A nonlinear precoding method and system for MIMO-NOMA system
By introducing dynamic channel correlation threshold and channel gain difference in MIMO-NOMA system for user clustering, and using alternating direction multiplier method for weighted equivalent channels and precoding joint optimization, the system calculation complexity and transmission reliability are solved. Problem, improving bit error rate performance and spectrum efficiency.
Patent Information
- Application Number
- CN202111131712.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-09-26
- Publication Date
- 2024-11-22
- Estimated Expiration
- 2041-09-26
AI Technical Summary
When the number of users increases, the MIMO-NOMA system faces high computational complexity for user clustering, power distribution and beamforming, and low transmission reliability. The existing linear precoding scheme cannot effectively improve the overall performance of the system.
Using nonlinear precoding method, users clustering by introducing dynamic channel correlation thresholds, combining channel gain differences for power allocation, and minimizing intercluster interference as the optimization goal, weighted equivalent channels and Precoding joint optimization.
It improves the system's bit error rate performance and transmission reliability, reduces the bit error rate, enhances in-cluster coverage and inter-cluster interference control, and achieves higher spectrum efficiency.
Smart Images

Figure CN113973204B_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of beamforming in wireless communication systems, and in particular relates to a nonlinear precoding method and system for a MIMO-NOMA system. Background Art
[0002] In the face of the future 5G massive traffic access and ultra-large capacity requirements, especially when spectrum resources are scarce, how to significantly improve spectrum efficiency and transmission capacity to meet the requirements of large traffic, huge connections and high reliability in all-scenario applications of future wireless communication systems is a key issue that wireless communication research institutes must urgently solve. In the process of mobile communication technology development, the pursuit of spectrum utilization has been ongoing. 5G NOMA technology adds a dimension to OFDM - the power domain, using the different path losses of each user to achieve multi-user multiplexing, thereby improving signal gain and providing higher spectrum efficiency than traditional orthogonal multiple access, so that all mobile devices in the same cell coverage area can obtain the maximum accessible bandwidth, which can solve the network challenges caused by connections.
[0003] The combination of NOMA technology and MIMO technology can integrate the advantages of these two technologies, and can simultaneously explore the degrees of freedom in the power domain and the spatial domain, significantly increase the system throughput, improve the system's peak rate and spectrum efficiency, and effectively cope with the explosive growth of traffic. The use of MIMO-NOMA systems has also become an inevitable trend. However, with the increase in the number of users in the MIMO-NOMA system, various contradictions have become more prominent, such as the contradiction between the transmission efficiency and the propagation of the accumulated error of SIC reception when users are clustered, and the contradiction between the enhancement of intra-cluster beam coverage and the suppression of inter-cluster interference. Therefore, precoding technology is required to align the final transmission signal with the user cluster. Since the user signals of the same cluster in the NOMA system are superimposed and then share the same beam, intra-cluster interference and inter-cluster interference are inevitable. Therefore, before sending the signal, it is necessary to perform user clustering and power allocation based on channel information such as the channel gain difference and correlation of each user, and then the transmitter uses precoding to reduce inter-cluster interference and improve spectrum efficiency. MIMO-NOMA and downlink precoding technology complement each other to better improve system performance.
[0004] The combination of MIMO technology and NOMA technology will lead to a significant increase in the number of antennas, a doubling of the number of service users, and a more dense distribution of users, which also brings challenges to precoding. It is necessary to ensure intra-cluster coverage while reducing inter-cluster and intra-cluster interference. The number of users accessing the MIMO-NOMA system will increase exponentially compared to the traditional MIMO-NOMA. How to fully exploit the spatial freedom brought by MIMO to achieve access to a large number of users while avoiding the ultra-high computational complexity of user clustering, power allocation, and beamforming joint optimization processing, as well as effectively reducing power consumption and hardware costs to achieve higher benefits, will become the difficulty in the design of MIMO-NOMA system solutions. Traditional linear precoding schemes such as the minimum mean square error algorithm and the zero-forcing algorithm only consider self-interest or altruism unilaterally, and cannot obtain ideal overall system performance. Directly optimizing the precoding matrix with the maximum sum rate as the optimization goal is an NP-hard problem. It is difficult to directly obtain the optimal solution, and the implementation complexity is too high. In addition, existing research shows that in the current research on MIMO-NOMA systems, few scholars pay attention to the system's bit error rate performance, that is, transmission reliability. Most of the research focuses on the system's spectrum utilization, system capacity, etc., but transmission reliability is the basis of communication. In order to better apply these technologies in practice, further exploration and performance optimization are urgently needed. According to preliminary simulation results, it can be seen that the current situation of low transmission reliability of MIMO-NOMA systems needs to be solved urgently. Summary of the invention
[0005] The technical problem to be solved by the present invention is to provide a nonlinear precoding method and system for a MIMO-NOMA system in view of the deficiencies in the above-mentioned prior art, so as to effectively improve the bit error rate performance of the system and improve the transmission reliability of the system.
[0006] The present invention adopts the following technical solutions:
[0007] A nonlinear precoding method for a MIMO-NOMA system comprises the following steps:
[0008] S1. Introduce a dynamic channel correlation threshold, select user pairs that meet the correlation threshold condition, and perform user clustering in combination with consideration of channel gain differences to obtain the user clustering result;
[0009] S2, using a fractional power allocation method to allocate power to the user clustering result obtained in step S1;
[0010] S3, based on the user clustering result in step S1 and the power allocation result in step S2, construct the superposition signal in the MIMO-NOMA system, and then establish the weighted equivalent channel and precoding joint optimization problem with minimizing inter-cluster interference as the optimization goal and the transmit power as the constraint condition;
[0011] S4. Use the alternating direction multiplier method to solve the weighted equivalent channel and precoding joint optimization problem established in step S3, with the input being the superimposed NOMA signal and the output being the precoded signal.
