CN101039290A

CN101039290A - Method for estimating MIMO related channel based on self-adaptive training sequence

Info

Publication number: CN101039290A
Application number: CN 200710017709
Authority: CN
Inventors: 李建东; 庞继勇; 赵林靖; 吕卓; 陈亮; 董伟
Original assignee: Xidian University
Current assignee: Xidian University
Priority date: 2007-04-19
Filing date: 2007-04-19
Publication date: 2007-09-19
Anticipated expiration: 2027-04-19
Also published as: CN101039290B

Abstract

This invention discloses a relative channel estimation method of MIMO based on self-adaptive training sequence. The process is that, the receptor determines the best length of the training sequence according to relative known information of the channel and transfers the value of the length and relative information of the channel through feedback link to the transmitter. The transmitter uses this feedback information to compute the optimal training sequence correspondent to current state of the channel according to the training sequence expression St=UD1/2tU*t designed by this model and transfers this optimal training sequence to the receptor through the forward link. Then the training cycle begins and the transmitter launches the training sequence to wireless channel. The receptor estimates the channel parameters of current time according to the known training sequence and the receipted signal in the training cycle and by using the minimum mean square error estimation criteria. The training sequence designed by this invention can make self-adaptive adjustment according to the needs of the actual system and the change of relative information of the channel. The invention has advantages of high estimation performance and strong robustness, thus can be used in wireless communication system with muti-antenna MIMO.

Description

MIMO related channel estimation method based on self-adaptive training sequence

Technical Field

The invention belongs to the field of communication signal processing, and relates to a channel estimation method which can be used for a multi-antenna MIMO wireless communication system.

Background

In the last decade, information communication technology and application systems thereof have been rapidly developed, presenting an unprecedented prosperous scene. The development and maturity of mobile communication, wireless communication, multimedia information service and internet show a good prospect for anyone to be able to perform any kind of information interaction at any time and any place. The widespread use of internet technology and the increasing reliance and demand of people on data transmission services have impacted and pushed the continuous update of mobile communication technology at the end of the twentieth century. The existing third generation mobile communication system has difficulty in meeting the requirements of high-speed, multi-service and high-quality communication and data transmission of future mobile communication, so that the concept of super 3G or 4G is proposed, and a plurality of international standardization organizations and forums are actively developing the research of future mobile communication. For example, the international telecommunications union-radio communication sector ITU-R indicates in the future development of the global standard IMT-2000 for third generation wireless communication and documents of the beyond-IMT-2000 system: the capability of the IMT-2000 terrestrial wireless interface will extend to nearly 30Mbps in around 2005; the new system envisioned to exceed IMT-2000 around 2010 would support a peak rate of about 100Mbps under high speed mobile conditions and about 1Gbps under low speed mobile conditions.

The key of the currently widely agreed technology for supporting the high-rate requirement of future mobile communication is a multiple-input multiple-output (MIMO) system, and the MIMO technology can improve the channel capacity of a wireless channel by times on the premise of not increasing the system bandwidth and the transmission power. According to the research result of information theory, if the channel fading between different transmitting-receiving antenna pairs are independent, one has N under the same transmitting power and bandwidth_tA transmitting antenna and N_rThe channel capacity of the MIMO system with a plurality of receiving antennas can reach min (N) of the prior single-antenna system_t，N_r) And thus provides capacity boosting potential that is incomparable with other current technologies. The MIMO system is considered to be a key to realizing future mobile communicationOne of the key technologies.

However, the premise of the MIMO system for achieving the multiple channel capacity is that the receiver can acquire the channel fading state information as accurate as possible, thereby effectively demodulating and decoding the received signal. Therefore, the channel estimation technique is a key for realizing effective transmission of wireless communication, and is a technical problem that must be solved first in the practical process of the MIMO system.

At present, the MIMO channel estimation methods mainly include three types, i.e., a training sequence/pilot based channel estimation algorithm, a blind channel estimation algorithm, and a semi-blind channel estimation algorithm. In consideration of the maturity, robustness and complexity of the algorithm, in combination with various standard specifications, an estimation method based on a training sequence is generally adopted. The method is fast, simple and effective, is favorable for separating and processing the channel estimation and signal detection processes, and can greatly simplify the design of a receiver.

