JP6474690B2 - Wireless communication system, wireless communication method, and wireless communication program - Google Patents

Description

  In a communication wireless network in which radio cells accommodating a plurality of radio stations are mixed and a plurality of small cells are mixed in a large cell having a large communication area, a large number of radio stations transmit at the same frequency and the same time. The present invention relates to a radio communication system, a radio communication method, and a radio communication program that use interference alignment (IA) technology and clustering technology for avoiding interference caused by radio signals.

  In conventional radio communication systems, radio resources are divided by frequency axis, time axis, space axis, or a combination thereof in order to avoid a decrease in throughput due to interference between radio signals transmitted by a large number of radio stations. However, interference with other radio stations is avoided by performing communication within the range of radio wave resources to which each radio station is assigned.

  On the other hand, in recent years, an interference alignment (IA) technique has been proposed as a new approach for interference control (Non-Patent Document 1). The IA technology does not divide radio wave resources that can be used by many radio stations as in the past, but allows interference that occurs between radio stations so that each radio station can use the entire radio wave resources. Recognize that the radio signal transmitted by each radio station instead of being divided is interference with other radio stations, and transmit so that a large number of interference signals received by individual radio stations are finally adjusted. The signal processing of the wireless station on the receiving side and the wireless station on the receiving side is devised. New ideas for interference control in IA technology are attracting attention from both academic and engineering fields, and their theoretical and application effectiveness in interference control is being proven.

  Communication scenarios to which the IA technology is applied are divided into a wide variety of cases, but those that target interference between radio stations (Non-patent Document 2), those that target interference between radio cells (Non-Patent Document 3), and larger There are devices (Non-Patent Documents 4 and 5) that target interference between heterogeneous system communication networks in which cells and small cells coexist.

  On the other hand, the amount of required processing and the amount of shared information increase in proportion to the increase in the number of radio stations or radio cells to which interference control by IA technology is applied. Therefore, in order to control interference generated between radio stations or between radio cells using IA technology in practice, the radio station or the entire radio cell to be applied is divided into several clusters, and radios accommodated by individual clusters are accommodated. It is necessary to control the required processing amount and shared information amount of IA technology by the number of stations or the number of wireless cells (Non-patent Document 6).

V. Cadambe and S. Jafar, “Interference alignment and degrees of freedom of the K-user interference channel,” IEEE Trans. Infomation Theory, vol.54, no.8, pp.3425-3441, Aug. 2008 K. Gomadam, V. R. Cadambe and S. A. Jafar, “A Distributed Numerical Approach to Interference Alignment and Applications to Wireless Interference Networks,” IEEE Trans. Inf. Theory, vol.57, no.6, pp.3309-3322, June 2011 C. Suh, M. Ho, and D. Tse, “Downlink interference alignment,” IEEE Trans. Commun., Vol.59, no.9, pp.2616-2626, Sep. 2011 W. Shin, W. Noh, K. Jang, and H.-H. Choi, “Hierarchical interference alignment for downlink heterogeneous networks,” IEEE Trans. Wireless Communication, vol.11, no.12, pp.4549-4559, Dec. 2012 Q. Niu, Z. Zeng, T. Zhang, Q. Gao, and S. Sun, “Joint Interference Alignment and Power Allocation in Heterogeneous Networks,” IEEE International Symposium on Personal, Indoor and Mobile Radio Communications (PIMRC), pp. 811-815, Sep. 2014 Sujie Chen, Roger S. Cheng, "Clustering for Interference Alignment in Multiuser Interference Networks", IEEE Transaction on Vehicular Technology, vol.63, No.6, July 2014

  However, with the conventional clustering technology that considers the application of the IA technology proposed in Non-Patent Document 6, all information such as the number of clusters, cluster size, and cluster elements included in the cluster pattern can be obtained from an unknown state. Since the entire cluster pattern is searched and an appropriate cluster pattern is selected, the required calculation amount, shared information amount, and hardware size increase in proportion to the number of radio stations to be clustered.

  Also, with the conventional clustering technology, it is not possible to control the number of all clusters included in a cluster pattern or the number of unit elements (cluster size) that can be accommodated by each cluster. However, if it deviates from the configuration of the actual system, its practical use becomes difficult.

  Furthermore, the conventional clustering technique is a clustering technique limited to a radio communication system in which a large number of single radio stations are mixed, and cannot be applied to clustering using radio cells including a plurality of radio stations as unit elements.

  In the present invention, a radio cell including a single or a plurality of radio stations is also a unit element, and not only a global search but also a local search in each cluster can be performed to determine a cluster pattern. An object of the present invention is to provide a wireless communication system, a wireless communication method, and a wireless communication program that realize clustering that enables control.

