A MIMO pilot allocation method, device, apparatus and storage medium

Abstract

This invention relates to the field of communication technology and discloses a MIMO pilot allocation method, apparatus, device, and storage medium. The method includes: acquiring interference strength information, which includes the interference strength generated between each user when multiple users reuse the same pilot resource; dividing each user into multiple pilot reuse groups that satisfy the optimization objective based on an improved Markov chain Monte Carlo-Simulated Annealing optimization algorithm; the number of pilot reuse groups is the same as the number of pilot resources of the base station, and users in the same pilot reuse group are allocated to use the same pilot resource of the base station; the optimization objective is determined based on the interference strength information to minimize the interference strength between users in the same pilot reuse group and maximize the interference strength between users in different pilot reuse groups. This invention suppresses pilot pollution and improves channel estimation accuracy and system transmission performance by acquiring interference strength information between users reusing the same pilot and dividing users into groups with the same number of pilot resources.

Description

Technical Field

[0001] This invention relates to the field of communication technology, and specifically to a MIMO pilot allocation method, apparatus, device, and storage medium. Background Technology

[0002] With the rapid development of fifth-generation and future communication technologies, Multiple-Input Multiple-Output (MIMO) technology is widely used in wireless communication systems to improve system capacity and spectral efficiency. In multi-user MIMO systems, before the base station officially transmits data to user equipment, it needs to accurately estimate the wireless channel between the base station and each user equipment using pilot signals. The accuracy of the channel estimation directly affects the system's transmission performance.

[0003] However, limited by coherence time and spectrum resources, the number of pilot resources available for channel estimation in the system is limited. In multi-user scenarios, different users often need to reuse pilot sequences, leading to pilot interference, or pilot pollution, which reduces channel estimation accuracy and consequently affects beamforming and data transmission performance, becoming one of the key factors restricting the performance improvement of MIMO systems. Currently, traditional pilot allocation schemes such as fixed pilot allocation, random pilot allocation, or heuristic allocation schemes are relatively simple to implement but suffer from low allocation efficiency and slow convergence speed. Meanwhile, relatively complex schemes based on deep learning have weak interpretability, are inconvenient to maintain, and require a very large amount of data to train deep learning models, making deployment difficult. Summary of the Invention

[0004] This invention provides a MIMO pilot allocation method, apparatus, device, and storage medium to solve the problems of low search efficiency and difficulty in escaping local optima in existing pilot allocation methods when solving interference optimization problems.

[0005] In a first aspect, the present invention provides a MIMO pilot allocation method, which obtains interference strength information, including the interference strength generated between each user when multiple users reuse the same pilot resource; based on an improved Markov chain Monte Carlo-simulated annealing optimization algorithm, each user is divided into multiple pilot reuse groups that satisfy the optimization objective; the number of pilot reuse groups is the same as the number of pilot resources of the base station, and users in the same pilot reuse group are assigned to use the same pilot resource of the base station; the optimization objective is determined based on the interference strength information, so as to minimize the interference strength between users in the same pilot reuse group and maximize the interference strength between users in different pilot reuse groups.

[0006] In one optional implementation, the improved Markov chain Monte Carlo-simulated annealing optimization algorithm, which divides users into several pilot reuse groups that satisfy the optimization objective, includes: generating an objective function for the optimization objective; constructing a discriminant using a temperature parameter and the objective function, wherein the temperature parameter is adjusted by the simulated annealing algorithm, and the discriminant is used to determine whether the adjusted users in the pilot reuse groups are closer to the optimization objective; initializing and assigning each user to each pilot reuse group; adjusting the users in each pilot reuse group, and determining whether the adjustment scheme is closer to the optimization objective using the discriminant; if the adjustment scheme is closer to the optimization objective, retaining the current adjustment; if the adjustment scheme is far from the optimization objective, determining whether to retain the current adjustment based on the output result of the discriminant; iterating back to the step of adjusting the users in each pilot reuse group until a stopping condition is met, and outputting the pilot reuse group allocation scheme.

[0007] In one optional implementation, the discriminant outputs a value greater than or equal to 1 when it is closer to the optimization target, and outputs a value less than 1 when it is far from the optimization target. If the adjustment scheme is far from the optimization target, the process of determining whether to retain the current adjustment based on the output result of the discriminant includes: generating a random number; if the random number is less than the output result of the discriminant, then abandoning the current adjustment; if the random number is greater than or equal to the output result of the discriminant, then retaining the current adjustment.

[0008] In one optional implementation, the objective function is:

[0009]

[0010] In the formula, This indicates the pilot group label for user v. For the set of all user pairs, Let be the interference strength when user i and user j share the same pilot signal. For indicator functions, The discriminant is:

[0011] In the formula, This indicates the original allocation scheme. This represents the new allocation scheme, in which , Denotes the objective function, It is the temperature parameter, This indicates that the user set was selected when the new allocation scheme was adjusted to the original allocation scheme. The probability, This indicates that a user set was selected when adjusting from the original allocation scheme to the new allocation scheme. The probability, Indicates user From the original group Switch to new group The probability, Indicates user From the original group Switch to new group The probability of.

[0012] In one optional implementation, the step of adjusting the temperature parameter using the simulated annealing algorithm includes: decreasing the temperature parameter by a preset step size after each preset number of iterations.

[0013] In one optional implementation, adjusting the users in each pilot multiplexing group includes: determining the number of users to be adjusted; determining the adjustment score for each user, the adjustment score being obtained by summing a first score and a second score, the first score being determined by the sum of the interference intensity of the current user and all other users, and the second score being determined by the change in the objective function after the current user is adjusted to the new allocation scheme; and determining the target user with the highest adjustment score from among the users based on the number of users for adjustment.

[0014] In one optional implementation, determining the number of users to be adjusted this time includes updating the number of users to be adjusted this time based on the output result of the discriminant after the last adjustment and the number of users in the last adjustment.