A great deal of literature is available to make extensive studies on the design of training sequences, and it is generally considered that the estimated mean square error of orthogonal training sequences is the smallest. These studies typically assume that the channel fading coefficients between each transceiver antenna pair are independently and equally distributed, or that a suboptimal estimation criterion is adopted that does not take into account channel-related information. However, it is known that the spatial fading correlation of signals caused by insufficient angular spread of incoming waves and small antenna element spacing in a wireless mobile channel, the correlation between multi-paths determined by the pulse forming filters at the transmitting and receiving ends and the physical channel response, the time correlation introduced by doppler shift and the existence of direct paths all seriously affect the characteristics of the MIMO system, and these factors must be considered.

The related information of the channel belongs to the second-order statistical characteristic of the channel state, the variable speed of the channel is much slower than the fading coefficient of the channel, and the related information of the channel can be obtained and tracked even under the condition that the coherence time of the channel is short. The receiving end can use the relevant information to assist channel parameter estimation and the transmitting end can use the relevant information to design the optimal training sequence, thus effectively improving the accuracy of channel estimation. In addition, the correlation of channel fading will reduce the number of independent coefficients in the channel matrix, and reduce the dimension of the channel coefficient space. For the channel estimation mechanism based on the training sequence, the number of parameters to be estimated directly determines the length of the training period, thereby limiting the transmission efficiency of the system. Therefore, under the channel estimation mechanism, the correlation reduces the number of independent parameters to be estimated, and also reduces the estimation workload and shortens the length of the training sequence, thereby theoretically improving the system capacity.

In recent years, there are also a few articles that consider the problem of training sequence design under the relevant MIMO channel in combination with channel-related information on the basis of orthogonal training sequences. However, the training sequences designed in these only few documents are only one of pursuit of minimizing the mean square error of channel estimation, and obviously, the longer the training sequence is, the better the estimation performance is, and no consideration is given to the trade-off between the length of the training sequence and the length of data transmission, and no consideration is given to the problem from the perspective of the effective transmission rate of the system under the training mechanism. It is clear that the effect of the training sequence is only to estimate the channel, and not to have any information transfer effect, and how much the occupied/consumed channel transmission time will directly affect the practical availability of the system. Moreover, since the structure of the existing training sequence is fixed and invariable, the length of the training sequence cannot be adaptively adjusted according to the spatial domain related information of the current channel, and the known channel related information cannot be used more effectively, which results in a problem of low robustness of the estimation performance.

Disclosure of the invention

The invention aims to overcome the defects of fixed training sequence structure and poor training length adaptive adjustment performance in the prior art, and provides a training sequence-based MIMO related channel adaptive estimation method.

The technical idea for realizing the purpose of the invention is as follows: aiming at an MIMO flat block fading channel with spatial fading correlation, on the basis of assuming that the spatial correlation of the channel obeys a Kronecker separation correlation model, a training sequence generation method which can effectively utilize channel correlation information and has controllable length is designed, the compromise between estimation accuracy and system transmission capacity is comprehensively considered in combination with channel coherence time, and the self-adaptive channel estimation of the MIMO correlation channel is realized by the Minimum Mean Square Error (MMSE) estimation criterion and the optimization principle, and the specific process is as follows:

(1) receiver pair known transmitting end space domain correlation matrix R_tAnd receiving end space domain correlation matrix R_rPerforming matrix direct product operation to obtain second-order statistical correlation information R of the current channel_hI.e. R_h＝R_r*R_t* denotes the direct product between the matrices, the channel correlation matrix R_hRank K of (2) is denoted as K ═ rank (R)_h)；

(2) Determining the training sequence length for channel estimation from the value of K, i.e.

