Pilot frequency distribution method for cellular-removed large-scale MIMO (Multiple Input Multiple Output) system based on topology awareness

Abstract

The invention relates to the technical field of wireless communication, in particular to a topology awareness-based pilot frequency distribution method for a cellular-removed large-scale MIMO (Multiple Input Multiple Output) system, which comprises the following steps of: establishing a system channel model, and calculating a large-scale fading coefficient between an AP (Access Point) and a user; constructing an inter-user interference topological graph based on a large-scale fading coefficient, and reserving strong interference connection through sparse processing; extracting loops in the topological graph, and screening significant loops according to lengths and weights; extracting a significant loop in the topological graph and calculating a weight; distributing a gradient phase based on the loop weight; generating an orthogonalization phase modulation pilot frequency matrix in combination with a user channel intensity priority; and calculating channel estimation parameters, a user downlink reachable rate and an MSE (Mean Square Error), and realizing maximization of system performance. And the actual application performance of the honeycomb-removed large-scale MIMO system is effectively improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of wireless communication, in particular to a pilot allocation method for a topology-aware cell-free massive MIMO system. BACKGROUND

[0002] The cell-free massive MIMO system has the advantages of no switching cost, wide coverage, and high transmission reliability due to the characteristics of cooperative work of distributed access points (APs). The system provides services for all users through a large number of APs, and the user receives signals are superimposed by the transmission signals of multiple APs, which effectively improves the system capacity and signal coverage quality, and has an irreplaceable position in the next generation of mobile communication.

[0003] However, pilot contamination is the core bottleneck that restricts the performance of the cell-free massive MIMO system. Because the pilot resource is limited (the pilot length is usually much smaller than the number of users K), multiple users have to reuse the same pilot sequence, resulting in serious inter-user interference in the channel estimation process. This interference directly reduces the channel estimation accuracy, and then affects the precoding matrix design and downlink transmission rate, which becomes a key factor limiting the performance improvement of the system.

[0004] The existing pilot allocation schemes cannot fundamentally solve the pilot contamination problem:

[0005] Random pilot allocation (RPA): completely ignoring the interference correlation between users, the pilot sequence is randomly allocated, resulting in the most serious pilot contamination and the worst system performance;

[0006] Greedy pilot allocation (GPA): only adjusting the pilot through local rate optimization, lacking consideration of the global interference topology, easy to fall into local optimization, and the optimization effect is limited;

[0007] Graph-based pilot allocation (GB): only minimizing the intra-group interference through user grouping, without considering the internal structure of the loop interference in the group, and the interference suppression is not complete;

[0008] Location-based greedy pilot allocation (LBGPA): using geographical location as the grouping basis, without combining the actual channel interference characteristics, the performance is unstable in complex channel environment (such as multi-shielding, multi-reflection scene).

[0009] In addition, auxiliary technologies such as reconfigurable intelligent surface (RIS) can improve signal coverage, but cannot fundamentally solve the interference problem caused by pilot reuse. Therefore, there is an urgent need for a pilot allocation method that can accurately capture the interference topology relationship of users and adaptively suppress pilot contamination, in order to break through the technical bottleneck of the existing technology and improve the practical application performance of the cell-free massive MIMO system. SUMMARY

[0010] The application aims to provide a pilot allocation method for a topology-aware de-cell large-scale MIMO system, which can realize accurate suppression of pilot contamination, improve channel estimation accuracy, maximize user and rate of the de-cell large-scale MIMO system, and ensure engineering feasibility and universality of the scheme by mining the topology structure characteristics of inter-user interference, combining phase optimization and orthogonal pilot design.

[0011] To solve the above technical problems, the application provides the following technical scheme:

[0012] A pilot allocation method for a topology-aware de-cell large-scale MIMO system, the method comprising:

[0013] S100, establishing a system channel model to calculate the large-scale fading coefficient between APs and users ;

[0014] S200, constructing an inter-user interference topology graph based on the large-scale fading coefficient, and retaining strong interference connections through sparse processing;

[0015] S300, extracting loops in the topology graph, and screening significant loops according to length and weight;

[0016] S400, extracting significant loops in the topology graph and calculating weights; and assigning a gradual phase based on the loop weight ;

[0017] S500, generating an orthogonal phase modulation pilot matrix in combination with user channel strength priority ;

[0018] S600, calculating channel estimation parameters, user downlink reachable rate and MSE to maximize system performance.

