CN101834814B - Time-variant TDD-MIMO communication channel reciprocity compensation method based on channel prediction - Google Patents

Abstract

The invention discloses a time-variant TDD-MIMO communication channel reciprocity compensation method based on channel prediction, mainly aiming at solving the problem of channel reciprocity loss caused by channel time variation in a TDD-MIMO communication system. The method comprises the steps of: using a base station (BS) for carrying out channel estimation on a subframe of an upstream link, and obtaining channel state information of the subframe of the upstream link; then, according to the obtained channel state information of the subframe of the upstream link, predicting the subsequent channel state information of a subframe of a downstream link; according to the predicted channel state information, using the BS for carrying out recoding treatment on the subsequent subframe of the downstream link, and maintaining the channel reciprocity of the TDD-MIMO and the downstream link. The method does not need feedback, and uses the base station to obtain the channel state information of the downstream link in the TDD-MIMO system, thus reducing the cost of the system, being applicable to compensating channel reciprocity loss in the TDD-MIMO system caused by channel time variation.

Description

Based on channel estimating the time become the compensation method of TDD-MIMO communication channel reciprocity

Technical field

The invention belongs to communication technical field, relate to the signal reciprocity compensation of TDD-MIMO communication system, specifically at becoming the compensation method that the channel reciprocity that causes is lost this problem in the TDD-MIMO system during by channel, thereby the raising power system capacity is in the TDD-MIMO mobile communication system that becomes when being applied to have channel.

Background technology

Mimo system becomes one of key technology of radio communication in the future because of the capacity that need not increase frequency spectrum resource and antenna transmitted power and but can significantly improve wireless channel.In the MIMO down link, if can know the channel condition information of making a start, then can significantly improve link capacity.In traditional FDD system, the uplink and downlink link works in different frequencies, and in order to know downlink channel status information making a start, receiving end need feed back to and make a start, but along with the increase of antenna number, feedback quantity will increase with exponential form.Therefore and in the TDD system, the uplink and downlink link uses same frequency, and uplink and downlink link channel characteristic unanimity is called and has channel reciprocity.In the TDD-MIMO system, downlink channel status information can obtain from up link by the reciprocity of channel, even known uplink channel state matrix H _U, then downlink channel state matrix can by

Obtain, base station BS can be made a start preliminary treatment to obtain maximum system capacity, as shown in Figure 1 according to the channel condition information of this acquisition.This process does not need to use special feedback channel, has reduced overhead, so the reciprocity of TDD system also is one of Inherent advantage than the FDD system.

In some realization suggestions of MIMO communication system, channel condition information not only is used for the decoding of receiving end, and also for precoding or the preliminary treatment of making a start, and the supposition of reciprocity is widely accepted and be used for estimating effectively channel.Be patent method, transceiver and the MIMO communication system of channel reciprocity " obtain " of CN 101444054A as publication number, provided a kind of method of in the MIMO communication system, determining the uplink downlink communication channel characteristics, so that computation complexity is minimum, and minimum to the needs of communication channel characteristics feedback.But, in the reality, because the factor of receiver/transmitter own or external environment condition tend to cause the reciprocity of channel to keep to factors such as communication link exert an influence.Publication number is the patent " realizing the calibration steps of bi-directional communication channel reciprocity " of CN 1910879A, provided a kind of transmitter---under the receiver chain non-ideality emission is received the method that the sum of errors difference in the chain compensates, make the reciprocity of channel to be employed, guarantee that valuable channel resource can not be wasted in the unnecessary signal transmission, improves capacity gain.But this method requires characteristic of channel not change, and is namely constant during channel.When channel, becoming, owing to have time delay between the transfer of data of the channel estimating of up link and down link, so uplink channel state information and next downlink channel status information constantly that current time estimates do not satisfy

Relation, the CSI that grasps of namely making a start this moment is out-of-date, the no longer reciprocity of uplink and downlink channel.If this method is still with the preliminary treatment of making a start of this out-of-date uplink channel state information, not only can not improve power system capacity, also can cause the big error of receiving end data generation.In this case, the influence that becomes during channel must be considered emphatically and compensate, otherwise the reciprocity of TDD not only can not effectively use, and also can have a strong impact on systematic function.