1st invention is the radio | wireless communications system which reduces the interference between cells which arise when there are several cells which a base station and a terminal station communicate, and each cell performs radio | wireless communication at the same frequency and the same time mutually, The transmission rate loss matrix is constructed on the basis of the propagation path information, and the cluster number N and individual clusters are selected based on the transmission rate loss matrix when selecting the cluster pattern subject to interference control in units of cells. after determining the cluster size is the number of cells included, e Bei cluster determining means for determining a cell included in each cluster, the cluster determining means from all the cluster pattern that satisfies the cluster number N and the cluster size When selecting a cluster pattern that maximizes the total amount of rate loss due to internal interference of all clusters, In order of the first to N−1 clusters with the largest N, local maximization processing is performed N−1 times for maximizing the total amount of rate loss due to internal interference for each cluster, and rate loss due to internal interference is performed for each cluster. In this configuration, the cell having the maximum total amount is determined.

The second invention is, in a wireless communication system cell base station and a terminal station communicates there is a plurality, for reducing inter-cell interference caused when each cell performs radio communication at the same frequency and the same time to each other, each cell The transmission rate loss matrix is constructed on the basis of the propagation path information, and the cluster number N and individual clusters are selected based on the transmission rate loss matrix when selecting the cluster pattern subject to interference control in units of cells. After determining the cluster size, which is the number of included cells, a cluster determining means for determining the cells included in each cluster, the cluster determining means, from among all cluster patterns satisfying the number of clusters N and the cluster size, when the total amount of rate loss due to internal interference of all clusters to select a cluster pattern that maximizes the cluster Sai Sequentially from the first large clusters of up to N-1 th, the Tsuboneiki maximization process to maximize the total amount of rate loss due to internal interference performs N-1 times for each cluster, the first cell in each cluster Determines the cell having the maximum amount of mutual interference with other cells among the candidate cells, and the second and subsequent cells determine the cell having the maximum amount of mutual interference with the already determined cell, and determines the cell. In this configuration, the order of each cell in the corresponding cluster is reflected in the transmission rate loss matrix to determine the cells included in each cluster.

According to a third aspect of the present invention, there is provided a radio communication method for reducing inter-cell interference that occurs when a plurality of cells communicate with a base station and a terminal station and each cell performs radio communication at the same frequency and at the same time. The transmission rate loss matrix is constructed on the basis of the propagation path information, and the cluster number N and individual clusters are selected based on the transmission rate loss matrix when selecting the cluster pattern subject to interference control in units of cells. after determining the cluster size is the number of cells included, have a cluster determination procedure for determining a cell included in each cluster, the cluster determination procedure from all the cluster pattern that satisfies the cluster number N and the cluster size When selecting a cluster pattern that maximizes the total amount of rate loss due to internal interference of all clusters, In order from the first cluster to the (N-1) th cluster, local maximization processing is performed N-1 times for maximizing the total amount of rate loss due to internal interference for each cluster. The cell with the maximum total amount is determined.

A fourth invention, cell base station and a terminal station communicates there is a plurality, in a wireless communication method for reducing inter-cell interference caused when each cell performs radio communication at the same frequency and the same time to each other, each cell The transmission rate loss matrix is constructed on the basis of the propagation path information, and the cluster number N and individual clusters are selected based on the transmission rate loss matrix when selecting the cluster pattern subject to interference control in units of cells. After determining the cluster size, which is the number of cells included, a cluster determination procedure for determining the cells included in each cluster, the cluster determination procedure is selected from among all cluster patterns satisfying the number of clusters N and the cluster size. when the total amount of rate loss due to internal interference of all clusters to select the cluster pattern having the maximum cluster size Order of a cluster from the first listening to (N-1) -th, the Tsuboneiki maximization process to maximize the total amount of rate loss due to internal interference performs N-1 times for each cluster, the first cell in each cluster Among the candidate cells, the cell having the maximum amount of mutual interference with other cells is determined, the second and subsequent cells are determined to determine the cell having the maximum amount of mutual interference with the already determined cell, and for each cell determination. The cells included in each cluster are determined by reflecting the order of the cells in the corresponding cluster in the transmission rate loss matrix.

A wireless communication program of a fifth invention causes a computer to execute a cluster determination procedure of the wireless communication method of the third or fourth invention, and determines a cell included in each cluster.

  Since the present invention constructs a clustering method using not only wireless stations but also wireless cells including single or multiple wireless stations as unit elements, not only communication networks in which wireless stations are mixed but also wireless cells are mixed. It is also possible to apply to communication networks, and the communication scenarios to which clustering technology is applied can be greatly expanded.

  Since the present invention proposes not only a global search but also a local search to determine the cluster pattern, the merit of the characteristics of the global search, the amount of calculation by the local search, the processing delay, the amount of shared information, the hardware size, etc. The trade-off with the merits of the clustering technology can be achieved, and the practicality of the clustering technology can be greatly enhanced.

