CN106788643A - A kind of safe method for precoding of multi-user to end-to-end relay system - Google Patents

Abstract

Multi-user is used for the Precoding Design method of safety of physical layer to end-to-end relay system, via node is using the otherness between channel, multi-user interference is eliminated at corresponding destination node by precoded signal treatment technology, ZF treatment is carried out in eavesdropping node direction simultaneously again, safe information transmission is realized.On the one hand, this method is eliminated and multi-user interference and prevents information leakage, the detection complexity for reducing destination node receiver and the safe transmission that ensure that information.On the other hand, this method concentrates on fixation means relaying major complexity, simplifies the hardware requirement to mobile subscriber, and this method relates only to the via node docking collection of letters number and does simple addition and multiplication operation, and hardware realization is simple, and practicality is stronger.

Description

A kind of safe method for precoding of multi-user to end-to-end relay system

Technical field

Multi-user couple is directed to the present invention relates to a kind of Precoding Design method for improving safety of physical layer, more particularly to one kind The recoding processing method of via node in end-to-end relay system, the method can be prevented while multiuser interference is eliminated Only information leakage.

Background technology

The extensive utilization of radio communication, although bring huge facility to people's lives, but the broadcast of radio communication is special The opening of property and wireless media, causes information security threats increasingly to highlight seriously, particularly the field such as ecommerce, and this is caused Information security issue is increasingly subject to extensive concern.The method for traditionally solving this safety problem is the high level more than physical layer In take encryption key system, distributed by carrying out to validated user key and realize authentication and secure communication.However, with Number of users abruptly increase, the complexity and key length of key distribution turn into encryption mechanism two big challenges urgently to be resolved hurrily.Based on letter Breath opinion, the safety of physical layer technology approach new for communication security is opened.Therefore, with reference to existing encryption mechanism, safety of physical layer Technology can provide more solid safety guarantee for communication security.

The main thought of safety of physical layer is to utilize channel diversity, and primary link is increased by various signal processing technologies The channel capacity of (channel between source node and destination node), and deteriorate eavesdropping link (source node and eavesdropping node as far as possible simultaneously Between channel) channel capacity.A.D.Wyner " The wire-tap channel.Bell Syst.Tech.J., Show in the texts of Jan.1975, pp.1355-1387. "：As long as the channel capacity of primary link is more than the channel capacity of eavesdropping link, The safe transmission of information can just be realized.Therefore, the research of safety of physical layer provides new approach for safe information transmission.It is many Antenna technology can be overcome in a single aerial system when tapping channel quality is better than main channel by introducing multidimensional processiug, Safe capacity is zero limitation, again can exponentially lifting system transmission rate.Therefore, multi-antenna technology is improving physical layer Safely, ensure information security and played an important role on transmitting.Multidimensional processiug mode mainly includes line in multi-antenna technology Property precoding technique and artificial plus two kinds of technology of making an uproar.Linear Precoding is mainly carried out at linear change by transmission signal Reason, is received with reaching enhancing destination node, reduces the information content for being leaked to eavesdrop node, so as to expand primary link with eavesdropping link Between channel capacity difference.In addition, the artificial plus technology of making an uproar is also that multi-antenna technology improves the important hand of safety of physical layer One of section.Under conditions of destination node normal received signal is not influenceed, manually made an uproar by being added in the kernel of primary link Sound deteriorates eavesdropping node receptivity to reach, so as to realize improving the purpose of safety of physical layer performance.

Relay system not only has and overcomes the advantages such as telecommunication, and via node is by using precoding technique The income of safety of physical layer improvement can be obtained.Because relay system concentrates on via node in fact signal precoding treatment It is existing, greatly simplifie the hardware configuration requirement to source node and destination node transceiver.Accordingly, with respect to relay system precoding The research of technology is increasingly subject to widely studied.Particularly, multi-user is to relay communications system (Multiuser Peer-to-Peer Relay Network) in Precoding Design problem：How to design efficient pre-coding scheme can either be while support multigroup use To multiplex data stream parallel transmission, can efficiently solve again interfere problem between multi-user and safe information transmission is asked at family Topic.For this safe Precoding Design problem, X.Gong et al. is in " Secure communication design for multi-user peer-to-peer wireless relay networks.IET Commun.Jan.2016,pp.770- 777. " solved from following thinking in a text：On the premise of ensureing validated user to secure communication, multi-user couple is maximized In minimum safe transmission rate, so as to by safe precoding problem be converted into " minimum safe transmission rate " be optimization aim Semi definite programming problem, and effectively solved using interior point method.However, this solution mode is merely capable of obtaining one closely As suboptimal solution, and calculating process is extremely cumbersome.And, this settling mode is, it is necessary to what joint-detection institute active node sent Source information can determine the private information transmitted by source node, and the signal detection complexity of destination node is big, and cannot have Effect ground suppresses influence of the multi-user interference to destination node.

