CN110311719B - Beam selection method and device applied to millimeter wave large-scale MIMO system - Google Patents

Beam selection method and device applied to millimeter wave large-scale MIMO system Download PDF

Info

Publication number
CN110311719B
CN110311719B CN201910705982.7A CN201910705982A CN110311719B CN 110311719 B CN110311719 B CN 110311719B CN 201910705982 A CN201910705982 A CN 201910705982A CN 110311719 B CN110311719 B CN 110311719B
Authority
CN
China
Prior art keywords
bird
nests
bird nest
fitness
nest
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201910705982.7A
Other languages
Chinese (zh)
Other versions
CN110311719A (en
Inventor
李晓辉
汪银
张红伟
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Anhui University
Original Assignee
Anhui University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Anhui University filed Critical Anhui University
Priority to CN201910705982.7A priority Critical patent/CN110311719B/en
Publication of CN110311719A publication Critical patent/CN110311719A/en
Application granted granted Critical
Publication of CN110311719B publication Critical patent/CN110311719B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B7/00—Radio transmission systems, i.e. using radiation field
    • H04B7/02—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas
    • H04B7/04—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas
    • H04B7/0413—MIMO systems
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B7/00—Radio transmission systems, i.e. using radiation field
    • H04B7/02—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas
    • H04B7/04—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas
    • H04B7/0413—MIMO systems
    • H04B7/0456—Selection of precoding matrices or codebooks, e.g. using matrices antenna weighting
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B7/00—Radio transmission systems, i.e. using radiation field
    • H04B7/02—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas
    • H04B7/04—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas
    • H04B7/06—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the transmitting station
    • H04B7/0686—Hybrid systems, i.e. switching and simultaneous transmission
    • H04B7/0695—Hybrid systems, i.e. switching and simultaneous transmission using beam selection
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B7/00—Radio transmission systems, i.e. using radiation field
    • H04B7/02—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas
    • H04B7/04—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas
    • H04B7/08—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the receiving station
    • H04B7/0868—Hybrid systems, i.e. switching and combining
    • H04B7/088—Hybrid systems, i.e. switching and combining using beam selection

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Radio Transmission System (AREA)
  • Mobile Radio Communication Systems (AREA)

Abstract

The invention discloses a wave beam selection method and a device thereof applied to a millimeter wave large-scale MIMO system, wherein the wave beam selection method comprises the following steps: step S1, defining the bird nest number, bird nest discovery probability, binary coding control parameters and maximum iteration times, initializing a plurality of bird nests, and calculating the fitness of the bird nests; step S2, performing binary code mixed updating, repairing abnormal codes, calculating the fitness of a plurality of newly generated bird nests, and screening the bird nests by reserving the bird nests with larger fitness; step S3, comparing the bird nest discovery probability with the random number, copying the bird nest with the global optimal solution to replace one of the discovered bird nests, and randomly changing the positions of the other discovered bird nests; and step S4, judging whether the iteration times reach the maximum iteration times, if so, outputting a global optimal solution, and otherwise, executing step S2. Compared with a full-digital precoding algorithm, the method does not cause obvious performance loss, reduces the complexity of the algorithm and obtains near-excellent system performance.

Description

Beam selection method and device applied to millimeter wave large-scale MIMO system
Technical Field
The invention relates to a beam selection method in the technical field of mobile communication, in particular to a beam selection method applied to a millimeter wave large-scale MIMO system, and relates to a beam selection device applied to the millimeter wave large-scale MIMO system.
Background
With the rapid development of internet services, people have increasingly increased demands on various application fields of wireless networks, and increasingly tense spectrum resources cannot meet the demands of people on communication. Millimeter wave large-scale input and output, which can achieve higher data rate and higher spectral efficiency through wider signal bandwidth, is considered as a key technology of future 5G wireless communication. The traditional full digital beam forming scheme requires that each antenna corresponds to an independent radio frequency link, and with the increasing number of base station antennas and the number of cell users, the number of required radio frequency links is also increasing, and although the performance is excellent, the hardware cost and the realization difficulty are increased.
