The content of the invention
For the deficiency of existing method, the present invention proposes a kind of energy distributing method of ZF MIMO communication system, more
In the MIMO communication system environment of user, the piecewise linearity of multi-user's energy distribution can be made full use of under gross energy constraints
Efficiency preferably energy output result is redistributed in characteristic, realization, and its overall calculation amount is small, and application is wider.
To achieve the above object, the energy distributing method of ZF MIMO communication system proposed by the present invention, multi-user's
In MIMO communication system environment, efficiency preferably energy output result is redistributed under gross energy constraints, under it includes
Row step:
S1 total number of users and the antenna amount and its BTS channel status information of each user) are obtained, builds each use
The information matrix at family;
S2 all information matrixs) are subjected to singular value decomposition respectively, produce the subchannel gain value corresponding to each user;
S3 all subchannel gain values) are inputted into breakpoint function and produce plural break value;
S4) initialize the antenna energy apportioning cost of all users and establish the temporary set of plural number;
S5) take out a break value sequentially to compare with all subchannel gain values, and comparison result is updated to correspondingly
Temporary set in;
S6) according to numerical value structure efficiency function caused by abovementioned steps;
S7) break value is inputted in efficiency function as variable and produces end value, whether judged result value is more than threshold
Value, if then return to step S5, until all break values are all calculated and terminated;Step S8 is carried out if end value is less than threshold value;
S8 the root of the efficiency function) is taken out, the new day of all users is calculated by root input energy partition function
Heat input apportioning cost simultaneously exports its energy.
Further, information matrix is N × MkThe antenna amount of the matrix of dimension, wherein N expression base station, and MkRepresent k-th
The antenna amount of user.By information matrix H in step S2kSingular value decomposition is carried out to obtainWherein ()HRepresent altogether
Yoke transposition,Represent Mk×MkThe singular value matrix of dimension, λkmRepresent HkM-th of singular value, Uk
And VkN × M is represented respectivelykLeft the singular vector matrix and M of dimensionk×MkThe right singular vector matrix of dimension.
Further, step S3 interruptions point function isWherein K is
Total number of users, MkRepresent the antenna amount of k-th of user, γkmTo represent that the subchannel of m-th of antenna of k-th of user increases
Benefit value.And the quantity of break value is Q,μkDetermined by water flood, that is, solve equationIts
In (x)+=max { x, 0 },Represent the higher limit of k-th of user transmitter energy.
μ is recited as by the above-mentioned break value generated(1), μ(2)... μ(Q), and arrange in descending order.
Further, the temporary set of plural number comprises at least the first temporary set, the second temporary set and the 3rd in step S4
Temporary set, respectively with A, B and EkRepresent.One of break value is taken out in step S5 with all subchannel gain values to be compared
It is right, following judgement is carried out according to the break value:IfWherein k ' ∈ { 1,2 ..., K }, m ' ∈ { 1,2 ..., Mk,
Then the temporary collection of renewal the 3rd is combined into Ek=Ek∪{m′};Ifμ(q)=μk′, wherein k ' ∈ { 1,2 ..., K }, then it is temporary to update first
Collection is combined into A=A ∪ { k ' }, and the temporary collection of renewal second is combined into B=B/ { k ' }.
Further, the efficiency function is in step S6Wherein coefficient a,
B, c are defined as:
Wherein PrThe fixed loss of the communication system in addition to emitted energy is represented,Represent the conversion efficiency of energy.
Further, threshold value is set as 0 in step S7.
Further, the optimal solution is in step S8Its
Middle μ*For the root of the efficiency function.
The present invention through obtain multi-user MIMO communication system in total number of users and each user antenna amount and
After its BTS channel status information, via abovementioned steps, it can reach under conditions of gross energy constraint, make full use of multi-user's energy
The piece wire characteristic of distribution is measured, efficiency preferably energy output result is redistributed in realization, compared with conventional method, overall meter
Calculation amount is small, mitigates the burden of hardware device, and application is wider.