During multi-user is to end-to-end relay system, can all enter in transmitting procedure per the signal of antenna all the way in contract network Row superposition.If carrying out linear predictive coding treatment to combined channel, big dimension data is introduced, increase spatial gain, i.e. in relaying Linear Precoding is taken at node, due to the increase of spatial gain, this mode can eliminate the same of multi-user interference When prevent information leakage.Meanwhile, linear predictive coding treatment can also simplify destination node receiver structure, reduce its detection again The miscellaneous availability of frequency spectrum that system is obviously improved while spend.But correlative study is but seldom seen in report.

The content of the invention

Technical problem：The technical problems to be solved by the invention are：For multi-user proposes one kind to end-to-end relay system For the Precoding Design scheme of safety of physical layer.On the one hand this method can prevent information leakage, realize safe information transmission, On the other hand multiuser interference can be eliminated, is simplified destination node receiver structure and is reduced its detection complexity.

Technical scheme：For by a via node R, K source node S₁,S₂,...,S_k,...,S_K, K destination node D₁,D₂,...,D_k,...,D_KWith J eavesdropping node E₁,E₂,...,E_j,...,E_JThe multi-user of composition is to end-to-end relay system (source node is consistent with the number of destination node in system, and a source node and a corresponding destination node constitute one group of user It is right) in system, K source node and K destination node realize telecommunication under the assistance of relaying R, each source node S_k K-th corresponding destination node D of difference_kOne group of user is constituted to { S_k,D_k, K group users are constituted altogether to { S₁,D₁},{S₂, D₂},...,{S_k,D_k},...,{S_K,D_K}.K groups user's centering, each destination node D_kOnly want to obtain source node S_k The private information of transmission, and the transmission information of other source nodes is considered as interference.The K groups user couple and J eavesdropping node are equal It is the legal mobile subscriber (such as mobile phone) in cellular network, due to the limitation of mobile device size, it is assumed that it is equipped with single antenna； And via node R is fixation means, comprising N root antennas R₁,R₂,...,R_n,...,R_N, the radical antenna N is for positive integer and expires Sufficient N >=K+J, wherein, the user is positive integer to number K and eavesdropping node number J, and k is any whole in the range of [1, K] Number, represents each group of user to { S_k,D_kNumbering；J is any integer in the range of [1, J], represent it is described each steal Listen node E_jNumbering；N is any integer in the range of [1, N], represents each antenna R_nNumbering.It is above-mentioned to solve Technical problem, the pre-coding scheme that the present invention is used is as follows：

The first step：Channel estimation：The via node R receives whole users to { S first₁,D₁},{S₂,D₂},...,{S_k, D_k},...,{S_K,D_KAnd all eavesdrop node E₁,E₂,...,E_k,...,E_JThe positional information of feedback；Then according to the position Information calculates described each source node S of user's centering respectively_kThe each antenna R in the via node_nChannel gainEach antenna R in the via node_nTo each destination node of user's centering D_kChannel gainAnd Each antenna R in the via node_nTo each eavesdropping node E_jChannel gainAgain will described each source Node S_kTo each antenna R_nChannel gainIt is combined as source node channel gain vectors By each antenna R_nTo each destination node D_kChannel gainIt is combined as destination node channel gain vectorsBy each antenna R_nTo each eavesdropping node E_jChannel gainCombination It is eavesdropping node channel gain vectorWherein, described source node channel gain vectorsDestination node channel gain vectorsWith eavesdropping node Channel gain vectorsIt is the vector of N × 1 dimension.

Second step：Receive broadcast message：The user is to { S₁,D₁},{S₂,D₂},...,{S_k,D_k},...,{S_K,D_KIn Each source node S_kRespectively by via node to its corresponding destination node D_kSend private information x_k, then section is relayed Every antenna R in point_nThe signal for receiving isWherein, P_kRepresent described each source node S_kThe power of private information is sent,Represent n-th antenna R in the via node_nThe additive white Gaussian noise at place；Then will Each antenna R in the via node_nThe signal for receiving is combined as antenna and receives signal phasorWhereinIt is antenna additive Gaussian white noise Acoustic vector.Wherein, antenna receives signal phasor y_RWith antenna additive white Gaussian noise vector v_RN × 1 n dimensional vector n is,

3rd step：Precoding processing：The via node docking collection of letters numberEach element carry out Linear superposition computing, and it is combined into new precoding vectors signal By_RIt is sent on transmitting antenna.Wherein, B prelists for N × N-dimensional Code matrix, pre-coding matrix B pth row q column elements are designated as B (l, q), and wherein l and q is the integer in the range of [1, N], and note is pre- Coding vector b is Ns of the pre-coding matrix B as obtained by row storehouse²× 1 dimensional vector, the generation step of the precoding vector b is such as Under：

