WO2010124488A1

WO2010124488A1 - Configuration method and apparatus for pre-coding weight in multi-cell

Info

Publication number: WO2010124488A1
Application number: PCT/CN2009/073793
Authority: WO
Inventors: 魏巍
Original assignee: 中兴通讯股份有限公司
Priority date: 2009-04-30
Filing date: 2009-09-07
Publication date: 2010-11-04
Also published as: CN101877607A; CN101877607B

Abstract

A configuration method and apparatus for pre-coding weight in multi-cell are provided in the present invention, the method involves: for each user of multiple cells, the first weight of the user is obtained based on the principle of maximizing signal to leakage and noise ratio(S102); for each user, the game function of the user is obtained according to the action set and the utility function of the user(S104); for each user, the first weight of the user is configured according to the game function of the user and the first weight of the other user, who interferes this user in multiple cells, in which, when the first weight configured for the user makes the utility function of the user maximization, and the utility function satisfies the nash equilibrium, the configured first weight is used as the pre-coding weight of the user(S106). With the invention, we can choose a logical pre-coding weight based on the selection criteria and method of game theory, obtain better effect of interference suppression in multi-cell multi-user, avoid the questions in related technology that the effect of interference suppression is poor, the interference is difficult to achieve, and the influence on the structure of the existing system is great.

Description

TECHNICAL FIELD The present invention relates to the field of communications, and in particular, to a method and apparatus for configuring a multi-cell precoding weight. BACKGROUND OF THE INVENTION After nearly a decade of development, multi-input multi-output (MIMO) technology has evolved from point-to-point single-link communication to single-cell multi-user communication. And, with the advancement of next-generation mobile communication technologies and standards, MIMO technology has begun to move from real-to-the-room to practical applications. However, the neighbor cell interference problem in multi-antenna multi-cells affects the application of MIMO technology in extended multi-cell systems. Through a series of system-level simulation and theoretical verification, it is found that when multi-antenna transmission and reception technology is applied in a multi-cell communication environment with co-channel interference, the frequency efficiency of the system will be severely limited, thus making the system An interference-limited system with a more serious interference situation. For multi-cell MIMO systems, since the system uses MIMO technology, many interference suppression technologies for single-user MIMO and multi-user MIMO can be extended to multi-cell systems. The most commonly used trick is based on single-user MIMO and Multi-user MIMO is especially good! ³ methods of technology. These methods can be generally divided into two types: interference suppression technology based on the receiving end and interference suppression technology based on the transmitting end. In practical applications, the application of interference suppression technology based on the transmitting end is relatively common. The interference-based interference suppression technology mainly uses multi-cell joint pre-coding to suppress neighbor cell interference at the transmitting end. The specific suppression strategy is usually based on the following assumptions: All user data and channel information to be transmitted can be realized between the base stations. Fully cooperative, that is, a distributed base station system whose application environment is fully cooperative, can be considered as a direct promotion of single-link MIMO to multi-cell systems (equivalent to distributing transmit antennas to different base stations). Based on the above assumptions, although inter-base station cooperation can obtain higher spectral efficiency, the following problems still exist in the actual system implementation process: (1) In order to achieve joint precoding, each base station must acquire real-time real-time of all communication users in the system. Channel information, which is more in the number of users and users It is difficult to achieve in a high-speed mobile environment; (2) Linear joint precoding is very sensitive to the characteristics of the channel matrix. When the channel is not full or the condition number of the channel is large (the variance of the channel matrix singular value distribution is large), system performance It will drop significantly, and even in the case of serious conditions, it may even lead to the inability to use the ZF method; (3) The transmitted signal from multiple base stations usually generates a large time offset when it arrives at the receiver. The joint coding technique requires the base station to transmit data before it Each user channel is pre-time compensated, and for high data rate transmission or receiver location away from the cell boundary, it may be difficult for the base station to obtain the timing of the real-time channel. Research shows that the signal synchronization problem in the joint coding system leads to joint coding. The performance of the system is seriously degraded. If you want to avoid this problem, you need to use more sophisticated sender control technology, which will increase the complexity of the system. In addition, in order to enable each user to communicate with the surrounding base stations simultaneously, the multi-cell joint coding system also faces a problem: the system needs to make a large adjustment to the link design of the existing system, and even needs to change the existing communication. The topology of the system, the power control of each base station in the cell is also a very complicated problem, which will undoubtedly increase the cost of the mobile operator. In view of the difficulty in implementing interference suppression in a multi-cell MIMO environment in the related art, the implementation cost and complexity are high, and even the existing topology of the network is affected, and an effective solution has not been proposed yet. SUMMARY OF THE INVENTION The present invention has been made in view of the related art in the multi-cell MIMO environment, in which the implementation of the interference is particularly difficult, the implementation cost and complexity are high, and even the existing topology of the network is affected. The purpose is to provide a method and device for configuring multi-cell precoding weights. According to an aspect of the present invention, a method for configuring a multi-cell precoding weight is provided. The method for configuring a multi-cell precoding weight according to the present invention includes: obtaining, for each user of the plurality of cells, a first weight of the user based on a maximum signal to noise ratio; for each user, according to the The user's policy set and the user's revenue function obtain the user's game function; for each user, configure the user's number according to the game function and the first weight of other users in the plurality of cells that cause interference to the user a weight, wherein, when the first weight configured for the user maximizes the user's income function and the benefit function satisfies the Nash balance, the configured first weight is used as the precoding weight of the user. Wherein, for each user, obtaining the first weight of the user based on the maximum information leakage ratio principle includes:

The definition of the noise-to-noise ratio is believed to be: and the minimum mean square error criterion is 1.

w _i ∞EV((a ² I + [H _l --H _i _ _i , H _M --H _t T[H _l --H _i _ _l ,H _i+l --H _t ]r ¹ H: H _i ). Let _; =[ 〜 _ ₁ , ₊₁ 〜^] denote the set of all interfering user channel matrices of the i-th user, and get:

Maximizing SLNR based on the minimum mean square error criterion yields:

Where " ^νΛ '· indicates the letter-to-noise ratio of the i-th user, ' is the channel matrix of the i-th user, ^ is the first weight of the i-th user, and ^σ ' is the noise power of the i-th user. For the channel matrix of the dry 4 user in the cell where the user is located and in the neighboring cell, K is the number of all the users of the dry 4 user.

^EV W represents the eigenvector decomposition of the matrix. The method may further include: obtaining a channel matrix of the i-th user in advance by one of: in a time division duplex system, estimating an uplink channel of each user by utilizing reciprocity of the channel, to obtain an equivalent The channel matrix of each user; in the frequency division duplex system, the channel information fed back by each user is obtained through limited feedback, and the eigenvector decomposition of the channel information of each user and the channel information of the user of the user is obtained. The user's channel matrix. And, the expression of the game function G of each user is:

Where N is a set of users with co-channel interference in each cell, and

For the income function of the i-th user, S is the policy set of user i, and the expression of ^ύ '· is as follows:

Where ^ ^ is the channel matrix of the i-th user, ^W "is the first weight of the i-th user, ^σ ' is the noise power of the i-th user, K is the number of all interfering users, and is the number of interfering users a weight value. Configuring a first weight of the user according to a game function of each user and a first weight of a user in the plurality of cells that causes interference to the user includes: bringing the expression into the user's revenue Function, get the user's income function with the user's u^S^ ² '■ w _i

The relationship between the first weights is as follows: , and the derivation is obtained:

Based on O relative to ^W 'derivative, further solution

, „ M, (S, )