Embodiment
Fig. 1 is the step schematic diagram of the present invention.The step is overall during realization is subject to through computer program
Amount of calculation is small, and processing speed is much better than conventional method, and can therefore mitigate the burden and cost of hardware device.In figure, it will walk
Rapid S1~S8 is arranged signal, and step S5~S7 is circular treatment, and its object is to calculate to finish by all break values
New antenna energy apportioning cost is obtained, wherein step S7 judges for decision-making, caused after judging break value input efficiency function
Whether end value is more than threshold value, and according to result return to step S5 and until all break values all calculate end or into step
S8.After all calculating are completed, it will new antenna energy apportioning cost is exported to each antenna, it is optimal to reach efficiency
Distribution, further overall calculation amount is small, therefore mitigates the burden of hardware device.
Fig. 2 is the schematic flow sheet of the present invention.Step after subdivision is naturally also to be realized through computer program.
In step s 201, the antenna amount and BTS channel status information of total number of users and each user are obtained, its
In total number of users be defined as K, MkRepresent the antenna amount of k-th of user, wherein k=1,2 ..., K.
In step S202, the information matrix of each user is built, wherein information matrix is defined as Hk, information matrix HkFor N
×MkThe matrix of dimension, MkExplicitly defined in above-mentioned, and N represents the antenna amount of base station.
In step S203, all information matrixs are subjected to singular value decomposition, each information matrix H respectivelykThrough unusual
Value is decomposedWherein ()HRepresent conjugate transposition,Represent Mk×MkDimension
Singular value matrix, and λkmRepresent HkM-th of singular value, UkRepresent N × MkThe left singular vector matrix of dimension, and VkRepresent Mk×Mk
The right singular vector matrix of dimension.
In step S204, the subchannel gain value corresponding to each user is produced, wherein m-th of son of k-th of user
Channel yield value is defined as γkm.Subchannel gain value γkmCalculation it is specifically as follows, first define ZF filtering matrix
G=(UHU)-1UH, wherein U=[U1, U2..., UK], then build subchannel gain value and beWherein σ2Represent high
The variance of this white noise, ukmRepresenting matrix (UHU)-1 Individual diagonal entry, LkRepresent that k-th of user is corresponding
Electromagnetic Wave Propagation loss.
In step S205, all subchannel gain values are inputted into breakpoint function and produce plural break value, its point of interruption function
ForWherein K is foregoing total number of users, MkFor k-th foregoing of use
The antenna amount at family.And the quantity of break value is Q,In addition, the μ in breakpoint functionkDetermined by water flood, i.e.,
Solve equationWherein (x)+=max { x, 0 },Represent the upper limit of k-th of user transmitter energy
Value.
In step S206, the antenna energy apportioning cost of all users is initialized, in order to ensure follow-up computing will not be by
Mistake is produced to the influence of numerical value before, is initialized, and the value initialized is zero, further, initialization is pkm
=0, k=1,2 ..., K, m=1,2 ..., Mk, wherein pkmRepresent the energy value distributed on m-th of antenna of k-th of user.
In step S207, the temporary set of plural number is established, its temporary set is respectively the first temporary set A, second kept in
Set B and the 3rd keeps in set Ek, and the temporary set is respectivelyWherein k=1,
2 ..., K.
In step S208, counting variable q=1 is set, sequentially done to take out break value successively with subchannel gain value
Compare, its comparison method does following judgements according to the source of break value:
IfWherein k ' ∈ { 1,2 ..., K }, m ' ∈ { 1,2 ..., Mk, the temporary collection of renewal the 3rd is combined into Ek=
Ek∪{m′};
If μ(q)=μk′, wherein k ' ∈ { 1,2 ..., K }, then update the first temporary collection and be combined into A=A ∪ { k ' }, and renewal the
Two temporary collection are combined into B=B/ { k ' }.
In step S209, comparison result is updated in corresponding temporary set.
In step S210, it is according to numerical value structure efficiency function, efficiency function caused by abovementioned stepsWherein coefficient a, b, c are respectively defined as:
Wherein PrThe fixed loss of the communication system in addition to emitted energy is represented, andRepresent the conversion efficiency of energy.
In step S211, break value is inputted in efficiency function as variable and produces end value, i.e., by break value μ(q+1)
As generation end value f (μ in variable x input efficiency function f (x)(q+1)), i.e., end value according to efficiency function f (x) in x=μ(q+1)
When value size.
