A millimeter wave massive MIMO hybrid precoding design method based on a bee colony algorithm

Abstract

The application discloses a millimeter wave massive MIMO hybrid precoding design method based on a bee colony algorithm, and belongs to the technical field of coding. Firstly, the method adopts zero-forcing precoding to design a digital precoding matrix in a baseband part, separately designs an analog precoding matrix in an analog precoding part, takes the phase of an analog phase shifter as a target variable, takes the condition of maximum system reachable rate as a target function, uses an ABC algorithm to optimize the problem, and obtains the best precoding matrix as possible. The method greatly improves the convergence speed and reduces the operation amount. The application adopts an artificial bee colony (ABC) algorithm to design the hybrid precoding, equivalently designs the hybrid precoding as a random optimization problem by means of the idea of the swarm intelligence algorithm, does not involve complex operations such as matrix inversion, has low operation amount, and achieves good performance.

Description

Technical Field

[0001] This invention relates to the field of coding technology, specifically a millimeter-wave massive MIMO hybrid precoding design method based on the bee colony algorithm. Background Technology

[0002] Millimeter waves, due to their short wavelength, make it possible to package a large number of antennas within a small aperture. Simultaneously, the massive antenna array provides significant beamforming gain, compensating for the path loss of millimeter waves. Therefore, the combination of millimeter waves and massive MIMO technology is one of the key technologies of 5G. Traditional digital precoding methods require an RF chain number equal to the number of antennas, leading to a sharp increase in hardware cost and power consumption. Traditional analog precoding methods use phase shifters instead of RF chains, but this results in system performance degradation. Therefore, a hybrid precoding method combining digital and analog precoding has been proposed. The core idea of ​​hybrid precoding is to decompose the traditional digital precoder into a low-dimensional digital precoder consisting of a small number of RF chains and an analog precoder consisting of a large number of phase shifters, reducing hardware cost and performance loss. Currently, massive MIMO mostly uses precoding methods such as OMP, SVD, and SIC. With the increase in the number of base stations and cell users, the computational complexity also increases. Summary of the Invention

[0003] The purpose of this invention is to provide a millimeter-wave massive MIMO hybrid precoding design method based on bee colony algorithm to solve the problems mentioned in the background art.

[0004] To achieve the above objectives, the present invention provides the following technical solution:

[0005] A hybrid precoding design method for millimeter-wave massive MIMO based on bee colony algorithm includes the following steps:

[0006] Step 1: First, create the analog precoding matrix and the millimeter-wave massive MIMO system model;

[0007] Step 2, Input: NP: Population size; Maxiter: Maximum number of iterations; H: Channel matrix; N t θ: Dimension of vector θ; K: Number of receiving antennas;

[0008] Step 3: Initialize relevant parameters. The current iteration number t = 0. Randomly generate NP dimensions of N. t The θ vector is denoted as Calculate the objective function value for each θ. Right now

[0009] Step 4, use ξ iCalculate the fitness values ​​of NP individuals. Half of the individuals with better fitness values ​​in the population form the leader bee population, and the other half form the follower bee population.

[0010] Step 5: The leader bees and follower bees search in sequence to generate new leader bee populations and follower bee populations;

[0011] Step 6: Evaluate NP function values ​​based on the objective function value, retain the best-performing θ vector, and record it as θ. elite When t < Maxiter, if but If the number of iterations t = t + 1, and the number of iterations is less than the maximum number of iterations, return to step 5; otherwise, proceed to step 7.

[0012] Step 7: After reaching the maximum number of iterations, use θ elite Generate an analog precoding matrix, i.e. Output F RF =diag{f1,f2,...,f N}

[0013] As a further technical solution of the present invention, step 1 specifically involves: firstly, using zero-forcing precoding in the baseband section to design a digital precoding matrix, and separately designing an analog precoding matrix in the analog precoding section. The phase of the analog phase shifter is taken as the target variable, and the condition that maximizes the system's reach and speed is taken as the objective function. The ABC algorithm is used to optimize this problem and obtain the best possible precoding matrix.

[0014] As a further technical solution of the present invention, in the millimeter-wave massive MIMO system model, N is first processed by a digital pre-encoder. s Each data stream undergoes digital precoding. After passing through the corresponding RF link, each data stream is precoded analogically by an analog precoder consisting of L phase shifters. Following analog precoding, each data stream is transmitted through a sub-antenna array with only L antennas connected to the corresponding RF link. The total number of transmit antennas is N. t =LN, the number of receiving users is K, and each user has a single antenna. Therefore, the received signal y∈C at the user end is... K×1 It can be represented as:

[0015] y = HF RF F BB s+n=HFs+n (1)

[0016] Where H = [h1, h2, ..., h K ] T ∈C K×LN Let F be the channel matrix. RF=diag{f1,f2,...,f N}∈C LN×N Represents the analog precoding matrix. Represents the digital precoding matrix, F = F RF F BB Represents a hybrid precoding matrix. It is a signal to be sent, n∈C K×1 This represents additive white Gaussian noise, i.e., n ~ CN(0,σ). 2 I K ), σ 2 F represents the variance. RF and F BB Satisfy the total transmit power constraint, i.e.

