A kind of collaboration communication method based on many district dynamic cluster
Technical field
The invention belongs to the wireless cellular network communications field, specifically, relate to a kind of collaboration communication method based on many district dynamic cluster.
Background technology
Along with the sharp increase of userbase and increasing of wireless traffic, cellular communications networks faces a significant challenge---improve the utilance of limited band resource.For satisfying the demands, the base station will be set up more thick and fast, frequency repeat utilization ratio will improve, even takies the available band of whole distribution.(inter-cell interference, ICI), (Quality of Service QoS) will be affected, particularly Cell Edge User to make QoS of customer but this will further cause more serious presence of intercell interference.Collaboration communication is a kind of effective means of being devoted to alleviate presence of intercell interference, and 3GPP formally includes the collaboration communication technology in the LTE-Advanced standard, makes it to become the key technology that improves following Radio Network System performance.
Collaboration communication provides possibility for abundant development system spatial degrees of freedom, and its basic thought is in down link, and handle the combined signal that will send a plurality of residential quarters, thereby make home cell suppress the interference signal of other residential quarters.Base station cooperation in a cooperation unit forms one bunch, and by cooperation, the interference of minizone will be effectively suppressed in bunch like this.In theory, more many with number of base stations in the cluster, presence of intercell interference is suppressed more obvious; But in practicality, more cell cooperative means more feedback overhead, computing cost and backbone network transport overhead, and this situation makes can not be too many with number of cells in the cluster.
At choosing which this problem of base station cluster, some fruitful algorithms are suggested, what consider mostly is according to the static policies of base station geography close position cluster, namely no longer change behind the selected cluster base station, yet the wireless channel time-varying characteristics may make this method improve limited to performance.To this, the base station of some dynamic clustering methods---cluster is suggested according to continuous variation of channel information, makes the static relatively clustering method of systematic function have clear improvement.Yet, existing dynamic clustering algorithm be mostly all cluster may in search thoroughly and calculate have maximum system capacity bunch as a result of, this process is quite complicated; There is the scholar that this class algorithm has been made slight change, calculates the cluster result that target is changed to all possible bunch channel gain and selects the overall gain maximum; But these methods all need the possibility of all clusters of limit.When network size was big, even base station number is less in each bunch, amount of calculation was also quite big, may calculate finish after, channel condition changes already.It is also inapplicable when therefore, this exhaustive search method is in real network.
Summary of the invention
Technical problem: technical problem to be solved by this invention is: a kind of collaboration communication method based on many district dynamic cluster is provided, this communication means makes each base station to determine whether and its cluster according to the degree of association size with other base stations, greatly accelerate dynamic clustering speed, thereby improved communication efficiency.
Technical scheme: for solving the problems of the technologies described above, the technical solution used in the present invention is:
A kind of collaboration communication method based on many district dynamic cluster, these many residential quarters comprise at least three residential quarters, are provided with travelling carriage and base station in each residential quarter, this communication means may further comprise the steps:
The first step: set up station list and calculate the minizone matrix of mutual interference mutually:
Initialization candidate base station tabulation Candidacy={1,2 ..., B} has selected station list
B represents the sum of base station in many residential quarters, sets the base station maximum quantity that comprises in each bunch and is B
c, B
c<B, the number in many residential quarters bunch is
Wherein, symbol
Expression rounds downwards;
Feed back to the channel information matrix H of each base station according to all travelling carriages of many residential quarters, by formula (1) the measuring and calculating mutual interference matrix F in minizone:
Formula (1)
Wherein, h
IjChannel gain between the transmitting antenna of expression user i and base station j, h
JiChannel gain between the transmitting antenna of expression user j and base station i, f
IjBe the capable j column element of i of mutual interference matrix F, f
IjMutual interference degree between expression base station i and the base station j;
Second step: in the candidate base station tabulation, select a cell base station i who is subjected to other area interference maximums
