A flexible rate division multiple access method

Abstract

The application discloses a flexible rate division multiple access method. Different from the commonly used rate division multiple access system, the application considers flexible selection of part of users to adopt rate division, avoids the problem that rate is limited by the worst user channel caused by forced implementation of rate division on all users, and can effectively improve the effective throughput of the system. In addition, the application considers a more practical limited packet length condition, and can also effectively improve the effective throughput of the system. Based on this, the application designs a flexible rate division multiple access method for a downlink multi-user multi-antenna system.

Description

Technical Field

[0001] This invention relates to the field of wireless communication technology, and in particular to a flexible rate division multiple access method. Background Technology

[0002] With the continuous increase in data transmission volume, traditional Orthogonal Multiple Access (OMA) methods are gradually failing to meet the needs of future communication systems, leading to the proposal of various novel multiple access schemes. In multi-antenna systems, Space Division Multiple Access (SDMA) is the most commonly used. However, in overloaded systems where the number of users exceeds the number of transmission antennas, SDMA performance degrades significantly due to severe interference. Non-Orthogonal Multiple Access (NOMA) is another novel candidate scheme. NOMA allows multiple users to occupy the same orthogonal time-frequency resource block, thus accommodating more users than traditional OMA. However, in multi-antenna systems, NOMA may lead to inefficient use of spatial resources and result in extremely high complexity.

[0003] Rate-Splitting Multiple Access (RSMA) has become a widely discussed new multiple access scheme. At the downlink transmitter, all users' signals are split and reassembled into a common data stream and several private data streams. At the receiver, all users first decode the common data stream, then use Successive Interference Cancellation (SIC) to decode their respective private data streams. Finally, the common and private data streams are reassembled to obtain the individual user signals. By splitting and reassembling user signals, RSMA can flexibly adjust inter-user interference, thus achieving higher spectral efficiency and degrees of freedom than NOMA and SDMA. Furthermore, RSMA can simultaneously support a large number of users with different quality of service requirements and is more robust to different network bearers and imperfect channel information.

[0004] However, current research on RSMA schemes relies on two common assumptions: first, that all users adopt rate segmentation. In this case, all users must decode the common data stream, limiting its rate to the user with the worst channel quality, thus failing to fully realize the advantages of RSMA; second, that packet length is infinite. Since most current RSMA scheme research uses the Shannon capacity formula, which is based on the assumption of infinite packet length, this differs significantly from real-world systems. These two assumptions prevent current RSMA schemes from flexibly selecting users for rate segmentation, resulting in schemes that are not realistic and have poor performance. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a flexible rate division multiple access method, which can flexibly select some users to be rate divided, thereby effectively improving the overall effective throughput of the system.

[0006] To address the aforementioned technical problems, this invention provides a flexible rate division multiple access method, comprising the following steps:

[0007] (1) Process the target data streams of all users in sequence according to the user's selected scheme: if the user adopts rate segmentation, jump to step (2); if the user does not adopt rate segmentation, jump to step (3). After processing all users, jump to step (4).

[0008] (2) Divide the user's target data stream into a public part and a private part, the private part being the user's private data stream. Jump to step (1).

[0009] (3) This user has no public data stream; the target data stream is the private data stream. Jump to step (1).

[0010] (4) Merge the common parts of all users who use rate segmentation into a single common data stream.

[0011] (5) Beamforming of user signals: After multiplying the public data stream and the private data streams of all users by the beamforming vector, all signals are superimposed and transmitted by the base station.

[0012] (6) Perform decoding on all users in sequence: if a user uses rate segmentation, proceed to step (7); if a user does not use rate segmentation, proceed to step (8). Continue until all users have completed decoding.

[0013] (7) The user first treats all private data streams as interference, decodes the public data stream from the received signal, and then separates the user's public part from it; then subtracts the corresponding part of the public data stream from the received signal, and decodes the user's private part based on the subtracted signal; the decoded public part and private part of the user are recombined to obtain the user's target data stream. Jump to step (6).

[0014] (8) The user treats all other data streams as interference and directly decodes the private data stream, i.e., the user's target data stream, from the received signal. Jump to step (6).

[0015] Preferably, in step (4), the data transmission rate of the public data stream is equal to the sum of the data transmission rates of the public portion of all users using rate segmentation.

[0016] Preferably, in step (1), the specific operations for optimizing the user's selection scheme are as follows:

[0017] (1-1) Randomly generate user selection schemes and calculate the system's effective throughput corresponding to the user selection scheme.

[0018] (1-2) Randomly select a user and change the user selection scheme, that is: if the user originally used rate segmentation, then the user will not use rate segmentation after the change; if the user originally did not use rate segmentation, then the user will use rate segmentation after the change.