Summary of the invention

The purpose of this invention is to provide a kind of based on channel estimating the time become the compensation method of TDD-MIMO communication channel reciprocity, make the TDD-MIMO system channel reciprocity the time kept under the changing environment, the DL-CSI of the downlink subframe that BS still can need observe according to the DL-CSI forecasting institute that the prediction of the up UL-CSI that estimates and front obtains and need not use special based on feedback link, thereby saving overhead, improve power system capacity, the Inherent advantage of TDD system is kept.

Purpose performing step of the present invention is: at first BS estimates the channel condition information of current uplink sub-frames, be uplink channel state information, then according to the uplink channel state information of estimating to obtain progressively dope adjacent downlink subframe downlink channel status information, further try to achieve the downlink channel status information of the downlink subframe of required observation, and according to the downlink channel status information that this prediction the obtains preliminary treatment of making a start, thereby become the forfeiture of the TDD-MIMO channel reciprocity that causes during compensate for channel, its step comprises as follows:

(1) the mobile station MS transmit data frames is to base station BS, and BS utilizes the Frame that receives to carry out channel estimating, obtain i (i=0,1 ..., ∞) the uplink channel state matrix of individual Frame correspondence is H _{U (i)}

(2) preserve BS and estimate the individual uplink sub-frames of the previous K (1≤K≤3) that the obtains channel state matrix H of correspondence respectively _{U (i)}(i=0,1 ..., (K-1));

(3) BS is according to the corresponding uplink channel matrix H of each uplink sub-frames that preserves _U, the downlink channel state matrix of first downlink subframe that uplink sub-frames is adjacent when utilizing AR model one-step prediction to obtain with i=0

(4) determine channel state matrix according to the value of i:

If i=1, then prediction obtains the channel state matrix of first downlink subframe

If i=2 is then according to known uplink channel state matrix H _UAnd the channel state matrix of first downlink subframe The recurrence prediction obtains the channel state matrix of second downlink subframe

If 2≤i≤L-K-1, L is frame length, then according to known uplink channel state matrix H _UAnd first channel state matrix to (i-1) individual downlink subframe

The recurrence prediction obtains the channel state matrix of i downlink subframe

The present invention has the following advantages:

1) the present invention avoids using special based on feedback link owing to replaced channel feedback with channel estimating, has reduced the complexity of system, has saved a large amount of overheads simultaneously;

2) the present invention is for the similarity of actual uplink downlink channel state information and actual downstream downlink channel state information, downlink channel status information and actual downstream downlink channel state information that the BS prediction obtains are more approaching, the preliminary treatment of making a start on this basis can effectively improve power system capacity, and the reciprocity of TDD is not fully exerted.

Description of drawings

In the TDD-MIMO communication system under the existing time varying channel of Fig. 1 based on the transmission principle schematic diagram of singular value decomposition;

Fig. 2 is the channel reciprocity compensation process figure based on channel estimating that the present invention proposes;

Fig. 3 the present invention compares with conventional method, when moving velocity of terminal v=30km/h, and the comparison diagram of power system capacity when adopting channel estimating to carry out the reciprocity compensation and do not adopt channel estimating to carry out the reciprocity compensation.

Embodiment

With reference to Fig. 2, the channel reciprocity compensation that the present invention proposes comprises the steps:

Step 1, mobile station MS transmit data frames are to base station BS, and BS utilizes the Frame that receives to carry out channel estimating.

The channel that channel estimating employing MMSE method is estimated to obtain during the emission data, adopts Kalman filtering or LMS algorithm to carry out channel tracking as initial value of tracking, obtain i (i=0,1 ..., ∞) the uplink channel state matrix of individual Frame correspondence is H _{U (i)}

Step 2 is preserved BS and is estimated the individual uplink sub-frames of the previous K (1≤K≤3) that the obtains channel state matrix H of correspondence respectively _{U (i)}(i=0,1 ..., (K-1)).