[0012] Specifically, in step S1, consider the downlink of a single-cell multi-user MIMO-NOMA system, where the base station is equipped with N t Root transmitting antenna, M users are selected from all users in the cell for clustering, and each cluster has K users, with a total of M / K clusters, recorded as C clusters; the power domain NOMA transmission structure is adopted within the cluster, and users in the same cluster share the same beam; the correlation threshold is set to δ cor , calculate the correlation between each user, and select the users with a correlation greater than δ cor If the number of user pairs is greater than the number of clusters C, select the C user pairs with the largest sum of channel correlation and channel gain difference and set them as a cluster respectively; if the number is not enough, set the correlation threshold to α cor The step size gradually decreases until the above conditions are met.
[0013] Specifically, in step S2, the signal transmission power of the nth cluster is set to P n , the user set is U, and the power P(m) allocated to the mth user is calculated as follows:
[0014]
[0015] Where, m∈U, h(m) / n(m) represents the channel quality of the mth user, h(i) / n(i), i∈U represents the channel quality of the ith user, μ FTPA is the attenuation factor.
[0016] Specifically, in step S3, the weighted equivalent channel and precoding joint optimization problem is specifically:
[0017]
[0018] xT H XP t =0
[0019] in, is the superimposed data signal of each cluster of users, represents the weighted channel of each cluster user, X is the digital baseband signal corresponding to each RF port, P t is the total transmit power, and β is the channel weighting factor matrix.
[0020] Specifically, in step S4, the alternating direction multiplier method is used to solve the joint optimization problem as follows:
[0021] Firstly, the auxiliary variable Z is added, and then the Lagrange multiplier method and the penalty function method are combined to obtain the augmented Lagrangian function. Then, the iterative update formula of the optimization variable is obtained through gradient descent and spherical optimization methods until the iteration stop condition is met.
[0022] Furthermore, by adding the auxiliary variable Z, the original optimization problem is expressed as:
[0023]
[0024] st||XZ||=0
[0025] in, is the superimposed data signal of each cluster of users, represents the weighted channel of each cluster user, X is the digital baseband signal corresponding to each RF port, and Z H is the conjugate transpose of Z, P t is the total transmit power.
[0026] Furthermore, the augmented Lagrangian function is:
[0027]
[0028] Among them, λ is the Lagrangian factor, ρ is the penalty parameter, is the superimposed data signal of each cluster of users, represents the weighted channel of each cluster user, X is the digital baseband signal corresponding to each RF port, and Z H is the conjugate transpose of Z, P t is the total transmit power.
[0029] Furthermore, the iterative update of the optimization variables is obtained through gradient descent and spherical optimization method:
[0030]
[0031] in
[0032] λ k+1 =λ k -ρ(X k+1 -Z k+1 )
[0033]
[0034]
[0035]
[0036] Among them, X k+1 , Z k+1 ,k+1 are the results of the k+1th iteration of X, Z, and λ, respectively, ij represents the channel weighting factor of the jth user in the i-th cluster, represents the channel vector of the jth user in the i-th cluster, X is the precoded signal, that is, the digital baseband signal corresponding to each RF port, represents the weighted channel of each cluster user, is the NOMA superposition signal of each cluster user, is the NOMA superposition signal sent to the c-th cluster user, m k is the result of the kth iteration of m, ρ is the penalty parameter, Z H is the conjugate transpose of Z, P t is the total transmit power, C is the number of clusters, and K is the number of users in each cluster.
[0037] Furthermore, the condition for stopping iterative optimization is the original residual r k and the dual residual s k are all less than a minimum value, and the iterative process is convergent, specifically:
[0038] r k =||X k -Z k ||≤ε
[0039] s k =||Z k -Z k-1 ||≤ζ
[0040] Where k represents the kth iteration, X k is the result of the kth iteration of X, Z k is the result of the kth iteration of Z.
[0041] Another technical solution of the present invention is a nonlinear precoding system of a MIMO-NOMA system, comprising:
[0042] The clustering module introduces a dynamic channel correlation threshold, selects user pairs that meet the correlation threshold condition, and clusters users by considering the channel gain difference to obtain the user clustering result;
[0043] The power module uses a fractional power allocation method to allocate power to the user clustering results obtained by the clustering module;
[0044] The optimization module constructs the superposition signal in the MIMO-NOMA system based on the user clustering results in the clustering module and the power allocation results in the power module. Then, the weighted equivalent channel and precoding joint optimization problem is established with the minimization of inter-cluster interference as the optimization goal and the transmit power as the constraint condition.
[0045] The precoding module uses the alternating direction multiplier method to solve the joint optimization problem of weighted equivalent channel and precoding established by the optimization module. The input is the superimposed NOMA signal and the output is the precoded signal.
[0046] Compared with the prior art, the present invention has at least the following beneficial effects:
[0047] The present invention discloses a nonlinear precoding method for a MIMO-NOMA system, which realizes better user clustering by introducing a dynamic correlation threshold; allocates power according to the channel quality of users in the same cluster, which is beneficial to correct SIC decoding of receiving-end users; controls inter-cluster interference by minimizing inter-cluster interference as an optimization goal, considers equivalent weighted channels (that is, considers the channel quality of each user in the same cluster with a certain weight factor), adopts nonlinear precoding to generate precoded signals, which is more beneficial to intra-cluster coverage, and is designed to achieve a lower bit error rate.
