CN107682289B - A Marker Symbol Design Method for Marker-Assisted Clustering Receivers - Google Patents

Abstract

The invention belongs to the technical field of communication, and in particular relates to a marking symbol design method for a marking-assisted clustering receiver. The main method of the present invention is at the transmitting end: before sending the data information, the transmitter inserts the marking symbols, the matrix formed by the marking symbols is φ, and the content of the marking symbols is known at the receiving end; at the receiving end: the received The marker signal corresponding to the marker symbol is marked as Y, and the marker obtained through marker reconstruction is Y _R , and the received signal is formed into multiple categories through a clustering algorithm; the sending symbol is restored according to the clustering result and the marker symbol; it is characterized in that , the design method of the marker symbol is to minimize the error of the marker Y _R reconstructed from Y relative to the real center point. The beneficial effect of the present invention is that the present invention proposes an optimal marking symbol design method, which can improve the accuracy and clustering performance of reconstructed labels, thereby improving the detection performance of the receiver.

Description

A Marker Symbol Design Method for Marker-Assisted Clustering Receivers

technical field

The invention belongs to the technical field of communication, and in particular relates to a marking symbol design method for a marking-assisted clustering receiver.

Background technique

Traditional receivers perform symbol detection step by step. Firstly, the channel estimation is performed according to the pilot signal at the transmitting end, and then the transmitted symbols are recovered through a maximum likelihood detector or other simpler detectors. The marker-assisted clustering receiver directly clusters the signals received by the receiver, and uses the marker symbols sent by the transmitter to determine the corresponding transmitted symbols. Therefore, it is necessary to design markers that can improve the accuracy and clustering performance of reconstructed tags, thereby improving the detection performance of the receiver.

Contents of the invention

The purpose of the present invention is to address the above problems and propose a method for designing marker symbols for marker-assisted clustering receivers. The method proposed in the present invention is mainly aimed at marker-assisted clustering receivers. After receiving a signal, the receiver , first perform tag reconstruction, then use the clustering method to classify the received signals, and finally use the reconstructed tags to indicate the transmitted symbols corresponding to each category, so as to realize symbol detection.

Technical scheme of the present invention is as follows:

A marker symbol design method for marker-assisted clustering receivers, including:

At the transmitting end: before sending the data information, the transmitter inserts marking symbols, the matrix formed by the marking symbols is φ, and the content of the marking symbols is known by the receiving end;

At the receiving end: mark the signal corresponding to the received mark symbol as Y, mark the mark obtained through mark reconstruction as Y _R , form multiple categories through the clustering algorithm for the received signal; according to the clustering result and mark symbol resume sending symbols;

It is characterized in that the design method of the mark symbol is to minimize the error of the mark Y _R reconstructed from Y.

In the general technical solution of the present invention, for the marker-assisted clustering receiver, after receiving the signal, the receiver first performs marker reconstruction, the matrix of the marker symbols sent by the invention is specified as φ, and the marker symbols received at the receiving end The corresponding marker signal is Y, and the reconstructed marker at the receiving end is Y _R . Since the clustering receiver needs the initial center point to be closer to the real center point to achieve better convergence performance, the present invention aims to design the optimal φ so that the mark error reconstructed by Y is the smallest.

The beneficial effect of the present invention is that the present invention proposes an optimal marking symbol design method, which can improve the accuracy and clustering performance of reconstructed labels, thereby improving the detection performance of the receiver.

Description of drawings

Fig. 1 shows the schematic diagram of clustering receiver work that the present invention proposes;

Fig. 2 shows the marking symbol design method in OFDM of the present invention;

Fig. 3 shows the marking symbol design method in MIMO of the present invention;

Fig. 4 has shown the performance contrast of best and worst design method of OFDM system of the present invention;

Fig. 5 shows the performance comparison between the optimal marking scheme and the suboptimal marking scheme of the MIMO system of the present invention.

