2D) communication is a kind of new technique of directly communicating between the terminal of local close together of allowing.Under the control of existing mobile communication system, can increase frequency spectrum efficiency of cellular communication system by multiplexing local resource, reduce terminal transmit power, solve to a certain extent the problem of wireless communication system frequency spectrum resource scarcity.With respect to technology such as existing wireless local local area network (LAN) and bluetooths, the D2D communication device works is in licensed band, thereby not only can adopt flexibly the multiple resources allocation strategy can also effectively suppress to disturb the reliability that guarantees communication, promotes user's service quality.Although the D2D communication technology has lot of advantages, owing to can produce interference to former cell signal at the D2D signal that transmits with frequency range, even might reduce cell capacity thereby cause the availability of frequency spectrum to descend.

In recent years, multi-antenna technology (MIMO) has been subject to researcher's extensive concern, this technology can utilize spatial degrees of freedom that many antennas of transmitting terminal and receiving terminal provide to if having time and frequency resource carry out multiplexing on the space, thereby significantly improve the utilization ratio of frequency spectrum resource and the message capacity of system.On the other hand, multi-antenna technology can utilize spatial degrees of freedom to suppress co-channel interference.

In traditional multi-user MIMO system, several optimized algorithms based on MIMO technology inhibition multi-user interference are widely used in point-to-point channel and two hop channels.The system index of optimizing includes the mean square error (MSE) that receives signal, power system capacity (Sum-Rate) etc.The algorithm that also has some suboptimums is zero forcing algorithm (ZF) and maximize letter and let out and make an uproar than (SLNR) algorithm for example, zero forcing algorithm is mapped completely to each user's signal in the kernel of interference channel by preliminary treatment, can eliminate like this other interference user fully to targeted customer's interference.SLNR then will be traditionally to the optimization of Signal to Interference plus Noise Ratio (SINR) index change into to letter reveal make an uproar than optimization, with can be in the hope of closed solutions after the optimization problem uncoupling, thereby greatly reduce the computation complexity of optimization problem, the performance of SLNR is not a lot of losses but.

The interference that seldom has research to pay close attention to based on multi-antenna technology for the communication scenes that comprises D2D suppresses problem.Recently, there is the researcher to design a kind of base station that suppresses fully based on ZF thought D2D is received the scheme that the user disturbs.The thought of this scheme is exactly that the base station is sent to mobile radio network user's signal map in the interference channel kernel of base station to D2D reception user.Owing to suppressed to come from the interference of mobile radio network link fully, the D2D link can maximum power send data.This scheme Shortcomings is exactly to consider that not D2D sends the user to mobile radio network user's interference problem.This patent will utilize all transmission channels and interference channel information, synthetically optimize transmission pre-coding matrix and the receiving matrix of mobile radio network link and D2D link, reach and minimize the residential quarter and receive the signal mean square error to reduce the target that receives bit error rate and maximize two links and speed.Performance to real system improves a lot.

Summary of the invention

The present invention is directed to the deficiencies in the prior art, the optimum transmit-receive combination processing method of communication between devices under a kind of mobile communications network is provided, thereby the mean square error of the reception signal that the specific implementation of combined optimization can be by making mobile radio network link and D2D link minimizes and reduces bit error rate, and that perhaps optimizes whole system makes it maximization with speed.

The technical solution used in the present invention is: the optimum transmit-receive combination processing method of communication between devices under a kind of mobile communications network, and the method is carried out according to the following steps:

1) D2D (Device-to-Device) transmission user UE1 (User Equipment) and D2D reception user UE2 have respectively N ₁Root transmitting antenna and M ₁The root reception antenna, the base station has N ₂The root transmitting antenna, network user UE3 has M ₂The root reception antenna.The transmitted power that D2D sends user and base station is respectively P ₁, P ₂, they send respectively multi-stream data x ₁And x ₂(dimension is respectively S ₁* 1 and S ₂* 1) gives separately receiver.S ₁And S ₂The data fluxion that represents respectively D2D link and mobile radio network link transmission.

2) D2D link and mobile radio network link adopt nonopiate mode to share running time-frequency resource, namely D2D equipment between physical channel and certain mobile radio network user's the physical resource of communication overlapping.In TDD (Time Division Duplexing) system, realize optimum transmitting-receiving design in order to make the base station obtain complete channel information, the base station can obtain the base station to the down channel H of UE3 by uplink channel estimation ₂And the base station is to the descending interference channel H of UE2 _2,1By above-mentioned identical channel estimating mode, UE1 can obtain it equally to the transmission channel H of UE2 ₁And the interference channel H between it and the UE3 _1,2Last UE1 is with H ₁And H _1,2Feed back to the base station.

3) base station receives signal mean square error reduction bit error rate or the optimization of maximum channel channel capacity criterion base station, the pre-coding matrix of UE1 and the receiving matrix of UE2, UE3 according to the four complete class channel informations that obtain by minimizing.Because the protruding problem of above-mentioned optimization problem right and wrong and do not have closed solutions is so this patent provides two kinds of optional iterative algorithms.Be convenient narration, will minimize reception signal mean square error iterative algorithm and be called algorithm one, and maximum channel capacity algorithm is called algorithm two.The iterative calculation method of algorithm one is as follows:

Step 1: make iterations variable n=0, according to transmission channel H ₁And H ₂The singular value decomposition initializes sends pre-coding matrix

U=1,2; Subscript u is the variable of sign link, and wherein u=1 and u=2 represent respectively D2D link (communication link between UE1 and the UE2) and mobile network's down link (base station is to the communication link of UE3).

