CN104113398A - MIMO blind channel estimation fuzziness removal method based on orthogonal space-time block codes - Google Patents

Abstract

The invention discloses a MIMO blind channel estimation ambiguity removal method based on orthogonal space-time block coding, comprising the following steps: establishing a MIMO space-time system, analyzing training sequence signals, grouping and processing data signals before transmission, and jointly Blind channel estimation is performed on the first data block by using blind and training sequence-based channel estimation methods, and the first data block is decoded by joint signal and channel two-dimensional search for maximum likelihood decoding, and the training sequence obtained by zero-forcing equalization is used. For the estimated value, channel estimation based on the training sequence is performed on the subsequent data blocks, and the maximum likelihood decoding is performed on the subsequent data blocks by one-dimensional search of the transmitted signal. The benefit of the present invention lies in that it adopts the joint two-dimensional search of signal and channel, utilizes the correlation between signals in orthogonal space-time coding, eliminates the inherent phase ambiguity of channel estimation results on the basis of blind channel estimation, and reduces the possibility of channel estimation. The solution range provides convenience for determining the phase of the decoded signal.

Description

Ambiguity Removal Method for MIMO Blind Channel Estimation Based on Orthogonal Space-Time Block Coding

technical field

The invention relates to a method for removing ambiguity of MIMO blind channel estimation, in particular to a method for removing ambiguity of MIMO blind channel estimation based on orthogonal space-time block coding, and belongs to the technical field of wireless communication. the

Background technique

Channel estimation refers to the process and method for the receiver to obtain channel state information. Since channel equalization and decoding at the receiving end usually require known channel state information to complete, the accuracy of channel estimation seriously affects the receiving performance and data transmission quality, which also makes wireless channel estimation and identification an important task in wireless communication signal processing. research field. Traditional channel estimation generally implements channel estimation by designing training sequences or inserting pilots in data packets. The disadvantage of these methods is that the channel capacity and spectrum utilization are obviously reduced. Although this loss is very small for a quasi-static channel, it cannot be ignored in high-speed wireless communications where the channel is time-varying. In addition, in a cooperative communication system, under the condition of ensuring effective cooperation with the transmitting end, the receiving end knows all or part of the training sequence, and non-blind or semi-blind channel estimation can be used, while in a non-cooperative communication system, since the receiving end The content of the training sequence used by the terminal to the transmitter is completely unknown. In order to realize the detection of MIMO signals, a blind channel estimation method must be used to obtain channel state information. the

In the blind channel estimation problem of MIMO system, it is impossible to realize the complete identification of the channel matrix and the recovery of the source signal only based on the observed signal, and there is a certain degree of ambiguity between the obtained channel response value and the actual channel. The ambiguity under the SIMO system is a scalar factor, and there are only amplitude and phase ambiguities, while under the MIMO system, the ambiguity is expressed as a matrix, including order ambiguity, phase ambiguity and amplitude ambiguity, that is, the estimated different antennas The sequence of the channels on the channel is misaligned, and the phase rotation and the overall scaling of the amplitude of the signal constellation diagram obtained by equalization and the original constellation diagram occur. For a communication system with constant modulus modulation, the amplitude ambiguity has no effect on signal detection, but if the sequence ambiguity and phase ambiguity cannot be removed, it will seriously affect the signal detection. In existing algorithms, part of the training sequence or pilot is generally inserted to correct the channel estimation value to solve the ambiguity problem, but in the blind signal detection scenario, the receiving end cannot know the training sequence or pilot, so the ambiguity The problem of degree is a major difficulty in blind channel estimation. the

Aiming at the problem of blind channel estimation in MIMO systems, many solutions based on time domain or frequency domain have been proposed. As a representative of time-domain methods, subspace methods have been extensively studied because of their simple structure and good performance. Gao F, Nallanathan A et al. proposed a subspace-based blind channel estimation algorithm in the article "Subspace-based blind channel estimation for SISO, MISO and MIMO-OFDM systems", and used the precoding matrix to solve the ambiguity problem. However, in non-cooperative communication, the precoding matrix is designed by the transmitter and unknown to the receiver, so this method cannot solve the ambiguity problem in non-cooperative MIMO communication. the