[0019] Preferably, S100 comprises:

[0020] Establishing a channel model for the de-cell large-scale MIMO system, the channel characteristics are determined by path loss and shadow fading, and the large-scale fading coefficient between the mth AP and the kth user is calculated according to the formula , which is used to reflect the long-term attenuation characteristics of the channel:

[0021] ;

[0022] wherein, represents the path loss from the mth AP to the user k, and the 3GPP standard path loss model is adopted: the distance between the AP and the user satisfies:

[0023] If , ;

[0024] like , ;

[0025] like , ;

[0026] in, This represents the near-field distance (usually taken as 0.01km). This indicates the mid-to-near field distance (usually taken as 0.05km).

[0027] The shadow fading represents a Gaussian distribution with a mean of 0 and a standard deviation of 8 dB.

[0028] L represents the path loss constant, determined by the AP height. (15 m), User height (1.65 m), and communication frequency (1900 MHz), the calculation yields: .

[0029] Preferably, S200 includes:

[0030] Based on the channel correlation reflected by the large-scale fading coefficient, an inter-user interference topology map is constructed, focusing on key interference relationships:

[0031] S201, Interference Weight Definition: Define User With users The interference weight between them is Then, the intensity of interference between users is quantified: ;

[0032] in, To represent the minimum value, usually take ), used to avoid the denominator being 0, when hour, ;

[0033] S202. Topology Sparsity Processing: To reduce computational complexity, focus on key interfering relationships, retain strong interfering connections with weights above the 70th percentile, and reset the weights of weak interfering connections to 0. ;

[0034] in, Indicates user With users The larger the interference weight, the more severe the pilot pollution will be if users reuse the same pilot signal. This means that after pulling all K×K weights into a single vector, the 70th percentile is taken as the "strong interference" threshold;

[0035] S203. Construction of Undirected Topology Graph: Constructing an undirected user interference topology graph based on sparse interference weights. Among them, vertex set edge set ;

[0036] Among them, in the vertex set Each vertex in the edge set represents a user. Each edge in the diagram represents a strong interference relationship between a pair of users.

[0037] Preferably, S300 includes:

[0038] The significant loops that have the greatest impact on pilot contamination are extracted from the interference topology diagram to provide a basis for subsequent phase optimization.

[0039] S301. Loop Extraction: The loop basis algorithm is used to extract all closed loops in the topology graph, requiring the loop length to be specified. (That is, it must contain at least 3 users), thus obtaining the loop set cc;

[0040] S302. Significant Loop Screening: Calculate the average weight of each loop, reflecting the overall interference intensity of the loop.

[0041] ;

[0042] in, Indicates the first The average weight of each loop This represents the length of the i-th loop. Indicates the first in the loop One vertex, Indicates the first in the loop The next vertex of the first vertex. This represents the edge weight connecting two adjacent users in the sparsified topology graph.

[0043] Then, loops with average weights greater than the 60th percentile of non-zero weights are retained to obtain the set of significant loops. The corresponding weight vector is .

[0044] Preferably, S400 includes:

[0045] Pilot multiplexing efficiency is improved by canceling pilot interference from users within the loop through phase optimization.

[0046] S401, Initialize phase vector (Used to record the phase value for each user) and counting vector (used to record the number of times each user is covered by a significant loop);

[0047] S402, Weighted phase allocation: for each significant loop, allocate a gradual phase based on the loop weight, so that the pilot signals of users within the loop have a phase difference, thereby canceling the interference:

[0048] ;

[0049] ;

[0050] wherein, represents the th vertex of the loop , the greater the loop weight, the greater the contribution of the phase gradient to the interference suppression, represents the number of times the user is included in a significant loop, and the counter is incremented by one each time a user appears in a certain significant loop;

[0051] S403, Average phase optimization: for users covered by multiple loops, calculate the average phase to balance multi-loop interference and ensure the rationality of phase modulation: ;

[0052] wherein, represents the total phase accumulation value of the user , represents the number of times the user is included in a significant loop, and the average phase of the user is obtained by dividing the number of times the user is included in a significant loop.

[0053] Preferably, S500 comprises:

[0054] In combination with the orthogonal pilot and phase optimization, the final pilot matrix is generated to ensure the orthogonality of the pilot and the interference suppression capability;

[0055] S501, User priority sorting: calculate the total channel strength of each user , and sort in descending order to obtain a priority sequence , ensuring that strong channel users have priority in obtaining orthogonal pilot resources;

[0056] S502, Basic pilot index allocation: based on the pilot length , allocate a basic index of the orthogonal pilot to each user, ensuring that the first high-priority users obtain mutually non-repeating orthogonal pilots: ;

[0057] S503, the orthogonal phase modulation pilot generation: the basic pilot and the gradual phase are combined, the final pilot vector is generated and normalized, and the pilot power is ensured to be consistent: ;

[0058] Among them, represents the set of orthogonal pilot sequences, and satisfies , represents order unit matrix, represents the phase modulation factor.