  In the present invention, the number of clusters and the size of each cluster can be controlled in consideration of practical use, so that it becomes possible to determine a cluster pattern suitable for the configuration of the actual system, and greatly enhance the practicality of the clustering technique. be able to.

It is a figure which shows the structural example 1 of the radio | wireless communications system of this invention. It is a figure which shows the structural example 2 of the radio | wireless communications system of this invention. 3 is a flowchart illustrating a processing procedure according to the first embodiment. 10 is a flowchart illustrating a processing procedure according to the second embodiment. 10 is a flowchart illustrating a processing procedure according to the third embodiment.

FIG. 1 shows a configuration example 1 of a wireless communication system of the present invention.
In FIG. 1, cell 1 to cell 8 are each composed of a base station 11 and at least one terminal station 12 communicating with the base station 11. When cells communicate with each other at the same frequency and at the same time, inter-cell interference occurs between cells, and interference control is necessary. In order to reduce the amount of calculation processing and the amount of shared information necessary for the interference control, clustering processing of each cell is performed using the technique of the present invention.

  In the clustering process in the configuration example 1, the base stations 11 of the cells 1 to 8 collect propagation path information and the like necessary for interference control in the central processing unit 100 via a wired or wireless path, and the central processing unit 100 Based on the information, a cluster pattern in which several cells are bundled is determined and notified to the base stations 11 of the cells 1 to 8. Here, individual clusters are indicated by dotted lines. The central processing unit 100 can be installed inside some or all of the cell base stations. In that case, some of the paths in FIG. 1 can be omitted.

FIG. 2 shows a configuration example 2 of the wireless communication system of the present invention.
In FIG. 2, each of the small cell 1 to the small cell 8 includes a small cell base station 21 and at least one small cell terminal station 22 that communicates with the small cell base station 21. The large cell has a large area including the small cell 1 to the small cell 8 and includes a large cell base station 31 and at least one large cell terminal station 32 communicating with the large cell base station 31. When each small cell and large cell communicate with each other at the same frequency and at the same time, inter-cell interference occurs between the cells, so that interference control is necessary. In order to reduce the amount of calculation processing and the amount of shared information necessary for the interference control, clustering processing of each cell is performed using the technique of the present invention.

  In the clustering process in Configuration Example 2, each small cell base station 21 and large cell base station 31 collects propagation path information and the like necessary for interference control in the central processing unit 100 via a wired or wireless path, and the central processing unit 100 Determines a cluster pattern in which several small cells are bundled based on the information, and notifies the small cell base station 21 and the large cell base station 31 of the small cells 1 to 8. Here, individual clusters are indicated by dotted lines. The central processing unit 100 can be installed inside some or all of the cell base stations. In that case, some of the paths in FIG. 2 can be omitted.

  Hereinafter, the clustering process of the present invention will be described using communication in an interference environment in a communication network having a total of B cells as in the configuration example shown in FIG.

First, the definition of mathematical symbols in the following equations is shown.
A (a: b, c: d) means the elements of rows A to b and columns c to d of the matrix A.
A ^T means transposition of the matrix A.
A ^* means the conjugate transpose of the matrix A.
det (A) means a determinant of the matrix A.
‖A‖ ² refers to the norm of the matrix A.
su (A) means the total value of all elements included in the matrix A.
diag (x _1, x _2, ..., x _N) is x _1, x _2, ..., means a diagonal matrix with the x _N diagonal elements.
_{blockdiag (x 1, x 2,} ..., x M) are x _1, x _2, ..., it means a block diagonal matrix with the x _M block diagonal.

Each cell has one base station and L terminal stations. Each base station has _NT communication degrees of freedom. That is, the total of the number of spatial antennas T _S , the number of time slots T _T, and the number of frequency subcarriers T _F is N _T = T _S · T _T · T _F. Similarly, each terminal station has N _R communication degrees of freedom. That is, the total of the number of spatial antennas R _S , the number of time slots R _T, and the number of frequency subcarriers R _F is N _R = R _S · R _T · R _F.

  However, in the case where the number L of terminals included in each cell is 1, the above communication configuration is the same as the configuration in an interference environment between B radio stations and B radio stations. Therefore, the present invention is applicable in both communication configurations.