Step 3, BS is according to the corresponding uplink channel matrix H of each uplink sub-frames that preserves _U, the downlink channel state matrix of first downlink subframe that uplink sub-frames is adjacent when utilizing AR model one-step prediction to obtain with i=0

Utilize the AR model, channel state matrix

Determine by following formula:

${\tilde{H}}_{\mathrm{<figure-callout\; id="0"\; label="D"\; filenames="HSA00000131399100011.png,HSA00000131399100021.png"\; state="\{\{state\}\}">D</figure-callout>}0}=\underset{i=0}{\overset{K-1}{\mathrm{\&Sigma;}}}{a}_i{H}_{U(-i)}$

Wherein, the span of K is 1≤K≤3; Model coefficient a=[a ₁, a ₂..., a _i] ^TDetermine by following formula:

$J_0\left(2\mathrm{\&pi;}{f}_d\mathrm{Ti}\right)=\underset{l=1}{\overset{i}{\mathrm{\&Sigma;}}}{J}_0\left(2\mathrm{\&pi;}{f}_{d}T\right|i-l\left|\right)a_l$

J wherein ₀(2 π f _DT|i-l|) be the time auto-correlation coefficient of time varying channel, J ₀() is first kind zeroth order Bei Saier function; T constant duration when being channel, the 3GPP standard thinks that channel remains unchanged in a subframe, so T is expressed as the duration of a subframe; Note ρ _i=J ₀(2 π f _dTi), ρ ₀=1, model coefficient a=[a then ₁, a ₂..., a _i] ^TCan be exchanged into and find the solution following equation group and obtain:

f _dBe maximum doppler frequency, be expressed as

V is the MS translational speed, f _cBe carrier frequency, c is the light velocity.

Step 4, determine channel state matrix according to the value condition of i:

If i=1 then utilizes the AR model, the channel state matrix that prediction obtains first downlink subframe is Namely predict the downlink channel status information of adjacent first downlink subframe with the up link channel condition information;

If i=2 is then according to known uplink channel state matrix H _UAnd the channel state matrix of first downlink subframe The recurrence prediction obtains the channel state matrix of second downlink subframe

If 2≤i≤L-K-1, L is frame length, then according to known uplink channel state matrix H _UAnd first channel state matrix to (i-1) individual downlink subframe

The recurrence prediction obtains the channel state matrix of i downlink subframe

When i＞1, the channel state matrix of i downlink subframe can obtain by the prediction of following formula recurrence

${\tilde{H}}_{\mathrm{Di}}=\underset{l=1}{\overset{i-1}{\mathrm{\&Sigma;}}}{a}_l{\tilde{H}}_{D(i-l)}+\underset{i=0}{\overset{K-1}{\mathrm{\&Sigma;}}}{a}_i{H}_{U(-i)}$

K is the number of uplink data subframe in the TDD-MIMO system.

Make a start according to channel condition information

I downlink subframe for adjacent uplink sub-frames carried out precoding processing, thereby realized the channel reciprocity compensation under the time varying channel condition.

With uplink sub-frames number K=3, observing i=3 downlink subframe is example below, and the AR model coefficient is expressed as a _{(i, l)}, i represents the downlink subframe that will predict, l represents the corresponding coefficient of different channels state matrix, provides implementation example of the present invention:

(1) stores 3 corresponding channel status squares of uplink sub-frames difference that the BS estimation obtains

H _U(i)(i＝0，-1，-2)；

(2) channel state matrix of the 1st downlink subframe of the adjacent uplink sub-frames of prediction

${\tilde{H}}_{\mathrm{<figure-callout\; id="1"\; label="D"\; filenames="HSA00000131399100011.png,HSA00000131399100021.png"\; state="\{\{state\}\}">D</figure-callout>}1}=a_{\left(\mathrm{0,1}\right)}{H}_{U\left(0\right)}+a_{\left(\mathrm{1,1}\right)}{H}_{U(-1)}+a_{\left(\mathrm{1,2}\right)}{H}_{U(-2)}$

(3) channel state matrix of the 2nd downlink subframe of the adjacent uplink sub-frames of prediction

${\tilde{H}}_{\mathrm{<figure-callout\; id="2"\; label="D"\; filenames="HSA00000131399100021.png"\; state="\{\{state\}\}">D</figure-callout>}2}=a_{\left(\mathrm{2,1}\right)}{\tilde{H}}_{D1}+a_{\left(\mathrm{2,2}\right)}{H}_{U\left(0\right)}+a_{\left(\mathrm{2,3}\right)}{H}_{U(-1)}+a_{\left(\mathrm{2,4}\right)}{H}_{U(-2)}$