In the prior art, a discrete lens array with negligible performance loss is adopted to convert a traditional spatial channel into a beam spatial channel so as to obtain channel sparsity under millimeter wave frequency. However, due to the sparsity of the beam-space MIMO channel, the number of RF links can be reduced by selecting only a small number of suitable antennas, and no significant system performance loss is incurred. Meanwhile, the currently used schemes also include selecting a beam that causes the least capacity loss to be eliminated, selecting a beam that contributes the most in terms of system capacity, a maximum amplitude beam selection algorithm, an interference-aware beam selection scheme, and an ant colony optimization-based scheme. The selection scheme for eliminating the beam causing the least capacity loss and the selection scheme for selecting the beam contributing the most in the system capacity are both required to perform sequential searching, and the complexity is too high. The maximum amplitude beam selection algorithm is the simplest scheme, but has the problems of multi-user interference and RF chain waste caused by different RF chains selecting the same beam. Interference aware beam selection schemes reselect beams for interfering users, but their complexity is affected by the number of interfering users. The ant colony optimization-based scheme is similar to the interference perception beam selection scheme, and is based on the standard of amplitude maximization instead of direct optimization and rate, so that the performance of the system is limited. Therefore, the existing beam selection schemes have the problems of high complexity or energy loss, too long calculation time and inapplicability to practical systems.
Disclosure of Invention
Aiming at the prior technical problem, the invention provides a beam selection method and a beam selection device applied to a millimeter wave large-scale MIMO system, and solves the problems that the existing beam selection schemes have high complexity or energy loss, overlong calculation time and are not suitable for practical systems.
The invention is realized by adopting the following technical scheme: a wave beam selection method applied to a millimeter wave massive MIMO system comprises the following steps:
step S1, defining the nest number, the nest discovery probability, the binary coding control parameter, the maximum iteration number, the antenna number and the user number of a millimeter wave large-scale MIMO system, initializing a plurality of nests, selecting a wave beam with the largest channel amplitude value and no repetition for a user by each nest, calculating the fitness of the plurality of nests, and taking the maximum fitness of the current nest as a global optimal solution;
step S2, after calculating the fitness of a plurality of bird nests, firstly, performing binary code mixed updating on the plurality of bird nests, repairing abnormal codes, then calculating the fitness of the newly generated plurality of bird nests, and finally screening the bird nests by reserving the bird nests with larger fitness;
step S3, after all bird nests are screened, comparing the bird nest discovery probability with random numbers obeying uniform distribution, if the random numbers are greater than the bird nest discovery probability, copying the bird nest of the global optimal solution to replace one of the found bird nests, randomly changing the positions of the rest of the found bird nests, reserving the global optimal solution to accelerate the convergence speed of the algorithm, and determining the position and the optimal value of the bird nest of the current optimal solution;
and step S4, after determining the position and the optimal value of the bird nest, judging whether the iteration times reach the maximum iteration times, if so, outputting the global optimal solution, otherwise, executing step S2.
As a further detailed description of the above scheme, in step S2, the binary-coded hybrid update method includes the steps of:
step S21, judging whether the random number of the system is not larger than the binary coding control parameter;
when the random number is not greater than the binary coding control parameter, executing step S22, mapping a real number into discrete binary data by using a sigmoid function, and calculating a state value of the nth antenna of the kth user in the (m + 1) th iteration through the following formula
Figure BDA0002150041800000031
Figure BDA0002150041800000032
Figure BDA0002150041800000033
When the random number is greater than the binary coding control parameter, executing step S23, and determining whether the Levy jump path is a positive number;
when the Levy hop path is not positive, step S24 is executed to calculate the state value of the nth antenna of the kth user in the (m + 1) th iteration through the following formula
Figure BDA0002150041800000034
Figure BDA0002150041800000035
Figure BDA0002150041800000036
When the hop path is positive, step S25 is executed to calculate the state value of the nth antenna of the kth user in the (m + 1) th iteration through the following formula
Figure BDA0002150041800000037
Figure BDA0002150041800000038
Figure BDA0002150041800000039
Wherein rand is the random number.
Further, the method for repairing the abnormal code comprises the following steps:
step S26, calculating the shape of the nth antenna of the kth user of the mth bird nestValue of state xnkThat is, calculating the state value of the n row at the k column of the m matrix and determining the state value xnkWhether or not it is 1;
at the state value xnkIf it is 1, step S27 is executed to calculate the corresponding channel gain gnk；
All state values x in each column of the binary coded matrix are judgednkThen, step S28 is executed to select the beam and the beam set of the user by the following formula:
Figure BDA0002150041800000041
Figure BDA0002150041800000042
wherein, betakIs the beam of the k-th user,
Figure BDA0002150041800000043
a set of beams that has been selected for the user.