Hereinafter, in order to simplify the notation, it is assumed that each cell has the same number of terminals L, each base station has the same degree of freedom of communication _NT , and each terminal station has the same degree of freedom of communication N _R. . However, the present invention can also be applied to a communication network in which each base station and each terminal station have different degrees of freedom of communication when each cell has a different number of terminals. In that case, the number of terminals in each cell, the degree of freedom of communication of each base station, and the degree of freedom of communication of each terminal station may be defined for each cell as L _b , N _{b, T} , N _{b, R.}

The reception signal y in each cell includes a transmission signal s = [s ^T ₁ ... S ^T _B ] ^T , interference components n = [n ^T ₁ ... N ^T _B ] ^T due to noise and signal sources other than the transmission signal, and a total of B A certain cell is expressed as follows using a channel response matrix ρ including a channel response coefficient between each cell-to-cell.

The propagation path response matrix ρ is expressed as follows. Here, ρ ^l _{b, a} (i, j) is a channel response coefficient between the j-th transmission degree of freedom of the base station of cell a and the i-th degree of freedom of reception of terminal l of cell b. Show. However, a = 1, ..., B , b = 1, ..., B, l = 1, ..., L, i = 1, ..., N R, j = 1, ..., and N _T.

Other than the above, if the values of the channel response coefficients for all N _T degrees of freedom of each base station are the same or close, and are the values of the channel response coefficients of all the N _R degrees of freedom for each terminal also the same? In the near case, the channel response matrix can be simplified as follows. Here, the channel response coefficient does not depend on individual degrees of freedom of communication of the base station and the terminal, but depends only on the transmitting cell, the receiving cell, and the receiving terminal.

  Here, the propagation path response may be a response by only the physical propagation path or a response by an effective propagation path in which the physical propagation path and the signal processing on the transmission side and the signal processing on the reception side are combined. There are various definitions of the channel response here. For example, from the viewpoint of distance, there is a propagation path loss in the wide area between the transmission side and the reception side or fading attenuation in the local area. Further, there is a long-term average response or a short-term instantaneous response between the transmission side and the reception side from the viewpoint of time. The present invention can be implemented based on a channel response matrix obtained by an appropriate definition, and the scope of the present invention is not limited by these definitions.

  As shown in FIG. 1 and FIG. 2, clustering in the present invention is a method in which several cells are bundled to form individual clusters, so that one interference communication network possessed by many cells is reduced to a small number for each cluster. This is a process of disassembling into a plurality of interference communication networks including only cells (reduced in scale).

  As shown in FIGS. 1 and 2, clustering is also a process of determining a specific cluster pattern from among many cluster patterns. In the example of FIGS. 1 and 2, a total of four clusters are formed, and each cluster includes three, two, three, and one element (cell). In this configuration, each cluster includes {1, 3, 6}, {2, 5}, {4, 7}, and {8} -th cells.

  As described above, an arbitrary cluster pattern includes two types of information, that is, a cluster configuration and cluster elements. The cluster configuration is the number of all clusters (hereinafter referred to as “number of clusters”) and the number of elements (corresponding to cells in the present invention) included in each cluster (hereinafter referred to as “cluster size”). Information). A cluster element is information on an element included in each cluster. In order to determine the cluster pattern, the cluster configuration and cluster elements need only be determined.

Organized by mathematical formulas, C means a specific cluster pattern including information such as cluster configuration and cluster elements. Ω represents a set of all cluster patterns C. N means the number of clusters in the cluster pattern C.

C ⁿ denotes the n-th cluster, | C ⁿ | is the number of elements included in C ^n, i.e. means cluster size of C ^n. b _m ⁿ represents the m-th element belonging to the cluster C ⁿ . However, for simplification of the display in the following, unless it is found that the b _m ⁿ before and after the context belonging to the cluster _{^{C n, b m n = b}} m (m = 1,2, ..., | C n | ).

  The goal of clustering is to reduce the required amount of signal processing, the amount of shared information, the hardware scale, and the like, and to minimize inter-cluster interference caused by clustering, compared to when not clustering.

For the clustering process, here, based on the channel coefficient matrix, when a certain cell b (b = 1, 2,..., B) is not subject to interference from other cells, The communication transmission rate _Rb is defined as follows.

P _b represents a transmission power control matrix including transmission power limit values P _b ¹ ,..., P _b ^NT in each communication degree of cell b. In the formula, P is written in bold and P is written in thin. P _b is the sum of transmission power limit values P _b ¹ ,..., P _b ^NT in each communication degree of cell b, and is assumed to be equal to or less than the transmission power upper limit value P _T.

On the other hand, when there is interference from another cell a (a ≠ b, a = 1, 2,..., B) where the cell b is present, the communication transmission rate of the cell b is defined as follows.

Next, a transmission rate loss Δ _{b, a} caused by receiving interference from the cell a when the cell b communicates is defined as follows. Note that the transmission rate R _b , R _{b, a} in the cell b is expressed as follows using the equation (6).

Similarly, for the cell a which is the interference source of the cell b, when the cell a does not receive interference from other cells, the communication transmission rate of the cell a is defined as follows.

