CN103354658A - Energy efficiency optimization power distribution algorithm of multi-cell multiple-input multiple-output system considering same frequency interference suppression - Google Patents

Abstract

The invention provides an energy efficiency optimization power distribution algorithm of a multi-cell multiple-input multiple-output (MIMO) system considering same frequency interference suppression. The algorithm comprises: A) initializing; B) sorting sub-channels; C)determining high-power channel sets and low-power channel sets; D) and distributing power for each sub-channel. The algorithm is advantaged in that in a MIMO multi-cell system, optimization of energy efficiency by using a two-element collection algorithm is superior to a conventional average power algorithm. Even though when the number of the sub-channels is relatively low, the optimized performance is better than average power optimization algorithm.

Description

Many residential quarters multi-input multi-output system is considered the optimized for energy efficiency power distribution algorithm that co-channel interference suppresses

Technical field

The present invention relates at many residential quarters multiple-input and multiple-output (Multi-input Multi-output; MIMO) consider the optimized for energy efficiency power distribution algorithm that co-channel interference suppresses in the system.

Background technology

In the MIMO communication system, especially when carrying out Space Time Coding, decoding, want to estimate very accurately the current sub-channel information of channel, and analyze accurately the performance of system, so high efficiency this Important Parameters of frequency domain response coefficient that utilizes is very crucial.

Development of Wireless Communications is to today, and people improve constantly for the speed of Wireless Data Transmission and the requirement of its service quality.In contrast to mainly in order to transmitting first and second wireless communication system of speech business in generation, third and fourth generation the multimedia broad band data service rate that can support of wireless communication system significantly improve.So high system bandwidth has just proposed requirements at the higher level to the channeling efficient of many minizones.Full rate is multiplexing to be the strongest technology that satisfies the demand in theory, yet under the multiplexing environment of full rate, the interference between the residential quarter will have a strong impact on transmission rate and its QoS of customer of its communication system.So in the system of next generation mobile communication, generally adopted SC-FDMA, OFDM and multiple access technique, guaranteeing the orthogonality between the user within the residential quarter, and interference that can inside, establishment residential quarter.

Interference between the residential quarter also just becomes the main interference source of future mobile communication system.In MIMO multi cell cellular system, because the characteristic that ICI disturbs, the user who is positioned at cell edge is subject to the seriousness that ICI disturbs will be higher than non-edge customer, so that the throughput performance of system reduces, does not usually reach its user's communication requirement.

To sum up, how can suppress better the interference of minizone, and the energy efficiency of elevator system and Cell Edge User will be one very meaningful, simultaneously also be the problem that extremely has challenge.

The interference mitigation technology of conventional cell mainly comprises interference elimination, interference coordination, interference randomization etc.Interference cancellation techniques is obtained at the form of signal, is had a lot of defectives synchronously and aspect the distribution of resource, and Receiver Design is very complicated.And the effect of interference randomization technology in practice is very limited.

So mainly utilize the interference coordination technique of minizone to improve total energy efficiency in the Cell Edge User system.

But the cutting procedure of carrier wave is to carry out within the communication system of whole MIMO, rather than in some residential quarters, adjust separately, the expense that this just needs a lot of extra signalings may cause the efficient of system seriously to reduce, and has also played very large negative interaction for the performance that improves system.Simultaneously, in the interval in each transmission time, the subcarrier that carries out is cut apart the propagation delay time that also can increase system.

So for the coordination technique of above-mentioned interference, the efficient of its frequency spectrum is lower.The scheme of seeking novel many coordinating district interferences is necessary.

Summary of the invention

The method according to this invention only need to be known the channel condition information of current subchannel, even can know the subchannel number of current system activity, so just greatly reduces the complexity of algorithm.

According to an aspect of the present invention, by the modeling to the many cell environments of MIMO, analyze its minizone co-channel interference characteristic, considering to optimize efficiency of energy utilization in the situation that the minizone co-channel interference suppresses, save the communication resource.

According to an aspect of the present invention, the optimized for energy efficiency power distribution algorithm that provides a kind of many residential quarters multi-input multi-output system of knowing clearly to consider that co-channel interference suppresses is characterized in that comprising:

A) initialization;

B) sort to subchannel;

C) determine powerful channel set

With low power channel set

D) distribute power for every sub-channels.

Description of drawings

Fig. 1 has shown the system block diagram that can use MIMO communication system according to an embodiment of the invention.

Fig. 2 and 3 has shown the transmitting-receiving model of demixing time space.

Fig. 4 has shown the variation tendency along with the growth of subchannel number M.

Fig. 5 be according to an embodiment of the invention, based on the flow chart of the power control algorithm of energy efficiency.

Fig. 6 and 7 is the relevant emulation of embodiment.