[0019] (1-3) Calculate the system's effective throughput after changing the user's selected scheme.

[0020] (1-4) Determine whether the system's effective throughput after changing the user's selected scheme is greater than the system's effective throughput before changing the user's selected scheme. If yes, proceed to step (1-5); otherwise, proceed to step (1-6).

[0021] (1-5) Approve the changes to the user selection scheme in step (1-2). Proceed to step (1-7).

[0022] (1-6) Disapprove the change to the user selection scheme in this step (1-2) and revert the user selection scheme to the scheme before this change.

[0023] (1-7) Determine whether changing any user's indicator variable will not improve the system's effective throughput. If yes, proceed to (1-8); otherwise, jump to step (1-2).

[0024] (1-8) Output the user's selected option.

[0025] Preferably, in steps (1-1) and (1-3), when calculating the effective throughput of the system, the case of a finite packet length is considered. The effective throughput of the system should be determined by the packet length, user channel coefficient, data transmission rate, user selection scheme, and beamforming vector.

[0026] The beneficial effects of this invention are as follows: The rate division multiple access method proposed in this invention can flexibly select some users to perform rate division, avoiding the problem that the rate is limited by the worst user channel caused by forcing all users to perform rate division, and can improve the effective throughput of the system; The rate division multiple access method proposed in this invention can improve the effective throughput of the system by considering the case of finite packet length that is closer to reality. Attached Figure Description

[0027] Figure 1 This is a schematic diagram of the signal flow at the transmitter of the present invention.

[0028] Figure 2 This is a schematic diagram of the signal flow at the receiving end of the present invention.

[0029] Figure 3 This is a flowchart illustrating the process of optimizing the user selection scheme, data transmission rate, and beamforming vector in a specific implementation method of the present invention.

[0030] Figure 4 A schematic diagram comparing the effective throughput of the flexible RSMA proposed in this invention with traditional RSMA, SDMA, and NOMA under different user numbers.

[0031] Figure 5 This diagram illustrates the comparison of the effective throughput of the flexible RSMA proposed in this invention with traditional RSMA, SDMA, and NOMA under different base station antenna counts. Detailed Implementation

[0032] The following section uses a specific example of a K-user system to illustrate the flexible rate division multiple access method proposed in this invention. Figure 3 As shown, it includes the following steps:

[0033] (1) In the downlink system, the base station is equipped with N t There are K users, each with a single antenna. The user set is denoted as K = {1, ..., K}.

[0034] (2) Optimize beamforming vector, data transmission rate, and user selection scheme. The user selection scheme is a:={a k |k∈κ}, where a k Let ∈{0,1} be the indicator variable for user k, and a k =0 indicates that user k does not use rate segmentation, ak =1 indicates that user k uses rate segmentation.

[0035] (3) Based on the user's selected scheme, perform the following operations sequentially for users k = 1, ..., K: If user k uses rate segmentation, proceed to step (4); if user k does not use rate segmentation, proceed to step (5). After performing the operation for all users, proceed to step (6).

[0036] (4) Transfer the target data stream s of user k k Divided into common parts s k，c and private parts k，p Based on the optimized data transmission rate scheme, the data transmission rate of the common part is R. k，c The data transmission rate of the private portion (i.e., user k's private data stream) is R. k，p Jump to step (3).

[0037] (5) User k has no public data stream; the target data stream is the private data stream s. k，p Based on the optimized data transmission rate scheme, the data transmission rate is R. k，p Jump to step (3).

[0038] (6) Merge the common data portions of all users using rate segmentation into a single common data stream s. c .

[0039] (7) Beamforming the signal based on the optimized beamforming vector: The base station transmits signal x = w c s c +∑ k∈κ w k，p s k，p , where w c w is the beamforming vector for the public data stream. k，p Beamforming vectors for each user's private data stream.

[0040] (8) Decode the signals sequentially for users k = 1, ..., K: The received signal for user k is Among them, h k Let z be the channel vector of user k. k The noise for user k has a mean of 0 and a variance of . Additive white Gaussian noise. If user k uses rate segmentation, proceed to step (9); if user k does not use rate segmentation, proceed to step (10). Continue until all users have finished decoding.

[0041] (9) User k first treats all private data streams as interference, from y k Decoding common data streams in Chinac The decoding result is recorded as After solving, separate the common part of user k. Then from y k Subtract public data flow The corresponding part is obtained according to Decode user k's private part The common part obtained from decoding and private parts Reorganize to obtain the target data stream for user k. Jump to step (8).

[0042] (10) User k treats all other data streams as interference and directly receives data from y. k Decoding private data streams That is, the target data stream for user k. Jump to step (8).

[0043] In step (2), the specific operations for optimizing the beamforming vector, data transmission rate, and user selection scheme are as follows:

[0044] (2-1) Parameter initialization: Initialize the optimal system effective throughput T * :=0; Initialize the original system effective throughput T (old) :=0; Initialize iteration number t:=0.

[0045] (2-2) Randomly generate user selection scheme The superscript (0) indicates the number of iterations.

[0046] (2-3) Update user selection scheme a (new) :=a (t) Update the indicator variable for user k. Where k:=mod(t,K)+1.

[0047] (2-4) Optimize the beamforming vector and data transmission rate to obtain the effective system throughput T. (new) .

[0048] (2-5) Determine if T is true (new) >T (old) If yes, then execute (2-6); otherwise, execute (2-7).

[0049] (2-6) Update T (old) :=T (new) and a (t+1) :=a (new) If T (new) >T * Then update T* =T (new) Jump to execution (2-8).

[0050] (2-7) Keep the indicator variable unchanged, i.e., a (t+1) :=a (t) .

[0051] (2-8) Determine whether changing any user's indicator variable will not improve the system's effective throughput. If yes, execute (2-9); otherwise, jump to execute (2-3).

[0052] (2-9) Output the optimal user selection scheme a (t) The corresponding system effective throughput, beamforming vector, and data transmission rate.

[0053] In steps (2-4), the specific operations for optimizing the beamforming vector and data transmission rate are as follows:

[0054] (2-4-1) Parameter Initialization: Initialize the iteration number q:=0, and the error tolerance δ2>0. Initialize the system's effective throughput T. (0) = 0.

[0055] (2-4-2) Set the initial beamforming vector for the common data stream Set the initial beamforming vector for the private data stream of all users k∈κ. Where P is the total transmit power, and 1 is N. t A vector of all 1s in dimension.

[0056] (2-4-3) Update q: = q+1.

[0057] (2-4-4) Optimize data transmission rate, including: optimizing the data transmission rate R of the private portion for all users. k，p For users employing rate segmentation, optimize the data transmission rate R of the common portion. k，c .

[0058] (2-4-5) Optimize beamforming vector.

[0059] (2-4-6) Computational system effective throughput T (q) .

[0060] (2-4-7) Judgment: If |T (q) -T (q-1) If |≤δ2, then proceed to step (2-4-8); if |T (q) -T (q-1) If |>δ2, then proceed to step (2-4-3).

[0061] (2-4-8) Output beamforming vector, data transmission rate, and system effective throughput T (q) .

[0062] In steps (2-4) and (2-4-6), the formula for calculating the effective throughput of the system is as follows:

[0063]

[0064] Among them, R c =∑ k∈κ a k R k，c It is the sum of the rates of the common part information of users using rate segmentation, where L is the code block length and γ is the sum of the rates. k，c The signal-to-interference-plus-noise ratio (SIR) of the user common area using rate segmentation is:

[0065]

[0066] γ k，p The signal-to-interference-plus-noise ratio (SIR) of the user's private data:

[0067]

[0068] In step (2-4-4), the specific operation for optimizing the data transmission rate of the private part is as follows: For any user k∈κ, the optimal data transmission rate of the private part is calculated by the following formula:

[0069]

[0070] in, It is the equation Φ k，p The solution to equation Φ(x) = 0 k，p (x) is

[0071]

[0072] In step (2-4-4), the specific operations for optimizing the data transmission rate of the common part are as follows:

[0073] (2-4-4-1) Parameter initialization: Set the error tolerance δ3 and initialize the iteration number t: = 0.

[0074] (2-4-4-2) Let κ1 be the set of all users who use rate segmentation. For all users k∈κ1 who use rate segmentation, initialize...

[0075] (2-4-4-3) Update t: = t+1;

[0076] (2-4-4-4) Update

[0077]

[0078] in, It is the equation Φ k，c The solution to equation Φ(x) = 0 k，c (x) is

[0079]

[0080] (2-4-4-5) Judgment: If Then proceed to step (2-4-4-6); if Then proceed to step (2-4-4-3).

[0081] (2-4-4-6) Output the optimal common part data transmission rate

[0082] In steps (2-4-5), the specific operations for optimizing the beamforming vector are as follows:

[0083] (2-4-5-1) Parameter initialization: Initialize the penalty factor η, growth factor β2>1, and error tolerance δ4; initialize the iteration number n:=0.

[0084] (2-4-5-2) Define the parameter vector v = [v i，c , i∈κ1, v k，p ,k∈κ] T and initialized to v (0) .

[0085] (2-4-5-3) Define the beamforming matrix W c and W k，p Let the set of beamforming matrices be W = {W c W k， p|k∈κ}, and initialized to W (0) .