As a further detailed description of the above scheme, the beam selection method further includes the steps of:
step S0, constructing the millimeter wave large-scale MIMO system; the construction method of the millimeter wave large-scale MIMO system comprises the following steps:
step S01, preliminarily defining an expression formula of the receiving signal of the kth user;
step S02, adopting a discrete lens array, converting the traditional space channel into a beam space channel through a space Fourier transform matrix, and redefining an expression formula of a receiving signal of the kth user according to the beam space channel;
and step S03, only selecting partial wave beams according to sparsity of the MIMO channel of the wave beam space, and establishing an objective function and constraint conditions selected by the antenna.
Further, in the millimeter wave massive MIMO system of the present embodiment, the expression formula of the received signal of the kth user is initially defined as:
y＝HHWs+n
where H is the channel matrix, H ═ H1,h2,...,hK]，hkIs the channel vector between the kth user and the base station; s is the original signal vector, s belongs to CK×1And normalizing the power E (ss)H)＝IK(ii) a W is a precoding matrix of size NxK, and tr (WW) is satisfiedH) Rho is not more than rho, and rho is total emission power; n is additive white Gaussian noise having a size of K × 1, and n to CN (0, σ)2IK)。
Still further, the calculation formula of the channel vector is as follows:
Figure BDA0002150041800000044
in the formula (I), the compound is shown in the specification,
Figure BDA0002150041800000051
for the line-of-sight path of the kth user,
Figure BDA0002150041800000052
a non-line-of-sight path for the kth user; gkFor complex gain,. psikIn the form of a spatial orientation,
Figure BDA0002150041800000053
is an array response vector.
Still further, the spatial fourier transform matrix is:
Figure BDA0002150041800000054
in the formula (I), the compound is shown in the specification,
Figure BDA0002150041800000055
thus, in the millimeter wave massive MIMO system, the expression formula of the reception signal of the k-th user is defined again as:
Figure BDA0002150041800000056
wherein the content of the first and second substances,
Figure BDA0002150041800000057
is a beam space channel, and the calculation formula is:
Figure BDA0002150041800000058
Figure BDA0002150041800000059
is the beam space channel vector of the kth user, and K is 1.
Still further, the expression formula of the downlink received signal is:
Figure BDA00021500418000000510
in the formula (I), the compound is shown in the specification,
Figure BDA00021500418000000511
for the channel formed by the selected beam, Wr∈CK×KAnd is a reduced dimensionality digital precoding matrix.
Still further, the objective function is:
Figure BDA00021500418000000512
in the formula, xnkThe state value of the nth antenna of the kth user; rkThe achievable average rate for the kth user; wherein, the calculation formula of the average speed is as follows:
Figure BDA00021500418000000513
in the formula, σ2In order to be able to measure the power of the noise,
Figure BDA0002150041800000061
Wr∈CK×K(ii) a Alpha is a scaling factor and satisfies
Figure BDA0002150041800000062
ρ is the total transmitted power.
The invention also provides a beam selection device applied to the millimeter wave large-scale MIMO system, which applies any of the beam selection methods applied to the millimeter wave large-scale MIMO system, and comprises the following steps:
the fitness calculation module is used for firstly defining the number of bird nests, the bird nest discovery probability, the binary coding control parameters, the maximum iteration times, the number of antennas and the number of users of a millimeter wave large-scale MIMO system, then initializing a plurality of bird nests, enabling each bird nest to select a wave beam with the maximum channel amplitude and no repetition for the users, finally calculating the fitness of the plurality of bird nests, and taking the maximum fitness of the current bird nest as a global optimal solution; the calculation formula of the fitness of the bird nests is as follows:
Figure BDA0002150041800000063
in the formula, xnkThe state value of the nth antenna of the kth user; rkThe achievable average rate for the kth user; wherein, the calculation formula of the average speed is as follows:
Figure BDA0002150041800000064
in the formula, σ2Is the noise power; alpha is a scaling factor and satisfies
Figure BDA0002150041800000065
Rho is total transmitting power, and K is the number of users; h is the channel matrix, HkIs the channel vector between the kth user and the base station;
Figure BDA0002150041800000066
forming channels for the selected beams;
Figure BDA0002150041800000067
Wr∈CK×Kand is a reduced dimensionality digital precoding matrix;
the bird nest screening module is used for firstly carrying out binary code mixed updating on the plurality of bird nests after the fitness calculation module calculates the fitness of the plurality of bird nests, repairing abnormal codes, calculating the fitness of the plurality of newly generated bird nests, and finally screening the bird nests by reserving the bird nests with higher fitness;
a bird nest position replacement module for comparing the bird nest discovery probability with random numbers subject to uniform distribution after the bird nest screening module screens all bird nests; if the random number is greater than the bird nest discovery probability, the bird nest position replacement module copies the bird nest of the global optimal solution to replace one found bird nest, and randomly changes the positions of the rest found bird nests to determine the bird nest position and the optimal value of the current optimal solution; and the iteration frequency judging module is used for judging whether the iteration frequency reaches the maximum iteration frequency after the bird nest position replacing module determines the bird nest position and the optimal value, if so, outputting the global optimal solution, and otherwise, driving the bird nest screening module to work.