Then step 2:n=n+1 calculates the receiving matrix of the n+1 time iterative process link u according to following formula (2) $F_u^{(n+1)};$

Step 3: step 2 is calculated

Formula (3) is tried to achieve parameter below the substitution power constraint condition

Wherein

That guaranteed output constraints is set up and the Lagrange's multiplier of introducing;

Step 4: with what calculate in step 2 and the step 3

With

Formula below the substitution (1) can calculate the transmission pre-coding matrix in the iterative process the n+1 time

Step 5: repeating step 2 to step 4 until With Convergence.

$W_u^{(n+1)}={(H_u^H{F}_u^{(n+1)}{F}_u^{H,(n+1)}{H}_u+H_{u,m}^H{F}_m^{(n+1)}{F}_m^{H,(n+1)}{H}_{u,m}+{\mathrm{\μ}}_u^{(n+1)}I)}^{-1}{H}_u^H{F}_u^{(n+1)}---\left(1\right)$

$F_u^{(n+1)}={(H_u{W}_u^{\left(n\right)}{W}_u^{H,\left(n\right)}{H}_u^H+H_{m,u}^H{W}_m^{\left(n\right)}{W}_m^{H,\left(n\right)}{H}_{m,u}+{\mathrm{\σ}}_u^{2}I)}^{-1}{H}_u{W}_u^{\left(n\right)}---\left(2\right)$

$\underset{i=1}{\overset{N_u}{\mathrm{\Σ}}}\frac{G_u^{\left[\mathrm{tt}\right],(n+1)}}{[{\mathrm{\μ}}_u^{(n+1)}+{\mathrm{\Δ}}_u^{\left[\mathrm{tt}\right],(n+1)}]}=\mathrm{Pu}---\left(3\right)$

Formula (1) all is variable and the u of sign link to the middle subscript u of formula (3) and m, m ∈ 1,2}, u ≠ m, and n represents the iterations variable, the conjugate transpose operation of capitalization H representing matrix, I representation unit battle array.

With

Represented respectively matrix

With

T element on the leading diagonal.Wherein, $G_u^{(n+1)}=U_u^{H,(n+1)}{H}_u^H{F}_u^{(n+1)}{F}_u^{H,(n+1)}{H}_u^H{U}_u^{(n+1)},$

With

Be respectively

The diagonal matrix of feature decomposition and unitary matrice.

The iterative calculation method of algorithm two is as follows:

Step 1: make iterations variable n=0, select enough little positive number τ and thresholding (be generally minimum positive number, for example get 0.001), according to transmission channel H ₁And H ₂The singular value decomposition initializes sends pre-coding matrix

And receiving matrix

And formula (4) below their substitutions obtained the Initial Channel Assignment capacity and be designated as cap (0);

Then step 2:n=n+1 calculates in the n+1 time iteration according to (5) formula quadratic programming and upgrades vectorial z; The expression formula of z is

Wherein,

With Represent respectively link u pre-coding matrix and the receiving matrix updating value in n+1 the iterative process in ground; U=1,2 represents respectively D2D link (communication link between UE1 and the UE2) and mobile network's down link (base station is to the communication link of UE3); Represent stretching computing, T representing matrix transposition.

Step 3: utilize the renewal vector that obtains in the step 2 and introduce linear search parameter β, β changes at interval [0,1] neutral line, simultaneously note

With

And their substitution formulas (4) are calculated the cap (n+1, β) that contains ginseng, finally try to achieve the β of maximization cap (n+1, β) ^*

Step 4: upgrade

With

And the channel capacity that obtains optimum in the n+1 iterative process is designated as cap (n+1)=cap (n+1, β ^*).

Step 5: repeating step 2 is to step 4 until cap (n+1)-cap (n)≤thresholding;

Step 6: last, according to the power constraint condition to W _uLinear weighted function;

The computing formula of maximum channel capacity criterion shown in following formula (4), below formula (5) be the approximate expression of formula (4) and for quadratic programming problem;

$\underset{{\mathrm{\Δ}}_{W,u}^{(n+1)},{\mathrm{\Δ}}_{F,u}^{(n+1)}}{\max \mathrm{imize}}\underset{u=1}{\overset{2}{\mathrm{\Σ}}}\underset{k=1}{\overset{S_u}{\mathrm{\Σ}}}{\log }_2(1+{\mathrm{\γ}}_{u,k}^{(n+1)})$ (4)

$s.t.{\mathrm{\γ}}_{u,k}^{(n+1)}=\frac{f_{u,k}^{H,(n+1)}{H}_u{w}_{u,k}^{(n+1)}{w}_{u,k}^{H,(n+1)}{H}_u^H{f}_{u,k}^{(n+1)}}{f_{u,k}^{H,(n+1)}(H_u\underset{i\≠k}{\overset{S_u}{\mathrm{\Σ}}}{w}_{u,i}^{(n+1)}{w}_{u,i}^{H,(n+1)}{H}_u^H+\underset{m\≠u}{\overset{2}{\mathrm{\Σ}}}\frac{{\mathrm{\ρ}}_m}{{\mathrm{\ρ}}_u}{H}_{m,u}{W}_m^{(n+1)}{W}_m^{H,(n+1)}{H}_{m,u}^H+\frac{{\mathrm{\σ}}_u^2}{{\mathrm{\ρ}}_u}I)f_{u,k}^{(n+1)}}$

$\mathrm{tr}\left(W_m^{(n+1)}{W}_m^{H,(n+1)}\right)=\frac{P_u}{{\mathrm{\ρ}}_u};u=\mathrm{1,2};k=1,...,S_u$

