Summary of the invention
The technical problem to be solved is to provide and a kind of uses brand-new design framework method, it is possible to is effectively improved and is
The MU-MIMO system true environment low complex degree user choosing method of system effective throughput.
The present invention is to solve above-mentioned technical problem by the following technical solutions: the present invention devises a kind of MU-MIMO system
System true environment low complex degree user choosing method, for having the base station of MU-MINO system, carries out user's selection;Described MU-
Mimo system true environment low complex degree user choosing method comprises the steps:
Step 001. initializes and selects family set S is empty set, and initialization alternative user set K includes base station MU-
Active user in MINO system down link and non real time user, and randomly generate in the most corresponding alternative user set K
The channel matrix of each user, meanwhile, initializing set Q to be selected is empty set, and parameter m=1, subsequently into step 002;
Step 002. is respectively directed to each user in alternative user set K, and the channel matrix calculating user is conjugated with it
The determinant of the product matrix of matrix, and obtain the determinant corresponding to this determinant, and then obtain in alternative user set K
The determinant of each user, then by alternative user set K, the user corresponding to maximum determinant value moves into and has selected family
Gather in S, and calculate the kernel of this user, in alternative user set K, delete this user, subsequently into step 003 simultaneously;
Step 003. judges to select in family set S whether there is a user, has selected in family set S in addition to it it
The combined channel rank of matrix of he all users, more than the antenna number of base station, is, user selects to terminate;Otherwise enter step
004;
After step 004. assumes that the m-th user in alternative user set K replicates immigration has selected family set S, to
Select each user in family set S to carry out block diagonalization process, and calculate this and be newly added user pre-selecting family set S
Encoder matrix, further according to the pre-coding matrix of this user, calculates the equivalent channel square of this user in conjunction with the channel matrix of this user
Battle array;Equivalent channel matrix and the product matrix determinant of its conjugate matrices then according to this user, it is thus achieved that the speed of this user
Low complex degree token state;Subsequently into step 005;
Step 005. continues assuming that the m-th user in alternative user set K replicates immigration has selected family set S's
On the basis of, continue to calculate and selected in family set S in addition to this user, the equivalent channel matrix of other each users, and according to
The equivalent channel matrix of this each user, and the transition matrix of this each user, calculate the most corresponding speed of this each user
The low complex degree token state of rate variable quantity, the low complex degree then according to the most corresponding speed variable quantity of this each user characterizes
Amount, updates and has selected the speed low complex degree token state of active user in family set S, subsequently into step 006;
Step 006. continues assuming that the m-th user in alternative user set K replicates immigration has selected family set S's
On the basis of, it may be judged whether select the speed low complex degree token state of all active users in family set S to be both greater than it and preset speed
Rate low complex degree token state lower limit, is the m-th user in alternative user set K to be copied in set Q to be selected, and enter
Step 007;Otherwise abandon replicating m-th user in alternative user set K immigration and select family set S's it is assumed that and enter
Step 007;
Step 007. has judged whether for the traversal of all users in alternative user set K, is to enter step
008;Otherwise traverse user sequence number m adds 1, and gives m by this result, is then back to step 004;
Step 008. is chosen a user's immigration making user choose benchmark maximum from set Q to be selected and has been selected family to gather
In S, in alternative user set K, delete this user simultaneously by alternative user set K deletes this user simultaneously, and empty to be selected
Set Q, subsequently into step 009;
Step 009. is newly added the channel matrix of user according to selecting in family set S, and has selected family to gather S Central Plains
There is the pre-coding matrix of each user, calculate and selected the transition matrix of each user original in family set S, further according to selecting
The transition matrix of each user original in family set S, updates and has selected the pre-coding matrix of each user corresponding in family set S,
It is then back to step 003.
As a preferred technical solution of the present invention: also include that step 010 is as follows, in described step 003, user
After selection terminates, enter step 010;
Step 010., according to selecting the channel matrix of each user in family set S, calculates to obtain and has selected in family set S
The pre-coding matrix of each user, and utilize block diagonalization method to eliminate the interference selected in family set S between each user.
As a preferred technical solution of the present invention: in described step 010, for any one in alternative user set K
Individual user j, passes throughEliminate other users interference, for making pre-coding matrix VjThere is untrivialo solution, then the side of being required to meet
The number of journey is less than the number of variable, it may be assumed that
Therefore for any user of access base station, only other user's combined channel ranks of matrix are less than base station and send out
Penetrate summation M of antenna number, pre-coding matrix V could be there isjEnsureing that each user is not disturbed by other users, this is also true ring
To the restriction of user capacity | S | selecting family set S when using block diagonalization technology under border.