Step 301：Calculate each destination node D_k(the K-1) × N at place²Tie up equivalent interference matrix F_k=[f_1,k,f_2,k,..., f_k-1,k,f_k+1,k,...,f_K,k]^T, wherein, except k-th source node S_kEach source node S in addition_iTo k-th destination node D_k Equivalent disturbance regime be expressed as source node S_iTo k-th destination node D_kEquivalent interference vector Table Show that right Kronecker (Kronecker) accumulates computing, i ∈ [1, K], i ≠ k.Now, k-th destination node D_kThe equivalent interference at place Link isFor k-th destination node D_k, it is only needed to from k-th source node S_kThe information of transmitting, and other sources are saved Point S_iThe information of (wherein i ≠ k) transmitting, to k-th destination node D_kFor be accordingly to be regarded as interference signal.Therefore, k-th section of mesh Point D_kThe equivalent interfering link at place is except source node S_kOutside institute's active node to destination node D_kEquivalent periodic line set, note It is F_kB, wherein F_k=[f_1,k,f_2,k,...,f_k-1,k,f_k+1,k,...,f_K,k]^TIt is (K-1) × N²Dimension matrix.

Step 302：Calculate each eavesdropping node E_jThe equivalent eavesdropping matrix G at place_j=[g_1,j,g_2,j,...,g_K,j]^T, wherein, Each source node S_kTo j-th eavesdropping node E_jEquivalent eavesdropping situation be expressed as source node S_kTo j-th eavesdropping node E_j's Equivalent eavesdropping vectorNow, j-th eavesdropping node E_jPlace equivalent eavesdropping link beTherefore, J eavesdropping node E_jThe equivalent eavesdropping link at place is institute's active node to eavesdropping node E_jEquivalent periodic line set, be designated as G_jB, its Middle G_j=[g_1,j,g_2,j,...,g_K,j]^TIt is K × N²Dimension matrix.

Step 303, calculates the kernel T of equivalent interference matrix and equivalent eavesdropping matrix：By all equivalent interference matrix F_k =[f_1,k,f_2,k,...,f_k-1,k,f_k+1,k,...,f_K,k]^TWith equivalent eavesdropping matrix G_j=[g_1,j,g_2,j,...,g_K,j]^TMerge into Dimension is (K+J-1) K × N²Interference eavesdropping matrixAnd to interference eavesdropping matrix Φ carries out singular value decompositionWherein, dimension is N²×N²Matrix V_ΦIt is the right feature of interference eavesdropping matrix Φ Vector matrix；Dimension is (K+J-1) K × N²Matrix Σ_ΦIt is the eigenvalue matrix of interference eavesdropping matrix Φ, dimension is (K+J- 1) matrix U of K × (K+J-1) K_ΦIt is the left eigenvector matrix of interference eavesdropping matrix Φ；The kernel T of matrix is eavesdropped in then interference It is by right eigenvectors matrix V_ΦN²The column elements of-K (K+J-1)+1 are to N²The N that column element is constituted²×(N²-K²- KJ+K) dimension square Battle array, is represented by：T=V_Φ(:,[N²-K(K+J-1)+1:N²])；

Step 304, precoded signal treatment eliminates multi-user interference and information leakage：Precoding is calculated according to the following equation VectorWherein P_RIt is the maximum of repeating device power；For N × N-dimensional matrix, represents the covariance of via node received signal power, wherein P_kIt is k-th source node S_kSend private information Power,It is source node channel gain vectorsConjugate matrices,It is source node channel gain vectorsTurn Put matrix, I_NIt is N × N-dimensional unit matrix, σ²It is the variance of additive white Gaussian noise；(N²-K²- KJ+K) × 1 equivalent the wave beam forming of dimension Vectorial p is all purposes node receiving power covariance matrixThe corresponding unit character vector of eigenvalue of maximum, Wherein, described each source node S of user's centering_kTo corresponding destination node D_kThe equivalent situation of direct transferring be expressed as source node S_kTo corresponding destination node D_kThe equivalent vector that direct transfers·^HRepresent associate matrix fortune Calculate,^TIt is transposed matrix computing.Pre-coding matrix B is entered according to B (l, q)=b (qN-N+l) according to precoding vector b finally Row assignment；

4th step：Forward signal：Each antenna R in the via node R_nAccording toTurn Hair precoding vectors signal.

By k-th destination node D_kSignal is received to be designated asIt is represented by: WhereinRepresent destination node D_kThe additive white Gaussian noise at place, it is that zero, variance is σ to obey average²Multiple Gauss distribution.Receive SignalIn multi-user interference be eliminated, only include user D_kDesired private information x_kWith noise two parts.Therefore, user D_kCan be according to reception signalCarry out independent detection private information x_k, without to institute's active node sending signal joint-detection, Oneself desired information is extracted again.