Wherein, subscript u and m are variable and the u of sign link in the formula (4), m ∈ { 1,2}, u ≠ m, the conjugate transpose operation of capitalization H representing matrix, I representation unit battle array.w _{U, k}And f _{U, k}Respectively to send pre-coding matrix W _uWith receiving matrix F _uThe k column vector; ρ _uBe the transmitted power constraint factor of link u, the guaranteed output perseverance is P _u

Be the reception Signal to Interference plus Noise Ratio of k the data flow of link u, can be with the following protruding optimization problem that is converted into of the optimization problem equivalence of (4) formula with ignoring behind the higher order term with the single order polynomial approximation after its Taylor expansion:

$\underset{z}{\min \mathrm{imize}}{z}^T\left(\underset{u=1}{\overset{2}{\mathrm{\Σ}}}\underset{k=1}{\overset{S_u}{\mathrm{\Σ}}}\frac{1}{2{(1+{\mathrm{\γ}}_{u,k}^{(n+1)})}^2}{p}_{u,k}{p}_{u,k}^T\right)z-\left(\underset{u=1}{\overset{2}{\mathrm{\Σ}}}\underset{k=1}{\overset{S_u}{\mathrm{\Σ}}}\frac{1}{1+{\mathrm{\γ}}_{u,k}^{(n+1)}}{p}_{u,k}^T\right)z$

s.t. Qz＝e (5)

-τ1≤z≤τ1

For convenience of description, some has been merged into following listed intermediate variable in the formula (5):

$p_{u,k}=(\frac{1}{y_{u,k}^{\left(n\right)}}{g}_{x,u,k}-\frac{x_{u,k}^{\left(n\right)}}{y_{u,\mathrm{<figure-callout\; id="2"\; label="k"\; filenames="120928104227.png"\; state="\{\{state\}\}">k</figure-callout>}}^{2,\left(n\right)}}{g}_{y,u,k}),$

$Q=\left[\begin{array}{c}{q}_1^T\\ q_2^T\end{array}\right],---\left(6\right)$

e=[P ₁/ρ ₁ P ₂/ρ ₂] ^T

p _{U, k}In the expression formula

g _{X, u, k}And g _{Y, u, k}It is as follows,

$x_{u,k}^{\left(n\right)}=f_{u,k}^{H,\left(n\right)}{H}_u{w}_{u,k}^{\left(n\right)}{w}_{u,k}^{H,\left(n\right)}{H}_u^H{f}_{u,k}^{\left(n\right)}$

$y_{u,k}^{\left(n\right)}=f_{u,k}^{H,\left(n\right)}(H_u\underset{i\&NotEqual;k}{\overset{S_u}{\mathrm{\&Sigma;}}}{w}_{u,i}^{\left(n\right)}{w}_{u,i}^{H,\left(n\right)}{H}_u^H+\underset{m\&NotEqual;u}{\overset{2}{\mathrm{\&Sigma;}}}\frac{{\mathrm{\&rho;}}_m}{{\mathrm{\&rho;}}_u}{H}_{m,u}{W}_m^{\left(n\right)}{W}_m^{H,\left(n\right)}{H}_{m,u}^H+\frac{{\mathrm{\&sigma;}}_u^2}{{\mathrm{\&rho;}}_u}I)f_{u,k}^{\left(n\right)}---\left(7\right)$

In the formula (7)

With

Be respectively

Molecule and denominator, variable a _{U, i}, b _{U, i}, c _mAnd d _{U, k}Be defined as follows

$a_{u,i}=\mathrm{vec}\left(H_u^H{f}_{u,k}^{\left(n\right)}{f}_{u,k}^{H,\left(n\right)}{H}_u{w}_{u,i}^{\left(n\right)}\right),i=1,...,S_u;$

$b_{u,i}=\mathrm{vec}\left(H_u{w}_{u,k}^{\left(n\right)}{w}_{u,k}^{H,\left(n\right)}{H}_u^H{f}_{u,k}^{\left(n\right)}\right),i=1,...,S_u;$

$c_m=\mathrm{vec}\left(\frac{{\mathrm{\&rho;}}_m}{{\mathrm{\&rho;}}_u}{H}_{m,u}^H{f}_{u,k}^{\left(n\right)}{f}_{u,k}^{H,\left(n\right)}{H}_{m,u}{W}_m^{\left(n\right)}\right),m=1,2;---\left(9\right)$

$d_{u,k}=\mathrm{vec}\left((\underset{m\&NotEqual;u}{\overset{2}{\mathrm{\&Sigma;}}}\frac{{\mathrm{\&rho;}}_m}{{\mathrm{\&rho;}}_u}{H}_{m,u}{W}_m^{\left(n\right)}{W}_m^{H,\left(n\right)}{H}_{m,u}^H+\frac{{\mathrm{\&sigma;}}_u^2}{{\mathrm{\&rho;}}_u}I)f_{u,k}^{\left(n\right)}\right).$

Capable vector in the formula (6) among the matrix Q is

Beneficial effect: 1) the inventive method has considered that more all sidedly the base station sends the user to the interference of mobile radio network link to interference and the D2D of D2D link.Take full advantage of all transmission channels and interference channel information, transmission pre-coding matrix and the receiving matrix of combined optimization mobile radio network link and D2D link.

2) adopting the transmission pre-coding matrix that calculates in the inventive method and receiving matrix to reach to minimize the residential quarter to receive the signal mean square error receives bit error rate and and maximizes two links and target speed to reduce.Performance to real system improves a lot.

Description of drawings

Fig. 1 is the system block diagram of the optimum transmit-receive combination processing method of communication between devices under the mobile communications network that proposes of the present invention.