As a preferred technical solution of the present invention: in described step 004 to step 006, it is assumed that by alternative user collection
A user in conjunction K replicates immigration and has selected in family set S, and in step 008, chooses one and make from set Q to be selected
User chooses the maximum user's immigration of benchmark and has selected in family set S, and analysis subscriber channel is respectively at rich scattering and dissipates with non-lipid
When penetrating environment, family set S has been selected to be newly added user to being serviced the equivalent channel matrix dimension of user and the impact of order, its
In for the pre-coding matrix V of any user j in set SjDimension beUse nj (|S|)Represent
When being | S | as selecting family set S user capacity, user j will not be produced the spatial degrees of freedom number of inter-user interference
Mesh, then in richness scattering and low scattering the true environment deposited, nj (|S|)Expression formula be:
The equivalent channel matrix of user j isBy above to pre-coding matrix VjThe analysis of dimension understands,ThenOrderAssume in the process carrying out user's selection algorithm
In, the number of users that time a certain, etching system services the most simultaneously is | S |, orderRepresent and selected in family set S
The combined channel matrix of all users, then NeSRepresent combined channel matrix H eSOrder, NeSWithBetween there is relation: Specifically carry out value as follows;
It is derived from
As a preferred technical solution of the present invention: in described step 004 to step 006, the speed of described user is low
Complexity token state, obtains this sign by the method utilizing iterative recursive, it is assumed that number of users that may be served is | S |, is simultaneously
System general power is P0Base station MU-MIMO system in, on each transmit power averaging distribution, then can be by the number of user j
According to speedIt is expressed as:
Wherein σ2Represent white Gaussian noise power, Represent corresponding to selecting in family set S
User j'sAll nonzero eigenvalues long-pending;P0Represent that power is always launched in base station;Represent and select
User gathers the equivalent channel matrix of user j and the product matrix of its conjugate matrices in S,Represent and selected family collection
Close the equivalent channel matrix of any user j in SOrderAnd then obtainValue is incorporated as along with new user's:
User j during (| S |+1) secondary iteration is gone to disturb pre-coding matrix Vj (|S|+1), can be by the side of iterative recursive
Method obtains this value:
Vj (|S|+1)=Vj (|S|)Gj (|S|+1) (21)
Product angle concept thus according to subspace can obtainWithBetween relational expression be:
In formula, θlRepresentation spaceWithBetween l to base vector elAnd alBetween basic angle.
Space can be defined asAnd spaceBetween product angle.
As a preferred technical solution of the present invention: in described step 008, from set Q to be selected, choose a use
Family is chosen the maximum user of benchmark and is moved into and selected in family set S, delete in alternative user set K simultaneously this user simultaneously by
Alternative user set K deletes this user, and empties set Q to be selected, wherein, be respectively directed to each user in set Q to be selected
Q, it is thus achieved that the value of equation below corresponding to each user:And select
Take the user wherein corresponding to maximum, i.e. on the basis of maximum user, move it into and select in family set S, simultaneously standby
Select the while that set K in family deleting this user by alternative user set K deletes this user, and empty set Q to be selected;Wherein,
PavgRepresent the average emitted power on every, base station transmitting antenna, i.e.
MU-MIMO system true environment low complex degree user choosing method of the present invention uses above technical scheme with existing
There is technology to compare, have following technical effect that MU-MIMO system true environment low complex degree user's selecting party designed by the present invention
Method, is analyzed based on real wireless transmission environments (rich scattering and non-lipid scattering all exist), has obtained scattering ring in non-lipid
The dynamic capacity upper limit that in border, user selects, and made a concrete analysis of user's equivalent channel matrix dimension and order along with user's selection repeatedly
The change that the increase in generation is occurred, and by this change application to user's selection reference;It addition, the present invention is by user's equivalent channel
Be multiplied with its conjugate matrices nonzero eigenvalue of the matrix obtained of matrix is long-pending, i.e. low complex degree token state selects base as user
Accurate so that it is to characterize the admissible rate of user to a certain extent, this product value can be obtained by last iterative value recurrence,
Reduce the computation complexity of algorithm, simultaneously on the basis of the differentiated service QoS that becomes more meticulous, improve effectively handling up of system
Amount.
Detailed description of the invention
Below in conjunction with Figure of description, the detailed description of the invention of the present invention is described in further detail.
