CN108112047B

CN108112047B - User-centered semi-dynamic clustering method

Info

Publication number: CN108112047B
Application number: CN201810052835.XA
Authority: CN
Inventors: 李兴旺; 许春梅; 陈慧; 田心记
Original assignee: Henan University of Technology
Current assignee: Henan University of Technology
Priority date: 2018-01-19
Filing date: 2018-01-19
Publication date: 2020-12-11
Anticipated expiration: 2038-01-19
Also published as: CN108112047A

Abstract

The invention discloses a semi-dynamic clustering method taking users as centers, which is suitable for a super-dense network adopting JT-CoMP. Grouping of small base stations based on interference patternsKClustering and selecting cluster head of each cluster, usingQ _kIs shown askThe number of the clusters is small,k=1,2,…,Kfinding the interference patternQ _kOutside and withQ _kSmall base stations with small base stations interfering with each other are arranged inP _kIn the process, theQ _kSmall base station in andP _kthe small base station in the middle is taken as aQ _kAnd the cluster head firstly adds one of the service base stations to be selected to the CoMP service base station cluster of the user, and then determines whether to add other service base stations to be selected to the CoMP service base station cluster of the user according to the throughput of the system.

Description

User-centered semi-dynamic clustering method

Technical Field

The invention relates to the field of communication, in particular to a user-centered semi-dynamic clustering method in an ultra-dense network adopting JT-CoMP.

Background

An Ultra Dense Network (UDN) deploys hundreds of small cells in the coverage area of a macro cell, and there are hundreds of small base stations, and frequency bands or sub-channels between the small base stations are reused, so that there is severe interference between the small base stations. The adoption of Coordinated multipoint transmission (CoMP) in the UDN can reduce the interference between small base stations.

There are three main implementations of CoMP: coordinated scheduling/coordinated beamforming, dynamic Transmission point selection, and Joint Transmission (JT). In the JT transmission mode, multiple small base stations transmit data for the same user on the same time-frequency resource block, and convert interference signals between small cells into useful signals, which is one of the three implementation modes with the best reliability. In JT-CoMP, small base stations exchange channel state information and user data, and if all small base stations cooperate with each other, this requires too high backhaul link overhead, and for this reason, the selection of a cooperation cluster is very important.

The clustering method in JT-CoMP is divided into three types: static clustering, semi-dynamic clustering, and dynamic clustering. The static clustering is carried out according to the topological structure of the network, and can not change along with the change of the small base station or the position of a user, so that the complexity is lower, and the signaling overhead is less. Semi-dynamic clustering is provided on the basis of static clustering, small base stations are clustered at first, and then a plurality of service small base stations are dynamically selected for each user in each base station cluster. Semi-dynamic clustering improves system performance compared to static clustering, with the disadvantage of increased complexity and signaling overhead. The dynamic clustering is the best performance among the three clustering methods, is also the highest in complexity, and can change along with the change of the small base station or the user position. The dynamic clustering is divided into network-centered dynamic clustering, user-centered dynamic clustering, and hybrid clustering. In the dynamic clustering method taking users as centers, a CoMP service base station cluster suitable for each user is selected, the complexity is highest, and the performance is optimal. The document "Semi-dynamic user-specific clustering for downlink clustered radio access network" (Liu Dong, Han Shengqian, Yang Chengyang, Zhang Qian, IEEE Transactions on Vehicular Technology,2016) proposes a user-centric Semi-dynamic clustering method, which, however, requires traversing all possible CoMP serving base station clusters of all users, with the number of iterations being proportional to the square of the total number of users, and with more users, the complexity is very high.

Disclosure of Invention

In summary, in order to solve the problems in the prior art, the present invention provides a user-centered semi-dynamic clustering method, which is suitable for an ultra-dense network using JT-CoMP and has low complexity.