On the other hand, by j-th eavesdropping node E_jThe signal of reception is designated asIt is represented by： WhereinRepresent eavesdropping node E_jThe additive white Gaussian noise at place, it is that zero, variance is σ to obey average²Multiple Gauss distribution.Due to Eavesdropping node only includes noise in receiving signal, therefore, all eavesdropping nodes do not get any private information, multi-user couple Secure communication between end-to-end is effectively ensured.

Further, in the above method, using the channel estimation mode meter based on least mean-square error in the first step Calculate the source node channel gain vectorsDestination node channel gain vectorsWith eavesdropping node channel gain vectorDue to K groups User couple and J eavesdropping node are legal mobile subscribers in cellular network, and these user nodes are anti-to via node in real time Present the positional information of itself.Therefore, via node estimates their letters and via node between from their feedback information Channel state information, for example, using the channel estimation mode based on least mean-square error, its detailed process is referring to " Georgios B.Giannakis.Signal Processing Advances in Wireless and Mobile Communications Volume 1:Trend in Channel Estimation and Equalization.Beijing:Post& Telecommunications Press,2002,11.”.Focus of the present invention assumes that channel estimation is not in channel estimation Accurately.

Further, in the above method, source node channel gain vectors described in the first step Destination node channel gain vectors described in the first stepNode is eavesdropped described in the first step Channel gain vectorsAntenna described in second step receives signal phasorAnd the vector of antenna additive white Gaussian noise described in second stepIn element arranged from small to large according to antenna number.

Further, in methods described, it is that zero, variance is σ that additive white Gaussian noise obeys average²Multiple Gauss point Cloth.

The present invention also provide simultaneously it is a kind of with the multi-user to being used for the pre- of safety of physical layer in end-to-end relay system The corresponding pre-coding apparatus of coding method, including N root reception antennas R₁,R₂,...,R_n,...,R_NWith N root transmitting antennas R₁', R₂',...,R_n',...,R_N', it is characterised in that the N roots reception antenna R₁,R₂,...,R_n,...,R_NWith N root transmitting antennas R₁',R₂',...,R_n',...,R_N' between be provided with N group signal samplings unit, N × N number of multiplying module and N number of superposition Module：The N roots reception antenna R₁,R₂,...,R_n,...,R_NFor reception antenna data, each reception antenna it is defeated Go out input of the end respectively with one group of signal sampling unit to be connected, the output end of each group of signal sampling unit is respectively It is connected with the first input end of N number of multiplying module simultaneously, the second input of the N × N number of multiplying module is successively An element in precoding vector b is received, the output end of the N × N number of multiplying module presses reception antenna data respectively Multiplication result numbering, the multiplication result of reception antenna data is exported to corresponding superposition module respectively Input, described each superposition module is calculated and relays signal all the way, described each superposition module Output end connect a transmitting antenna, the transmitting antenna R respectively₁',R₂',...,R_n',...,R_N' for forward it is described in After signal.

By taking n-th reception antenna as an example, n-th reception antenna after n-th group signal sampling unit sampling by obtaining dataThe output end of n-th group signal sampling unit first input end simultaneously with n-th × N-N+1 to n-th × N multiplying modules Connection, the second input of the n-th × N-N+1 to n-th × N multiplying modules receives precoding vector b in order successively In n-th × N-N+1 to n-th × N element, the output end of the n-th × N-N+1 to n-th × N number of multiplying module connects successively (other inputs of the 1st to N superposition module receive it respectively to connect the n-th input of the 1st to N superposition module The data of the multiplying module corresponding to his reception antenna), the described 1st to N superpositions module respectively by superposition Result export to the 1st to N transmitting antennas that (the relaying signal that now, n-th transmitting antenna is forwarded isWherein, m is the integer between 1 to N).

Beneficial effect

1st, have a safety feature, the present invention is applied to many eavesdropping node scenes using precoding technique.As long as meet condition N >= K+J, i.e. via node antenna number is not less than destination node and eavesdropping node number sum, it becomes possible to ensure relaying precoding Matrix dimensionality is sufficiently large, the kernel that there is equivalent interfering link and equivalent eavesdropping link, so as to secret signal is concentrated on It is transmitted in this kernel.This method is used, as long as via node number of antennas is enough, it is possible to ensure the exhausted of transmission information To safety.