Compared with the existing beam selection scheme, the beam selection method and the device thereof applied to the millimeter wave large-scale MIMO system have the following beneficial effects:
the beam selection method applied to the millimeter wave large-scale MIMO system considers the beam selection as solving the {0-1} knapsack problem and solves the problem, reduces the number of required radio frequency links without causing obvious performance loss, can avoid overhigh energy loss caused by adopting a large number of radio frequency links required by full digital coding, and can obtain near-excellent system performance on the premise of reducing complexity. Because abnormal coding is easy to occur in the solving process, the abnormal solutions are repaired by the method, and the calculated solutions are guaranteed to be feasible solutions. Meanwhile, the invention copies the globally optimal bird nest to replace the found bird nest, so that the optimal bird nest is saved, the convergence of the algorithm is accelerated, and in addition, the interference among users is not considered, and the optimization and the speed rate are directly taken as targets, so the realization and the speed rate performance are better compared with the prior several wave beam selection schemes. In addition, the embodiment adds application scenes to the discrete cuckoo search algorithm.
In the scheme of the invention, the beam selection method takes the selected beam as a selected article to be loaded into the backpack, the maximum sum rate of the system is taken as the problem of solving the maximum capacity of the backpack, an improved discrete cuckoo search algorithm is adopted to obtain a near-optimal solution, and a heuristic greedy algorithm is adopted to repair abnormal codes generated in a Levy flight discretization result of the cuckoo algorithm. Meanwhile, the invention introduces the copy in the genetic algorithm into the DCS algorithm, copies the globally optimal bird nest to replace the found bird nest, accelerates the convergence speed of the algorithm, and proves that the beam selection scheme of the invention is superior to the existing beam selection scheme in the aspect of improving the sum rate of the system through analysis.
Drawings
Fig. 1 is a flowchart of a beam selection method applied to a millimeter wave massive MIMO system in embodiment 1 of the present invention.
Fig. 2 is a system model diagram constructed by the beam selection method applied to the millimeter wave massive MIMO system in embodiment 2 of the present invention.
Fig. 3 is a graph illustrating the convergence of the algorithm of the beam selection method provided in embodiment 2 of the present invention.
Fig. 4 is a simulation diagram of the implementation and rate comparison of the beam selection method provided in embodiment 2 of the present invention with different beam selection algorithms in a sparse system.
Fig. 5 is a simulation diagram of the beam selection method provided in embodiment 2 of the present invention implemented in a dense system and compared with different beam selection algorithms in terms of speed.
Fig. 6 is a simulation diagram comparing the speed with different numbers of users in the beam selection method provided in embodiment 2 of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
Example 1
Referring to fig. 1, the present embodiment provides a beam selection method applied to a mmwave massive MIMO system, which can be used to select a beam in a mmwave massive MIMO system. In this embodiment, for the optimization problem of the beam selection model, a model is proposed, in which the beam selection problem is regarded as solving a {0-1} knapsack problem, the selected beam is regarded as loading an article into a knapsack, the maximum sum rate of the system is regarded as the maximum capacity problem of the knapsack, and a discrete cuckoo algorithm is used for solving the problem. The CS algorithm is based on the following three assumptions:
(1) randomly selecting one nest for each cuckoo and only producing the next cuckoo egg;
(2) the best bird nest will be retained to the next generation;
(3) the number of bird nests is fixed, and the probability that the brook bird eggs in the bird nests are found by host birds is Pa∈[0,1]。
The formula of the Levy flight jump path is simulated by adopting Mantegna, and the basic cuckoo algorithm formula of Levy flight is as follows:
Figure BDA0002150041800000091
the discrete cuckoo algorithm requires binary code transformation of the jump path updated every time the Levy flight is in position. Based on this, the beam selection method applied to the millimeter wave massive MIMO system of the present embodiment includes the following steps (steps S1-S4).