The technical idea for realizing the invention is as follows: dividing the small base station into K clusters based on the interference graph, selecting the cluster head of each cluster, and using Q_kDenotes the kth cluster, K is 1,2, …, K, and Q is found in the interference graph_kAnd is not with Q_kSmall base stations in the cell interfering with each other, placing the small base stations in P_kIn (1), Q_kSmall cell in and P_kThe small base station in middle is regarded as Q_kAnd the cluster head firstly adds one of the service base stations to be selected to the CoMP service base station cluster of the user, and then determines whether to add other service base stations to be selected to the CoMP service base station cluster of the user according to the throughput of the system.

In order to realize the technical thought, the semi-dynamic clustering method taking the user as the center, which is provided by the invention, is suitable for the ultra-dense network adopting JT-CoMP, and comprises the following steps:

a, carrying out non-overlapping clustering on small base stations in the network, selecting a cluster head of each cluster, dividing the cluster heads into K clusters, and using Q_kIs represented by K ═ 1,2, …, K;

b, representing the set of all the small base stations in the network by Q, and using Q as the reference

Represents Q_kRelative complement in Q, based on the interference graph constructed in step A1, at

In the column and Q_kConnected to a small base station inSmall base stations placed in set P_kK is 1,2, …, K is the total number of base station clusters contained in the network;

C，Q_km user slave Q_kAnd P_kSelecting the service base station to be selected to form a set U_kmAnd sends the set to one of the small base stations, which sends all the sets it receives to the cluster head of the cluster, m 1,2, …, I_k，I_kIs the kth cluster, i.e. Q_kThe total number of users contained, K is 1,2, …, K is the total number of base station clusters contained in the network;

d, the cluster head of the 1 st cluster is connected with U_1i1 st small base station BS in_1i,1JT-CoMP service base station cluster of I-th user in the cluster, I is 1,2, …, I₁，I₁Is the total number of users in the 1 st cluster, the cluster head calculates the sum of the throughputs of all users in the cluster at the moment and assigns a value to R₁；

E, the cluster head of the 1 st cluster determines whether to connect U_1iThe 2 nd to the last small base station in the cluster are added into a JT-CoMP service base station cluster of the ith user in the cluster, I is 1,2, …, I₁，I₁Is the total number of users contained in the 1 st cluster;

and F, selecting a JT-CoMP service base station cluster for each user in the cluster by adopting the same method as the cluster head of the 1 st cluster through the cluster heads of the 2 nd to the Kth clusters, wherein K is the total number of the base station clusters contained in the network.

Further, the step a specifically includes:

a1, constructing an interference graph corresponding to the network, wherein nodes in the graph correspond to small base stations, edges correspond to the distance between the small base stations or the large-scale fading between the small base stations, if the distance between the small base stations is smaller than a preset distance threshold or the large-scale fading between the small base stations is smaller than a preset fading threshold, edges exist between the nodes corresponding to the two small base stations, the length of the edges is equal to the distance between the two base stations or the large-scale fading between the two base stations, and if the distance between the small base stations is larger than a preset distance threshold or the large fading scale between the small base stations is larger than a preset fading threshold, no edge exists between the nodes corresponding to the two small base stations;

a2, based on the interference diagram constructed in step A1, using K-means method or other clustering method to cluster the small base stations without overlapping, putting the small base stations with large interference into one cluster and selecting the cluster head of each cluster to divide into K clusters, using Q to divide into K clusters_kDenotes, K ═ 1,2, …, K.

Further, the step C specifically includes:

C1，Q_kand P_kThe small base station in the middle transmits a reference signal to Q_kPer user in, Q_kM user measurement Q_kAnd P_kTo its instantaneous received signal power, m 1,2, …, I_k，I_kIs the total number of users contained in the kth cluster, K is 1,2, …, K is the total number of base station clusters contained in the network;

C2，Q_kthe mth user compares the power of the plurality of instantaneous received signals with a preset power threshold, selects a plurality of small base stations corresponding to the received signals larger than the power threshold, and sequentially places the small base stations corresponding to the instantaneous received signals in a set U according to the sequence of the power from large to small_kmIn, U_km＝{BS_km,1,BS_km,2,…,BS_km,nkm}，U_kmThe element in (1) represents a small base station, m is 1,2, …, I_k，I_kIs the total number of users contained in the kth cluster, K is 1,2, …, K is the total number of base station clusters contained in the network, nkm is U_kmThe total number of elements in;

C3，Q_km users of (2) will set U_kmSending to the small base station BS_km,1Small base station BS_km,1Will be set U_kmThe cluster head sent to the cluster, m 1,2, …, I_k，I_kIs the total number of users contained in the kth cluster, K is 1,2, …, K is the total number of base station clusters contained in the network.