2nd, multiuser interference is eliminated.The present invention when the design of pre-coding matrix B is carried out, by introduce it is equivalent disturb to AmountConstitute equivalent interference matrix F_k=[f_1,k,f_2,k,...,f_k-1,k,f_k+1,k,...,f_K,k]^TSo that prelist Precoding vector b and equivalent interference vector corresponding to code matrix BIt is orthogonal, thus eliminate it is equivalent disturb to AmountInfluence to receiving signal end at destination node, eliminates the multiuser interference at destination node.By This, destination node only needs to the source information needed for detecting itself, and carries out joint-detection without to institute's active node sending signal. This greatly reduces the detection complexity of destination node receiver.

3rd, information leakage is prevented.The present invention is vectorial by introducing equivalent eavesdropping when the design of pre-coding matrix B is carried outThe equivalent eavesdropping matrix G of composition_j=[g_1,j,g_2,j,...,g_K,j]^TSo that it is pre- corresponding to pre-coding matrix B Coding vector b and equivalent eavesdropping vectorIt is orthogonal, so as to eliminate eavesdropping node E_jThe private information that place receives x_k, prevent source node private information from revealing.

4th, realize simple.The present invention uses the Linear Precoding, mainly the reception signal to via node to carry out letter Single addition and multiplying.Only in conjunction with existing channel estimation technique, corresponding precoding vector b is generated, then to N group signals Sampling unit, N × N number of multiplying module and N number of superposition module are attached according to computing requirement, you can realize this hair Bright described precoding computing.Hardware realizes that simply practicality is stronger.

Brief description of the drawings

Fig. 1 is multi-user of the present invention to end-to-end relaying secure communication schematic diagram of a scenario；

Fig. 2 is signal processing flow figure of the present invention for safety of physical layer precoding；

Fig. 3 is relaying precoded signal processing structure exemplary plot.

Fig. 4 be the present invention carry safe pre-coding scheme relay antenna number be 4, eavesdropping interstitial content be 1 and 2 group of use In the case of family pair, system take input signal respectively BPSK, QPSK, 8PSK and 16PSK when resulting and safe transmission Speed is with signal to noise ratio situation of change；

Fig. 5 is that the present invention carries safe pre-coding scheme in QPSK inputs and relay antenna number in the case of 7,2 groups User is to the resulting and safe transmission speed when eavesdropping interstitial content is respectively 1,2,3 with signal to noise ratio situation of change.

Specific embodiment

Technical scheme implementation process is as follows：

Multi-user shown in Fig. 1 is to end-to-end relay communications system, including a via node R, K source node S₁, S₂,...,S_k,...,S_K, K destination node D₁,D₂,...,D_k,...,D_KWith J eavesdropping section E₁,E₂,...,E_j,...,E_JConstitute Multi-user to end-to-end relay system, in system, K source node and K destination node realize long distance under the assistance of relaying R From communication, each source node S_kK-th corresponding destination node D of difference_kOne group of user is constituted to { S_k,D_k, constitute altogether K group users are to { S₁,D₁},{S₂,D₂},...,{S_k,D_k},...,{S_K,D_K}.K groups user's centering, each destination node D_kOnly want to obtain source node S_kThe private information of transmission, and the transmission information of other source nodes is considered as interference.The K groups are used Family pair and J eavesdropping node are legal mobile subscribers (such as mobile phone) in cellular network, due to the limitation of mobile device size, Assuming that it is equipped with single antenna；And via node R is fixation means, comprising N root antennas R₁,R₂,...,R_n,...,R_N, the day Line radical N is for positive integer and meets N >=K+J, wherein, the user is positive integer to number K and eavesdropping node number J, and k is Any integer in the range of [1, K], represents each group of user to { S_k,D_kNumbering；J is any whole in the range of [1, J] Number, represents each eavesdropping node E_jNumbering；N is any integer in the range of [1, N], represents each antenna R_nNumbering.

Fig. 2 show multi-user to end-to-end trunking traffic flow chart, including the reception of channel estimation, signal, signal transacting With signal forwarding etc. process.

The first step：Channel estimation：The via node R receives whole users to { S first₁,D₁},{S₂,D₂},...,{S_k, D_k},...,{S_K,D_KAnd all eavesdrop node E₁,E₂,...,E_k,...,E_JThe positional information of feedback；Then according to the position Information calculates described each source node S of user's centering respectively_kThe each antenna R in the via node_nChannel gainEach antenna R in the via node_nTo each destination node of user's centering D_kChannel gainAnd Each antenna R in the via node_nTo each eavesdropping node E_jChannel gainAgain will it is described each Source node S_kTo each antenna R_nChannel gainIt is combined as source node channel gain vectorsBy each antenna R_nTo each destination node D_kChannel gainIt is combined as mesh Node channel gain vectorBy each antenna R_nTo each eavesdropping node E_j's Channel gainIt is combined as eavesdropping node channel gain vectorWherein, described source node Channel gain vectorsDestination node channel gain vectors With eavesdropping node channel gain vectorIt is the vector of N × 1 dimension.