The mimo system model

Suppose that a MIMO communication system has n in adopting the discrete-time system model _TTransmit antennas, n _RRoot reception antenna, system block diagram are as shown in Figure 1.

In Gaussian channel, at the signal that transmitting terminal sends, it submits to Gaussian Profile.The covariance matrix that transmits is:

R _xx＝E{xx ^H} (1.1)

Wherein x is the independent synchronously distribution gaussian variable of zero-mean.

A ^HHermitian (Hermitian) transposed matrix of expression A matrix, i.e. the complex-conjugate transpose matrix of A.Wherein, E{.} represents average; Number of transmit antennas M is no matter how much numerical value is, following formula can be expressed as:

P＝tr(R _xx) （1.2)

Wherein, tr (A) represents the mark of matrix A; P is the restriction of total transmitting power; Tr (A) can use the diagonal element of summation A usually to obtain.

If, at the signal of each antenna transmission of supposition, its power P/n _TAll equate, and at transmitting terminal, channel condition information (Channel State Information, CSI) is unknown.

The covariance matrix that transmits is:

$R_{\mathrm{xx}}=\frac{P}{n_T}{I}_{n_T}---\left(1.3\right)$

Wherein,

N _T* n _TThe unit square.Complex matrix H n _T* n _{R_}Channel is described.The noise n of receiving end is n _R* 1 column vector.In matrix H, use h _IjRepresent the i * j element, its concrete meaning is the coefficient of the channel fading from the j transmit antennas to i root reception antenna.It is to add up independently zero-mean, has independently, variance equates real part and the imaginary part of multiple Gaussian random variable.

Be used for representing to receive the covariance matrix of noise, receiving signal y is n _R* 1 column vector.Use linear model, can be shown as receiving vector table:

y＝Hx+n （1.4）

Use E{yy ^HDefinition receives the covariance matrix of signal, so tr (R _Yy) represent total received signal power; The covariance matrix that receives signal is expressed as:

R _yy＝HR _xxH ^H+R _nn （1.5）

The energy efficiency model

In a communication system, definition according to the Shannon-Hartley theorem, no matter be Ad-hoc network or the mobile network of cellular system, no matter be the communication system of single antenna or the communication system of many antennas, suppose that transmitting signal is that Time Continuous is unlimited, and the complexity that detects is regarded as certain, all available following maximum size equation expression and calculate maximum size all be higher than the communication network of any reality ^[20], specific as follows:

R<Blog ₂(1+γ _s) （2.1）

The capacity of this equation expression is Gaussian distributed, and wherein, B can represent the available bandwidth of this communication system, and the signal to noise ratio snr of the receiving terminal of system (Signal-to-Noise Ratio) can be used γ _sThe capacity that represents this system represents that with R its unit is: bps(bits per second).

Take top formula as the basis, spectrum efficiency is important in our research process variable, especially in the design and optimization of whole system.We can further study: a ratio that allows the ratio of the capacity of system and its bandwidth obtain just is defined as the spectrum efficiency of this system.Present stage, along with we progressively go deep into the research of system optimization aspect, a new efficiency Model is concerned, and this efficient is exactly " energy efficiency ".List of references definition energy efficiency is: the bit number (TNEU:the number of bits per thermal noise energy unit) that the per unit energy is produced is defined as a model v who weighs energy efficiency _p, it is user side spectrum efficiency v _bWith user side Signal to Interference plus Noise Ratio γ _sRatio:

V wherein _b=log ₂(1+ γ _s) (2.2)

From above-mentioned formula, because we wish and can observe in mobile communication system with more blunt more intuitively angle, the every Bit data of transmission, its unit of the energy that its system will consume is (cost per bit), is that angle from the user defines so we will define transmitting energy efficient.As follows:

$\mathrm{\&eta;}=f\left(P_i\right)=\frac{\mathrm{\&Sigma;C}\left(p_i\right)}{\mathrm{\&Sigma;}{p}_i}---\left(2.3\right)$

Wherein, in many residential quarters, the spectrum efficiency of i residential quarter can be used C (p _i) represent; In cell mobile communication systems, the transmitted power of i cell base station can be used p _iThe unit that represents him is watt (Watt), above-mentioned formula has clearly been expressed a kind of relation of input and output, be specifically described as follows: drop on the output ratio of the transmitted power of base station and just can be expressed as energy efficiency η, his unit is bps/hertz/watt (bits/s/Hz/W); From above-mentioned formula, we can find out that base-station transmitting-power is argument of function, and minute subrepresentation output: it is all spectrum efficiency sums of system, and denominator represents to drop into: represent with all base-station transmitting-power sums.