Fig. 2 is the error bit ability curve of the optimization mean square error algorithm (algorithm one) that proposes of the present invention and ZF, maximization SLNR algorithm.Among the figure, base station and UE1 have 4 transmitting antennas, and UE2 and UE3 have 2 reception antennas.Article two, link all transmits the single current data.

Fig. 3 is that maximized system capacity algorithm (algorithm two), ZF (ZF), the maximization letter that the present invention proposes revealed to make an uproar and contained the capacity curve that the D2D link is in the residential quarter than (SLNR) algorithm.Fig. 3 has also comprised the capacity curve the when residential quarter does not contain the D2D link.Among the figure, base station and UE1 have 4 transmitting antennas, and UE2 and UE3 have 2 reception antennas.Article two, link all transmits double-current data.

Embodiment

Below in conjunction with accompanying drawing the present invention is further described.

As shown in Figure 1: implementation step of the present invention is as follows:

1) D2D (Device-to-Device) transmission user UE1 (User Equipment) and D2D reception user UE2 have respectively N ₁Root transmitting antenna and M ₁The root reception antenna, the base station has N ₂The root transmitting antenna, network user UE3 has M ₂The root reception antenna.The transmitted power that D2D sends user and base station is respectively P ₁, P ₂, they send respectively multi-stream data x ₁And x ₂(dimension is respectively S ₁* 1 and S ₂* 1) gives separately receiver.S ₁And S ₂The data fluxion that represents respectively D2D link and mobile radio network link transmission.

2) D2D link and mobile radio network link adopt nonopiate mode to share running time-frequency resource, namely D2D equipment between physical channel and certain mobile radio network user's the physical resource of communication overlapping.In TDD (Time Division Duplexing) system, realize optimum transmitting-receiving design in order to make the base station obtain complete channel information, the base station can obtain the base station to the down channel H of UE3 by uplink channel estimation ₂And the base station is to the descending interference channel H of UE2 _2,1By above-mentioned identical channel estimating mode, UE1 can obtain it equally to the transmission channel H of UE2 ₁And the interference channel H between it and the UE3 _1,2Last UE1 is with H ₁And H _1,2Feed back to the base station.

3) base station receives signal mean square error reduction bit error rate or the optimization of maximum channel channel capacity criterion base station, the pre-coding matrix of UE1 and the receiving matrix of UE2, UE3 according to the four complete class channel informations that obtain by minimizing.Because the protruding problem of above-mentioned optimization problem right and wrong and do not have closed solutions is so this patent provides two kinds of optional iterative algorithms.U=1, u=2 represent respectively D2D link and mobile radio network link.Be convenient narration, will minimize reception signal mean square error iterative algorithm and be called algorithm one, as shown in Figure 2, and maximum channel capacity algorithm is called algorithm two, as shown in Figure 3.The iterative calculation method of algorithm one is as follows:

Step 1: make iterations variable n=0, according to transmission channel H ₁And H ₂The singular value decomposition initializes sends pre-coding matrix

U=1,2; Subscript u is the variable of sign link, and wherein u=1 and u=2 represent respectively D2D link (communication link between UE1 and the UE2) and mobile network's down link (base station is to the communication link of UE3).

Then step 2:n=n+1 calculates the receiving matrix of the n+1 time iterative process link u according to following formula (2) $F_u^{(n+1)};$

Step 3: step 2 is calculated

Formula (3) is tried to achieve parameter below the substitution power constraint condition

Wherein

That guaranteed output constraints is set up and the Lagrange's multiplier of introducing;

Step 4: with what calculate in step 2 and the step 3 With

Formula below the substitution (1) can calculate the transmission pre-coding matrix in the iterative process the n+1 time

Step 5: repeating step 2 to step 4 until

With

Convergence.

$W_u^{(n+1)}={(H_u^H{F}_u^{(n+1)}{F}_u^{H,(n+1)}{H}_u+H_{u,m}^H{F}_m^{(n+1)}{F}_m^{H,(n+1)}{H}_{u,m}+{\mathrm{\&mu;}}_u^{(n+1)}I)}^{-1}{H}_u^H{F}_u^{(n+1)}---\left(1\right)$

$F_u^{(n+1)}={(H_u{W}_u^{\left(n\right)}{W}_u^{H,\left(n\right)}{H}_u^H+H_{m,u}^H{W}_m^{\left(n\right)}{W}_m^{H,\left(n\right)}{H}_{m,u}+{\mathrm{\&sigma;}}_u^{2}I)}^{-1}{H}_u{W}_u^{\left(n\right)}---\left(2\right)$

$\underset{i=1}{\overset{N_u}{\mathrm{\&Sigma;}}}\frac{G_u^{\left[\mathrm{tt}\right],(n+1)}}{[{\mathrm{\&mu;}}_u^{(n+1)}+{\mathrm{\&Delta;}}_u^{\left[\mathrm{tt}\right],(n+1)}]}=\mathrm{Pu}---\left(3\right)$

Formula (1) all is variable and the u of sign link to the middle subscript u of formula (3) and m, m ∈ 1,2}, u ≠ m, and n represents the iterations variable, the conjugate transpose operation of capitalization H representing matrix, I representation unit battle array.

With

Represented respectively matrix

With

T element on the leading diagonal.Wherein, $G_u^{(n+1)}=U_u^{H,(n+1)}{H}_u^H{F}_u^{(n+1)}{F}_u^{H,(n+1)}{H}_u^H{U}_u^{(n+1)},$

With

Be respectively

The diagonal matrix of feature decomposition and unitary matrice.