Step S1: first defining the bird nest number and bird nest discovery probability of a millimeter wave large-scale MIMO systemThe method comprises the steps of firstly, initializing a plurality of bird nests according to binary coding control parameters, the maximum iteration times, the number of antennas and the number of users, then, initializing the plurality of bird nests, selecting a wave beam with the maximum channel amplitude and no repetition for the users by each bird nest, finally, calculating the fitness of the plurality of bird nests, and taking the maximum fitness of the current bird nest as a global optimal solution. In this embodiment, the parameters of the beam selection method are first set: number of bird nests M, bird nest discovery probability PaBinary coding of the control parameter prMaximum number of iterations TmaxThe number of antennas N and the number of users K. Then, M bird nests are initialized, each bird nest selects a beam with the largest channel amplitude value and no repetition for a user, and if multiple users select repeated beams, non-repeated beams are randomly selected from the rest beams for the multiple users. And finally, calculating the fitness of the M bird nests, namely calculating an objective function selected by the antenna, and taking the maximum value of the fitness of the current bird nest as a global optimal solution. The objective function represents maximizing the system and rate under constraints.
Step S2: after the fitness of a plurality of bird nests is calculated, binary coding mixed updating is carried out on the plurality of bird nests firstly, abnormal codes are repaired, the newly generated fitness of the plurality of bird nests is calculated, and finally the bird nests are screened by reserving the bird nests with larger fitness. In this embodiment, under the parameter condition given in step S1, binary coding hybrid updating is performed on the Levy flight path, and a heuristic greedy algorithm is used to repair the abnormal code. Then, calculating the fitness of newly generated M bird nests, replacing the original bird nest if the fitness is higher than that of the original bird nest, and discarding if the fitness is lower than that of the original bird nest; and if the current bird nest fitness is larger than the global optimal solution, taking the current bird nest fitness as the global optimal solution. In the present embodiment, the binary-coded hybrid update method includes the following steps (steps S21-25):
step S21, judging whether the random number of the system is not larger than the binary coding control parameter;
when the random number is not greater than the binary coding control parameter, executing step S22, mapping the real number into discrete binary data by using a sigmoid function, and calculating the nth antenna of the kth user in the (m + 1) th iteration through the following formulaStatus value
Figure BDA0002150041800000101
Figure BDA0002150041800000102
Figure BDA0002150041800000103
When the random number is greater than the binary coding control parameter, executing step S23, and determining whether the Levy jump path is a positive number;
when the Levy hop path is not positive, step S24 is executed to calculate the state value of the nth antenna of the kth user in the (m + 1) th iteration through the following formula
Figure BDA0002150041800000104
Figure BDA0002150041800000105
Figure BDA0002150041800000106
When the hop path is positive, step S25 is executed to calculate the state value of the nth antenna of the kth user in the (m + 1) th iteration through the following formula
Figure BDA0002150041800000107
Figure BDA0002150041800000108
Figure BDA0002150041800000109
Wherein rand is a random number, pr∈[0,1]。prThe larger the algorithm is, the stronger the global diversity of the discrete cuckoo algorithm is; p is a radical ofrThe smaller the discrete cuckoo algorithm, the stronger the convergence.
In addition, the binary coding method of Levy flight updates is prone to produce abnormal coding under constraints, such as when the kth user in the mth nest of the ith iteration selects more than one beam. In order to ensure that the solutions are feasible, a certain coding repair strategy must be adopted, and a heuristic greedy algorithm is used for repairing abnormal codes. In this embodiment, the amplitude of the item encoded as 1 updated by Levy flight is calculated first, and the kth user selects the first k-1 unselected beams with the largest amplitude value, and sequentially selects them until all users select one beam. Before repair, the inputs are: the output of a binary coding matrix with the size of NxK after restoration is as follows: the size of the composite constraint is a binary coded matrix of N × K. Specifically, the repair method for abnormal codes includes the following steps (steps S26-28).
Step S26, sequentially calculating the state value x of the nth antenna of the kth user of the mth bird nestnkThat is, calculating the state value of the n row at the k column of the m matrix and determining the state value xnkWhether or not it is 1;
at the state value xnkIf it is 1, step S27 is executed to calculate the corresponding channel gain gnk；
All state values x in each column of the decision binary coding matrixnkThen, step S28 is executed to select the beam and the beam set of the user by the following formula:
Figure BDA0002150041800000111
Figure BDA0002150041800000112
wherein, betakIs the beam of the k-th user,
Figure BDA0002150041800000113
a set of beams that has been selected for the user.
