CN-113541746-B - Millimeter wave large-scale MIMO mixed precoding design method based on swarm algorithm

Abstract

The invention discloses a millimeter wave large-scale MIMO mixed precoding design method based on a bee colony algorithm, which belongs to the technical field of coding, and firstly adopts zero-forcing precoding to design a digital precoding matrix in a baseband part, the analog precoding matrix is designed independently in the analog precoding part, the phase of the analog phase shifter is used as a target variable, the condition that the system is enabled to reach the maximum speed is used as a target function, and the ABC algorithm is used for optimizing the problem, so that the precoding matrix which is as good as possible is obtained. The method greatly improves the convergence rate and reduces the operand, the invention adopts artificial bee colony (ARTIFICIAL BEE COLONY) algorithm to design the mixed precoding, and the mixed precoding design is equivalent to the problem of random optimization by means of the thought of the swarm intelligence algorithm, does not involve complex operations such as matrix inversion and the like, has lower operand and obtains better performance.

Inventors

• TAN FANGQING
• XIAO QING
• CUI ZIJIAN
• XU XU

Assignees

• 桂林电子科技大学

Dates

Publication Date

20260512

Application Date

20210603

Claims (5)

1. 1. A millimeter wave large-scale MIMO mixed precoding design method based on a bee colony algorithm is characterized by comprising the following steps: Step 1, firstly, manufacturing a simulation precoding matrix and a millimeter wave large-scale MIMO system model; Step 2, inputting, wherein NP is population number, maxiter is maximum iteration number, and H is channel matrix; Vector of Is a dimension of (2); The number of the receiving antennas; Step 3, initializing related parameters, wherein the current iteration time t=0, and randomly generating NP dimensions as follows A kind of electronic device Vectors, noted as , Calculate each Corresponding objective function value I.e. (5); Representing an analog precoding matrix; Step 4, use Calculating fitness values of NP individuals, wherein half of individuals with better fitness values in the population form a leading bee population, and the other half of individuals form a following bee population; Step 5, leading bees and following bees are searched in sequence to generate a new leading bee population and a new following bee population; step 6, evaluating NP to obtain function values according to the objective function values, and keeping better performance Vectors, recorded as When t < Maxiter, the time of the transition, If (3) < Then = , < Returning to the step 5 when the iteration times t=t+1 are smaller than the maximum iteration times, otherwise, entering the step 7; Step 7, after the maximum iteration number is reached Generating an analogue precoding matrix, i.e Output of ; In the millimeter wave large-scale MIMO system model, a digital precoder pair is used for encoding the millimeter wave large-scale MIMO system model The data streams are digitally precoded, and each data stream passes through a corresponding radio frequency link and then is coded by The analog precoder composed of phase shifters performs analog precoding on the data stream, and after analog precoding, each data stream passes through only the data stream connected with the corresponding radio frequency link Sub-antenna array of the root antenna is transmitted, total transmitting antenna number The number of the receiving end users is The user is a single antenna, so that the user receives signals Expressed as: (1) Wherein, the In the form of a channel matrix, The analog precoding matrix is represented by a matrix of analog precoding, A digital pre-coding matrix is represented and, A hybrid pre-coding matrix is represented and, Is the transmission of a signal that, Representing an additive gaussian white noise and, , The variance is represented as a function of the variance, And The total transmit power constraint is satisfied and, ; The system reachability and rate are expressed as: (2) Wherein, the (3) The ABC algorithm optimized objective function is: (4)。
2. 2. the method is characterized in that step 1 specifically comprises the steps of firstly adopting zero-forcing precoding in a baseband part to design a digital precoding matrix, independently designing an analog precoding matrix in an analog precoding part, taking the phase of an analog phase shifter as a target variable, taking the condition of maximizing the system accessibility and the speed as a target function, and optimizing the problem by using an ABC algorithm to obtain the precoding matrix as good as possible.
3. 3. The method for designing the millimeter wave massive MIMO hybrid precoding based on the swarm algorithm according to claim 1, wherein the leading bee searches for the current first step One target individual in the population of the lead bee Randomly selecting individuals Performing cross search dimension by dimension, generating new individuals according to (6) (6); For new born individuals And a target individual And (3) evaluating the fitness, comparing the fitness value of the population with the fitness value of the population, and selecting individuals with better fitness values to enter the population of the leading bees.
4. 4. The method for designing millimeter wave massive MIMO mixed precoding based on the swarm algorithm according to claim 3, wherein following bees search in a roulette manner, namely Selection of preferred target individuals in a novel population of lead bees , With randomly selected individuals Searching according to formula (6) to generate new individuals , A population of following bees is formed.
5. 5. The method for designing the millimeter wave large-scale MIMO mixed precoding based on the swarm algorithm according to claim 4, wherein after the leading bee search and the following bee search are completed, the two bee swarms are combined to generate an iterative population, whether the behavior of the scout bee occurs is judged, if the behavior of the scout bee occurs, the scout bee search is performed according to the formula (6) to generate a new individual, and after the new individual is compared with the original individual one by one, the individual with the better fitness value is selected to enter the iterative population.