[0017] The reachability and rate of the system are expressed as:

[0018] in,

[0019]

[0020] The objective function for optimizing the ABC algorithm is:

[0021] As a further technical solution of the present invention, the guiding bee search involves identifying a target individual within the current t-th generation guiding bee population. Randomly select individuals r1∈{1,2,...,NP / 2} and perform cross-search dimension by dimension to generate new individuals according to equation (6).

[0022]

[0023] For newly generated individuals and target individuals A fitness evaluation is conducted, and the fitness values ​​of the two individuals are compared. The individual with the better fitness value is selected to enter the leader bee colony.

[0024] As a further technical solution of the present invention, follow-bee search: follow-bees follow a roulette-like method (i.e. Selecting superior target individuals from new leader bee colonies. A new individual is generated by searching with a randomly selected individual j according to equation (6). They form a follower bee colony.

[0025] As a further technical solution of the present invention, after the lead bee search and follower bee search are completed, an iterative population is generated by combining the two bee colonies. It is determined whether scout bee behavior occurs. If it occurs, a scout bee search is performed according to formula (6) to generate new individuals. After comparing them one-to-one with the original individuals, individuals with better fitness values ​​are selected to enter the iterative population.

[0026] Compared with the prior art, the beneficial effects of the present invention are: the present invention uses the Artificial Bee Colony algorithm to design hybrid precoding, and uses the idea of ​​swarm intelligence algorithm to make the hybrid precoding design equivalent to a random optimization problem, which does not involve complex operations such as matrix inversion, has a low computational load, and achieves better performance. Attached Figure Description

[0027] Figure 1 This is a model for a millimeter-wave large-scale MIMO system. Detailed Implementation

[0028] 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, and 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.

[0029] Please see Figure 1 Example 1: A hybrid precoding design method for millimeter-wave massive MIMO based on bee colony algorithm, comprising the following steps:

[0030] Step 1: First, a zero-forcing precoding method is used to design the digital precoding matrix in the baseband section. A separate analog precoding matrix is ​​designed for the analog precoding section. The phase of the analog phase shifter is used as the objective variable, and the condition that maximizes the system's reach and speed is used as the objective function. The ABC algorithm is used to optimize this problem, obtaining the best possible precoding matrix. This method significantly improves the convergence speed and reduces the computational load.

[0031] Step 2, Input: NP: Population size; Maxiter: Maximum number of iterations; H: Channel matrix; N t θ: Dimension of vector θ; K: Number of receiving antennas;

[0032] Step 3: Initialize relevant parameters. The current iteration number t = 0. Randomly generate NP dimensions of N. t The θ vector is denoted as Calculate the objective function value for each θ. Right now

[0033] Step 4, use ξ i Calculate the fitness values ​​of NP individuals. Half of the individuals with better fitness values ​​in the population form the leader bee population, and the other half form the follower bee population.

[0034] Step 5: The leader bees and follower bees search in sequence to generate new leader bee populations and follower bee populations;

[0035] Leader bee search: A target individual in the current t-th generation leader bee population. Randomly select individuals r1∈{1,2,...,NP / 2} and perform cross-search dimension by dimension to generate new individuals according to equation (6).

[0036]

[0037] For newly generated individuals and target individuals A fitness evaluation is conducted, and the fitness values ​​of the two individuals are compared. The individual with the better fitness value is selected to enter the leader bee colony.

[0038] Follow-the-bee search: Follow-the-bees follow a roulette-like method (i.e. Selecting superior target individuals from new leader bee colonies. A new individual is generated by searching with a randomly selected individual j according to equation (6). Forming a follower bee colony;

[0039] After the lead bee search and follower bee search are completed, an iterative population is generated by combining the two bee colonies. It is then determined whether scout bee behavior has occurred. If it has, a scout bee search is performed according to formula (6) to generate new individuals. After comparing them one-to-one with the original individuals, individuals with better fitness values ​​are selected to enter the iterative population.

[0040] Step 6: Evaluate NP function values ​​based on the objective function value, retain the best-performing θ vector, and record it as θ. elite When t < Maxiter, if but If the number of iterations t = t + 1, and the number of iterations is less than the maximum number of iterations, return to step 5; otherwise, proceed to step 7.