Detailed ways

The technical solutions of the present invention will be described in detail below in conjunction with the drawings and embodiments.

Fig. 1 shows the workflow of the clustering receiver proposed by the present invention. The transmitted symbols include marker symbols and data symbols, corresponding to marker signals and data signals at the receiving end. The marker signal is first reconstructed with markers, and the reconstructed markers are used to mark each category and serve as the initial center point of the clustering algorithm. After the algorithm converges, the estimated category center point and variance are obtained, and the corresponding categories are also kept marked. The data signals clustered into the same class are detected and judged as the transmitted symbols corresponding to the class.

Example 1

As shown in Fig. 2, this example is a marker symbol design method taking the OFDM system as an example.

For the OFDM system, φ is a vector, defined as s _L , correspondingly, Y is also a vector, defined as y _L , and Y _R is also a vector y _R . Assume that there are M subcarriers in total, and there are L paths to reach the receiving end. These time-domain channels are expressed as h ₀ , h ₁ ,…,h _L-1 , and the receiving end is performing cyclic prefix removal and fast discrete Fourier transform (FFT) After the process, the frequency domain signal at the nth moment can be expressed as

y _n ＝Λs _n +u _n ,

Among them, s _n refers to an OFDM symbol, s=[s _n,0 ,s _n,1 ,…,s _n,M-1 ] ^T , s _n,m represents the nth moment on the mth subcarrier Symbol, and selected from the symbol set S, S is the set of symbols that can be sent by the sender, such as the sender adopts QPSK modulation, then Λ=diag(H ₀ ,H ₁ ,…,H _M-1 ), represents the frequency domain response of the channel, and

This example illustrates the design scheme of the optimal mark from three aspects: ① minimum number of sent marks, ② mark reconstruction, ③ selection of mark position and mark symbol.

①Minimum number of sent tags: In order to reduce the number of sent symbols, this example uses the relationship between subcarriers and the relationship between sent symbols to perform tag reconstruction. The relationship between the frequency domain channel responses of each subcarrier can be expressed as

Designate marked subcarriers as p ₁ , p ₂ ,...p _T , where T is the number of marked subcarriers. These marked subcarriers can be expressed as

in

in l=0,...L-1, p _t =p ₁ ,...p _T . It can be proved that when T=L, B is an invertible matrix, so the frequency domain channel response on subcarrier m can be written as

H _m = w _m B ^-1 d (3)

in and

It can be seen from the formula (3) that when T=L, the marked subcarrier frequency domain channel response can represent the frequency domain channel response on all subcarriers, so at least L subcarriers must be marked at the sending end.

②Mark reconstruction: At the receiving end, the receiver obtains K=MQ classes through clustering, where Q represents the modulation order, for example, if the OFDM system adopts QPSK modulation, then Q=4. The real mean value of the signal received by the receiving end is μ _k =H _m S _q , where k=m+Mq, m=0,...,M-1, q=0,...,Q-1, S _q represents the An element of , it can be seen from formula (3), μ _k ＝H _m S _q ＝w _m B ^-1 dS ₀ f _q ＝a _k dS ₀ , where a _k ＝w _m B ^-1 f _q , In order to identify the groups obtained by clustering by the receiver, the present invention obtains K labels by way of label reconstruction. The token sent by the sender is The tokens received at the receiver can be expressed as The operator ° represents the multiplication of the corresponding elements of the matrix or vector, Preprocess the received tokens, in Then the reconstructed markup is in

③Selection of marker position and marker symbol: In order to determine the position of the transmitted symbol and what kind of marker symbol to send, the present invention needs to minimize the reconstruction error, namely

in where the optimal solution is l is a constant, where l∈{0,...M-1}. where operator Indicates rounding down, and mod(·,M) indicates modulo M addition operation. It can be seen from the optimal solution that it is optimal to select the symbols to be transmitted with the largest energy, and it is optimal to select the farthest distance between the subcarriers for transmitting marks.

