Joint resource allocation method for cellular MIMO based on distributed-centralized hybrid processing architecture
By alternately solving power control coefficients and centralized AP indicators in the distributed-centralized hybrid precoding architecture, combining slack and alternating optimization, the problems of user fairness and limited fronthaul link capacity are solved, and the spectrum efficiency is improved and improvements in system performance.
Patent Information
- Application Number
- CN202311307999.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-10-10
- Publication Date
- 2024-11-29
- Estimated Expiration
- 2043-10-10
AI Technical Summary
The existing distributed-centralized hybrid precoding architecture cannot effectively guarantee user fairness under the condition of limited prefamily link capacity. Moreover, the selection of centralized precoding access point sets and power allocation are independently optimized, ignoring the transmission power and execution concentration. The coupling relationship between the set of precoded access points results in a degradation in system performance.
By obtaining channel information between the user and the access point, measuring the signal reception power, generating a connection matrix, and establishing optimization problems based on maximizing the minimum spectrum efficiency as the optimization goal, alternating the power control coefficient and centralized AP indicator flags, combined with Slack and alternating optimization methods reduce calculation complexity and realize resource allocation.
While reducing the preamble overhead, the spectrum efficiency is improved, and the spectrum efficiency and the preamble link overhead are compromised, ensuring user fairness and reducing computational complexity.
Smart Images

Figure CN117500080B_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of wireless communications, and in particular relates to a decellular MIMO joint resource allocation method in a distributed-centralized hybrid processing architecture. Background Art
[0002] As one of the potential key technologies of B5G (Beyond 5G) and 6G, the decellularized massive MIMO (Multiple Input Multiple Output) system has received widespread attention. This network breaks the traditional cell architecture. Multiple access points (Access Points) deployed in a dispersed manner are managed by the CPU (Central Process Unit) in a unified manner. They use the same time and frequency resources and collaborate with each other to provide services to users, thus realizing distributed massive MIMO transmission. The decellularized massive MIMO network fully combines the advantages of massive MIMO technology, joint multi-point transmission technology, and ultra-dense microcellular network technology. It can significantly improve network throughput and enhance user service fairness, and has become the basic form of the next generation network.
[0003] For decellularized large-scale MIMO networks that use a CPU-access point dual-layer network architecture, resource management is usually divided into two modes: distributed processing and centralized processing, depending on the functions of the access point and the CPU and the type of interactive information. When distributed processing is used, each access point performs precoding based on the local CSI (channel state information) and performs centralized power allocation at the CPU. At this time, the access point only uploads large-scale CSI information to the CPU, and the CPU only sends the power control coefficient to each access point. When centralized processing is used, the CPU performs centralized precoding and power allocation based on the global CSI information. At this time, each access point needs to upload complete CSI information to the CPU, and in addition to the power allocation coefficient, the CPU also needs to send the precoding matrix to each access point. Centralized processing can achieve better interference management performance than distributed processing, but the fronthaul overhead is large.
[0004] In actual deployment, although a large-capacity optical fiber connection is used between the access point and the CPU, the transmission capacity of the fronthaul link is still limited. Regardless of distributed processing or centralized processing, the limited fronthaul capacity will lead to a degradation of system performance. To address this problem, Liutong Du proposed a precoding scheme that mixed CB (Conjugate beamforming) and ZF (zero forcing), that is, some access points use centralized ZF precoding to effectively eliminate interference, and the remaining access points use distributed CB precoding. Minhyun Kim proposed another resource allocation strategy under distributed-centralized hybrid precoding processing, in which the precoding scheme is that some access points use centralized partial minimum mean square error precoding, and the remaining access points use distributed local partial minimum mean square error precoding.