Description

Millimeter wave large-scale MIMO mixed precoding design method based on swarm algorithm Technical Field The invention relates to the technical field of coding, in particular to a millimeter wave large-scale MIMO mixed precoding design method based on a bee colony algorithm. Background The millimeter wave has a short wavelength, so that a large number of antennas can be packaged in a small aperture, and meanwhile, a huge antenna array provides remarkable beam forming gain and can compensate the path loss of the millimeter wave. Thus, millimeter wave is one of the key technologies of 5G in combination with massive MIMO technology. Conventional digital precoding requires the same number of RF chains as the number of antennas, but results in a drastic increase in hardware cost and power consumption. Conventional analog precoding uses phase shifters instead of RF chains, but results in system performance loss. Thus, hybrid precoding, which combines digital precoding with analog precoding, has been proposed, and the core idea of hybrid precoding is to decompose a conventional 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 costs and performance loss. At present, large-scale MIMO adopts a OMP, SVD, SIC precoding mode, and the operation complexity is increased along with the increase of the number of days of a base station and the increase of cell users. Disclosure of Invention The invention aims to provide a millimeter wave large-scale MIMO mixed precoding design method based on a bee colony algorithm, which aims to solve the problems in the background technology. In order to achieve the above purpose, the present invention provides the following technical solutions: a millimeter wave large-scale MIMO mixed precoding design method based on a bee colony algorithm comprises the following steps: Step 1, firstly, manufacturing a simulation precoding matrix and a millimeter wave large-scale MIMO system model; Step 2, inputting NP, maxiter, H, N t, the dimension of vector theta, K, the number of receiving antennas, wherein NP is the number of population, maxiter is the maximum iteration number, H is the channel matrix; Step 3, initializing related parameters, randomly generating NP theta vectors with N t dimensions according to the current iteration times t=0, and recording the NP vectors as Calculating the objective function value corresponding to each θI.e. Calculating fitness values of NP individuals by using xi i, wherein half of individuals with better fitness values in the population form a leading bee population, and the other half of individuals form a following bee population; Step 5, leading bees and following bees are searched in sequence to generate a new leading bee population and a new following bee population; And step 6, evaluating NP to obtain function values according to the objective function values, reserving a theta vector with better performance, and recording as theta elite. When t < Maxiter is reached, If it isThenReturning to the step 5 when the iteration times t=t+1 are smaller than the maximum iteration times, otherwise, entering the step 7; Step 7, generating a simulation precoding matrix by using theta elite after the maximum iteration number is reached, namely Output F RF＝diag{f1,f2,...,fN. The step 1 is to design a digital precoding matrix by zero forcing precoding at a baseband part, design an analog precoding matrix at an analog precoding part independently, take the phase of an analog phase shifter as a target variable, take the condition of maximizing the system accessibility and the speed as a target function, and use an ABC algorithm to optimize the problem to obtain a precoding matrix as good as possible. As a further technical scheme of the invention, in the millimeter wave large-scale MIMO system model, firstly, N s paths of data streams are subjected to digital precoding through a digital precoder. After each path of data flow passes through the corresponding radio frequency link, an analog precoder formed by L phase shifters performs analog precoding on the data flow. After analog precoding, each data stream is transmitted through a sub-antenna array with only L antennas connected to a corresponding radio frequency link. Total transmit antenna number N t =ln, number of receiving end users is K, user is single antenna, and then received signal y e C K×1 at user end can be expressed as: y=HFRFFBBs+n=HFs+n (1) Where h= [ H 1,h2,...,hK]T∈CK×LN ] is the channel matrix, F RF＝diag{f1,f2,...,fN}∈CLN×N represents the analog precoding matrix, Representing a digital precoding matrix, f=f RFFBB representing a hybrid precoding matrix,Is the transmitted signal, n ε C K×1 represents additive Gaussian white noise, i.e., n-CN (0, σ 2IK),σ2 represents variance. F RF and F BB satisfy the total transmit power constraint, i.e The system reachability and rate are