Second step：Receive broadcast message：The user is to { S₁,D₁},{S₂,D₂},...,{S_k,D_k},...,{S_K,D_KIn Each source node S_kRespectively by via node to its corresponding destination node D_kSend private information x_k, then section is relayed Every antenna R in point_nThe signal for receiving isWherein, P_kRepresent described each source node S_kThe power of private information is sent,Represent n-th antenna R in the via node_nThe additive white Gaussian noise at place；Then will Each antenna R in the via node_nThe signal for receiving is combined as antenna and receives signal phasorWhereinIt is antenna additive Gaussian white noise Acoustic vector.Wherein, antenna receives signal phasor y_RWith antenna additive white Gaussian noise vector v_RN × 1 n dimensional vector n is,

3rd step：Precoding processing：The via node docking collection of letters numberEach element carry out Linear superposition computing, and it is combined into new precoding vectors signal By_RIt is sent on transmitting antenna.Wherein, B prelists for N × N-dimensional Code matrix, pre-coding matrix B l row q column elements are designated as B (l, q), and wherein l and q is the integer in the range of [1, N], and note is pre- Coding vector b is Ns of the pre-coding matrix B as obtained by row storehouse²× 1 dimensional vector, the generation step of the precoding vector b is such as Under：

Step 301：Calculate each destination node D_k(the K-1) × N at place²Tie up equivalent interference matrix F_k=[f_1,k,f_2,k,..., f_k-1,k,f_k+1,k,...,f_K,k]^T, wherein, except k-th source node S_kEach source node S in addition_iTo k-th destination node D_k Equivalent disturbance regime be expressed as source node S_iTo k-th destination node D_kEquivalent interference vector Table Show right Kronecker (Kronecker) accumulate computing, i ∈ [1, K [, i ≠ k.Now, k-th destination node D_kThe equivalent interference at place Link isFor k-th destination node D_k, it is only needed to from k-th source node S_kThe information of transmitting, and other sources are saved Point S_iThe information of (wherein i ≠ k) transmitting, to k-th destination node D_kFor be accordingly to be regarded as interference signal.Therefore, k-th section of mesh Point D_kThe equivalent interfering link at place is except source node S_kOutside institute's active node to destination node D_kEquivalent periodic line set, note It is F_kB, wherein F_k=[f_1,k,f_2,k,...,f_k-1,k,f_k+1,k,...,f_K,k]^TIt is (K-1) × N²Dimension matrix.

Step 302：Calculate each eavesdropping node E_jThe equivalent eavesdropping matrix G at place_j=[g_1,j,g_2,j,...,g_K,j]^T, wherein, Each source node S_kTo j-th eavesdropping node E_jEquivalent eavesdropping situation be expressed as source node S_kTo j-th eavesdropping node E_j's Equivalent eavesdropping vectorNow, j-th eavesdropping node E_jPlace equivalent eavesdropping link beTherefore, J eavesdropping node E_jThe equivalent eavesdropping link at place is institute's active node to eavesdropping node E_jEquivalent periodic line set, be designated as G_jB, its Middle G_j=[g_1,j,g_2,j,...,g_K,j]^TIt is K × N²Dimension matrix.

Step 303, calculates the kernel T of equivalent interference matrix and equivalent eavesdropping matrix：By all equivalent interference matrix F_k =[f_1,k,f_2,k,...,f_k-1,k,f_k+1,k,...,f_K,k]^TWith equivalent eavesdropping matrix G_j=[g_1,j,g_2,j,...,g_K,j]^TMerge into Dimension is (K+J-1) K × N²Interference eavesdropping matrixAnd matrix Φ is carried out very Different value is decomposedWherein, dimension is N²×N²Matrix V_ΦIt is the right characteristic vector square of interference eavesdropping matrix Φ Battle array；Dimension is (K+J-1) K × N²Matrix Σ_ΦIt is the eigenvalue matrix of interference eavesdropping matrix Φ, dimension is (K+J-1) K × (K + J-1) K matrix U_ΦIt is the left eigenvector matrix of interference eavesdropping matrix Φ；Then the kernel T of interference eavesdropping matrix is by the right side Eigenvectors matrix V_ΦN²The column elements of-K (K+J-1)+1 are to N²The N that column element is constituted²×(N²-K²- KJ+K) dimension matrix, can table It is shown as：T=V_Φ(:,[N²-K(K+J-1)+1:N²])。

Step 304, precoded signal treatment eliminates multi-user interference and information leakage：Setting precoding vector isWherein P_RIt is the maximum of repeating device power； It is N × N-dimensional matrix, represents the covariance of via node received signal power；I_NIt is N × N-dimensional unit matrix；(N²-K²-KJ+K)×1 Equivalent wave beam forming vector p is tieed up for all purposes node receiving power covariance matrixEigenvalue of maximum correspondence Unit character vector, wherein, described each source node S of user's centering_kTo corresponding destination node D_kThe equivalent shape that direct transfers Condition is expressed as source node S_kTo corresponding destination node D_kThe equivalent vector that direct transfers·^HRepresent altogether Yoke transposed matrix computing,^TIt is transposed matrix computing.Then pre-coding matrix B is assigned according to B (l, q)=b (qN-N+l) Value.