, and jHhi determines the maximum value of '', where is the user's cost factor, which is a steep factor; configures ¹ ^ ', and when the configured ^ ' makes the corresponding ^ ¹ maximum satisfy the Nash equilibrium The first weight of the configuration is used as the precoding weight of the user. In addition, for each user, the gain function satisfies the Nash equilibrium means: ^U i ₊ l ~ ^U i < ε Among them, is the income function of the i-th user, and ε represents the convergence precision. Preferably, the foregoing policy set includes at least one of the following: a signal to noise ratio, a signal to noise ratio, and a signal to interference and noise ratio. According to another aspect of the present invention, a configuration apparatus for multi-cell precoding weights is provided. The apparatus for configuring a multi-cell precoding weight according to the present invention includes: a first obtaining module, configured to obtain, for each user of the plurality of cells, a first weight of the user based on a maximum information leakage noise ratio principle; An obtaining module, configured to obtain, for each user, a game function of the user according to the policy set of the user and the income function of the user; a configuration module, configured, for each user, according to the game function of the user and multiple a first weight of another user that interferes with the user in the cell, configured with a first weight of the user, where the first weight configured for the user maximizes a revenue function of the user, and the revenue function satisfies When Nash balances, the first weight of the configuration is used as the precoding weight of the user. The processing by the first obtaining module for obtaining the first weight of the user for each user includes:

The first acquisition module is based on the definition of the leakage ratio:

And the minimum mean square error criterion is obtained:

w ? ( ² +[ r" — +Γ'ΆΠ ... , +Γ'Ά]) — ¹ ,); Let ^^[^'''H Hw''^] denote all interfering user channels of the i-th user The collection of matrices, get:

SL muscle

+||H, I . Maximize SLNR based on the minimum mean square error criterion, yield: w. = arg max ~ ^WiHiHi ~, ~ = max EV((Na ² I + Η*Η) ¹ Η'Η.)

Where ^SLNR i represents the information leakage ratio of the i-th user, which is the channel matrix of the i-th user. ^W ' is the first weight of the i-th user, and ^σ ' is the noise power of the i-th user, which is the channel matrix of the user in the cell where the user is located and in the neighboring cell, K is all the dry 4 The number of users,

^EV ^ represents the eigenvector decomposition of the matrix. And, the expression of the game function G of each user is:

Where N is a set of users with co-channel interference in each cell, and ¹ '''·!; is the revenue function of the i-th user, which is the policy set of user i, and the expressions are as follows:

Where ^ is the channel matrix of the i-th user, ^w ' is the first weight of the i-th user, ^σ ' is the noise power of the i-th user, Κ is the number of all interfering users, ^ is the number of interfering users A weight. Through the above technical solution of the present invention, the weight selection method based on the principle of maximizing the signal leakage ratio in the cell and optimizing the policy set principle based on the game theory in the cell is not changed, so as not to change the topology structure and the communication mode of the existing cell. As a premise, the selection criteria and methods based on game theory select reasonable precoding weights, so that better multi-cell multi-user dry suppression can be obtained, and the feedback overhead and computation complexity can be improved while minimizing feedback overhead and computational complexity. The system performance and anti-interference ability avoid the problem that the interference suppression effect in the related technology is poor, the interference is difficult to be realized, and the structure of the existing system is greatly affected. BRIEF DESCRIPTION OF THE DRAWINGS The accompanying drawings, which are set to illustrate,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, In the drawings: FIG. 1 is a flowchart of a method for configuring a multi-cell precoding weight according to an embodiment of a method of the present invention; FIG. 2 is a schematic diagram of a processing example of a method for configuring a multi-cell precoding weight according to an embodiment of the method of the present invention; FIG. 3 is a configuration of a multi-cell precoding weight according to an embodiment of the method of the present invention; Schematic diagram of signal leakage in the method; FIG. 4 is a schematic block diagram of a system based on hierarchical beamforming applicable to a method for configuring multi-cell precoding weights according to an embodiment of the method of the present invention; FIG. 5 is a schematic diagram of an apparatus according to an embodiment of the present invention. A block diagram of a device for configuring a cell precoding weight. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT The traditional weight-to-interference (SINR)-based weighting method for overcoming co-channel interference (such as ZF, minimum mean square error (MMSE) method, etc.) has certain limitations on the number of antennas ( Generally, the transmitting antenna is larger than the total number of receiving antennas of all users. Therefore, it is impossible to generate a reasonable weight, which results in poor interference suppression, difficulty in achieving interference, and even affects the structure of the existing system. Based on the above problems, the present invention provides a pre-coding weight determination method that is easy to implement and has low complexity. The invention proposes to optimize the signal to interference and noise ratio (SINR) based on the game theory based on the principle of maximizing the signal leakage ratio (the ratio of the signal to the energy and noise of the signal leaked to other users, referred to as SLNR) in the cell. ), the signal-to-noise ratio (SNR), and/or the signal-to-leakage-to-noise ratio (SLNR) principle weight selection idea, based on the premise of not changing the existing cell topology and communication mode, based on game theory selection criteria and method selection Reasonable precoding weights can achieve better multi-cell multi-user interference suppression, and improve system performance and anti-interference ability while minimizing feedback overhead and computational complexity. Embodiments of the present invention will be described in detail below. It should be noted that the embodiments in the present application and the features in the embodiments may be combined with each other without conflict. Method Embodiment In this embodiment, a method for configuring a multi-cell precoding weight is provided. 1 is a flowchart of a method for configuring a multi-cell precoding weight according to an embodiment of the present invention, and it should be noted that The steps described in the following methods may be performed in a computer system such as a set of computer executable instructions, and although the logical order is illustrated in FIG. 1, in some cases, may be in a different order than jt匕Perform the steps shown or described. As shown in FIG. 1 , the method for configuring a multi-cell precoding weight according to this embodiment includes the following processing: Step S102: For each user of multiple cells, obtain the first user according to a maximum information leakage ratio a weight value; Step S104, for each user, obtaining a game function of the user according to the policy set of the user and the income function of the user; Step S106, for each user, according to the user's game function and a plurality of cells The first weight of the user causing the interference, the first weight of the user is configured, wherein when the first weight configured for the user maximizes the user's income function and the income function satisfies the Nash balance The first weight of the configuration is used as the precoding weight of the user. For example, for the user i, the user needs to be configured according to the first weight of the user who interferes with the user (ie, user i) in multiple cells. The first weight of i, if the first weight of user i makes the income function of user i the largest and can satisfy the Nash balance, the first weight at this time is taken as the precoding weight of user i The method of the present embodiment will be described in detail below. Step 1: First, perform the first training in the cell based on the principle of maximizing the signal leakage ratio; wherein, the expression of the signal to noise ratio is as follows:

SLNR _t = k= Σ ^k≠i Ι ² ( 1 ) where SLNR represents the information leakage noise ratio of the i-th user, and _Wi and σ are respectively the channel matrix of the i-th user (this matrix can be used as the user's Channel information), weight vector and noise power,

^Hk is the channel matrix of the ^inter -cell and neighboring cell interfering users (here, the interfering user can be a strong interfering user with strong interference), and K is the number of all interfering users. With the aim of maximizing (1) above, based on the MMSE criterion, the following formula (2) can be obtained: ^∞ Εν{(σ ² Ι + [H _r - , H _i+l -H IH,... H , H _M ... H _t ]) - ¹ HH,. )

(2) In the formula (2), (·) represents the eigenvector decomposition of the matrix. It can be seen from the formula (2) that the weight vector is related to the channel information of the target user and the interfering user. Specifically, for the acquisition of the channel information (that is, the channel matrix described above), the following two methods may be used: In the time division duplex system, by using the reciprocity of the channel, the uplink channel of each user can be estimated on the base station side to obtain the channel information of the downlink channel in an equivalent manner. Method 2: In the frequency division duplex system, the base station Through the limited feedback, the channel information fed back by each user is obtained, and the channel information of the target user i and the interfering user is subjected to feature vector decomposition according to the formula (2). After that, let = ... ^H ", ^H i"'' Ά ] represent the set of all interfering user channel matrices of the i user, then (1) can be obtained:

Targeting the maximization of (3), based on the MMSE criterion, looking for precoding weights, you can get:

(4) The result of the first weight training of the user i in the cell can be obtained by using the above formula (4) (ie: the first weight value described above). The processing of the above step one may correspond to step S102 in FIG. Step 2: Perform a second training on the 'J, interval same frequency users based on the game theory; In this step two, each user in the cell will get a weight ^, and k represents all users in the cell. If all users of the multi-cell adopt the same method, malicious competition will occur between the inter-cells, especially the same-frequency users in different cells. Therefore, it is necessary to perform the weighting of the same-frequency users. Second training based on game theory. From the perspective of game theory, interference coordination between multiple cells can be regarded as a non-cooperative game process. In the case of co-channel interference, each user strives to maximize the utility. Corresponding to game theory, we can use (5) to represent the game function G of each user:

G = [V, 7 ?.,{^(-)}] (5) where, the function represents a multi-cell non-cooperative power allocation discard model, W^ ¹ ' ² ''"' ⁷ } for each cell has the same a set of users of frequency interference (ie, a set of participants); a set of policies of user i, wherein the set of policies may also be other parameters, for example, in addition to SINR, SNR SLNR may be used as a policy set, and Described using a similar expression.

In the case of SINR as a policy set, the expression of the policy set is as shown in equation (6), which is the revenue function of user i.

\\H _: w _:

SINR _; κ ( 6 ) As shown in equation ( 6 ), for the ^S leg ' of user i , the weight of the target user ^ is a contradiction with the weight ^w of the interfering user, and can be used by the following formula ( 7 ) The income function ^ (to indicate: SJ1VR

IN^ + a ( 7 ) In (7), “is a steep factor whose value can be configured according to the actual situation of the system; as a cost factor constant, it is used to define the cost to the user when it is disturbed. And, for all users, "can be the same. Based on the following formula (8), the equation (7) can be made The largest income function:

SINR

Arg max u _t (SINR _t ) ² - arg max( '■ w _t )

SINR, + a

(8) In order to obtain the income function satisfying the formula (8), we can derive the (7) formula for ^w ': du) _ du djSINR,) _ _Q

_{! dw _ d (SINR t)} dw t _ (9) of the equation (9) is solved to give:

Equation (10) further illustrates that the weight selection between users is an interactive, dynamic process, so in order to maximize its SINR and minimize the interference between users, it is necessary to find a Nash equilibrium point. , to achieve the best steady state of the system. According to the Nash equilibrium point definition, when the income function satisfies (11), the Nash equilibrium point exists and is unique:

^U M ~ ^U - ^S (ii) where ^£ represents the convergence precision and is a minimum value. Based on the above processing, it is possible to adjust ^ according to (10) and in combination with (4) until the obtained formula (11) is satisfied, and at this time, the pre-coding weight of the i-th user, the second training ends. The processing of step 2 may correspond to step S104 and step S106 in FIG. 1, and by means of the game principle in step two, malicious contention between users of the same frequency can be effectively avoided. As shown in FIG. 2, in practice, the weight configuration method of the present invention may include the following steps: (21) According to the above manner 1 and/or mode 2, the primary station collects channel information from the target user and the dry 4 user;