The iterative calculation method of algorithm two is as follows:

Step 1: make iterations variable n=0, select enough little positive number τ and thresholding (be generally minimum positive number, for example get 0.001), according to transmission channel H ₁And H ₂The singular value decomposition initializes sends pre-coding matrix

And receiving matrix

And formula (4) below their substitutions obtained the Initial Channel Assignment capacity and be designated as cap (0);

Then step 2:n=n+1 calculates in the n+1 time iteration according to following (5) formula quadratic programming and upgrades vectorial z; The expression formula of z is

Wherein, With

Represent respectively link u pre-coding matrix and the receiving matrix updating value in n+1 the iterative process in ground; U=1,2 represents respectively D2D link (communication link between UE1 and the UE2) and mobile network's down link (base station is to the communication link of UE3);

Represent stretching computing, T representing matrix transposition.

Step 3: utilize the renewal vector that obtains in the step 2 and introduce linear search parameter β, β changes at interval [0,1] neutral line, simultaneously note

With

And their substitution formulas (4) are calculated the cap (n+1, β) that contains ginseng, finally try to achieve the β of maximization cap (n+1, β) ^*

Step 4: upgrade

With

And the channel capacity that obtains optimum in the n+1 iterative process is designated as cap (n+1)=cap (n+1, β ^*).

Step 5: repeating step 2 is to step 4 until cap (n+1)-cap (n)≤thresholding;

Step 6: last, according to the power constraint condition to W _uLinear weighted function;

The computing formula of maximum channel capacity criterion is suc as formula shown in following (4), below formula (5) be the approximate expression of formula (4) and be quadratic programming problem;

$\underset{{\mathrm{\&Delta;}}_{W,u}^{(n+1)},{\mathrm{\&Delta;}}_{F,u}^{(n+1)}}{\max \mathrm{imize}}\underset{u=1}{\overset{2}{\mathrm{\&Sigma;}}}\underset{k=1}{\overset{S_u}{\mathrm{\&Sigma;}}}{\log }_2(1+{\mathrm{\&gamma;}}_{u,k}^{(n+1)})$ (4)

$s.t.{\mathrm{\&gamma;}}_{u,k}^{(n+1)}=\frac{f_{u,k}^{H,(n+1)}{H}_u{w}_{u,k}^{(n+1)}{w}_{u,k}^{H,(n+1)}{H}_u^H{f}_{u,k}^{(n+1)}}{f_{u,k}^{H,(n+1)}(H_u\underset{i\&NotEqual;k}{\overset{S_u}{\mathrm{\&Sigma;}}}{w}_{u,i}^{(n+1)}{w}_{u,i}^{H,(n+1)}{H}_u^H+\underset{m\&NotEqual;\mathrm{<figure-callout\; id="2"\; label="u"\; filenames="120928104227.png"\; state="\{\{state\}\}">u</figure-callout>}}{\overset{2}{\mathrm{\&Sigma;}}}\frac{{\mathrm{\&rho;}}_m}{{\mathrm{\&rho;}}_u}{H}_{m,u}{W}_m^{(n+1)}{W}_m^{H,(n+1)}{H}_{m,u}^H+\frac{{\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="u"\; filenames="120928104227.png"\; state="\{\{state\}\}">u</figure-callout>}}^2}{{\mathrm{\&rho;}}_u}I)f_{u,k}^{(n+1)}}$

$\mathrm{tr}\left(W_m^{(n+1)}{W}_m^{H,(n+1)}\right)=\frac{P_u}{{\mathrm{\&rho;}}_u};u=\mathrm{1,2};k=1,...,S_u$

Wherein, subscript u and m are variable and the u of sign link in the formula (4), m ∈ { 1,2}, u ≠ m, the conjugate transpose operation of capitalization H representing matrix, I representation unit battle array.w _{U, k}And f _{U, k}Respectively to send pre-coding matrix W _uWith receiving matrix F _uThe k column vector; ρ _uBe the transmitted power constraint factor of link u, the guaranteed output perseverance is P _u Be the reception Signal to Interference plus Noise Ratio of k the data flow of link u, can be with the following protruding optimization problem that is converted into of the optimization problem equivalence of (4) formula with ignoring behind the higher order term with the single order polynomial approximation after its Taylor expansion:

$\underset{z}{\min \mathrm{imize}}{z}^T\left(\underset{u=1}{\overset{2}{\mathrm{\&Sigma;}}}\underset{k=1}{\overset{S_u}{\mathrm{\&Sigma;}}}\frac{1}{2{(1+{\mathrm{\&gamma;}}_{u,k}^{(n+1)})}^2}{p}_{u,k}{p}_{u,k}^T\right)z-\left(\underset{u=1}{\overset{2}{\mathrm{\&Sigma;}}}\underset{k=1}{\overset{S_u}{\mathrm{\&Sigma;}}}\frac{1}{1+{\mathrm{\&gamma;}}_{u,k}^{(n+1)}}{p}_{u,k}^T\right)z$

s.t. Qz＝e (5)

-τ1≤z≤τ1

For convenience of description, some has been merged into following listed intermediate variable in the formula (5):

$p_{u,k}=(\frac{1}{y_{u,k}^{\left(n\right)}}{g}_{x,u,k}-\frac{x_{u,k}^{\left(n\right)}}{y_{u,\mathrm{<figure-callout\; id="2"\; label="k"\; filenames="120928104227.png"\; state="\{\{state\}\}">k</figure-callout>}}^{2,\left(n\right)}}{g}_{y,u,k}),$