4th step：Forward signal：Each antenna R in the via node R_nAccording toTurn Hair precoding vectors signal.

By k-th destination node D_kSignal is received to be designated asIt is represented by WhereinRepresent destination node D_kThe additive white Gaussian noise at place, it is that zero, variance is σ to obey average²Multiple Gauss distribution.Receive SignalIn multi-user interference be eliminated, only include user D_kDesired private information x_kWith noise two parts.Therefore, user D_kCan be according to reception signalCarry out independent detection private information x_k, without to institute's active node sending signal joint-detection, Oneself desired information is extracted again.

On the other hand, by j-th eavesdropping node E_jThe signal of reception is designated asIt is represented by： WhereinRepresent eavesdropping node E_jThe additive white Gaussian noise at place, it is that zero, variance is σ to obey average²Multiple Gauss distribution.Due to Eavesdropping node only includes noise in receiving signal, therefore, all eavesdropping nodes do not get any private information, multi-user couple Secure communication between end-to-end is effectively ensured.

Fig. 4 and Fig. 5 sets forth Precoding Design method of the institute of the present invention drawings in safety of physical layer in different inputs The graph of a relation of under the conditions of signal and different eavesdropping interstitial contents and safe transmission speed and signal to noise ratio.As shown in Figure 4, the present invention Method is directed to different limited character input signals, ensure that safe transmission speed is issued to its capacity in high s/n ratio Boundary.Fig. 5 shows that the inventive method can guarantee that the safe transmission of information in the case of different eavesdropping interstitial contents.

Claims (5)

1. a kind of multi-user in end-to-end relay system be used for safety of physical layer method for precoding, the system include one Via node (R), K source node (S₁,S₂,...,S_k,...,S_K), K destination node (D₁,D₂,...,D_k,...,D_K) and J Eavesdropping node (E₁,E₂,...,E_j,...,E_J), each source node (S_k) k-th corresponding destination node of difference (D_k) one group of user is constituted to ({ S_k,D_k), K group users are constituted altogether to ({ S₁,D₁},{S₂,D₂},...,{S_k,D_k},...,{S_K, D_K), the via node (R) includes N root antennas (R₁,R₂,...,R_n,...,R_N), the radical antenna N is positive integer and expires Sufficient N >=K+J, wherein, the user is positive integer to number K and eavesdropping node number J, and k is any whole in the range of [1, K] Number, represents each group of user to ({ S_k,D_k) numbering；J is any integer in the range of [1, J], represent it is described each Eavesdropping node (E_j) numbering；N is any integer in the range of [1, N], represents each antenna (R_n) numbering, it is special Levy and be, precoding step is as follows：

The first step, channel estimation：The via node (R) receives whole users couple first

({S₁,D₁},{S₂,D₂},...,{S_k,D_k},...,{S_K,D_K) and all eavesdrop node (E₁,E₂,...,E_k,...,E_J) The positional information of feedback；Then each source node of user's centering (S is calculated according to the positional information respectively_k) described in Each antenna (R in via node_n) channel gainEach antenna (R in the via node_n) arrive the user Each destination node of centering (D_k) channel gainAnd each antenna (R in the via node_n) to it is described each Eavesdropping node (E_j) channel gainAgain will each source node (S_k) arrive each antenna (R_n) channel gainIt is combined as source node channel gain vectorsBy each antenna (R_n) arrive each Individual destination node (D_k) channel gainIt is combined as destination node channel gain vectors By each antenna (R_n) eavesdrop node (E to each_j) channel gainIt is combined as eavesdropping node channel gain arrow Amount

$h_{{\mathrm{RE}}_j}={\[h_{R_1{E}_j},h_{R_2{E}_j},...,h_{R_N{E}_j}\]}^T;$

Second step：Receive broadcast message：The user couple

({S₁,D₁},{S₂,D₂},...,{S_k,D_k},...,{S_K,D_K) in each source node (S_k) respectively by relaying Node is to its corresponding destination node (D_k) send private information x_k, then every antenna (R in via node is remembered_n) letter that receives Number it isBy each antenna (R in the via node_n) signal that receives is combined as antenna and receives signal phasor