l, as when the guiding base station of prevariety, guide base station i
lSatisfy formula (2):
Formula (2)
According to formula (3) with i
lDeletion from candidate base station tabulation, thus upgrade the candidate base station tabulation, according to formula (4) with i
lAdd to and select in the station list, selected station list thereby upgrade:
Candidacy1=Candidacy-{i
lFormula (3)
Preclusion1=Preclusion+{i
lFormula (4)
Wherein, the tabulation of candidate base station after Candidacy1 represents to upgrade is first selected station list after Preclusion1 represents to upgrade first;
The 3rd step: select one to the guiding base station i when prevariety in the candidate base station tabulation after renewal
lDisturb maximum cell base station i ' adding to work as prevariety, base station i ' satisfies
f
Ii 'Annoyance level between expression base station i and the base station i '; With deletion the candidate base station tabulation of base station i ' after last update, upgrade the candidate base station tabulation; Add base station i ' to selecting in the station list after the last update, upgrade and selected station list, entered for the 4th step then;
The 4th step: returned for the 3rd step, equal B until the number of base stations in prevariety
c, entered for the 5th step then;
The 5th step: the candidate base station tabulation after will upgrading through the 4th step is tabulated as candidate base station, station list has been selected in the station list conduct of selecting after will upgrading through the 4th step, returned for second step, carry out next bunch base station selected, the candidate base station tabulation after once upgrading as of late is empty set
After the last update select station list be 1,2 ..., B}, that sets up all bunches in many residential quarters selects station list;
The 6th step: system in all base stations by all bunches of the 5th step foundation select station list, carry out collaboration communication by the cluster result.
Beneficial effect: compared with prior art, technical scheme of the present invention has following beneficial effect:
1. the measuring and calculating interaction matrix is simple to operate, and only a sub-addition just can obtain, and this matrix can accurately be weighed the degree that influences each other between the residential quarter, but not independent influence to the other side.
2. adopt the guiding base station selecting method, different base station has different priority in bunch, the randomness of having got rid of the conventional dynamic clustering method has to a certain extent improved the information rate performance of collaboration communication, the formation that the cell base station that can accurately select influenced maximum is dominated bunch.
3. guiding base station cell annoyance level being come finally got rid of the randomness of conventional dynamic clustering method for the different priority of other base station configurations, thus make cluster bunch in disturb minimum.The present invention determines whether and its cluster according to the degree of association size of base station and other base stations, has greatly accelerated dynamic clustering speed.
Description of drawings
Fig. 1 is the system schematic of the embodiment of the invention, and wherein, solid line is represented the useful signal link, and dotted line is represented the interference signal link, the base station cluster in two residential quarters of light-colored part, the base station cluster in three residential quarters of dark part.
Fig. 2 be bunch in the base station count B
c=2 o'clock, the method for the embodiment of the invention, static clustering method and do not have the contrast string diagram of system's average size of obtaining respectively of cooperation clustering method.
Fig. 3 be bunch in the base station count B
c=3 o'clock, the method for the embodiment of the invention, static clustering method and do not have the contrast string diagram of system's average size of obtaining respectively of cooperation clustering method.
Fig. 4 be bunch in the base station count B
c=4 o'clock, the method for the embodiment of the invention, static clustering method and do not have the contrast string diagram of system's average size of obtaining respectively of cooperation clustering method.
Fig. 5 be bunch in the base station count B
c=2 o'clock, the method for the embodiment of the invention, tradition were based on the dynamic clustering method of searching mode thoroughly and do not have the contrast string diagram of system's average size of obtaining respectively of cooperation clustering method.
Fig. 6 be bunch in the base station count B
c=3 o'clock, the method for the embodiment of the invention, tradition were based on the dynamic clustering method of searching mode thoroughly and do not have the contrast string diagram of system's average size of obtaining respectively of cooperation clustering method.
Fig. 7 be bunch in the base station count B
c=4 o'clock, the method for the embodiment of the invention, tradition were based on the dynamic clustering method of searching mode thoroughly and do not have the contrast string diagram of system's average size of obtaining respectively of cooperation clustering method.
Embodiment
Below in conjunction with drawings and Examples, technical scheme of the present invention is described in further detail.