[0059] Preferably, S600, comprising:

[0060] Based on the generated pilot matrix, the channel estimation parameter and the system performance index are calculated, and the pilot allocation optimization is completed;

[0061] S601, according to the formula, the channel estimation gain is calculated : ;

[0062] Among them, represents the pilot power, represents the pilot multiplexing interference term, represents the large scale fading coefficient vector of the mth AP and all users;

[0063] S602, the AP power allocation coefficient is calculated: ;

[0064] Ensure that the power allocation of each AP satisfies the total power constraint, while maximizing the system rate;

[0065] S603, user rate and MSE calculation:

[0066] According to the formula, the user downlink reachable rate is calculated: ;

[0067] Among them, represents the downlink transmission power, represents the pilot multiplexing interference matrix element;

[0068] According to the formula, the channel estimation mean square error is calculated, which is used to measure the channel estimation accuracy: .

[0069] Compared with the prior art, the beneficial effects achieved by the present application are:

[0070] The present application can improve channel estimation accuracy and system transmission rate without complex hardware modification, is significantly superior to the traditional pilot allocation scheme, and is suitable for various de-clustering large-scale MIMO communication scenarios, and provides an efficient pilot optimization solution for the next generation of mobile communication. BRIEF DESCRIPTION OF DRAWINGS

[0071] The accompanying drawings are included to provide a further understanding of the present application, and constitute a part of the specification, illustrate the present application, and are used to explain the present application together with the embodiments of the present application, and do not constitute a limitation on the present application. In the drawings:

[0072] Fig. 1 is a structural schematic diagram of a de-clustering MIMO system of the present application;

[0073] Fig. 2 is a rate CDF comparison diagram of different pilot allocation schemes;

[0074] Fig. 3 is a MSE CDF comparison diagram of different pilot allocation schemes. DETAILED DESCRIPTION

[0075] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0076] Please refer to Figs. 1-3 , the present application provides technical solutions:

[0077] Embodiment 1: The system parameters are set according to Table 1:

[0078] Table 1 System parameter setting table

[0079]

[0080] A topology-aware de-clustering large-scale MIMO system pilot allocation method, specifically comprising:

[0081] 1. Position generation: generate the position information of AP and user, including the central area and 8 adjacent extended areas, calculate the shortest distance between AP and user to avoid the influence of boundary effect.

[0082] 2. Shadow fading calculation: Based on the distance correlation between APs and users, the shadow fading vector is generated by Cholesky decomposition to meet the correlation With , the joint shadow fading matrix is constructed .

[0083] 3. Large-scale fading calculation: The large-scale fading coefficient matrix is calculated according to the segmentation rule , reflecting the channel attenuation characteristics between APs and users.

[0084] According to the formula: ;

[0085] Where, represents the path loss from the mth AP to user k, using the 3GPP standard path loss model: the distance between AP and user satisfies:

[0086] If , ;

[0087] If , ;

[0088] If , ;

[0089] Where, represents the near-field distance, represents the near-field distance;

[0090] represents the shadow fading, which obeys the Gaussian distribution with mean 0 and standard deviation 8 dB;

[0091] L represents the path loss constant, calculated by the AP height , user height , and communication frequency : .

[0092] 4. Topology construction and sparsification: Calculate the interference weight between users , achieve sparsification through 70% quantile threshold, and construct the undirected user interference topology graph.

[0093] Define the interference weight between user and user as , then quantify the interference intensity between users: ;

[0094] Where, represents the minimum value, used to avoid the denominator being 0, when Time, ;

[0095] Topology sparsification: keep the strong interference connections whose weights are higher than the 70th percentile, and set the weights of weak interference connections to 0: ;

[0096] where, represents the 70th percentile of all KxK weights after being pulled into a vector, which is taken as the "strong interference" threshold;

[0097] Undirected topology graph construction: based on the sparsified interference weights, construct an undirected user interference topology graph , where the vertex set , and the edge set ;

[0098] where each vertex in the vertex set represents a user, and each edge in the edge set represents a strong interference relationship between a pair of users.