Further, the step E specifically includes:

e1, if | U₁₁| is more than or equal to 2, |, represents the potential of the set, then the cluster head will be the small base station BS_11,2When acting as JT-CoMP service base station for user 1, user 1 has two service base stations, clustersHead calculates I in the cluster₁The sum of throughputs of the users, if the sum of throughputs is greater than R₁Then cluster head will small base station BS_11,2Adding the CoMP service base station cluster of the user 1, and assigning the sum of the throughputs to R₁Otherwise, the small base station BS is not considered later_11,2As the JT-CoMP serving base station for user 1;

e2, if | U₁₂| is more than or equal to 2, |, represents the potential of the set, then the cluster head will be the small base station BS_12,2When the user 2 is used as JT-CoMP service base station, the user 2 has two service base stations, and the cluster head calculates I in the cluster₁The sum of throughputs of the users, if the sum of throughputs is greater than R₁Then cluster head will small base station BS_12,2Adding JT-CoMP service base station cluster of user 2, and assigning the sum of throughput to R₁Otherwise, the small base station BS is not considered later_12,2As the JT-CoMP serving base station for user 2;

e3, if | U_1i| is ≧ 2, |, represents the potential of the set, the cluster head decides whether to use the small base station BS in the same way as step E1 and step E2_1i,2Adding JT-CoMP service base station cluster of user i, if small base station BS_1i,2Adding JT-CoMP service base station cluster of user I, then I in the cluster at the moment₁The sum of throughputs of the users is assigned to R₁，i＝3,4,…,I₁，I₁Is the total number of users contained in the 1 st cluster;

e4, if | U₁₁| is more than or equal to 3, |, represents the potential of the set, then the cluster head will be the small base station BS_11,3As the JT-CoMP serving base station of the user 1, the cluster head calculates I in the cluster₁The sum of throughputs of the users, if the sum of throughputs is greater than R₁Then cluster head will small base station BS_11,3Adding JT-CoMP service base station cluster of user 1, and assigning the sum of throughput to R₁Otherwise, the small base station BS is not considered later_11,3As the JT-CoMP serving base station for user 1;

e5, if | U₁₂| is more than or equal to 3, |, represents the potential of the set, then the cluster head will be the small base station BS_12,3As the JT-CoMP serving base station for user 2, the cluster head calculates I in the cluster₁Sum of throughputs of individual usersIf the sum of the throughputs is greater than R₁Then cluster head will small base station BS_12,3Adding JT-CoMP service base station cluster of user 2, and assigning the sum of throughput to R₁Otherwise, the small base station BS is not considered later_12,3As the JT-CoMP serving base station for user 2;

e6, if | U_1i| ≧ 3, |, represents the potential of the set, the cluster head decides whether to use the small base station BS in the same way as step E4 and step E5_1i,3Adding JT-CoMP service base station cluster of user i, if small base station BS_1i,3Adding JT-CoMP service base station cluster of user I, then I in the cluster at the moment₁The sum of throughputs of the users is assigned to R₁，i＝3,4,…,I₁，I₁Is the total number of users contained in the 1 st cluster;

e7, the cluster head of the 1 st cluster uses the same method as the steps E1 to E6 to decide whether to switch U_1iThe other small base stations in the group are added into JT-CoMP service base station cluster of user I, I is 1,2, …, I₁，I₁Is the total number of users contained in the 1 st cluster.

