A channel estimation method and device for OTFS modulation

Abstract

The present invention discloses a channel estimation method and device for OTFS modulation. The method comprises the following steps: obtaining a transmission symbol matrix X, a maximum Doppler frequency shift, a delay-Doppler grid dimension value, and a sampling time interval of the OTFS modulation; constructing a basis function coefficient matrix solution model based on the transmission symbol matrix X, the maximum Doppler frequency shift, the delay-Doppler grid dimension value, and the sampling time interval; and solving the basis function coefficient matrix solution model based on the transmission symbol matrix X to obtain a time domain channel response matrix of the OTFS modulation.

Description

Technical Field

[0001] The present invention relates to the fields of OTFS modulation technology, high-reliability communication and channel estimation, and in particular to a channel estimation method and device for OTFS modulation. Background Art

[0002] Ensuring the reliability and throughput of information transmission in high-speed mobility scenarios is a hot research topic. While the high-speed mobile channels associated with these scenarios exhibit rapidly varying characteristics in the time-frequency domain, they can be represented as time-invariant channels in the delay-Doppler domain using impulse responses. Therefore, delay and Doppler frequency offset can be easily estimated in the delay-Doppler domain. OTFS modulation effectively mitigates these rapidly varying channel characteristics by representing signals in the delay-Doppler domain.

[0003] In actual communication scenarios, since the number of reflectors is necessarily limited, the number of multipaths is also limited. The channel has a sparse characteristic in the delay-Doppler domain, which greatly reduces the overhead of pilot symbols used for channel estimation and solves the problem of reduced throughput caused by increasing pilot symbols. OTFS signals are subject to interference in dual-selection channels, and the receiver must perform channel estimation to correctly recover the signal. Channel estimation is crucial to the normal operation of the receiver, and its performance directly affects the subsequent equalization and signal detection results. Currently, the more popular method is the channel estimation algorithm based on embedded pilot assistance, but this method has a high peak-to-average ratio, making it difficult to use in actual engineering applications. Summary of the Invention

[0004] The present invention mainly solves the problem of how to achieve accurate estimation of a fast time-varying channel based on OFTS modulation. The present invention discloses a channel estimation method and device for OTFS modulation.

[0005] In a first aspect, an embodiment of the present invention discloses a channel estimation method for OTFS modulation, including:

[0006] S1, obtain the transmit symbol matrix X, maximum Doppler frequency shift, delay-Doppler grid dimension value, and sampling time interval of OTFS modulation;

[0007] S2, constructing a basis function coefficient matrix solution model based on the transmit symbol matrix X, the maximum Doppler frequency shift, the delay-Doppler grid dimension value and the sampling time interval;

[0008] S3: Solve the basis function coefficient matrix solution model based on the transmit symbol matrix X to obtain a time domain channel response matrix of OTFS modulation.

[0009] The method of constructing a basis function coefficient matrix solution model based on the transmit symbol matrix X, the maximum Doppler frequency shift, the delay-Doppler grid dimension value, and the sampling time interval includes:

[0010] S21, calculating a dimension value Q based on the maximum Doppler shift, the delay-Doppler grid dimension value, and the sampling time interval;

[0011] S22, constructing a first information matrix B based on the transmitted symbol matrix X p ;

[0012] S23, based on the first information matrix B p , a basis function coefficient matrix solution model is constructed; the expression of the basis function coefficient matrix solution model is:

[0013]

[0014] Among them, G is the basis function coefficient matrix to be solved, is the partial time domain channel matrix.

[0015] The first information matrix B is constructed based on the transmitted symbol matrix X. p ,include:

[0016] S221, performing pilot extraction processing on the transmission symbol matrix X to obtain a pilot matrix X p ;

[0017] S222, extract and obtain the pilot matrix X p Column vector of ;

[0018] S223, repeating square loop operations on all extracted column vectors to obtain a first information matrix B p .

[0019] The pilot extraction process is performed on the transmission symbol matrix X to obtain the pilot matrix X p ,include:

[0020] S2211, using the transmission symbol matrix X, calculate and obtain the reception symbol matrix Z;

[0021] S2212, using the received symbol matrix Z, calculate the pilot matrix X p ;

[0022] The expression of the received symbol matrix Z is:

[0023]

[0024] Z p [i,n]=Z[mp +i,n],i=0,…,l max ,

[0025] Where Z[m,n] is the element in the mth row and nth column of the received symbol matrix Z. is the channel feature matrix, W P The matrix dimension is l max ×N noise matrix, l max It represents the ratio of the maximum delay of the channel to the sampling time interval, M and N are the matrices H DD The row and column dimensions of the matrix, vec means arranging all column vectors of the matrix into one vector, m p =M / 2,m p is the intermediate position factor, Z p represents the first symbol matrix, Z p [i,n] represents the element in the i-th row and n-th column of the first symbol matrix, and X[|mk|,|nl|] represents the element in the |mk|-th row and |nl|-th column of the matrix X.