As it is shown in figure 1, a kind of MU-MIMO system true environment low complex degree user choosing method, pin designed by the present invention
To having the base station of MU-MINO system, carry out user's selection;In the middle of actual application, step specific as follows:
Step 001. initializes and selects family set S is empty set, and initialization alternative user set K includes base station MU-
Active user in MINO system down link and non real time user, and randomly generate in the most corresponding alternative user set K
The channel matrix of each user, meanwhile, initializing set Q to be selected is empty set, and parameter m=1, subsequently into step 002.
Step 002. is respectively directed to each user in alternative user set K, and the channel matrix calculating user is conjugated with it
The determinant of the product matrix of matrix, and obtain the determinant corresponding to this determinant, and then obtain in alternative user set K
The determinant of each user, then by alternative user set K, the user corresponding to maximum determinant value moves into and has selected family
Gather in S, and calculate the kernel of this user, in alternative user set K, delete this user, subsequently into step 003 simultaneously.
Step 003. judges to select in family set S whether there is a user, has selected in family set S in addition to it it
The combined channel rank of matrix of he all users, more than the antenna number of base station, is, user selects to terminate, and enters step 010;
Otherwise enter step 004.
In above-mentioned steps 003, analyze user capacity upper limit when being in rich scattering and non-lipid scattering environments and equivalent channel
Situation, orderRepresent and selected in the collection S of family the combined channel square of all users in addition to user j
Battle array, andThenIt it is oneDimension matrix.
RightCarry out singular value decomposition to obtain:
Wherein,ForUnitary matrice, its column vector isCharacteristic vector;∑jFor
The diagonal matrix of dimension, the element on diagonal matrix isSingular value;For M × M rank unitary matrice;Correspond to ∑j
InThe characteristic vector of individual singular value, is oneThe matrix of dimension;OrderThat represent is pre-coding matrix TjIn
For eliminating the part of the multi-user interference of user j, it is the kernel of other users, VjDimension be
For user j, pre-coding matrix TjCan be by the product gained of two matrixes, i.e. Tj=VjBj.We can lead to
Cross and makeEliminate other users interference, for making VjThere is untrivialo solution, be then required to meet the number of equation less than variable
Number, it may be assumed that
Therefore for any user of access base station, only other user's combined channel ranks of matrix are less than base station and send out
Penetrate summation M of antenna number, pre-coding matrix V could be there isjEnsureing that each user is not disturbed by other users, this is also true ring
Restriction to user capacity | S | when block uses diagonalization technology under border.
Under rich scattering environments, due toSo the upper limit of system user capacity becomes:
This is the user capacity restrictive condition that the utilization BD generally existed in the middle of existing research carries out the mimo system of precoding;And low
In the environment of scattering, due toSo user capacity now: | S |Low scattering≥|S|Rich scattering.If therefore for
The channel being under low-scattering environments processes according to rich scattering environments, then can affect the user capacity upper limit.
Therefore, when considering based on rich scattering environments, | S | is by antenna for base station number M and user effective reception antenna number (nr=N1
=N2==N|K|) restriction, be a definite valueAnd the user capacity upper limit | the S | under non-lipid scattering environments
Need specifically to calculate.
After step 004. assumes that the m-th user in alternative user set K replicates immigration has selected family set S, to
Select each user in family set S to carry out block diagonalization process, and calculate this and be newly added user pre-selecting family set S
Encoder matrix, further according to the pre-coding matrix of this user, calculates the equivalent channel square of this user in conjunction with the channel matrix of this user
Battle array;Equivalent channel matrix and the product matrix determinant of its conjugate matrices then according to this user, it is thus achieved that the speed of this user
Low complex degree token state;Subsequently into step 005.
In above-mentioned steps 004 and below step 005, it is required for calculating the equivalent channel matrix of user, for set S
For middle user j, along with the addition of user m, the equivalent channel matrix of the user j in S in set KOrder can change,
Thus have influence on its data rate Rj.For replicating the S in the user m in K to set S, calculate each user in S etc.
The method of effect channel matrix (step 004 and 005) is all intended to each user in set S is carried out block diagonalization process, and rightMatrix dimensionality be analyzed, and obtain the change condition of its order, and provide user rate R on this basisj。
Analyzed from step 003, VjDimension beUse nj (|S|)RepresentCan be by it
Regard as when access customer number is | S |, user j will not be produced the spatial degrees of freedom number of inter-user interference.Then dissipate in richness
Penetrate and in low scattering the true environment deposited, nj (|S|)Expression formula be:
By finding VjMakeTherefore user j is at the received signal vector of down channelBelieve with sending
Number vectorRelation be represented by:
yj=HjTjxj+nj=HjVjBjxj+nj (6)
Wherein, the equivalent channel matrix of user j isBy above to VjThe analysis of matrix dimensionality understands,ThereforeOrderIt is N for its concrete valuejAlso
It is nj (|S|), need us situation to be divided to discuss.