[0015] Secondly, the present invention provides a MIMO pilot allocation device, the device comprising: an interference strength acquisition module for acquiring interference strength information, the interference strength information including the interference strength generated between each user when multiple users reuse the same pilot resource; and a pilot partitioning module for partitioning each user into multiple pilot reuse groups that satisfy an optimization objective based on an improved Markov chain Monte Carlo-simulated annealing optimization algorithm; wherein the number of pilot reuse groups is the same as the number of pilot resources of the base station, and users in the same pilot reuse group are allocated to use the same pilot resource of the base station; the optimization objective is determined based on the interference strength information to minimize the interference strength between users in the same pilot reuse group and maximize the interference strength between users in different pilot reuse groups.

[0016] Thirdly, the present invention provides an electronic device, comprising: a memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the computer instructions to perform a MIMO pilot allocation method of the first aspect or any corresponding embodiment described above.

[0017] Fourthly, the present invention provides a computer-readable storage medium storing computer instructions for causing a computer to perform a MIMO pilot allocation method according to the first aspect or any corresponding embodiment described above.

[0018] This invention provides a quantitative basis for pilot optimization allocation by acquiring interference intensity information generated when users reuse the same pilot. Then, based on an improved Markov chain Monte Carlo-simulated annealing optimization algorithm, each user is divided into multiple pilot reuse groups with the same number of pilot resources as the base station. This allows users in the same group to reuse the same pilot resources. Based on the interference intensity information, optimization objectives are set, ultimately achieving the technical effect of minimizing the interference intensity between users in the same pilot reuse group and maximizing the interference intensity between users in different pilot reuse groups. This suppresses pilot pollution and improves channel estimation accuracy and system transmission performance. Attached Figure Description

[0019] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0020] Figure 1 This is a schematic diagram of the first type of MIMO pilot allocation method according to an embodiment of the present invention; Figure 2 This is a base station-user schematic diagram of the MIMO pilot allocation method according to an embodiment of the present invention; Figure 3 This is a schematic diagram of pilot contamination and spectral interference in the MIMO pilot allocation method according to an embodiment of the present invention; Figure 4 This is an undirected weighted schematic diagram of the MIMO pilot allocation method according to an embodiment of the present invention; Figure 5 This is a schematic diagram of Markov chain space transfer in the MIMO pilot allocation method according to an embodiment of the present invention; Figure 6 This is a structural block diagram of an apparatus for a MIMO pilot allocation method according to an embodiment of the present invention; Figure 7 This is a schematic diagram of the hardware structure of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] It is understood that before using the technical solutions disclosed in the various embodiments of the present invention, users should be informed of the types, scope of use, and usage scenarios of the personal information involved in the present invention and their authorization should be obtained in accordance with relevant laws and regulations through appropriate means.

[0023] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0024] While multiple-input multiple-output (MIMO) technology is widely used in wireless communication systems to improve system capacity and spectral efficiency, its performance is highly dependent on the accuracy of channel estimation. In multi-user MIMO systems, base stations need to estimate the channel using pilot signals before data transmission. However, limited by coherence time and spectral resources, the number of pilot resources available for channel estimation is limited. Different users often need to reuse pilot sequences, which leads to pilot pollution, severely reducing channel estimation accuracy and affecting beamforming and data transmission performance, becoming a key bottleneck restricting the performance improvement of MIMO systems. To address pilot pollution, pilot allocation schemes in related technologies mainly suffer from two types of defects: one is traditional schemes such as fixed pilot allocation, random pilot allocation, or heuristic allocation. Although relatively simple to implement, these schemes are inefficient and have slow convergence speeds, making it difficult to achieve global optimization in complex and variable user interference environments. The other is complex schemes based on deep learning. While these schemes theoretically have performance potential, their interpretability is weak, the amount of data required for model training is enormous, leading to deployment difficulties and maintenance inconvenience, making them difficult to promote and apply in practical communication systems. Therefore, how to optimize the allocation of pilot resources while maintaining high efficiency and interpretability has become an urgent technical problem to be solved.

[0025] According to an embodiment of the present invention, a MIMO pilot allocation method embodiment is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0026] This embodiment provides a MIMO pilot allocation method, which can be used in the aforementioned electronic devices. Figure 1 This is a flowchart of a MIMO pilot allocation method according to an embodiment of the present invention, such as... Figure 1 As shown, the process includes the following steps: Step S101: Obtain interference strength information, which includes the interference strength generated between each user when they reuse the same pilot resource.

[0027] Step S102: Based on the improved Markov chain Monte Carlo-simulated annealing optimization algorithm, each user is divided into multiple pilot reuse groups that meet the optimization objective. The number of pilot resources in each pilot reuse group is the same as that in the base station. Users in the same pilot reuse group are assigned to use the same pilot resource of the base station. The optimization objective is determined based on interference strength information to minimize the interference strength between users in the same pilot reuse group and maximize the interference strength between users in different pilot reuse groups.

[0028] Specifically, firstly, it is necessary to obtain the interference intensity information between different users in the entire system to form the basis data for subsequent optimization and partitioning. The interference intensity information can quantify the potential conflict between users, representing the strength of mutual interference between any two different users when they are assigned to use the same pilot resources. The interference intensity information usually exists in the form of a matrix or network diagram, where each value corresponds to the interference weight between a pair of users.

[0029] Next, the core task of the whole method lies in how to reasonably divide all users into several pilot reuse groups based on the improved optimization algorithm. The concept of pilot reuse groups is strictly corresponding to the number of pilot resources available to the base station. That is to say, the number of user groups needs to be divided according to the number of orthogonal pilot resources the base station has. Users in each group will be forced to share the same pilot resources. For example, suppose a cell base station is configured with a total of 4 orthogonal pilots. Then all users need to be divided into exactly 4 different groups. Users in group 1, group 2, group 3, and group 4 each use the same pilot. During the division process, the optimization objective of the algorithm is driven entirely by the interference intensity information obtained earlier. The objective includes two aspects: First, the interference intensity between users divided into the same pilot reuse group should be as low as possible, because sharing the same pilot will directly lead to pilot pollution. Low interference means that even if resources are shared, the impact on system performance is limited. Second, the interference intensity between users divided into different pilot reuse groups should be as high as possible, so that the users with the strongest interference are separated, thereby eliminating the strongest interference source globally.