[0099] 5. Significant loop extraction: extract the loops in the topology graph, and filter the significant loops according to the length and weight (higher than the 60th percentile of non-zero weights).

[0100] Loop extraction: use the loop-based algorithm to extract all closed loops in the topology graph, requiring the loop length , to obtain the loop set cc;

[0101] Significant loop filtering: calculate the average weight of each loop, reflecting the overall interference strength of the loop:

[0102] ;

[0103] where, represents the average weight of the i-th loop, represents the length of the i-th loop, represents the j-th vertex in the loop, represents the next vertex of the j-th vertex in the loop, represents the edge weight connecting two adjacent users in the sparsified topology graph; Then, the loops with an average weight greater than the 60th percentile of non-zero weights are kept, and the significant loop set is obtained, and the corresponding weight vector is .

[0104] .

[0105] ​​6. Phase distribution: distribute the phase based on the significant loop weight, calculate the average phase for the users covered by multiple loops;

[0106] Initialize phase vector and counter vector ;

[0107] Weighted phase distribution: distribute the phase based on the loop weight for each significant loop, make the pilot signals of the users in the loop have phase difference, so as to offset the interference:

[0108] ;

[0109] ;

[0110] wherein, represents the i-th vertex of the loop , represents the user , the number of times the user is included in the significant loop, and the counter of the user is increased by one each time the user appears in a certain significant loop;

[0111] Average phase optimization: for the users covered by multiple loops, calculate the average phase to balance the multi-loop interference and ensure the rationality of the phase modulation: ;

[0112] wherein, represents the total phase accumulation value of the user , represents the number of times the user is included in the significant loop, and the average phase of the user is obtained by dividing the number of times.

[0113] 7. Pilot generation: sort the users in descending order of channel intensity, distribute the basic pilot index, and generate the phase modulation pilot matrix and normalize.

[0114] User priority sorting: calculate the total channel intensity of each user , sort in descending order to obtain the priority sequence , and ensure that the strong channel users obtain the orthogonal pilot resources in priority; Basic pilot index distribution: based on the pilot length

[0115] , distribute the basic index of the orthogonal pilot for each user, and ensure that the first high-priority users obtain non-repeated orthogonal pilots: ;

[0116] ​Orthogonal phase modulation pilot generation: combine the base pilot with the gradient phase, generate the final pilot vector and perform normalization processing to ensure the consistency of pilot power ;

[0117] where, denotes the set of orthogonal pilot sequences, satisfying , denotes the order identity matrix, denotes the phase modulation factor.

[0118] 8. Performance calculation: calculate the channel estimation gain of each user , AP power allocation coefficient , downlink achievable rate and MSE, and compare with the traditional scheme;

[0119] According to the formula, the channel estimation gain is calculated: ;

[0120] where, denotes the pilot power, denotes the pilot multiplexing interference term, denotes the large-scale fading coefficient vector of the mth AP and all users;

[0121] Calculate the AP power allocation coefficient: ; ensure that the power allocation of each AP satisfies the total power constraint, while maximizing the system rate;

[0122] According to the formula, the downlink achievable rate of the user is calculated: ;

[0123] where, denotes the downlink transmission power, denotes the pilot multiplexing interference matrix element;

[0124] According to the formula, the channel estimation mean square error is calculated to measure the channel estimation accuracy: .

[0125] Simulation result analysis: as shown in Fig. 2 and Fig. 3 , the rate and accuracy of the pilot allocation method of the topology-aware de-cell large-scale MIMO system are better than those of the several traditional schemes mentioned above.

[0126] Finally, it should be noted that the above only describes the preferred embodiments of the present application and is not intended to limit the present application. Although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art will appreciate that the technical solutions described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalent ones. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A pilot allocation method for a topology-aware decellularized massive MIMO system, characterized in that: The method includes: S100. Establish a system channel model and calculate the large-scale fading coefficient between the AP and the user. ; S200: Construct an inter-user interference topology graph based on large-scale fading coefficients, and retain strong interference connections through sparsification. S300. Extract loops from the topology graph, sorted by length. Significant loops were selected based on weights; S400. Extract significant loops from the topology graph and calculate their weights; assign gradient phases based on the loop weights. ; S500, combined with user channel strength priority, generates an orthogonalized phase modulation pilot matrix. ; The S600 calculates channel estimation parameters, user downlink reachable rate, and MSE to maximize system performance.