Assuming during carrying out user's selection algorithm, the number of users that time a certain, etching system services the most simultaneously is | S |, orderRepresent the combined channel matrix of all users, then Ne in SSRepresent combined channel matrix H eSOrder,
NeSWithBetween there is relation:
1. whenTime, due toTherefore can meet with lower inequality:
Therefore necessarily haveAccording to (4) formula, thenCertainly exist untrivialo solution.It addition, for nj (|S|)Have:
Therefore can obtain:
Now to equivalent channel matrixDo singular value decomposition to have:
(10) in formula,Be byNon-zero singular value composition Nj×NjDimension diagonal matrix;ForA left side unusual
Vector,For corresponding toIn NjThe right singular vector of individual singular value.
2. whenTime, with situation 1. shown in, due toTherefore exist:
ThereforeCertainly exist untrivialo solution.Situation 2. under, andOrder cannot be judged as nj (|S|)Or Nj, only
Can calculate user j's by concreteValue solvesAnd by nj (|S|)With NjCompare.
3. as M < NeS, andTime, the most necessarily meetThere is untrivialo solution.Due at true ring
The channel matrix H of user j under borderjVector withThere may exist dependency, therefore meet following formula:
Therefore n can be obtained by (12) formulaj (|S|)With NjRelational expression be:
nj (|S|)< Nj (13)
Therefore:
Now to equivalent channel matrixDo singular value decomposition to have:
In above formula,Be dimension be nj (|S|)×nj (|S|)Diagonal matrix,ForLeft singular vector,For right singular vector.
Can obtain from above analysis,Order be NjOr nj (|S|)Depend on NeS,And the size between M three
Contrast, is specifically summarized as follows:
Step 005. continues assuming that the m-th user in alternative user set K replicates immigration has selected family set S's
On the basis of, continue to calculate and selected in family set S in addition to this user, the equivalent channel matrix of other each users, and according to
The equivalent channel matrix of this each user, and the transition matrix of this each user, calculate the most corresponding speed of this each user
The low complex degree token state of rate variable quantity, the low complex degree then according to the most corresponding speed variable quantity of this each user characterizes
Amount, updates and has selected the speed low complex degree token state of active user in family set S, subsequently into step 006.
Step 006. continues assuming that the m-th user in alternative user set K replicates immigration has selected family set S's
On the basis of, it may be judged whether select the speed low complex degree token state of all active users in family set S to be both greater than it and preset speed
Rate low complex degree token state lower limit, is the m-th user in alternative user set K to be copied in set Q to be selected, and enter
Step 007;Otherwise abandon replicating m-th user in alternative user set K immigration and select family set S's it is assumed that and enter
Step 007.
In above-mentioned steps 004 to step 006, the speed low complex degree token state of described user, by utilizing iterative recursive
Method obtain this sign, it is assumed that simultaneously number of users that may be served be | S |, system total power be P0Base station MU-MIMO system
In, power averaging distribution on each transmit, then can be by the data rate of user jIt is expressed as:
Wherein σ2Represent white Gaussian noise power, Represent corresponding to selecting in family set S
User j'sAll nonzero eigenvalues long-pending;P0Represent that power is always launched in base station;Represent and select
User gathers the equivalent channel matrix of user j and the product matrix of its conjugate matrices, min (N in Sj,nj (|S|)) represent and select
The equivalent channel matrix of any user j in family set SOrderAnd then obtainValue is incorporated as along with new user's:
User j during (| S |+1) secondary iteration is gone to disturb pre-coding matrix Vj (|S|+1), can be by the side of iterative recursive
Method obtains this value:
Product angle concept thus according to subspace can obtainWithBetween relational expression be:
In formula, θlRepresentation spaceWithBetween l to base vector elAnd alBetween basic angle.
Space can be defined asAnd spaceBetween product angle.It is when user capacity is | S |
Low complex degree speed token state.