The optimization of many cell energies of MIMO

Two-element set

As a rule, all be arbitrarily value, i.e. P in interval [P min, P max] for the value of the power of power division _Min≤ P≤P _Max, and the two-element set two-element set refers to only have two elements, or namely:

p _i∈ Ω ^B, Ω ^B={ p|p=p _MinOr p=p _Max(3.1)

According to prior art, show that according to the analysis result of mathematics exist when system under many channel situation, two-element set is applied in the optimization of systematic function, although its resulting result is not strict global optimization, also be very close to the optimization of the overall situation.

But, take different systematic functions as optimizing purpose, how better to use the two-element set method, remain a problem.If the N subchannel is arranged in the system, use two-element set, under the condition of not adding any constraint, may will the mode of plurality of distribution appear.

For target and the noisy communication channel of the given error rate (BER), the energy of signal is diffused into the performance that (the sub-carrier frequency of frequency division multiplexing) on all possible time (time division multiplexing subchannel) and the bandwidth dimension can improve system.

Will be to the gross power mean allocation to 1 of fixed system, 2, when 4 and 8 channels, and suppose that the situation of every channel is identical, the output of system-total capacity situation of change.Can find out that the capacity of system has growth along with the increase of the number of channel that is assigned with.

The thresholding value of decision gate

Consider many cell mimo-ofdm systems, the dual-mode antenna number of system is respectively n _RAnd n _TTotal spectrum efficiency is C _Total, total transmitted power is P _Total, the relational expression between them is:

$C_{\mathrm{total}}=\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{\log }_2(1+\frac{P_{\max 1}}{n_i}{\left|\right|H_i\left|\right|}_F^2)$

${\mathrm{\&eta;}}_{\mathrm{total}}=C_{\mathrm{total}}/P_{\mathrm{total}}=\frac{\underset{i=1}{\overset{M}{\mathrm{\&Sigma;}}}{\log }_2(1+\frac{P_{\max 1}}{n_i}{\left|\right|H_i\left|\right|}_F^2)}{p_{\mathrm{total}\max }+\underset{j=1}{\overset{N}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{M}{\mathrm{\&Sigma;}}}{p}_{j,i}{\left|\right|H_{j,i}\left|\right|}_F^2}---\left(3.2\right)$

$P_{\mathrm{total}}=p_{\mathrm{total}\max }+\underset{j=1}{\overset{N}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{M}{\mathrm{\&Sigma;}}}{p}_{j,i}{\left|\right|H_{j,i}\left|\right|}_{\mathrm{<figure-callout\; id="2"\; label="F"\; filenames="HDA00003441913200022.png"\; state="\{\{state\}\}">F</figure-callout>}}^2$

Wherein $P_{\max 1}=\frac{P_{\max \mathrm{total}}}{M},$ ${\mathrm{<figure-callout\; id="2"\; label="P\; max"\; filenames="HDA00003441913200022.png"\; state="\{\{state\}\}">P</figure-callout>}}_{\max }=\frac{P_{\max \mathrm{total}}}{M-1}$

Wherein expression: the transmitted power P that subchannel i is assigned to _iExpression; Local noise power n _iN represents the adjacent cell number, and M represents residential quarter subchannel number, and j represents the sequence number of adjacent cell;

Expression: the transmission matrix norm of subchannel i, wherein, H _iBe n _T* n _RMatrix.

And, consider to exist orthogonality and hypothesis user side to take the ZFBF strategy between the subchannel of MIMO-OFDM system that then the interference between each sub-channels is eliminated.

, be located in the system that comprises set CH here _iIn comprise whole N sub-channels:

${\mathrm{CH}}_i\&Element;K,K=\left\{\mathrm{CH}\right|\mathrm{CH}=\underset{i=1}{\overset{N}{\&cup;}}{\mathrm{CH}}_i\}---\left(3.3\right)$

Adopt the method for two-element set, distribute the power of every sub-channels

P _i∈Ω ^B，Ω ^B＝{P|P＝P _minorP＝P _max} （3.4）

Above two-element set has been done detailed introduction, we will split into two son set to the set K of channel with the principle of two-element set, and the relation of these two son set is supplementary sets each other.And these two set are distributed two power, maximum power and minimum power:

At powerful channel set

In, including only a performance number is P _MaxAt low power channel set

In, including only a performance number is P _Min, as follows:

${\mathrm{CH}}_i\&Element;K_{p\max }^M,K_{p\max }^M=\left\{\mathrm{CH}\right|\mathrm{CH}=\underset{i=1}{\overset{M}{\&cup;}}{\mathrm{CH}}_i,P_i=P_{\max },M\&le;N\}$ （3.5）