[0030] To achieve the partitioning, this invention proposes a novel optimization algorithm that integrates the Markov chain Monte Carlo method and the simulated annealing mechanism. This algorithm performs an iterative search in the solution space and converges to a user grouping scheme that simultaneously satisfies the dual objectives of "minimizing intra-group interference" and "maximizing inter-group interference" through random state transitions and probabilistic judgments.

[0031] In one example, a system and mathematical model are established in a single-cell MU-MIMO scenario, such as... Figure 2 As shown, consider L cells, such as the 7 hexagonal cells in the diagram. Each cell has a base station at its center, and several users are distributed around each base station. Under the constraint of a finite number of reusable pilot sequences, severe pilot pollution and spectrum interference will occur when different users reuse the same pilot. For example... Figure 3 As shown, multiple colors represent multiple sets of reusable pilots. Users using the same pilot (i.e., the same color) will cause mutual interference, which will occur during the channel estimation process. This is known as pilot pollution, which is a factor that limits the improvement of system performance.

[0032] To characterize the interference relationships between users, the "co-channel interference intensity" is abstracted as an undirected weighted graph, such as... Figure 4 As shown

[0033] The vertex set Indicates user (e.g.) Figure 4 In the context of U1~U5), W includes edge weights. Indicates user The interference intensity generated when the same pilot is reused between devices, for example Figure 4 In ~ E represents… Pilot assignment is formalized as a maximum k-cut problem: assigning a label to each user. This is used to indicate which group each user is assigned to, so that the sum of cross-group edge weights is maximized and interference within the same group is minimized.

[0034] This invention, through the introduction of interference intensity information and the optimization allocation of pilot resources based on this information, assigns users with high mutual interference intensity to different pilot reuse groups. This fundamentally cuts off the impact of strong interference sources on channel estimation, significantly reducing pilot pollution and enabling base stations to obtain more accurate channel state information, thus laying a reliable foundation for data transmission. Furthermore, an improved Markov chain Monte Carlo-Simulated Annealing optimization algorithm achieves a dual optimization objective: minimizing interference between users within the same pilot group and maximizing interference isolation between users in different pilot groups. This suppresses co-channel interference among multiple users, improves the signal-to-interference-plus-noise ratio, and achieves higher system throughput and spectral efficiency with limited spectrum resources. This fusion algorithm combines the stochastic state transition characteristics of the Markov chain Monte Carlo method with the probabilistic jump capability of the simulated annealing mechanism, enabling it to search within a large solution space, avoiding local optima, and ensuring the algorithm's adaptability and robustness in different network environments. Furthermore, by abstracting complex wireless environment interference into the form of an undirected weighted graph, the physical layer interference problem is transformed into the maximum k-cut problem in graph theory. This makes the algorithm independent of specific channel models or antenna configurations, and the algorithm framework has good scalability and applicability regardless of changes in the number of base station antennas or users.

[0035] In one optional implementation, step S102 includes: Step a1: Generate the objective function for the optimization objective; Step a2: Construct a discriminant using temperature parameters and objective function. The temperature parameters are adjusted using the simulated annealing algorithm. The discriminant is used to determine whether the adjustments made by users in the pilot multiplexing group are closer to the optimization objective. Step a3: Initialize and assign each user to their respective pilot multiplexing group; Step a4: Adjust the users in each pilot multiplexing group, and determine whether the adjustment scheme is closer to the optimization goal through discriminant analysis; Step a5: If the adjusted plan is closer to the optimization goal, then retain this adjustment; Step a6: If the adjustment scheme is far from the optimization target, determine whether to retain the current adjustment based on the discriminant output result; Step a7: Iterate through the steps of adjusting users in each pilot multiplexing group until the stopping condition is met, and output the pilot multiplexing group allocation scheme.

[0036] Specifically, the entire optimization process is broken down into several specific steps. First, it is necessary to construct an objective function to quantify the optimization goal. This function is essentially a mathematical model that transforms the dual requirements of "minimizing the interference intensity between users within the same pilot reuse group" and "maximizing the interference intensity between users in different pilot reuse groups" into a calculable numerical expression.

[0037] For example, the objective function can be designed as the weighted sum of interference intensity for all users within a group, minus the weighted sum of interference intensity for all users between groups, and the search direction can be guided by minimizing or maximizing this function.

[0038] The algorithm then introduces an annealing mechanism, which uses the temperature parameter and the newly established objective function to construct a discriminant. The temperature parameter is not physical temperature, but rather a control variable that gradually decays over time in the simulated annealing algorithm, responsible for adjusting the probability of the algorithm accepting suboptimal solutions. The discriminant checks each user adjustment to determine the direction of change of the adjusted group relative to the optimization objective. Specifically, if the objective function value output by the discriminant becomes more convergent after adjustment, approaching the final target value, the result of the discriminant indicates that the adjusted group is better. If the adjusted function value becomes divergent, moving away from the final target value, the discriminant will not directly reject it, but will accept a suboptimal solution with a certain probability, based on the current temperature parameter. This probabilistic acceptance mechanism allows the algorithm to escape local optima during the search process.

[0039] The first step in the formal operation is to initialize and assign users to various pilot multiplexing groups. The initial state can be generated randomly. For example, in a scenario with 4 pilot resources and 16 users, the system first randomly assigns the 16 users to 4 groups, with exactly 4 user devices in each group. The initial grouping may be far from the optimal solution and requires repeated adjustments later.

[0040] The adjustment process forms the core loop of the entire algorithm. In each iteration, the algorithm attempts to fine-tune the users in each pilot multiplexing group, such as randomly swapping users in two different groups or moving a user from one group to another. After each adjustment, the system uses the previously constructed discriminant to evaluate the new solution: if the new solution makes the objective function better, then the adjustment is accepted, and the new grouping replaces the old grouping.

[0041] However, not every adjustment brings direct improvement. If the adjusted solution causes the objective function value to diverge, the algorithm will not simply discard the adjusted solution. In this case, the discriminant will output a judgment result based on the current temperature parameter: combining the principle of simulated annealing, if the temperature parameter is high, the algorithm has a higher probability of accepting the inferior solution, allowing user groups to temporarily "retreat"; if the temperature has dropped significantly, the probability of accepting the inferior solution also decreases. This mechanism enables the algorithm to have strong exploration capabilities at higher temperatures and gradually converge as the temperature decreases.

[0042] The entire process of adjustment, judgment, and retention is repeated continuously, forming an iterative loop. After each round of adjustment, the temperature parameter is usually reduced according to a predetermined strategy. The iterative process continues until a preset stopping condition is met. The stopping condition can be that the temperature parameter drops to near zero, the objective function value no longer changes significantly in multiple consecutive iterations, or the preset maximum number of iterations is reached. When the loop terminates, the user grouping scheme finally output by the algorithm is the pilot reuse group allocation result that satisfies the dual objectives of minimizing intra-group interference and maximizing inter-group interference after random search and probability judgment.

[0043] This invention improves algorithm performance during pilot allocation through an optimization mechanism. By constructing a quantified objective function with dual optimization goals, the physical objectives of "minimizing intra-group interference" and "maximizing inter-group interference" are transformed into clear mathematical guidance, providing a precise evaluation benchmark for subsequent iterative searches. The introduction of a discriminant with a temperature parameter cleverly integrates the time-dependent decay of temperature parameters in simulated annealing, giving the algorithm high fault tolerance in the early stages of iteration. It can accept adjustments that worsen the objective function with a certain probability, thus avoiding the pitfalls of traditional greedy algorithms that easily get trapped in local optima. As iteration progresses, the temperature gradually decreases, reducing the probability of the algorithm accepting inferior solutions and gradually converging towards the global optimum. Based on the initial allocation, the algorithm repeatedly adjusts users within the group and dynamically evaluates them using the discriminant. Each adjustment, regardless of optimization, is scientifically determined. Adjustments in the optimization direction are directly retained to accelerate convergence, while adjustments far from the optimization goal are probabilistically retained based on the temperature parameter. This gives the algorithm the ability to fully explore the solution space, ensuring both the correctness of the search direction and maintaining population diversity. The entire iterative process, through the gradual annealing of temperature parameters and the continuous action of discriminant, enables the user grouping scheme to continuously evolve until the final scheme is output when the stopping condition is met. This scheme not only achieves the design goal of minimizing the interference intensity between users in the same group and maximizing the interference intensity between users in different groups, but also ensures the stability and reliability of the allocation results through rigorous mathematical logic. It provides a robust technical path for multi-user MIMO systems to combat pilot pollution and improve spectral efficiency in actual deployment.

[0044] In one optional implementation, the output is greater than or equal to 1 when the discriminant is closer to the optimization target, and less than 1 when the discriminant is far from the optimization target. Step a6 includes: Step a61: Generate random numbers.

[0045] In step a62, if the random number is less than the output of the discriminant, then abandon this adjustment.

[0046] Step a63: If the random number is greater than or equal to the output of the discriminant, then retain this adjustment.

[0047] Specifically, when a user's adjustment causes the new solution to deviate further from the optimization target compared to the original solution, the discriminant will output a value less than 1, representing the probability of accepting the worse solution at the current temperature. To decide whether to retain the adjustment, the algorithm introduces a random decision: first, a random number uniformly distributed between 0 and 1 is generated, and then this random number is compared with the output of the discriminant. If the random number is less than the output value of the discriminant, it means that the adjustment is accepted, even though it worsens the solution, and the algorithm will retain the adjustment and replace the original solution with the new one; conversely, if the random number is greater than or equal to the output value of the discriminant, the adjustment is abandoned, and the original solution remains unchanged.

[0048] This invention introduces a random decision mechanism. When user adjustments cause the solution to deviate from the optimization target, the algorithm accepts the inferior solution with a probability value output by the discriminant, thereby avoiding the algorithm from getting trapped in local optima. This mechanism, combined with temperature parameters, allows the algorithm to tolerate inferior solutions with a higher probability during high-temperature stages, enhancing global exploration capabilities. During low-temperature stages, the acceptance probability decreases, ensuring stable convergence of the algorithm and achieving a dynamic balance between exploration and utilization. Under the premise of ensuring the detailed balance conditions of the Markov chain, the search efficiency of the solution space is improved, ultimately resulting in a better pilot allocation scheme.

[0049] In one alternative implementation, the objective function is:

[0050] In the formula, This indicates the pilot group label for user v. For the set of all user pairs, Let be the interference strength when user i and user j share the same pilot signal. It is an indicator function; The discriminant is:

[0051] In the formula, This indicates the original allocation scheme. This represents the new allocation scheme, in which , Describe the objective function. It's a temperature parameter. This indicates that the user set was selected when the new allocation scheme was adjusted to the original allocation scheme. The probability, This indicates that a user set was selected when adjusting from the original allocation scheme to the new allocation scheme. The probability, Indicates user From the original group Switch to new group The probability, Indicates user From the original group Switch to new group The probability of that. Specifically, the objective function described above is constructed to maximize the sum of interference intensities between all cross-group user pairs.

[0052] This represents the set of pilot group labels for all users, where User The assigned labels, such as group 1, group 2, or group 3, for the entire set. This represents the complete allocation plan. It is the set of all user pairs, each pair of users They are all considered in the set. It is a known value, representing what would happen if the user... and users The magnitude of interference that will occur between devices using the same pilot resource is determined through measurement or calculation before the algorithm runs. It is an indicator function, when and The function value is 1 if the two users are assigned to different pilot groups, and 0 otherwise. The summation symbol represents iterating through all user pairs and summing the values ​​of each pair. Added together, the outermost layer This means finding a solution among all possible allocation schemes that maximizes the cumulative result.

[0053] Traverse all user pairs Check each user individually to see if they are in different groups. If so, adjust the interference strength between them. The sum is added to the total, and the ultimate optimization goal is to maximize the sum to reduce pilot interference. Pilot interference only occurs between users using the same pilot resource, that is, users within the same group will interfere with each other. If two users are assigned to different groups and use different pilots, there will be no interference. Therefore, by splitting user pairs with high interference intensity into different groups, they will no longer affect each other. The objective function only accumulates user pairs that cross groups, maximizing the objective function value to separate user pairs with strong interference as much as possible. When all strong interference edges cross groups, the remaining user pairs with weak interference intensity within the same group will have their overall pilot interference reduced to a minimum.

[0054] The algorithm then enters the iteration phase. Since the feasible solution space grows exponentially with the number of users, direct solution is not feasible. Therefore, a stochastic search method based on Markov chain Monte Carlo is introduced, using a stochastic process (Markov chain) to implement transitions in the transition space. A specific example of a transition chain is shown below. Figure 5 As shown in the diagram, the two circles S1 and S2 represent two states of the system, i.e., two user scheduling strategies. The arrows and numbers represent the transition probabilities from one state to another. State S1 is a self-loop (S1→S1): probability 0.2, meaning that if you are currently in state S1, the probability of remaining in S1 next time is 20%. Transitioning to S2 (S1→S2): probability 0.8, meaning that if you are currently in state S1, the probability of switching to S2 next time is 80%. Transitioning from state S2 to S1 (S2→S1): probability 0.6, meaning that if you are currently in state S2, the probability of switching to S1 next time is 60%. Self-loop S2→S2: probability 0.4, meaning that if you are currently in state S2, the probability of remaining in S2 next time is 40%. The goal of this framework is to construct... A Markov chain in state , such that it is stationary. It is exactly the following distribution: in Temperature parameter controls the steepness of the distribution; For distribution The constants are normalized. This process will... Larger convert The smaller the value, the higher the probability that a high-quality solution with low energy (i.e., high objective function value) will be accessed.

[0055] Then, the Metropolis-Hastings sampling mechanism is used to generate a discriminant by driving state transitions through proposal distribution and acceptance-rejection rules. Specifically, from the current state... Departure, based on the proposed distribution Generate candidate states Then, with the probability of acceptance Decide whether to transfer to This acceptance probability stems from detailed equilibrium conditions. To ensure the stable distribution of the chain as the target distribution .

[0056] In the actual implementation, the design of the proposal distribution is crucial. Suppose a single adjustment involves a set of users. The proposal probability can then be decomposed into: in In the current plan Select user set The probability, For users From the original group Move to new group The probability of a reverse proposal is... Similarly, the probability of a reverse proposal is... Substituting the acceptance probability, we obtain the specific discriminant:

[0057] In other words, when adjusting user groups, if the objective function If it becomes larger, then The calculated value will be greater than 1, according to the discriminant above. calculate If the final output value is 1, then the user adjustment will be accepted directly.

[0058] When adjusting user groups, if the objective function If it becomes smaller, that is, if the adjustment scheme deviates from the optimization target, then... The calculated value will be less than 1, according to the discriminant above. The final output value is less than 1. Then, based on this value less than 1, a decision is made on whether to accept the user's adjustment. For example, assuming the output value is 0.6, the control system randomly generates a random number. If the random number is greater than or equal to 0.6, the user's adjustment is accepted; if the random number is less than 0.6, the user's adjustment is abandoned. This technique helps prevent the user's adjustment process from getting stuck in local optima.

[0059] In the discriminant calculation process described above, to improve search efficiency, the algorithm further integrates the concept of simulated annealing. Temperature parameters in Perform annealing scheduling. Initial temperature. The default value is set to a relatively large value, at which point the differences in acceptance probabilities among the various solutions are small, allowing the chain to extensively explore the solution space. In other words, when the user adjusts the solution away from the optimization objective, although the discriminant... The output value is less than 1, but it can output a larger value closer to 1, ensuring that the adjustment scheme is preserved. In the early stages of adjustment, it is easier to accept poor adjustment results, allowing the user grouping scheme to try a wider range of possibilities and preventing the user grouping from getting trapped in local optima. As iteration progresses, the temperature gradually decreases according to a predetermined strategy, and the stable distribution of the chain gradually concentrates in the low-energy region, eventually converging to a near-global optimum in the low-temperature region. In other words, as the number of adjustments increases, the user grouping scheme should have moved far away from the local optimum, and the current adjustment direction is theoretically very close to the global optimum. Therefore, as the number of adjustments increases, the temperature parameter... Gradually decrease, so that when the adjustment scheme deviates from the optimization target, not only the discriminant... The output value is less than 1, and it is made to be closer to a smaller value than 0, so that bad adjustment schemes are discarded. In the later stage of scheme adjustment, the user grouping scheme is guaranteed to stably approach the global optimum, and bad user grouping is avoided.

[0060] The entire iterative process is controlled by a double-layered loop: the outer loop is responsible for temperature updates, while the inner loop executes at a fixed temperature. After several Metropolis-Hastings samplings, when the temperature drops to a preset threshold or the optimal solution is not improved after multiple iterations, the algorithm terminates and outputs the historical best allocation scheme, which is the pilot reuse group partitioning result that satisfies the optimization objective.

[0061] Furthermore, the adjustment operation itself has an adaptive mechanism in each inner iteration. The algorithm first adjusts the parameters based on the current step size. Determine the number of users to be adjusted this time (i.e., whether it's adjusting 1, 2, or more user groups). This parameter is dynamically updated based on historical acceptance frequency: if the acceptance rate is high, increase it. To increase the scope of exploration; if the acceptance rate is low, then reduce it. A refined search is then conducted. After determining the number of adjustments, the system calculates an adjustment score for each user. This score consists of two parts: first, the sum of the interference intensity of the current user and all other users in the same group, reflecting the degree of conflict in the current group; second, the change in the objective function after the user moves to the new group, reflecting the potential benefit of the adjustment. Based on the score ranking, the corresponding number of users with the highest scores are selected to perform the adjustment operation, generating candidate solutions.

[0062] This invention transforms the pilot allocation problem into a maximum cut problem in graph theory by constructing a function that maximizes the sum of cross-group interference intensities, thus aligning the optimization objective with the physical requirement of suppressing pilot contamination. Building upon this, a Markov chain Monte Carlo method based on Metropolis-Hastings sampling is introduced to construct a Markov chain with a stationary Boltzmann distribution. A discriminant, derived from the ratio of the objective function to the ratio of the proposed distribution, is used as the state transition acceptance criterion, ensuring that the Markov chain can sample with the target distribution as a stationary distribution, thereby possessing global search capability in an exponentially growing solution space. The algorithm further integrates simulated annealing, using annealing scheduling based on temperature parameters. By controlling the randomness of the search process, the acceptance probabilities of each scheme are relatively small in the initial high-temperature stage, allowing the chain to explore the solution space extensively. As the temperature gradually decreases, the stable distribution of the chain gradually concentrates in the low-energy region, eventually converging to an approximate global optimum. The entire iterative process is controlled by an inner and outer double-layer loop. The outer loop is responsible for temperature updates, while the inner loop performs multiple Metropolis-Hastings samplings at a fixed temperature. The algorithm terminates when the temperature drops to a preset threshold or when multiple iterations fail to improve the optimal solution, outputting the historical best allocation scheme. This scheme achieves the dual optimization objectives of minimizing interference between users within the same group and maximizing interference between users across groups, fundamentally reducing mutual pollution caused by pilot reuse and improving channel estimation accuracy and system throughput.

[0063] In one alternative implementation, step a1 includes: Step b1: After each preset number of iterations, adjust the preset step size of the temperature parameter.

[0064] Specifically, during the algorithm's operation, whenever the cumulative number of iterations reaches a pre-set threshold (i.e., the preset number of iterations), the current temperature parameter is adjusted. An update operation is performed, with the magnitude and direction of the update determined by a preset step size. In the classic simulated annealing framework, this update typically manifests as a decrease in temperature to achieve the "annealing" effect. The preset step size can be linearly decaying, for example... ,in It can be a fixed decrease; or it can be a geometric decrease (or exponential decrease), for example... ,in The attenuation factor can be used; other more complex cooling strategies, such as logarithmic decay, can also be employed.

[0065] This invention achieves effective control over the simulated annealing process by adjusting the temperature parameter by a preset step size after each preset number of iterations. This mechanism gradually reduces the temperature according to a preset scheduling strategy, keeping the algorithm at a relatively high temperature in the early stages of iteration. At this point, the probability of accepting inferior solutions is higher, allowing the Markov chain to fully explore the solution space and avoid prematurely getting trapped in local optima. As the number of iterations increases, the temperature decreases by a preset step size, gradually reducing the probability of accepting inferior solutions. The algorithm's search behavior shifts from extensive exploration to fine-grained convergence, ensuring stable convergence to a region of high-quality solutions with lower energy. The configurability of the preset number of iterations and step size allows this cooling strategy to flexibly adapt to user scenarios of different scales, improving the algorithm's convergence accuracy while maintaining search efficiency. This provides a foundation for annealing control, enabling the subsequent output of pilot grouping schemes that meet the optimization objectives.

[0066] In one alternative implementation, step a4 includes: Step c1: Determine the number of users to be adjusted; Step c2: Determine the adjustment score for each user during the adjustment process. The adjustment score is obtained by summing the first score and the second score. The first score is determined by the sum of the interference intensity of the current user and all other users. The second score is determined by the change in the objective function after the current user is adjusted to the new allocation scheme. Step c3: Based on the number of users, identify the target user with the highest adjustment score from among all users and make adjustments accordingly.

[0067] Specifically, adjusting the number of users within a pilot reuse group can be broken down into three sequentially executed steps. First, the number of users involved in this adjustment needs to be determined. This number can be a pre-set fixed value or dynamically adjusted based on the acceptance of the previous iteration; its purpose is to control the disturbance amplitude of a single adjustment. Next, the system calculates an adjustment score for each user. This score is the sum of two parts: the first part is the first score, which is the sum of the interference intensity between the current user and all other users in the same pilot reuse group. This value directly reflects the degree of conflict for the user in the current group, i.e., the total interference generated when the user shares the same pilot with other users in the group. The second part is the second score, obtained by simulating the change in the objective function value after adjusting the user to another pilot reuse group. Specifically, the system will attempt to move the user from the current group to another candidate group and calculate the increment of the objective function value before and after the move. A larger increment indicates a higher benefit from the adjustment. Adding the two scores yields an index that comprehensively measures the user's "urgency for adjustment" and "adjustment benefit." Finally, based on the number of users determined in the first step, the corresponding number of users with the highest adjusted scores are selected from all users. Actual grouping adjustments are then performed on these users, such as moving them to a new group pre-defined during the scoring process, thereby generating new candidate allocation schemes. Through this score-guided adjustment mechanism, each iteration prioritizes processing the users with the greatest impact on the objective function, thus improving search efficiency.

[0068] As a specific example of the above scoring mechanism, we can define the first... Adjustment rating for individual users

[0069] in, Indicates user The weighting, which is the sum of the interference strength between the user and all other users, corresponds to the first score; Indicates user The change in the objective function when adjusting from the current group to the "most advantageous group" corresponds to the second score. In actual calculations, and All coefficients need to be normalized to avoid scaling issues. Preferences used to balance two factors: when When the score depends entirely on the local gain estimate, it degenerates into focusing only on the adjusted gain; when At that time, the rating depended entirely on the weighting, degenerating into focusing only on the degree of conflict among users within the current group. This can be addressed by adjusting... It allows for flexible control over the emphasis of the scoring mechanism on different factors.

[0070] To increase the randomness of the search, noise can be introduced into the probability to prevent the algorithm from getting trapped in local optima too early. Then, based on the current step size parameter... The number of users confirmed for this adjustment Sampled without replacement from this probability distribution A set of users As the target users of this adjustment, the probability sampling method based on ratings retains the bias of rating guidance (users with higher ratings are more likely to be selected) while improving the exploration ability through randomness, compared to simply selecting the highest-rated users. Compared to individual users, it can better balance exploration and utilization.

[0071] This invention achieves refined local search of pilot grouping schemes through an adjustment mechanism. First, by determining the number of users to be adjusted, the perturbation amplitude of each iteration is controlled. Second, an adjustment score is constructed for each user, obtained by summing a first score and a second score. The first score is based on the sum of the interference intensity of the current user and other users in the same group, intuitively reflecting the degree of conflict for the user in the current group. The second score is based on the change in the objective function after the user is adjusted to the new group, quantifying the potential benefit of the adjustment operation to the overall optimization objective. The combination of these two scores allows the score to comprehensively measure the urgency and optimization value of the user's adjustment. Based on this, the target user with the highest score is selected for adjustment according to a preset number of users, ensuring that each iteration prioritizes the user with the greatest impact on the objective function, thus concentrating computational resources on the operations most likely to improve solution quality. As a further optimization of this mechanism, a coefficient α is introduced to balance the weights of the first and second scores, allowing for flexible adjustment of the preference for the current degree of conflict and the benefit of adjustment according to the actual scenario. Simultaneously, a probability sampling method based on the score replaces hard selection, introducing randomness while retaining the score-guided tendency, balancing exploration and exploitation, and preventing the algorithm from prematurely falling into local optima. Overall, this adjustment mechanism improves the targeting and efficiency of iterative search, providing key support for the algorithm to quickly converge to a high-quality pilot grouping scheme.

[0072] In one alternative implementation, step c1 includes: Step d1: Update the number of users for this adjustment based on the output of the discriminant after the previous adjustment and the number of users after the previous adjustment.

[0073] Specifically, the algorithm dynamically updates the number of users for the current adjustment based on the output of the discriminant after the previous adjustment and the number of users in the previous adjustment. In each inner iteration, the algorithm records the number of users involved in the previous adjustment and the corresponding discriminant output, i.e., the acceptance probability. Acceptance probability This reflects the adoption of the previous adjustment, and its value is between 0 and 1. A higher value indicates that the previous adjustment direction matched the target distribution well, and the current step size is appropriate; if A lower value indicates that the step size may be too large, leading to frequent rejections of candidate solutions. Based on the feedback, the algorithm adjusts the number of users μ for this adjustment using a preset update rule.

[0074] In one example, the expression for adjusting the number of users is as follows:

[0075] in It is the output of the discriminant after the last adjustment, i.e. The output result; The preset target acceptance rate (e.g., 0.574) is used as the benchmark for comparison. It's the learning rate, which controls the magnitude of each adjustment.

[0076] For example when hour, μ increases; when hour, The decrease in μ is completely consistent with the qualitative description. This is achieved by introducing a target acceptance rate. and learning rate The formula can precisely control the speed and direction of step size adjustment, automatically adjusting the actual acceptance rate. By converging, a balance is achieved between exploration (leapfrog) and exploitation (meticulous search). The value of 0.574 is derived from the optimal theoretical acceptance rate of Markov chains under certain proposed distributions, which helps the algorithm maintain a hybrid approach.

[0077] This invention achieves adaptive control of the search step size by dynamically updating the number of users for the current adjustment based on the output of the discriminant after the previous adjustment and the number of users in the previous adjustment. This mechanism uses an acceptance flag A to reflect the adoption status of the previous adjustment and adjusts the number of users for the current adjustment according to a preset update rule: when the acceptance flag is high, the step size is appropriately increased to expand the search range; when the acceptance flag is low, the step size is appropriately decreased to perform a finer search. This is achieved by introducing a target acceptance rate A. And the learning rate η, this rule enables the actual acceptance rate to automatically adjust to A. By converging, a dynamic balance is maintained between exploration and utilization. This avoids wasting computational resources due to frequent rejection of candidate solutions caused by excessively large step sizes, and also prevents the search from getting stuck in a local stagnation due to excessively small step sizes. This adaptive update mechanism is embedded in an inner and outer double-layer loop control. The outer layer is responsible for temperature annealing, while the inner layer iteratively performs sampling and step size updates at a fixed temperature. This allows the algorithm to continuously optimize its search behavior based on real-time feedback, improve convergence speed and solution quality, and provide iterative support for the final output of a pilot grouping scheme that meets the optimization objective.

[0078] This embodiment also provides a MIMO pilot allocation device for implementing the above embodiments and preferred embodiments; details already described will not be repeated. As used below, the term "module" can refer to a combination of software and / or hardware that performs a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, hardware implementations, or a combination of software and hardware, are also possible and contemplated.

[0079] This embodiment provides a MIMO pilot allocation device, such as Figure 6 As shown, it includes: The interference strength acquisition module 601 acquires interference strength information, which includes the interference strength generated between each user when they reuse the same pilot resource. The pilot segmentation module 602 uses an improved Markov chain Monte Carlo-simulated annealing optimization algorithm to divide each user into multiple pilot reuse groups that meet the optimization objective. The number of pilot reuse groups is the same as the number of pilot resources of the base station, and users in the same pilot reuse group are assigned to use the same pilot resource of the base station. The optimization objective is determined based on the interference intensity information to minimize the interference intensity between users in the same pilot reuse group and maximize the interference intensity between users in different pilot reuse groups.

[0080] The MIMO pilot allocation device provided in this embodiment of the invention can execute the MIMO pilot allocation method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects for executing the method. Further functional descriptions of the various modules and units described above are the same as in the corresponding embodiments described above, and will not be repeated here.

[0081] Figure 7 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention.

[0082] The following is a detailed reference. Figure 7This diagram illustrates a suitable structural schematic for implementing an electronic device according to embodiments of the present invention. The electronic device may include a processor (e.g., a central processing unit, graphics processor, etc.) 701, which can perform various appropriate actions and processes based on a program stored in a read-only memory (ROM) 702 or a program loaded from memory 707 into random access memory (RAM) 703. The RAM 703 also stores various programs and data required for the operation of the electronic device. The processor 701, ROM 702, and RAM 703 are interconnected via a bus 704. An input / output (I / O) interface 705 is also connected to the bus 704.

[0083] Typically, the following devices can be connected to I / O interface 705: input devices 706 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 707 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; memory devices 708 including, for example, magnetic tapes, hard disks, etc.; and communication devices 709. Communication device 709 allows electronic devices to exchange data via wireless or wired communication with other devices. Although Figure 7 Electronic devices with various devices are shown, but it should be understood that it is not required to implement or have all of the devices shown, and more or fewer devices may be implemented or have instead.

[0084] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In embodiments, the computer program can be downloaded and installed from a network via a communication device 709, or installed from a memory 708, or installed from a ROM 702. When the computer program is executed by the processor 701, it performs the functions defined in the MIMO pilot allocation method of the embodiments of the present invention.

[0085] Figure 7 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of use of the embodiments of the present invention.

[0086] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code. When the software or computer code is accessed and executed by the computer, processor, or hardware, the MIMO pilot allocation method shown in the above embodiments is implemented.

[0087] A portion of this invention can be applied as a computer program product, such as computer program instructions, which, when executed by a computer, can invoke or provide the methods and / or technical solutions according to the invention through the operation of the computer. Those skilled in the art will understand that the forms in which computer program instructions exist in a computer-readable medium include, but are not limited to, source files, executable files, installation package files, etc. Correspondingly, the ways in which computer program instructions are executed by a computer include, but are not limited to: the computer directly executing the instructions, or the computer compiling the instructions and then executing the corresponding compiled program, or the computer reading and executing the instructions, or the computer reading and installing the instructions and then executing the corresponding installed program. Here, the computer-readable medium can be any available computer-readable storage medium or communication medium accessible to a computer.

[0088] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A MIMO pilot allocation method, characterized in that, The method includes: Obtain interference strength information, which includes the interference strength generated between each user when they reuse the same pilot resource; Based on the improved Markov chain Monte Carlo-simulated annealing optimization algorithm, each user is divided into multiple pilot reuse groups that meet the optimization objective. The number of pilot reuse groups is the same as the number of pilot resources of the base station, and users in the same pilot reuse group are assigned to use the same pilot resource of the base station. The optimization objective is determined based on the interference intensity information to minimize the interference intensity between users in the same pilot reuse group and maximize the interference intensity between users in different pilot reuse groups.

2. The method according to claim 1, characterized in that, The improved Markov chain Monte Carlo-simulated annealing optimization algorithm divides each user into several pilot multiplexing groups that meet the optimization objective, including: Generate the objective function for the optimization objective; A discriminant is constructed using temperature parameters and the objective function. The temperature parameters are adjusted using a simulated annealing algorithm. The discriminant is used to determine whether the adjustments made by users in the pilot multiplexing group are closer to the optimization objective. Initialize and assign each user to their respective pilot multiplexing group; Adjust the users in each pilot multiplexing group, and determine whether the adjustment scheme is closer to the optimization goal through the discriminant; If the proposed adjustment is closer to the optimization goal, then the adjustment will be retained. If the adjustment scheme deviates from the optimization target, then it is determined whether to retain the current adjustment based on the discriminant output result; The process of adjusting the users in each pilot multiplexing group is repeated iteratively until the stopping condition is met, and the pilot multiplexing group allocation scheme is output.

3. The method according to claim 2, characterized in that, The discriminant outputs a value greater than or equal to 1 when it is closer to the optimization target, and outputs a value less than 1 when it is far from the optimization target. If the adjustment scheme is far from the optimization target, a determination is made based on the output of the discriminant to determine whether to retain the current adjustment, including: Generate random numbers; If the random number is less than the output result of the discriminant, then this adjustment is abandoned; If the random number is greater than or equal to the output of the discriminant, then this adjustment is retained.

4. The method according to claim 3, characterized in that, The objective function is: In the formula, This indicates the pilot group label for user v. For the set of all user pairs, Let be the interference strength when user i and user j share the same pilot signal. It is an indicator function; The discriminant is: In the formula, This indicates the original allocation scheme. This represents the new allocation scheme, in which , Denotes the objective function, It is the temperature parameter, This indicates that the user set was selected when the new allocation scheme was adjusted to the original allocation scheme. The probability, This indicates that a user set was selected when adjusting from the original allocation scheme to the new allocation scheme. The probability, Indicates user From the original group Switch to new group The probability, Indicates user From the original group Switch to new group The probability of.

5. The method according to claim 2, characterized in that, The steps for adjusting the temperature parameters using a simulated annealing algorithm include: After a preset number of iterations, the temperature parameter is adjusted down by a preset step size.

6. The method according to claim 2, characterized in that, The adjustment of users in each pilot multiplexing group includes: Determine the number of users to be adjusted in this instance; Determine the adjustment score for each user during the adjustment process. The adjustment score is obtained by summing a first score and a second score. The first score is determined by the sum of the interference intensity of the current user and all other users. The second score is determined by the change in the objective function after the current user is adjusted to the new allocation scheme. Based on the number of users, the target user with the highest adjustment score is determined from among all users and the adjustment is carried out.

7. The method according to claim 6, characterized in that, Determining the number of users for this adjustment includes: The number of users for this adjustment is updated based on the output of the discriminant after the previous adjustment and the number of users after the previous adjustment.

8. A MIMO pilot allocation device, characterized in that, The device includes: An interference strength acquisition module is used to acquire interference strength information, which includes the interference strength generated between each user when they reuse the same pilot resource. The pilot segmentation module is used to divide users into multiple pilot reuse groups that meet the optimization objective based on an improved Markov chain Monte Carlo-simulated annealing optimization algorithm. The number of pilot reuse groups is the same as the number of pilot resources of the base station, and users in the same pilot reuse group are assigned to use the same pilot resource of the base station. The optimization objective is determined based on the interference intensity information to minimize the interference intensity between users in the same pilot reuse group and maximize the interference intensity between users in different pilot reuse groups.

9. An electronic device, characterized in that, include: A memory and a processor are communicatively connected, the memory stores computer instructions, and the processor executes the computer instructions to perform a MIMO pilot allocation method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing a computer to perform a MIMO pilot allocation method according to any one of claims 1 to 7.

Images