Channel adaptive hybrid superposition coding modulation method

Abstract

The invention belongs to the technical field of wireless communication, and particularly relates to a channel-adaptive hybrid superposition coding modulation method. According to the method provided by the invention, a traditional superposition coded modulation system is improved, zero filling is carried out on a transmitting end by utilizing a channel feedback technology and combining channel information, and compared with the original superposition coded modulation system, the bit error rate is reduced and the spectrum efficiency is flexibly realized at the same time; the method has the potential of being combined with other detection algorithms.

Description

Technical Field

[0001] The present invention belongs to the technical field of wireless communication, and specifically relates to a channel adaptive hybrid superposition coding modulation method. Background Art

[0002] With the continuous increase of wireless communication requirements, spectrum resources have become increasingly tense. As an efficient modulation technology, Superposition Coded Modulation (SCM) linearly superimposes multiple independent signals in the same channel, making the transmitted signal approximately Gaussian distributed and effectively improving the spectrum utilization rate. Compared with traditional coding modulation methods, such as Trellis Coded Modulation and Bit-Interleaved Coded Modulation with iterative decoding, SCM has the advantages of high transmission efficiency, high diversity gain, adaptable code rate adjustment, low receiving detection complexity and high reliability. The transmitted signal of the 1D-SCM system is a real-valued symbol, and the 2D-SCM system introduces the complex domain, which has a higher spectrum efficiency. The receiver of the SCM system adopts a low-complexity iterative detection structure similar to the Turbo type. The Elementary Signal Estimator (ESE) and the Decoder (DEC) are important components of the receiver.

[0003] In a communication system, Channel State Information (CSI) is very important. CSI reflects the real-time characteristics of the channel and can help the system optimize the coding and modulation methods. However, traditional SCM methods often can only superimpose a fixed number of layers for a specific spectrum efficiency and fail to fully utilize the channel feedback information for dynamic adjustment, which may lead to a decrease in transmission efficiency and lack of channel adaptability in poor channel conditions. To address these problems, based on the poor performance of the SCM system in high-layer superposition, the present invention combines different-layer SCM systems through hybrid superposition according to a specific spectrum efficiency and channel feedback, further reducing the bit error rate of SCM. Summary of the Invention

[0004] The present invention proposes a new method for hybrid superposition of SCM layers based on spectrum efficiency and channel feedback for traditional SCM systems, simply referred to as the Channel Adaptive-Hybrid SCM (CA-Hybrid SCM) system. Compared with traditional SCM systems, by adjusting the superposition layer through CSI feedback, the present invention can maintain a high transmission rate when the channel condition is good, and reduce signal interference through Zero Padding (ZP) to reduce the bit error rate of the system and improve the transmission performance when the channel condition is poor, with wide applicability.

[0005] For easy understanding, the zero-padding scheme and the new method of channel feedback mechanism adopted in the present invention are described as follows:

[0006] The block diagram of the traditional K-layer 2D-SCM system is as shown Figure 1 In the transmitter, the binary data sequence d is divided into K sub-sequences {d k}, and the k-th sub-sequence d k is encoded by the binary forward error correction code (FEC-k) of the k-th layer to generate a coded bit sequence c k ={c k (t)}, where c k (t) ∈ (0, 1), and T is the length of the chip sequence. Then, c k is randomly interleaved by the interleaver (INTL-k) of the k-th layer into a sequence v k of length 2T. Then, the sequence v k is mapped to a QPSK symbol sequence using Gray mapping where

[0007] The output signal at time t is a linear superposition of K coded symbols:

[0008]

[0009] where {β k} are constant weight factors. The total rate R k is the coding rate of the binary component code of the k-th layer. It can be seen from (1) that {β k} is a key parameter of SCM, and different constellations with different shapes and performances can be generated by optimizing it. Each |β k | 2 can be regarded as the power allocated to the symbols of the k-th layer.

[0010] At the receiver of the 2D-SCM system, after the received data is sent into the ESE estimator, the ESE estimator uses the received signal, channel information, and the prior information fed back by the decoder to obtain the extrinsic information of the received signal. After Turbo iteration, the extrinsic information is sent into the DEC to output the final decision result.

[0011] The technical solution of the present invention is as follows:

[0012] A multi-layer SCM hybrid superposition scheme according to spectral efficiency and combined with channel feedback. It is defined that the required spectral efficiency of the 2D-SCM system is η = KR′, where K is the total number of layers of the 2D-SCM system, R′ is the coding rate of each layer encoder, the total number of input information bits is KN bits, and N is the number of input information bits of each layer. The modulation method includes the following steps:

[0013] Transmitter:

[0014] S1. At the transmitter, first, according to the spectral efficiency η, the average number of layers of the 2D - SCM system can be obtained The transmitted data is divided into two parts, K1 - layer and K2 - layer, which are mixed and superimposed, and K1 and K2 satisfy the following formula:

[0015]

[0016] where,

[0017] Therefore, the k - th layer bit sequence after interleaving is divided into two parts with a size of This part of the bits needs to be zero - padded. with a size of This part of the bits does not need to be operated on.

[0018] S2. The SCM system with a high number of layers superimposed is equivalent to an SCM system with a low number of layers superimposed. First, After BPSK modulation, it becomes {b k} = {b k (t)}, where b k (t) ∈ (-1, 1).

[0019] S3. Then, perform zero - padding operation on to get {u k}}. {b k} is a matrix B with a size of K2×2M, and after zero - padding, it becomes a matrix U with a size of K2×2L, where The relationship between the element u m,n in U and the element b m,n in B is:

[0020]

[0021] where, 0 < m ≤ K2, 0 < n ≤ 2L.

[0022] After zero - padding, the superposition transmission of the K2 - layer data {u k} can be equivalent to the superposition of the K1 - layer data G = {g l , l = 1, 2,..., K1}. The relationship between the element g p,q in G and the element u p,q in U is:

[0023]

[0024] where, 0 < p ≤ K1, 0 < q ≤ 2L.

[0025] S4. Perform QPSK modulation on the bit sequences {g l (t), t = 1, 2,..., 2L} and respectively to obtain the modulated symbol sequences {z l (t), t = 1, 2,..., L} and {x k (t), t = L + 1, L + 2,..., L + W}.

[0026] S5. Through channel feedback technology, obtain the channel state matrix {H(t), t = 1, 2,..., L + W}. Calculate the Frobenius norm:

[0027] h(t) = ||H(t)|| F (5)

[0028] Sort h′(t) = {h(t1) ≤ h(t2) ≤... ≤ h(t L+W )} in ascending order according to the norm size, and then divide it into {h1′(t i ), i = 1, 2,..., L} and {h2′(t j ), j = L + 1, L + 2,..., L + W}, and let S1 = {t i , i = 1, 2,..., L}, S2 = {t j , j = L + 1, L + 2,..., L + W}.

[0029] Adjust the transmission symbol order according to the magnitude of the channel gain strength. After adjustment, the symbol sequence {z l (t)} becomes {z l (t), t ∈ S1}, and the new transmission order of the symbol sequence {x k (t)} after adjustment is {x k (t), t ∈ S2}.

[0030] Receiver:

[0031] S6. The received signal of the SCM system after hybrid superposition can be expressed as:

[0032] y(t) = H(t)s(t) + n(t), t = 1, 2,..., L + W (6)

[0033] where n(t) represents additive white Gaussian noise (AWGN), and the transmitted signal s(t) is:

[0034]

[0035] Among them, {α l} and {β k} are the power allocations when the number of layers is K1 and K2.

[0036] S7. The MMSE signal detection can maximize the SINR after detection. First, let its weighting matrix be:

[0037]

[0038] Among them, is the noise variance, and I is the identity matrix. Using the MMSE weighting matrix in Equation (8) for estimation, the estimation result can be obtained as:

[0039]

[0040] Among them, is the estimated value of the transmitted signal, is the noise after detection.

[0041] According to the noise expected value formula given in "Yong, Soo Cho. MIMO - OFDM Wireless Communication Technology and MATLAB Implementation", it can be calculated that:

[0042]

[0043] Among them, is the noise variance, is the singular value of the channel matrix H(t).

[0044] S8. According to different channel conditions, is sent into the ESE estimator of the K1 - layer 2D - SCM system to obtain the output e ESE (z l ) = {e ESE (z l (t)), t ∈ S1}, is sent into the ESE estimator of the K2 - layer 2D - SCM system to obtain the output e ESE (x k ) = {e ESE (x k (t)), t ∈ S2}. E = e ESE (z l ) is a matrix of size K1×2L. It is restored to a matrix F = e′ ESE (z l ). The relationship between the element f i,j in F and the element e i,j in E is:

[0045]

[0046] wherein 0 < i ≤ K2, 0 < n ≤ 2M.

[0047] Finally, the outputs of all ESE estimators are

[0048] The beneficial effects of the present invention are as follows. The present invention comprehensively considers channel adaptation and the hybrid superposition method of the SCM layer number, applies channel feedback and zero-padding schemes in the SCM system, and reduces the interference between different user layers through zero-padding under poor channel conditions, improving the bit error rate performance of the original SCM detection. At the same time, the present invention also has good scalability and is applicable to multi-user scenarios in various wireless communication systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 is a block diagram of a traditional SCM system;

[0050] Figure 2 is a block diagram of the channel-adaptive CA-hybrid SCM system proposed by the present invention;

[0051] Figure 3 is a schematic diagram comparing the bit error rate performance of a traditional SCM system and a CA-hybrid SCM system when the system modulation method is QPSK, the channel coding rate R' = 0.5, and the spectral efficiency is 2.4;

[0052] Figure 4 is a schematic diagram comparing the bit error rate performance of a traditional SCM system and a CA-hybrid SCM system when the system modulation method is QPSK, the channel coding rate R' = 0.5, and the spectral efficiency is ; DETAILED DESCRIPTION OF THE INVENTION

[0053] The technical method of the present invention will be described in detail below in conjunction with the drawings and embodiments:

[0054] Taking the spectral efficiency η = 2.4, the information bit length of each layer is 512 bits, the channel coding is a convolutional code with a code rate of 0.5, and the modulation method is QPSK as an example, assuming that the channel is a flat Rayleigh fading channel, and both the transmitter and the receiver can perfectly obtain the channel state information H(t).

[0055] Transmitter:

[0056] S1. At the transmitter, first, according to the spectral efficiency η, the average number of layers of the 2D-SCM system can be obtained The transmitted data is divided into a hybrid superposition of K1 layers and K2 layers, and K1 and K2 satisfy the following formula:

[0057]

[0058] Among them, K1 = 2 and K2 = 3.

[0059] Therefore, the k-th layer bit sequence after being interleaved is divided into two parts with a size of This part needs to be zero-padded. with a size of This part does not need any operation.

[0060] S2. Equivalent the SCM system with high-layer superposition to the SCM system with low-layer superposition. First, for After BPSK modulation, it becomes {b k} = {b k (t)}, where b k (t) ∈ (-1, 1).

[0061] S3. Then, perform zero-padding operation on to obtain {u k}}. {b k} is the matrix B with a size of 3 × 512, and after zero-padding, it becomes the matrix U with a size of 3 × 768. The relationship between the element u m,n in U and the element b m,n in B is:

[0062]

[0063] where 0 < m ≤ 3, 0 < n ≤ 768.

[0064] After zero-padding, the superposition transmission of the 3-layer data {u k} can be equivalent to the superposition of the 2-layer data G = {g l , l = 1, 2}. The relationship between the element g p,q in G and the element u p,q in U is:

[0065]

[0066] where 0 < p ≤ 2, 0 < n ≤ 768.

[0067] S4. Respectively perform QPSK modulation on the bit sequences {g l (t), t = 1, 2,..., 768} and to obtain the modulated symbol sequences {z l (t), t = 1, 2,..., 384} and {x k (t), t = 385, 386,..., 640}.

[0068] S5. Obtain the channel state matrix {H(t), t = 1, 2, ..., 640} through channel feedback technology. Calculate the Frobenius norm:

[0069] h(t) = ||H(t)|| F (15)

[0070] Sort h′(t) = {h(t1) ≤ h(t2) ≤... ≤ h(t 680 )} in ascending order according to the norm size, and then divide it into {h1′(t i ), i = 1, 2, ..., 384} and {h2′(t j ), j = 385, 386, ..., 640}, and let S1 = {t i , i = 1, 2, ..., 384}, S2 = {t j , j = 385, 386, ..., 640}.

[0071] Adjust the transmission symbol order according to the magnitude of the channel gain strength. After adjustment, the symbol sequence {z l (t)} becomes {z l (t), t ∈ S1}, and the new transmission order of the symbol sequence {x k (t)} after adjustment is {x k (t), t ∈ S2}.

[0072] Receiving end:

[0073] S6. The received signal of the SCM system after hybrid superposition can be expressed as:

[0074] y(t) = H(t)s(t) + n(t), t = 1, 2, ..., 640 (16)

[0075] where n(t) represents AWGN, and the transmitted signal s(t) is:

[0076]

[0077] According to "Tong, J. Superposition Coded Modulation. PhD diss., City University of Hong Kong, 2009", when the number of layers is 2 and 3, the power allocations {α l} = {1, 1.51} and {β k} = {1, 1.4717, 2.138} are used respectively.

[0078] S7. MMSE signal detection can maximize the SINR after detection. First, let its weighting matrix be:

[0079]

[0080] where is the noise variance and I is the identity matrix. Using the MMSE weighting matrix of Equation (18) for estimation, the estimation result can be obtained as:

[0081]

[0082] where is the estimated value of the transmitted signal, is the detected noise.

[0083] Calculated according to the noise expected value formula given in "Yong, Soo Cho. MIMO-OFDM Wireless Communication Technology and MATLAB Implementation":

[0084]

[0085] where is the noise variance, are the singular values of the channel matrix H(t).

[0086] S8. According to different channel conditions, is sent to the ESE estimator of the 2-layer 2D-SCM system to obtain the output e ESE (z l ) = {e ESE (z l (t)), t ∈ S1}, is sent to the ESE estimator of the 3-layer 2D-SCM system to obtain the output e ESE (x k ) = {e ESE (x k (t)), t ∈ S2}. E = e ESE (z l ) has a size of 2 × 768, and it is restored to a matrix F = e′ ESE (z l ) with a size of 3 × 512. The relationship between the element f i,j in F and the element e i,j in E is:

[0087]

[0088] where 0 < i ≤ 3, 0 < j ≤ 512.

[0089] Finally, the outputs of all ESE estimators are

Claims

1. A channel adaptive hybrid superposition coding modulation method, defining that the spectral efficiency required by the 2D-SCM system is η = KR′, where K is the total number of layers of the 2D-SCM system, R′ is the coding rate of each layer encoder, the total number of input information bits is KN bits, and N is the number of information bits input for each layer. It is characterized in that, The modulation method includes the following steps: S1. Obtain the average number of layers of the 2D-SCM system according to the spectral efficiency η Divide the transmitted data into a hybrid superposition of K1 layers and K2 layers, where K1 and K2 satisfy the following formula: Among them, Correspondingly, the k-th layer bit sequence after being interleaved by the interleaver is divided into two parts with a size of with a size of S2. After performing BPSK modulation on we obtain where b k (t) ∈ (-1, 1), is the matrix B of size K2 × 2M; S3. For perform zero-padding operation to obtain {u k}, after zero-padding, matrix B becomes matrix U of size K2×2L, where the element u m,n in U and the element b m,n in B have the following relationship: Among them, 0 < m ≤ K2, 0 < n ≤ 2L; After zero-padding, the superposition transmission of the data {u k} in the K2 layer is equivalent to the superposition of the data G = {g l , l = 1, 2,..., K1} in the K1 layer. The relationship between the element g p,q in G and the element u p,q in U is as follows: Among them, 0 < p ≤ K1, 0 < q ≤ 2L; S4. Perform QPSK modulation on the bit sequences {g l (t), t = 1, 2, ..., 2L} and respectively to obtain the modulated symbol sequences {z l (t), t = 1, 2, ..., L} and {x k (t), t = L + 1, L + 2, ..., L + W}; S5. Obtain the channel state matrix {H(t), t = 1, 2,..., L + W} through channel feedback technology, and calculate the Frobenius norm: h(t) = ||H(t)|| F , Sort \(h′(t)=\{h(t1)\leq h(t2)\leq...\leq h(t L+W )\) in ascending order according to the norm size, and then divide it into \(\{h1′(t i ),i = 1,2,...,L\}\) and \(\{h′2(t j ),j = L + 1,L + 2,...,L + W\}\). Let \(S1=\{t i ,i = 1,2,...,L\}\), \(S2=\{t j ,j = L + 1,L + 2,...,L + W\}\); Adjust the transmission symbol order according to the magnitude of the channel gain intensity. After adjustment, the symbol sequence {z l (t)} becomes {z l (t), t ∈ S1}, and the new transmission order of the symbol sequence {x k (t)} after adjustment is {x k (t), t ∈ S2}; S6. The received signal of the 2D-SCM system after hybrid superposition is expressed as: y(t) = H(t)s(t) + n(t), t = 1, 2,..., L + W, where n(t) represents additive white Gaussian noise (AWGN), and the transmitted signal s(t) is: Among them, {α l} and {β k} are the power allocations when the number of layers is K1 and K2; S7. The weighted matrix for MMSE estimation is: Among them, is the noise variance, and I is the identity matrix; The estimation result is: Among them, is the estimated value of the transmitted signal, is the noise after detection; It can be calculated according to the noise expected value formula: Among them, is the noise variance, are the singular values of the channel matrix H(t); S8. According to different channel conditions, is sent to the ESE estimator of the K1-layer 2D-SCM system to obtain the output e ESE (z l ) = {e ESE (z l (t)), t ∈ S1}, is sent to the ESE estimator of the K2-layer 2D-SCM system to obtain the output e ESE (x k ) = {e ESE (x k (t)), t ∈ S2}; E = e ESE (z l ) is a matrix of size K1×2L, which is restored to a matrix F = e' ESE (z l ) of size K2×2M. The relationship between the element f i,j in F and the element e i,j in E is: Among them 0 < i ≤ K2, 0 < n ≤ 2M; Finally, the output of all ESE estimators is

Images