(22) Perform the first training in the cell based on the principle of maximizing the signal leakage ratio; the expression of the signal leakage ratio is as shown in the above formula (1), and the principle is as shown in Fig. 3, ^, ^w '. ^σ ' ² is the channel matrix, weight vector and noise power of the i user respectively, which is the channel matrix of the user and the neighboring cell strongly interfering with the user, and K is the number of all interfering users. In order to maximize the above formula (1), based on the MMSE criterion, the channel information of the target user and the interfering user is decomposed by using (4). In this way, the result of the first weight training of the user i in the cell (the first weight mentioned above) can be obtained. (23) Performing a second training on inter-cell co-frequency users based on game theory; after the above step (22), each user in the cell will get a weight ^w , and k represents all users in the cell. If all users of the multi-cell adopt the same method, malicious competition will occur between the inter-cells, especially the same-frequency users in different cells. Therefore, it is necessary to perform the weighting of the intra-cell co-frequency users. Second training based on game theory. From the perspective of game theory, interference coordination between multiple cells can be regarded as a non-cooperative game process. In the case of co-channel interference, each user strives to maximize the utility. Corresponding to the game theory, the second weight training can be performed by the multi-cell non-cooperative power allocation game model of (5), wherein the signal to interference and noise ratio function and the income function are respectively shown in equations (6) and (7). The goal of the game is to maximize the benefit function shown in equation (7) and adjust it according to (10). (10) shows that the weight selection between users is an interactive, dynamic process. To maximize its SINR and minimize interference between users, it is necessary to find a Nash equilibrium point to achieve optimal steady state. According to the Nash equilibrium point definition, when the income function satisfies (11), the Nash equilibrium point exists and is unique. Therefore, it can be adjusted according to (10) and combined with (4) until the formula (11) is satisfied, and the second training ends. (24) As shown in FIG. 4, the pre-processing performed at the origin may be performed by using the weights obtained above, including: encoding and modulating the input bit stream, performing serial-to-parallel conversion, multi-cell pre-processing, and then Multiple transmit antennas (XI to XMt) are transmitted; after that, the receiver can receive signals through multiple receive antennas (yl to yNt) for equivalent channel estimation, space-time detection, parallel-to-serial conversion, demodulation and decoding. , get the output bit stream. Through the above processing, a reasonable precoding weight can be selected based on the selection criterion and method of the game theory, so that a better multi-cell multi-user interference suppression effect can be obtained, and the feedback overhead and the computation complexity are minimized. Improve system performance and anti-interference ability. According to an embodiment of the present invention, there is also provided a computer readable medium, wherein the computer readable means stores computer executable instructions that, when executed by a computer or processor, cause the computer or processor to execute As with the processing of the respective steps shown in FIGS. 1 and 2, preferably, one or more of the above embodiments may be performed. Apparatus Embodiment In this embodiment, a multi-cell precoding weight setting apparatus is provided to perform the above processing. FIG. 5 is a block diagram of a configuration apparatus of multi-cell precoding weights according to the present embodiment. As shown in FIG. 5, the apparatus for configuring multi-cell precoding weight according to the embodiment includes: a first ear 4, for each user in multiple cells, based on maximum information leakage noise Obtaining the first weight of the user by the principle; the second obtaining module 2 is connected to the first obtaining module 1 for obtaining, for each user, the game function of the user according to the policy set of the user and the income function of the user The configuration module 3 is connected to the second acquisition module 2, and for each user, configures the user according to the game function of the user and the first weight of other users in the plurality of cells that cause the user to be dry. a first weight, wherein, when the first weight configured for the user maximizes a profit function of the user, and the benefit function satisfies a Nash balance, the configured first weight is used as the precoding weight of the user . The process of obtaining, by the first obtaining module 1 for each user, the first weight of the user based on the maximum information leakage ratio principle may include: It is believed that the definition of the noise to noise ratio is:

Guidelines

_Wl ∞EV((a ² IHH _l --H _l _ _i , H _M --H _k r[H _i --H _i _ _l ,H _M --H _k ]r ^l H;H _i ) Let ^... !^,^...!^ indicates the acquisition of all interfering user channel matrices for the first user:

SLNR, = ~

^ ² +||H, ,.|| . Maximize SLNR based on MMSE criteria, get w. = arg max ~ ^WiHiHi ~ , ~ = max EV((Na ² I + Η'Η Υ ¹ H'H )

Where " ^νΛ '· indicates the letter-to-noise ratio of the i-th user, ' is the channel matrix of the i-th user, ^w ' is the first weight of the i-th user, and ^σ ' ² is the noise of the i-th user The power is the channel matrix of the interfering users in the cell where the user is located and in the neighboring cell, where K is the number of all interfering users.

^EV W represents the eigenvector decomposition of the matrix. In addition, the expression of the game function G of each of the above users is:

N is a set of users with co-channel interference in each cell, and 1, , , ¹ ; ^ > is the revenue function of the i-th user, which is the policy set of user i, and the expression of S is as follows:

Where ^ is the channel matrix of the i-th user, w ' is the first weight of the i-th user, ^σ ' is the noise power of the i-th user, K is the number of all interfering users, ¹ ^ is the interference user The first weight. The configuration module 3 can process the first weight of the user according to the game function of each user, the first weight of the user, and the first weight of the user in the plurality of cells that is caused by the user. Including: The expression is brought into the user's income function, and the relationship between the user's income function and the user's first weight is as follows:

, and for ^W ' to get:

Based on ^UiiS relative to the derivative, further solution

And the maximum value determined by jtb, where is the user's cost factor, ^α is the steep factor; configure for ¹ ^, and when the configured ^w ' is such that the corresponding maximum value satisfies the Nash equilibrium, the first configuration will be The weight is used as the precoding weight of the user. In the above description, for each user, its income function satisfies the Nash balance.