(4) channel state matrix of the 3rd downlink subframe of the adjacent uplink sub-frames of prediction

${\tilde{H}}_{\mathrm{<figure-callout\; id="3"\; label="D"\; filenames="HSA00000131399100021.png"\; state="\{\{state\}\}">D</figure-callout>}3}=a_{\left(\mathrm{3,1}\right)}{\tilde{H}}_{\mathrm{<figure-callout\; id="2"\; label="D"\; filenames="HSA00000131399100021.png"\; state="\{\{state\}\}">D</figure-callout>}2}+a_{\left(\mathrm{3,2}\right)}{\tilde{H}}_{\mathrm{<figure-callout\; id="1"\; label="D"\; filenames="HSA00000131399100011.png,HSA00000131399100021.png"\; state="\{\{state\}\}">D</figure-callout>}1}++a_{\left(\mathrm{3,3}\right)}{H}_{U\left(0\right)}+a_{\left(\mathrm{3,4}\right)}{H}_{U(-1)}+a_{\left(\mathrm{3,5}\right)}{H}_{U(-2)}$

So far, think that the channel condition information of transmitting-receiving two-end is all known, making a start can be according to the known channel condition information of making a start

The 3rd downlink subframe for adjacent uplink sub-frames carried out precoding processing, thereby realized the channel reciprocity compensation under the time varying channel condition.

Effect of the present invention can further specify by following principle and emulation:

1) existing reciprocity compensation method principle based on singular value decomposition

Supposing has the M transmit antennas, N root reception antenna in the TDD-MIMO system.X is the incoming symbol vector, H _D3Be the channel state matrix of the 3rd downlink subframe of reality of needs observation, n is the additive white Gaussian noise vector AWGN of receiving end, and the symbolic vector that down link MS receives can be expressed as

y＝H _D3x+n，

Suppose that n is the multiple Gaussian noise of zero-mean, and noise is independent on every reception antenna, has

E (nn ^H)=I _M, n ^HConjugate transpose for n

Be BS making a start of down link, to the channel state matrix H of i=0 uplink sub-frames of next-door neighbour's downlink subframe _{U (0)}Carrying out the SVD branch solves

H _U(0)＝U _U(0)D _U(0)V _U(0) ^H

Wherein, U and V are respectively unitary matrix, and D is diagonal matrix, and diagonal element is the characteristic value of H, and arrange by order from big to small.

According to the reciprocity of channel, make a start earlier to x V _{U (0)}Carry out precoding, by actual channel H _D3, again at receiving end U _D3 ^HSeparate precoding, finally obtain

y＝U _D3 ^H(H _D3V _U(0)x+n)

The down link of this moment differs the 3T time with the up link that estimates channel status, namely 3 subframe lengths make channel that bigger variation take place, and the variation that channel is very little in the reality just can cause that pre-coding matrix reconciliation pre-coding matrix produces bigger skew, make it no longer to mate, thereby use V _{U (0)}Will produce bigger error, thereby cause power system capacity to descend.

The reciprocity compensation method based on SVD when 2) use is of the present invention

By above analysis as can be known, in time-varying system, carry out channel estimating, obtain the channel state matrix of i=3 downlink subframe of required observation

Carrying out the SVD branch solves

${\tilde{H}}_{\mathrm{<figure-callout\; id="3"\; label="D"\; filenames="HSA00000131399100021.png"\; state="\{\{state\}\}">D</figure-callout>}3}={\tilde{U}}_{D3}{\tilde{D}}_{D3}{{\tilde{V}}_{D3}}^H$

This moment, the channel condition information of the downlink subframe that BS is known was

Therefore adopt

The precoding of making a start is by actual channel H _D3, use in receiving end again Separate precoding, finally obtain the symbolic vector that the MS of i=3 down link of required observation receives

$y={U_{D3}}^H(H_{D3}{\tilde{V}}_{D3}x+n)$

The BS precoding of making a start is separated precoding in conjunction with receiving end MS, can carry out corresponding reciprocity compensation at the variation of channel, thus channel variation the influence of transmission performance is eliminated fully, power system capacity is improved.Be example with MIMO 2 * 2 antennas, getting MS speed is 30km/h, signal to noise ratio snr=10dB, i=3, K=3 has carried out emulation, and the channel capacity of existing reciprocity uncompensation is carried out emulation, the result as shown in Figure 3, from Fig. 3 as seen, the present invention adopt channel estimating carry out channel reciprocity compensation after power system capacity obviously improve, thereby effectively realized channel reciprocity compensation under the time varying channel condition.