$Q=\left[\begin{array}{c}{q}_1^T\\ q_2^T\end{array}\right],---\left(6\right)$

e=[P ₁/ρ ₁ P ₂/ρ ₂] ^T

p _{U, k}In the expression formula

g _{X, u, k}And g _{Y, u, k}It is as follows,

$x_{u,k}^{\left(n\right)}=f_{u,k}^{H,\left(n\right)}{H}_u{w}_{u,k}^{\left(n\right)}{w}_{u,k}^{H,\left(n\right)}{H}_u^H{f}_{u,k}^{\left(n\right)}$

$y_{u,k}^{\left(n\right)}=f_{u,k}^{H,\left(n\right)}(H_u\underset{i\&NotEqual;k}{\overset{S_u}{\mathrm{\&Sigma;}}}{w}_{u,i}^{\left(n\right)}{w}_{u,i}^{H,\left(n\right)}{H}_u^H+\underset{m\&NotEqual;\mathrm{<figure-callout\; id="2"\; label="u"\; filenames="120928104227.png"\; state="\{\{state\}\}">u</figure-callout>}}{\overset{2}{\mathrm{\&Sigma;}}}\frac{{\mathrm{\&rho;}}_m}{{\mathrm{\&rho;}}_u}{H}_{m,u}{W}_m^{\left(n\right)}{W}_m^{H,\left(n\right)}{H}_{m,u}^H+\frac{{\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="u"\; filenames="120928104227.png"\; state="\{\{state\}\}">u</figure-callout>}}^2}{{\mathrm{\&rho;}}_u}I)f_{u,k}^{\left(n\right)}---\left(7\right)$

In the formula (7)

With

Be respectively Molecule and denominator, variable a _{U, i}, b _{U, i}, c _mAnd d _{U, k}Be defined as follows

$a_{u,i}=\mathrm{vec}\left(H_u^H{f}_{u,k}^{\left(n\right)}{f}_{u,k}^{H,\left(n\right)}{H}_u{w}_{u,i}^{\left(n\right)}\right),i=1,...,S_u;$

$b_{u,i}=\mathrm{vec}\left(H_u{w}_{u,k}^{\left(n\right)}{w}_{u,k}^{H,\left(n\right)}{H}_u^H{f}_{u,k}^{\left(n\right)}\right),i=1,...,S_u;$

$c_m=\mathrm{vec}\left(\frac{{\mathrm{\&rho;}}_m}{{\mathrm{\&rho;}}_u}{H}_{m,u}^H{f}_{u,k}^{\left(n\right)}{f}_{u,k}^{H,\left(n\right)}{H}_{m,u}{W}_m^{\left(n\right)}\right),m=1,2;---\left(9\right)$

$d_{u,k}=\mathrm{vec}\left((\underset{m\&NotEqual;\mathrm{<figure-callout\; id="2"\; label="u"\; filenames="120928104227.png"\; state="\{\{state\}\}">u</figure-callout>}}{\overset{2}{\mathrm{\&Sigma;}}}\frac{{\mathrm{\&rho;}}_m}{{\mathrm{\&rho;}}_u}{H}_{m,u}{W}_m^{\left(n\right)}{W}_m^{H,\left(n\right)}{H}_{m,u}^H+\frac{{\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="u"\; filenames="120928104227.png"\; state="\{\{state\}\}">u</figure-callout>}}^2}{{\mathrm{\&rho;}}_u}I)f_{u,k}^{\left(n\right)}\right).$

Capable vector in the formula (6) among the matrix Q is

The above only is preferred implementation of the present invention; be noted that for those skilled in the art; under the prerequisite that does not break away from the principle of the invention, can also make some improvements and modifications, these improvements and modifications also should be considered as protection scope of the present invention.

Claims (3)

1. the optimum transmit-receive combination processing method of communication between devices under the mobile communications network, it is characterized in that: the method is carried out according to the following steps:

1) D2D transmission user UE1 and D2D reception user UE2 have respectively N ₁Root transmitting antenna and M ₁The root reception antenna, the base station has N ₂The root transmitting antenna, network user UE3 has M ₂The root reception antenna, the transmitted power that D2D sends user and base station is respectively P ₁, P ₂, they send respectively multi-stream data x ₁And x ₂Give receiver separately, their dimension is respectively S ₁* 1 and S ₂* 1, S ₁And S ₂The data fluxion that represents respectively D2D link and mobile radio network link transmission;

2) D2D link and mobile radio network link adopt nonopiate mode to share running time-frequency resource, namely D2D equipment between physical channel and certain mobile radio network user's the physical resource of communication overlapping; In the TDD system, realize optimum transmitting-receiving design in order to make the base station obtain complete channel information, the base station can obtain the base station to the down channel H of UE3 by uplink channel estimation ₂And the base station is to the descending interference channel H of UE2 _2,1By above-mentioned identical channel estimating mode, UE1 can obtain it equally to the transmission channel H of UE2 ₁And the interference channel H between it and the UE3 _1,2, last UE1 is with H ₁And H _1,2Feed back to the base station;

3) base station receives signal mean square error reduction bit error rate or the optimization of maximum channel channel capacity criterion base station, the pre-coding matrix of UE1 and the receiving matrix of UE2, UE3 according to the four complete class channel informations that obtain by minimizing;

4) base station is respectively W according to minimizing the mean square error that receives signal or transmission pre-coding matrix and the receiving matrix that maximum channel capacity criterion calculates link 1, link 2 ₁, F ₁And W ₂, F ₂, the base station is with W ₁, F ₁Pass to D2D communication equipment pair by control channel;

5) last, UE1 utilizes and sends pre-coding matrix W ₁Weighted data stream x ₁Send to UE2, and send pre-coding matrix W ₂Then be used for weighting downstream signal x by the base station ₂And send to UE3.