${\mathrm{CH}}_i\&Element;K_{p\min }^{N-M},K_{p\min }^{N-M}=\left\{\mathrm{CH}\right|\mathrm{CH}=\underset{i=M+1}{\overset{N}{\&cup;}}{\mathrm{CH}}_i,P_i=P_{\min },M\&le;N\}$

Suppose powerful channel set

In the power summation of each channel be P _Max, low power channel set

In the power summation of each channel be P _Min, the relation between each power and power can be set up an equation group, concerns as follows:

$\left\{\begin{array}{c}{P}_{\mathrm{total}}=P_{\mathrm{total}\max }+P_{\mathrm{total}\min }\\ P_{\mathrm{total}\max }=M\&times;P_{\max }\\ P_{\mathrm{total}\min }=(N-M)\&times;P_{\min }\end{array}\right.---\left(3.6\right)$

We can find out according to the formula of top, want to solve equation group, only need can find the solution by definite any 3 variablees wherein from 6 variablees.

According to condition, we can provide a constraints, and it is constraint about the gross power aspect of system: total consumed power P _TotalBe definite value (definite variable).And as long as determine We just can know small-power collection and total power collection that we are required, because

With

It is complementary each other relation.

We will meanwhile, also will analyze according to the concrete condition of each subchannel from the algorithm about the energy efficiency aspect, judge comprehensively whether it can be assigned to high-power P _MaxWith channel set In.

Exist simultaneously the condition of second restriction in the set:

The condition of first restriction of this algorithm is: in set, and the consumed power P that channel is total _TotalValue be certain (and this has exactly just generated the condition of second restriction).

We will be from the angle of energy efficiency, derive formula and the algorithm of concrete energy efficiency, that is: we will screen one by one to each subchannel, judge whether he meets the condition of high-power collection, concrete grammar is as follows, we will guarantee that the energy efficiency after the adding of this sub-channels is not less than the energy efficiency summation before adding at least, and we could add set with it

In go, otherwise this sub-channels can only be put into set

To sum up, will can obtain following equation group after the constraint equation arrangement recited above:

Restrictive condition 1:p _TatalBe definite value, that is:

M is set

The sub-channels number

Restrictive condition 2: η 1 _{I ∈ N}〉=η 2 _{I ∈ N, i ≠ k}, that is:

$\left\{\begin{array}{c}{\mathrm{\&eta;}1}_{i\&Element;N}=\frac{B\underset{i=1}{\overset{M}{\mathrm{\&Sigma;}}}{\log }_2(1+\frac{P_{\max 1}}{n_i}{\left|\right|H_i\left|\right|}_F^2)}{p_{\mathrm{total}\max }+\underset{j=1}{\overset{N}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{M}{\mathrm{\&Sigma;}}}{p}_{j,i}}\\ {\mathrm{\&eta;}2}_{i\&Element;N}=\frac{B\underset{i=1,i\&NotEqual;k}{\overset{M}{\mathrm{\&Sigma;}}}{\log }_2(1+\frac{P_{\max 2}}{n_i}{\left|\right|H_i\left|\right|}_F^2)}{p_{\mathrm{total}\max }+\underset{j=1}{\overset{N}{\mathrm{\&Sigma;}}}\underset{i=1,i\&NotEqual;k}{\overset{M}{\mathrm{\&Sigma;}}}{p}_{j,i}}\end{array}\right.---\left(3.7\right)$

Wherein $P_{\max 1}=\frac{P_{\mathrm{total}\max }}{M},$ ${\mathrm{<figure-callout\; id="2"\; label="P\; max"\; filenames="HDA00003441913200022.png"\; state="\{\{state\}\}">P</figure-callout>}}_{\max }=\frac{P_{\mathrm{total}\max }}{M-1}.$

Among this inequality, the left side of the sign of inequality is illustrated in the above In energy efficiency during total M sub-channels (comprising subchannel k), the expression of inequality the right

In the energy efficiency of (not comprising subchannel k) when the M-1 sub-channels is arranged.

The physical significance of this formula is: subchannel k only in the situation that whole set energy efficiency is reduced, just can join

Can find out by following formula, the algorithm that only directly calculate according to this inequality, the condition that know is:

1: the subchannel matrix norm of every sub-channels

2: the local noise power n of every sub-channels _i

3:

The number of subchannels M(that has had supposes that the subchannel number of each residential quarter is identical);

Therefore, because the gross power P that exists _TotalmaxUnder this constraints of definite value, top inequality just can abbreviation be:

$\frac{B\underset{i=1}{\overset{M}{\mathrm{\&Sigma;}}}{\log }_2(1+\frac{P_{\max 1}}{n_i}{\left|\right|H_i\left|\right|}_F^2)}{P_{\mathrm{total}\max }+\underset{j=1}{\overset{N}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{M}{\mathrm{\&Sigma;}}}{p}_{j,i}}\&GreaterEqual;\frac{B\underset{i=1,i\&NotEqual;k}{\overset{M}{\mathrm{\&Sigma;}}}{\log }_{2(1+\frac{P_{\max 2}}{n_i}{\left|\right|H_i\left|\right|}_F^2)}}{P_{\mathrm{total}\max }+\underset{j=1}{\overset{N}{\mathrm{\&Sigma;}}}\underset{i=1,i\&NotEqual;k}{\overset{M}{\mathrm{\&Sigma;}}}{p}_{j,i}}---\left(3.8\right)$

When M was enough large, we can ignore other adjacent cells to the interference power that subchannel k produces, and just obtain

$\underset{j=1}{\overset{N}{\mathrm{\&Sigma;}}}\underset{i=1,i\&NotEqual;k}{\overset{M}{\mathrm{\&Sigma;}}}{p}_{j,i}\&ap;\underset{j=1}{\overset{N}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{M}{\mathrm{\&Sigma;}}}{p}_{j,i}---\left(3.9\right)$

Again because

$\underset{i=1}{\overset{M}{\mathrm{\&Sigma;}}}{\log }_2(1+\frac{P_{\max 1}}{n_i}{\left|\right|H_i\left|\right|}_F^2)={\mathrm{<figure-callout\; id="2"\; label="log"\; filenames="HDA00003441913200022.png"\; state="\{\{state\}\}">log</figure-callout>}}_2\underset{i\&Element;M}{\mathrm{\&Pi;}}(1+\frac{P_{\max 1}}{n_i}{\left|\right|H_i\left|\right|}_F^2)$

$={\log }_2(1+\frac{P_{\max 1}}{n_k}{\left|\right|H_k\left|\right|}_F^2)+{\mathrm{<figure-callout\; id="2"\; label="log"\; filenames="HDA00003441913200022.png"\; state="\{\{state\}\}">log</figure-callout>}}_2\underset{i\&Element;M,i\&NotEqual;k}{\mathrm{\&Pi;}}(1+\frac{P_{\max 1}}{n_i}{\left|\right|H_i\left|\right|}_F^2)---\left(3.10\right)$

$\underset{i=1}{\overset{M}{\mathrm{\&Sigma;}}}{\log }_2(1+\frac{P_{\max 2}}{n_i}{\left|\right|H_i\left|\right|}_F^2)={\mathrm{<figure-callout\; id="2"\; label="log"\; filenames="HDA00003441913200022.png"\; state="\{\{state\}\}">log</figure-callout>}}_2\underset{i\&Element;M,i\&NotEqual;k}{\mathrm{\&Pi;}}(1+\frac{P_{\max 2}}{n_i}{\left|\right|H_i\left|\right|}_F^2)$

Then former inequality is derived as follows:

$\frac{B\underset{i=1}{\overset{M}{\mathrm{\&Sigma;}}}{\log }_2(1+\frac{P_{\max 1}}{n_i}{\left|\right|H_i\left|\right|}_F^2)}{P_{\mathrm{total}\max }+\underset{j=1}{\overset{N}{\mathrm{\&Sigma;}}}\underset{i=1}{\overset{M}{\mathrm{\&Sigma;}}}{p}_{j,i}}\&GreaterEqual;\frac{B\underset{i=1,i\&NotEqual;k}{\overset{M}{\mathrm{\&Sigma;}}}{\log }_{2(1+\frac{P_{\max 2}}{n_i}{\left|\right|H_i\left|\right|}_F^2)}}{P_{\mathrm{total}\max }+\underset{j=1}{\overset{N}{\mathrm{\&Sigma;}}}\underset{i=1,i\&NotEqual;k}{\overset{M}{\mathrm{\&Sigma;}}}{p}_{j,i}}$

$\&DoubleDownArrow;$

${\log }_2(1+\frac{P_{\max 1}}{n_k}{\left|\right|H_k\left|\right|}_F^2)+{\mathrm{<figure-callout\; id="2"\; label="log"\; filenames="HDA00003441913200022.png"\; state="\{\{state\}\}">log</figure-callout>}}_2\underset{i\&Element;M,i\&NotEqual;k}{\mathrm{\&Pi;}}(1+\frac{P_{\max 1}}{n_i}{\left|\right|H_i\left|\right|}_F^2)\&GreaterEqual;{\log }_2\underset{i\&Element;M,i\&NotEqual;k}{\mathrm{\&Pi;}}(1+\frac{P_{\max 2}}{n_i}{\left|\right|H_i\left|\right|}_F^2)---\left(3.11\right)$