2. The pilot allocation method for a topology-aware decellularized massive MIMO system as described in claim 1, characterized in that, The S100 includes: Calculate the large-scale fading coefficient between the m-th AP and the k-th user using the formula. : ； in, This represents the path loss from the m-th AP to user k, using the 3GPP standard path loss model: the distance between the AP and the user. satisfy: like , ; like , ; like , ; in, Indicates near-field distance. Indicates the mid-near field distance; The shadow fading represents a Gaussian distribution with a mean of 0 and a standard deviation of 8 dB. L represents the path loss constant, determined by the AP height. User height and communication frequencies The calculation yields: .

3. The pilot allocation method for a topology-aware decellularized massive MIMO system as described in claim 1, characterized in that, The S200 includes: S201, Interference Weight Definition: Define User With users The interference weight between them is Then, the intensity of interference between users is quantified: ; in, To represent the minimum value, used to avoid the denominator being 0, when hour, ; S202. Topology Sparsity Processing: To reduce computational complexity, focus on key interfering relationships, retain strong interfering connections with weights above the 70th percentile, and reset the weights of weak interfering connections to 0. ; in, Indicates user With users Interference weights between This means that after pulling all K×K weights into a single vector, the 70th percentile is taken as the "strong interference" threshold; S203. Construction of Undirected Topology Graph: Constructing an undirected user interference topology graph based on sparse interference weights. Among them, vertex set edge set ; Among them, in the vertex set Each vertex in the edge set represents a user. Each edge in the diagram represents a strong interference relationship between a pair of users.

4. The pilot allocation method for a topology-aware decellularized massive MIMO system as described in claim 1, characterized in that, The S300 includes: S301. Loop Extraction: The loop basis algorithm is used to extract all closed loops in the topology graph, requiring the loop length to be specified. This yields the set of loops cc; S302. Significant Loop Screening: Calculate the average weight of each loop, reflecting the overall interference intensity of the loop. ； in, Indicates the first The average weight of each loop This represents the length of the i-th loop. Indicates the first in the loop One vertex, Indicates the first in the loop The next vertex of the first vertex. This represents the edge weight connecting two adjacent users in the sparsified topology graph. Then, loops with average weights greater than the 60th percentile of non-zero weights are retained to obtain the set of significant loops. The corresponding weight vector is .

5. The pilot allocation method for a topology-aware decellularized massive MIMO system as described in claim 1, characterized in that, The S400 includes: S401, Initialize phase vector and counting vector ; S402, Weighted Phase Allocation: For each significant loop, a gradually changing phase is allocated based on the loop weight, causing a phase difference in the pilot signals of users within the loop, thereby canceling out interference. ； ； in, Indicates a loop The One vertex, Indicates user The counter for the number of times a user appears in a salient loop is incremented by one each time a user appears in a salient loop. S403, Average Phase Optimization: For users covered by multiple loops, the average phase is calculated to balance multi-loop interference and ensure the rationality of phase modulation. ; in, Indicates user Total phase accumulation value, Indicates user The number of times a data point is contained in a significant loop is divided to obtain the user's score. The average phase.

6. The pilot allocation method for a topology-aware decellularized massive MIMO system as described in claim 1, characterized in that, The S500 includes: S501, User Priority Ranking: Calculate the total channel strength for each user. ,according to Priority sequence obtained by sorting in descending order This ensures that users with strong channels have priority access to orthogonal pilot resources; S502, Basic Pilot Index Assignment: Based on Pilot Length Assign a base index for orthogonal pilots to each user to ensure the preceding Each high-priority user receives a unique orthogonal pilot signal: ; S503, Quadrature Phase Modulation Pilot Generation: Combines the basic pilot with the gradient phase to generate the final pilot vector and performs normalization processing to ensure consistent pilot power. ; in, Describes a set of orthogonal pilot sequences that satisfy... , express An identity matrix of order 1. This represents the phase modulation factor.

7. The pilot allocation method for a topology-aware decellularized massive MIMO system as described in claim 1, characterized in that, The S600 includes: S601. Calculate the channel estimation gain according to the formula. : ; in, Indicates pilot power. This indicates the pilot multiplexing interference term. This represents the large-scale fading coefficient vector of the m-th AP and all users; S602, Calculate AP power allocation factor: ; Ensure that the power allocation of each AP meets the total power constraint while maximizing the system rate; S603, User Rate and MSE Calculation: Calculate the user's downlink reachable speed using the formula: ; in, Indicates downlink transmission power. Indicates the elements of the pilot multiplexing interference matrix; The mean square error of channel estimation is calculated according to the formula and is used to measure the accuracy of channel estimation: .

Images