P _a represents a transmission power control matrix including transmission power limit values P _a ¹ ,..., P _a ^NT in each communication degree of the cell a. P _a transmission in each communication flexibility of the cell b power limit value P _a ^1, ..., the sum of P _a ^NT, assume that equal to or less than maximum transmission power P _T.

On the other hand, when an interference from another cell b (b ≠ a) is received by a cell a, the communication transmission rate of the cell a is defined as follows.

Next, the transmission rate loss Δ _{a, b} caused by receiving interference from the cell b when the cell a communicates is defined as follows. Note that the transmission rates R _a and R _{a, b} in the cell a are expressed as follows using the equation (6).

  However, the definition of the transmission rate and the definition of the transmission rate loss due to interference can take various forms, but the method of the present invention can be implemented based on the transmission rate loss obtained by an appropriate definition. The scope of the method of the present invention is not limited by these definitions.

Example 1
In the first embodiment, a rate loss matrix Δ is constructed, and based on the rate loss matrix Δ, collective cluster determination is performed by maximizing the entire area for all clusters.

Here, in all B cells, the following rate loss matrix Δ can be constructed based on the transmission rate loss (due to interference) between the two cells as described above. A row index of Δ represents a cell to be communicated, and a column index of Δ represents a cell causing interference. Here, Δ _{b, b} = 0 (b = 1, 2,..., B). Since it is assumed that each cell does not interfere with its own communication, there is no transmission rate loss due to its own interference.

Next, the mutual rate loss w (a, b), which is an index representing the amount of mutual interference between the cell a and the cell b, is defined as follows.

w (a, b) represents the total transmission rate loss when cell a and cell b are assigned to different clusters. When determining the elements of each cluster, w (a, b) is an important judgment index. Further, the value of w (a, b) depends only on the matrix components related to cell a and cell b, and these matrix components are located symmetrically on the diagonal of Δ. Paying attention to the structural relationship between w (a, b) and matrix Δ, the total value su (Δ) of all components of Δ is along the diagonal of Δ as shown in the following equation (24). In addition, w (a, b) can be expressed as an accumulation from a = 1 to b-1.

Next, when a specific cluster pattern C in the rate loss matrix Δ is reflected, the following matrix Δ (C) is obtained.

Δ _D (C) is a partial matrix of Δ (C) of the block diagonal structure, and the number of diagonal blocks included in Δ _D (C) is the number N of clusters. Further, the n-th diagonal blocks delta _D (C ⁿ⁾ of delta _D (C) corresponds to the n-th cluster C ^n, the number of delta _D cells associated with rate loss value included in (C ⁿ⁾ Is the number of elements included in cluster C ⁿ (cluster size) | C ⁿ |.

Thus, a specific cluster pattern C can be completely associated with the block diagonal sub-matrix Δ _D (C) of the rate loss matrix Δ (C) reflecting the pattern. In order to reflect C in Δ, the rows and columns of Δ may be rearranged based on C. Therefore, the optimization of the clustering process, that is, the process of determining the optimal cluster pattern can be performed based on the rate loss matrix Δ.

On the other hand, delta _D ^- (C) is the submatrix of the delta from (C) Δ _D (C) removal of the delta (C) (× moiety of formula), is a co-matrix of delta _D (C).

A component included in Δ _D (C) indicates a transmission rate loss due to intra-cluster interference (hereinafter referred to as “internal interference rate loss”). Interference control techniques such as interference alignment can be applied within individual clusters, and as a result, internal interference rate loss can be greatly reduced.

On the other hand, delta _D ^- component contained in (C) is the transmission rate loss due to inter-cluster interference (. Hereinafter referred to as "external interference Rate Loss") indicates, not sufficiently control the interference countermeasure in the individual clusters. Therefore, in the clustering process as described above, it is desired to suppress the inter-cluster interference caused by clustering and minimize the external interference rate loss resulting therefrom.

Attention should be paid to the following relationship.

  su (Δ) represents the total amount of transmission rate loss due to interference received from other cells when all B cells simultaneously communicate. Although the matrix Δ (C) and the component (rate loss value) included in the matrix Δ have different positions in the matrix depending on the rearrangement for reflecting the cluster pattern C, the value of the rate loss value itself is exactly the same. Therefore, the total amount su (Δ (C)) of the rate loss due to the sum of these components is a fixed value that does not depend on the clustering process.

On the other _{hand, su (Δ D (C)} ) represents the total amount of the internal interference rate ^{_{loss, su (Δ D - (C}} )) represents the total amount of external interference rate loss. For su (Δ (C)) is a fixed value, su maximize (Δ _D (C)) and su ^- The minimization of (Δ _D (C)), is equivalent as follows.

Therefore, in the following, maximization of su (Δ _D (C)) is used as a clustering criterion to minimize external interference rate loss.