In step S212, whether judged result value is more than threshold value, and wherein threshold value is set as 0.When end value is more than
Threshold value, i.e. f (μ(q+1)) >=0, then return to step S208, terminates up to all break values all calculate;And when end value is less than door
Threshold value, i.e. f (μ(q+1)) < 0, then carry out step S213.
In step S213, the root of the efficiency function is taken out, that is, takes out the root of Equation f (x)=0, as solves equation f
(x) root=0, root are designated as μ*。
In step S214, root input energy partition function, the new antenna energy apportioning cost of all users is calculated, and it is defeated
Go out its energy value, wherein optimal solution isAnd above-mentioned root is
It on the premise of energy efficiency priority is optimum solution to be, after the completion of after the calculating of energy apportioning cost, its energy apportioning cost will be exported to day
Line.
In the above description, step S201~S207 is that step S1~S4 in Fig. 1 is described in detail, and step S208
~S214 is that step S5~S8 in Fig. 1 is described in detail, is elaborated for each definition that each step is previously mentioned, thereby
Correlation between being more readily understood by.
Fig. 3 is a kind of step module map of schematically embodiment aspect of the present invention, specific and simplify the skill for showing the present invention
Art emphasis, make more quick and clear and definite in overall understanding.
In step S301, the information matrix H of each user is builtk, the preposition work that is previously mentioned in the step S1 in as Fig. 1
Industry is completed in the lump, the follow-up required information matrix H used of outputk。
In step S302, by each information matrix HkSingular value decomposition is carried out, produces the subchannel corresponding to each user
Yield value γkm, what is be previously mentioned in the step S2 in as Fig. 1 seeks subchannel gain value γkm, for making during follow-up calculating break value
With.
In step S303, Q breakpoint is calculated, and is arranged in descending order, is designated as μ respectively(1), μ(2)... μ(Q), its object is to
Descending is adopted to be arranged with beneficial to follow-up calculating processing.
In step S304, the antenna energy apportioning cost of all users is initialized as 0, establishes temporary set A, B, Ek.Such as
With foregoing, the energy apportioning cost of antenna is initialized as 0, is value interference before avoiding, and establishes following cycle calculating and is used
Temporary set A, B, the E arrivedk, and the counting variable q of circulation is set to 1.
In step S305, according to break value μ(q)Value source, corresponding renewal temporary set A, B, Ek。
In step S306, temporary set A, B, E for being updated according to step S305kBuild efficiency function f (x).
In step S307, judge whether f (μ(q+1)) >=0, is whether judged result value is more than threshold value, if greater than setting
Fixed threshold value, then it is return to step S305 after q=q+1 to update counting variable, if less than threshold value, then carries out step
S308。
In step S308, the root of Equation f (x)=0 is taken out, as solves equation the root of f (x)=0, root is designated as μ*.So
Afterwards, root input energy partition function, the new antenna energy apportioning cost of all users is calculated, and exports its energy value, wherein can
Measuring partition function isAnd above-mentioned root is the premise of energy efficiency priority
It is down optimum solution, after the completion of after the calculating of energy apportioning cost, its energy apportioning cost will be exported to antenna.In other words, it is
The optimal antenna energy apportioning cost of efficiency is achieved, and result is partitioned energy into each antenna.
Herein for convenience of explanation, in figure 3 not repeatedly be confirmed whether complete all break values, only
Whether illustrated for f (x) value less than this part of threshold value 0, thus can be understood as single treatment just complete it is all
Antenna energy apportioning cost calculates.
Fig. 4 be carry out 1000 Monte Carlo Experiments under the conditions of, when transceiver distance is changed by 0.1 to 1.5 kms, this hair
The bright energy valid value comparison schematic diagram obtained with conventional method.In order to illustrate the performance of this method, it is assumed that there is a multiuser MIMO
Communication system, number of users K=3, antenna for base station quantity N=6, the antenna amount of each user is Mk=2, k=1,2 ...,
K.Each element of channel state information matrix obeys the independent Gaussian distribution of zero mean unit variance, and the band of system is a width of
10MHz, noise power are -130dBm/Hz, the conversion efficiency of energyFixed loss Pr=140dBm, the energy of each user
Amount constraints is that the transmitter energy upper limit is 0.15W, and the propagation loss of electromagnetism is 128.1+37.61g (d), and wherein d represents to receive
The distance between hair machine, its unit are kms.
According to the multi-user MIMO communication system of above-mentioned hypothesis, it is as follows to carry out energy distributing method of the invention shown in Fig. 3:
1) in step S301, reception and transmission range d=0.1 kms, it is assumed that estimate the channel status letter of each user obtained
Ceasing matrix is:
2) in step s 302, by matrix H1, H2, H3Singular value decomposition is carried out, produces the subchannel gains of each user
Value, is obtained:
γ11=0.6531, γ12=0.5958
γ21=1.9467, γ22=1.8082
γ31=1.6445, γ32=1.0933
3) in step S303, it can calculate and obtain 8 break values:
Arrange in descending order, obtain 8 break values:
μ(1)=0.5137, μ(2)=0.5530,
μ(3)=0.6081, μ(4)=0.6084,
μ(5)=0.7581, μ(6)=1.5311,
μ(7)=1.6783, μ(8)=1.6797
4) it is zero in step s 304, to initialize all energy apportioning costs, i.e. p11=p12=p21=p22=p31=p32=
0;
The temporary set of initializationCounting variable q=1 is set.
5) in step S305, as q=1,Therefore, temporary set E is updated2={ 1 }.
6) in step S306, efficiency function is builtWherein a, b, c's
Value is respectively:
A=0.9611
B=-0.1137
C=1
7) in step S307, because f (μ(2))=f (0.5530)=1.0396 > 0, therefore q=2 is made, and return to step
S305。
8) in step S305, as q=2,Therefore, temporary set E is updated2={ 1,2 }.
9) in step S306, efficiency function is builtWherein a, b, c's
Value is respectively:
A=1.8156
B=-0.6667
C=2
10) in step S307, because f (μ(3))=f (0.6081)=1.8465 > 0, therefore q=3 is made, and return to step
Rapid S305.
11) in step S305, as q=3,Therefore, temporary set E is updated3={ 1 }.
12) in step S306, efficiency function is builtWherein a, b, c's
Value is respectively:
A=2.5332
B=-1.2748
C=3
13) in step S307, because f (μ(4))=f (0.6084)=2.7681 > 0, therefore q=4 is made, and return to step
Rapid S305.
14) in step S305, as q=4, μ(q)=0.6084=μ2, therefore, update temporary set A={ 2 } and B=
{ 1,3 }.
15) in step S306, efficiency function is builtWherein a, b, c's
Value is respectively:
A=1.0992
B=-0.0581
C=1
16) in step S307, because f (μ(5))=f (0.7581)=0.6324 > 0, therefore q=5 is made, and return to step
Rapid S305.
17) in step S305, as q=5, μ(q)=0.7581=μ3, therefore, update temporary set A={ 2,3 } and B
={ 1 }.
18) in step S306, efficiency function is builtWherein a, b, c
Value be respectively:
A=0.6997
B=0.7000
C=0
19) in step S307, because f (μ(6))=f (1.5311)=0 >=0, therefore q=6 is made, and return to step
S305。
20) in step S305, as q=6,Therefore, temporary set E is updated1={ 1 }.
21) in step S306, efficiency function is builtWherein a, b, c's
Value is respectively:
A=0.0851
B=-0.8311
C=1
22) in step S307, because f (μ(7))=f (1.6783)=- 0.1038 < 0, carry out step S308.
23) in step S308, the root of Equation f (x)=0 is solved, obtains root μ*=1.5311, the distribution of output energy is most
Result is afterwards:
p11=0, p12=0,
p21=0.0947, p22=0.0553,
p31=0.1500, p32=0
It is to carry out step S305~S307 cycle calculations that step 5 is can see in above-mentioned to 22.And pass through the present invention
Energy distributing method, the good and bad difference with conventional method can be will become apparent from by the curvilinear motion in Fig. 4, realizes antenna energy energy
Imitate the allocative decision of optimization.
Embodiments of the invention are the foregoing is only, are not intended to limit the invention, it is all in the spirit and principles in the present invention
Within, any modification, equivalent substitution and improvements made etc., it should be included in the scope of the protection.