2. the optimum transmit-receive combination processing method of communication between devices under the mobile communications network according to claim 1, it is characterized in that: minimize reception signal mean square error iterative algorithm in the described step 3) and be called algorithm one, the iterative calculation method of this algorithm is as follows:

Step 1: make iterations variable n=0, according to transmission channel H ₁And H ₂The singular value decomposition initializes sends pre-coding matrix

U=1,2; Subscript u is the variable of sign link, and wherein u=1 and u=2 represent respectively D2D link and mobile network's down link;

Then step 2:n=n+1 calculates the receiving matrix of the n+1 time iterative process link u according to following formula (2)

Step 3: step 2 is calculated Formula (3) is tried to achieve parameter below the substitution power constraint condition

Wherein That guaranteed output constraints is set up and the Lagrange's multiplier of introducing;

Step 4: with what calculate in step 2 and the step 3

With

Formula below the substitution (1) can calculate the transmission pre-coding matrix in the iterative process the n+1 time

Step 5: repeating step 2 to step 4 until With

Convergence;

$W_u^{(n+1)}={(H_u^H{F}_u^{(n+1)}{F}_u^{H,(n+1)}{H}_u+H_{u,m}^H{F}_m^{(n+1)}{F}_m^{H,(n+1)}{H}_{u,m}+{\mathrm{\&mu;}}_u^{(n+1)}I)}^{-1}{H}_u^H{F}_u^{(n+1)}---\left(1\right)$

$F_u^{(n+1)}={(H_u{W}_u^{\left(n\right)}{W}_u^{H,\left(n\right)}{H}_u^H+H_{m,u}^H{W}_m^{\left(n\right)}{W}_m^{H,\left(n\right)}{H}_{m,u}+{\mathrm{\&sigma;}}_u^{2}I)}^{-1}{H}_u{W}_u^{\left(n\right)}---\left(2\right)$

$\underset{i=1}{\overset{N_u}{\mathrm{\&Sigma;}}}\frac{G_u^{\left[\mathrm{tt}\right],(n+1)}}{[{\mathrm{\&mu;}}_u^{(n+1)}+{\mathrm{\&Delta;}}_u^{\left[\mathrm{tt}\right],(n+1)}]}=\mathrm{Pu}---\left(3\right)$

Formula (1) all is variable and the u of sign link to the middle subscript u of formula (3) and m, m ∈ 1,2}, u ≠ m, and n represents the iterations variable, the conjugate transpose operation of capitalization H representing matrix, I representation unit battle array;

With

Represented respectively matrix

With

T element on the leading diagonal; Wherein, $G_u^{(n+1)}=U_u^{H,(n+1)}{H}_u^H{F}_u^{(n+1)}{F}_u^{H,(n+1)}{H}_u^H{U}_u^{(n+1)},$

With

Be respectively The diagonal matrix of feature decomposition and unitary matrice.

3. the optimum transmit-receive combination processing method of communication between devices under the mobile communications network according to claim 1, it is characterized in that: maximum channel capacity algorithm is called algorithm two in the described step 3), and the iterative calculation method of this algorithm is as follows:

Step 1: make iterations variable n=0, select enough little positive number τ and thresholding, according to transmission channel H ₁And H ₂The singular value decomposition initializes sends pre-coding matrix

And receiving matrix

And formula (4) below their substitutions obtained the Initial Channel Assignment capacity and be designated as cap (0);

Then step 2:n=n+1 calculates in the n+1 time iteration according to following formula (5) quadratic programming and upgrades vectorial z; The expression formula of z is

Wherein,

With

Represent respectively link u pre-coding matrix and the receiving matrix updating value in n+1 the iterative process in ground; U=1,2 represents respectively D2D link and mobile network's down link;

Represent stretching computing, T representing matrix transposition;

Step 3: utilize the renewal vector that obtains in the step 2 and introduce linear search parameter β, β changes at interval [0,1] neutral line, simultaneously note

With

And formula (4) below their substitutions calculated the cap (n+1, β) that contains ginseng, finally try to achieve the β of maximization cap (n+1, β) ^*

Step 4: upgrade

With

And the channel capacity that obtains optimum in the n+1 iterative process is designated as cap (n+1)=cap (n+1, β ^*);

Step 5: repeating step 2 is to step 4 until cap (n+1)-cap (n)≤thresholding;

Step 6: last, according to the power constraint condition to W _uLinear weighted function;

The computing formula of maximum channel capacity criterion shown in following formula (4), below formula (5) be the approximate expression of following formula (4) and for quadratic programming problem;

$\underset{{\mathrm{\&Delta;}}_{W,u}^{(n+1)},{\mathrm{\&Delta;}}_{F,u}^{(n+1)}}{\max \mathrm{imize}}\underset{u=1}{\overset{2}{\mathrm{\&Sigma;}}}\underset{k=1}{\overset{S_u}{\mathrm{\&Sigma;}}}{\log }_2(1+{\mathrm{\&gamma;}}_{u,k}^{(n+1)})$ (4)

$s.t.{\mathrm{\&gamma;}}_{u,k}^{(n+1)}=\frac{f_{u,k}^{H,(n+1)}{H}_u{w}_{u,k}^{(n+1)}{w}_{u,k}^{H,(n+1)}{H}_u^H{f}_{u,k}^{(n+1)}}{f_{u,k}^{H,(n+1)}(H_u\underset{i\&NotEqual;k}{\overset{S_u}{\mathrm{\&Sigma;}}}{w}_{u,i}^{(n+1)}{w}_{u,i}^{H,(n+1)}{H}_u^H+\underset{m\&NotEqual;u}{\overset{2}{\mathrm{\&Sigma;}}}\frac{{\mathrm{\&rho;}}_m}{{\mathrm{\&rho;}}_u}{H}_{m,u}{W}_m^{(n+1)}{W}_m^{H,(n+1)}{H}_{m,u}^H+\frac{{\mathrm{\&sigma;}}_u^2}{{\mathrm{\&rho;}}_u}I)f_{u,k}^{(n+1)}}$