Advantageous effects

The invention firstly divides the small base stations into a plurality of clusters and selects the cluster heads, then finds out the small base stations outside each cluster and interfering with the cluster, and uses the small base stations and the small base stations in the clusters as the alternative service base stations of the users in the clusters, thereby reducing the search range. The cluster head firstly adds one alternative service base station into the CoMP service base station cluster of the user according to the received signal power, and then sequentially considers whether other small base stations are added into the CoMP service base station cluster of the user or not, the iteration times are in direct proportion to the total number of the users, and the calculation complexity is reduced.

Drawings

FIG. 1 is an interference diagram of an embodiment of the present invention;

FIG. 2 is a clustering diagram of an embodiment of the present invention;

fig. 3 is a flow chart of the present invention.

Detailed Description

The present invention will be described in further detail with reference to an embodiment. Consider an ultra-dense network comprising several small base stations and a plurality of users, both randomly distributed within the network. Each small base station is connected with the central controller through a return link, and all the small base stations adopt the same frequency band. For each user, a plurality of small base stations serve the user at the same time in a joint transmission mode, and the base stations are put together to form a CoMP service base station cluster of the user. Overlapping clusters are considered here, i.e. the same small base station can serve multiple users simultaneously.

The central controller first constructs an interference graph corresponding to the network, as shown in fig. 1, nodes in the graph, namely circles, represent small base stations, numbers in the circles represent serial numbers of the small base stations, edges correspond to distances between the small base stations or large-scale fading between the small base stations, if the distance between the small base stations is smaller than a preset distance threshold or the large-scale fading between the small base stations is smaller than a preset fading threshold, edges exist between the nodes corresponding to the two small base stations and the length of the edges is equal to the distance between the two base stations or the large-scale fading between the two base stations, and if the distance between the small base stations is larger than a preset distance threshold or the large-scale fading between the small base stations is larger than a preset fading threshold, no edge exists between the nodes corresponding to the two small base stations.

The central controller adopts K-means method or other clustering method to cluster the small base stations without overlapping, the small base stations with large interference are placed in one cluster, the cluster head of each cluster is selected, the cluster heads are divided into K clusters, and Q is used_kDenotes, K ═ 1,2, …, K. By way of example, the small base stations are here divided into 4 clusters, as shown in fig. 2, Q₁＝{1,2,3,6,7,8,11,12,13}，Q₂＝{4,5,9,10,14,15,16,18,32}，Q₃＝{19,20,21,22,27,28,29,30,31}，Q₄＝{17,23,24,25,26}。

If only cluster Q is present_kThe small base station in the middle is cluster Q_kUser service in, cluster Q_kAnd Q in the interference diagram_kSmall base stations connected to the inner small base station may be paired to the cluster Q_kThe inner boundary users cause large interference, and the small base station 8 in fig. 2 may cause interference to the boundary users of the second cluster, so that it is necessary to use the cluster Q_kOuter and withCluster Q_kSmall base station with large interference of inner small base station is added into cluster Q_kAlternative serving base stations for the inner users. Q denotes the set of all small base stations in the network, and

represents Q_kRelative complement in Q, based on interference patterns, in

In the column and Q_kThe small base stations are connected with each other and are placed in the set P_kWhere K is 1,2, …, and K is the total number of base station clusters included in the network. Cluster Q_kAnd set P_kAll the small base stations in the middle are cluster Q_kAnd the alternative serving small base station of the inner user. As an example, P₁＝{9,14,18,19,20,21}，P₂＝{8,13,23,24,17}，P₃＝{11,12,13,23,26}，P₄＝{22,18,16,32,27,31}。

Below with cluster Q₁For example, a method for selecting a CoMP serving base station cluster for each user by a cluster head is given.

Cluster Q₁And set P₁The small base station in the middle sends a reference signal to a cluster Q₁Per user in (2), cluster Q₁The ith user of (2) measures the cluster Q₁And set P₁I-1, 2, …, I-to its instantaneous received signal power₁，I₁Is the total number of users contained in the 1 st cluster.

