High-reliability secure transmission method based on OTFS

Abstract

The invention discloses an OTFS-based high-reliability secure transmission method, which comprises the following steps of: sequentially carrying out LDPC (Low Density Parity Check) coding, QAM (Quadrature Amplitude Modulation) mapping, SVD (Singular Value Decomposition) pre-coding, inverse symplectic finite Fourier transform, Hisenberg transform, R-CP (Reverse-CP) addition, D / A (Digital-to-Analog) transform and up-conversion on an information source sequence of a transmitting end to obtain a transmitting sequence, and transmitting the transmitting sequence through a wireless channel; after a receiving end of a legal user obtains a receiving sequence, down-conversion, A / D conversion, R-CP removal, Wigner conversion, symplectic finite Fourier transform, SVD decoding, single-tap equalizer processing, QAM demapping and LDPC decoding are carried out to obtain an information sink sequence. Through a precoding technology based on an OTFS equivalent matrix and an error control coding technology based on LDPC, high-reliability physical layer secure transmission is realized in a high-dynamic communication scene, the total throughput and the spectrum efficiency of a system are improved, the peak-to-average power ratio and the transmission bit error rate of the system are reduced, and meanwhile, overhigh calculation complexity and time-frequency resource overhead are not brought.

Description

Technical Field

[0001] The present invention belongs to wireless communication technology, and in particular relates to a highly reliable and secure transmission method based on OTFS. Background Art

[0002] Orthogonal Frequency Division Multiplexing (OFDM) is a multi-carrier modulation technology widely used in 4G and 5G. Its core idea is to convert high-speed serial data streams into low-speed parallel data streams by multiplexing orthogonal subcarriers, and can effectively combat frequency selective fading and overcome inter-symbol interference (ISI). However, with the development of high-speed transportation and mobile communication services, in high-mobility scenarios, due to the Doppler effect, the subcarrier orthogonality of traditional OFDM will be destroyed, thereby introducing inter-carrier interference (ICI), which seriously affects the performance of the communication system.

[0003] To combat the high Doppler effect in high mobility scenarios, Orthogonal Time Frequency Space (OTFS) was proposed as a new multi-carrier modulation scheme. It converts the fast time-varying time-frequency (TF) channel into a time-invariant sparse delay-Doppler (DD) channel through a pair of two-dimensional orthogonal transformations. Its characteristic is that each DD domain symbol is flattened into the entire TF domain, thereby obtaining full time-frequency diversity gain, thereby effectively combating the ICI caused by the Doppler effect and achieving better bit error rate performance than traditional OFDM in high mobility scenarios.

[0004] Low Density Parity Check (LDPC) code is a linear block code with strong error correction capability. The parity check matrix contains relatively few zero elements. This sparsity makes its decoding complexity and minimum code distance grow only linearly with the code length, while for other typical block codes they grow exponentially. Therefore, the code length of LDPC code can be designed to be very long without causing excessive decoding complexity.

[0005] Due to the open nature of wireless channels, OTFS waveforms are easily detected and intercepted by external eavesdroppers during air interface transmission, resulting in the leakage of private information. Therefore, the OTFS communication system urgently needs a secure physical layer air interface transmission method while ensuring that the system transmission has good bit error rate performance. Summary of the invention

[0006] In order to solve the problem of OTFS physical layer security transmission and further improve the reliability of transmission, the present invention provides a high-reliability and secure transmission method based on OTFS. Through the precoding technology based on the OTFS equivalent matrix and the error control coding technology based on LDPC, high-reliability physical layer security transmission can be achieved in high dynamic communication scenarios, the system total throughput and spectrum efficiency can be improved, the system peak-to-average ratio and transmission bit error rate can be reduced, and at the same time, excessive computational complexity and time-frequency resource overhead will not be brought about.

[0007] In order to achieve the above tasks, the present invention adopts the following technical solutions:

[0008] A highly reliable and secure transmission method based on OTFS, for a sending end, the transmission method includes:

[0009] Step 1: The source bit sequence m[i] sent by each data frame is constructed as a 1×k matrix m, where k represents the number of bits contained in each data frame; then the matrix m is combined with the k×n standard generator matrix G LDPC Perform linear matrix multiplication to map the source bit sequence to the sparse code domain of LDPC and obtain the new codeword c = mG LDPC ;

[0010] Step 2, performing QAM constellation mapping on the new codeword c to obtain a mapped discrete constellation symbol d[i], and constructing d[i] into a matrix d;

[0011] Step 3: Perform singular value decomposition on the DD domain equivalent channel matrix of the legal channel to obtain Among them U Bob represents the left singular value matrix, E Bob represents a diagonal matrix, V Bob represents the right singular value matrix, (·) H represents the conjugate transposed matrix; then use the matrix V Bob Linear precoding is performed on the matrix d to obtain the coding matrix x DD ;

[0012] Step 4: transform the encoding matrix x DD By first mapping to the DD domain grid through serial-to-parallel conversion, we get the matrix X DD ;

[0013] Step 5: For the matrix X DD Perform inverse symplectic finite Fourier transform to obtain the time-frequency domain signal matrix X TF ;

[0014] Step 6: TF Perform the Heisenberg transform and first obtain the matrix S in the time-delay domain TD , and then column vectorization is performed to obtain the time domain vector s T;

[0015] Step 7: Use the time domain plus CP matrix to process the time domain vector to obtain the time domain signal vector

[0016] Step 8: transform the time domain signal vector The transmission signal is obtained after digital-to-analog conversion and up-conversion, and finally sent to the wireless legal channel through the antenna for air interface transmission.

[0017] Furthermore, for the receiving end of the legitimate user:

[0018] Step 1′: After the legitimate user obtains the received signal from the legitimate channel, the baseband digital signal vector is obtained through analog-to-digital conversion and down-conversion.

[0019] Step 2′, for the vector Perform the CP removal operation to obtain the time domain signal vector r T,Bob ;

[0020] Step 3′, the time domain signal vector r T,Bob Perform inverse column vectorization to reconstruct the matrix R TD,Bob ; For the TD domain matrix R TD,Bob Perform Wigner transformation to obtain the TF matrix Y TF,Bob ;

[0021] Step 4′, for the matrix Y TF,Bob Perform symplectic finite Fourier transform to obtain the DD domain signal matrix Y DD,Bob , then for Y DD,Bob Column vectorization is performed to obtain y DD,Bob ;

[0022] Step 5′, perform singular value decomposition on the DD domain matrix of the legal channel to obtain the left singular value matrix U Bob Perform SVD decoding and get

[0023] Step 6′, the diagonal matrix E is obtained by performing singular value decomposition on the DD domain matrix of the legal channel Bob For the y DD,Bob Perform single-tap equalization and obtain the matrix d Bob ;

[0024] Step 7', the signal d Bob Perform QAM constellation demapping to obtain codeword c Bob ;

[0025] Step 8′, for the code word c Bob Perform LDPC decoding to obtain m Bob , mBob Converted into the destination bit sequence m Bob [i]; when the decoding is completely correct, this sequence is the same as the source bit sequence m[i].

[0026] Furthermore, G LDPC =[I k ,P k×(n-k) ], I k represents the k×k identity matrix, P k×(n-k) represents a k×(nk) sparse matrix, and n represents the encoded code length.

[0027] Furthermore, the pair matrix X DD Perform inverse symplectic finite Fourier transform to obtain the time-frequency domain signal matrix X TF , expressed as:

[0028] The inverse symplectic finite Fourier transform is obtained by combining M-point DFT by column and N-point IDFT by row, that is, in represents the normalized discrete Fourier transform matrix of M points, e represents a natural constant, j represents an imaginary unit, k and l represent the matrix F M The row and column index values ​​in the matrix F N Represents the N-point normalized discrete Fourier transform matrix.

[0029] Further, the transmit filter matrix is ​​expressed as:

[0030] G tx =diag[g tx (0),g tx (T / M),...,g tx ((M-1)T / M)]

[0031] in,(·) H represents the conjugate transposed matrix, g tx (t) represents a transmission filter function, T=1 / Δf represents a time grid interval, Δf represents a subcarrier interval, and M represents the number of points of a normalized discrete Fourier transform.

[0032] Furthermore, the time domain plus CP matrix is ​​expressed as:

[0033]

[0034] Among them, the matrix is the MN×MN identity matrix I MN The last l CP The matrix constructed by the submatrix of the row, where the length of CP is l CP It is generally designed to be larger than the maximum channel delay tap l max .

[0035] Furthermore, the pair of vectors Perform the CP removal operation to obtain the time domain signal vector r T,Bob , expressed as:

[0036] r T,Bob =H T,Bob s T +w T,Bob

[0037] in represents the time domain equivalent channel matrix before and after CP addition and removal, represents the permutation matrix for the forward cyclic shift, represents a diagonal matrix, P represents the number of multipath channels between the transmitter and the receiver of the legitimate user, and h i , and denote the complex gain coefficient, integer delay tap and integer Doppler tap of the ith path respectively, j is an imaginary unit, e is a natural constant, It represents a time domain noise vector that obeys a Gaussian distribution, with a mean of 0 and a variance of the noise power

[0038] Furthermore, the receiver filter matrix is ​​expressed as:

[0039] G rx =diag[g rx (0),g rx (T / M),...,g rx ((M-1)T / M)]

[0040] Among them, g rx (t) represents the receiving filter function, F M Represents the M-point normalized discrete Fourier transform matrix.

[0041] Furthermore, any legal codeword c B ob all satisfy in(·) T represents the transposed matrix, H LDPC Represents the generator matrix G of the transmitter LDPC The corresponding check matrix.

[0042] Compared with the prior art, the present invention has the following technical features:

[0043] 1. Compared with traditional OFDM, the OTFS communication system adopted in the present invention has excellent performance in high mobility scenarios and can effectively resist the severe Doppler effect in high mobility channels.

[0044] 2. The present invention utilizes the reciprocity characteristics of the wireless channel of the duplex system to perform singular value decomposition on the DD domain equivalent channel matrix unique to OTFS. Based on the difference in the channel matrices between the legitimate user and the eavesdropper, a secure precoder is designed, which makes the transmission waveform have the characteristic of low interception, thereby ensuring the security of wireless physical layer transmission.

[0045] 3. Facing the high-order cumulative amount detection of the eavesdropper, the phase and amplitude distribution of the air interface transmission waveform of the present invention is close to Gaussian white noise, and has a good covert transmission characteristic.

[0046] 4. The present invention combines LDPC coding with SVD precoding to form serial cascade coding, and simultaneously utilizes the grouping characteristics of OTFS and LDPC coding. At the transmitting end, the signal is subjected to linear operations of matrix multiplication with the LDPC generation matrix and the SVD precoding matrix in succession, and at the receiving end, a single-tap linear equalizer is designed based on the diagonal matrix decomposed by SVD, so that the cascade encoder and the single-tap equalizer of the system of the present invention both have the characteristics of low complexity.

[0047] 5. The present invention makes full use of the low-density characteristics of the LDPC code domain to reduce the peak-to-average ratio of the OTFS system. The powerful error detection and correction capabilities of LDPC iterative decoding greatly improve the transmission reliability of the system of the present invention.

[0048] 6. The present invention adopts the R-CP model unique to the OTFS system, which further improves the spectrum efficiency and saves a lot of time-frequency resource overhead compared to the traditional F-CP model used in OFDM. While improving the total system throughput, it can also maintain similar bit error rate performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 It is a block diagram of a legal communication link;

[0050] Figure 2 The distribution diagram of the phase-amplitude statistical characteristics of the transmission waveform, where (a) is the amplitude distribution and (b) is the phase distribution;

[0051] Figure 3 The constellation diagram received by the eavesdropper, where (a) is the primary eavesdropping user and (b) is the secondary eavesdropping user;

[0052] Figure 4 This is the system bit error rate curve simulation diagram. DETAILED DESCRIPTION

[0053] The present invention proposes a highly reliable and secure transmission method based on OTFS, which combines LDPC coding with OTFS precoding technology to form a serial internal and external cascade coding between the code domain and the symbol domain, and fully combines the powerful error correction performance of LDPC code and the reliable security performance of precoding. Based on this, a set of OTFS air interface transmission solutions with both high reliability and high security are designed; the specific technical solutions of the present invention are:

[0054] The sending end includes the following steps:

[0055] Step 1, the source bit sequence sent by each data frame is m[i] (i=0,1,...,k-1), where k represents the number of bits contained in each data frame, and i represents the i-th data bit; the source bit sequence m[i] is constructed as a 1×k matrix m, and then the matrix m is combined with the k×n standard generator matrix G LDPC Perform linear matrix multiplication to map the source bit sequence to the sparse code domain of LDPC and obtain the new codeword c = mG LDPC , where G LDPC =[I k ,P k×(n-k) ], I k represents the k×k identity matrix, P k×(n-k) represents a k×(nk) sparse matrix, n represents the encoded code length, and the dimension parameter setting principle of the generator matrix and its submatrices is to ensure that the source matrix m and the generator matrix G LDPC The dimensions of conform to the linear matrix multiplication rules.

[0056] Step 2, perform QAM constellation mapping on the new codeword c to obtain a mapped discrete constellation symbol d[i] (i=0, 1, ..., R-1), where R represents the number of constellation symbols contained in each data frame, and i represents the i-th constellation symbol; construct the d[i] as a matrix d=[d 0 ,d 1 ,…d R-1 ] T , where element d i =d[i](i=0,1,...,R-1), the superscript T indicates the matrix transpose.

[0057] Step 3: According to the reciprocity of the wireless channel, both the legitimate transmitter and the receiver of the legitimate user of the time division duplex system can obtain channel state information. The DD domain equivalent channel matrix of the legitimate channel (i.e., the wireless channel between the transmitter and the receiver of the legitimate user) is subjected to singular value decomposition (SVD) to obtain Among them U Bob represents the left singular value matrix, E Bob represents a diagonal matrix, VBob represents the right singular value matrix, (·) H represents the conjugate transposed matrix; in general, the equivalent channel matrix The acquisition can be estimated by receiving a known pilot sequence, and then using the matrix V Bob Linear precoding is performed on the matrix d in step 2 to obtain the coding matrix x DD =V Bob d.

[0058] Step 4: For the subsequent two-dimensional transformation of OTFS modulation in the DD domain, the coding matrix By serial-to-parallel conversion, it is first mapped to the M×N DD domain grid to obtain the matrix represents an M×N dimensional complex space, the matrix X DD Each row of represents the constellation symbol information carried by a subcarrier, so there are M subcarriers, and the entire matrix X DD The constellation symbol information carried is used as an OTFS symbol, so the OTFS symbol length is N.

[0059] Step 5: For the matrix X DD Perform an inverse symplectic finite Fourier transform (ISFFT) to obtain the time-frequency domain signal matrix X TF , which can be obtained by combining M-point DFT by column and N-point IDFT by row, that is in represents the normalized discrete Fourier transform matrix of M points, e represents a natural constant, j represents an imaginary unit, k and l represent the matrix F M The row and column index values ​​in the matrix F N The same applies;

[0060] Step 6: TF Perform the Heisenberg transform and first obtain the matrix in the time-delay (TD) domain: Among them G tx =diag[g tx (0),g tx (T / M),...,g tx ((M-1)T / M)] represents the transmit filter matrix, (·) H represents the conjugate transposed matrix, g tx (t) represents the transmission filter function, which can be flexibly designed by referring to the existing pulse shaping function in the common communication system, T = 1 / Δf represents the time grid interval, Δf represents the subcarrier interval; then column vectorization is performed to obtain the time domain vector sT =vec(S TD ), where vec(·) represents a column quantization operation. Steps 5 and 6 at the transmitting end combine the standard OTFS modulation process, transform the signal from the delay-Doppler domain to the time-frequency domain and the time domain successively, and retain the intermediate process of transforming to the time-frequency domain, which is convenient for combining with other technologies applicable to traditional OFDM, and to a certain extent ensures the compatibility of the scheme of the present invention with other technologies, which is the same as the corresponding steps 3 and 4 at the receiving end below.

[0061] Step 7: To resist ICI and ISI caused by time expansion, the R-CP (Reduced Cyclic Prefix) model with less overhead than the F-CP (Full Cyclic Prefix) model in the traditional OFDM system is used to improve the spectrum efficiency and system throughput. Specifically, the time domain vector is processed using the time domain plus CP matrix to obtain the time domain signal vector: in Represents the time domain plus CP matrix, matrix is the MN×MN identity matrix I MN The last l CP The matrix constructed by the submatrix of the row, where the length of CP is l CP It is generally designed to be slightly larger than the maximum channel delay tap l max .

[0062] Step 8: transform the time domain signal vector After the digital-to-analog conversion and up-conversion, the transmission signal is obtained, and finally sent to the wireless channel through the antenna for air interface transmission. The present invention is mainly described by taking the multipath fading channel as an example.

[0063] For the receiving end of a legitimate user, the following steps are included:

[0064] Step 1′: After the legitimate user obtains the received signal from the legitimate channel, the baseband digital signal vector is obtained through analog-to-digital conversion and down-conversion.

[0065] Step 2′, for the vector Perform the CP removal operation to obtain the time domain signal vector r T,Bob =H T,Bob s T +w T,Bob ,in represents the time domain equivalent channel matrix before and after CP addition and removal, represents the permutation matrix for the forward cyclic shift, represents a diagonal matrix, P represents the number of multipath channels between the transmitter and the receiver of the legitimate user, and h i , and denote the complex gain coefficient, integer delay tap and integer Doppler tap of the ith path respectively, It represents a time domain noise vector that obeys a Gaussian distribution, with a mean of 0 and a variance of the noise power

[0066] Step 3′, the time domain signal vector r T,Bob Perform inverse column vectorization to reconstruct into an M×N matrix R TD,Bob =vec -1 (r T,Bob );For the TD domain matrix R TD,Bob Perform Wigner transformation to obtain the TF matrix Y TF,Bob =F M G rx R TD,Bob , where the receiver filter matrix can be expressed as G rx =diag[g rx (0),g rx (T / M),...,g rx ((M-1)T / M)],g rx (t) represents the receiving filter function, F M Represents the M-point normalized discrete Fourier transform matrix.

[0067] The two-step cascade transformation method used in this scheme takes into account:

[0068] From the perspective of the standard modulation process, without loss of generality, the two-step transformation is a standard modulation transmission process of the OTFS technology; from the perspective of transform domain signal processing, at the transmitting end, the delay-Doppler domain signal is transformed into the time-frequency domain through ISFFT, and then the signal is transformed into the time domain through Heisenberg transform, and some traditional OFDM processing techniques can be used in the time-frequency domain, thereby ensuring compatibility with other technologies to a certain extent; the same is true for the receiving end.

[0069] Step 4′, for the matrix Y TF,Bob Perform a symplectic finite Fourier transform (SFFT) to obtain the DD domain signal matrix Then perform column vectorization to get y DD,Bob =vec(Y DD,Bob ); At this time, the DD domain signal reception vector can be expressed as where x DD =vec(X DD ), X DD is the matrix described in step 4 at the sending end, and represent the equivalent channel matrix and noise vector of the legal channel in the DD domain respectively.

[0070] Step 5′, the matrix U obtained in step 3 at the sending end Bob Perform SVD decoding and get

[0071] Step 6′, the matrix E obtained in step 3 at the sending end Bob For the y DD,Bob Perform single-tap equalization and get After further simplification, the signal can be obtained The second term is a noise term that follows a Gaussian distribution.

[0072] Step 7', the signal d Bob Perform QAM constellation demapping to obtain codeword c Bob .

[0073] Step 8′, for the code word c Bob Perform LDPC decoding to obtain m Bob , so that any legal codeword c Bob All meet in(·) T represents the transposed matrix, Represents the matrix G generated by the sender in step 1 LDPC The corresponding check matrix, where the parameters n, k and matrix P are the same as step 1 at the sending end; the decoded m Bob Converted into the destination bit sequence m Bob [i] (i = 0, 1, ..., k-1), when the decoding is completely correct, this sequence is the same as the source bit sequence m[i] (i = 0, 1, ..., k-1).

[0074] The LDPC decoding algorithm can use soft decision decoding and hard decision decoding. In order to further reduce the decoding complexity, the present invention mainly uses the hard decision decoding method as an example. In fact, if soft decision decoding is used, a higher coding gain will be obtained at the cost of increased decoding complexity, thereby further improving the reliability of system transmission.

[0075] The safety performance analysis of the scheme of the present invention is carried out as follows:

[0076] The design of the eavesdropping model in the scheme of the present invention takes into account two levels of eavesdroppers, namely, the first-level eavesdropper and the second-level eavesdropper. Among them, for the first-level eavesdropper, considering the situation that is very unfavorable to the legitimate link, it is assumed that it has prior knowledge of the OTFS transmission method and system parameters of the legitimate link, but is unaware of the cascade coding method proposed in the present invention; for the second-level eavesdropper, we consider the situation that is extremely unfavorable to the legitimate link, assuming that it is fully aware of the secure transmission method based on cascade coding proposed in the present invention.

[0077] The receiving process of the first-level eavesdropper is as follows:

[0078] Steps 1-4, the first-level eavesdropper can perform the same processing as the legitimate user's receiving end from step 1' to step 4', and obtain the baseband digital signal vector The time domain signal vector r after removing CP T,E ve 1 =H T,E ve 1 s T +w T, E ve 1 , TD domain signal matrix R TD,Eve1 =vec -1 (r T,Eve1 ), TF domain signal matrix Y TF,Eve1 =F M G rx R TD,Eve1 , DD domain signal matrix DD domain signal vector The DD equivalent channel matrix and DD equivalent noise vector of the first-level eavesdropper are and w T,Eve1 represents the time domain noise vector of the first-level eavesdropper, H T,Eve1 Represents the time domain equivalent channel matrix of the first-level eavesdropper channel;

[0079] Step 5: DD domain vector y of the rectangular wave received signal TF,Eve1 For DD domain equalization, Zero Forcing (ZF) equalization and Minimum Mean Square Error (MMSE) equalization can generally be used. The ZF equalization formula is: The ZF equilibrium matrix is It represents the signal vector obtained by the first-level eavesdropper after ZF equalization. The MMSE equalization formula is The MMSE equalization matrix is represents the signal vector obtained by the first-level eavesdropper after MMSE equalization, and ρ represents the average signal-to-noise ratio. The present invention mainly uses a more optimal MMSE equalization method as an example;

[0080] Step 6: equalize the signal Perform QAM constellation demapping to obtain the received codeword c Eve1 ;

[0081] Step 7: Receive codeword c Eve1Perform LDPC decoding in the same way as in step 8′ at the receiving end of the legitimate user to obtain m Eve1 ; The decoded m Eve1 Converted into the destination bit sequence m Eve1 [i](i=0,1,...,k-1). So far, the first-level eavesdropper has completed all the processing procedures at its receiving end. Since it is unable to know the concatenated coding transmission method proposed in the present invention, even if a better MMSE equalization is used, it is still unable to recover the source bit sequence m[i](i=0,1,...,k-1), that is, the bit sequence m[i] obtained by the first-level eavesdropper decoding. Eve1 [i] (i = 0, 1, ..., k-1) is different from the source bit sequence m[i] (i = 0, 1, ..., k-1), so the eavesdropping fails.

[0082] The receiving process of the secondary eavesdropper is as follows:

[0083] Steps 1-8, the secondary eavesdropper can perform the same processing flow as the legitimate user's receiving end, thereby obtaining the receiving signals of each level under the eavesdropping channel conditions, which are the baseband digital signal vectors in order. The time domain signal vector r after removing CP T,Eve2 =H T,Eve2 s T +w T,Eve2 , TD domain signal matrix R TD,Eve2 =vec -1 (r T,Eve2 ), TF domain signal matrix Y TF,Eve2 =F M G rx R TD,Eve2 , DD domain signal matrix DD domain signal vector y DD,Eve2 =vec(Y DD,Eve2 ), signal vector after SVD decoding Signal vector after single tap equalization For the d Eve2 Further processing yields The DD equivalent channel matrix and DD equivalent noise vector of the secondary eavesdropper are and w T,Eve2 represents the time domain noise vector of the secondary eavesdropper, H T,Eve2 The time domain equivalent channel matrix of the secondary eavesdropper channel is represented by the SVD decomposition of the DD domain equivalent channel matrix of the secondary eavesdropper's eavesdropping channel. Due to the difference between the equivalent channel matrix of the eavesdropping channel and the legitimate channel, the secondary eavesdropper can only obtain the downgraded version of d Eve2 and d Eve2, and it is impossible to further eliminate the transmitter according to the legal channel matrix The designed precoding matrix V Bob Finally, the secondary eavesdropper will perform LDPC decoding in the same way as the legitimate user's receiving end step 8' to obtain the destination bit sequence m Eve2 [i](i=0,1,...,k-1). So far, the secondary eavesdropper has completed all the same processing procedures as the legitimate user's receiving end, but still cannot successfully recover the source bit m[i](i=0,1,...,k-1), that is, the bit sequence m[i] obtained by the secondary eavesdropper's decoding Eve2 [i] (i = 0, 1, ..., k-1) is different from the source bit sequence m[i] (i = 0, 1, ..., k-1), so the eavesdropping fails.

[0084] The solution of the present invention supports OTFS transmission under arbitrary waveforms by selecting different transmission filter matrices G tx and the receive filter matrix G rx The transmission waveform can be changed appropriately. In order to fit the actual communication system implementation and simulation analysis of engineers, the simulation uses rectangular waves to simplify the Heisenberg transform and Wigner transform. Under the rectangular wave condition, there is a relationship G tx =G rx =I M , where I M represents the M×M identity matrix;

[0085] Depend on Figure 2 and Figure 3 It can be seen that the amplitude distribution of the transmission waveform of the present invention is similar to Rayleigh Distribution, and its phase distribution is similar to Homogenous Distribution, which makes the phase-amplitude distribution characteristics of its constellation diagram extremely similar to Gaussian white noise, making it impossible for an eavesdropper to distinguish the waveform of a useful signal from Gaussian white noise.

[0086] Figure 4 The transmission reliability from the legitimate sender to the legitimate user, the first-level eavesdropper and the second-level eavesdropper was simulated. The simulation conditions adopted the EVA channel model in the 3GPP standard. The multipath delay and relative power were [0 30 150 310370 710 1090 1730 2510]ns and [0 -1.5 -1.4 -3.6 -0.6 -9.1 -7-12-16.9]dB, respectively. The maximum speed of the simulated channel was 500km / h. The simulation results show that the scheme of the present invention has good transmission reliability and security. Neither the first-level eavesdropper nor the second-level eavesdropper can decipher the information transmitted in the legitimate link.

[0087] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A highly reliable and secure transmission method based on OTFS, characterized in that: For the transmitting end, the transmission method includes: Step 1: The source bit sequence m[i] sent by each data frame is constructed as a 1×k matrix m, where k represents the number of bits contained in each data frame; then the matrix m is combined with the k×n standard generator matrix G LDPC Perform linear matrix multiplication to map the source bit sequence to the sparse code domain of LDPC and obtain the new codeword c = mG LDPC ; Step 2, performing QAM constellation mapping on the new codeword c to obtain a mapped discrete constellation symbol d[i], and constructing d[i] into a matrix d; Step 3: Perform singular value decomposition on the DD domain equivalent channel matrix of the legal channel to obtain Among them U Bob represents the left singular value matrix, E Bob represents a diagonal matrix, V Bob represents the right singular value matrix, (·) H represents the conjugate transposed matrix; then use the matrix V Bob Linear precoding is performed on the matrix d to obtain the coding matrix x DD ; Step 4: transform the encoding matrix x DD By first mapping to the DD domain grid through serial-to-parallel conversion, we get the matrix X DD ; Step 5: For the matrix X DD Perform inverse symplectic finite Fourier transform to obtain the time-frequency domain signal matrix X TF ; Step 6: TF Perform the Heisenberg transform and first obtain the matrix S in the time-delay domain TD , and then column vectorization is performed to obtain the time domain vector s T ; Step 7: Use the time domain plus CP matrix to process the time domain vector to obtain the time domain signal vector Step 8: transform the time domain signal vector The transmission signal is obtained after digital-to-analog conversion and up-conversion, and finally sent to the wireless legal channel through the antenna for air interface transmission.

2. A highly reliable and secure transmission method based on OTFS, characterized in that: For the receiving end, the transmission method includes: Step 1′, after the receiving end of the legal user obtains the received signal from the legal channel, it obtains the baseband digital signal vector through analog-to-digital conversion and down-conversion Step 2′, for the vector Perform the CP removal operation to obtain the time domain signal vector r T,Bob ; Step 3′, the time domain signal vector r T,Bob Perform inverse column vectorization to reconstruct the matrix R TD,Bob ; For the TD domain matrix R TD,Bob Perform Wigner transformation to obtain the TF matrix Y TF,Bob ; Step 4′, for the matrix Y TF,Bob Perform symplectic finite Fourier transform to obtain the DD domain signal matrix Y DD,Bob , then for Y DD,Bob Column vectorization is performed to obtain y DD,Bob ; Step 5′, perform singular value decomposition on the DD domain matrix of the legal channel to obtain the left singular value matrix U Bob Perform SVD decoding and get Step 6′, the diagonal matrix E is obtained by performing singular value decomposition on the DD domain matrix of the legal channel Bob For the y DD,Bob Perform single-tap equalization and obtain the matrix d Bob ; Step 7', the signal d Bob Perform QAM constellation demapping to obtain codeword c Bob ; Step 8′, for the code word c Bob Perform LDPC decoding to obtain m Bob , m Bob Converted into the destination bit sequence m Bob [i]; when the decoding is completely correct, this sequence is the same as the source bit sequence m[i].

3. The highly reliable and secure transmission method based on OTFS according to claim 1, characterized in that: G LDPC =[I k ,P k×(n-k) ], I k represents the k×k identity matrix, P k×(n-k) represents a k×(nk) sparse matrix, and n represents the encoded code length.

4. The highly reliable and secure transmission method based on OTFS according to claim 1, characterized in that: The pair matrix X DD Perform inverse symplectic finite Fourier transform to obtain the time-frequency domain signal matrix X TF , expressed as: The inverse symplectic finite Fourier transform is obtained by combining M-point DFT by column and N-point IDFT by row, that is, in represents the normalized discrete Fourier transform matrix of M points, e represents a natural constant, j represents an imaginary unit, k and l represent the matrix F M The row and column index values ​​in the matrix F N Represents the N-point normalized discrete Fourier transform matrix.

5. The highly reliable and secure transmission method based on OTFS according to claim 1, characterized in that: The transmit filter matrix is ​​expressed as: G tx =diag[g tx (0),g tx (T / M),...,g tx ((M-1)T / M)] in,(·) H represents the conjugate transposed matrix, g tx (t) represents a transmission filter function, T=1 / Δf represents a time grid interval, Δf represents a subcarrier interval, and M represents the number of points of a normalized discrete Fourier transform.

6. The highly reliable and secure transmission method based on OTFS according to claim 1, characterized in that: The time domain plus CP matrix is ​​expressed as: Among them, the matrix is the MN×MN identity matrix I MN The last l CP The matrix constructed by the submatrix of the row, where the length of CP is l CP It is generally designed to be larger than the maximum channel delay tap l max .

7. The highly reliable and secure transmission method based on OTFS according to claim 1, characterized in that: The pair of vectors Perform the CP removal operation to obtain the time domain signal vector r T,Bob , expressed as: r T,Bob =H T,Bob s T +w T,Bob in represents the time domain equivalent channel matrix before and after CP addition and removal, represents the permutation matrix for the forward cyclic shift, represents a diagonal matrix, P represents the number of multipath channels between the transmitter and the receiver of the legitimate user, and h i , and denote the complex gain coefficient, integer delay tap and integer Doppler tap of the ith path respectively, j is an imaginary unit, e is a natural constant, It represents a time domain noise vector that obeys a Gaussian distribution, with a mean of 0 and a variance of the noise power 8. The highly reliable and secure transmission method based on OTFS according to claim 2, characterized in that: The receiver filter matrix is ​​expressed as: G rx =diag[g rx (0),g rx (T / M),…,g rx ((M-1)T / M)] Among them, g rx (t) represents the receiving filter function, F M Represents the M-point normalized discrete Fourier transform matrix.

9. The highly reliable and secure transmission method based on OTFS according to claim 1, characterized in that: Any legal codeword c B ob all satisfy in(·) T represents the transposed matrix, H LDPC Represents the generator matrix G of the transmitter LDPC The corresponding check matrix.

Images