In "Frequency domain blind MIMO system identification based on second and higher order statistics", Binning Chen et al proposed a frequency domain channel estimation method using the second-order and higher-order statistics of the received signal, and studied the ambiguity problem , so it was cited by many later articles to deal with the ambiguity problem. However, this algorithm can only unify the phase deflection of the channel estimation value into a frequency-independent value, that is to say, there is still an unknown phase ambiguity, which needs to be corrected by using the real channel. In non-cooperative communication, the actual channel information is unknown, so this method cannot be applied to blind channel estimation in non-cooperative communication. the

In STBC-MIMO system, the ambiguity of blind channel estimation can be dealt with to a certain extent with the help of information provided by space-time coding. Choqueuse V et al. proposed a channel estimation method using higher-order statistics in "Blind channel estimation for STBC systems using higher-order statistics", and used STBC coding information to use the channel of several specific code patterns The estimated ambiguity matrix is limited to a set, but the ambiguity matrix cannot be determined accurately, so this method cannot solve the ambiguity problem of blind channel estimation in non-cooperative communication systems. the

Contents of the invention

In order to solve the deficiencies in the prior art, the object of the present invention is to provide a method for removing the ambiguity of MIMO blind channel estimation, which is based on Orthogonal Space-Time Block Code (OSTBC), which can effectively solve the non-cooperative Ambiguity Problems for Blind Channel Estimation in MIMO Communications. the

In order to achieve the above object, the present invention adopts the following technical solutions:

A MIMO blind channel estimation ambiguity removal method based on orthogonal space-time block coding, is characterized in that, comprises the following steps:

(1) Establish, analyze and process the MIMO space-time system model:

(1) Establish a MIMO space-time system with N _t transmitting antennas and N _r receiving antennas. Each frame signal transmitted by the transmitting antennas is composed of training sequence, cyclic prefix and data. Cross-training sequence, the channel is a flat and slow-varying Rayleigh fading channel, assuming that the continuous N _f frame signals experience the same channel state;

(2) Analyze the time-domain orthogonal training sequence signal:

Assuming that the time-domain orthogonal training sequence signal of any antenna at time k is represented by N _t ×1-dimensional complex vector s _tr (k), the received signal is expressed as:

y _tr (k)=Hs _tr (k)+v(k)

In the formula, H is the N _r ×N _t Virelli fading channel response matrix, y _tr (k) is the N _r ×1-dimensional received signal vector, v(k) is the N _r ×1-dimensional noise vector, and the aforementioned noise obeys the mean is 0, and the variance is Gaussian distribution;

(3) Group data signals before transmission:

Let s(k)=[s ₁ (k)s ₂ (k)...s _N (k)] ^T is the kth data packet composed of N symbols to be transmitted, and each symbol is independently and identically distributed;

(4) Process the kth data packet s(k):

Map s(k) to N _t ×L-dimensional coding matrix C(k) after space-time coding:

$\mathrm{<span\; class="notranslate">\; C</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; N</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; A</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; n</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; j</span>}{\mathrm{<span\; class="notranslate">\; B</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; In</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span>$

In the formula, A _n and B _n are the encoding matrices corresponding to the real part (s _Rn (k)) and the imaginary part (s _In (k)) of the nth symbol s _n (k) respectively, and when L is the encoding matrix The number of slots, the received signal is expressed as:

Y(k)＝HC(k)+V(k)

In the formula, Y(k) is the N _r ×L dimensional received signal matrix, V(k) is the N _r ×L dimensional noise matrix, and Y(k) and V(k) obey the mean value of 0 and the variance of Gaussian distribution;

(2), channel estimation and decoding:

(1) Output the estimated signal of the first data block:

First, blind channel estimation is performed on the first data block by joint use of blind channel estimation method and channel estimation method based on time-domain orthogonal training sequence; then, maximum likelihood is performed on the first data block by joint two-dimensional search of signal and channel Decode and output the estimated signal of the first data block;

(2) Output the estimated signal of subsequent data blocks:

Firstly, the channel estimation based on the time-domain orthogonal training sequence for subsequent data blocks is performed using the time-domain orthogonal training sequence estimation value obtained by zero-forcing equalization, so that the estimated channel of the subsequent data blocks has the same phase ambiguity as the first data block ; Then, perform maximum likelihood decoding on the subsequent data blocks by using one-dimensional search of the transmitted signal, and output estimated signals of the subsequent data blocks. the

The aforementioned MIMO blind channel estimation ambiguity removal method based on orthogonal space-time block coding is characterized in that, when performing blind channel estimation on the first data block, the chu sequence is used as the time-domain orthogonal training sequence, and Alamouti is used to send plan. the

The aforementioned MIMO blind channel estimation ambiguity removal method based on orthogonal space-time block coding is characterized in that, when performing blind channel estimation on the first data block, the QPSK modulation sequence is used as the time-domain orthogonal training sequence, and the Alamouti Send proposal. the

The advantage of the present invention is that by using the joint two-dimensional search of the signal and the channel, the correlation between the signals in the orthogonal space-time coding is used to eliminate the inherent phase ambiguity of the channel estimation result on the basis of the blind channel estimation, and reduce the channel size. range of possible solutions, so as to provide convenience for determining the phase of the decoded signal; in addition, this scheme uses blind and time-domain orthogonal training sequence-based channel estimation methods in different data blocks to decode data from different data blocks sent at one time The phase deviation of the phase deviation is unified to a fixed value, which solves the problem of decoding errors caused by the ambiguity of the blind channel in non-cooperative communication; the method of the present invention can be applied to blind identification and blind detection of various multi-antenna signals and cooperative communication signals . the

Description of drawings

Figure 1 is a structural diagram of sending information;

Fig. 2 is a flow chart of channel estimation and decoding;

Fig. 3 is the maximum likelihood decoding flowchart of the first frame;

Fig. 4 is the maximum likelihood decoding flow chart of follow-up frame;

Fig. 5 is the constellation diagram of the system receiving signal when the chu sequence is used as the time-domain orthogonal training sequence and the Alamouti transmission scheme is adopted;

Fig. 6 is the constellation diagram of the system equalization signal when the chu sequence is used as the time-domain orthogonal training sequence and the Alamouti transmission scheme is adopted;

Figure 7 is the decoding performance curve of the Alamouti transmission scheme when the chu sequence is used as the time-domain orthogonal training sequence;

Fig. 8 is the constellation diagram of the system receiving signal when the QPSK modulation sequence is used as the time-domain orthogonal training sequence and the Alamouti transmission scheme is adopted;

Figure 9 is a constellation diagram of the system equalized signal when the QPSK modulation sequence is used as the time-domain orthogonal training sequence and the Alamouti transmission scheme is adopted;

Figure 10 is the decoding performance curve of the Alamouti transmission scheme when the QPSK modulation sequence is used as the time-domain orthogonal training sequence;

FIG. 11 is a comparison diagram of the decoding performance of the Alamouti transmission scheme and the system with fully known channel information when the QPSK modulation sequence is used as the time-domain orthogonal training sequence. the

Detailed ways

The present invention will be specifically introduced below in conjunction with the accompanying drawings and specific embodiments. the

1. Establish, analyze and process MIMO space-time system model

1. Establish MIMO space-time system model

Establish a MIMO space-time system with N _t transmitting antennas and N _r receiving antennas. Each frame signal transmitted by the transmitting antenna is composed of training sequence, cyclic prefix and data. The channel is a flat and slow-varying Rayleigh fading channel. It is assumed that the continuous N _f frame signals experience the same channel state, that is, a data block is composed of N _f frames, as shown in Fig. 1 .

The training sequence in the wireless communication system is divided into a time domain orthogonal training sequence, a frequency domain orthogonal training sequence and a code domain orthogonal training sequence. In the space-time system model established by the present invention, the training sequence adopts the time domain orthogonal training sequence. sequence. the

2. Analyze time-domain orthogonal training sequence signals

Assuming that the time-domain orthogonal training sequence signal of any antenna at time k is represented by N _t ×1-dimensional complex vector s _tr (k), the received signal is expressed as:

y _tr (k)=Hs _tr (k)+v(k)

In the formula, H is the N _r ×N _t Virelli fading channel response matrix, y _tr (k) is the N _r ×1-dimensional received signal vector, v(k) is the N _r ×1-dimensional noise vector, where the noise obeys The mean is 0 and the variance is Gaussian distribution.