[0086] (2-4-5-4) For a given v (n) and W (n) Solve the optimization problem to obtain v (n+1) and W (n+1 The value of ).

[0087] (2-4-5-5) Update n:=n+1 and η:=β2η.

[0088] (2-4-5-6) Judgment: If ||W (n) -W (n-1) If ||≤δ4, then execute step (2-4-5-7); if ||W (n) -W (n-1) If ||>δ4, then proceed to step (2-4-5-4).

[0089] (2-4-5-7) W (n) The beamforming matrix W in c and W k，p Perform eigenvalue decomposition on each, and denote the largest eigenvalue as λ. c and λ k，p Its corresponding eigenvector is q c and q k，p Output beamforming vector and

[0090] In step (2-4-5-4), the specific operations for solving the optimization problem are as follows:

[0091] (2-4-5-4-1) Define the set of slack variables τ={τ i，c , τ k，p |i∈κ1,k∈κ}

[0092] (2-4-5-4-2) For a given v (n) and W (n) Solve the following convex optimization problem using the CVX toolkit:

[0093]

[0094]

[0095]

[0096]

[0097]

[0098]

[0099]

[0100] log2(1+v k，p )+τ k，p ≥R k，p ,k∈κ

[0101] in, Given the channel matrix of user k, the function Defined as

[0102]

[0103]

[0104]

[0105] Function I i，c (W)=∑ j∈κ tr(H i W j，p ), i∈κ1, function I k，p (W)=(1-a k )tr(H k W c )+∑ j∈κ\{k} tr(H k W j，p ), k∈κ.

[0106] The above description is merely a preferred embodiment of the present invention and should not be construed as limiting the scope of the invention. Therefore, any equivalent variations made in accordance with the claims of the present invention are still within the scope of the present invention.

Claims

1. A flexible rate division multiple access method, characterized in that, Includes the following steps: (1) Process the target data streams of all users in sequence according to the user selection scheme: if the user adopts rate segmentation, then jump to step (2); if the user does not adopt rate segmentation, then jump to step (3); after processing all users, jump to step (4); the specific operation of optimizing the user selection scheme is as follows: (1-1) Randomly generate user selection schemes and calculate the system's effective throughput corresponding to the user selection schemes; (1-2) Randomly select a user and change the user selection scheme, that is: if the user originally used rate segmentation, then the user will not use rate segmentation after the change; if the user originally did not use rate segmentation, then the user will use rate segmentation after the change. (1-3) Calculate the system's effective throughput after changing the user's selected scheme; (1-4) Determine whether the effective throughput of the system after the change of the user's selected scheme is greater than the effective throughput of the system before the change of the user's selected scheme; if yes, proceed to step (1-5); if no, proceed to step (1-6). (1-5) Approve the changes to the user selection scheme in step (1-2); Jump to execution steps (1-7); (1-6) Do not approve the change of the user selection scheme in this step (1-2), and revert the user selection scheme to the scheme before this change; (1-7) Determine whether changing any user's indicator variable will not improve the system's effective throughput; if yes, proceed to (1-8); if no, proceed to step (1-2). (1-8) Output the user's chosen solution; (2) Divide the user’s target data stream into a public part and a private part, the private part being the user’s private data stream; Jump to execution step (1); (3) This user has no public data stream; the target data stream is the private data stream. Jump to execution step (1); (4) Merge the common parts of all users using rate segmentation into a single common data stream; (5) Beamforming the user signal: After multiplying the public data stream and the private data streams of all users by the beamforming vector, all signals are superimposed and sent by the base station; (6) Decode all users in sequence: If a user uses rate segmentation, proceed to step (7); if a user does not use rate segmentation, proceed to step (8); until all users have completed decoding. (7) The user first treats all private data streams as interference, decodes the public data stream from the received signal, and then separates the user's public part from it; then subtracts the part corresponding to the public data stream from the received signal, and decodes the user's private part based on the subtracted signal; the decoded public part and private part of the user are recombined to obtain the user's target data stream; Jump to step (6); (8) The user treats all other data streams as interference and directly decodes the private data stream from the received signal, which is the user's target data stream; jump to step (6).

2. The flexible rate division multiple access method as described in claim 1, characterized in that, In step (4), the data transmission rate of the public data stream is equal to the sum of the data transmission rates of the public portion of all users using rate segmentation.

3. The flexible rate division multiple access method as described in claim 1, characterized in that, In steps (1-1) and (1-3), when calculating the effective throughput of the system, the case of a finite packet length is considered. The effective throughput of the system should be determined by the packet length, user channel coefficient, data transmission rate, user selection scheme, and beamforming vector.

Images