Uplink multi-user random modulation and OAMP detection method

Abstract

The invention discloses an uplink multi-user random modulation and OAMP detection method, which comprises the following steps of: independently and randomly modulating each user, transmitting a signal, and receiving the sum of the signals of each user; performing linear estimation according to the prior distribution of the received signal and the transmitted signal of each user, the state information of the channel matrix, the estimation error of the channel matrix and the channel noise power; nonlinear estimation is carried out according to the prior distribution of the signals transmitted by the users and an estimator and a variance obtained through linear estimation, and a nonlinear estimator and a variance in the iteration are obtained; and carrying out continuous iteration on the linear estimation and the non-linear estimation, and carrying out hard decision according to a linear estimator and a variance obtained by final iteration to obtain a final decoding result. According to the method, independent random modulation is carried out on all users, channel estimation errors are taken into consideration, the whole equivalent channel matrix does not need to have the right unitary invariant characteristic, the performance loss is reduced, and the detection performance obviously superior to that of an existing OAMP is obtained.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of signal detection, and in particular to an uplink multi-user random modulation and OAMP detection method. BACKGROUND

[0002] The Orthogonal Approximate Message Passing (OAMP) algorithm is a kind of algorithm commonly used in the receiving end signal detection of the current communication scenario. As a variant of the Approximate Message Passing (AMP) algorithm, the OAMP algorithm solves the limitation that the AMP algorithm can only be applied to the Independent and Identically Distributed (IID) Gaussian matrix, and expands the matrix range that the algorithm can be applied to the right unitary invariant matrix.

[0003] However, even though the application range of the OAMP algorithm is expanded compared to the original AMP algorithm, it still cannot be well applied to the actual communication scenario, because the channel matrix in the actual communication scenario often does not have the right unitary invariant feature. Therefore, in order to enable the OAMP algorithm to be applied to signal detection in the actual communication scenario, a random modulation technology can be used to precode the transmitted signal. After random modulation, the equivalent channel matrix of the system will have the right unitary invariant feature, so that the OAMP algorithm can be used for signal detection.

[0004] However, the above-mentioned OAMP algorithm is generally only used for single-user signal detection. In a multi-user communication scenario, users are relatively independent, and cannot jointly randomize the transmitted signals of all users, and need to randomize independently between users. Therefore, when the Channel State Information (CSI) between multiple users in the system is asymmetric, the entire equivalent channel matrix of the system relative to all users does not have the right unitary invariant feature, and the OAMP algorithm will have a significant performance loss. In addition, the OAMP algorithm for signal detection previously assumes that the CSI of the channel is accurate, but in actual situations, the estimated channel CSI will always deviate from the true CSI of the channel, and when the CSI deviation is relatively obvious, the OAMP algorithm will also have a significant performance loss.

[0005] The commonly used multi-user signal detection at present stage often needs to modulate the signals of each user by using orthogonal frequency division multiple access (OFDMA) technology at the transmitting end, and performs minimum mean square error (MMSE) detection at the receiving end. However, this detection method can only be applied to static multipath channels, because if the communication channel is a time-varying multipath channel, the OFDMA technology will not diagonalize the equivalent channel matrix, so that the MMSE detection cannot be performed.

[0006] In the uplink multi-user scenario, joint random modulation cannot be performed on all users, and each user needs to perform random modulation independent of each other. When the CSI between users is asymmetric, the equivalent channel matrix does not have the right unitary invariant property, and the OAMP algorithm cannot be applied.

[0007] When the estimated channel CSI is significantly deviated from the real CSI, the OAMP algorithm will have a significant performance loss.

[0008] The multi-user MMSE detection based on the OFDMA technology will fail in the time-varying multipath channel. SUMMARY

[0009] The present application aims at the deficiencies of the prior art, and provides an uplink multi-user random modulation and OAMP detection method.

[0010] The purpose of the present application is achieved by the following technical scheme: an uplink multi-user random modulation and OAMP detection method, comprising:

[0011] S1, transmitting the signal after independently modulating each user at the transmitting end;

[0012] S2, receiving the sum of the signals of each user from the receiving end;

[0013] S3, performing linear estimation according to the received signal, the state information of the channel matrix, the estimation error of the channel matrix and the channel noise power;

[0014] calculating the orthogonal parameter according to the channel state information, the estimation error of the channel matrix and the channel noise power;

[0015] calculating the linear estimation quantity of the current iteration process according to the orthogonal parameter, the channel state information, the estimation error of the channel matrix, the channel noise power and the nonlinear estimation quantity in the last iteration process;

[0016] The linear estimation variance in the current iteration process is obtained according to the orthogonal parameters, channel state information, estimation error of the channel matrix, channel noise power and the nonlinear estimation variance in the last iteration process;

[0017] S4, performing nonlinear estimation according to the prior distribution of the signal transmitted by each user, the estimated value and the variance obtained by linear estimation, to obtain the nonlinear estimated value and the variance in the current iteration process; continuously iterating the linear estimation and the nonlinear estimation until convergence or the maximum iteration number is reached, to obtain the linear estimated value and the variance in the last iteration;

[0018] S5, performing hard decision according to the obtained linear estimated value and the variance, to obtain the final decoding result.

[0019] On the other hand, the specification of the present application also provides an uplink multi-user random modulation and OAMP detection device, comprising a memory and one or more processors, wherein the memory stores executable codes, and when the processor executes the executable codes, the uplink multi-user random modulation and OAMP detection method is realized.

[0020] On the other hand, the specification of the present application also provides a computer readable storage medium, which stores a program, and when the program is executed by a processor, the uplink multi-user random modulation and OAMP detection method is realized.

[0021] The present application has the following beneficial effects: the previous OAMP algorithm needs to perform joint random modulation on all users, which is not feasible in an actual multi-user communication scenario. However, the present application requires different users to perform independent random modulation on each other in a multi-user communication scenario, which is a more practical and feasible solution.