3. Dealing with MIMO space-time system model

Since the data signal needs to be space-time coded, it is grouped before transmission, let s(k)=[s ₁ (k)s ₂ (k)...s _N (k)] ^T is the N symbols to be transmitted The k-th data group composed of , and each symbol is independently and identically distributed.

Next, the k-th data packet s(k) is processed. Specifically, s(k) is mapped to an N _t ×L-dimensional coding matrix C(k) through space-time coding:

$\mathrm{<span\; class="notranslate">\; C</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; N</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; A</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; n</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; j</span>}{\mathrm{<span\; class="notranslate">\; B</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; In</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span>$

In the formula, A _n and B _n are the encoding matrices corresponding to the real part (s _Rn (k)) and the imaginary part (s _In (k)) of the nth symbol s _n (k) respectively, and when L is the encoding matrix number of slots. Then the received signal can be expressed as:

Y(k)＝HC(k)+V(k)

In the formula, Y(k) is the N _r ×L dimensional received signal matrix, V(k) is the N _r ×L dimensional noise matrix, and Y(k) and V(k) obey the mean value of 0 and the variance of Gaussian distribution.

In addition, because the channel changes slowly, the channels of adjacent data block n+1 and data block n follow the following relationship:

${\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; =</span>\sqrt{\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; \&beta;</span>}}<span\; class="notranslate">\; \&Center\; Dot;</span>{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; +</span>\sqrt{\mathrm{<span\; class="notranslate">\; \&beta;</span>}}<span\; class="notranslate">\; \&Center\; Dot;</span>\mathrm{<span\; class="notranslate">\; \&Delta;H</span>}$

H _n and △H obey the complex Gaussian distribution, β is a constant coefficient, and the smaller the β value, the smaller the channel change.

2. Channel estimation and decoding

In previous blind channel estimation algorithms, the phase ambiguity generated by each channel estimation is different and random, which brings great difficulties to correct decoding. the

Next, referring to Fig. 2, the channel estimation and decoding scheme involved in the present invention will be introduced in detail. the

1. Perform channel estimation and decoding on the first data block

Channel estimation part: jointly use the blind channel estimation method and the channel estimation method based on the time-domain orthogonal training sequence to perform blind channel estimation on the first data block, and unify the phase ambiguity of the first data block to a fixed value. Values are determined by semantic decisions. the

Decoding part: Due to the inherent ambiguity of blind channel estimation, there are multiple sets of solutions for channel estimation. In the decoding part, the present invention uses the joint two-dimensional search of the signal and the channel, utilizes the correlation between the signals in the space-time coding, and performs maximum likelihood decoding on the first data block, and further narrows down possible solutions of the channel on the basis of channel estimation. range, which facilitates the determination of the final phase ambiguity. the

In order to unify the phase deflection of the decoded signal caused by the blind channel estimation ambiguity of all data blocks, we only perform the two-dimensional search of the signal and channel in the maximum likelihood decoding of the first data block. The specific process is shown in Figure 3. the

When performing blind channel estimation on the first data block, since the blind channel estimation is completed by using time-domain orthogonal training sequences, the performance of the scheme is different for different time-domain orthogonal training sequences. the

In recent years, many multi-antenna space-time processing schemes for different communication environments have emerged. Among them, the space-time block code (STBC) scheme improves connection stability and increases data transmission rate by providing diversity gain, and is a very representative space-time scheme. However, its performance depends heavily on the accuracy of channel estimation. the

Taking the classic Alamouti scheme as an example, its encoding matrix can be expressed as $C_{\mathrm{STBC}}=\left[\begin{array}{cc}{\mathrm{the\; s}}_1& -{\mathrm{the\; s}}_2^{*}\\ {\mathrm{the\; s}}_2& {\mathrm{the\; s}}_1^{*}\end{array}\right].$ Therefore, in the following, we will take two transmitting antennas as an example to discuss the channel estimation and decoding problems of the Alamouti transmission scheme using the chu sequence and the QPSK modulation sequence as the time-domain orthogonal training sequence.