[0048] Furthermore, for user clustering in the NOMA system, the channels of users within the same cluster are highly correlated and the coverage within the cluster is enhanced. Therefore, when users share the same beam, the performance of each user can be guaranteed to a certain extent. In addition, the channel gains of users within the same cluster vary greatly, which can ensure that users can perform correct SIC decoding at the receiving end to a certain extent.
[0049] Furthermore, for the NOMA system, the base station sends the superimposed signals of users in the same cluster. At this time, more than half of the power on each antenna is allocated to users with poor channel conditions to facilitate the correct decoding of weak users at the receiving end; for strong users, it is necessary to decode the weak users first, then eliminate the interference of the weak user signals, and then decode their own signals. From the above SIC decoding process, it can be seen that weak users need to be decoded under the signal interference of strong users, so more power needs to be allocated to weak users for correct decoding.
[0050] Furthermore, an optimization problem is established, with the goal of minimizing inter-cluster interference and the transmit power as a constraint, and a joint optimization problem of weighted equivalent channel and precoding.
[0051] Furthermore, the alternating direction method of multipliers (ADMM) algorithm (hereinafter referred to as the ADMM algorithm) is used. This algorithm is suitable for distributed convex optimization problems. It decomposes the large optimization problem into multiple smaller sub-problems, and obtains the solution to the entire problem through the solutions to the sub-problems.
[0052] Furthermore, after adding the auxiliary variable Z using the ADMM algorithm, X in the constraint is replaced by Z, and it is transformed and added to the objective function, and a new constraint is added that the binary norm of the difference between X and Z is 0, so that the two are consistent. This step can be understood as the optimization problem approaching the optimal solution from both X and Z directions, which can improve the convergence rate and reduce the computational complexity to a certain extent.
[0053] Furthermore, according to the optimization problem obtained after adding the auxiliary variable Z, we can get the augmented Lagrangian function. The augmented Lagrangian function method adds a quadratic penalty term to the Lagrangian multiplier method, which is a combination of the Lagrangian multiplier method and the penalty function method. Compared with the above two methods, the augmented Lagrangian function method converges faster than the penalty function method and has a wider range of application than the Lagrangian multiplier method (the Lagrangian multiplier method requires the optimization problem to have a local convex structure).
[0054] Furthermore, the ADMM algorithm can be used to obtain the iterative update process of each variable. According to this process, the gradient descent method and the spherical optimization method are used to solve the iterative update formula of each variable. The implementation based on the gradient descent method is simple, and the negative gradient direction of the current position is used as the search direction, because this direction is the fastest descent direction of the current position, gradually approaching / reaching the optimal solution. Spherical optimization can map the optimization variable to a sphere, where the power of the auxiliary variable (which can also be understood as the transmitted signal) is mapped to a sphere with the transmission power as the radius. In this way, the iterative update formula of the optimization variable can be obtained.
[0055] Furthermore, for the iteration stopping condition, the commonly used stopping iteration condition in the ADMM algorithm is adopted here to ensure the feasibility of the scheme. Specifically, two aspects are considered: the original residual r k and the dual residual s k Here, the original residual is defined as the binary norm of the difference between the optimization variable X and the auxiliary variable Z to ensure the consistency of the two to a certain extent; the dual residual is defined as the binary norm of the two changes of the auxiliary variable twice, which ensures that the optimization process is convergent and stable, and prevents the iterative optimization process from stopping when there is a large change or no convergence.
[0056] In summary, this method has advantages in both the system and the optimization process. From the system level, a dynamic correlation threshold is introduced, and the correlation threshold is adjusted according to the correlation of multi-user real-time channels to perform user clustering; the power on each antenna is allocated according to the channel quality of users in the same cluster, which is conducive to the correct SIC decoding of the receiving end user; the inter-cluster interference is controlled by minimizing the inter-cluster interference as the optimization goal, and the equivalent weighted channel (that is, the channel quality of each user in the same cluster is considered with a certain weight factor) and the pre-coded signal is generated by nonlinear precoding, which is more conducive to intra-cluster coverage. From the optimization process level, the ADMM algorithm, augmented Lagrangian function method, gradient optimization, spherical optimization and other methods and related designs are used to ensure convergence and accelerate the convergence speed. In summary, the above two aspects of the design achieve a lower bit error rate in the end.
[0057] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS
[0058] Figure 1 is the MIMO-NOMA system model of the present invention;
[0059] Figure 2 It is a downlink transmission block diagram of the MIMO-NOMA system of the present invention;
[0060] Figure 3 The present invention selects 8 users from 100 users under a general Rayleigh fading channel model and measures the average bit error rate performance of each precoding algorithm;
[0061] Figure 4 The present invention selects 8 users from 300 users under a general Rayleigh fading channel model and measures the average bit error rate performance of each precoding algorithm;
[0062] Figure 5 The present invention compares the strong and weak bit error rate performance of each precoding algorithm for 8 users selected from 100 users under a general Rayleigh fading channel model;
[0063] Figure 6 The present invention selects 8 users from 100 users under the Ray-based channel model and measures the average bit error rate performance of each precoding algorithm;
[0064] Figure 7 The present invention selects 8 users from 300 users under the Ray-based channel model and measures the average bit error rate performance of each precoding algorithm;
[0065] Figure 8In the Ray-based channel model, the present invention selects 8 users from 100 users to compare the strong and weak bit error rate performances of various precoding algorithms. DETAILED DESCRIPTION
[0066] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0067] In the description of the present invention, it should be understood that the terms “include” and “comprises” indicate the presence of described features, wholes, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or collections thereof.