Wherein N is_rWhich indicates the number of receive antennas to be used,

represents rounding up;

(3) the receiver will determine the training sequence length T through a feedback link between the receiver and the transmitter_τAnd channel related information R_hFeeding back to the transmitter;

(4) the transmitter determines a diagonal matrix D in the training sequence structure according to the feedback information_τ，

When T is_τ≥N_tWhen it is taken

When T is_τ＜N_tWhen it is taken

Wherein N is_tDenotes the number of transmitting antennas, 0_(Tτ-Nt)×NtRepresents a (T)_τ-N_t)×N_tAll 0 matrix, superscript, of dimension^1/2Represents the square root of the matrix, D_{[1：Tτ，：]} ^1/2Represents to take D^1/2Front T of_τRow, matrix D is one N_t×N_tA dimensional diagonal matrix;

(5) from the above matrix D_τGenerating a training sequence S_τI.e. by

Wherein U is an arbitrary T_τ×T_τA dimensional unitary matrix; u shape_tIs a transmit-end correlation matrix R_tN in eigenvalue decomposition of_t×N_tDimensional eigenvector unitary matrix, superscript^*A conjugate transpose operation representing a matrix;

(6) the transmitter informs the receiver of the currently used training sequence S via the forward link from the transmitter to the receiver_τThe specific numerical values of (a);

(7) starting the training period, the transmitter transmits a training sequence S_τTransmitted into a radio channel H via a transmitting antenna, the receiver being at T_τThe signal received in one symbol period may be represented as a T_τ×N_tMatrix X of dimensions_τ；

(8) The receiver receives the signal X according to the above_τAnd the signalled training sequence S_τAnd obtaining the estimated value of the current channel coefficient by adopting a Minimum Mean Square Error (MMSE) estimation criterion according to the following formula

Wherein,

\hat{h} = vec (\hat{H}),

x_τ＝vec(X_τ) H ═ vec (h) denotes channel estimation matrices, respectively

Received signal matrix X_τAnd the column-piled-up vector, p, of the channel matrix H_τRepresents the average transmit power over the training period,

<math> <mrow> <msub> <mover> <mi>S</mi> <mo>~</mo> </mover> <mi>τ</mi> </msub> <mo>=</mo> <mrow> <mo>(</mo> <msub> <mi>I</mi> <msub> <mi>N</mi> <mi>r</mi> </msub> </msub> <mo>&CircleTimes;</mo> <msub> <mi>S</mi> <mi>τ</mi> </msub> <mo>)</mo> </mrow> <mo>,</mo> <msub> <mi>I</mi> <msub> <mi>N</mi> <mi>r</mi> </msub> </msub> </mrow> </math>

represents N_rDimension unit array, superscript^-1Representing a matrix inversion operation;

(9) after the training is finished, the transmission of real useful data is started. After one frame transmission is finished, if the channel related information R is_hIf the value of the training sequence is changed, returning to the first step to re-determine a new training sequence, and performing channel estimation and data transmission of the next frame; if the channel related information R_hIf the value of (2) is not changed, returning to the step (7) to start the channel estimation and data transmission of the next frame.

The above channel estimation method, wherein N_t×N_tThe dimension diagonal matrix D may be normalized to N by the traces satisfying matrix D_tT_τIs obtained by minimizing the following formula:

wherein d is_iIs the ith diagonal element of D, λ_i(R_t) Representing a transmit-side spatial correlation matrix R_tOf the ith characteristic value, λ_j(R_r) Represents the receiving end spatial correlation matrix R_rThe jth eigenvalue of (1).

Compared with the prior art, the invention has the following advantages:

(1) the training sequence structure used by the channel estimation has universality and generalization, and can contain the existing specific training sequence under the same channel condition and model, such as the traditional orthogonal training sequence;

(2) the length of a training sequence generated by a training sequence structure in the channel estimation has self-adaptive characteristics, and the training time can be adjusted according to different channel airspace correlation conditions and different channel estimation error requirements;

(3) on the premise of the same training length, the estimation error of channel estimation through the training sequence is obviously smaller than that when the orthogonal training sequence is adopted;

(4) on the premise of the same channel estimation error, the length of the training sequence is smaller than that of the orthogonal training sequence, so that the training overhead is reduced, and the utilization rate of the system bandwidth is improved;

(5) the training sequence design idea has universality and can be effectively combined with an MIMO precoding technology and a self-adaptive transmission technology.