$\&DoubleDownArrow;$

${\log }_2(1+\frac{P_{\max 1}}{n_k}{\left|\right|H_k\left|\right|}_F^2)\&GreaterEqual;{\log }_2\frac{\underset{i\&Element;M,i\&NotEqual;k}{\mathrm{\&Pi;}}(1+\frac{P_{\max 2}}{n_i}{\left|\right|H_i\left|\right|}_F^2)}{\underset{i\&Element;M,i\&NotEqual;k}{\mathrm{\&Pi;}}(1+\frac{P_{\max 1}}{n_i}{\left|\right|H_i\left|\right|}_F^2)}$

In the following formula, this moment, we supposed:

Local noise power n _iBe a less value, and all equate; N is the adjacent cell number; M is residential quarter subchannel number; I represents subchannel; J represents adjacent cell;

Because $P_{\max 1}=\frac{P_{\mathrm{total}\max }}{M},$ ${\mathrm{<figure-callout\; id="2"\; label="P\; max"\; filenames="HDA00003441913200022.png"\; state="\{\{state\}\}">P</figure-callout>}}_{\max }=\frac{P_{\mathrm{total}\max }}{M-1}$

Thus inequality just can further become for:

$(1+\frac{P_{\max 1}}{n_k}{\left|\right|H_k\left|\right|}_F^2)\&GreaterEqual;\frac{\underset{i\&Element;M,i\&NotEqual;k}{\mathrm{\&Pi;}}(1+\frac{P_{\mathrm{total}}}{M-1}{\left|\right|H_i\left|\right|}_F^2)}{\underset{i\&Element;M,i\&NotEqual;k}{\mathrm{\&Pi;}}(1+\frac{P_{\mathrm{total}}}{M}{\left|\right|H_i\left|\right|}_F^2)}---\left(3.12\right)$

For a given sub-channels i, the matrix norm of its subchannel

All equate, so we can obtain:

$(1+\frac{P_{\max 1}}{n_k}{\left|\right|H_k\left|\right|}_F^2)\&GreaterEqual;\frac{\underset{i\&Element;M,i\&NotEqual;k}{\mathrm{\&Pi;}}\left(\frac{1}{M-1}\right)}{\underset{i\&Element;M,i\&NotEqual;k}{\mathrm{\&Pi;}}\left(\frac{1}{M}\right)}---\left(3.13\right)$

$\&DoubleDownArrow;$

$(1+\frac{P_{\max 1}}{n_k}{\left|\right|H_k\left|\right|}_F^2)\&GreaterEqual;{\left(\frac{M}{M-1}\right)}^{M-1}$

Like this, by this algorithm abbreviation be: as long as know current

The number of subchannels M that exists and the state information that newly arrives of its subchannel just can be judged power division.

When M is tending towards infinite,

Limiting value, for:

$\underset{M\&RightArrow;\&infin;}{\lim }{\left(\frac{M}{M-1}\right)}^{M-1}=e---\left(3.14\right)$

The threshold value expression formula

Along with the variation tendency of the growth of subchannel number M as shown in Figure 4, when changing in less interval of subchannel, eg. is in interval [0,10], it is very fast that threshold value increases, but when number of subchannels during above some, the growth of threshold value is just very slow.When the channel number of system rose to 40 by 2, threshold value had just risen near limiting value e from 2, is the growth along with the subchannel number, monotonic increase, as shown in Figure 4:

To sum up, just draw subchannel k and be selected into set

Threshold condition:

When existing number of subchannels M 〉=35, then we can regard as:

$\underset{M\&RightArrow;\&infin;}{\lim }{\left(\frac{M}{M-1}\right)}^{M-1}=e$

Threshold condition 1 is:

Former formula $(1+\frac{P_{\max 1}}{n_k}{\left|\right|H_k\left|\right|}_F^2)\&GreaterEqual;{\left(\frac{M}{M-1}\right)}^{M-1}$ Become:

$({\mathrm{SNR}}_k=\frac{P_{\max 1}}{n_k}{\left|\right|H_k\left|\right|}_F^2)\&GreaterEqual;e-1,$

or （3.16）

${\mathrm{\&eta;}}_k=\frac{{\log }_2(1+{\mathrm{SNR}}_k)}{P_{\max 1}}\&GreaterEqual;\frac{{\log }_{2}e}{P_{\max 1}}$

When existing number of subchannels M＜35, threshold condition 2 is:

$({\mathrm{SNR}}_k=\frac{P_{\max 1}}{n_k}{\left|\right|H_k\left|\right|}_F^2)\&GreaterEqual;e-1,$

or （3.16）

${\mathrm{\&eta;}}_k=\frac{{\log }_2(1+{\mathrm{SNR}}_k)}{P_{\max 1}}\&GreaterEqual;\frac{{\log }_{2}e}{P_{\max 1}}$

Therefore, just can select according to this condition, we only need know the channel condition information of current subchannel, even can know the number of the subchannel of current active, have so just reduced greatly the complexity of this algorithm.