[0022] When the users perform independent random modulation on each other, if there is a significant gap in the CSI between the users, the entire equivalent channel matrix of the system no longer has the right unitary invariant property, and the ordinary OAMP detection technology will have a large performance loss. However, the MU-OAMP detection technology proposed in the present application does not require the entire equivalent channel matrix to have the right unitary invariant property, and the detection performance is obviously better than that of the ordinary OAMP.

[0023] The previous OAMP algorithm assumes that the CSI of the channel is accurate when applied, but when there is a significant deviation between the estimated channel CSI and the real CSI of the channel, the detection performance will decrease significantly. However, the MU-OAMP detection technology proposed in the present application takes into account the channel estimation error, and by modifying the LMMSE filter, the detection performance is obviously better than that of the OAMP.

[0024] The MMSE detection based on the OFDMA technology commonly used in the communication field for multi-user signal detection at the present stage can be realized only under the static channel, and the MU-OAMP detection technology provided by the application can obtain better detection performance under the dynamic channel, and even if under the static channel, the detection performance is obviously superior to the former. BRIEF DESCRIPTION OF DRAWINGS

[0025] Figure 1 It is a multi-user random modulation flowchart;

[0026] Figure 2 It is a multi-user OAMP (MU-OAMP) detection technology flowchart;

[0027] Figure 3 It is a comparison chart of system error rates of multi-user OAMP and ordinary OAMP with the same transmission length under accurate channel estimation;

[0028] Figure 4 It is a comparison chart of system error rates of multi-user OAMP and ordinary OAMP with different transmission lengths under accurate channel estimation;

[0029] Figure 5 It is a comparison chart of system error rates of signal recovery by using unmodified LMMSE filter and modified LMMSE filter respectively under the condition that channel estimation error matrices are all IID Gaussian matrices;

[0030] Figure 6 It is a comparison chart of system error rates of signal recovery by using matched filter and modified LMMSE filter respectively under the condition that channel estimation error matrices are all IID Gaussian matrices;

[0031] Figure 7 It is a comparison chart of system error rates of signal recovery by using unmodified LMMSE filter and modified LMMSE filter respectively under the condition that channel estimation error matrices are all IID Gaussian matrices and the user transmission signal lengths are different;

[0032] Figure 8 It is a comparison chart of system error rates of signal recovery by using matched filter and modified LMMSE filter respectively under the condition that channel estimation error matrices are all IID Gaussian matrices and the user transmission signal lengths are different;

[0033] Figure 9 It is a comparison chart of system error rates of OFDMA technology and multi-user OAMP under the condition of accurate channel estimation and user static state;

[0034] Figure 10 It is a schematic diagram of the uplink multi-user random modulation and OAMP detection device provided by the embodiment of the application. DETAILED DESCRIPTION

[0035] The specific embodiments of the present invention are further described in detail below with reference to the accompanying drawings.

[0036] Example 1

[0037] like Figure 1 and Figure 2 As shown, the present invention provides an uplink multi-user random modulation and OAMP detection method. Before detection, it is necessary to perform independent random modulation on each user at the transmitting end and then transmit a signal. The sum of the signals of each user is received from the receiving end. Here, an uplink communication system with random modulation containing U users is considered:

[0038]

[0039] in It is the sum of the signals received by the base station from U users after being transmitted through their respective channels and interfered by noise; is the IID Gaussian channel noise vector, is the transmitted signal vector from the u-th user, represents the Multi-Input-Multi-Output (MIMO) channel matrix corresponding to the u-th user, where the dimension L is the number of subcarriers. Signal x u is a randomly modulated signal, expressed as

[0040] x u =Ξ u s u ,

[0041] in The dimension is L×N u A random unitary matrix, The dimension is N u The unmodulated input signal vector of ×1. The input signal vector of U users s u and the random unitary matrix Ξ u They must be independent of each other and generated in an independent manner. It should be noted that for the u-th user, when the equivalent channel matrix A u =H u Ξ u Only when it belongs to the universal matrix, can x u =Ξ u s u called pairs u The universal matrix needs to meet the following conditions:

[0042] 1)A u is spectrally convergent and has a bounded spectral norm, satisfying

[0043] 2) For any fixed constant and constant ∈ > 0, there exists

[0044]

[0045] In theory, the unitary matrix completely randomly selected from the entire feasible region is strictly optimal; however, experimental results show that most randomly generated unitary matrices are good modulation matrices, so a random unitary matrix that is easy to implement can be selected, such as:

[0046] Matrix Ξ u The QR decomposition matrix of the Haar matrix or the IID Gaussian matrix can be selected;

[0047] Matrix Ξ u It can be constructed by multiplying a random permutation matrix Π u and a unitary matrix T u , that is, Ξ u = Π u T u , where T u The Discrete Fourier Transform (DFT) matrix, Hadamard- Walsh transform matrix, discrete sine transform matrix, and normalized IID Gaussian transform matrix can be selected.

[0048] Matrix Ξ u It can also be constructed using a multi-layer row permutation unitary matrix, denoted as Ξ u = Π u,1 T u,1 Π u,2 T u,2 … Π u, Q T u,Q , where each pair {Π u,q , T u,q} is randomly selected.

[0049] MU-OAMP algorithm: linear estimation and nonlinear estimation are performed according to the channel information of the received signal transmitted by each user; the linear estimation and nonlinear estimation are continuously iterated until convergence or the maximum number of iterations is reached, and the linear estimation quantity and variance of the last iteration are obtained;

[0050] For the convenience of description, only the u-th user is described here, and the signal estimation method of other users is completely the same.