$\mathrm{tr}\left(W_m^{(n+1)}{W}_m^{H,(n+1)}\right)=\frac{P_u}{{\mathrm{\&rho;}}_u};u=\mathrm{1,2};k=1,...,S_u$

Wherein, subscript u and m are variable and the u of sign link in the formula (4), m ∈ { 1,2}, u ≠ m, the conjugate transpose operation of capitalization H representing matrix, I representation unit battle array; w _{U, k}And f _{U, k}Respectively to send pre-coding matrix W _uWith receiving matrix F _uThe k column vector; ρ _uBe the transmitted power constraint factor of link u, the guaranteed output perseverance is P _u Be the reception Signal to Interference plus Noise Ratio of k the data flow of link u, can be with the following protruding optimization problem that is converted into of the optimization problem equivalence of (4) formula with ignoring behind the higher order term with the single order polynomial approximation after its Taylor expansion:

$\underset{z}{\min \mathrm{imize}}{z}^T\left(\underset{u=1}{\overset{2}{\mathrm{\&Sigma;}}}\underset{k=1}{\overset{S_u}{\mathrm{\&Sigma;}}}\frac{1}{2{(1+{\mathrm{\&gamma;}}_{u,k}^{(n+1)})}^2}{p}_{u,k}{p}_{u,k}^T\right)z-\left(\underset{u=1}{\overset{2}{\mathrm{\&Sigma;}}}\underset{k=1}{\overset{S_u}{\mathrm{\&Sigma;}}}\frac{1}{1+{\mathrm{\&gamma;}}_{u,k}^{(n+1)}}{p}_{u,k}^T\right)z$

s.t. Qz＝e (5)

-τ1≤z≤τ1

Some has been merged into following listed intermediate variable in the formula (5):

$p_{u,k}=(\frac{1}{y_{u,k}^{\left(n\right)}}{g}_{x,u,k}-\frac{x_{u,k}^{\left(n\right)}}{y_{u,k}^{2,\left(n\right)}}{g}_{y,u,k}),$

$Q=\left[\begin{array}{c}{q}_1^T\\ q_2^T\end{array}\right],---\left(6\right)$

e=[P ₁/ρ ₁ P ₂/ρ ₂] ^T

p _{U, k}In the expression formula g _{X, u, k}And g _{Y, u, k}It is as follows,

$x_{u,k}^{\left(n\right)}=f_{u,k}^{H,\left(n\right)}{H}_u{w}_{u,k}^{\left(n\right)}{w}_{u,k}^{H,\left(n\right)}{H}_u^H{f}_{u,k}^{\left(n\right)}$

$y_{u,k}^{\left(n\right)}=f_{u,k}^{H,\left(n\right)}(H_u\underset{i\&NotEqual;k}{\overset{S_u}{\mathrm{\&Sigma;}}}{w}_{u,i}^{\left(n\right)}{w}_{u,i}^{H,\left(n\right)}{H}_u^H+\underset{m\&NotEqual;u}{\overset{2}{\mathrm{\&Sigma;}}}\frac{{\mathrm{\&rho;}}_m}{{\mathrm{\&rho;}}_u}{H}_{m,u}{W}_m^{\left(n\right)}{W}_m^{H,\left(n\right)}{H}_{m,u}^H+\frac{{\mathrm{\&sigma;}}_u^2}{{\mathrm{\&rho;}}_u}I)f_{u,k}^{\left(n\right)}---\left(7\right)$

In the formula (7)

With Be respectively

Molecule and denominator, variable a _{U, i}, b _{U, i}, c _mAnd d _{U, k}Be defined as follows

$a_{u,i}=\mathrm{vec}\left(H_u^H{f}_{u,k}^{\left(n\right)}{f}_{u,k}^{H,\left(n\right)}{H}_u{w}_{u,i}^{\left(n\right)}\right),i=1,...,S_u;$

$b_{u,i}=\mathrm{vec}\left(H_u{w}_{u,k}^{\left(n\right)}{w}_{u,k}^{H,\left(n\right)}{H}_u^H{f}_{u,k}^{\left(n\right)}\right),i=1,...,S_u;$

$c_m=\mathrm{vec}\left(\frac{{\mathrm{\&rho;}}_m}{{\mathrm{\&rho;}}_u}{H}_{m,u}^H{f}_{u,k}^{\left(n\right)}{f}_{u,k}^{H,\left(n\right)}{H}_{m,u}{W}_m^{\left(n\right)}\right),m=1,2;---\left(9\right)$

$d_{u,k}=\mathrm{vec}\left((\underset{m\&NotEqual;u}{\overset{2}{\mathrm{\&Sigma;}}}\frac{{\mathrm{\&rho;}}_m}{{\mathrm{\&rho;}}_u}{H}_{m,u}{W}_m^{\left(n\right)}{W}_m^{H,\left(n\right)}{H}_{m,u}^H+\frac{{\mathrm{\&sigma;}}_u^2}{{\mathrm{\&rho;}}_u}I)f_{u,k}^{\left(n\right)}\right).$

Capable vector in the formula (6) among the matrix Q is

Images