Cluster Q₁The ith user compares the power of a plurality of instantaneous received signals with a preset power threshold, selects a plurality of small base stations corresponding to the received signals larger than the power threshold, and sequentially puts the small base stations corresponding to the instantaneous received signals in a set U according to the sequence of the power from large to small_1iIn, U_1i＝{BS_1i,1,BS_1i,2,…,BS_1i,n1i}，U_1iThe element in (1) represents a small base station, I is 1,2, …, I₁N1i is U_1iThe total number of elements in (c).

Cluster Q₁Will set U_1iSending to the small base station BS_1i,1Small base station BS_1i,1Will be set U_1iAnd sending to the cluster head of the cluster.

Cluster head of 1 st cluster is U_1i1 st small base station BS in_1i,1JT-CoMP service base station cluster of I-th user in the cluster, I is 1,2, …, I₁The cluster head calculates the sum of the throughputs of all users in the cluster at that time and assigns a value to R₁。

If U₁₁| is more than or equal to 2, |, represents the potential of the set, then the cluster head will be the small base station BS_11,2When the user 1 is used as JT-CoMP service base station of the user 1, the user 1 has two service base stations, and the cluster head calculates I in the cluster₁The sum of throughputs of the users, if the sum of throughputs is greater than R₁Then cluster head will small base station BS_11,2Adding the CoMP service base station cluster of the user 1, and assigning the sum of the throughputs to R₁Otherwise, the small base station BS is not considered later_11,2As the JT-CoMP serving base station for user 1;

if U₁₂| is more than or equal to 2, |, represents the potential of the set, then the cluster head will be the small base station BS_12,2When the user 2 is used as JT-CoMP service base station, the user 2 has two service base stations, and the cluster head calculates I in the cluster₁The sum of throughputs of the users, if the sum of throughputs is greater than R₁Then cluster head will small base station BS_12,2Adding JT-CoMP service base station cluster of user 2, and assigning the sum of throughput to R₁Otherwise, the small base station BS is not considered later_12,2As the JT-CoMP serving base station for user 2;

if U_1iIf | ≧ 2, the cluster head adopts the same method to decide whether to use the small base station BS_1i,2Adding JT-CoMP service base station cluster of user i, if small base station BS_1i,2Adding JT-CoMP service base station cluster of user I, then I in the cluster at the moment₁The sum of throughputs of the users is assigned to R₁，i＝3,4,…,I₁。

If U₁₁| is more than or equal to 3, |, represents the potential of the set, then the cluster head will be the small base station BS_11,3As the JT-CoMP serving base station of the user 1, the cluster head calculates I in the cluster₁The sum of throughputs of the users is larger thanR₁Then cluster head will small base station BS_11,3Adding JT-CoMP service base station cluster of user 1, and assigning the sum of throughput to R₁Otherwise, the small base station BS is not considered later_11,3As the JT-CoMP serving base station for user 1.

If U₁₂| is more than or equal to 3, |, represents the potential of the set, then the cluster head will be the small base station BS_12,3As the JT-CoMP serving base station for user 2, the cluster head calculates I in the cluster₁The sum of throughputs of the users, if the sum of throughputs is greater than R₁Then cluster head will small base station BS_12,3Adding JT-CoMP service base station cluster of user 2, and assigning the sum of throughput to R₁Otherwise, the small base station BS is not considered later_12,3As the JT-CoMP serving base station for user 2.

If U_1iIf | ≧ 3, | · | represents the potential of the set, the cluster head adopts the same method to decide whether to use the small base station BS_1i,3Adding JT-CoMP service base station cluster of user i, if small base station BS_1i,3Adding JT-CoMP service base station cluster of user I, then I in the cluster at the moment₁The sum of throughputs of the users is assigned to R₁，i＝3,4,…,I₁。

The cluster head of the 1 st cluster determines whether to convert U in the same way as above_1iThe other small base stations in the group are added into JT-CoMP service base station cluster of user I, I is 1,2, …, I₁。

And selecting a JT-CoMP service base station cluster for each user in the cluster by adopting the same method as the cluster head of the 1 st cluster from the cluster heads of the 2 nd cluster to the 4 th cluster.

With reference to the flowchart of the present invention, i.e., fig. 3, the user-centered semi-dynamic clustering method includes the following specific steps:

To representQ_kRelative complement in Q, based on the interference graph constructed in step A1, at

In the column and Q_kThe small base stations are connected with each other and are placed in the set P_kK is 1,2, …, K is the total number of base station clusters contained in the network;

The above embodiments are merely illustrative of the present invention, and those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if such modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include such modifications and variations.

Claims

1. The semi-dynamic clustering method taking users as centers is characterized in that the method is suitable for a super-dense network adopting JT-CoMP and comprises the following steps:

a, carrying out non-overlapping clustering on small base stations in the network, selecting a cluster head of each cluster, dividing the cluster heads into K clusters, and using Q_kThe expression, K is 1,2, … and K, and the specific process is as follows:

a2, based on the interference diagram constructed in step A1, using K-means method to cluster the small base stations without overlapping, putting the small base stations with large interference into one cluster and selecting the cluster head of each cluster, dividing into K clusters altogether, and using Q to use Q to divide the cluster heads into K clusters_kIs represented by K ═ 1,2, …, K;

b, Q is used to represent the set of all small base stations in the network, C is used_QQ_kRepresents Q_kRelative complement in Q, based on the interference graph constructed in step A1, at C_QQ_kIn the column and Q_kThe small base stations are connected with each other and are placed in the set P_kK is 1,2, …, K is the total number of base station clusters contained in the network;

C，Q_km user slave Q_kAnd P_kSelecting the service base station to be selected to form a set U_kmAnd sends the set to one of the small base stations, which sends all the sets it receives to the cluster head of the cluster, m 1,2, …, I_k，I_kIs the kth cluster, i.e. Q_kIncluding usersThe total number K is 1,2, …, and K is the total number of the base station clusters contained in the network, and the specific process is as follows:

C3，Q_km users of (2) will set U_kmSending to the small base station BS_km,1Small base station BS_km,1Will be set U_kmThe cluster head sent to the cluster, m 1,2, …, I_k，I_kIs the total number of users contained in the kth cluster, K is 1,2, …, K is the total number of base station clusters contained in the network;

E, the cluster head of the 1 st cluster determines whether to connect U_1iThe 2 nd to the last small base station in the cluster are added into a JT-CoMP service base station cluster of the ith user in the cluster, I is 1,2, …, I₁，I₁Is the total number of users contained in the 1 st cluster, and the specific processThe following were used:

e1, if | U₁₁| is more than or equal to 2, |, represents the potential of the set, then the cluster head will be the small base station BS_11,2When the user 1 is used as JT-CoMP service base station of the user 1, the user 1 has two service base stations, and the cluster head calculates I in the cluster₁The sum of throughputs of the users, if the sum of throughputs is greater than R₁Then cluster head will small base station BS_11,2Adding the CoMP service base station cluster of the user 1, and assigning the sum of the throughputs to R₁Otherwise, the small base station BS is not considered later_11,2As the JT-CoMP serving base station for user 1;

e5, if | U₁₂| is more than or equal to 3, |, represents the potential of the set, then the cluster head will be the small base station BS_12,3As the JT-CoMP serving base station for user 2, the cluster head calculates I in the cluster₁The sum of throughputs of the users, if the sum of throughputs is greater than R₁Then cluster head will small base station BS_12,3Adding JT-CoMP service base station cluster of user 2, and assigning the sum of throughput to R₁Otherwise, the small base station BS is not considered later_12,3As the JT-CoMP serving base station for user 2;

e7, the cluster head of the 1 st cluster uses the same method as the steps E1 to E6 to decide whether to switch U_1iThe other small base stations in the group are added into JT-CoMP service base station cluster of user I, I is 1,2, …, I₁，I₁Is the total number of users contained in the 1 st cluster;