[0068] It should also be understood that the terms used in the present specification are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the present specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.
[0069] It should be further understood that the term "and / or" used in the present description and the appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes these combinations.
[0070] Various structural schematic diagrams of the embodiments disclosed in the present invention are shown in the accompanying drawings. These figures are not drawn to scale, and some details are magnified and some details may be omitted for the purpose of clear expression. The shapes of various regions and layers shown in the figures and the relative sizes and positional relationships therebetween are only exemplary, and may deviate in practice due to manufacturing tolerances or technical limitations, and those skilled in the art may additionally design regions / layers with different shapes, sizes, and relative positions according to actual needs.
[0071] The present invention provides a nonlinear precoding method for a MIMO-NOMA system. By introducing a dynamic correlation threshold, user clustering is performed by jointly considering channel correlation and channel gain difference. Through channel state information, equivalent weighted channels and nonlinear precoding are jointly optimized, which is more conducive to intra-cluster coverage. Inter-cluster interference is controlled by minimizing inter-cluster interference as an optimization goal to achieve a lower bit error rate. The present invention proposes a solution for jointly optimizing channel weighting coefficients and nonlinear precoding for beamforming, which can improve the transmission reliability of the system and is superior to other comparative solutions.
[0072] The present invention provides a nonlinear precoding method for a MIMO-NOMA system, comprising the following steps:
[0073] In S1, user clustering module, a dynamic channel correlation threshold is first introduced, and then user clustering is performed in combination with consideration of channel gain differences;
[0074] The present invention is based on a single-cell MIMO-NOMA system model, firstly performs user clustering, introduces a dynamic correlation threshold, and then judges and selects channel gain differences for user pairs that meet the correlation threshold condition.
[0075] See also Figure 1 and Figure 2 ,In the downlink transmission of the MIMO-NOMA system, the transmitter includes user clustering, power allocation, and precoding modules.
[0076] Considering that the base station is equipped with N t A transmitting antenna serves M single-antenna users. M users are selected from all users in the cell for clustering, and each cluster has K users. There are M / K clusters in total, recorded as C clusters. The power domain NOMA transmission structure is adopted within the cluster, and users in the same cluster share the same beam.
[0077] The Rayleigh fading channel is considered in the simulation, and the real and imaginary parts are all subject to normal distribution. In addition, the 3D-MIMO channel modeling based on the Ray-based channel model is also considered in the system. The base station deploys UPA antennas on the yz plane, and the number of vertical antennas is N. v , the antenna spacing is d1, and the number of antennas in the horizontal direction is N h , the antenna spacing is d2, and the channel contains L u Scattering paths.
[0078] For simplicity, the mechanical downtilt angle of the array antenna is not considered. φ represents the horizontal incident angle of the array antenna, θ represents the vertical incident angle of the array antenna, σ represents the standard deviation of the horizontal angle spread of the array antenna, and ξ represents the standard deviation of the vertical angle spread of the array antenna. For each scattering path, the random complex gain g can be expressed as α is the amplitude, is the phase, then the Ray-based channel vector from the kth user to the base station is expressed as:
[0079]
[0080]
[0081]
[0082] Among them, hk is a 1×M-dimensional small-scale channel vector, b(v k,l ) represents the vertical array response, a(u k,l ) represents the horizontal array response;
[0083]
[0084]
[0085] Where: λ represents the carrier wavelength, Δθ k,l represents the vertical angle expansion of the kth user's lth path, which follows the normal distribution Δθ k,l ~N(0,σ), Δθ of different antenna elements k,l Independent of each other, represents the horizontal angle expansion of the kth user's lth path, which follows a normal distribution Different antenna elements Independent of each other.
[0086] First, a larger correlation threshold is set as δ cor (close to 1), calculate the correlation between each user, and select a certain number of users with a correlation greater than δ cor The number of user pairs is usually greater than M / K to ensure that M users can be selected without duplication, and then the channel gain difference of the selected user pairs is considered. After comprehensive consideration, the M / K user pairs with the largest sum of channel correlation and channel gain difference are selected and set as a cluster; if the number is insufficient, the correlation threshold is set to α cor The step size gradually decreases until the above conditions are met.
[0087] S2, using the user clustering result obtained in step S1, and adopting the fractional power allocation method to perform power allocation;
[0088] The fractional power allocation method is used for power allocation. Let the signal transmission power of the nth cluster be P n , the user set is U, then the power allocated to the mth (m∈U) user is:
[0089]
[0090] Where h(m) / n(m) represents the channel quality of the mth (m∈U)th user. FTPA =0, the power allocated to each user in the system is the same; μ FTPA When the value of increases, the power allocated to users with poor channel quality will increase accordingly.
[0091] In the power domain NOMA system, more than half of the transmit power of each antenna is usually allocated to weak users, and the remaining power is allocated to strong users (users with poor channel conditions in the same cluster are called weak users, and users with better channel conditions are called strong users). This is beneficial for the receiving end to use serial interference cancellation and enable correct decoding.