Drawings

FIG. 1 is a schematic block diagram of a training sequence based MIMO channel estimation/data detection system according to the present invention

FIG. 2 is a schematic diagram of training sequence design concept of the present invention

FIG. 3 is a flow chart of the adaptive channel estimation of the present invention

FIG. 4 is a graph comparing the estimation performance of the training sequence of the present invention and the conventional orthogonal training sequence

FIG. 5 is a graph of the length adaptation of training sequence and the estimation accuracy of the present invention

Detailed Description

The technical solution of the present invention is described in further detail below with reference to the accompanying drawings.

Referring to fig. 1, the principle of the MIMO channel estimation/data detection system based on training sequence of the present invention is to transmit a certain number of training symbols, called training sequence S, known to both the transmitter and the receiver at the beginning of each data frame/packet_τ. The receiving end receives the signal X according to the training period_τAnd a training sequence S_τEstimating the channel parameter matrix in the frame by adopting the Minimum Mean Square Error (MMSE) estimation criterion

After the training period, transmission of the actual information data symbols S is started_dThe receiving end utilizes the estimated channel parameters

Receiving signal X from data transmission stage_dIntermediate demodulation/decoding of original data information

The transmitting and receiving ends are realized by an error-free bidirectional link independent of data transmission, namely a forward link and a feedback linkThe interaction between information sharing and parameter control is performed, so that an adaptive channel estimation mechanism is established.

Referring to fig. 2, the starting point of the training sequence design of the present invention is to fully and effectively utilize the known channel related information to improve the performance and robustness of channel estimation. The important point of consideration is the spatial fading correlation of the channel, and the spatial fading correlation is modeled by a Kronecker correlation model. The main factor reflecting the self-adaptive principle in the training sequence structure is the controllability of the training length, the evaluation criterion adopted by the training sequence design is the minimum mean square error MMSE estimation criterion, and in the calculation of deriving the minimum mean square error, the matrix theory and the optimization principle are utilized to derive the optimal training sequence.

Referring to fig. 3, the system for channel estimation of the present invention is a MIMO system having N therein_tA transmitting antenna and N_rA receiving antenna, currently N_t×N_rThe dimensional channel state matrix is H, the training sequence length is T_τT over the entire training period_τ×N_tThe dimension training sequence matrix is written as S_τ，T_τ×N_rDimension received signal matrix is X_τ，T_τ×N_rThe received complex Gaussian white noise matrix is N_τ. The system has a bidirectional link between the transmitter and the receiver, namely a forward link from the transmitter to the receiver and a feedback link from the receiver to the transmitter, the bidirectional link is independent from a wireless communication link between the transceiving antennas and is only used for transmitting a small amount of interactive control information, a communication frequency band of the bidirectional link is independent from a service data transmission frequency band, and the transmission rate is low and can be considered as error-free transmission. The receiver in the system can obtain the second-order statistical correlation information of the channel, namely the transmitting terminal space domain correlation matrix R of the channel_tAnd receiving end space domain correlation matrix R_rAnd can better track the change of the channel related information. The specific channel estimation process is as follows:

1. obtaining second-order statistical relevant information R of current channel_h

Receiver pair known transmitting end space domain correlation matrix R_tAnd receiving end space domain correlation matrix R_rPerforming matrix direct product operation to obtain second-order statistical correlation information R of the current channel_hI.e. R_h＝R_r*R_t* denotes the direct product between the matrices; the channel correlation matrix R_hRank of (c) is K, denoted as K ═ rank (R)_h)；

2. Selecting training sequence length T for channel estimation_τ′

Training sequence length T is selected by using theory of incoherent capacity and requirement of receiving end on estimation error performance_τ', i.e. according to N_t×N_rThe number of independent parameters in the dimensional channel matrix H is equal to R_hAccording to the system capacity maximization criterion, selecting