The many residential quarters that the present invention relates to comprise at least three residential quarters, are provided with travelling carriage and base station in each residential quarter.A kind of collaboration communication method based on many district dynamic cluster of the present invention may further comprise the steps:
The first step: set up station list and calculate the minizone matrix of mutual interference mutually:
Initialization candidate base station tabulation Candidacy={1,2 ..., B} has selected station list
B represents the sum of base station in many residential quarters, sets the base station maximum quantity that comprises in each bunch and is B
c, B
c<B, the number in many residential quarters bunch is
Wherein, symbol
Expression rounds downwards;
Feed back to the channel information matrix H of each base station according to all travelling carriages of many residential quarters, by formula (1) the measuring and calculating mutual interference matrix F in minizone:
Formula (1)
Wherein, h
IjChannel gain between the transmitting antenna of expression user i and base station j, h
JiChannel gain between the transmitting antenna of expression user j and base station i, f
IjBe the capable j column element of i of mutual interference matrix F, f
IjMutual interference degree between expression base station i and the base station j.As preferred version, B
cIn that { value among 2,3, the 4}, the quantity of B is less than or equal to 37.
Second step: in the candidate base station tabulation, select a cell base station i who is subjected to other area interference maximums
l, as when the guiding base station of prevariety, guide base station i
lSatisfy formula (2):
Formula (2)
According to formula (3) with i
lDeletion from candidate base station tabulation, thus upgrade the candidate base station tabulation, according to formula (4) with i
lAdd to and select in the station list, selected station list thereby upgrade:
Candidacy1=Candidacy-{i
lFormula (3)
Preclusion1=Preclusion+{i
lFormula (4)
Wherein, the tabulation of candidate base station after Candidacy1 represents to upgrade is first selected station list after Preclusion1 represents to upgrade first;
The 3rd step: select one to the guiding base station i when prevariety in the candidate base station tabulation after renewal
lDisturb maximum cell base station i ' adding to work as prevariety, base station i ' satisfies
f
Ii 'Annoyance level between expression base station i and the base station i '.With deletion the candidate base station tabulation of base station i ' after last update, upgrade the candidate base station tabulation; Add base station i ' to selecting in the station list after the last update, upgrade and selected station list, entered for the 4th step then.
The 4th step: returned for the 3rd step, equal B until the number of base stations in prevariety
c, entered for the 5th step then.
The 5th step: the candidate base station tabulation after will upgrading through the 4th step is tabulated as candidate base station, station list has been selected in the station list conduct of selecting after will upgrading through the 4th step, returned for second step, carry out next bunch base station selected, the candidate base station tabulation after once upgrading as of late is empty set
After the last update select station list be 1,2 ..., B}, that sets up all bunches in many residential quarters selects station list;
The 6th step: system in all base stations by all bunches of the 5th step foundation select station list, carry out collaboration communication by the cluster result.
Collaboration communication method based on many district dynamic cluster provided by the invention, for reducing the common-channel interference of minizone, the LTE-A standard has been introduced a kind of mechanism of multi-cell cooperating, by a plurality of base stations associated treatment receiving and transmitting signal (set of the base station in an IWU be called bunch), to disturb signal be useful signal thereby become dry.At how to choose this problem of base station cluster, existing method generally according to geographical far and near as sole criterion between the base station, can't adapt to the wireless channel time-varying characteristics.The dynamic clustering method allows the different base station cluster adaptively according to channel variation, existing dynamic clustering method all by search thoroughly to find the solution make the channel gain maximum cluster may, the huge amount of calculation of this mode has hindered its practicality.The friend-making mechanism of communication means simulating human of the present invention has proposed the dynamic clustering method based on the friend-making algorithm.Existing relatively dynamic clustering method, dynamic clustering method of the present invention have the low advantage of complexity.Method of the present invention has been constructed the mutual interference matrix of each minizone according to the channel condition information of each travelling carriage feedback, and directly determines which base station cluster according to the annoyance level between each residential quarter, can the disposable cluster result that tries to achieve.This calculation side makes and the base station collaboration cluster that common-channel interference is serious disturbs by ZF precoding counteracting, thereby improves power system capacity.The present invention is based on the characteristic of human friend-making activity, be applied to multi-cell cooperating dynamic clustering problem, make each base station to determine whether and its cluster according to the degree of association size with other base stations, greatly accelerated cluster speed, and the gained performance is almost identical with the mode of searching thoroughly, and this method will be well suited for practical application.