Further, in the optimization of clustering of the present invention, that is, the determination of the optimal cluster pattern, the cluster configuration and the cluster elements are determined separately. First, the number of clusters and the cluster size corresponding to the cluster configuration are determined, and then suitable cluster elements are searched based on the determined cluster configuration to find an optimal cluster pattern. Therefore, the maximization of su (Δ _D (C)) in equation (29) is further rewritten as follows.

Here, the number N of clusters corresponding to the cluster configuration and the individual cluster sizes | C ¹ |,..., | C ^N | are all determined in advance according to the configuration of the actual system, Through the global maximization problem with respect to su (Δ _D (C)) corresponding to the cluster, element search is performed for all the clusters at once.

FIG. 3 shows a processing procedure of the first embodiment.
First, a rate loss matrix, a list of candidate elements, the number of clusters N, and cluster sizes | C ¹ | to | C ^N | are determined in advance. Next, based on the information, batch cluster determination is performed by maximizing the entire area for all clusters. That is, in the first cluster C ¹ N th cluster C ^N, the number of clusters and from all the cluster pattern that satisfies the cluster size, the cluster pattern that the total amount is the maximum rate loss due to internal interference of all clusters select. Finally, the determined cluster pattern is output and the clustering process is terminated.

  The square shown on the right side of FIG. 3 shows an operation image of the clustering process based on the rate loss matrix. By predetermining the number of clusters and the cluster size, perform processing corresponding to Equation (30) or Equation (31), and search all cluster patterns that satisfy the given number of clusters and cluster size. It can be seen that element determination of all clusters is performed.

(Example 2)
In the second embodiment, sequential cluster determination is performed by local maximum for each cluster.
In the first embodiment, a rate loss matrix is constructed, and the clustering optimization problem is determined based on the cluster configuration in advance, and the global maximization problem regarding su (Δ _D (C)) corresponding to all clusters ( In other words, it is simplified to the collective element search problem of all clusters. Furthermore, Equation (31) is rewritten as follows.

In the equation (32), one-time maximization for su (Δ _D (C)) corresponding to all clusters is performed by su (Δ _D (C ⁿ ) corresponding to each cluster (from the first to the Nth). ) For N local maxima. That is, instead of determining the elements of all N clusters (corresponding to N diagonal blocks of Δ _D (C)) in a lump, each cluster (Δ _D (C) in each diagonal block is sequentially determined. Will be determined).

Here, if clusters C ¹ to C ^N-1 are determined, the elements (cells) that enter the last cluster C ^N are uniquely determined from all the remaining candidate elements that have not yet been selected, so there is room for optimization. Disappears. Therefore, an equivalent and office area maximization process of each cluster sequentially from the cluster C ¹ to cluster C ^N-1, and sequentially station area maximization process of each cluster from the cluster C ¹ to cluster C ^N, the formula ( As shown in equation (33), the number of times of local maximization was reduced from N times to N-1 times. By such approximation, the required search amount can be greatly reduced.

  In the clustering process, the number of candidate elements that can be selected is large and the degree of freedom of optimization is high at an early stage, but the number of candidate elements that can be selected is low and the degree of freedom of optimization is low. Therefore, for a cluster having a large number of elements (cluster size), it is desired to perform element optimization at a stage where the degree of freedom is as high as possible. A cluster having a large number of elements is also considered to have a relatively large amount of interference with other clusters. In the second embodiment in which element optimization is performed sequentially on individual clusters based on these ideas, element optimization is performed preferentially on clusters having a large cluster size.

FIG. 4 shows a processing procedure of the second embodiment.
First, as an input, a rate loss matrix, a list of candidate elements, the number of clusters N, and a large to small cluster size | C ¹ | to | C ^N | are determined in advance. Next, based on the information, sequential cluster determination is performed by local maximization for each cluster from the cluster C ¹ to the cluster C ^{N-1 in} descending order of the cluster size. That is, in the n-th cluster C ⁿ in descending order of the cluster size, the total amount of rate loss due to internal interference of the cluster C ⁿ becomes the maximum among all element patterns including | C ⁿ | candidate elements. Select an element pattern. Further, after the determination of the cluster C ^N−1 , all of the remaining candidate elements are set as elements of the Nth cluster C ^N. Finally, the determined cluster pattern is output and the clustering process is terminated.

  The square shown on the right side of FIG. 4 shows an operation image of the clustering process based on the rate loss matrix. It can be seen that sequential processing is performed for each cluster in descending order of cluster size.

(Example 3)
In the third embodiment, sequential cluster element determination is performed by further local maximization for each element of the cluster.