[0026] The expression for the repeated square loop operation process is:

[0027]

[0028] in, Denotes the pilot matrix X p is the i-th column vector of , N is the column dimension of the pilot matrix, and circ(·) represents a square circulant matrix operation.

[0029] The step of solving the basis function coefficient matrix solution model based on the transmit symbol matrix X to obtain the time domain channel response matrix of OTFS modulation includes:

[0030] S31, constructing a first information matrix B based on the transmitted symbol matrix X p ;

[0031] S32, performing eigenvector calculation processing on the channel characteristic matrix and the noise matrix to obtain a first vector and a second vector;

[0032] S33, performing partial channel response estimation processing on the first vector and the second vector to obtain a partial time domain channel matrix

[0033] S34, based on partial time domain channel matrix Solve the basis function coefficient matrix G to be solved and obtain the solution value of the basis function coefficient matrix G;

[0034] S35, performing channel response calculation on the solved value of the basis function coefficient matrix G to obtain a time domain channel response matrix of OTFS modulation;

[0035] The expression for calculating the channel response is:

[0036] H=G·B,

[0037] Where H is the time domain channel response matrix of OTFS modulation, and B is the basis function matrix.

[0038] The expression for the feature vector calculation process is:

[0039]

[0040] in, is the channel feature matrix, W p is the noise matrix, and are the first vector and the second vector respectively;

[0041] The expression for the partial channel response estimation process is:

[0042]

[0043] in, For the general The dimension of the rearranged columns is l max ×N matrix, is the third vector, is the Fourier transform matrix, the noise vector for The corresponding row vector, is the N×N dimensional inverse discrete Fourier transform matrix, and the upper right corner H is a conjugate symmetric operation.

[0044] According to a second aspect of an embodiment of the present invention, a channel estimation device for OTFS modulation is disclosed, the device comprising:

[0045] a memory storing executable program code;

[0046] a processor coupled to the memory;

[0047] The processor calls the executable program code stored in the memory to execute the channel estimation method for OTFS modulation.

[0048] According to a third aspect of an embodiment of the present invention, a computer-storable medium is disclosed. The computer-storable medium stores computer instructions. When the computer instructions are called by a computer, the computer instructions are used to execute the channel estimation method for OTFS modulation.

[0049] According to a fourth aspect of the embodiments of the present invention, an information data processing terminal is disclosed. The information data processing terminal is used to implement the channel estimation method for OTFS modulation.

[0050] The beneficial effects of the present invention are:

[0051] This paper proposes an improved embedded pilot channel estimation algorithm for OTFS modulation. Compared with traditional methods, the improved algorithm achieves a 10dB peak-to-average ratio gain at the cost of a 1.5dB performance loss at a bit error rate of 1e-4, making it more effectively applicable in practical engineering applications.

[0052] This method first estimates the local channel response using a delay-Doppler domain embedded pilot scheme, then converts it to the time domain and recovers the complete time-domain channel response using an optimized BEM method. This method effectively addresses the high peak-to-average ratio issue of traditional embedded pilot channel estimation algorithms, making it more practical in engineering applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 Flow chart for the implementation of the method of the present invention;

[0054] Figure 2 The CCDF performance comparison chart of different channel estimation algorithms;

[0055] Figure 3 This is a comparison chart of bit error rates for different channel estimation algorithms;

[0056] Figure 4 Schematic diagram of the construction of the double-circular block matrix B. DETAILED DESCRIPTION

[0057] In order to better understand the content of the present invention, an embodiment is given here.

[0058] Figure 1 4 is an implementation flow chart of the method of the present invention. Figure 2 The CCDF performance comparison chart of different channel estimation algorithms; Figure 3 The figure is a comparison of the bit error rates of different channel estimation algorithms. Figure 4 Schematic diagram of the construction of the double-circular block matrix B.

[0059] In a first aspect, an embodiment of the present invention discloses a channel estimation method for OTFS modulation, including:

[0060] S1, obtain the transmit symbol matrix X, maximum Doppler frequency shift, delay-Doppler grid dimension value, and sampling time interval of OTFS modulation;

[0061] S2, constructing a basis function coefficient matrix solution model based on the transmit symbol matrix X, the maximum Doppler frequency shift, the delay-Doppler grid dimension value, and the sampling time interval;

[0062] S3: Solve the basis function coefficient matrix solution model based on the transmit symbol matrix X to obtain a time domain channel response matrix of OTFS modulation.