The present invention is not limited to the embodiment of said system, utilizes principle of the present invention and scheme, and those skilled in the art can make various modifications or remodeling, but these remodeling and application are all within protection scope of the present invention.

Claims (1)

1. One kind based on channel estimating the time become the compensation method of TDD-MIMO communication channel reciprocity, comprise the steps:

   (1) the mobile station MS transmit data frames is to base station BS, and BS utilizes the Frame that receives to carry out channel estimating, and the uplink channel state matrix that obtains i Frame correspondence is H _{U (i)}, i=0,1 ..., ∞;

   (2) preserve previous K the corresponding channel state matrix H of uplink sub-frames difference that the BS estimation obtains _{U (i)}, i=0,1 ..., (K-1), 1≤K≤3;

   (3) BS is according to the corresponding uplink channel matrix H of each uplink sub-frames that preserves _U, utilize the AR model to carry out one-step prediction, the downlink channel state matrix of first downlink subframe that uplink sub-frames is adjacent when obtaining with i=0

   

   ${\tilde{H}}_{D0}=\underset{i=0}{\overset{K-1}{\mathrm{\&Sigma;}}}{a}_i{H}_{U(-i)},$

   Wherein, a=[a ₀, a ₁..., a _K-1] ^TBe model coefficient, the span of K is 1≤K≤3, model coefficient a=[a ₁, a ₂..., a _i] ^TDetermine by following formula:

   $J_0\left(2\mathrm{\&pi;}{f}_d\mathrm{Ti}\right)=\underset{l=1}{\overset{i}{\mathrm{\&Sigma;}}}{J}_0\left(2\mathrm{\&pi;}{f}_{d}T\right|i-l\left|\right)a_l,$

   J wherein ₀(2 π f _dT|i-l|) be the time auto-correlation coefficient of time varying channel, J ₀() is first kind zeroth order Bei Saier function; T constant duration when being channel, the 3GPP standard thinks that channel remains unchanged in a subframe, so T is expressed as the duration of a subframe; Note ρ _i=J ₀(2 π f _dTi), ρ ₀=1, model coefficient a=[a then ₁, a ₂..., a _i] ^TCan be exchanged into and find the solution following equation group and obtain:

   

   f _dBe maximum doppler frequency, be expressed as

   

   V is the MS translational speed, f _cBe carrier frequency, c is the light velocity;

   (4) determine channel state matrix according to the value condition of i:

   If i=1, then prediction obtains the channel state matrix of first downlink subframe

   

   If i=2 is then according to known uplink channel state matrix H _UAnd the channel state matrix of first downlink subframe

   

   Obtain the channel state matrix of second downlink subframe by following formula recurrence prediction

   

   ${\tilde{H}}_{\mathrm{Di}}=\underset{l=1}{\overset{i-1}{\mathrm{\&Sigma;}}}{a}_l{\tilde{H}}_{D(i-l)}+\underset{i=0}{\overset{K-1}{\mathrm{\&Sigma;}}}{a}_i{H}_{U(-i)},$

   Wherein, a=[a ₁, a2 ..., a _i] ^TBe model coefficient; 2≤i≤L-K-1, L determines that by the frame structure in the 3GPP standard of selecting for use K represents the number of uplink sub-frames in the frame, and protection of 1 expression is subframe at interval, and L-K-1 represents the number of downlink subframe in the frame;

   If 2≤i≤L-K-1, L is frame length, then according to known uplink channel state matrix H _UAnd first channel state matrix to i-1 downlink subframe

   

   The recurrence prediction obtains the channel state matrix of i downlink subframe

   

Images