(1), using the chu sequence as a time-domain orthogonal training sequence

When noise is not considered, there is a phase ambiguity of 180° or 0° in the channel matrix of the entire MIMO system. The specific derivation process is as follows:

Let h ₁ and h ₂ represent the real sub-channels of the two transmitting antennas respectively, and s ₁ and s ₂ represent the actual transmitted signal, then the received signal is:

$\left[\begin{array}{cc}{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}& {\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}\end{array}\right]\left[\begin{array}{cc}{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}& <span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}^{<span\; class="notranslate">\; *</span>}\\ {\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}& {\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}^{<span\; class="notranslate">\; *</span>}\end{array}\right]<span\; class="notranslate">\; =</span>\left[\begin{array}{cc}{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}& <span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}^{<span\; class="notranslate">\; *</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}^{<span\; class="notranslate">\; *</span>}\end{array}\right]$

make represents the channel estimate, Indicates the final decoding result. If the sub-channel estimates of the two antennas have a 180° phase deviation, when decoding with such a channel, we can get:

$\left[\begin{array}{cc}\stackrel{<span\; class="notranslate">\; ^</span><span\; class="notranslate">\; ^</span>}{{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}}& {\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}\end{array}\right]\left[\begin{array}{cc}\stackrel{<span\; class="notranslate">\; ^</span><span\; class="notranslate">\; ^</span>}{{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}}& <span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}^{<span\; class="notranslate">\; *</span>}\\ \stackrel{<span\; class="notranslate">\; ^</span><span\; class="notranslate">\; ^</span>}{{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}}& {\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}^{<span\; class="notranslate">\; *</span>}\end{array}\right]<span\; class="notranslate">\; =</span>\left[\begin{array}{cc}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}\stackrel{<span\; class="notranslate">\; ^</span><span\; class="notranslate">\; ^</span><span\; class="notranslate">\; ^</span><span\; class="notranslate">\; ^</span>}{{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; -</span>}}{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}& {\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}^{<span\; class="notranslate">\; *</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}^{<span\; class="notranslate">\; *</span>}\end{array}\right]$

make Then α ₁ =α ₂ =-1, the result is the same as the received signal, and the error vector is 0, so in this case, the decoded signal will have a phase shift of 180°.

It can be seen from the derivation that for the Alamouti transmission scheme, the decoding part has a certain selection effect on the phase ambiguity, so that there are only two possible final decoding results, namely (s ₁ s ₂ ) and (-s ₁ -s ₂ ), where , (s ₁ s ₂ ) is the correct decoding.

(2), using the QPSK modulation sequence as a time-domain orthogonal training sequence

When the noise is not considered, the channel matrix of the whole MIMO system has a phase ambiguity of 0°, 180°, 90° or -90°. The specific derivation process is as follows:

Let h ₁ and h ₂ represent the real sub-channels of the two transmitting antennas respectively, and s ₁ and s ₂ represent the actual transmitted signal, then the received signal is:

$\left[\begin{array}{cc}{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}& {\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}\end{array}\right]\left[\begin{array}{cc}{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}& <span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}^{<span\; class="notranslate">\; *</span>}\\ {\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}& {\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}^{<span\; class="notranslate">\; *</span>}\end{array}\right]<span\; class="notranslate">\; =</span>\left[\begin{array}{cc}{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}& <span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}^{<span\; class="notranslate">\; *</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}^{<span\; class="notranslate">\; *</span>}\end{array}\right]$

make represents the channel estimate, Indicates the final decoding result. When decoding with such a channel, α ₁ and α ₂ can be 1, -1, i or -i respectively, except for the case of correct decoding (α ₁ =α ₂ =1), there are three cases we Need to consider:

① If the sub-channel estimates of the two antennas have a 180° phase deflection, we can get:

$\left[\begin{array}{cc}\stackrel{<span\; class="notranslate">\; ^</span><span\; class="notranslate">\; ^</span>}{{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}}& {\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}\end{array}\right]\left[\begin{array}{cc}\stackrel{<span\; class="notranslate">\; ^</span><span\; class="notranslate">\; ^</span>}{{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}}& <span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}^{<span\; class="notranslate">\; *</span>}\\ \stackrel{<span\; class="notranslate">\; ^</span><span\; class="notranslate">\; ^</span>}{{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}}& {\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}^{<span\; class="notranslate">\; *</span>}\end{array}\right]<span\; class="notranslate">\; =</span>\left[\begin{array}{cc}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}\stackrel{<span\; class="notranslate">\; ^</span><span\; class="notranslate">\; ^</span><span\; class="notranslate">\; ^</span><span\; class="notranslate">\; ^</span>}{{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; -</span>}}{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}& {\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}^{<span\; class="notranslate">\; *</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}^{<span\; class="notranslate">\; *</span>}\end{array}\right]$

make Then α ₁ =α ₂ =-1, the result is the same as the received signal, and the error vector is 0, so in this case, the decoded signal will have a phase shift of 180°.

② If there is a 90° phase shift in the first sub-channel estimate and a -90° phase shift in the second sub-channel estimate, we can get:

$\left[\begin{array}{cc}{\mathrm{<span\; class="notranslate">\; i\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}& <span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; i\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}\end{array}\right]\left[\begin{array}{cc}\stackrel{<span\; class="notranslate">\; ^</span><span\; class="notranslate">\; ^</span>}{{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}}& <span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}^{<span\; class="notranslate">\; *</span>}\\ \stackrel{<span\; class="notranslate">\; ^</span><span\; class="notranslate">\; ^</span>}{{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}}& {\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}^{<span\; class="notranslate">\; *</span>}\end{array}\right]<span\; class="notranslate">\; =</span>\left[\begin{array}{cc}{\mathrm{<span\; class="notranslate">\; i\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}\stackrel{<span\; class="notranslate">\; ^</span><span\; class="notranslate">\; ^</span><span\; class="notranslate">\; ^</span><span\; class="notranslate">\; ^</span>}{{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; -</span>}}{\mathrm{<span\; class="notranslate">\; i\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}& <span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; i\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}^{<span\; class="notranslate">\; *</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; i\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}^{<span\; class="notranslate">\; *</span>}\end{array}\right]$

make Then α ₁ =-i, α ₂ =i, the result is the same as the received signal, and the error vector is 0, so in this case, the decoded signal of antenna 1 has a -90° phase deflection, and the decoded signal of antenna 2 has a 90° phase deflection.

③ If there is a -90° phase shift in the first sub-channel estimate and a 90° phase shift in the second sub-channel estimate, we can get:

$\left[\begin{array}{cc}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; i\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}& {\mathrm{<span\; class="notranslate">\; i\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}\end{array}\right]\left[\begin{array}{cc}\stackrel{<span\; class="notranslate">\; ^</span><span\; class="notranslate">\; ^</span>}{{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}}& <span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}^{<span\; class="notranslate">\; *</span>}\\ \stackrel{<span\; class="notranslate">\; ^</span><span\; class="notranslate">\; ^</span>}{{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}}& {\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}^{<span\; class="notranslate">\; *</span>}\end{array}\right]<span\; class="notranslate">\; =</span>\left[\begin{array}{cc}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; i\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}\stackrel{<span\; class="notranslate">\; ^</span><span\; class="notranslate">\; ^</span><span\; class="notranslate">\; ^</span><span\; class="notranslate">\; ^</span>}{{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; +</span>}}{\mathrm{<span\; class="notranslate">\; i\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}& {\mathrm{<span\; class="notranslate">\; i\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}^{<span\; class="notranslate">\; *</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; i\; h</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}^{<span\; class="notranslate">\; *</span>}\end{array}\right]$

make Then α ₁ =i, α ₂ =-i, the result is the same as the received signal, and the error vector is 0, so in this case, the decoded signal of antenna 1 has a 90° phase deflection, and the decoded signal of antenna 2 has a -90° phase deflection.

To sum up, when using the QPSK modulation sequence as the time-domain orthogonal training sequence to perform channel estimation and decoding on the Alamouti transmission scheme, the decoding results have the following four possibilities: (s ₁ s ₂ ), (-s ₁ -s ₂ ) , (-is ₁ is ₂ ), (is ₁ -is ₂ ).