[0092] S3, using weighted equivalent channel and precoding problem modeling, the optimization goal is to minimize inter-cluster interference, and the constraint condition is the transmission power constraint;
[0093] Taking minimizing inter-cluster interference as the optimization goal and transmit power as the constraint, the joint optimization problem is established as follows:
[0094]
[0095] xT H XP t =0
[0096] in, is the superimposed data signal of each cluster user, which is specifically expressed as
[0097]
[0098] in, represents the NOMA superposition signal sent by the base station to the cth cluster, p ij (i=1,...,C;j=1,...,K) represents the power allocated to the jth user in the i-th cluster, s ij (i=1, ..., C; j=1, ..., K) represents the data signal sent by the base station to the jth user of the i-th cluster (QPSK modulation is used here), and K represents the number of users in each cluster.
[0099] X is the digital baseband signal corresponding to each RF port, that is, the precoded transmission signal vector. Represents the weighted channel of each cluster user, expressed as:
[0100]
[0101] Where K represents the number of users in each cluster, represents the equivalent weighted channel of the cth cluster, β ij represents the channel weighting factor of the jth user in the i-th cluster, represents the channel vector of the jth user in the i-th cluster, and K represents the number of users in each cluster.
[0102] S4. Use the alternating direction multiplier method to solve the joint optimization problem of weighted equivalent channel and precoding. The input is the superimposed NOMA signal and the output is the precoded signal.
[0103] The alternating direction multiplier method suitable for solving large-scale distributed problems is used to solve the joint optimization problem, specifically:
[0104] First, the auxiliary variable Z is added, and then the Lagrange multiplier method and the penalty function method are combined to obtain the augmented Lagrangian function. Then, the iterative update formula of the optimization variable is obtained through gradient descent, spherical optimization and other methods until the iteration stopping condition is met, that is, the original error and the dual error are less than a certain minimum value or the preset number of iterations is reached.
[0105] Furthermore, by adding the auxiliary variable Z, the original optimization problem is expressed as:
[0106]
[0107] st||XZ||=0
[0108] Furthermore, the Lagrange multiplier method and the penalty function method are combined to obtain the corresponding augmented Lagrangian function:
[0109]
[0110] Among them, λ is the Lagrangian factor (dual factor) and ρ is the penalty parameter.
[0111] Further, by Definition: From the weighted channels of users in each cluster, it can be seen that there is no relationship between the channel weighting coefficients of users in each cluster, so the optimization of the channel weighting factors of each cluster can be carried out simultaneously.
[0112] The update steps of ADMM are optimized separately as follows:
[0113]
[0114]
[0115]
[0116]
[0117]
[0118] λ k+1 =λ k -ρ(X k+1 -Z k+1 )
[0119] Furthermore, the update formula of each variable obtained by gradient descent and spherical optimization method is:
[0120]
[0121] in
[0122] λ k+1 =λ k -ρ(X k+1 -Z k+1 )
[0123]
[0124]
[0125]
[0126] Among them, X k+1 , Z k+1 , k+1 are the results of the k+1th iteration of X, Z, and λ, respectively, i,1 and β i,2 are the channel weight coefficients of the strong and weak users in the i-th cluster, represents the channel vector of the jth user in the i-th cluster, X is the precoded signal, that is, the digital baseband signal corresponding to each RF port, represents the weighted channel of each cluster user, is the NOMA superposition signal of each cluster user, is the NOMA superposition signal sent to the c-th cluster user, m k is the result of the kth iteration of m, ρ is the penalty parameter, Z H is the conjugate transpose of Z, P t is the total transmit power, C is the number of clusters, and K is the number of users in each cluster.
[0127] Furthermore, in order not to introduce additional channel gain, ij Normalize:
[0128]
[0129]
[0130]
[0131] Among them, β ij represents the channel weighting factor of the jth user in the ith cluster, and K is the number of users in each cluster.
[0132] The condition for stopping iterative optimization is the original residual r k and the dual residual s k are all smaller than a very small number (the k in the upper right corner represents the kth iteration), and the iterative process is convergent, and the expression is:
[0133] rk =||X k -Z k ||≤ε
[0134] s k =||Z k -Z k-1 ||≤ζ
[0135] Among them, ε and ζ are both very small numbers, taking 10 -3 ~10 -4 .
[0136] In another embodiment of the present invention, a nonlinear precoding system for a MIMO-NOMA system is provided, which can be used to implement the nonlinear precoding method for the above-mentioned MIMO-NOMA system. Specifically, the nonlinear precoding method system for the MIMO-NOMA system includes a clustering module, a power module, an optimization module and a precoding module.
[0137] The clustering module introduces a dynamic channel correlation threshold, selects user pairs that meet the correlation threshold condition, and performs user clustering in combination with consideration of channel gain differences to obtain user clustering results;
[0138] The power module uses a fractional power allocation method to allocate power to the user clustering results obtained by the clustering module;
[0139] The optimization module constructs the superposition signal in the MIMO-NOMA system based on the user clustering results in the clustering module and the power allocation results in the power module. Then, the weighted equivalent channel and precoding joint optimization problem is established with the minimization of inter-cluster interference as the optimization goal and the transmit power as the constraint condition.
[0140] The precoding module uses the alternating direction multiplier method to solve the joint optimization problem of weighted equivalent channel and precoding established by the optimization module. The input is the superimposed NOMA signal and the output is the precoded signal.
[0141] In another embodiment of the present invention, a terminal device is provided, which includes a processor and a memory, wherein the memory is used to store a computer program, the computer program includes program instructions, and the processor is used to execute the program instructions stored in the computer storage medium. The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, which is suitable for implementing one or more instructions, and is specifically suitable for loading and executing one or more instructions to implement the corresponding method flow or corresponding function; the processor described in the embodiment of the present invention can be used for the operation of the nonlinear precoding method of the MIMO-NOMA system, including:
[0142] A dynamic channel correlation threshold is introduced, and user pairs that meet the correlation threshold condition are selected. User clustering is performed in combination with consideration of channel gain differences to obtain user clustering results. The fractional power allocation method is used to allocate power to the user clustering results. Based on the user clustering results and power allocation results, the superimposed signal in the MIMO-NOMA system is constructed, and then the weighted equivalent channel and precoding joint optimization problem is established with minimizing inter-cluster interference as the optimization goal and the transmit power as the constraint condition. The alternating direction multiplier method is used to solve the established weighted equivalent channel and precoding joint optimization problem, with the input being the superimposed NOMA signal and the output being the precoded signal.
[0143] In another embodiment of the present invention, the present invention further provides a storage medium, specifically a computer-readable storage medium (Memory), which is a memory device in a terminal device for storing programs and data. It can be understood that the computer-readable storage medium here can include both the built-in storage medium in the terminal device and the extended storage medium supported by the terminal device. The computer-readable storage medium provides a storage space, which stores the operating system of the terminal. In addition, one or more instructions suitable for being loaded and executed by the processor are also stored in the storage space, and these instructions can be one or more computer programs (including program codes). It should be noted that the computer-readable storage medium here can be a high-speed RAM memory, or a non-volatile memory (non-volatile memory), such as at least one disk storage.
[0144] One or more instructions stored in a computer-readable storage medium may be loaded and executed by a processor to implement the corresponding steps of the nonlinear precoding method for the MIMO-NOMA system in the above embodiment; one or more instructions in the computer-readable storage medium are loaded and executed by the processor as follows:
[0145] A dynamic channel correlation threshold is introduced, and user pairs that meet the correlation threshold condition are selected. User clustering is performed in combination with consideration of channel gain differences to obtain user clustering results. The fractional power allocation method is used to allocate power to the user clustering results. Based on the user clustering results and power allocation results, the superimposed signal in the MIMO-NOMA system is constructed, and then the weighted equivalent channel and precoding joint optimization problem is established with minimizing inter-cluster interference as the optimization goal and the transmit power as the constraint condition. The alternating direction multiplier method is used to solve the established weighted equivalent channel and precoding joint optimization problem, with the input being the superimposed NOMA signal and the output being the precoded signal.
[0146] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. The components of the embodiments of the present invention described and shown in the drawings here can usually be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.
[0147] The specific process of step S1 is as follows:
[0148] 1. First generate N u The channel HH of each user, the array index is initialized to be empty, and the number of clusters is N clu =0 (where N u >M)
[0149] 2. Calculate N u The correlation between the user channels and select the channels with a correlation greater than δ cor There are p pairs of users in total;
[0150] 3.while(the number of user pairs p<T)
[0151] (T is a set value, usually greater than M / K, usually set to more than twice M / K to ensure that M non-duplicate users can be selected)
[0152] δ cor With α cor is the step size decreasing update, that is, δ cor =δ cor -α cor ;
[0153] Then again select the correlation greater than δ cor of users;
[0154] Until the number of selected user pairs is greater than T, jump out of the loop
[0155] End while
[0156] 4. Add the correlation and channel gain difference of the selected user pairs and store the result in Sum;
[0157] (x, y, D, Sum have the same length, length is represented by L)
[0158] 5.For p=1:L
[0159] if cluster number N u / K, to stop the cycle;
[0160] Select the user pair corresponding to the largest value in Sum;
[0161] If the corresponding user in the user pair does not exist in the index, it is stored in the index; and N clu =N clu +1; otherwise, proceed to the next user pair judgment;
[0162] End For.
[0163] Each channel realization is random. The introduction of a dynamic correlation threshold can find a better one when the channel correlation of users is generally good, and can also select the corresponding number of users as much as possible when the channel correlation of users is generally low, so that the correlation threshold value can still be adjusted as the channel changes.
[0164] Example
[0165] Consider a single-cell MIMO-NOMA scenario, in which the nonlinear precoding scheme for the MIMO-NOMA system proposed in the present invention is used. Detailed simulation parameters are shown in Table 1.
[0166] Table 1 Simulation parameters
[0167]
[0168]
[0169] Comparison plan:
[0170] Comparative solution 1: fractional power allocation is adopted, user clustering adopts the method in step S1, and precoding adopts ZF precoding;
[0171] Comparative solution 2: fractional power allocation is adopted, user clustering adopts the method in step S1, and precoding adopts MMSE precoding.
[0172] During the simulation implementation process, in addition to the above-mentioned user clustering, power allocation, and precoding, serial interference cancellation is also required at the receiving end. That is, weak users can directly decode their own signals, while strong users need to first decode the weak user's signal, then eliminate the interference caused by the weak user's signal, and then decode to obtain their own signal.
[0173] See also Figure 3 ,Under the general Rayleigh fading channel model, the user clustering method in S1 is used to select 8 users from 100 users, and then the channel weight coefficient and precoding are jointly optimized. The receiving end uses serial interference cancellation. The results of the proposed method are compared with ZF precoding and MMSE precoding. The results are shown in Figure 3 shown.
[0174] See also Figure 4 ,Under the general Rayleigh fading channel model, the user clustering method in S1 is used to select 8 users from 300 users, and then the channel weight coefficient and precoding are jointly optimized. The receiving end uses serial interference cancellation. The results of the proposed method are compared with ZF precoding and MMSE precoding. The results are shown in Figure 4As shown in the figure, it can be seen that the user selection range becomes larger, and user pairs with higher channel correlation and larger channel gain difference are selected, which is more suitable for the user clustering principle of the NOMA system, so it has better bit error rate performance.