The rounding-up is expressed, namely the optimal compromise between the estimated reliability and the maximization of the data transmission rate can be ensured;

3. determining the final training sequence length T for channel estimation_τ

The above threshold value

The length of the training sequence is ideally optimal, but in specific application, the length of the training sequence needs to be adjusted according to the limit of the practical system on the training time and the requirement of the receiver on the performance of the current channel estimation mean square error. The basic principle of the adjustment is to reduce the training overhead as much as possible on the premise of ensuring that the performance requirement of the estimation error and the data transmission rate required by the service are met, and generally, the method is to take

4. The receiver will finally determine the training sequence length T through the feedback link from the receiver to the transmitter_τAnd channel phaseInformation of interest R_hFeeding back to the transmitter;

5. determining a diagonal matrix D in a training sequence structure_τ

The transmitter determines the matrix D according to the feedback information according to the following two conditions_τThe concrete structure of (1):

when T is_τ≥N_tWhen it is taken

When T is_τ＜N_tWhen it is taken

Get D immediately^1/2Front T of_τThe rows of the image data are, in turn,

wherein the matrix D is an N_t×N_tAnd (5) maintaining a diagonal matrix. The matrix D can be obtained by minimizing the following equation, i.e.

<math> <mrow> <mi>tr</mi> <msup> <mrow> <mo>(</mo> <msubsup> <mi>D</mi> <mi>h</mi> <mrow> <mo>-</mo> <mn>1</mn> </mrow> </msubsup> <mo>+</mo> <mfrac> <msub> <mi>ρ</mi> <mi>τ</mi> </msub> <msub> <mi>N</mi> <mi>t</mi> </msub> </mfrac> <msub> <mi>I</mi> <msub> <mi>N</mi> <mi>r</mi> </msub> </msub> <mo>&CircleTimes;</mo> <mi>D</mi> <mo>)</mo> </mrow> <mrow> <mo>-</mo> <mn>1</mn> </mrow> </msup> </mrow> </math>

<math> <mrow> <mo>=</mo> <mi>tr</mi> <msup> <mrow> <mo>(</mo> <msup> <mrow> <mo>(</mo> <msub> <mi>D</mi> <mi>r</mi> </msub> <mo>&CircleTimes;</mo> <msub> <mi>D</mi> <mi>t</mi> </msub> <mo>)</mo> </mrow> <mrow> <mo>-</mo> <mn>1</mn> </mrow> </msup> <mo>+</mo> <mfrac> <msub> <mi>ρ</mi> <mi>τ</mi> </msub> <msub> <mi>N</mi> <mi>t</mi> </msub> </mfrac> <msub> <mi>I</mi> <msub> <mi>N</mi> <mi>r</mi> </msub> </msub> <mo>&CircleTimes;</mo> <mi>D</mi> <mo>)</mo> </mrow> <mrow> <mo>-</mo> <mn>1</mn> </mrow> </msup> </mrow> </math>

Where tr denotes the trace of the matrix, d_iIs the ith diagonal element of D, λ_i(R_t) Representing a transmit-side spatial correlation matrix R_tOf the ith characteristic value, λ_j(R_r) Represents the receiving end spatial correlation matrix R_rIs the jth characteristic value of (1) < p >_τRepresenting the average transmit power over the training period.

The above formula can be solved iteratively by the optimization theory, and the constraint condition of the optimization is that the trace of the matrix D is normalized to N_tT_τ. For four different conditions of high signal-to-noise ratio and low signal-to-noise ratio, only receiving end spatial correlation and only transmitting end spatial correlation, the above formula can obtain a closed solution by Lagrange number multiplication and a limit approximation method;

6. from the above diagonal matrix D_τGenerating a training sequence S_τI.e. by

The training sequence structure consists of three parts. Wherein U is an arbitrary T_τ×T_τThe dimensional unitary matrix such as a Fourier matrix and the 'arbitrary' binary ensures the universality and the wide inclusion of the structure; u shape_tIs a transmit-end correlation matrix R_tN in eigenvalue decomposition of_t×N_tA dimensional eigenvector unitary matrix; matrix D_τThe number of rows is equal to the length of the training sequence, so the training sequence structure can generate the training sequence with any length and has self-adaptability.