Additionally by above-mentioned user's j data rateExpression formula, we can be by the number after twice iteration of user j
Being analyzed according to speed difference, can obtain the changes in data rate in true environment is:
In above formula, due toSoBy analyzing, along with adding of user
Enter, the order of the equivalent channel matrix of user jOnly may decline, i.e.Therefore can obtainShould
Formula shows, system often accesses a new user, and new user can take a part of space resources of segmenting system, causes previous system
The data rate of the user accessed likely can be weakened.
Especially, when mimo channel is in rich scattering environments, (38) formula can be changed into:
Under rich scattering environments, due to the restrictive condition of user capacity, it is impossible to occurFeelings
Condition.From (36) formula we, when effective reception antenna number of user each in system is identical, (| S |+1) individual new user
Selection only can affectIn expression formula, nowThe low complex degree of speed variable quantity
Sign isAnd under non-lipid scattering environments,Expression formula be exactly (35) formula.
Step 007. has judged whether for the traversal of all users in alternative user set K, is to enter step
008;Otherwise add 1 for the value corresponding to m, and give m by this result, be then back to step 004.
Step 008. is chosen a user's immigration making user choose benchmark maximum from set Q to be selected and has been selected family to gather
In S, in alternative user set K, delete this user simultaneously by alternative user set K deletes this user simultaneously, and empty to be selected
Set Q, subsequently into step 009.
In above-mentioned steps 008, from set Q to be selected, choose a user's immigration making user choose benchmark maximum select
In family set S, in alternative user set K, delete this user simultaneously by alternative user set K deletes this user simultaneously, and clearly
Empty set Q to be selected, wherein, is respectively directed to each user q in set Q to be selected, it is thus achieved that equation below corresponding to each user
Value:And choose wherein user corresponding to maximum, it is
The user that benchmark is maximum, moves it into and selects in family set S, delete in alternative user set K simultaneously this user simultaneously by
Alternative user set K deletes this user, and empties set Q to be selected;Wherein, PavgRepresent putting down on every, base station transmitting antenna
All launch power, i.e.
Step 009. is newly added the channel matrix of user according to selecting in family set S, and has selected family to gather S Central Plains
There is the pre-coding matrix of each user, calculate and selected the transition matrix of each user original in family set S, further according to selecting
The transition matrix of each user original in family set S, updates and has selected the pre-coding matrix of each user corresponding in family set S,
It is then back to step 003.
Step 010., according to selecting the channel matrix of each user in family set S, calculates to obtain and has selected in family set S
The pre-coding matrix of each user, and utilize block diagonalization method to eliminate the interference selected in family set S between each user.
Wherein, in step 010, for any one user j in alternative user set K, pass throughEliminate other
User disturbs, for making pre-coding matrix VjThere is untrivialo solution, be then required to meet the number number less than variable of equation, it may be assumed that
Therefore for any user of access base station, only other user's combined channel ranks of matrix are less than base station and send out
Penetrate summation M of antenna number, pre-coding matrix V could be there isjEnsureing that each user is not disturbed by other users, this is also true ring
To the restriction of user capacity | S | selecting family set S when using block diagonalization technology under border.
Under rich scattering environments, due toSo having selected the upper limit of family aggregate user capacity to become:
This is the user capacity restrictive condition that the utilization BD generally existed in the middle of existing research carries out the mimo system of precoding;And low
In the environment of scattering, due toSo user capacity now: | S |Low scattering≥|S|Rich scattering.If therefore for place
Channel under low-scattering environments processes according to rich scattering environments, then can affect and select the family aggregate capacity upper limit.
MU-MIMO system true environment low complex degree user choosing method designed by technique scheme, based on truly
Wireless transmission environments (rich scattering and non-lipid scattering all exist) be analyzed, obtained user in non-lipid scattering environments and selected
The dynamic capacity upper limit, and made a concrete analysis of user's equivalent channel matrix dimension and order and select the increase of iteration to be sent out along with user
Raw change, and by this change application to user's selection reference;It addition, user's equivalent channel matrix is conjugated by the present invention with it
The nonzero eigenvalue of the matrix that matrix multiple obtains is long-pending, and i.e. low complex degree token state is as user's selection reference so that it is can be
Characterizing the admissible rate of user to a certain extent, this product value can be obtained by last iterative value recurrence, reduces algorithm
Computation complexity, simultaneously on the basis of the differentiated service QoS that becomes more meticulous, improve the effective throughput of system.
Above in conjunction with accompanying drawing, embodiments of the present invention are explained in detail, but the present invention is not limited to above-mentioned enforcement
Mode, in the ken that those of ordinary skill in the art are possessed, it is also possible on the premise of without departing from present inventive concept
Make a variety of changes.