Algorithm flow

As above-mentioned, as shown in Figure 5, according to an embodiment of the invention, comprise based on the power control algorithm of energy efficiency:

At first, K ^NIn the N subchannel with descending arranged sequentially, the restrictive condition whether wherein i subchannel satisfies decision gate is calculated in beginning successively, last 1 i value that satisfies condition is just for gathering

In high-power subchannel number M(determined the 3rd variable).

The concrete steps of power control algorithm are as follows:

Step 1

Initialization:

1) determines the gross power p of system _Total, and

In every sub-channels power and be p _TotalmaxLow power set of sub-channels

In every sub-channels power and be p _Totalmin

2) the set K of generation N sub-channels: $K=\left\{\mathrm{CH}\right|\mathrm{CH}{=}^{\&prime;}\underset{i=1}{\overset{N}{\&cup;}}{\mathrm{CH}}_i\};$

3) powerful set of sub-channels

And low power set of sub-channels

For undetermined: $K_{p\max }^M=\mathrm{\&Phi;}$ (empty set); With $K_{p\min }^{N-M}=\mathrm{\&Phi;}$ (empty set).

Step 2

Sort to subchannel:

1) the N sub-channels of set among the K according to the descending arrangement of the norm value of its channel, as follows:

$\tilde{K}=\left\{\mathrm{CH}\right|\mathrm{CH}{=}^{\&prime;}\underset{i=1}{\overset{N}{\&cup;}}{\mathrm{CH}}_i;\&ForAll;(1\&le;i\&le;k\&le;N):{\left|\right|H_i\left|\right|}_{\mathrm{<figure-callout\; id="2"\; label="F"\; filenames="HDA00003441913200022.png"\; state="\{\{state\}\}">F</figure-callout>}}^2\&GreaterEqual;{\left|\right|H_k\left|\right|}_F^2\};$

Step 3

Determine set With

Be directed to every sub-bar channel CH for i=1:N( _i∈ K)

1) establishes $p_{\max }=\frac{p_{\mathrm{total}\max }}{i};$

2) for the i subchannel.

2.1) if i 〉=35 judge then whether its Signal to Interference plus Noise Ratio is satisfied with the condition 2 of thresholding:

$({\mathrm{SNR}}_i=\frac{p_{\max 1}}{n_i}{\left|\right|H_i\left|\right|}_F^2)\&GreaterEqual;e-1;$

2.2) if i＜35 judge then whether its Signal to Interference plus Noise Ratio is satisfied with the condition 1 of thresholding:

$({\mathrm{SNR}}_i=\frac{p_{\max 1}}{n_i}{\left|\right|H_i\left|\right|}_F^2)\&GreaterEqual;{\left(\frac{i}{i-1}\right)}^{i-1}-1;$

3) if its signal to noise ratio is satisfied with threshold value, then it is added powerful channel set And the circulation of continuation down hop.

4) if its signal to noise ratio does not satisfy above-mentioned threshold value, then remembers M=i-1, and gathering i bar remaining in the K to the N bar, the N+1-i subchannel all adds low power channel set altogether

In, and again jump in the step 3.(annotate: before this we with all subchannels according to descending arranged sequentially of matrix norm, matrix norm is larger, illustrate that channel status is better, that is to say like this according to channel status by to bad order all subchannels being lined up well, if there is a sub-channels not satisfy the condition of threshold value, that just illustrates that later subchannel does not satisfy the threshold value condition)

Step 4

Distribute power for every sub-channels

1) determines $p_{\max }=\frac{p_{\mathrm{total}\max }}{M},$ $p_{\min }=\frac{p_{\mathrm{total}\min }}{N-M};$

2)

In every subchannel all distribute power P _Max,

In every subchannel all distribute power p _Min

3) all subchannels all begin communication;

Finish algorithm.

Fig. 6 shown along with the number of subchannels purpose increases, and two-element set algorithm and traditional two kinds of algorithms of average power are for the performance comparison of optimized for energy efficiency, wherein transmitting antenna n _T=4, reception antenna n _R=4, subchannel number M ∈ [8,128], total consumed power p _Total=1W.

Fig. 6 shown subchannel for two-element set algorithm and traditional average power two kinds of algorithms in the impact aspect the optimized for energy efficiency, can clearly find out: under the system of the many residential quarters of MIMO, the algorithm of two-element set for the optimization of energy efficiency be algorithm apparently higher than traditional average power, even when subchannel number was lower, the performance of its optimization all was higher than the optimized algorithm of average power.

Fig. 7 shown in the cellular communication system of the many residential quarters of MIMO, and total power constraint is for the impact of optimized for energy efficiency, above a curve be that expression is as gross power definite value p _TotalThe curve of the energy efficiency during=0.6W and subchannel number; Below curve be that explanation is as gross power definite value p _TotalThe curve of the energy efficiency during=1.3W and subchannel number, wherein transmitting antenna n _T=4, reception antenna n _R=4, subchannel number M ∈ [8,128].

As seen from Figure 7, the energy efficiency of system increases along with the increase of subchannel number; But the increase of system's gross power restriction is so that the energy efficiency reduction of system.Therefore can draw: use this power control algorithm, in the situation that improves the total power constraint value, can not improve the energy efficiency of system; But same, the increase of number of subchannels, namely channel is multiplexing, can increase the energy efficiency of system.

Advantage of the present invention comprises:

Under the system of the many residential quarters of MIMO, the algorithm of two-element set is better than the algorithm of traditional average power for the optimization of energy efficiency, even when subchannel number is lower, the performance of its optimization all is higher than the optimized algorithm of average power.

Claims (6)

1. many residential quarters multi-input multi-output system is considered the optimized for energy efficiency power distribution algorithm that co-channel interference suppresses, and it is characterized in that comprising:

A) initialization;

B) sort to subchannel;

C) determine powerful channel set

With low power channel set

D) distribute power for every sub-channels.

2. method according to claim 1 is characterized in that described steps A) comprising:

A1) determine the gross power p of system _Total, determine powerful channel set In the power summation p of each channel _Totalmax, determine low power channel set

In the power summation p of each channel _Totalmin,

Wherein, powerful channel set

And low power channel set (empty set) and $K_{p\min }^{N-M}=\mathrm{\&Phi;}$ (empty set),

A2) the set K of generation N sub-channels: $K=\left\{\mathrm{CH}\right|\mathrm{CH}{=}^{\&prime;}\underset{i=1}{\overset{N}{\&cup;}}{\mathrm{CH}}_i\}.$

3. method according to claim 2 is characterized in that described step B) comprising:

B1) as follows according to the descending arrangement of the norm value of its channel the N sub-channels among the set K:

$\tilde{K}=\left\{\mathrm{CH}\right|\mathrm{CH}{=}^{\&prime;}\underset{i=1}{\overset{N}{\&cup;}}{\mathrm{CH}}_i;\&ForAll;(1\&le;i\&le;k\&le;N):{\left|\right|H_i\left|\right|}_F^2\&GreaterEqual;{\left|\right|H_k\left|\right|}_F^2\};$

The transmission matrix norm of expression subchannel i.

4. method according to claim 3 is characterized in that described step C) comprising:

Be directed to every channel CH _i∈ K is for i=1:N

C1) establish $p_{\max }=\frac{p_{\mathrm{total}\max }}{i};$

C2) for i bar channel

C2.1) if i 〉=35 judge then whether its Signal to Interference plus Noise Ratio is satisfied with the condition 2 of thresholding:

$({\mathrm{SNR}}_i=\frac{p_{\max 1}}{n_i}{\left|\right|H_i\left|\right|}_F^2)\&GreaterEqual;e-1;$

C2.2) if i＜35 judge then whether its Signal to Interference plus Noise Ratio is satisfied with the condition 1 of thresholding:

$({\mathrm{SNR}}_i=\frac{p_{\max 1}}{n_i}{\left|\right|H_i\left|\right|}_F^2)\&GreaterEqual;{\left(\frac{i}{i-1}\right)}^{i-1}-1;$

C3) if above-mentioned steps C2) judge that described signal to noise ratio is satisfied with the respective doors limit value, then the i subchannel is added powerful set of sub-channels

And the circulation of continuation down hop;

C4) if above-mentioned steps C2) judge that described signal to noise ratio does not satisfy the respective threshold value, then remember M=i-1, and i bar remaining in the set K is all added low power set of sub-channels to N bar altogether (N+1-i) subchannel

In, and again jump to step C) in.

5. method according to claim 4 is characterized in that described Step D) comprising:

D1) determine $p_{\max }=\frac{p_{\mathrm{total}\max }}{M},$ $p_{\min }=\frac{p_{\mathrm{total}\min }}{N-M};$

D2)

In every subchannel all distribute power P _Max,

In every subchannel all distribute power p _Min

D3) all subchannels all begin communication.

6. method according to claim 3 is characterized in that

Since with all subchannels according to descending arranged sequentially of matrix norm, matrix norm is larger, illustrate that channel status is better, so according to channel status by to bad order all subchannels being lined up well, do not satisfied threshold condition if judged a sub-channels, just the subchannel after the explanation does not satisfy threshold condition.

Images