In the second embodiment, the clustering optimization problem is first simplified to the global maximization problem for su (Δ _D (C)) (that is, the batch element search problem for all clusters) after the cluster configuration is determined in advance. . _{Next, su (Δ D (C)} ) about once more the entire maximization problem ^{_{su (Δ D (C n)}} ) N-1 times the station area maximization issues (i.e. successive elements search problem of each cluster ).

Furthermore, T ⁿ is defined, and the local area maximization in each diagonal block Δ _D (C ⁿ ) included in the equation (33) is rewritten as follows.

Equation (36) can be easily obtained from Equation (24). Further, since the diagonal component of Δ is 0, Expression (37) and Expression (38) are equivalent. Accordingly, the determination of the elements from the second to the | C ⁿ | th of each cluster can be sequentially optimized by the above method.

In equation (38), the su (delta _D (C ⁿ⁾⁾ regarding single station area maximization corresponding to a cluster C ^n, the C ⁿ (the second to | C ⁿ | until th) to each element corresponding cross rate loss total amount Σw (a, b _m) (accumulator from a = T ^n-1 +1 to ^{T n-1 + m-1} )
Is approximated by | C ⁿ | −1 further local maximization. That is, instead of determining all the elements of the cluster C ⁿ at once, the elements from the second to the | C ⁿ | th of the cluster C ⁿ are sequentially determined.

Specifically, first, the element having the largest mutual rate loss with the first element in the cluster is set as the second element. Next, the element with the largest sum of the mutual rate loss with the first element and the second element is set as the third element. Thus analogy to the last | C ⁿ | th element from the first element | C ⁿ | -1-th element to the sum of the cross-rate loss largest ones of | C ⁿ | th Element. This further local approximation can further reduce the required search amount.

However, as can be seen from the equation (38), the above method can be applied to the determination of each element from the second to the | C ⁿ | th of the cluster C ⁿ , but not the determination of the first element. Therefore, the following concept is applied to the determination of the first element of each cluster.

As described above, the optimization of the cluster elements of the present invention is realized through maximization of the total rate loss su (Δ _D (C)). Therefore, if an element having a large mutual rate loss with other elements is preferentially selected and the remaining elements in the cluster are sequentially determined from that element, su (Δ _D (C)) It can contribute to maximization.

In line with that idea, in equation (39), in each cluster, among the remaining candidate elements not yet selected, the total amount of mutual rate loss between itself and all other remaining candidate elements Σw (a , b _m ) (a = accumulation from T ^n-1 +1 to B)
Is selected as the first element.

  In addition to the above method, one element may be arbitrarily selected from the remaining candidate elements that have not been selected without optimizing the first element of each cluster.

As described above, when the elements included in the cluster C ⁿ are sequentially determined, if a certain element b _m is determined, it is related to the selected element so that the determination of the next element does not cause confusion. The rows and columns of Δ are rearranged so that the rows and columns of Δ to become the b _m- th rows and columns of Δ.

  In the clustering process, the number of candidate elements that can be selected is large and the degree of freedom of optimization is high at an early stage, but the number of candidate elements that can be selected is low and the degree of freedom of optimization is low. Therefore, for a cluster having a large number of elements (cluster size), it is desired to perform element optimization at a stage where the degree of freedom is as high as possible. A cluster having a large number of elements is also considered to have a relatively large amount of interference with other clusters. In the second embodiment in which element optimization is performed sequentially on individual clusters based on these ideas, element optimization is performed on a cluster having a large cluster size preferentially.

FIG. 5 shows a processing procedure of the third embodiment.
First, as an input, a rate loss matrix, a list of candidate elements, the number of clusters N, and a large to small cluster size | C ¹ | to | C ^N | are determined in advance. Next, based on the information, sequential cluster determination is performed by local maximization for each cluster from the cluster C ¹ to the cluster C ^{N-1 in} descending order of the cluster size. Here, in the determination of each cluster, sequential cluster element determination is performed by further local maximum for each element from the first to the | C ⁿ | th.

That is, in the n-th cluster C ⁿ in descending order of the cluster size, the one having the maximum total amount of mutual rate loss with the remaining candidate elements of its own / others is the first among the remaining candidate elements. Select as an element. Further, the element having the maximum total amount of the mutual rate loss with the already determined first to m−1 elements is selected as the mth element. The rows and columns are such that the row and column of the rate loss matrix Δ associated with the selected element is the mth row and column of the rate loss matrix so that the order is reflected each time a cluster element is determined. Sort the. Further, after the determination of the cluster C ^N−1 , all of the remaining candidate elements are set as elements of the Nth cluster C ^N. Finally, the determined cluster pattern is output and the clustering process is terminated.

  The square on the right side of FIG. 5 shows an operation image of the clustering process based on the rate loss matrix. It can be seen that sequential processing is performed for each cluster and each element.