[0175] See also Figure 5 In the general Rayleigh fading channel model, 8 users are selected from 100 users using the user clustering method in S1, and then the channel weight coefficient and precoding are jointly optimized. The receiving end uses serial interference cancellation. The simulation observes the bit error rate performance of strong users and weak users under the three algorithms of ZF precoding and MMSE precoding. The results are shown in Figure 2. Figure 5 It can be seen that the method proposed in the present invention has a significant improvement in the bit error rate performance for strong users, and can be improved by multiple orders of magnitude.
[0176] See also Figure 6 ,Under the Ray-based channel model, the user clustering method in S1 is used to select 8 users from 100 users, and then the channel weight coefficient and precoding are jointly optimized. The receiving end uses serial interference cancellation. The results of the proposed method are compared with ZF precoding and MMSE precoding. The results are shown in Figure 6 shown.
[0177] See also Figure 7 ,Under the Ray-based channel model, the user clustering method in S1 is used to select 8 users from 300 users, and then the channel weight coefficient and precoding are jointly optimized. The receiving end uses serial interference cancellation. The results of the proposed method are compared with ZF precoding and MMSE precoding. The results are shown in Figure 7 Similarly, this case has better bit error rate performance.
[0178] See also Figure 8 In the Ray-based channel model, 8 users are selected from 100 users using the user clustering method in S1, and then the channel weight coefficient and precoding are jointly optimized. The receiving end uses serial interference cancellation. The simulation observes the bit error rate performance of strong users and weak users under the three algorithms of ZF precoding and MMSE precoding. The results are as follows: Figure 5 shown.
[0179] In summary, the nonlinear precoding method and system of the MIMO-NOMA system of the present invention also significantly improves the bit error rate performance for strong users under the Ray-based channel model, and can also be improved by multiple orders of magnitude, which also reflects the robustness of the scheme.
[0180] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.
[0181] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0182] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.
[0183] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.
[0184] The above contents are only for explaining the technical idea of the present invention and cannot be used to limit the protection scope of the present invention. Any changes made on the basis of the technical solution in accordance with the technical idea proposed by the present invention shall fall within the protection scope of the claims of the present invention.
Claims
1. A nonlinear precoding method for a MIMO-NOMA system, characterized in that: The following steps are involved: S1. Introduce a dynamic channel correlation threshold, select user pairs that meet the correlation threshold condition, and perform user clustering in combination with the consideration of channel gain differences. The user clustering result is obtained. Considering the downlink of a single-cell multi-user MIMO-NOMA system, the base station is equipped with Root transmitting antennas are selected from all users in the cell The users are clustered, and each cluster users, total Cluster, denoted as cluster; the power domain NOMA transmission structure is adopted within the cluster, and users in the same cluster share the same beam; the correlation threshold is set to , calculate the correlation between each user, and select the users with a correlation greater than The number of user pairs is greater than the number of clusters. , select the channel correlation and the channel gain difference with the largest sum Each user pair is set as a cluster; if the number is insufficient, the correlation threshold is set to The step size gradually decreases until the above conditions are met; S2, using a fractional power allocation method to allocate power to the user clustering result obtained in step S1; S3. Based on the user clustering result in step S1 and the power allocation result in step S2, the superposition signal in the MIMO-NOMA system is constructed, and then the weighted equivalent channel and precoding joint optimization problem is established with minimizing inter-cluster interference as the optimization goal and the transmit power as the constraint condition. The weighted equivalent channel and precoding joint optimization problem is specifically: in, is the superimposed data signal of each cluster of users, represents the weighted channel of each cluster user, is the digital baseband signal corresponding to each RF port, is the total transmit power, is the channel weighting factor matrix; S4. Use the alternating direction multiplier method to solve the weighted equivalent channel and precoding joint optimization problem established in step S3. The input is the superimposed NOMA signal, and the output is the precoded signal. The alternating direction multiplier method is used to solve the joint optimization problem as follows: First add the auxiliary variables , then the Lagrange multiplier method and the penalty function method are combined to obtain the augmented Lagrangian function, and then the iterative update formula of the optimization variable is obtained through gradient descent and spherical optimization methods until the iteration stop condition is met; Add auxiliary variables , the original optimization problem is expressed as: in, is the superimposed data signal of each cluster of users, represents the weighted channel of each cluster user, It is the digital baseband signal corresponding to each RF port. for The conjugate transpose of is the total transmit power; The augmented Lagrangian function is: in, is the Lagrangian factor, is the penalty parameter, is the superimposed data signal of each cluster of users, represents the weighted channels of each cluster, It is the digital baseband signal corresponding to each RF port. for The conjugate transpose of is the total transmit power; The iterative update of the optimization variables obtained by gradient descent and spherical optimization method is: in, , , They are , , No. The result of the iterations, Representative tudi The channel weighting factor for each user, Representative tudi The channel vector of each user is It is a precoded signal, that is, the digital baseband signal corresponding to each RF port. represents the weighted channel of each cluster user, is the NOMA superposition signal of each cluster user, To send to The NOMA superposition signal of cluster users, , for No. The result of the iterations, is the penalty parameter, for The conjugate transpose of is the total transmit power, is the number of clusters, and K is the number of users in each cluster.
2. The method according to claim 1, characterized in that In step S2, set The signal transmission power of the cluster is , the user set is , calculate the first The power allocated to each user ,as follows: in, , Indicates The channel quality of each user, Indicates The channel quality of each user, is the attenuation factor.