Residential quarter as shown in Figure 1, the collaboration communication pattern that adopts method of the present invention to set up.Be provided with five residential quarters among Fig. 1, solid line is represented the useful signal link, dotted line is represented the interference signal link, the base station of two residential quarters of light-colored part (namely being positioned at leftmost two residential quarters) becomes one bunch, the cluster in base station of three residential quarters of dark part (namely being positioned at three residential quarters on middle part and the right).
Exemplify an embodiment below.
This embodiment chooses four layers of cell pattern, and total number of base is 37 in the system, i.e. B=37, and B is counted in the base station in bunch
cIn { value among 2,3, the 4}.It is 1km that radius of society is set, and travelling carriage is randomly dispersed in each position in the residential quarter, and poll strategy scheduling travelling carriage is adopted in the base station, and the base station number is fixed as B in bunch
c, remain not enough B
cThe independent cluster in individual base station.Channel model between travelling carriage i and the base station j is shown below:
Wherein, Γ
IjBe that variance is 1, obey the Rayleigh fading of multiple Gaussian Profile, G
IjLong-pending for the dual-mode antenna gain, α is path loss index, and β is the path loss constant, d
IjBe the distance between travelling carriage i and the base station j, r
IjFor obeying the shadow fading of logarithm Gaussian Profile.According to model that existing 3GPP gives (source is seen: LTE-A Evaluation Model, NTT DocoMo, Available online.), G is set
Ij=9dB, α=3.76, β=10
-14.84, r
IjAverage be 0dB, variance is 8dB.Given antenna for base station maximum transmission power is P
n=1W, the definition relative signal-to-noise ratio is the signal to noise ratio that cell edge is only considered path loss and Gaussian noise.Adopt the ZF precoding to carry out Monte-Carlo Simulation 3000 times.Simulation result such as table 1, Fig. 2 are to shown in Figure 7.Clustering method as a comparison has: (source is seen: S.Zhou to search tactful dynamic clustering method thoroughly, J.Gong, Z.Niu, Y.Jia, and P.Yang, " A decentralized framework for dynamic downlink base station cooperation; " in Proc.IEEE GLOBECOM, Dec.2009, pp.1 – 6.), the clustering method of static clustering method and nothing cooperation (C.3-Cell Network MIMO Architectures with Sectorization and Fractional Frequency Reuse[J of Wang L is seen in the source] .IEEE Journal on Selected Areas in Communications.2011.29 (6): 1185-1199.).
Table 1
In table 1, T
mRepresent the required CM number of times of cluster in the communication means of the present invention, T
aRepresent that the required plural number of cluster adds number of times, T in the communication means of the present invention
cThe number of times of representing the comparison operation that cluster in the communication means of the present invention is required; T
mThe required CM number of times of dynamic clustering method of the poor search method of mentioning in ' expression the background technology, T
aThe required plural number of dynamic clustering method of the poor search method of mentioning in ' expression the background technology adds number of times, T
cThe number of times of the comparison operation that the dynamic clustering method of the poor search method of mentioning in ' expression the background technology is required.As can be seen from Table 1, in having same cluster during number of base stations, T
mMuch smaller than T
m', T
aMuch smaller than T
a', T
cMuch smaller than T
c'.
From Fig. 2 to Fig. 4 as can be seen, compare with static clustering method, adopt dynamic clustering method of the present invention, the average size gain that system obtains has obtained bigger must the raising.From Fig. 5-Fig. 7 as can be seen, with based on search thoroughly the strategy the dynamic clustering method compare, adopt the current performance of dynamic clustering method of the present invention to be more or less the same, but its amount of calculation obviously reduces.