In the wireless communication system of the present invention, the clustering process based on the rate loss matrix, the list of candidate elements, the number of clusters N, and the large to small cluster sizes | C ¹ | to | C ^N | It can be realized by a program. This computer program can be stored in a computer-readable storage medium or provided via a network.

1 to 8 cells 11 base stations 12 terminal stations 21 small cell base stations 22 small cell terminal stations 31 large cell base stations 32 large cell terminal stations

Claims (5)

In a wireless communication system that reduces the inter-cell interference that occurs when there are multiple cells with which a base station and a terminal station communicate and each cell performs wireless communication at the same frequency and the same time,
When constructing a transmission rate loss matrix based on the propagation path information of each cell, and selecting a cluster pattern subject to interference control in units of the cell based on the transmission rate loss matrix, the number of clusters N and after determining the cluster size is the number of cells included in each cluster, Bei example cluster determining means for determining a cell included in each cluster,
The cluster determination means selects the cluster size from among all the cluster patterns satisfying the number of clusters N and the cluster size when the total amount of rate loss due to internal interference of all the clusters is maximized. In order of the first to N−1 clusters with the largest N, local maximization processing is performed N−1 times for maximizing the total amount of rate loss due to internal interference for each cluster, and rate loss due to internal interference is performed for each cluster. A wireless communication system, characterized in that the cell having the maximum total amount is determined . In a wireless communication system that reduces the inter-cell interference that occurs when there are multiple cells with which a base station and a terminal station communicate and each cell performs wireless communication at the same frequency and the same time ,
When constructing a transmission rate loss matrix based on the propagation path information of each cell, and selecting a cluster pattern subject to interference control in units of the cell based on the transmission rate loss matrix, the number of clusters N and After determining the cluster size, which is the number of cells included in each cluster, the cluster determining means for determining the cells included in each cluster,
The cluster determination means selects the cluster size from among all the cluster patterns satisfying the number of clusters N and the cluster size when the total amount of rate loss due to internal interference of all the clusters is maximized. sequentially from the first large clusters of up to N-1 th, the Tsuboneiki maximization process to maximize the total amount of rate loss due to internal interference performs N-1 times for each cluster, the first cell in each cluster Determines the cell having the maximum amount of mutual interference with other cells among the candidate cells, and the second and subsequent cells determine the cell having the maximum amount of mutual interference with the already determined cell, and determines the cell. This is characterized in that the cell included in each cluster is determined by reflecting the order of each cell in the corresponding cluster in the transmission rate loss matrix. Wireless communication system. In a wireless communication method for reducing inter-cell interference that occurs when there are multiple cells with which a base station and a terminal station communicate, and each cell performs wireless communication at the same frequency and the same time,
When constructing a transmission rate loss matrix based on the propagation path information of each cell, and selecting a cluster pattern subject to interference control in units of the cell based on the transmission rate loss matrix, the number of clusters N and after determining the cluster size is the number of cells included in each cluster, have a cluster determination procedure for determining a cell included in each cluster,
When the cluster determination procedure selects a cluster pattern that maximizes the total amount of rate loss due to internal interference of all the clusters from among all the cluster patterns that satisfy the number N of clusters and the cluster size, the cluster size In order of the first to N−1 clusters with the largest N, local maximization processing is performed N−1 times for maximizing the total amount of rate loss due to internal interference for each cluster, and rate loss due to internal interference is performed for each cluster. A wireless communication method characterized by determining a cell that maximizes the total amount . In a wireless communication method for reducing inter-cell interference that occurs when there are multiple cells with which a base station and a terminal station communicate, and each cell performs wireless communication at the same frequency and the same time,
When constructing a transmission rate loss matrix based on the propagation path information of each cell, and selecting a cluster pattern subject to interference control in units of the cell based on the transmission rate loss matrix, the number of clusters N and Having a cluster determination procedure for determining the cells included in each cluster after determining the cluster size, which is the number of cells included in each cluster;
When the cluster determination procedure selects a cluster pattern that maximizes the total amount of rate loss due to internal interference of all the clusters from among all the cluster patterns that satisfy the number N of clusters and the cluster size, the cluster size sequentially from the first large clusters of up to N-1 th, the Tsuboneiki maximization process to maximize the total amount of rate loss due to internal interference performs N-1 times for each cluster, the first cell in each cluster Determines the cell having the maximum amount of mutual interference with other cells among the candidate cells, and the second and subsequent cells determine the cell having the maximum amount of mutual interference with the already determined cell, and determines the cell. The wireless communication is characterized in that the cell included in each cluster is determined by reflecting the order of each cell in the corresponding cluster in the transmission rate loss matrix. Method. A wireless communication program that causes a computer to execute a cluster determination procedure of the wireless communication method according to claim 3 or 4 to determine a cell included in each cluster.

Images