Therefore, in this example, before sending data symbols, one marker symbol is respectively inserted on the L subcarriers that are farthest apart from each other. Taking M=64 as an example, the subcarriers that transmit markers should be separated by 32 subcarriers that do not transmit markers. Subcarriers, there is no requirement for the specific position of the subcarriers for sending markers. For example, marker symbols can be sent on the 1st and 33rd subcarriers, or marker symbols can be sent on the 25th and 57th subcarriers, but the subcarriers of the marker symbols are required to be sent The carriers are farthest apart. For the symbol to be sent, it needs to meet the maximum amplitude in the transmittable symbol set S. Taking the 16QAM method as an example in the OFDM system, the transmitted symbol can be in the set If the QPSK system is adopted, the sent marker symbols can be selected arbitrarily from the symbol set S.

Example 2

As shown in Figure 3, this example takes the MIMO system as an example, and the design method of the marking symbol is as follows:

For a general MIMO model with M _t transmit antennas and M _r receive antennas, the number of marker symbols T=M _t , is a matrix, correspondingly, It is also a matrix, Y _R =Y·V. Represents the signal received at time n, and the transmitted symbol s(n) is obtained from a finite codebook with K elements in total Selected from , p is the transmit power, H is the channel matrix from the transmitter to the receiver, its elements are independent of each other and is the symbol sent together by M _t transmit antennas, w(n) is the noise obeying the cyclic symmetric complex Gaussian (CSCG) distribution, namely w(n) and s(n) are independent of each other, then the signal received by the receiving antenna is expressed as:

Since the noise obeys the CSCG distribution, so that in a given Under the condition of , the signal received on the receiving antenna obeys the mean value CSCG distribution with variance ∑ _k , namely The distribution of signals received by all antennas of the receiver in this period satisfies

①Minimum number of sending marks: yes Doing the maximum rank decomposition has where U is a matrix of full rank. After passing through the channel, there is in If and only if U is a full-rank matrix, we have at this time can be uniquely recovered by φ, so M _t is the minimum number of marked symbols.

②Marker reconstruction: At the receiving end, the signal corresponding to the received marker symbol is a composition matrix in is the noise on the received signal. All other center points can be obtained by right multiplying the recovery matrix, namely

③ Selection of sent marker symbols: let the sum of the distance between the recovered center point and the real center point be the recovery error, then the optimal marker selection purpose is to minimize the recovery error, that is,

The optimal marker design criterion obtained by solving the problem is to minimize the marker matrix U In addition, if the number of transmitting antennas M _t is the m power of 2, and m is a natural number, then the optimal marking design criterion is from Select M _t orthogonal vectors with the largest energy to form the matrix U.

Therefore, in this example, before sending data symbols, M _t marker symbols are inserted, and each marker symbol vector is a possible sending symbol vector. Taking 4× ⁴ antenna array and 16QAM modulation as an example, there are 44 possible transmission symbol vectors. Since M _t =4 is 2 to the power of 2, the optimal marking design is to select 4 orthogonal vectors with the largest energy. Here is a selection strategy based on the Hadamard matrix, namely:

Fig. 4 has shown the performance comparison figure of the best and worst design method of the mark symbol that the present invention takes OFDM system as example, in this figure, the sending end adopts QPSK modulation mode, and M=64, N=5, L are set = 2, the worst design means that the subcarriers that send marker symbols are adjacent, in this simulation the 1st and 2nd subcarriers send marker symbols, the best design means that the subcarriers that send marker symbols are separated by 32 subcarriers, In this simulation, the 1st and 33rd subcarriers transmit marker symbols. It can be seen that there is a large performance difference between the two designs. This is because when the subcarriers transmitting marker symbols are close to each other, the Marker symbols, all reconstructed marker symbols have large errors.

Fig. 5 shows the performance comparison diagram of the best and worst design method of the marking symbol taking MIMO system as an example in the present invention, in this figure, M _t =M _r =2, the transmitting end adopts QPSK modulation mode, optimal design is the orthogonal vector with the maximum energy, and the suboptimal design is a matrix U composed of two non-orthogonal vectors with the same energy. Specifically, let the optimal marker design be The suboptimal marker design is Have

It can be seen that the performance of the cluster receiver with the optimal label design is significantly higher than that with the sub-optimal label.

Claims (1)

1. a method of tag symbol design for a tag assisted clustered receiver, comprising:

at the transmitting end: before a transmitter sends data information, inserting a marker symbol, wherein a matrix formed by the marker symbol is phi, and the content of the marker symbol is known by a receiving end;

at the receiving end: marking the mark signal corresponding to the received mark symbol as Y, and marking the mark signal obtained by mark reconstruction as Y_RForming a plurality of categories of the received signals through a clustering algorithm; recovering the sending symbol according to the clustering result and the mark symbol;

Wherein the design method of the mark symbol is that the mark Y reconstructed by Y is used_Rthe method comprises the following specific steps:

For an OFDM system, φ is a vector, defined as s_LAccordingly, Y is also a vector, defined as Y_L，Y_RIs also a vector, defined as y_RLet a total of M subcarriers, with L paths to the receiving end, these time-domain channels are denoted as h₀,h₁,…,h_L-1After the receiving end performs the processes of cyclic prefix removal and fast discrete fourier transform, the frequency domain signal at the nth time is represented as:

y_n＝Λs_n+u_n

Wherein s is_nRefers to an OFDM symbol, s ═ s_n,0,s_n,1,…,s_n,M-1]^T，s_n,mRepresents the symbol at the nth time instant on the mth subcarrier and is selected from a set of symbols, S, which is a set of symbols that the transmitting end can transmit, Λ ═ diag (H)₀,H₁,…,H_M-1) Which represents the frequency-domain response of the channel,And is

The design method of the mark symbol comprises the following steps:

The minimum number of transmitted markers: in order to reduce the number of transmitted symbols, the mark reconstruction is carried out by utilizing the relation between subcarriers and the relation between transmitted symbols, and the relation between frequency domain channel responses of the subcarriers is expressed as:

Designating the marked sub-carriers as p₁,p₂,…p_TWhere T is the number of marked subcarriers; these marked subcarriers are denoted as:

Wherein

WhereinWhen T ═ L, B is an invertible matrix, and the frequency domain channel response on subcarrier m is:

H_m＝w_mB^-1d (3)

WhereinAnd is

As shown in equation (3), when T is equal to L, the marked subcarrier frequency domain channel response may indicate frequency domain channel responses on all subcarriers, that is, the transmitting end should mark at least L subcarriers;

Secondly, label reconstruction: at a receiving end, a receiver obtains a class K as an MQ through clustering, wherein Q represents a modulation order; the true mean value of the received signal at the receiving end is mu_k＝H_mS_qWhere k is M + Mq, M is 0, …, M-1, Q is 0, …, Q-1, S_qRepresents an element of the set S, and is represented by formula (3):

μ_k＝H_mS_q＝w_mB^-1dS₀f_q＝a_kdS₀

Wherein a is_k＝w_mB^-1f_q，In order to identify the groups obtained by clustering of the receiver, K marks are obtained by using a mark reconstruction mode; the mark symbols sent by the sending end are as follows:

The mark received at the receiving end is denoted as

Wherein the operatorRepresenting the multiplication of corresponding elements of a matrix or vector,Preprocessing the received mark:

WhereinThe reconstructed mark isWherein

③ selecting the marking position and the marking symbol: minimizing reconstruction errors, i.e.

WhereinThe optimal solution is Operatorrepresents rounding down, mod (·, M) represents a modulo-M addition operation; it can be obtained from the optimal solution that the symbol to be transmitted is selected to be optimal with the largest energy, and the subcarriers for transmitting the marks are optimally located at the farthest positions from each other.