[0051] The MU-OAMP algorithm proposed in this application is applicable to signal detection when the channel CSI is not perfect. Let the channel matrix obtained by channel estimation be Then the equivalent channel matrix estimated is If the estimated channel is biased from the real channel, and let the channel estimation error matrix be The equivalent channel estimation error matrix is Let As the initial value of the estimation of the u-th user, the multi-user OAMP algorithm proposed in the present application is iterated as follows:

[0052]

[0053] According to the obtained linear estimation and variance, a hard decision is made to obtain the final decoding result.

[0054] For normalized QPSK signals Let Then The decision result of can be obtained by the following formula

[0055]

[0056] wherein represents the i-th element of the decoding decision result,

[0057]

[0058] is the i-th element of the linear estimation in the last iteration, T is the maximum number of iterations or the number of iterations when the iteration converges.

[0059] Since d k,q is too large or too small, MATLAB may have a numerical overflow problem when calculating Therefore, in the actual program, the upper and lower bounds of d k,q are 100 and -100, respectively, and let We have

[0060]

[0061] Embodiment 2

[0062] This embodiment is an embodiment when the estimated channel CSI has an error from the real channel CSI, and the channel estimation error matrix is an IID Gaussian matrix. The implementation case is: the communication channel is a time-varying multipath channel, and the CSI obtained by each user through channel estimation is different, which may have a significant difference; the CSI obtained by each user through channel estimation may have errors of different degrees from the real CSI of each user, and the channel estimation error matrix is an IID Gaussian matrix; and the length of the transmission signal of each user may be different.

[0063] Assume that the base station receives the transmitted normalized QPSK signals of 2 users (i.e., U = 2), and the lengths of the transmitted signals of user 1 and user 2 are N1 and N2, respectively; let the transmitted signals of user 1 and user 2 be and Then, we have

[0064] s1~ i.i.d. S1,

[0065] s2~ i.i.d. S2, wherein

[0066]

[0067] u = 1, 2,

[0068] I u represents the component in phase with the reference carrier, and Q u represents the component orthogonal to the reference carrier. The uniform distribution function randi provided by MATLAB can be used to generate them. S u is the constellation diagram of the QPSK signal, and Unif(·) represents the uniform distribution;

[0069] For the sake of brevity, assume that both user 1 and user 2 use the TDL-A channel in the 5G 3GPP standard (but in fact the channel categories of the two users can be completely different).

[0070] Set the number of subcarriers as L, the subcarrier spacing as Δf, and the carrier frequency as f c , so that the sampling frequency f s = LΔf and the sampling interval T s = 1 / f s can be calculated; set the total number of sampling symbols as s pa , and the oversampling rate as s ps ;

[0071] Assume that the relative motion speeds between user 1 and user 2 and the base station are v1 and v2, respectively, so that the maximum Doppler shifts of user 1 and user 2 can be calculated as v max,1 = f c v1 / c and v max,2 = f c v2 / c, and c = 3e 8 m / s represents the speed of light.

[0072] Assume that the power gains of user 1 and user 2 due to channel slow fading are p1 and p2, respectively;

[0073] The delay spreads of user 1 and user 2 are d1 and d2, respectively.

[0074] The channel matrix of the u-th user (u = 1, 2) is generated according to the following channel model​​

[0075]

[0076] wherein

[0077]

[0078] l, p are indices, τ k,u is the delay of the kth path of the u th user, v k,u is the Doppler shift of the kth path of the u th user,

[0079] Detailed description:

[0080] The parameters involved in the above channel model are:

[0081]

[0082] τ k,u = τ k,0 · d u ,

[0083]

[0084]

[0085] wherein the channel impulse response of the u th user is:

[0086]

[0087] After the transmitted signal passes through a transmitting filter h rc (t), the response of the entire equivalent channel is:

[0088]

[0089] wherein K is the number of channel paths, v k , τ k represent the Doppler shift and the delay of the kth path, respectively; is the time-varying complex amplitude gain of the kth path, φ k,u (t) is a random phase; h rc (t) is a Raised Cosine Filter (RCF):

[0090]

[0091] wherein β ∈ [0, 1] is the roll-off factor, T s is the symbol interval / sampling interval. It is assumed that the signal passes through the filter and is then oversampled, and the total number of sampling symbols is s pa , and the oversampling rate is sps , then the filter order is s pa ·s ps +1,

[0092] The actual amplitude gain of the u-th user |h k,u (t)| is proportional to the normalized amplitude gain |h k,0 (t)|, in order to distinguish the channels of different users, we give different power gains p u to the signals received at the receiving end of different users to simulate the difference caused by slow fading due to different distances, that is,

[0093]

[0094] u = 1, …, U.

[0095] The actual delay of the u-th user τ k,u is also proportional to the normalized delay τ k,0 . Assuming that the delay spread of the channel of the u-th user is d u , then the actual delay can be calculated by the following conversion formula

[0096] τ k,u = τ k,0 · d u ,

[0097] u = 1, …, U.

[0098] The normalized delay τ k,0 (k = 1, …, K) and the normalized power gain |h k,0 (t)| 2 of the TDL channel can be found in the 5G 3GPP standard, and the 5G toolbox provided by MATLAB can also directly generate these channel parameters.

[0099] If the relative motion speed between the u-th user and the base station is v u , then the maximum Doppler shift is

[0100]

[0101] where f c is the carrier frequency and c is the speed of light. Then under the premise that v max,u is known, the Doppler shift v k,u of all K paths can be further obtained according to the 5G 3GPP standard and the 5G toolbox provided by MATLAB.

[0102] In the 5G 3GPP standard, the TDL-A channel has 23 distinguishable paths, so K = 23;

[0103] |hk,0 (t)|, τ k,u and v k,u can be known when the delay spread d u and the maximum Doppler shift v max,u are known. The power gain p u and the random phase are artificially added after that. h k,u (t) can be obtained.

[0104] h rc [s] represents the impulse response of the raised cosine filter, which can be directly generated by the MATLAB rcode design function under the condition that the total number of sampling symbols s pa and the oversampling rate s ps are known. It can be noted that the value of h is equal to the value of the MATLAB generated discrete raised cosine filter impulse response at the th position.

[0105] Assume that the channel estimation error matrix ε u is an IID Gaussian matrix, where the variance of each item is that is,

[0106] ε u ~ N(0, σ i.i.d. 2), ε u ,

[0107]

[0108] ε u can be generated by MATLAB functions such as randn. Finally, the estimated channel matrix is obtained as

[0109] Two independent random modulation matrices and are generated respectively, and the real equivalent channel matrix and the estimated equivalent channel matrix of the u-th user are A u = H u Ξ u ,

[0110] Assume that the signal-to-noise ratio at the receiving end is snr, then the noise power is σ 2 = 1 / snr, and the signal received by the base station is

[0111]

[0112] The specific process of the MU-OAMP algorithm in this embodiment includes:

[0113] Since the iteration method is similar for all users, only the u-th user is taken as an example:

[0114] Initialize iteration number t = 1, maximum iteration number T; nonlinear estimator Nonlinear estimation variance

[0115] Linear estimation (LE):

[0116] Calculate intermediate variable z (t)

[0117]

[0118] Where ∑ (t) Can be various filter matrices, such as

[0119] 1) Matched filter (Matched filter):

[0120] ∑ (t) = I L ,

[0121] 2) Undevised LMMSE filter (Undevised LMMSE filter):

[0122]

[0123] 3) Devised LMMSE filter (Devised LMMSE filter):

[0124]

[0125] Where

[0126]

[0127] Calculate the orthogonal parameter

[0128]

[0129] Calculate the linear estimator

[0130]

[0131] Calculate the linear estimation variance

[0132]

[0133] In particular, when

[0134] 1) ∑ (t) is a matched filter

[0135]

[0136] 2)∑ (t) is the LMMSE filter

[0137]

[0138] Non-linear estimation (NLE):

[0139] Given the prior distribution s u ~ i.i.d. S i , compute the MMSE posterior estimate

[0140]

[0141] Compute the MMSE posterior estimation variance

[0142]

[0143] Compute the NLE estimation variance

[0144]

[0145] Compute the NLE estimate

[0146]

[0147] To demonstrate the practicality of the present application in the following, the bit error rate is calculated after each iteration:

[0148] Compute the bit error rate BER of the u-th user at the current t-th iteration u (t)

[0149]

[0150] where is the indicator function, taking 1 when the condition in (·) is satisfied, and 0 otherwise; s u,i denotes the i-th element of s u (i = 1,..., N u ), denotes the i-th element of the decoding decision result at the current t-th iteration s y is taken from the constellation i.e., there is and s u,i takes the value x k with probability q k = P(s u,i = x k ).

[0151] In particular, for normalized QPSK signals Record Then The decision result of can be obtained by the following formula

[0152]

[0153] Wherein

[0154]

[0155] The ith element of the linear estimator. Since the value overflow problem may occur in the calculation of when d k,j is too large or too small, we give the upper and lower bounds of d k,j in the actual program as 100 and-100, respectively, and record

[0156] We have

[0157] Update the iteration number t=t+1.

[0158] Repeat the iteration process until t=T+1.

[0159] According to the obtained linear estimator and variance, hard decision is made to obtain the final decoding result.

[0160] In addition, in order to ensure the practicability of the application, the embodiment also provides the state evolution (SE) of the algorithm, specifically including:

[0161] The state evolution (SE) of the algorithm is

[0162] The state evolution (SE) of the algorithm is

[0163]

[0164]

[0165] Wherein:

[0166] 1. and are the linear end and nonlinear end estimators in the tth iteration

[0167] 2. (Note that for an N×N matrix U N×N , the symbol

[0168] ​ denotes the empirical mean of the trace of the N x N matrix U N×N , while for a one-dimensional N x 1 vector r N×1 , the notation denotes the empirical mean of the N elements of the vector r N×1 ; the power of the signal s u

[0169] 3. Let then

[0170] and are the error vectors of the estimates of the linear and nonlinear ends and the true values in the tth iteration, respectively, and

[0171] is the error SE;

[0172] 4.

[0173] 5. is a deterministic, generally non-separable function. It is noted that these functions need to satisfy the divergence-free property:

[0174]

[0175] It is noted that, denotes the Jacobian matrix

[0176]

[0177] In particular, we can choose the function:

[0178]

[0179] where is the Minimum Mean Square Error (MMSE) posteriori estimation function

[0180]

[0181] where

[0182] 6.∑ (t) In algorithm analysis, it can be an almost arbitrary matrix, but its correct construction is crucial for achieving optimal performance. The simplest choice is the matched filter: ​

[0183] ∑ (t) = I L ,

[0184] The OAMP algorithm iteration formula is

[0185]

[0186] SE is

[0187]

[0188] From the perspective of computational complexity, this choice is particularly attractive because it makes the above OAMP formula iteration only involve low complexity matrix and vector multiplication. However, this simple choice can lead to poor algorithm performance.

[0189] Another commonly used choice is the linear minimum mean square error (LMMSE) estimator:

[0190]

[0191] The OAMP algorithm iteration formula is

[0192]

[0193] SE is

[0194]

[0195]

[0196] However, when the CSI is imperfect, the performance of the ordinary LMMSE estimator deteriorates as the CSI error increases, so if the best estimation performance is needed in the case of imperfect CSI, the ordinary LMMSE estimator needs to be modified additionally. In particular, we can consider the case where the error matrix is a complex Gaussian matrix, i.e.

[0197]

[0198] where is an arbitrary fixed positive semi-definite matrix. Here, we use the notation to represent that X is a random matrix with a complex Gaussian distribution, whose row and column covariance matrices are A and B, respectively. That is, where the entries in X are IID standard normal distribution random variables. It should be noted that in the analysis we assume that H u and the spectral radius of the matrix is bounded with high probability. Formally, we assume that the normalized scale

[0199]

[0200] where ||·||2denotes the largest singular value of a matrix. Note that if for some constant c > 0, we can always scale the observation vector by a factor of -c Hence, we assume that It is also for this reason that the scale of the row / column covariance matrix in u

[0201] Additionally, it can be shown that under the Gaussian channel estimation error model

[0202]

[0203] It can be noted that under this error model the OAMP algorithm has:

[0204] 1) the MMSE posterior estimation function is

[0205]

[0206] 2) the nonlinear estimation function is

[0207]

[0208] 3) the SE can be simplified to

[0209]

[0210]

[0211] where

[0212]

[0213] At this point, we modify (t) in the LMMSE estimator to:

[0214]

[0215] The OAMP algorithm iteration with the modified LMMSE estimator is then given by

[0216]

[0217] The SE can be further simplified to:

[0218] ​​

[0219] Further, the error matrix can be made At this time, the error matrix is an IID Gaussian matrix, and the variance is That is,

[0220] ε u ~ i.i.d. ε u ,

[0221] wherein

[0222]

[0223] At this time

[0224]

[0225] According to Figure 3 Fig. 6 shows the system bit error rate (BER) changes with the number of iterations when the ordinary OAMP algorithm and the multi-user OAMP algorithm are used for signal recovery in the uplink communication scenario with the channel model being TDL-A channel, the number of users being U = 2, the channel estimation being accurate (σ1 = σ2 = 0), the transmission signal being QPSK signal, the number of subcarriers being L = 4096, the lengths of the transmission signals of user 1 and user 2 being N1 = 4096 and N2 = 4096 respectively, the signal-to-noise ratio being snr = 13 dB, the moving speed of user 1 and user 2 relative to the base station being 100 km / h, the delay spread being 10 ns, and the power gain of user 1 and user 2 being 0 dB and -3 dB respectively. The dotted line and the solid line are the bit error rate curves of the ordinary OAMP algorithm and the multi-user OAMP algorithm respectively, the red line and the blue line are the bit error rate curves of user 1 and user 2 respectively, and the red circle and the blue circle represent the SE of user 1 and user 2 respectively when the multi-user OAMP algorithm is used for signal recovery.

[0226] According to Figure 4It is shown that: in the channel model is TDL-A channel, the number of users is U = 2, the channel estimation is accurate (σ1 = σ2 = 0), the transmit signal is QPSK signal, the number of subcarriers is L = 4096, the transmit signal length of user 1 and user 2 is N1 = 4096, N2 = 2048 respectively, the signal-to-noise ratio is snr = 13dB; the relative speed of user 1 and user 2 to the base station is 100km / h, the delay spread is 10ns; the power gain of user 1 and user 2 is 0dB, -3dB respectively. The change of system bit error rate (Bit Error Rate, BER) with iteration number when the ordinary OAMP algorithm and the multi-user OAMP algorithm are used for signal recovery in the uplink communication scene respectively. The dotted line and the solid line are the bit error rate curves of the ordinary OAMP algorithm and the multi-user OAMP algorithm respectively, the red line and the blue line are the bit error rate curves of user 1 and user 2 respectively, and the red and blue circles represent the SE of user 1 and user 2 when using the multi-user OAMP algorithm for signal recovery.

[0227] According to Figure 5 It is shown that: in the channel model is TDL-A channel, the number of users is U = 2, the channel estimation error matrix is IID Gaussian matrix and the error parameters are σ1 = 0.135, σ2 = 0.096 respectively, the transmit signal is QPSK signal, the number of subcarriers is L = 4096, the transmit signal length of user 1 and user 2 is N1 = 4096, N2 = 4096 respectively, the signal-to-noise ratio is snr = 18dB; the relative speed of user 1 and user 2 to the base station is 100km / h, the delay spread is 10ns; the power gain of user 1 and user 2 is 0dB, -3dB respectively. The change of system bit error rate (Bit Error Rate, BER) with iteration number when the undevised LMMSE filter and the devised LMMSE filter are used for signal recovery in the uplink communication scene respectively. The dotted line and the solid line are the bit error rate curves of the undevised LMMSE filter and the devised LMMSE filter respectively, the red line and the blue line are the bit error rate curves of user 1 and user 2 respectively, and the red and blue circles represent the SE of user 1 and user 2 when using the devised LMMSE filter for signal recovery.

[0228] According to Figure 6It is shown that: in the uplink communication scenario of TDL-A channel, U = 2, the channel estimation error matrix is an IID Gaussian matrix, the error parameters are respectively σ1 = 0.135, σ2 = 0.096, the transmitted signal is QPSK signal, the subcarrier number is L = 4096, the transmitted signal length of user 1 and user 2 is respectively N1 = 4096, N2 = 4096, the signal-to-noise ratio is snr = 18dB; the relative motion speed of user 1 and user 2 to the base station is 100km / h, the delay spread is 10ns; the power gain of user 1 and user 2 is respectively 0dB, -3dB, the system bit error rate (Bit Error Rate, BER) changes with the number of iterations when the matched filter (Matched filter) and the modified LMMSE filter (Devised LMMSE filter) are used for signal recovery respectively. The dotted line and the solid line are the bit error rate curves of the matched filter and the modified LMMSE filter respectively, the red line and the blue line are the bit error rate curves of user 1 and user 2 respectively, and the red and blue circles represent the SE of user 1 and user 2 when the modified LMMSE filter is used for signal recovery.

[0229] According to Figure 7 It is shown that: in the uplink communication scenario of TDL-A channel, U = 2, the channel estimation error matrix is an IID Gaussian matrix, the error parameters are respectively σ1 = 0.135, σ2 = 0.096, the transmitted signal is QPSK signal, the subcarrier number is L = 4096, the transmitted signal length of user 1 and user 2 is respectively N1 = 4096, N2 = 4096, the signal-to-noise ratio is snr = 18dB; the relative motion speed of user 1 and user 2 to the base station is 100km / h, the delay spread is 10ns; the power gain of user 1 and user 2 is respectively 0dB, -6dB, the system bit error rate (Bit Error Rate, BER) changes with the number of iterations when the undevised LMMSE filter (Undevised LMMSE filter) and the modified LMMSE filter (Devised LMMSE filter) are used for signal recovery respectively. The dotted line and the solid line are the bit error rate curves of the matched filter and the modified LMMSE filter respectively, the red line and the blue line are the bit error rate curves of user 1 and user 2 respectively, and the red and blue circles represent the SE of user 1 and user 2 when the modified LMMSE filter is used for signal recovery.

[0230] According to Figure 8It is shown that: in the channel model is TDL-A channel, the number of users is U=2, the channel estimation error matrix is an IID Gaussian matrix and the error parameters are respectively σ1=0.250, σ2=0.125, the transmitting signal is a QPSK signal, the number of subcarriers is L=4096, the transmitting signal length of user 1 and user 2 is respectively N1=4096, N2=2048, the signal-to-noise ratio is snr=18dB; the relative motion speed of user 1 and user 2 to the base station is 100km / h, and the delay spread is 10ns; the power gain of user 1 and user 2 is respectively 0dB, -6dB; the bit error rate (BER) of the system changes with the number of iterations when the matched filter and the devised LMMSE filter are used for signal recovery in the uplink communication scene. The dotted line and the solid line are the bit error rate curves of the matched filter and the devised LMMSE filter respectively, the red line and the blue line are the bit error rate curves of user 1 and user 2 respectively, and the red and blue circles represent the SE of user 1 and user 2 when the devised LMMSE filter is used for signal recovery.

[0231] Embodiment 3

[0232] Specifically, the embodiment of the application also provides a SISO time-varying (including static) multipath channel model used when simulating the transceiving process of the sending end and the receiving end:

[0233] We consider the Time Delay Line (TDL) channel in 5G 3GPP, and the channel model is the same as in embodiment 2. l=0,1,…,L-1;

[0234] The application provides an OAMP algorithm for signal detection in an uplink multi-user communication scene, which can still have good detection performance when there is a significant difference between the CSI obtained by each user through channel estimation and there is an error between the CSI obtained by each user through channel estimation and the real CSI of each user.

[0235] This embodiment is an embodiment of multi-user OAMP detection with signal extension and random modulation in a static channel (i.e., the channel Doppler shift is 0) and perfect channel estimation, and the implementation of the embodiment is as follows: the communication channel is a static multipath channel, the CSI obtained by each user through channel estimation is different and may have a significant difference, but the estimated channel is exactly the same as the real channel.

[0236] The specific steps of multi-user OAMP include:

[0237] S1, initialization:

[0238] S11, for the sake of brevity, it is assumed that both user 1 and user 2 adopt TDL-A channel in 5G 3GPP standard (but in fact the channel categories of the two users can be completely different).

[0239] S111, set the number of subcarriers as L, the subcarrier spacing as Δf, and the carrier frequency as f c , thus the sampling frequency can be calculated as s = LΔf, and the sampling interval T s = 1 / f s ; set the total number of sampling symbols as s pa , and the oversampling rate as s ps .

[0240] S112, it is assumed that there is no relative motion between user 1 and user 2 and the base station, i.e., the Doppler shifts are both 0.

[0241] S113, it is assumed that the power gains of user 1 and user 2 due to channel slow fading are p1 and p2, respectively.

[0242] S114, the delay spreads of user 1 and user 2 are d1 and d2, respectively.

[0243] S12,

[0244] S121, the channel matrix of the u-th user (u = 1, 2) is generated according to the following channel model

[0245]

[0246] wherein

[0247]

[0248] The parameters involved in the above channel model are described in Embodiment 1.

[0249] S13, it is assumed that the base station receives the transmitted QPSK signals of the two users (i.e., U = 2), and the lengths of the transmitted signals of user 1 and user 2 are N1 and N2, respectively; let the transmitted signals of user 1 and user 2 be and , then

[0250] s1~<0, s2~<0, i.i.d. S1,

[0251] s2~<0, s2~<0, i.i.d. S2,

[0252] wherein

[0253]

[0254] u = 1, 2,

[0255] I u ,Q u The uniform distribution function randi in MATLAB can be used to generate the signal amplitude. The signal amplitude in this example is not The reason is that the signal needs to be zero-padded and extended to make the total length of the signal consistent with the number of columns L of the channel matrix, i.e.

[0256]

[0257] In order to ensure that the average power of the signal is still 1, the amplitude of the signal needs to be adjusted.

[0258] S14, fast Fourier transform algorithm is used to perform fast Fourier transform on the two users respectively (equivalent to multiplying a fast Fourier transform matrix F), and then random interleaving between users is performed (equivalent to multiplying a random permutation matrix Π1 or Π2), and finally inverse Fourier transform is performed (equivalent to multiplying a fast inverse Fourier transform matrix F H ). The random modulation signals belonging to the two users obtained after such transformation are sent into static channels H1 and H2 respectively. Assuming that the signal-to-noise ratio at the receiving end is snr, the noise power is σ 2 = 1 / snr, and the signal received by the base station is

[0259]

[0260] The equivalent channel matrix obtained by performing Fourier transform on the received signal at the receiving end is A1 = A1Π1F and A2 = A2Π2F respectively. Where Λ1 = FH1F H , Λ2 = FH2F H are two diagonal matrices, i.e.

[0261]

[0262] S2, MU-OAMP algorithm

[0263] Since the MU-OAMP algorithm used here is exactly the same as that in Example 1, it will not be described again.

[0264] In addition to the channel model and the principle of the MU-OAMP algorithm mentioned in the technical principle of Example 1, the technical principle according to the above-mentioned OAMP algorithm for signal detection in an uplink multi-user static communication scenario includes:

[0265] In a static communication scenario, the channel is a static multipath channel (without Doppler shift). After adding a cyclic prefix (CP) at the transmitting end, the static multipath channel matrix can be diagonalized by multiplying a Fourier transform matrix on the left and an inverse Fourier transform matrix on the right.

[0266] In static scenario, we can let the random modulation matrix Ξ u be a square matrix as follows

[0267] Ξ u = F H Π u F,

[0268] where, is an interleaver. After Fourier transform of the received data, the equivalent channel becomes

[0269] A u = Λ u Π u F.

[0270] Therefore, by using the random modulation matrix above, the time complexity of the multi-user OAMP algorithm with LMMSE filter can be greatly reduced. Because the matrix inversion operation is performed on the diagonal matrix, the time complexity is reduced from to Moreover, due to the random interleaving FFT operation, the time complexity required for matrix and vector multiplication can be reduced from to

[0271] Before discussing the communication scenarios discussed above are dynamic scenarios with time-varying multipath channels. Among them, due to the existence of Doppler shift, the commonly used orthogonal frequency division multiple access (OFDMA) technology cannot diagonalize the channel matrix, so it cannot directly apply MMSE for signal detection. But in static scenarios, we can compare the detection performance of multi-user OAMP detection with random modulation and MMSE detection based on OFDMA technology. For the sake of simplicity, we can assume that each user has perfect CSI. At the same time, in order to ensure fairness, we should keep the amount of information carried by the signal consistent with the amount of information using OFDMA technology under the same channel size. Therefore, we extend the signal without changing the total amount of information carried by the signal, which is equivalent to zero padding in the transmitted signal, that is

[0272]

[0273] where the length (i.e. N u ) and power of s u should be consistent with s u when using OFDMA technology for signal transmission.

[0274] It needs to be particularly pointed out that in the dual-user static scenario, the principle of the OFDMA technology is as follows:

[0275]

[0276] wherein

[0277]

[0278] As Figure 9 shown in the figure, in an uplink communication scenario where a channel model is a TDL-A channel, a user number is U = 2, channel estimation is accurate (σ1 = σ2 = 0), a transmission signal is a QPSK signal, a subcarrier number is L = 4096, a transmission signal length of user 1 and user 2 is N1 = N2 = 2048, user 1 and user 2 are static relative to a base station, a delay spread is 10 ns, and a power gain of user 1 and user 2 is 0 dB and -3 dB respectively, a system bit error rate (BER) changes with a signal-to-noise ratio (SNR) in the case of (i) multi-user OAMP detection after signal expansion and with random modulation and (ii) MMSE detection using OFDMA technology for signal modulation. The solid line and the dashed line are respectively a bit error rate curve of (i) multi-user OAMP detection after signal expansion and with random modulation and (ii) MMSE detection using OFDMA technology for signal modulation, the red line and the blue line are respectively a bit error rate curve of user 1 and user 2, and the red circle and the blue circle represent SE of user 1 and user 2 respectively in the case of using multi-user OAMP detection after signal expansion and with random modulation.

[0279] In summary, the method of the present application can obtain a detection performance obviously better than MMSE detection based on the OFDMA technology in a static channel.

[0280] Corresponding to the foregoing embodiment of the uplink multi-user random modulation and OAMP detection method, the present application further provides an embodiment of an uplink multi-user random modulation and OAMP detection device.

[0281] Referring to Figure 10 , the embodiment of the present application provides an uplink multi-user random modulation and OAMP detection device, which comprises a memory and one or more processors, the memory stores executable codes, and the processor executes the executable codes to implement the uplink multi-user random modulation and OAMP detection method in the foregoing embodiment.

[0282] The embodiment of the uplink multi-user random modulation and OAMP detection device provided by the present application can be applied to any data processing device, which can be a device or apparatus such as a computer. The device embodiment can be realized by software, hardware or a combination of software and hardware. Taking software realization as an example, as a logical device, it is formed by reading the corresponding computer program instructions in the non-volatile memory into the memory for running by the processor of the data processing device. From the hardware level, as shown in Figure 10 Fig. 1 is a hardware structure diagram of the data processing device in which the uplink multi-user random modulation and OAMP detection device provided by the present application is located, in addition to the processor, memory, network interface and non-volatile memory shown in Figure 10 Fig. 1, the data processing device in which the device is located in the embodiment can also include other hardware according to the actual function of the data processing device, which will not be described here.

[0283] The implementation process of the functions and roles of each unit in the above device is specifically described in the implementation process of the corresponding steps in the above method, which will not be described here.

[0284] For the device embodiment, since it basically corresponds to the method embodiment, the related parts can be referred to the part of the method embodiment. The above described device embodiment is only schematic, and the units described as separate components can or can not be physically separated, and the components displayed as units can or can not be physical units, that is, they can be located in one place or distributed on multiple network units. According to the actual needs, part or all of the modules can be selected to achieve the purpose of the present application. Those skilled in the art can understand and implement it without creative labor.

[0285] The embodiment of the present application also provides a computer readable storage medium, which stores a program, and the program is executed by a processor to realize the uplink multi-user random modulation and OAMP detection method in the above embodiment.

[0286] The computer readable storage medium can be an internal storage unit of any of the aforementioned data processing capable devices, such as a hard disk or a memory. The computer readable storage medium can also be an external storage device of any of the aforementioned data processing capable devices, such as a plug-in hard disk, a Smart Media Card (SMC), an SD card, a Flash Card, etc. Further, the computer readable storage medium can include both an internal storage unit and an external storage device of any of the aforementioned data processing capable devices. The computer readable storage medium is used to store the computer program and other programs and data required by the aforementioned data processing capable devices, and can also be used to temporarily store data that has been output or is about to be output.

[0287] The application further provides a computer program product comprising a computer program which, when executed by a processor, implements the uplink multi-user random modulation and OAMP detection method.

[0288] Other embodiments of the application will be apparent to those skilled in the art from consideration of the specification and practice of the application disclosed herein. It is intended that the specification and examples be considered as exemplary only, with the true scope and spirit of the application being indicated by the following claims.

[0289] It is to be understood that the above general description and the following detailed description are exemplary and explanatory only and are not restrictive of the application. The application is not restricted to the precise construction and combinations of parts and steps described above and shown in the accompanying drawings, as such variations and modifications are intended to be included within the scope of the application. The scope of the application is limited only by the following claims.

Claims

1. An uplink multi-user random modulation and OAMP detection method, characterized in that: include: S1, at the transmitting end, each user is independently randomly modulated and then the signal is transmitted; S2, receiving the sum of the signals of each user from the receiving end; S3. Perform linear estimation based on the received signal and the state information of the channel matrix, the estimation error of the channel matrix and the channel noise power; Calculating orthogonal parameters based on channel state information, channel matrix estimation error, and channel noise power; The linear estimator of this iteration is calculated based on the orthogonal parameters, channel state information, estimation error of the channel matrix, channel noise power and the nonlinear estimator of the previous iteration; The linear estimation variance in this iteration is obtained based on the orthogonal parameters, channel state information, channel matrix estimation error, channel noise power and the nonlinear estimation variance in the previous iteration; S4. Perform nonlinear estimation based on the prior distribution of each user's transmitted signal, the estimator and variance obtained by linear estimation, to obtain the nonlinear estimator and variance in this iteration; The linear estimation and nonlinear estimation are continuously iterated until convergence or the maximum number of iterations is reached, and the linear estimator and variance of the last iteration are obtained; S5. Perform hard decision based on the obtained linear estimator and variance to obtain the final decoding result.

2. The uplink multi-user random modulation and OAMP detection method according to claim 1, characterized in that: The step of performing independent random modulation on each user and then transmitting the signal specifically includes: x u =Ξ u s u , And the equivalent channel matrix A u =H u Ξ u belongs to the universal matrix; The dimension is L×N u A random unitary matrix, The dimension is N u ×1 unmodulated input signal vector; Represents the multi-transmit and multi-receive channel matrix corresponding to the u-th user.

3. The uplink multi-user random modulation and OAMP detection method according to claim 1, characterized in that: The receiving of the sum of the signals of each user from the receiving end includes: in It is the sum of the signals received by the base station from U users after being transmitted through their respective channels and interfered by noise; is the IID Gaussian channel noise vector, is the transmitted signal vector from the u-th user, Represents the multi-transmit multi-receive channel matrix corresponding to the u-th user, where the dimension L is the number of subcarriers, and the signal x u It is a randomly modulated signal.

4. The uplink multi-user random modulation and OAMP detection method according to claim 1, characterized in that: The linear estimation includes: Calculate the intermediate variable z (t) where Σ (t) is the filter matrix; y is the sum of the received multi-user signals, To estimate the equivalent channel matrix, is the nonlinear estimation value of the u-th user, U is the total number of users; the estimated channel matrix is Calculate orthogonal parameters Compute linear estimators Calculate the linear estimate variance in, The dimension is N u ×1 unmodulated input signal vector.

5. The uplink multi-user random modulation and OAMP detection method according to claim 4, characterized in that: The filter matrix Σ (t) When is a matched filter, the linear estimation variance is: Among them I L represents the identity matrix of dimension L×L, where L is the number of subcarriers; the subscript u′ represents the u′th user, u′=1,…,U, and the filter matrix ∑ (t) When it is an LMMSE filter, the linear estimation variance is:

6. The uplink multi-user random modulation and OAMP detection method according to claim 4, characterized in that: The nonlinear estimation includes: Known prior distribution s u ~ i.i.d. S u , calculate the MMSE posterior estimator Calculate the MMSE posterior estimate variance Calculate the variance of nonlinear estimates Compute nonlinear estimators S24. Update the number of iterations t=t+1.

7. The uplink multi-user random modulation and OAMP detection method according to claim 1, characterized in that: When the channels at the transmitting and receiving ends are static channels, the random modulation process is to perform fast Fourier transform on the two users, then perform random interleaving independently between the users, and finally perform inverse Fourier transform to obtain the random modulated signal. At this time, the random modulation matrix Ξ u As shown below: X u =F H P u F, in, is the interleaver, and F is the Fourier transform.

8. The uplink multi-user random modulation and OAMP detection method according to claim 6, characterized in that: The obtained linear estimator and variance are used for hard decision making, specifically including: For the normalized QPSK signal remember So The judgment result is obtained by the following formula in Represents the i-th element of the decoding decision result, is the linear estimator in the last iteration The i-th element of ; Give d k,q The upper and lower bounds are set to 100 and -100 respectively.

9. An uplink multi-user random modulation and OAMP detection device, comprising a memory and one or more processors, wherein the memory stores executable code, characterized in that: When the processor executes the executable code, an uplink multi-user random modulation and OAMP detection method according to any one of claims 1 to 8 is implemented.

10. A computer-readable storage medium having a program stored thereon, characterized in that: When the program is executed by a processor, an uplink multi-user random modulation and OAMP detection method according to any one of claims 1 to 8 is implemented.

Images