It can be shown that S is present when there is no spatial correlation in the channel_τIs an orthogonal matrix, that is, the conventional orthogonal training sequence is an application of a specific case in the invention;

7. the transmitter informs the receiver of the currently used training sequence S via the forward link from the transmitter to the receiver_τThe specific numerical values of (a);

8. starting the training period, the transmitter transmits a training sequence S_τTransmitted into a radio channel H via a transmitting antenna, the receiver being at T_τThe signal received in one symbol period may be represented as a T_τ×N_tMatrix X of dimensions_τ. The mathematical description of the transceived signals of the training phase is as follows:

<math> <mrow> <mo>&DoubleRightArrow;</mo> <mi>vec</mi> <mrow> <mo>(</mo> <msub> <mi>X</mi> <mi>τ</mi> </msub> <mo>)</mo> </mrow> <mo>=</mo> <msqrt> <mfrac> <msub> <mi>ρ</mi> <mi>τ</mi> </msub> <msub> <mi>N</mi> <mi>t</mi> </msub> </mfrac> </msqrt> <mrow> <mo>(</mo> <msub> <mi>I</mi> <msub> <mi>N</mi> <mi>r</mi> </msub> </msub> <mo>&CircleTimes;</mo> <msub> <mi>S</mi> <mi>τ</mi> </msub> <mo>)</mo> </mrow> <mi>vec</mi> <mrow> <mo>(</mo> <mi>H</mi> <mo>)</mo> </mrow> <mo>+</mo> <mi>vec</mi> <mrow> <mo>(</mo> <msub> <mi>N</mi> <mi>τ</mi> </msub> <mo>)</mo> </mrow> </mrow> </math>

<math> <mrow> <mo>&DoubleRightArrow;</mo> <msub> <mi>x</mi> <mi>τ</mi> </msub> <mo>=</mo> <msqrt> <mfrac> <msub> <mi>ρ</mi> <mi>τ</mi> </msub> <msub> <mi>N</mi> <mi>t</mi> </msub> </mfrac> </msqrt> <msub> <mover> <mi>S</mi> <mo>~</mo> </mover> <mi>τ</mi> </msub> <mi>h</mi> <mo>+</mo> <msub> <mi>n</mi> <mi>τ</mi> </msub> <mo>,</mo> <mi>s</mi> <mo>.</mo> <mi>t</mi> <mo>.</mo> <mi>tr</mi> <mrow> <mo>(</mo> <msubsup> <mi>S</mi> <mi>τ</mi> <mo>*</mo> </msubsup> <msub> <mi>S</mi> <mi>τ</mi> </msub> <mo>)</mo> </mrow> <mo>=</mo> <msub> <mi>N</mi> <mi>t</mi> </msub> <msub> <mi>T</mi> <mi>τ</mi> </msub> </mrow> </math>

in the formula x_τ＝vec(X_τ) H ═ vec (h) denotes the received signal matrix X, respectively_τAnd the column-piled-up vector, p, of the channel matrix H_τRepresents the average transmit power over the training period,

<math> <mrow> <msub> <mover> <mi>S</mi> <mo>~</mo> </mover> <mi>τ</mi> </msub> <mo>=</mo> <mrow> <mo>(</mo> <msub> <mi>I</mi> <msub> <mi>N</mi> <mi>r</mi> </msub> </msub> <mo>&CircleTimes;</mo> <msub> <mi>S</mi> <mi>τ</mi> </msub> <mo>)</mo> </mrow> <mo>,</mo> </mrow> </math>

I_Nrrepresents N_rA dimensional unit array;

9. the receiver receives the signal X according to the above_τAnd the signalled training sequence S_τThe estimation value of the current channel coefficient is obtained by the following formula by adopting the Minimum Mean Square Error (MMSE) estimation criterion

In the formula

\hat{h} = vec (\hat{H})

Representing a channel estimation matrix

The corresponding estimated mean square error MSE is:

10. after the training is finished, the transmission of real useful data is started

After one frame transmission is finished, if the channel related information R is_hIf the value of the training sequence is changed, returning to the first step to re-determine a new training sequence, and performing channel estimation and data transmission of the next frame; if the channel related information R_hIf the value of (A) is not changed, the process returns to the step 8 to start the next frameChannel estimation and data transmission.