3rd step：Precoding processing：Via node receives signal phasor to antennaCarry out precoding Computing, obtains precoding vectors signal By_R, wherein B is N × N-dimensional pre-coding matrix, pre-coding matrix B l row q column elements B (l, q) is designated as, wherein l and q is the integer in the range of [1, N], and note precoding vector b presses row storehouse institute for pre-coding matrix B The N for obtaining²× 1 dimensional vector, the generation step of the precoding vector b is as follows：

Step 301：Calculate each destination node (D_k) place equivalent interference matrix

F_k=[f_1,k,f_2,k,...,f_k-1,k,f_k+1,k,...,f_K,k]^T, wherein, the i-th source node (S_i) to k-th destination node (D_k) Equivalent interference vectori∈[1,K],i≠k；

Step 302：Calculate each eavesdropping node (E_j) place equivalent eavesdropping matrix

G_j=[g_1,j,g_2,j,...,g_K,j]^T, wherein, kth source node (S_k) eavesdrop node (E to j-th_j) equivalent eavesdropping vector

Step 303：The kernel T of equivalent interference matrix and equivalent eavesdropping matrix is calculated, by all equivalent interference matrix F_k= [f_1,k,f_2,k,...,f_k-1,k,f_k+1,k,...,f_K,k]^TWith equivalent eavesdropping matrix G_j=[g_1,j,g_2,j,...,g_K,j]^TMerge into dry Disturb eavesdropping matrixSingular value decomposition is carried out to interference eavesdropping matrix ΦThe then kernel of equivalent interference matrix and equivalent eavesdropping matrix

T=V_Φ(:,[N²-K(K+J-1)+1:N²])；

Step 304：Precoded signal treatment eliminates multi-user interference and information leakage：Precoding vector is calculated according to the following equationWherein P_RIt is the maximum of repeating device power；N × N-dimensional matrix Q is via node The covariance matrix of received signal power,

Wherein P_kIt is k-th source node (S_k) send private information power,For source is saved Point channel gain vectorsConjugate matrices,It is source node channel gain vectorsTransposed matrix；I_NIt is N × N-dimensional Unit matrix；σ²It is the variance of additive white Gaussian noise；

(N²-K²- KJ+K) × 1 equivalent wave beam forming vector the p of dimension is all purposes node receiving power covariance matrixThe corresponding unit character vector of eigenvalue of maximum, wherein, source node (S_k) arrive corresponding destination node (D_k) The equivalent vector that direct transfersPre-coding matrix B (l, q)=b is finally obtained according to precoding vector b (qN-N+l)。

4th step：Forward signal：Each antenna (R in the via node (R)_n) according toTurn Hair precoding vectors signal.

2. multi-user as claimed in claim 1 in end-to-end relay system be used for safety of physical layer method for precoding, its It is characterised by, the source node channel gain is calculated using the channel estimation mode based on least mean-square error in the first step VectorDestination node channel gain vectorsAnd eavesdropping Node channel gain vector

3. multi-user as claimed in claim 2 in end-to-end relay system be used for safety of physical layer method for precoding, its It is characterised by, source node channel gain vectors described in the first stepMesh described in the first step Node channel gain vectorNode channel gain vector is eavesdropped described in the first stepAntenna described in second step receives signal phasor And the vector of antenna additive white Gaussian noise described in second stepIn element according to antenna compile Number arrange from small to large.

4. multi-user as claimed in claim 1 in end-to-end relay system be used for safety of physical layer method for precoding, its It is that zero, variance is σ to be characterised by that the additive white Gaussian noise obeys average²Multiple Gauss distribution.

5. a kind of and multi-user described in claim 1 is to the method for precoding pair in end-to-end relay system for safety of physical layer The pre-coding apparatus answered, including N root reception antennas (R₁,R₂,...,R_n,...,R_N) and N root transmitting antennas (R₁′,R₂′,..., R_n′,...,R_N'), it is characterised in that the N roots reception antenna (R₁,R₂,...,R_n,...,R_N) and N root transmitting antennas (R₁′, R₂′,...,R_n′,...,R_N') between be provided with N group signal samplings unit, N × N number of multiplying module and N number of superposition mould Block：

The N roots reception antenna (R₁,R₂,...,R_n,...,R_N) be used for reception antenna data, each reception antenna it is defeated Go out input of the end respectively with one group of signal sampling unit to be connected, the output end of each group of signal sampling unit is respectively It is connected with the first input end of N number of multiplying module simultaneously, the second input of the N × N number of multiplying module is successively An element in precoding vector b is received, the output end of the N × N number of multiplying module presses reception antenna data respectively Multiplication result numbering, the multiplication result of reception antenna data is exported to corresponding superposition module respectively Input, described each superposition module is calculated and relays signal all the way, described each superposition module Output end connect a transmitting antenna, the transmitting antenna (R respectively₁′,R₂′,...,R_n′,...,R_N') described for forwarding Relaying signal.