[0063] The method of constructing a basis function coefficient matrix solution model based on the transmit symbol matrix X, the maximum Doppler frequency shift, the delay-Doppler grid dimension value, and the sampling time interval includes:

[0064] S21, calculating a dimension value Q based on the maximum Doppler shift, the delay-Doppler grid dimension value, and the sampling time interval;

[0065] S22, constructing a first information matrix B based on the transmitted symbol matrix X p ;

[0066] S23, based on the first information matrix B p , a basis function coefficient matrix solution model is constructed; the expression of the basis function coefficient matrix solution model is:

[0067]

[0068] Among them, G is the basis function coefficient matrix to be solved, is the partial time domain channel matrix.

[0069] The first information matrix B is constructed based on the transmitted symbol matrix X. p ,include:

[0070] Perform pilot extraction processing on the transmission symbol matrix X to obtain the pilot matrix X p ;

[0071] Extract the pilot matrix X p Column vector of ;

[0072] Repeat the square loop operation on all extracted column vectors to obtain the first information matrix B p ;

[0073] The pilot extraction process is performed on the transmission symbol matrix X to obtain the pilot matrix X p ,include:

[0074] Using the transmit symbol matrix X, a receive symbol matrix Z is calculated;

[0075] Using the received symbol matrix Z, the pilot matrix X is calculated p ;

[0076] The expression of the received symbol matrix Z is:

[0077]

[0078] Z p [i,n]=Z[m p +i,n],i=0,…,l max ,

[0079] Where Z[m,n] is the element in the mth row and nth column of the received symbol matrix Z. is the channel feature matrix, W P The matrix dimension is l max ×N noise matrix, l max It represents the ratio of the maximum delay of the channel to the sampling time interval, M and N are the matrices H DD The row and column dimensions of the matrix, vec means arranging all column vectors of the matrix into one vector, m p =M / 2,m p is the intermediate position factor, Z p represents the first symbol matrix, Z p [i,n] represents the element in the i-th row and n-th column of the first symbol matrix, and X[|mk|,|nl|] represents the element in the |mk|-th row and |nl|-th column of the matrix X.

[0080] The expression for the repeated square loop operation process is:

[0081]

[0082] in, Denotes the pilot matrix X p The i-th column vector of , N is the column dimension of the pilot matrix, and circ(·) represents a square circulant matrix operation;

[0083] The step of solving the basis function coefficient matrix solution model based on the transmit symbol matrix X to obtain the time domain channel response matrix of OTFS modulation includes:

[0084] S31, constructing a first information matrix B based on the transmitted symbol matrix X p ;

[0085] S32, performing eigenvector calculation processing on the channel characteristic matrix and the noise matrix to obtain a first vector and a second vector;

[0086] S33, performing partial channel response estimation processing on the first vector and the second vector to obtain a partial time domain channel matrix

[0087] S34, based on partial time domain channel matrix Solve the basis function coefficient matrix G to be solved and obtain the solution value of the basis function coefficient matrix G;

[0088] S35, performing channel response calculation on the solved value of the basis function coefficient matrix G to obtain a time domain channel response matrix of OTFS modulation;

[0089] The expression for calculating the channel response is:

[0090] H=G·B,

[0091] Where H is the time domain channel response matrix of OTFS modulation, and B is the basis function matrix;

[0092] The expression b of the qth row and nth column of the basis function matrix B is q (n) is:

[0093]

[0094] Among them, M and N are the dimensions of the delay-Doppler grid, which is also the matrix H DD The row and column dimensions of , P is a preset constant factor, f max and T s are the maximum Doppler shift and sampling time interval, respectively, Indicates rounding up;

[0095] The calculation expression of the solution value of the basis function coefficient matrix G is:

[0096]

[0097] The calculating the dimension value Q based on the maximum Doppler shift, the delay-Doppler grid dimension value and the sampling time interval includes: f max and T s are the maximum Doppler shift and sampling time interval respectively.

[0098] The expression for the feature vector calculation process is:

[0099]

[0100] in, is the channel feature matrix, W p is the noise matrix, and are the first and second vectors respectively;

[0101] The expression for the partial channel response estimation process is:

[0102]

[0103] in, For the general The dimension of the rearranged columns is l max ×N matrix, is the third vector, is the Fourier transform matrix, the noise vector for The corresponding row vector, is the N×N dimensional inverse discrete Fourier transform matrix, and the upper right corner H is a conjugate symmetric operation.

[0104] Channel characteristic matrix It can be calculated by the following formula: Represents Z p The corresponding vector.

[0105] In order to improve the channel feature matrix The estimation accuracy can be calculated at multiple moments Perform precision improvement processing;

[0106] The expression of the precision improvement process is:

[0107]

[0108] Among them, h ij is the element of the i-th row and j-th column of the channel feature matrix after precision improvement, ω1 and ω2 are the first weighting factor and the second weighting factor respectively, and h k,ij is the element of the i-th row and j-th column of the channel feature matrix calculated at the k-th moment, is the mean of the elements in the i-th row and j-th column of the channel feature matrix at all times, K is the total number of calculation moments, and μ is the preset division factor, which can be obtained by calculation The average value of the elements in all rows and columns is obtained.

[0109] Matrix B p The rows of are the rows of matrix B, matrix B p The column of is the M / 2+M×k-th column of matrix B, k=0,...,N-1.

[0110] This example simulates the performance of the channel estimation algorithm of the present invention and a traditional embedded pilot algorithm under TU channel conditions, using CCDF and BER as evaluation metrics. The specific simulation parameters are shown in Table 1. For ease of notation and understanding, the algorithms proposed in this section are denoted as Proposed-BEM, Proposed-Linear, Proposed-Spline, and Proposed-Cubic, respectively.

[0111] Table 1 Channel estimation simulation parameters

[0112]

[0113] Since the embedded pilot needs to set a very high pilot signal-to-noise ratio, in this simulation the pilot power is set to be 40dB higher than the noise power. Figure 2 It is obvious that the maximum peak-to-average ratio of the traditional embedded pilot method is as high as 20dB, which is about 10dB higher than the original signal, which will inhibit the efficiency of the power amplifier. -6 The magnitude is increased by about 0.71dB compared to the original signal. The increase in peak-to-average ratio is due to the use of zero-sequence padded protection symbols in the embedded pilot, which reduces the average power of the signal.

[0114] In the simulation parameters of this section, the maximum multipath delay l max =10, then the total number of pilot symbols and guard symbols used is (2l max +1)×N. Let the energy of each modulation symbol be σ 2 , from the above analysis, the pilot energy can be obtained as Theoretically, the average signal power should be reduced by 10×log10(M / (M-2l max -1))=0.77dB, that is, the peak-to-average ratio increases by 0.77dB. According to the actual simulation results, the peak-to-average ratio of the improved algorithm increases by 0.71dB, which is almost consistent with the theoretical results. Overall, CCDF has a good performance when the probability is 10 -6 In terms of order of magnitude, the use of the improved algorithm can effectively reduce the peak-to-average ratio by about 10dB compared to the traditional embedded pilot channel estimation method.

[0115] Figure 3 This is a comparison of the bit error rates of the improved channel estimation algorithm and the traditional algorithm. As can be seen from the figure, the bit error rate is 10 -4Compared to ideal channel estimation, the performance loss of the traditional embedded pilot-based channel estimation algorithm is approximately 2.5dB, while the optimized BEM algorithm suffers a 1.5dB performance loss. The unoptimized BEM algorithm and the three interpolation methods achieve nearly identical bit error rates, with a performance loss of approximately 2dB compared to the traditional embedded pilot channel estimation algorithm. Taking CCDF into account, the proposed algorithm, while exhibiting a performance loss compared to the traditional embedded pilot algorithm, effectively addresses the high peak-to-average ratio issue and facilitates hardware implementation.

[0116] For B=circ[circ(x1),…,circ(x N )], circ(·) represents a square circulant matrix, the matrix B is a double circulant block matrix, x i is the i-th column vector of matrix X, the formation of matrix B, that is, the operation process of circ(·), is as follows Figure 4 shown.

[0117] According to a second aspect of an embodiment of the present invention, a channel estimation device for OTFS modulation is disclosed, the device comprising:

[0118] a memory storing executable program code;

[0119] a processor coupled to the memory;

[0120] The processor calls the executable program code stored in the memory to execute the channel estimation method for OTFS modulation.

[0121] According to a third aspect of an embodiment of the present invention, a computer-storable medium is disclosed. The computer-storable medium stores computer instructions. When the computer instructions are called by a computer, the computer instructions are used to execute the channel estimation method for OTFS modulation.

[0122] According to a fourth aspect of the embodiments of the present invention, an information data processing terminal is disclosed. The information data processing terminal is used to implement the channel estimation method for OTFS modulation.

[0123] The foregoing is merely an embodiment of the present invention and is not intended to limit the present invention. It will be apparent to those skilled in the art that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention are intended to be included within the scope of the claims of the present invention.

Claims

1. A channel estimation method for OTFS modulation, characterized in that: include: S1, obtain the transmit symbol matrix X, maximum Doppler frequency shift, delay-Doppler grid dimension value, and sampling time interval of OTFS modulation; S2, based on the transmit symbol matrix X, the maximum Doppler frequency shift, the delay-Doppler grid dimension value and the sampling time interval, constructing a basis function coefficient matrix solution model, including: S21, calculating a dimension value Q based on the maximum Doppler shift, the delay-Doppler grid dimension value, and the sampling time interval; S22, constructing a first information matrix B based on the transmitted symbol matrix X p ; S23, based on the first information matrix B p , a basis function coefficient matrix solution model is constructed; the expression of the basis function coefficient matrix solution model is: Among them, G is the basis function matrix to be solved, is the partial time domain channel matrix; S3, based on the transmit symbol matrix X, solving the basis function coefficient matrix solution model to obtain the time domain channel response matrix of OTFS modulation, including: S31, constructing a first information matrix B based on the transmitted symbol matrix X p ; S32, performing eigenvector calculation processing on the channel characteristic matrix and the noise matrix to obtain a first vector and a second vector; S33, performing partial channel response estimation processing on the first vector and the second vector to obtain a partial time domain channel matrix S34, based on partial time domain channel matrix Solve the basis function coefficient matrix G to be solved and obtain the solution value of the basis function coefficient matrix G; S35, performing channel response calculation on the solved value of the basis function coefficient matrix G to obtain a time domain channel response matrix of OTFS modulation; The expression for calculating the channel response is: H=G·B, Where H is the time domain channel response matrix of OTFS modulation, and B is the basis function matrix.

2. The channel estimation method for OTFS modulation according to claim 1, wherein The first information matrix B is constructed based on the transmitted symbol matrix X. p ,include: S221, performing pilot extraction processing on the transmission symbol matrix X to obtain a pilot matrix X p ; S222, extract and obtain the pilot matrix X p Column vector of ; S223, repeating square loop operations on all extracted column vectors to obtain a first information matrix B p .

3. The channel estimation method for OTFS modulation according to claim 2, wherein The pilot extraction process is performed on the transmission symbol matrix X to obtain the pilot matrix X p ,include: S2211, using the transmission symbol matrix X, calculate and obtain the reception symbol matrix Z; S2212, using the received symbol matrix Z, calculate the pilot matrix X p ; The expression of the received symbol matrix Z is: Z p [i,n]=Z[m p +i,n],i=0,…,l max , Where Z[m,n] is the element in the mth row and nth column of the received symbol matrix Z. is the channel feature matrix, W P The matrix dimension is l max ×N noise matrix, l max It represents the ratio of the maximum delay of the channel to the sampling time interval, M and N are the matrices H DD The row and column dimensions of the matrix, vec means arranging all column vectors of the matrix into one vector, m p =M / 2,m p is the intermediate position factor, Z p represents the first symbol matrix, Z p [i,n] represents the element in the i-th row and n-th column of the first symbol matrix, and X[|mk|,|nl|] represents the element in the |mk|-th row and |nl|-th column of the matrix X.

4. The channel estimation method for OTFS modulation according to claim 3, wherein The expression for the repeated square loop operation process is: in, Denotes the pilot matrix X p is the i-th column vector of , N is the column dimension of the pilot matrix, and circ(·) represents a square circulant matrix operation.

5. The channel estimation method for OTFS modulation according to claim 1, wherein The expression for the feature vector calculation process is: in, is the channel feature matrix, W p is the noise matrix, and are the first and second vectors respectively; The expression for the partial channel response estimation process is: in, For the general The dimension of the rearranged columns is l max ×N matrix, is the third vector, is the Fourier transform matrix, the noise vector for The corresponding row vector, is the N×N dimensional inverse discrete Fourier transform matrix, and the upper right corner H is a conjugate symmetric operation.

6. A channel estimation device for OTFS modulation, characterized in that: The device comprises: a memory storing executable program code; a processor coupled to the memory; The processor calls the executable program code stored in the memory to execute the channel estimation method for OTFS modulation according to any one of claims 1 to 5.

7. A computer storable medium, characterized in that The computer storable medium stores computer instructions, and when the computer instructions are called by a computer, they are used to execute the channel estimation method for OTFS modulation according to any one of claims 1 to 5.

8. An information data processing terminal, characterized in that: The information data processing terminal is used to implement the channel estimation method for OTFS modulation according to any one of claims 1 to 5.

Images