3. The method according to claim 1, characterized in that The condition for stopping iterative optimization is the original residual and the dual residual are all less than a minimum value, and the iterative process is convergent, specifically: in, Indicates Iterations, for No. The result of the iterations, for No. The result of the iteration.
4. A nonlinear precoding system for a MIMO-NOMA system, characterized in that: include: The clustering module introduces a dynamic channel correlation threshold, selects user pairs that meet the correlation threshold condition, and clusters users based on the consideration of channel gain differences. The user clustering results are obtained by considering the downlink of a single-cell multi-user MIMO-NOMA system. The base station is equipped with Root transmitting antennas are selected from all users in the cell The users are clustered, and each cluster users, total Cluster, denoted as cluster; the power domain NOMA transmission structure is adopted within the cluster, and users in the same cluster share the same beam; the correlation threshold is set to , calculate the correlation between each user, and select the users with a correlation greater than The number of user pairs is greater than the number of clusters. , select the channel correlation and the channel gain difference with the largest sum Each user pair is set as a cluster; if the number is insufficient, the correlation threshold is set to The step size gradually decreases until the above conditions are met; The power module uses a fractional power allocation method to allocate power to the user clustering results obtained by the clustering module; The optimization module constructs the superposition signal in the MIMO-NOMA system based on the user clustering results in the clustering module and the power allocation results in the power module. Then, the weighted equivalent channel and precoding joint optimization problem is established with the minimization of inter-cluster interference as the optimization goal and the transmit power as the constraint condition. The weighted equivalent channel and precoding joint optimization problem is as follows: in, is the superimposed data signal of each cluster of users, represents the weighted channel of each cluster user, is the digital baseband signal corresponding to each RF port, is the total transmit power, is the channel weighting factor matrix; The precoding module uses the alternating direction multiplier method to solve the joint optimization problem of weighted equivalent channel and precoding established by the optimization module. The input is the superimposed NOMA signal and the output is the precoded signal. The alternating direction multiplier method is used to solve the joint optimization problem as follows: First add the auxiliary variables , then the Lagrange multiplier method and the penalty function method are combined to obtain the augmented Lagrangian function, and then the iterative update formula of the optimization variable is obtained through gradient descent and spherical optimization methods until the iteration stop condition is met; Add auxiliary variables , the original optimization problem is expressed as: in, is the superimposed data signal of each cluster of users, represents the weighted channel of each cluster user, It is the digital baseband signal corresponding to each RF port. for The conjugate transpose of is the total transmit power; The augmented Lagrangian function is: in, is the Lagrangian factor, is the penalty parameter, is the superimposed data signal of each cluster of users, represents the weighted channels of each cluster, It is the digital baseband signal corresponding to each RF port. for The conjugate transpose of is the total transmit power; The iterative update of the optimization variables obtained by gradient descent and spherical optimization method is: in, , , They are , , No. The result of the iterations, Representative tudi The channel weighting factor for each user, Representative tudi The channel vector of each user is It is a precoded signal, that is, the digital baseband signal corresponding to each RF port. represents the weighted channel of each cluster user, is the NOMA superposition signal of each cluster user, To send to The NOMA superposition signal of cluster users, , for No. The result of the iterations, is the penalty parameter, for The conjugate transpose of is the total transmit power, is the number of clusters, K is the number of users in each cluster; Total transmit power, is the number of clusters, and K is the number of users in each cluster.
5. The nonlinear precoding system of the MIMO-NOMA system according to claim 4, characterized in that: In the power module, set The signal transmission power of the cluster is , the user set is , calculate the first The power allocated to each user ,as follows: in, , Indicates The channel quality of each user, Indicates The channel quality of each user, is the attenuation factor.
6. The nonlinear precoding system of the MIMO-NOMA system according to claim 4, characterized in that: In the optimization module, the superimposed data signals of each cluster of users for: in, Send to the first The NOMA superposition signal of the cluster, , Representatives assigned to tudi The power of each user, Send to the first tudi The data signal of each user, , Indicates the number of users per cluster.
7. The nonlinear precoding system of the MIMO-NOMA system according to claim 4, characterized in that: In the precoding module, the condition for stopping iterative optimization is the original residual and the dual residual are all less than a minimum value, and the iterative process is convergent, specifically: in, Indicates Iterations, for No. The result of the iterations, for No. The result of the iteration.
8. The nonlinear precoding system of the MIMO-NOMA system according to claim 4, characterized in that: Weighted channels of users in each cluster for in, represents the number of users in each cluster, Representative The equivalent weighted channel of the cluster, Representative tudi The channel weighting factor for each user, Representative tudi The channel vector of each user is Indicates the number of users per cluster.
9. The nonlinear precoding system of the MIMO-NOMA system according to claim 4, characterized in that: In the precoding module, in order not to introduce additional channel gain, Normalize: in, Indicates tudi The channel weighting factor of each user, K is the number of users in each cluster.
Citation Information
Patent Citations
User selection method of non-orthogonal multiple access system downlink under limited feedback
CN105337651A
Secondary user system capacity optimization method for cognitive MIMO network
CN107994933A
User matching and power distribution method of MIMO-NOMA (Multiple Input Multiple Output-Non-Orthogonal Multiple Output) downlink communication system
CN110932764A
Millimeter wave large-scale MIMO-NOMA system resource allocation method and system
CN113242602A
Method and system for effective wireless communication using combined modulation and coding rate based on data using software defined radio
IN201821006508A