[0005] The above-mentioned distributed-centralized hybrid precoding architecture provides a new solution for resource management under the condition of limited fronthaul link capacity, but the research on this architecture is still in its infancy. Existing methods are either only applicable to fully connected scenarios or cannot effectively guarantee user fairness. In addition, in existing methods, the selection of centralized precoding access point set and power allocation are optimized independently, ignoring the coupling relationship between transmission power and the execution of centralized precoding access point set. Summary of the invention
[0006] In order to solve the above problems existing in the prior art, the present invention provides a method for decellular MIMO joint resource allocation in a distributed-centralized hybrid processing architecture. The technical problem to be solved by the present invention is achieved through the following technical solutions:
[0007] The present invention provides a decellularized MIMO joint resource allocation method in a distributed-centralized hybrid processing architecture, which is applied to a decellularized multiple-input multiple-output MIMO system, wherein the decellularized MIMO system includes M access points AP and K users;
[0008] The method comprises:
[0009] Get the channel information h between user k and access point m m,k , m=1,2,...,M, k=1,2,...,K;
[0010] For each user, the signal reception power between the user and each access point is measured, and an access point that meets the preset conditions is selected to serve the user based on the signal reception power, and a connection matrix C is generated;
[0011] Based on the channel information h m,k and the connection matrix C, with maximizing the minimum spectrum efficiency as the optimization goal, to establish an optimization problem;
[0012] By solving the optimization problem, the M access points are divided into a centralized AP set Z c Or distributed AP set Z d , and allocate resources based on the optimization results.
[0013] In one embodiment of the present invention, for each user, the signal reception power between the user and each access point is measured, and an access point that meets a preset condition is selected based on the signal reception power to serve the user, and then the step of generating a connection matrix C includes:
[0014] Get the signal receiving power P when user k accesses access point m m,k ;
[0015] Determine the signal receiving power P m,k Whether the preset conditions are met; if so, select access point m for user k to connect, and obtain a connection matrix C; the element C in the connection matrix C m,k =1 or C m,k =0, where C m,k =1 means user k is connected to access point m, C m,k =0 means user k is not connected to access point m.
[0016] In one embodiment of the present invention, the preset condition is:
[0017] 10lg(P max / P m,k )≤β;
[0018] Where P max represents the maximum signal receiving power when user k accesses each access point, and β represents the preset access threshold, in dB.
[0019] In one embodiment of the present invention, the optimization problem is expressed as:
[0020]
[0021] subject to C1: m2∈Z d ,
[0022] C2:
[0023] C3:
[0024] C4:
[0025] C5:
[0026] C6:λ m ∈{0,1}
[0027] In the formula, C1, C2, C3, C4, C5 and C6 are constraints, SINR k represents the signal-to-interference-noise ratio of user k, η m,k represents the power control coefficient between user k and access point m, Represents user k and centralized AP set Z c The power control coefficient between the access points m1, Represents user k and distributed AP set Z d The power control coefficient between access points m2, λ m It is the centralized AP indicator. m =1, it indicates that access point m is a centralized AP, otherwise it indicates that access point m is a distributed AP. Indicates that user k is connected to centralized AP m1, Indicates that user k is connected to distributed AP m2, The number of centralized APs preset in the centralized AP set.
[0028] In one embodiment of the present invention, the optimization problem is solved according to the following steps:
[0029] Initialize related parameters, the related parameters include: Centralized AP selection threshold λ c , initial AP selection matrix λ (0) , number of iterations N, upper bound t of binary search upper With the lower bound t lower And the solution accuracy ε>0;
[0030] In each iteration, the power control coefficients of this iteration are solved alternately. and centralized AP indicator Until the preset number of iterations N is reached.
[0031] In one embodiment of the present invention, in each iteration process, the power control coefficients in this iteration process are alternately solved. and centralized AP indicator The steps include:
[0032] The initial AP selection matrix λ (0) As the current AP selection matrix λ (n-1) , n represents the current iteration number;
[0033] Select matrix λ for the current AP (n-1) The element λ in mAfter being sorted in descending order, the APs that are greater than or equal to the centralized AP selection threshold λ c The access points corresponding to the elements are placed in the centralized precoding AP set Z c and place the access points corresponding to the remaining elements in the distributed precoding AP set Z d middle;
[0034] For centralized AP set Z c The access point m1 in the central precoding is used, and the distributed AP set Z d The access point m2 in performs distributed precoding;
[0035] Keep centralized AP set Z c Unchanged, solve the power control coefficient in this iteration process
[0036] Maintain the power control coefficient Unchanged, solve the centralized AP indicator flag in this iteration process
[0037] In one embodiment of the present invention, the matrix λ is selected for the current AP. (n-1) The element λ in m After being sorted in descending order, the APs that are greater than or equal to the centralized AP selection threshold λ c The access points corresponding to the elements are placed in the centralized precoding AP set Z c and place the access points corresponding to the remaining elements in the distributed precoding AP set Z d The steps in include:
[0038] Select matrix λ for the current AP (n-1) The element λ in m Arrange in descending order and determine the APs that are greater than or equal to the centralized AP selection threshold λ c The number of elements in ;
[0039] Compare the number of elements with the preset number of centralized APs If the number of elements is greater than the preset number of centralized APs Select the previous The access points corresponding to the elements are placed in the centralized precoding AP set Z c Otherwise, it will be greater than or equal to the centralized AP selection threshold λ c The access points corresponding to the elements are placed in the centralized precoding AP set Z c middle;
[0040] Place the access points corresponding to the remaining elements in the distributed precoding AP set Z d middle.
[0041] In one embodiment of the present invention, a centralized AP set Z is maintained. c Unchanged, solve the power control coefficient The steps include:
[0042] Determine whether t is satisfied upper -t lower >ε; if not, then execute the method of maintaining the power control coefficient Unchanged, solve the centralized AP indicator flag in this iteration process Steps;
[0043] If so, solve the first feasibility problem as follows:
[0044]
[0045] If the first feasibility problem has a feasible solution, let t = t lower , On the contrary, let t = t upper ;
[0046] Returns whether the judgment satisfies t upper -t lower >ε steps.
[0047] In one embodiment of the present invention, the power control coefficient is maintained Unchanged, solve the centralized AP indicator flag in this iteration process The steps include:
[0048] Determine whether t is satisfied upper -t lower >ε; if not, return the current AP selection matrix λ (n-1) The element λ in m Perform the steps of sorting in descending order;
[0049] If so, solve the second feasibility problem as follows:
[0050]
[0051] If the second feasibility problem has a feasible solution, let t = t lower , On the contrary, let t = t upper ;
[0052] Returns whether the judgment satisfies t upper -t lower >ε steps.
[0053] Compared with the prior art, the present invention has the following beneficial effects:
[0054] The present invention provides a method for joint resource allocation of cellular MIMO under a distributed-centralized hybrid processing architecture. Since the present invention considers the coupling relationship between centralized AP selection and power allocation coefficient when solving the optimization problem, the two are jointly optimized with the goal of maximizing the spectrum efficiency of the minimum user. Therefore, compared with the step-by-step optimization scheme under the existing distributed-centralized hybrid precoding, the present invention can improve the spectrum efficiency while reducing the fronthaul overhead, achieving a compromise between the fronthaul link overhead and the spectrum efficiency, while ensuring user fairness. In addition, the present invention proposes to solve the joint optimization problem by using relaxation and alternating optimization, and greatly reduces the computational complexity by relaxing the sub-problems.
[0055] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS
[0056] Figure 1 It is a flow chart of a method for allocating cellular MIMO joint resources in a distributed-centralized hybrid processing architecture provided by an embodiment of the present invention;
[0057] Figure 2 is a schematic diagram of a decellularized MIMIO system provided by an embodiment of the present invention;
[0058] Figure 3a is a spectrum efficiency comparison diagram under different numbers of centralized APs provided by an embodiment of the present invention;
[0059] Figure 3b It is a comparison diagram of cumulative distribution function curves provided by an embodiment of the present invention. DETAILED DESCRIPTION
[0060] The present invention is further described in detail below with reference to specific embodiments, but the embodiments of the present invention are not limited thereto.
[0061] Figure 1 This is a flow chart of a method for allocating cellular MIMO joint resources in a distributed-centralized hybrid processing architecture provided by an embodiment of the present invention. Figure 1 As shown, an embodiment of the present invention provides a decellularized MIMO joint resource allocation method in a distributed-centralized hybrid processing architecture, which is applied to a decellularized multiple-input multiple-output MIMO system, wherein the decellularized MIMO system includes M access points AP and K users;
[0062] The above methods include:
[0063] S1. Obtain the channel information h between user k and access point m m,k , m=1,2,...,M, k=1,2,...,K;
[0064] S2. For each user, measure the signal receiving power between the user and each access point, and select an access point that meets the preset conditions to serve the user based on the signal receiving power, and then generate a connection matrix C;
[0065] S3, based on channel information h m,k and the connection matrix C, and establish the optimization problem with the optimization goal of maximizing the minimum spectrum efficiency;
[0066] S4. By solving the optimization problem, the M access points are divided into a centralized AP set Z c Or distributed AP set Z d , and allocate resources based on the optimization results.
[0067] Specifically, in the decellularized MIMO system, a large number of distributed APs at random locations are deployed in the area covered by the network. All APs are connected to the central processor through the fronthaul link for joint processing and participate in data transmission at the same frequency. The AP exchanges various variable information for network operation with the central processor through the fronthaul link, achieving the effect of participating in data processing and transmission with the central processor. Optionally, the decellularized MIMO system of this embodiment includes M access points and K users, and each access point has N t Antenna.
[0068] Optionally, in step S2, for each user, the signal reception power between the user and each access point is measured, and an access point that meets a preset condition is selected based on the signal reception power to serve the user, and then the step of generating a connection matrix C includes:
[0069] S201. Obtaining the signal receiving power P when user k accesses access point m m,k ;
[0070] S202, determine the signal receiving power P m,k Whether the preset conditions are met; if so, select access point m for user k to connect and obtain the connection matrix C; the element C in the connection matrix C m,k =1 or C m,k =0, where C m,k =1 means user k is connected to access point m, C m,k =0 means user k is not connected to access point m.
[0071] In this embodiment, the preset conditions are:
[0072] 10lg(P max / P m,k )≤β;
[0073] Where P maxrepresents the maximum signal receiving power when user k accesses each access point, and β represents the preset access threshold, in dB.
[0074] In this embodiment, maximizing the minimum spectrum efficiency is taken as the optimization goal, and the established optimization problem is expressed as:
[0075]
[0076] subject to C1: m2∈Z d ,
[0077] C2:
[0078] C3:
[0079] C4:
[0080] C5:
[0081] C6:λ m ∈{0,1}
[0082] In the formula, C1, C2, C3, C4, C5 and C6 are constraints, SINR k represents the signal-to-interference-noise ratio of user k, η m,k represents the power control coefficient between user k and access point m, Represents user k and centralized AP set Z c The power control coefficient between the access points m1, Represents user k and distributed AP set Z d The power control coefficient between access points m2, λ m It is the centralized AP indicator. m =1, it indicates that access point m is a centralized AP, otherwise it indicates that access point m is a distributed AP. Indicates that user k is connected to centralized AP m1, Indicates that user k is connected to distributed AP m2, The number of centralized APs preset in the centralized AP set.
[0083] Specifically, in the above optimization problem, constraint C1 is a constraint of centralized precoding, which is used to ensure that centralized precoding still has the ability to eliminate inter-user interference after power allocation, and constraint C2 is used to ensure that the number of centralized APs is less than It can be flexibly adjusted according to the network overhead. Constraints C3 and C5 are power constraints. Constraint C4 is used to ensure that when access point m is not connected to user k, the power allocated to it by access point m is 0. In addition, in order to reduce the computational complexity, this embodiment uses λ m Relax to a continuous variable in the interval [0,1], that is, construct constraint C6:λ m ∈[0,1].
[0084] Further, the optimization problem is solved as follows:
[0085] S401, initializing relevant parameters, the relevant parameters include: Centralized AP selection threshold λ c , initial AP selection matrix λ (0) , number of iterations N, upper bound t of binary search upper With the lower bound t lower And the solution accuracy ε>0;
[0086] S402, in each iteration process, alternately solving the power control coefficient of this iteration process and centralized AP indicator Until the preset number of iterations N is reached.
[0087] Optionally, in each iteration, the power control coefficients in this iteration are solved alternately. and centralized AP indicator The steps include:
[0088] S4021, initial AP selection matrix λ (0) As the current AP selection matrix λ (n-1) , n represents the current iteration number;
[0089] S4022. Select matrix λ for the current AP (n-1) The element λ in m After sorting in descending order, the APs that are greater than or equal to the centralized AP selection threshold λ c The access points corresponding to the elements are placed in the centralized precoding AP set Z c and place the access points corresponding to the remaining elements in the distributed precoding AP set Z d middle;
[0090] S4023, centralized AP set Z c The access point m1 in the central precoding is used, and the distributed AP set Z d The access point m2 in performs distributed precoding;
[0091] S4024, maintain centralized AP set Z cUnchanged, solve the power control coefficient in this iteration process
[0092] S4025, maintain power control coefficient Unchanged, solve the centralized AP indicator flag in this iteration process
[0093] In step S4022, the matrix λ is selected for the current AP. (n-1) The element λ in m After sorting in descending order, the APs that are greater than or equal to the centralized AP selection threshold λ c The access points corresponding to the elements are placed in the centralized precoding AP set Z c and place the access points corresponding to the remaining elements in the distributed precoding AP set Z d The steps in include:
[0094] Select matrix λ for the current AP (n-1) The element λ in m Sort in descending order and determine the APs that are greater than or equal to the centralized AP selection threshold λ c The number of elements of
[0095] Compare the number of elements with the preset number of centralized APs If the number of elements is greater than the preset number of centralized APs Select the previous The access points corresponding to the elements are placed in the centralized precoding AP set Z c Otherwise, it will be greater than or equal to the centralized AP selection threshold λ c The access points corresponding to the elements are placed in the centralized precoding AP set Z c middle;
[0096] Place the access points corresponding to the remaining elements in the distributed precoding AP set Z d middle.
[0097] In step S4024, the centralized AP set Z is maintained. c Unchanged, solve the power control coefficient The steps include:
[0098] Determine whether t is satisfied upper -t lower >ε; if not, execute the above power control coefficient Unchanged, solve the centralized AP indicator flag in this iteration process Steps;
[0099] If so, solve the first feasibility problem as follows:
[0100]
[0101] If the first feasibility problem has a feasible solution, let t = t lower , On the contrary, let t = t upper ;
[0102] Return to step S4024 to determine whether t is satisfied. upper -t lower >ε steps.
[0103] In step S4025, the power control coefficient is maintained Unchanged, solve the centralized AP indicator flag in this iteration process The steps include:
[0104] Determine whether t is satisfied upper -t lower >ε; if not, return the current AP selection matrix λ (n-1) The element λ in m Perform the steps of sorting in descending order;
[0105] If so, solve the second feasibility problem as follows:
[0106]
[0107] If there is a feasible solution to the second feasibility problem, let t = t lower , On the contrary, let t = t upper ;
[0108] Return to step S4025 to determine whether t is satisfied. upper -t lower >ε steps.
[0109] It should be understood that in this example, the centralized AP performs centralized interference-aware precoding, and the precoding coefficients of access point m and user k can be obtained. First, the users are divided into groups, assuming that there are K i users, then the channel matrix corresponding to this group of users and all the access points they connect to is Among them, the user is the selected K i users, and APs are all APs connected to the selected users. For the channels corresponding to other users connected to these APs, the precoding matrix for interference perception of this group of users is calculated as follows:
[0110]
[0111] in,
[0112] The precoding vectors of other groups of users are calculated in the same way, and the precoding vectors of the unconnected parts are set to zero to obtain the precoding vector w of each user group. k , the final precoding matrix is:
[0113]
[0114] For distributed APs, ZF precoding is used, and the precoding matrix of access point m is:
[0115]
[0116] Among them, H m =[h m1 ,h m2 ,…,h mK ].
[0117] When the hybrid precoding method of centralized interference-aware precoding and distributed ZF precoding is adopted, the signal received by user k is:
[0118]
[0119] At this time, the SINR of user k can be expressed as:
[0120]
[0121] Since λ in the above optimization problem m and η m,k The two variables are coupled to each other. Therefore, in order to reduce the complexity of the solution, this embodiment decomposes the optimization problem into two sub-problems: centralized AP selection and power allocation, and solves them through an alternating optimization iterative algorithm. Specifically, first initialize the number of centralized APs. Centralized AP selection threshold λ c ∈[0,1], preset number of iterations N, upper bound t of binary search upper With the lower bound t lower , the accuracy of subproblem solving ε>0 and the initial AP selection matrix
[0122] Next, the solution of the optimization problem is decomposed into two sub-problems and solved alternately in each iteration:
[0123] First, select the matrix λ for the current AP (n-1) Sort in descending order and select m ≥λ c Lambda m The corresponding AP is placed in the AP set Z that performs centralized precoding c If the number of APs that meet the conditions is greater than Select the previous APs are placed in set Z c The remaining APs are placed in the AP set Z that performs distributed precoding. d In. Z c The AP in Z performs centralized precoding. d The APs in perform distributed precoding.
[0124] Then, fix the centralized AP set and solve the power control coefficient If upper -t lower >ε, solve the following first feasibility problem:
[0125]
[0126] When there is a feasible solution to the first feasibility problem, let t = t lower , Otherwise, let t = t upper .
[0127] Furthermore, the fixed power control coefficient Solving the AP selection matrix If upper -t lower >ε, solve the following second feasibility problem:
[0128]
[0129] If there is a feasible solution to the second feasibility problem, let t = t lower , Otherwise, let t = t upper .
[0130] Next, the decellular MIMO joint resource allocation method based on the distributed-centralized hybrid processing architecture provided by the present invention is further described through simulation experiments.
[0131] 1. Simulation conditions:
[0132] The hardware platform of the simulation experiment is: Intel(R) Core(TM) i7-8700k CPU, the main frequency of the processor is 3.20GHz, and the memory is 16GB.
[0133] The simulation is implemented using MATLAB R2020 software, and the parameters are set as follows: the total number of APs M is 30, the number of antennas per AP N is 8, the total number of users K is 8, the number of cycles N is 8, and the upper bound of SINR t upper and the lower bound t lower 20 and 0 respectively, the sub-problem solving accuracy ε is 0.05, the grouping threshold β is 20dB, and the time slot τ of the user downlink data transmission in the coherent block dis 100 and the carrier frequency is 3.5GHz.
[0134] 2. Simulation content and result analysis:
[0135] Simulation 1. Under the above simulation conditions, the spectrum efficiency of the cellular MIMO joint resource allocation method under the distributed-centralized hybrid processing architecture provided by the present invention and the existing distributed-centralized hybrid precoding resource allocation method based on distributed optimization under different centralized AP numbers is simulated and compared. The results are as follows: Figure 3a shown.
[0136] like Figure 3a As shown, the horizontal axis and the vertical axis represent the number of centralized APs and the average user spectrum efficiency, respectively. Compared with the existing distribution-based optimization method, the spectrum efficiency of the present invention is significantly improved when the number of centralized APs is the same. When the number of centralized APs is 15, the performance advantage of the present invention is most obvious. Compared with the step-by-step optimization solution, the spectrum efficiency can be improved by up to 10%.
[0137] Simulation 2. Under the above simulation conditions, the spectrum efficiency of the cellular MIMO joint resource allocation method under the distributed-centralized hybrid processing architecture provided by the present invention and the existing distributed-centralized hybrid precoding resource allocation method based on distributed optimization when selecting the threshold is simulated. For the existing distributed optimization scheme, the number of centralized APs is set to 20, and for the present invention, the maximum number of centralized APs is set to Z. max is 20, threshold λ c is 0.9.
[0138] Table 1 shows the forwarding overhead results:
[0139] Table 1
[0140] plan Fronthaul overhead Centralized processing <![CDATA[0.8263×10 4 ]]> Distributed processing <![CDATA[0.6343×10 4 ]]> Distributed-centralized hybrid precoding (step-by-step optimization of resources <![CDATA[0.7623×10 4 ]]> The present invention <![CDATA[0.6619×10 4 ]]>
[0141] Figure 3b : is a comparison diagram of the cumulative distribution function curve provided by an embodiment of the present invention, wherein the horizontal axis and the vertical axis represent the user spectrum efficiency and the CDF (Cumulative Distribution Function, cumulative probability distribution function) respectively. Figure 3b It can be seen that the user spectrum efficiency of the present invention is higher than that of the existing step-by-step optimization scheme. Further combined with Table 1, the fronthaul overhead of centralized processing is the largest, the fronthaul overhead of distributed processing is the smallest, and the fronthaul overhead of distributed-centralized precoding is between the two. However, the scheme of the present invention can save about 13% of the fronthaul overhead compared with the existing step-by-step optimization scheme, that is, the present invention can improve the user spectrum efficiency while saving the fronthaul overhead.