2. Channel estimation and decoding of subsequent data blocks

Channel estimation part: use the estimated value of the time-domain orthogonal training sequence obtained by zero-forcing equalization to perform channel estimation based on the time-domain orthogonal training sequence for the subsequent data blocks, so that the estimated channel of the subsequent data blocks has the same Phase ambiguity, so that the phase ambiguity of different data blocks sent at one time is unified to a fixed value. the

Decoding part: perform maximum likelihood decoding on the subsequent data blocks by one-dimensional search of the transmitted signal, and output estimated signals of the subsequent data blocks. the

It can be seen that after further determining the channel (that is, after channel estimation and decoding of the first data block), we use the general maximum likelihood algorithm for decoding the subsequent data blocks, that is, use the one-dimensional search of the transmitted signal to do the maximum likelihood Then decode, so as to ensure that the phase offsets carried by the decoded signals of different data blocks are the same. The specific process is shown in FIG. 4 . the

3. Scheme simulation and performance analysis

In order to obtain the estimation performance, we use the time-domain orthogonal training sequence composed of chu sequence and the time-domain orthogonal training sequence composed of QPSK modulation sequence to estimate the channel by direct SVD decomposition method, the phase ambiguity is eliminated by the real channel, and the statistics Bit error rate, the results are as follows:

1. The chu sequence is used as a time-domain orthogonal training sequence, using the Alamouti transmission scheme

Simulation conditions: establish a MIMO space-time system with two transmissions and eight receptions, adopt the Alamouti transmission scheme, send and receive two data blocks, each data block has 4 frames. The time-domain orthogonal training sequence in the transmitted data uses the chu sequence, which consists of 32 symbols, and the data part consists of 64 symbols. A total of 1000 Monte Carlo simulations were performed. the

The constellation diagram of the received signal is shown in Figure 5. the

The constellation diagram of the equalized signal is shown in Figure 6. the

The decoding performance curve is shown in Figure 7. the

It can be seen that using the chu sequence as the time-domain orthogonal training sequence, under the Alamouti space-time coding scheme, the proposed algorithm can better balance the received signals that are aliased together due to the influence of the channel, so as to correctly recover the system transmitted signal. And the bit error rate decreases with the increase of the signal-to-noise ratio, reaching 10 ^-2 at 12dB.

2. The QPSK modulation sequence is used as a time-domain orthogonal training sequence, using the Alamouti transmission scheme

Simulation conditions: Establish a MIMO space-time system with two transmissions and eight receptions, using the Alamouti transmission scheme, sending and receiving two data blocks, each with 4 frames. The time-domain orthogonal training sequence in the transmitted data uses a QPSK modulation sequence, which consists of 32 symbols, and the data part consists of 64 symbols. A total of 1000 Monte Carlo simulations were performed to calculate the bit error rate. the

The constellation diagram of the received signal is shown in Figure 8. the

The constellation diagram of the equalized signal is shown in Figure 9. the

See Figure 10 for the decoding performance curve. the

It can be seen that using the QPSK modulation sequence as the time-domain orthogonal training sequence, under the Alamouti space-time coding scheme, the proposed algorithm can better balance the received signals that are aliased together due to the influence of the channel, so as to correctly restore the system transmitted signal , and the bit error rate decreases with the increase of the signal-to-noise ratio, reaching 10 ^-2 at 15dB.

When the QPSK modulation sequence is used as the time-domain orthogonal training sequence, the decoding performance of the Alamouti transmission scheme is compared with that of the system with completely known channel information, as shown in Fig. 11 . the

It can be seen that in the scenario considered in this patent, the receiving end cannot know the specific content of the time-domain orthogonal training sequence and the channel state information. There is a certain gap in the bit error rate performance of decoding, but the normal communication of the system can also be realized to a certain extent. the

In summary, the method of the present invention first uses the joint two-dimensional search of the signal and the channel, utilizes the correlation between the signals in the space-time coding, eliminates the inherent phase ambiguity of the channel estimation result on the basis of the blind channel estimation, and reduces the possibility of the channel. The scope of the solution, thus providing convenience for determining the phase of the decoded signal. the