[0041] Step 7: After reaching the maximum number of iterations, use θ elite Generate an analog precoding matrix, i.e. Output F RF =diag{f1,f2,...,f N}

[0042] Example 2, based on Example 1, in the millimeter-wave massive MIMO system model, firstly, N is processed by a digital pre-encoder. s Each data stream undergoes digital precoding. After passing through the corresponding RF link, each data stream is precoded analogically by an analog precoder consisting of L phase shifters. Following analog precoding, each data stream is transmitted through a sub-antenna array with only L antennas connected to the corresponding RF link. The total number of transmit antennas is N. t=LN, the number of receiving users is K, and each user has a single antenna. Therefore, the received signal y∈C at the user end is... K×1 It can be represented as:

[0043] y = HF RF F BB s+n=HFs+n (1)

[0044] Where H = [h1, h2, ..., h K ] T ∈C K×LN Let F be the channel matrix. RF =diag{f1,f2,...,f N}∈C LN×N Represents the analog precoding matrix. Represents the digital precoding matrix, F = F RF F BB Represents a hybrid precoding matrix. It is a signal to be sent, n∈C K×1 This represents additive white Gaussian noise, i.e., n ~ CN(0,σ). 2 I K ), σ 2 F represents the variance. RF and F BB Satisfy the total transmit power constraint, i.e.

[0045] The reachability and rate of the system are expressed as:

[0046] in,

[0047]

[0048] The objective function for optimizing the ABC algorithm is:

[0049] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

[0050] Furthermore, it should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

Claims

1. A millimeter-wave massive MIMO hybrid precoding design method based on bee colony algorithm, characterized in that, Includes the following steps: Step 1: First, create the analog precoding matrix and the millimeter-wave massive MIMO system model; Step 2, Input: NP: Population size; Maxiter: Maximum number of iterations; H: Channel matrix; :vector The dimension; Number of receiving antennas; Step 3: Initialize relevant parameters. The current iteration number t=0. Randomly generate NP dimensions. of Vector, denoted as , Calculate each Corresponding objective function value ,Right now (5); Represents the analog precoding matrix; Step 4, use Calculate the fitness values ​​of NP individuals. Half of the individuals with better fitness values ​​in the population form the leader bee population, and the other half form the follower bee population. Step 5: The leader bees and follower bees search in sequence to generate new leader bee populations and follower bee populations; Step 6: Evaluate NP function values ​​based on the objective function value, and retain the best performing ones. Vector, recorded as When t < Maxiter, ,if < ,but = , < The number of iterations is t = t + 1. If the number of iterations is less than the maximum number of iterations, return to step 5; otherwise, proceed to step 7. Step 7: After reaching the maximum number of iterations, ... Generate an analog precoding matrix, i.e. Output ; In the millimeter-wave massive MIMO system model, the digital precoder first... Each data stream undergoes digital precoding, and after passing through the corresponding RF link, it is... An analog pre-encoder composed of phase shifters performs analog precoding. After analog precoding, each data stream is connected to a single RF link. The sub-antenna array of the root antenna is transmitted, with a total number of transmitting antennas. The number of receiving users is The user has a single antenna, so the received signal at the user end... Represented as: （1） in, For the channel matrix, Represents the analog precoding matrix. Represents a digital precoding matrix. Represents a hybrid precoding matrix. It is sending a signal. This represents additive white Gaussian noise. , Represents variance. and To meet the total transmit power constraint, ; The reachability and rate of the system are expressed as: (2) in, (3) The objective function for optimizing the ABC algorithm is: (4).

2. The millimeter-wave massive MIMO hybrid precoding design method based on bee colony algorithm according to claim 1, characterized in that, Step 1 specifically involves: firstly, designing a digital precoding matrix using zero-forcing precoding in the baseband section, and then designing a separate analog precoding matrix in the analog precoding section. The phase of the analog phase shifter is taken as the objective variable, and the condition that maximizes the system's reach and speed is taken as the objective function. The ABC algorithm is used to optimize this problem and obtain the best possible precoding matrix.

3. The millimeter-wave massive MIMO hybrid precoding design method based on bee colony algorithm according to claim 1, characterized in that, Leading Bee Search: Current number A leader guides a target individual within a bee colony. Randomly select individuals Perform cross-search dimension by dimension and generate new individuals according to equation (6). （6）； For newly generated individuals and target individuals A fitness evaluation is conducted, and the fitness values ​​of the two individuals are compared. The individual with the better fitness value is selected to enter the leader bee colony.

4. The millimeter-wave massive MIMO hybrid precoding design method based on bee colony algorithm according to claim 3, characterized in that, Follow-the-bees search: Follow-the-bees search in a roulette-like manner, i.e. Select superior target individuals from the new leading bee colony. , , and randomly selected individuals The search is performed according to formula (6) to generate new individuals. , This forms a follower bee colony.

5. The millimeter-wave massive MIMO hybrid precoding design method based on bee colony algorithm according to claim 4, characterized in that, After the lead bee search and follower bee search are completed, an iterative population is generated by combining the two bee colonies. It is then determined whether scout bee behavior has occurred. If it has, a scout bee search is performed according to formula (6) to generate new individuals. After comparing them one-to-one with the original individuals, individuals with better fitness values ​​are selected to enter the iterative population.

Images