[0142] It can be seen from the above embodiments that the beneficial effects of the present invention are:
[0143] The present invention provides a method for joint resource allocation of cellular MIMO under a distributed-centralized hybrid processing architecture. Since the present invention considers the coupling relationship between centralized AP selection and power allocation coefficient when solving the optimization problem, the two are jointly optimized with the goal of maximizing the spectrum efficiency of the minimum user. Therefore, compared with the step-by-step optimization scheme under the existing distributed-centralized hybrid precoding, the present invention can improve the spectrum efficiency while reducing the fronthaul overhead, achieving a compromise between the fronthaul link overhead and the spectrum efficiency, while ensuring user fairness. In addition, the present invention proposes to solve the joint optimization problem by using relaxation and alternating optimization, and greatly reduces the computational complexity by relaxing the sub-problems.
[0144] In the description of the present invention, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.
[0145] The description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine different embodiments or examples described in this specification.
[0146] Although the present application is described herein in conjunction with various embodiments, in the process of implementing the present application for which protection is sought, a person skilled in the art may understand and implement other variations of the disclosed embodiments by reviewing the drawings, the disclosure, and the appended claims.
[0147] The above contents are further detailed descriptions of the present invention in combination with specific preferred embodiments, and it cannot be determined that the specific implementation of the present invention is limited to these descriptions. For ordinary technicians in the technical field to which the present invention belongs, several simple deductions or substitutions can be made without departing from the concept of the present invention, which should be regarded as falling within the protection scope of the present invention.
Claims
1. A method for decellular MIMO joint resource allocation based on a distributed-centralized hybrid processing architecture, characterized in that: The invention is applied to a decellularized multiple-input multiple-output (MIMO) system, wherein the decellularized multiple-input multiple-output (MIMO) system includes M access points (APs) and K users. The method comprises: Get the channel information h between user k and access point m m,k , m=1,2,...,M, k=1,2,...,K; For each user, the signal receiving power between the user and each access point is measured, and an access point that meets a preset condition is selected based on the signal receiving power to serve the user, and then a connection matrix C is generated; including: Get the signal receiving power P when user k accesses access point m m,k ; Determine the signal receiving power P m,k Whether the preset conditions are met; if so, select access point m for user k to connect, and obtain a connection matrix C; the element C in the connection matrix C m,k =1 or C m,k =0, where C m,k =1 means user k is connected to access point m, C m,k =0 means user k is not connected to access point m; Based on the channel information h m,k and the connection matrix C, with maximizing the minimum spectrum efficiency as the optimization goal, an optimization problem is established; the optimization problem is expressed as: C6:λ m ∈{0,1} In the formula, C1, C2, C3, C4, C5 and C6 are constraints, SINR k represents the signal-to-interference-noise ratio of user k, η m,k represents the power control coefficient between user k and access point m, Represents user k and centralized AP set Z c The power control coefficient between the access points m1, Represents user k and distributed AP set Z d The power control coefficient between access points m2, λ m It is the centralized AP indicator. m =1, it indicates that access point m is a centralized AP, otherwise it indicates that access point m is a distributed AP. Represents user k and centralized AP set Z c The access point m1 is connected, Represents user k and distributed AP set Z d In the access point m2 connection, is the number of centralized APs preset in the centralized AP set; By solving the optimization problem, the M access points are divided into a centralized AP set Z c Or distributed AP set Z d , and allocate resources based on the optimization results.
2. The method for decellular MIMO joint resource allocation based on the distributed-centralized hybrid processing architecture according to claim 1, characterized in that: The preset conditions are: 10lg(P max / P m,k )≤β; Where P max represents the maximum signal receiving power when user k accesses each access point, and β represents the preset access threshold, in dB.