The technical effects of the present invention will be described in further detail below with reference to the accompanying drawings.

Referring to fig. 4, assuming that a 4-transmission 4-reception MIMO system has the same spatial correlation coefficient at both transmission and reception ends, an exponential correlation coefficient generation model is used, and R is set_r(i，j)＝R_t(i，j)＝r^|i-j|And i, j is equal to 1, 2, 3 and 4, r is less than or equal to 1, and the larger the value of r is, the stronger the spatial correlation is. The training length is 4 symbol periods. In fig. 4, 'ort' represents a conventional orthogonal training sequence, and 'opt' represents a training sequence employed in the present invention. It can be seen that, although the mean square error of the channel estimation corresponding to the two training sequences is gradually reduced as the signal-to-noise ratio is increased. However, the estimation performance of the training sequence of the invention is better than that of the traditional orthogonal training sequence, and the performance is improved more obviously along with the enhancement of the spatial correlation of the channel.

Referring to fig. 5, still assume a 4-transmission 4-reception MIMO system, taking the following three cases as an example:

(1) assuming no spatial correlation, take T_τ＝4；

(2) Assuming no receiving correlation, the transmitting end has space-domain correlation and the correlation matrix is not full rank, and has three eigenvalues of 2, 1.5 and 0.5, and tr (R) is satisfied_t) When K is equal to 12, take

(3) Further ignoring R_hThe smaller 4 positive eigenvalues in (K) 8 are taken, in which case

As can be seen from the relationship curves of the estimated mean square error with the change of the signal-to-noise ratio in the three cases given in fig. 5, although the training overhead of case (2) is reduced by 25% compared with case (1), the estimated variance is still smaller than that of case (1); case (3) uses 50% less training overhead than case (1), and its estimated variance is better than case (1) at low signal-to-noise ratios. This shows that the estimation performance in the case of the correlated channel can still be obtained with less training symbols than in the case of the independent channel. Meanwhile, the training sequence of the invention can make self-adaptive adjustment to the length of the training sequence according to the variation of the channel correlation strength and the estimation performance of the receiver, so as to reduce the training overhead to the maximum extent on the basis of meeting the estimation performance requirement, achieve the maximization of the data transmission rate and improve the utilization rate of the frequency band.

Claims

1. A MIMO related channel estimation method based on self-adaptive training sequence includes the following steps:

(2) Determining channel estimation from K valuesLength of training sequence of meter, i.e.

Wherein N is_rWhich indicates the number of receive antennas to be used,

represents rounding up;

When T is_τ≥N_tWhen it is taken

When T is_τ＜N_tWhen it is taken

Wherein N is_tDenotes the number of transmitting antennas, 0_(Tτ-Nt)×NtRepresents a (T)_τ-N_t)×N_tAll 0 matrix, superscript, of dimension^1/2To representSquare root of matrix, D_[1:Tτ，:] ^1/2Represents to take D^1/2Front T of_τRow, matrix D is one N_t×N_tA dimensional diagonal matrix;

(5) from the above matrix D_τGenerating a training sequence S_τI.e. by

(7) starting the training period, the transmitter will train symbols S_τTransmitted into a radio channel H via a transmitting antenna, the receiver being at T_τThe signal received in one symbol period may be represented as a T_τ×N_tMatrix X of dimensions_τ；

，

Wherein,

\hat{h} = vec (\hat{H}),

2. The channel estimation method of claim 1, wherein N_t×N_tThe dimension diagonal matrix D may be normalized to N by the traces satisfying matrix D_tT_τIs obtained by minimizing the following formula: