Signal transmission method based on random unitary matrix coded modulation and cross-domain iterative detection

Abstract

The present invention discloses a signal transmission method based on random unitary matrix coding modulation and cross-domain iterative detection, including random unitary matrix coding modulation at a signal transmitting end and cross-domain iterative detection at a receiving end; random unitary matrix coding modulation includes random unitary modulation or random unitary precoding, and random unitary transformation is used to make the equivalent channel matrix satisfy the correct right unitary invariant assumption; the cross-domain iterative detection is to use a cross-domain iterative receiver to receive the signal modulated by the random unitary matrix coding. Random unitary matrix coding modulation can ensure that general OAMP / VAMP, UAMP and MAMP receivers achieve Bayesian optimal performance; the cross-domain iterative detection enables the receiver to utilize ultra-sparse time domain channels through cross-domain detection, including low-complexity MAMP and OAMP / VAMP, UAMP receivers.

Description

Technical Field

[0001] The present invention relates to the field of signal transmission, and in particular to a signal transmission method based on random unitary matrix coding modulation and cross-domain iterative detection. Background Art

[0002] With the rapid development of wireless communication technology, the communication demand for high-mobility scenarios such as high-speed railway systems and low-orbit satellites is growing. However, the Doppler shift caused by high-speed mobility will cause serious inter-carrier interference, destroying the orthogonality of subcarriers in Orthogonal Frequency Division Multiplexing (OFDM), making it difficult for OFDM to ensure reliable data transmission.

[0003] To mitigate the impact of Doppler shift, Orthogonal Time Frequency Space (OTFS) was proposed to multiplex information symbols in the Delay-Doppler (DD) domain, achieving diversity in channel delay and Doppler shift, thereby improving the reliability of signal transmission. Recently, Affine Frequency Division Multiplexing (AFDM) was proposed, which uses Inverse Discrete Affine Fourier Transform (IDAFT) to modulate information symbols into the time-frequency domain. In this way, AFDM has been shown to achieve full diversity in high-mobility communication systems.

[0004] The potential significant diversity gain of OTFS and AFDM can only be achieved through high-complexity Maximum Likelihood (ML) receivers, which severely limits their applications. There are many studies on low-complexity receivers for OTFS and AFDM. For OTFS, a Cross-Domain Orthogonal Approximate Message Passing (CD-OAMP) receiver is proposed in the prior art, in which the estimated signal is iteratively transmitted between the time domain and the DD domain. At the same time, a DD-OAMP receiver is also proposed for signal detection in the DD domain. However, since CD / DD-OAMP uses Linear Minimum Mean Square Error (LMMSE) estimation, the sparsity of the channel matrix is ​​not fully utilized due to matrix inversion, resulting in high complexity. To solve this problem, a DD-Doppler Domain Memory Approximate Message Passing (DD-MAMP) receiver is proposed in the prior art, which utilizes the sparsity of the DD-domain channel matrix through a memory matched filter (MF). However, DD-MAMP ignores the more sparse time-domain channel matrix. The development of AFDM receivers is still in its early stages, and there is currently no efficient receiver that can fully utilize the advantages of AFDM's significant diversity gain.

[0005] The main idea of ​​the existing multi-carrier modulation technology (OFDM, OTFS, AFDM, etc.) is to remove the interference between symbols as much as possible and obtain an equivalent sparse channel matrix, so as to recover the signal using a simplified low-complexity receiver. However, each signal only experiences a small amount of channel fading, which leads to the degradation of the detection performance of the existing multi-carrier modulation technology (OFDM, OTFS, AFDM, etc.) under actual receivers. Summary of the invention

[0006] The present invention aims to address the deficiencies in the prior art and to propose a signal transmission method based on random unitary matrix coding modulation and cross-domain iterative detection.

[0007] The object of the present invention is achieved through the following technical solutions: a signal transmission method based on random unitary matrix coding modulation and cross-domain iterative detection, the method comprising random unitary matrix coding modulation at a signal transmitting end and cross-domain iterative detection at a receiving end;

[0008] The random unitary matrix coded modulation is implemented by random unitary modulation or random unitary precoding;

[0009] The random unitary modulation includes: digitally modulating the information bit sequence at the transmitter, performing serial-to-parallel conversion on the modulated information and performing random unitary modulation through a random unitary transformation matrix, adding a cyclic prefix to the signal vector after random unitary modulation, and sending it to the channel after passing through a transmission filter;

[0010] The random unitary precoding includes: performing channel coding and digital modulation on the information bit sequence at the transmitter, performing unitary precoding on the modulated information through a random unitary transformation matrix to obtain a new symbol sequence, combining the new symbol sequence with an existing transmission scheme to obtain a transmission signal for transmission;

[0011] The cross-domain iterative detection is to use a cross-domain iterative receiver to receive a signal modulated by random unitary matrix coding.

[0012] Furthermore, in the random unitary modulation, the random unitary transformation matrix is ​​a unitary invariant matrix.

[0013] Furthermore, the use process of the cross-domain iterative receiver is specifically as follows: based on the received signal y and the a priori estimate x provided by the nonlinear detection t , linear detection output estimation signal r t , and after inverse random unitary transformation, the signal is obtained After further nonlinear detection, the output estimated signal s is obtained. t+1 , after random unitary transformation, x is generated t+1 And input to linear detection; iterate to estimate signal s t+1 After accurate recovery or reaching the preset maximum number of iterations, the entire detection process stops and the final estimated signal is output as the received signal.

[0014] Furthermore, the cross-domain iterative receiver uses a Bayesian optimal cross-domain iterative receiver, and the Bayesian optimal cross-domain iterative receiver specifically comprises: multiplying the sparse time domain channel matrix with the input signal by using the time domain memory MF in the linear detection of the cross-domain iterative receiver, and using orthogonal MMSE demodulation in the nonlinear detection.

[0015] Further, the transmitting end processing of the random unitary modulation in the MIMO scenario includes: performing signal segmentation after the random unitary modulation or performing signal segmentation before the random unitary modulation; the performing signal segmentation before the random unitary modulation also includes using a random permutation matrix segmentation before the random unitary modulation;

[0016] The corresponding receiving end processing is: receiving the signal through the receiving filter at the antenna of the receiving end respectively and combining the signals after deleting the cyclic prefix; after combining, using the cross-domain iterative receiver to receive the signal modulated by the random unitary matrix coding.

[0017] Further, the signal segmentation after random unitary modulation is specifically:

[0018] After the digital modulated signal is converted from serial to parallel and modulated by random unitary modulation, it is divided into several segments, and a cyclic prefix is ​​added to each segment. After the signal is obtained through the transmit filter, it is sent to the channel through the antenna.

[0019] The corresponding cross-domain iterative receiver processing is: based on the received signal y and the prior estimate x provided by the nonlinear detection t , linear detection output estimation signal r t , and after inverse random unitary transformation, the signal is obtained After further nonlinear detection, the output estimated signal s is obtained. t+1 , after random unitary transformation, x is generated t+1 And input to linear detection; iterate to estimate signal s t+1 After the accurate recovery can be achieved or the preset maximum number of iterations is reached, the entire detection process stops and the final recovered estimated signal is output as the received signal.

[0020] Furthermore, the signal segmentation before random unitary modulation is specifically performed as follows:

[0021] After the digital modulated signal is converted from serial to parallel, it is divided into several segments, and each segment is randomly unitary modulated and a cyclic prefix is ​​added. After the signal is obtained through the transmit filter, it is sent to the channel through the antenna.

[0022] The corresponding receiving end processing is: receiving the signal through the receiving filter at the antenna of the receiving end respectively and deleting the cyclic prefix and then combining the signals; after combining, using the cross-domain iterative receiver to receive the signal modulated by the random unitary matrix coding, wherein the implementation steps of the cross-domain iterative receiver include: converting the estimated signal r output by the linear detection t Divide into N t Segments, each segment is subjected to inverse random unitary modulation and then combined to obtain an estimated signal Input to the nonlinear detection; the estimated signal s for the nonlinear detection output t+1 Divide into N t Segments, each segment is randomly unitarily modulated and then combined to obtain the estimated signal x t+1 , input into the linear detection, iterate to estimate the signal s t+1 After the accurate recovery can be achieved or the preset maximum number of iterations is reached, the entire detection process stops and the final recovered estimated signal is output as the received signal.

[0023] Further, the use of random permutation matrix segmentation before random unitary modulation is specifically:

[0024] The digital modulated signal is divided into several segments after serial-to-parallel conversion and random permutation matrix. Each segment is randomly unitary modulated and a cyclic prefix is ​​added. The time domain signal is obtained through the transmit filter and sent to the channel through the antenna.

[0025] The corresponding receiving end processing is: receiving the signal through the receiving filter at the antenna of the receiving end respectively and deleting the cyclic prefix and then combining the signals; after combining, using the cross-domain iterative receiver to receive the signal modulated by the random unitary matrix coding, wherein the implementation steps of the cross-domain iterative receiver include: converting the estimated signal r output by the linear detection t Divide into N t Segments, each segment is subjected to inverse random unitary modulation and then merged into the random permutation inverse matrix Π H Get the estimated signal Input to the nonlinear detection; the estimated signal s for the nonlinear detection output t+1 Divide into N t segment, which is input into the random permutation matrix Π and then divided into N t Segments, each segment is randomly unitarily modulated, and then combined to obtain the estimated signal x t+1 , input into the linear detection; iterate to estimate the signal s t+1 After the accurate recovery can be achieved or the preset maximum number of iterations is reached, the entire detection process stops and the final recovered estimated signal is output as the received signal.

[0026] Furthermore, the random unitary matrix coded modulation uses random unitary precoding specifically as follows:

[0027] The information bit sequence is channel coded to generate sequence c, which is mapped into symbol sequence s through digital modulation. S is precoded with random unitary matrix U to obtain new symbol sequence b, i.e. b=Us, and the transmission sequence x is generated through the existing transmission scheme P, expressed as x=Pb=PUs. The sequence x is transmitted to the channel, and the signal y received by the receiving end is expressed as:

[0028] y=Hx+w=HPUs+w=H eff s+w,

[0029] Among them, H eff =HPU, H is the time domain channel matrix, H eff is the transform domain equivalent channel, w~CN(0,σ 2 I) is additive white Gaussian noise.

[0030] Beneficial effects of the present invention:

[0031] The present invention proposes a random unitary matrix coding modulation technology, including random unitary matrix modulation and random unitary precoding, to ensure high-reliability signal transmission in static multipath and dynamic time-varying communication systems. Different from existing multi-carrier modulation technologies (OFDM, OTFS, AFDM, etc.), the random unitary modulation realizes non-orthogonal transmission, and utilizes random unitary transformation to randomly superimpose all signals together, so that each signal can experience statistically stable channel fading, thereby ensuring that the maximum diversity gain can be obtained. At the same time, the random unitary transformation makes the equivalent channel matrix meet the standard right unitary invariant assumption, ensuring that general OAMP / VAMP, UAMP and MAMP receivers achieve Bayesian optimal performance.

[0032] The random unitary precoding can be directly applied to communication systems characterized by standard linear models, such as OFDM, OTFS, AFDM, RIS and other systems, so that the equivalent channel matrix can satisfy the standard right unitary invariant assumption, and then combined with general OAMP / VAMP, UAMP and MAMP type receivers to achieve Bayesian optimal performance.

[0033] However, the equivalent channel of random unitary modulation is completely dense, which brings great challenges to signal detection. To address this problem, the present invention is based on the characteristics of random unitary transformation, and through linear detection in the time domain, the receiver can utilize the sparsity of the time domain channel to achieve low-complexity signal detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 It is a flow chart of a signal transmission method based on random unitary matrix coding modulation and cross-domain iterative detection;

[0035] Figure 2 A comparison between a time domain channel and a transform domain equivalent channel provided by an embodiment of the present invention;

[0036] Figure 3 A flow chart of a signal transmission method of random unitary modulation scheme 1 in a MIMO scenario provided by an embodiment of the present invention;

[0037] Figure 4 A flow chart of a signal transmission method of random unitary modulation scheme 2 in a MIMO scenario provided by an embodiment of the present invention;

[0038] Figure 5 A flow chart of a signal transmission method of random unitary modulation scheme 3 in a MIMO scenario provided by an embodiment of the present invention;

[0039] Figure 6 A flow chart of a number transmission method based on random unitary precoding provided by an embodiment of the present invention;

[0040] Figure 7A comparison diagram of the estimated signal before and after random unitary transformation provided by an embodiment of the present invention;

[0041] Figure 8 A schematic diagram of a cross-domain MAMP receiver provided by an embodiment of the present invention;

[0042] Fig. 9 A BER comparison diagram of IFDM, OFDM, OTFS and AFDM in a SISO scenario when the mobile station speed is 300 km / h provided in an embodiment of the present invention;

[0043] Fig.10 A BER comparison diagram of IFDM, OFDM, OTFS and AFDM in a 4*4 MIMO scenario when the mobile station speed is 300km / h provided in an embodiment of the present invention;

[0044] Fig.11 A BER comparison diagram of IFDM, OFDM, OTFS and AFDM in a SISO scenario when the mobile station speed is 500 km / h provided in an embodiment of the present invention;

[0045] Fig.12 A comparison chart of the MAMP and OAMP running time of different modulation schemes in the MIMO scenario provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0046] The specific implementation modes of the present invention are further described in detail below with reference to the accompanying drawings.

[0047] The present invention provides a signal transmission method based on random unitary matrix coding modulation and cross-domain iterative detection, the method comprising random unitary matrix coding modulation at a signal transmitting end and cross-domain iterative detection at a receiving end;

[0048] The random unitary matrix coded modulation is implemented by random unitary modulation or random unitary precoding:

[0049] like Figure 1 As shown, the random unitary modulation is specifically: digitally modulating the information bit sequence at the transmitter, performing serial-to-parallel conversion on the modulated information and performing random unitary modulation through a random unitary transformation matrix, adding a cyclic prefix to the signal vector after random unitary modulation, and sending it to the channel after passing through a transmission filter;

[0050] like Figure 6 As shown, the random unitary precoding is specifically: channel coding and digital modulation are performed on the information bit sequence at the transmitter, the modulated information is unitarily precoded through a random unitary transformation matrix to obtain a new symbol sequence, and the transmission signal is obtained by combining the new symbol sequence and the existing transmission scheme for transmission;

[0051] The cross-domain iterative detection is to use a cross-domain iterative receiver to receive a signal modulated by random unitary matrix coding;

[0052] Embodiment 1: This embodiment is an implementation step when the random unitary matrix coded modulation uses random unitary modulation, specifically:

[0053] Sending side:

[0054] like Figure 1 As shown, at the transmitter, the information bit sequence is digitally modulated as The power constraint is The element symbols are selected from the constellation set S (such as QPSK, 16QAM and other digital modulations). After serial-to-parallel (S / P) conversion and random unitary modulation, the signal vector x is:

[0055] x=Us,

[0056] where U is a random unitary transformation matrix, and UU H = 1, 1 represents the unit matrix. In order to overcome the influence of inter-frame interference and multipath propagation, a cyclic prefix (Chirp-Periodic Prefix, CP) is added to x and sent to the channel after passing through the transmit filter. The received signal d[n] of the nth time slot is:

[0057]

[0058] Among them, w~CN(0,σ 2 I) is additive white Gaussian noise, and the time-varying channel response g[n,p] is:

[0059]

[0060] Where p = 0, ..., P-1, P is the channel segment, L is the number of multipaths, T s is the system sampling interval, {h i ,f i ,τ i} represent the complex gain, Doppler frequency shift and delay on the i-th path respectively. Let P is determined by τ max and the duration of the receive filter response, P rc [·] is a sampled raised cosine roll-off filter.

[0061] Receiver:

[0062] After receiving filter and CP removal, the received signal y is expressed as:

[0063] y=Hx+w=HUs+w=H eff s+w,

[0064] in is the time domain channel, is the transform domain equivalent channel. Figure 2 The time domain channel H and the transform domain equivalent channel H are given eff Compare to verify the effect of random unitary transformation. Note that H eff It is more stable than H and is conducive to signal recovery.

[0065] For the detection of random unitary modulation, the present invention proposes a Bayesian optimal cross-domain iterative receiver. Figure 1 As shown, it consists of linear detection in the time domain, inverse random unitary transform, nonlinear detection in the transform domain and random unitary transform. Its general form is expressed as:

[0066] Linear detection: r t =γ t (x t ),

[0067] Inverse random unitary transform:

[0068] Non-linear detection:

[0069] Random unitary transformation: x t+1 =Us t+1 ,

[0070] Where t represents the number of iterations. Specifically, in the t-th iteration process, based on the received signal y and the prior estimate x provided by the nonlinear detection t , linear detection output estimation signal r t , and after inverse random unitary transformation, the signal is obtained Furthermore, the output estimated signal s is obtained through nonlinear detection. t+1 , after random unitary transformation, x is generated t+1 And input to the linear detection. After multiple iterations, until the estimated signal s t+1 The entire detection process stops when accurate recovery can be achieved or the preset maximum number of iterations is reached.

[0071] Considering that the time-domain LMMSE detection in the existing CD-OAMP receiver does not fully utilize the sparsity of the time-domain channel due to matrix inversion, the complexity of CD-OAMP detection is still very high. Figure 1 The framework of the Bayesian optimal cross-domain iterative receiver is shown in Figure 1, and a low-complexity CD-MAMP receiver is proposed, in which the sparse time-domain channel matrix is ​​multiplied by the input signal by using the time-domain memory MF in linear detection, which fully utilizes the sparsity of the time-domain channel and achieves the Bayesian optimal performance with extremely low complexity. Figure 8As shown, the CD-MAMP receiver is composed of orthogonal memory MF, inverse random unitary transform, orthogonal MMSE demodulation and random unitary transform, in which linear detection uses orthogonal memory MF and nonlinear detection uses orthogonal MMSE demodulation. The details are as follows. Consider the tth iteration process,

[0072] 1) Orthogonal Memory MF: Based on the received signal y and the output estimate x of the random unitary transformation t , the output estimate of the orthogonal memory MF is:

[0073]

[0074] Where X t =[x1,...,x t ], and

[0075]

[0076] in It is used to ensure the convergence and orthogonality of MAMP. The specific expression is as follows:

[0077] p t =[p t,1 ,p t,2 ,...,p t,t ]

[0078] p t,i =-θ t,i w t-i

[0079]

[0080]

[0081]

[0082]

[0083] in W t ≡A H B t A, λ min and λ max AA H The smallest and largest eigenvalues ​​of γ. t The output estimation variance of (·) is:

[0084]

[0085] Since the estimated error of the current output is orthogonal to the estimated errors of all previous inputs, the estimated error has an asymptotically independent and identically distributed (IID) Gaussian property, that is, in is Gaussian noise independent of x.

[0086] 2) Inverse random unitary transformation: Through the inverse random unitary transformation U H , we can get the estimated value in the time domain and variance

[0087]

[0088]

[0089] Through the properties of random unitary transformation, we can enhance The IID Gaussian property of the error in the like Figure 7 As shown, Closer to the mean is s and the variance is Gaussian signal.

[0090] 3) Orthogonal MMSE demodulation: φ t (·) By symbol-by-symbol MMSE demodulation Orthogonalization and damping are combined, and its output estimate is s t+1 for:

[0091]

[0092] in is the normalization and orthogonality coefficient, ζ t+1 is the damping vector, φ t The output variance of (·) for:

[0093]

[0094] 4) Random unitary transformation: Through random unitary transformation U, the estimated value x of the transformation domain can be obtained t+1 and variance

[0095] x t+1 =Us t+1 ,

[0096]

[0097] Then the estimated value x t+1 and variance Feedback to the orthogonal memory MF.

[0098] The entire CD-MAMP receiver goes through multiple iterations until the estimated signal s t+1 The entire detection process stops when accurate recovery can be achieved or the preset maximum number of iterations is reached.

[0099] Embodiment 2: This embodiment is an implementation step when the random unitary matrix coded modulation selects random unitary modulation in a MIMO scenario:

[0100] Solution 1:

[0101] Sending end: Figure 3 As shown in the figure, the digital modulation signal s is divided into N t segments, and add CP in each segment, and obtain the time domain signal x through the transmit filter. j , and at the jth antenna (j=1,...,N r ) is sent to the channel. The channel impulse response from the jth transmitting antenna to the mth receiving antenna is:

[0102]

[0103] Where n=0,...,N-1,m=1,...,N r ,p=0,...,P m,j -1,L m,j is the number of multipaths between the jth transmitting antenna and the mth receiving antenna. m,j,i ,f m,j,i ,τ m,j,i denote the channel gain, Doppler shift and delay associated with the i-th path respectively.

[0104] Receiving end: The received signal on the mth antenna at the receiving end passes through the receiving filter and removes the CP to obtain the received signal in the time domain:

[0105]

[0106] Then the received signal of random unitary modulation in MIMO system can be expressed in the form of a matrix, that is:

[0107]

[0108] in

[0109] Based on the received signal After cross-domain iteration, the receiver recovers the estimated signal Among them, the cross-domain iterative receiver process is linear detection: t =γ t (x t ),

[0110] Inverse random unitary transform:

[0111] Non-linear detection:

[0112] Random unitary transformation: x t+1 =Us t+1 , where t represents the number of iterations. Specifically, in the t-th iteration process, based on the received signal and the prior estimate x provided by nonlinear detection t , linear detection output estimation signal r t , and after inverse random unitary transformation, the signal is obtained Furthermore, the output estimation signal is obtained through nonlinear detection Generate x through random unitary transformation t+1 And input to the linear detection. After multiple iterations, until the estimated signal s t+1 The entire detection process stops when accurate recovery can be achieved or the preset maximum number of iterations is reached.

[0113] The CD-MAMP receiver proposed in the present invention can be used for detection. Specifically,

[0114] 1) Orthogonal memory MF: based on received signal and the output estimate x of the random unitary transformation t , the output estimate of the orthogonal memory MF is:

[0115]

[0116] Where X t =[x1,...,x t ], and

[0117]

[0118] in Used to ensure the convergence and orthogonality of MAMP. It is used to ensure the convergence and orthogonality of MAMP. The specific expression is as follows:

[0119] p t =[p t,1 ,p t,2 ,…,p t,t ]

[0120] p t,i =-θ t,i w t-i

[0121]

[0122]

[0123]

[0124]

[0125] in W t ≡A H B t A, λ min and λ max AA H The smallest and largest eigenvalues ​​of γ. t The output estimation variance of (·) is:

[0126]

[0127] Since the estimated error of the current output is orthogonal to the estimated errors of all previous inputs, the estimated error has an asymptotically independent and identically distributed (IID) Gaussian property, that is, in is Gaussian noise independent of x.

[0128] 2) Inverse random unitary transformation: Through the inverse random unitary transformation U H , we can get the estimated value in the time domain and variance

[0129]

[0130]

[0131] Through the properties of random unitary transformation, we can enhance The IID Gaussian property of the error in the like Figure 7 As shown, Closer to the mean is s and the variance is Gaussian signal.

[0132] 3) Orthogonal MMSE demodulation: φ t (·) By symbol-by-symbol MMSE demodulation Orthogonalization and damping are combined, and its output estimate is s t+1 for:

[0133]

[0134] in is the normalization and orthogonality coefficient, ζ t+1 is the damping vector, φ t The output variance of (·) for:

[0135]

[0136] 4) Random unitary transformation: Through random unitary transformation U, the estimated value x of the transformation domain can be obtained t+1 and variance

[0137] x t+1 =Us t+1 ,

[0138]

[0139] Then the estimated value x t+1 and variance Feedback to the orthogonal memory MF.

[0140] The entire CD-MAMP receiver goes through multiple iterations until the estimated signal The entire detection process stops when accurate recovery can be achieved or the preset maximum number of iterations is reached.

[0141] Option 2:

[0142] Sending end: Figure 4 As shown, the digital modulation signal s is divided into N t Segments, each segment is randomly unitarily modulated (i.e., after a random unitary matrix ), and add CP and generate signals through transmit filters respectively

[0143] Receiver: The difference from the receiver of Scheme 1 is the inverse random unitary transform and random unitary change in the cross-domain iterative receiver. Specifically, the estimated signal r output by the linear detection is t Divide into N t Segment, each segment is subjected to inverse random unitary modulation (i.e., after the inverse random unitary matrix ) and then merge to get the estimated signal Input to the nonlinear detection. For the estimated signal s output by the nonlinear detection t+1 Divide into N t Segments, each segment is randomly unitarily modulated (i.e., after the inverse random unitary matrix ) and then merge to get the estimated signal x t+1 , input into the linear detection.

[0144] Option 3:

[0145] Sending end: Figure 5As shown, the digital modulation signal s is divided into N t Segments, each segment is randomly unitarily modulated (i.e., after a random unitary matrix ), and add CP and generate signals through transmit filters respectively

[0146] Receiver: The difference from the receiver of Scheme 1 is the inverse random unitary transform and random unitary change in the cross-domain iterative receiver. Specifically, the estimated signal r output by the linear detection is t Divide into N t Segment, each segment is subjected to inverse random unitary modulation (i.e., after the inverse random unitary matrix ) and then merge into the random permutation inverse matrix Π H Get the estimated signal Input to the nonlinear detection. For the estimated signal s output by the nonlinear detection t+1 Divide into N t segment, which is input into the random permutation matrix Π and then divided into N t Segments, each segment is randomly unitarily modulated (i.e., after a random unitary matrix ), and then merge to get the estimated signal x t+1 , input into the linear detection.

[0147] Embodiment 3: This embodiment is an implementation step when the random unitary matrix coded modulation uses random unitary precoding, specifically:

[0148] Sending side: Figure 6 A communication system using a random unitary matrix as a precoding scheme is presented. First, the information bit sequence is channel coded to generate a sequence c. Then, it is digitally modulated and mapped into a symbol sequence s. S is precoded using a random unitary matrix U to obtain a new symbol sequence b, i.e., b=Us, and a transmission sequence x is generated through an existing transmission scheme, expressed as x=Pb=PUs. Finally, the sequence x is transmitted to the channel.

[0149] Receiving end: The signal received by the receiving end is expressed as:

[0150] y=Hx+w=HPUs+w=H eff s+w,

[0151] Among them, H eff =HPU. Based on the prior information of y, channel matrix H, transmission scheme P, random unitary precoding matrix U and s, a cross-domain iterative receiver can be used to recover

[0152] Based on the received signal y, the receiver recovers the estimated signal after cross-domain iteration Among them, the cross-domain iterative receiver process is linear detection:t =γ t (x t ),

[0153] Inverse random unitary transform:

[0154] Non-linear detection:

[0155] Random unitary transformation: x t+1 =Us t+1 , where t represents the number of iterations. Specifically, in the t-th iteration process, based on the received signal and the prior estimate x provided by nonlinear detection t , linear detection output estimation signal r t , and after inverse random unitary transformation, the signal is obtained Furthermore, the output estimation signal is obtained through nonlinear detection Generate x through random unitary transformation t+1 And input to the linear detection. After multiple iterations, until the estimated signal s t+1 The entire detection process stops when accurate recovery can be achieved or the preset maximum number of iterations is reached.

[0156] The CD-MAMP receiver proposed in the present invention can be used for detection. Specifically,

[0157] 1) Orthogonal Memory MF: Based on the received signal y and the output estimate x of the random unitary transformation t , the output estimate of the orthogonal memory MF is:

[0158]

[0159] Where X t =[x1,...,x t ], and

[0160]

[0161] in It is used to ensure the convergence and orthogonality of MAMP. The specific expression is as follows:

[0162] p t =[p t,1 ,p t,2 ,…,p t,t ]

[0163] p t,i =-θ t,i w t-i

[0164]

[0165]

[0166]

[0167]

[0168] in λ min and λ max AA H The minimum and maximum eigenvalues ​​of . According to (17), γ t The output estimation variance of (·) is:

[0169]

[0170] Since the estimated error of the current output is orthogonal to the estimated errors of all previous inputs, the estimated error has an asymptotically independent and identically distributed (IID) Gaussian property, that is, in is Gaussian noise independent of x.

[0171] 2) Inverse random unitary transformation: Through the inverse random unitary transformation U H , we can get the estimated value in the time domain and variance

[0172]

[0173]

[0174] Through the properties of random unitary transformation, we can enhance The IID Gaussian property of the error in the like Figure 7 As shown, Closer to the mean is s and the variance is Gaussian signal.

[0175] 3) Orthogonal MMSE demodulation: φ t (·) By symbol-by-symbol MMSE demodulation Orthogonalization and damping are combined, and its output estimate is s t+1 for:

[0176]

[0177] in is the normalization and orthogonality coefficient, ζ t+1 is the damping vector, φt The output variance of (·) for:

[0178]

[0179] 4) Random unitary transformation: Through random unitary transformation U, the estimated value x of the transformation domain can be obtained t+1 and variance

[0180] x t+1 =Us t+1 ,

[0181]

[0182] Then the estimated value x t+1 and variance Feedback to the orthogonal memory MF.

[0183] The entire CD-MAMP receiver goes through multiple iterations until the estimated signal The entire detection process stops when accurate recovery can be achieved or the preset maximum number of iterations is reached.

[0184] The main idea of ​​existing multi-carrier modulation technologies (OFDM, OTFS, AFDM, etc.) is to remove the interference between symbols as much as possible and obtain an equivalent sparse channel matrix, so as to recover the signal using a simplified low-complexity receiver. However, each signal only experiences a small amount of channel fading, resulting in a degradation in detection performance.

[0185] In comparison, the random unitary modulation proposed in the present invention is a non-orthogonal transmission mode, which allows all signals to be randomly superimposed together so that each signal can experience statistically stable channel fading, thereby ensuring that the maximum diversity gain can be obtained.

[0186] In addition, the random unitary modulation proposed in the present invention can also be used as a precoding method and applied to a communication system characterized by a standard linear model.

[0187] In terms of the receiver, the present invention uses random unitary modulation of CD-MAMP to achieve performance far superior to OFDM, OTFS, and AFDM with extremely low complexity.

[0188] The numerical simulation results are given below, taking IFDM in random unitary modulation as an example.

[0189] For simplicity, the simulation parameters are set as follows: carrier frequency 4 GHz, subcarrier spacing Δf = 15 kHz, mobile station speed {300 km / h, 500 km / h}, maximum Doppler shift v max={1111,1852}Hz, the channel Doppler shift is generated by Jakes information, the RRC roll factor in the transceiver is set to 0.4, N = M = 512 (the number of subcarriers in OTFS is 32 and the number of time slots is 16), and the signal-to-noise ratio is defined as E{||Hx|| 2} / σ 2 .

[0190] Fig. 9 and Fig.10 The BER performance comparison of IFDM, OFDM, OTFS and AFDM using CD / DD-OAMP and CD / DD-MAMP receivers in SISO and 4*4MIMO scenarios is given to verify the performance advantage of IFDM. Fig.10 Medium r =N s =4. When BER is 10 -5 When the speed is 300km / h, in the SISO scenario, CD-MAMP's IFDM achieves 3dB and 9dB gains compared to OAMP's OTFS / AFDM and MAMP's OTFS / AFDM, respectively. In the MIMO scenario, CD-MAMP's IFDM achieves more than 5dB gains compared to OAMP's OTFS, AFDM and OFDM. Fig.11 As shown in the figure, when the BER is 10 -5 When the speed is 500km / h, in the SISO scenario, the IFDM of CD-MAMP obtains 2dB and 4dB gains compared with the OTFS / AFDM of OAMP.

[0191] Fig.12 Further, in the 4*4 MIMO scenario, the BER is 10 -4 When N r =N s = 4. Note that the running time of MIMO-IFDM of CD-MAMP is only 0.7% of OTFS / AFDM of CD / DD-OAMP, 8% of MIMO-OTFS of DD-MAMP, and slightly lower than MIMO-OTFS / OFDM of CD-MAMP.

[0192] To further verify the advantages of IFDM and CD-MAMP, the comparison of different modulation schemes is summarized in Table 1. Table 2 summarizes the comparison of different receivers, where T is the maximum number of iterations of the receiver.

[0193] Table 1 Comparison of OFDM, OTFS, AFDM and IFDM

[0194]

[0195] Table 2 Comparison of different receivers

[0196]

[0197] The above embodiments are used to illustrate the present invention rather than to limit the present invention. Any modification and change made to the present invention within the spirit of the present invention and the protection scope of the claims shall fall within the protection scope of the present invention.

Claims

1. A signal transmission method based on random unitary matrix coding modulation and cross-domain iterative detection, characterized in that: The method includes random unitary matrix coding modulation at a signal transmitting end and cross-domain iterative detection at a receiving end; The random unitary matrix coded modulation is implemented by random unitary modulation or random unitary precoding; The random unitary modulation includes: digitally modulating the information bit sequence at the transmitter, performing serial-to-parallel conversion on the modulated information and performing random unitary modulation through a random unitary transformation matrix, adding a cyclic prefix to the signal vector after random unitary modulation, and sending it to the channel after passing through a transmission filter; The random unitary precoding includes: performing channel coding and digital modulation on the information bit sequence at the transmitter, performing unitary precoding on the modulated information through a random unitary transformation matrix to obtain a new symbol sequence, combining the new symbol sequence with a transmission scheme to obtain a transmission signal for transmission; The cross-domain iterative detection is to use a cross-domain iterative receiver to receive a signal modulated by random unitary matrix coding. The use process of the cross-domain iterative receiver is specifically as follows: based on the received signal y and the prior estimate x provided by the nonlinear detection t , linear detection output estimation signal r t , and after inverse random unitary transformation, the signal is obtained After further nonlinear detection, the output estimated signal s is obtained. t+ 1. Generate x through random unitary transformation t+ 1 and input to linear detection; iterate to estimate signal s t+1 After accurate recovery or reaching the preset maximum number of iterations, the entire detection process stops and the final estimated signal is output as the received signal.

2. The signal transmission method based on random unitary matrix coding modulation and cross-domain iterative detection according to claim 1 is characterized in that: In the random unitary modulation, the random unitary transformation matrix is ​​a unitary invariant matrix.

3. The signal transmission method based on random unitary matrix coding modulation and cross-domain iterative detection according to claim 1 is characterized in that: The cross-domain iterative receiver uses a Bayesian optimal cross-domain iterative receiver. Specifically, the Bayesian optimal cross-domain iterative receiver multiplies the sparse time domain channel matrix with the input signal by using the time domain memory MF in the linear detection of the cross-domain iterative receiver, and uses orthogonal MMSE demodulation in nonlinear detection.

4. The signal transmission method based on random unitary matrix coding modulation and cross-domain iterative detection according to claim 1 is characterized in that: The transmitting end processing of the random unitary modulation in the MIMO scenario includes: performing signal segmentation after the random unitary modulation or performing signal segmentation before the random unitary modulation; the performing signal segmentation before the random unitary modulation also includes using a random permutation matrix segmentation before the random unitary modulation; The corresponding receiving end processing is: receiving the signal through the receiving filter at the antenna of the receiving end respectively and combining the signals after deleting the cyclic prefix; after combining, using the cross-domain iterative receiver to receive the signal modulated by the random unitary matrix coding.

5. The signal transmission method based on random unitary matrix coding modulation and cross-domain iterative detection according to claim 4 is characterized in that: The signal segmentation after random unitary modulation is specifically as follows: After the digital modulated signal is converted from serial to parallel and modulated by random unitary modulation, it is divided into several segments, and a cyclic prefix is ​​added to each segment. After the signal is obtained through the transmit filter, it is sent to the channel through the antenna. The corresponding cross-domain iterative receiver processing is: based on the received signal y and the prior estimate x provided by the nonlinear detection t , linear detection output estimation signal r t , and after inverse random unitary transformation, the signal is obtained After further nonlinear detection, the output estimated signal s is obtained. t+1 , after random unitary transformation, x is generated t+1 And input to linear detection; iterate to estimate signal s t+1 After the accurate recovery can be achieved or the preset maximum number of iterations is reached, the entire detection process stops and the final recovered estimated signal is output as the received signal.

6. The signal transmission method based on random unitary matrix coding modulation and cross-domain iterative detection according to claim 4 is characterized in that: The signal segmentation before random unitary modulation is specifically as follows: After the digital modulated signal is converted from serial to parallel, it is divided into several segments, and each segment is randomly unitary modulated and a cyclic prefix is ​​added. After the signal is obtained through the transmit filter, it is sent to the channel through the antenna. The corresponding receiving end processing is: receiving the signal through the receiving filter at the antenna of the receiving end respectively and deleting the cyclic prefix and then combining the signals; after combining, using the cross-domain iterative receiver to receive the signal modulated by the random unitary matrix coding, wherein the implementation steps of the cross-domain iterative receiver include: converting the estimated signal r output by the linear detection t Divide into N t Segments, each segment is subjected to inverse random unitary modulation and then combined to obtain an estimated signal Input to the nonlinear detection; the estimated signal s for the nonlinear detection output t+ 1 divided into N t Segments, each segment is randomly unitarily modulated and then combined to obtain the estimated signal x t+ 1, input to linear detection, iterate to estimate signal s t+1 After the accurate recovery can be achieved or the preset maximum number of iterations is reached, the entire detection process stops and the final recovered estimated signal is output as the received signal.

7. The signal transmission method based on random unitary matrix coding modulation and cross-domain iterative detection according to claim 4 is characterized in that: The use of random permutation matrix segmentation before random unitary modulation is specifically: The digital modulated signal is divided into several segments after serial-to-parallel conversion and random permutation matrix. Each segment is randomly unitary modulated and a cyclic prefix is ​​added. The time domain signal is obtained through the transmit filter and sent to the channel through the antenna. The corresponding receiving end processing is: receiving the signal through the receiving filter at the antenna of the receiving end respectively and deleting the cyclic prefix and then combining the signals; after combining, using the cross-domain iterative receiver to receive the signal modulated by the random unitary matrix coding, wherein the implementation steps of the cross-domain iterative receiver include: converting the estimated signal r output by the linear detection t Divide into N t Segments, each segment is subjected to inverse random unitary modulation and then merged into the random permutation inverse matrix Π H Get the estimated signal Input to the nonlinear detection; the estimated signal s for the nonlinear detection output t+1 Divide into N t segment, which is input into the random permutation matrix Π and then divided into N t Segments, each segment is randomly unitarily modulated, and then combined to obtain the estimated signal x t+1 , input into the linear detection; iterate to estimate the signal s t+1 After the accurate recovery can be achieved or the preset maximum number of iterations is reached, the entire detection process stops and the final recovered estimated signal is output as the received signal.

8. The signal transmission method based on random unitary matrix coded modulation and cross-domain iterative detection according to claim 1 is characterized in that: The random unitary matrix coded modulation uses random unitary precoding specifically as follows: The information bit sequence is channel coded to generate sequence c, which is mapped into symbol sequence s through digital modulation. S is precoded with random unitary matrix U to obtain new symbol sequence b, i.e. b=Us, and the transmission sequence x is generated through the existing transmission scheme P, expressed as x=Pb=PUs. The sequence x is transmitted to the channel, and the signal y received by the receiving end is expressed as: y=Hx+w=HPUs+w=H eff s+w, Among them, H eff =HPU, H is the time domain channel matrix, H eff is the transform domain equivalent channel, w~CN(0,σ 2 I) is additive white Gaussian noise.

Images