3. The method for decellular MIMO joint resource allocation based on the distributed-centralized hybrid processing architecture according to claim 1, characterized in that: Follow the steps below to solve the optimization problem: Initialize related parameters, the related parameters include: Centralized AP selection threshold λ c , initial AP selection matrix λ (0) , number of iterations N, upper bound t of binary search upper With the lower bound t lower And the solution accuracy ε>0; In each iteration, the power control coefficients of this iteration are solved alternately. and centralized AP indicator Until the preset number of iterations N is reached.
4. The method for decellular MIMO joint resource allocation based on the distributed-centralized hybrid processing architecture according to claim 3, characterized in that: In each iteration, the power control coefficients of this iteration are solved alternately. and centralized AP indicator The steps include: The initial AP selection matrix λ (0) As the current AP selection matrix λ (n-1) , n represents the current iteration number; Select matrix λ for the current AP (n-1) The element λ in m After being sorted in descending order, the APs that are greater than or equal to the centralized AP selection threshold λ c The access points corresponding to the elements are placed in the centralized precoding AP set Z c and place the access points corresponding to the remaining elements in the distributed precoding AP set Z d middle; For centralized AP set Z c The access point m1 in the central precoding is used, and the distributed AP set Z d The access point m2 in performs distributed precoding; Keep centralized AP set Z c Unchanged, solve the power control coefficient in this iteration process Maintain the power control coefficient Unchanged, solve the centralized AP indicator flag in this iteration process 5. The method for decellular MIMO joint resource allocation based on the distributed-centralized hybrid processing architecture according to claim 4, characterized in that: Select matrix λ for the current AP (n-1) The element λ in m After being sorted in descending order, the APs that are greater than or equal to the centralized AP selection threshold λ c The access points corresponding to the elements are placed in the centralized precoding AP set Z c and place the access points corresponding to the remaining elements in the distributed precoding AP set Z d The steps in include: Select matrix λ for the current AP (n-1) The element λ in m Arrange in descending order and determine the APs that are greater than or equal to the centralized AP selection threshold λ c The number of elements in ; Compare the number of elements with the preset number of centralized APs If the number of elements is greater than the preset number of centralized APs Select the previous The access points corresponding to the elements are placed in the centralized precoding AP set Z c Otherwise, it will be greater than or equal to the centralized AP selection threshold λ c The access points corresponding to the elements are placed in the centralized precoding AP set Z c middle; Place the access points corresponding to the remaining elements in the distributed precoding AP set Z d middle.
6. The method for decellular MIMO joint resource allocation based on the distributed-centralized hybrid processing architecture according to claim 4, characterized in that: Keep centralized AP set Z c Unchanged, solve the power control coefficient in this iteration process The steps include: Determine whether t is satisfied upper -t lower >ε; if not, then execute the method of maintaining the power control coefficient Unchanged, solve the centralized AP indicator flag in this iteration process Steps; If so, solve the first feasibility problem as follows: If the first feasibility problem has a feasible solution, let t = t lower , On the contrary, let t = t upper ; Returns whether the judgment satisfies t upper -t lower >ε steps.
7. The method for decellular MIMO joint resource allocation based on a distributed-centralized hybrid processing architecture according to claim 6, characterized in that: Maintain the power control coefficient Unchanged, solve the centralized AP indicator flag in this iteration process The steps include: Determine whether t is satisfied upper -t lower >ε; if not, determine whether the current number of iterations reaches the preset number of iterations N, and if the current number of iterations does not reach the preset number of iterations N, further determine whether t upper -t lower >ε; if not, return the current AP selection matrix λ (n-1) The element λ in m Perform the steps of sorting in descending order; If so, solve the second feasibility problem as follows: l m ∈[0,1] If the second feasibility problem has a feasible solution, let t = t lower , On the contrary, let t = t upper ; Return whether the judgment satisfies t upper -t lower >ε steps.
Citation Information
Patent Citations
Power distribution method in decellularized large-scale MIMO system based on PZF precoding
CN112468196A
De-cellular large-scale MIMO uplink spectrum efficiency optimization method
CN113179109A