In addition, the solution of the present invention unifies the phase deviations of decoded data of different data blocks sent at one time to a fixed value by using blind and time-domain orthogonal training sequence-based channel estimation methods in different data blocks. the

In a word, the method of the present invention solves the influence of blind channel estimation inherent phase ambiguity on system decoding. the

The method of the invention can be applied to blind identification and blind detection of various multi-antenna signals and cooperative communication signals. the

It should be noted that the above embodiments do not limit the present invention in any form, and all technical solutions obtained by means of equivalent replacement or equivalent transformation fall within the protection scope of the present invention. the

Claims (3)

1. MIMO blind channel estimation ambiguity removal method based on Orthogonal Space Time Block Coding, is characterized in that, comprises the following steps: (1) Establish, analyze and process the MIMO space-time system model: (1) Establish a MIMO space-time system with N _t transmit antennas and N _r receive antennas. Each frame signal transmitted by the transmit antennas consists of three parts: training sequence, cyclic prefix and data. The training sequence is the time domain Orthogonal training sequence, the channel is a flat and slow-varying Rayleigh fading channel, assuming that the continuous N _f frame signals experience the same channel state; (2) Analyze the time-domain orthogonal training sequence signal: Assuming that the time-domain orthogonal training sequence signal of any antenna at time k is represented by N _t ×1-dimensional complex vector s _tr (k), the received signal is expressed as: y _tr (k)=Hs _tr (k)+v(k) In the formula, H is the N _r ×N _t Virelli fading channel response matrix, y _tr (k) is the N _r ×1-dimensional received signal vector, v(k) is the N _r ×1-dimensional noise vector, and the noise obeys The mean is 0 and the variance is Gaussian distribution; (3) Packet the data signal before transmitting: Let s(k)=[s ₁ (k)s ₂ (k)...s _N (k)] ^T is the kth data packet composed of N symbols to be transmitted, and each symbol is independently and identically distributed; (4) Process the kth data packet s(k): Map s(k) to N _t ×L-dimensional coding matrix C(k) after space-time coding: $\mathrm{<span\; class="notranslate">\; C</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; N</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; A</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; Rn</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; j</span>}{\mathrm{<span\; class="notranslate">\; B</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}}{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; In</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span>$ In the formula, A _n and B _n are the encoding matrices corresponding to the real part (s _Rn (k)) and the imaginary part (s _In (k)) of the nth symbol s _n (k) respectively, and when L is the encoding matrix The number of slots, the received signal is expressed as: Y(k)=HC(k)+V(k) In the formula, Y(k) is the N _r ×L dimensional received signal matrix, V(k) is the N _r ×L dimensional noise matrix, and Y(k) and V(k) obey the mean value of 0 and the variance of Gaussian distribution; (2), channel estimation and decoding: (1) Perform channel estimation and decoding on the first data block: First, blind channel estimation is performed on the first data block by joint use of blind channel estimation method and channel estimation method based on time-domain orthogonal training sequence; then, maximum likelihood is performed on the first data block by joint two-dimensional search of signal and channel Decode and output the estimated signal of the first data block; (2) Perform channel estimation and decoding on subsequent data blocks: Firstly, the channel estimation based on the time-domain orthogonal training sequence for subsequent data blocks is performed using the time-domain orthogonal training sequence estimation value obtained by zero-forcing equalization, so that the estimated channel of the subsequent data blocks has the same phase ambiguity as the first data block ; Then, perform maximum likelihood decoding on the subsequent data blocks by using one-dimensional search of the transmitted signal, and output estimated signals of the subsequent data blocks. 2. the MIMO blind channel estimation ambiguity removal method based on orthogonal space-time block coding according to claim 1, is characterized in that, When performing blind channel estimation on the first data block, the chu sequence is used as the time-domain orthogonal training sequence, and the Alamouti transmission scheme is adopted. 3. the MIMO blind channel estimation ambiguity removal method based on orthogonal space-time block coding according to claim 1, is characterized in that, When performing blind channel estimation on the first data block, the QPSK modulation sequence is used as the time-domain orthogonal training sequence, and the Alamouti transmission scheme is adopted.