^{≤ £} , where ^W '' is the income function of the i-th user, and s is the convergence precision. Optionally, the foregoing policy set includes at least one of the following: a signal to noise ratio, a signal to noise ratio, and a signal to interference and noise ratio. Through the above device, a reasonable precoding weight can be selected based on the selection criteria and method of the game theory, so that a better multi-cell multi-user dry/non-suppression effect can be obtained, and feedback is minimized. Improve system performance and anti-interference ability with overhead and computational complexity. In summary, with the technical solution of the present invention, by using the criterion of maximizing the signal leakage ratio, the selection method of the weight is no longer limited by the number of antennas and can take into account the influence of noise without degrading performance. In the case, the application scenario of multi-cell multi-user interference cancellation based on the pre-coding method is extended; and the weight second training is performed based on the game theory between the inter-cell co-frequency users, thereby avoiding malicious between users in the cell. Contention and conflict; In addition, by obtaining how to obtain interference channel information of the user and the target user in different duplex modes, the channel information required by the base station is acquired without increasing the feedback overhead. In addition, the implementation of the present invention does not tamper with the system architecture and the current processing flow, is easy to implement, facilitates promotion in the technical field, and has strong industrial applicability. Obviously, those skilled in the art should understand that the above modules or steps of the present invention can be implemented by a general-purpose computing device, which can be concentrated on a single computing device or distributed over a network composed of multiple computing devices. Alternatively, they may be implemented by program code executable by the computing device, such that they may be stored in the storage device by the computing device, or they may be separately fabricated into individual integrated circuit modules, or they may be Multiple modules or steps are made into a single integrated circuit module. Thus, the invention is not limited to any specific combination of hardware and software. The above is only the preferred embodiment of the present invention, and is not intended to limit the present invention, and various modifications and changes can be made to the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and scope of the present invention are intended to be included within the scope of the present invention.

Claims

A method for configuring a multi-cell precoding weight, comprising:

For each of the plurality of cells, obtaining the first weight of the user based on a maximum signal to noise ratio principle;

For each of the users, #> obtaining a game function of the user according to the policy set of the user and the income function of the user;

For each user, the first weight of the user is configured according to the first weight of the game function and other users in the plurality of cells that cause interference to the user, where the user is configured The first weight value makes the user's income function maximum, and when the income function satisfies the Nash balance, the first weight value configured is used as the precoding weight value of the user.

2. The method of claim 1, wherein for each of the users, obtaining the first weight of the user based on a maximum information leakage noise ratio principle comprises:

According to the definition of the letter leakage ratio:

And the minimum mean square error criterion is obtained:

((σ ² / + [ ... ― _P H, ₊₁ ... ^ ... ― _P H, ₊₁ — H ^^ Let H^ Hi-'.H '.H^ denote all interfering user channel matrices of the ith user Collection, get:

SLNR, = ~

^ ² +||H ,|| . Maximize SLNR based on the minimum mean square error criterion, yield: w, - arg max ~ ^WiHiHi ~, ~ = max EV((N,al + H'H, y ^[ H' H, ) where SLNRi represents the information leakage ratio of the i-th user, which is the channel matrix of the i-th user, ¹ ^ is the first weight of the i-th user, and ^ is the noise power of the i-th user, The channel matrix of the interfering users in the cell where the user is located and in the neighboring cell, K is the number of all interfering users, and ^ W represents the eigenvector decomposition of the matrix. The method according to claim 2, further comprising:

Obtain the channel matrix of the i-th user in advance by one of the following methods:

In the time division multiplexing system, each user's uplink channel is estimated by utilizing the reciprocity of the channel to obtain the channel matrix of each user equivalently;

In the frequency division duplex system, the channel information situation fed back by each user is obtained through limited feedback, and the channel information of each user and the channel information of the user of the user are subjected to feature vector decomposition to obtain the user. Channel matrix. The method according to claim 2, wherein the expression of the game function G of each user is:

(?„ , ,{", (·)}] , where N is the set of users with co-channel interference in each cell, and ^ i ¹ ' ² ''''' } (W is the benefit of the i-th user Function, for user i's policy collection, and, the table

The formula is as follows:

Where ^ ^ is the channel matrix of the i-th user, the first weight of the i-th user, ^σ ' ² is the noise power of the i-th user, K is the number of all interfering users, and ^w is the number of interfering users A weight. The method according to claim 4, wherein the first weight of the user is configured according to a game function of each user and a first weight of a user in the plurality of cells that causes interference to the user include: Bring the expression of ⁰ ' into the user's income function to get the user's income function and use

'■ Xw _i

The relationship between the first weight of the household is as follows: S + " , and derives from ^W ': du _ du] djSINR^

Dw, ~ diSINR,) dw _l

Relative to the ^w 'derivative, further solution:

And thus determine the maximum value of '', where, for the user's cost factor, "for steepness factor; configuration for ¹ ^, and in the configured w' such that the corresponding ^u- person ^S ' maximum satisfies the Nash equilibrium And configuring the first weight as the precoding weight of the user. The method according to any one of claims 1 to 5, characterized in that, for each user, a gain function satisfying the Nash balance means: , . | ≤ f , where 1^ is the i The revenue function of each user, indicating convergence accuracy. The method according to any one of claims 1 to 5, wherein the policy set includes at least one of the following: a signal to noise ratio, a signal to noise ratio, and a signal to interference and noise ratio.

A device for configuring a multi-cell precoding weight, comprising:

a first acquiring module, configured to obtain, for each user of the multiple cells, a first weight of the user based on a maximum information leakage ratio;

a second obtaining module, configured to obtain, by each user, a game function of the user according to the policy set of the user and the revenue function of the user;

a configuration module, configured to configure, for each user, a first weight of the user according to a game function of the user and a first weight of another user in the plurality of cells that causes interference to the user, where And when the first weight configured for the user maximizes a profit function of the user, and the benefit function satisfies a Nash balance, the configured first weight is used as a precoding weight of the user.

The device according to claim 8, wherein the processing by the first obtaining module for obtaining the first weight of the user for each user comprises:

The first obtaining module is defined according to a letter to noise ratio:

K-n and the minimum mean square error criterion are obtained: w _i ∞ EV((a ² IHH --H _i _ _l , H _M -..H _k ]' H _i -..H _i _ _l ,H _M -. .H _k ]r ^l H;H _i ). Let: ^...^,^...^ denote the set of all interfering user channel matrices of the i-th user, and get:

SLNR _; = ~

w, ² +||H ,.|| . Maximize SLNR based on the minimum mean square error criterion, resulting in:

^{^{_{= Ar g m,, ΛΓ '}}} 2' - -. = Max EVHN ^ I + H i Η ^ ι Η ί H,) wi eC w t IN i a i I + H i H i \ w i where, SLNRi represents The information leakage ratio of the i-th user is the channel matrix of the i-th user, ¹ ^ is the first weight of the first user, and is the noise power of the i-th user, ^H « is the intra-cell and the user The channel matrix of the interfering users in the neighboring cell, K is the number of all interfering users, and ^£ W represents the eigenvector decomposition of the matrix. The apparatus according to claim 9, wherein the expression of the game function G of each user is:

Where N is a set of users with the same frequency interference in each cell, and ^2^ ¹ ' ² '...'^; is the income function of the i-th user, & is the policy set of user i, and the table of

The formula is as follows:

Where A is the channel matrix of the i-th user, ^W ' is the first weight of the i-th user, ^σ ' ² is the noise power of the i-th user, K is the number of all interfering users, ^ is the interference user The first weight.