Step S3: after all the bird nests are screened, comparing the bird nest discovery probability with random numbers which obey uniform distribution, if the random numbers are larger than the bird nest discovery probability, copying the bird nest of the global optimal solution to replace one of the found bird nests, and randomly changing the positions of the other found bird nests to determine the position and the optimal value of the bird nest of the current optimal solution. In this example, the probability of discovery P with foreign eggs is usedaAnd random number R epsilon [0,1 ] subjected to uniform distribution]Making a comparison if R > PaThen, the idea of replication in the Genetic Algorithm (GA) is adopted to replicate the globally optimal bird nest to replace one of the found bird nests, and the other found bird nests randomly change the position of the bird nest, so that the optimal bird nest can be stored, and the current optimal bird nest position and the optimal value can be determined.
Step S4: after determining the position and the optimal value of the bird nest, judging whether the iteration number reaches the maximum iteration number (T)max) If yes, the global optimum solution is output, otherwise, step S2 is executed.
In summary, the beam selection method applied to the millimeter wave massive MIMO system of the embodiment has the following advantages:
the beam selection method applied to the millimeter wave large-scale MIMO system considers the beam selection as solving the {0-1} knapsack problem and solves the problem, reduces the number of required radio frequency links without causing obvious performance loss, can avoid overhigh energy loss caused by adopting a large number of radio frequency links required by full digital coding, and can obtain near-excellent system performance on the premise of reducing complexity. Because abnormal coding is easy to occur in the solving process, the abnormal solutions are repaired by the method, and the calculated solutions are guaranteed to be feasible solutions. Meanwhile, the embodiment copies the globally optimal bird nest to replace the found bird nest, so that the optimal bird nest is saved, the convergence of the algorithm is accelerated, and in addition, the interference among users is not considered, and the optimization and the rate are directly taken as targets, so compared with the existing several beam selection schemes, the implementation and the rate performance are the best. In addition, the embodiment adds application scenes to the discrete cuckoo search algorithm.
In the embodiment, the beam selection method regards the selected beam as a selected article to be loaded into the backpack, considers the maximum sum rate of the system as solving the problem of the maximum capacity of the backpack, obtains a near-optimal solution by adopting an improved discrete cuckoo search algorithm, and repairs abnormal codes generated in a Levy flight discretization result of the cuckoo algorithm by adopting a heuristic greedy algorithm. Meanwhile, in the embodiment, the copy in the genetic algorithm is introduced into the DCS algorithm, the globally optimal bird nest is copied to replace the found bird nest, the convergence speed of the algorithm is accelerated, and analysis proves that the beam selection scheme is superior to the existing beam selection scheme in the aspect of improving the sum rate of the system.
Example 2
Referring to fig. 2, the present embodiment provides a beam selection method applied to the mm-wave massive MIMO system, which adds a step on the basis of embodiment 1 (step S0). Step S0 is: and constructing a millimeter wave large-scale MIMO system. In this embodiment, consider a mmwave massive MIMO single-cell system, assuming that the number of antennas equipped at the base station end is N, and the number of RF links is NRFSatisfy N > NRF. The base station serves K single-antenna users simultaneously. In order not to lose generality, the present embodiment assumes that K is equal to NRF. Therefore, the construction method of the millimeter wave massive MIMO system includes the following steps (steps S01-S04).
In step S01, an expression formula of the received signal of the kth user is preliminarily defined. In the millimeter wave massive MIMO system of the present embodiment, the expression formula of the received signal of the kth user is initially defined as:
y＝HHWs+n
where H is the channel matrix, H ═ H1,h2,...,hK]，hkIs the channel vector between the kth user and the base station; s is the original signal vector, s belongs to CK×1And normalizing the power E (ss)H)＝IK(ii) a W is a precoding matrix of size NxK, and tr (WW) is satisfiedH) Rho is less than or equal to the total emissionPower; n is additive white Gaussian noise having a size of K × 1, and n to CN (0, σ)2IK). In this embodiment, a Saleh-Valenzuela channel model widely used for millimeter wave communication is adopted, and a channel vector of user k is:
Figure BDA0002150041800000131
in the formula (I), the compound is shown in the specification,
Figure BDA0002150041800000132
for the line-of-sight path of the kth user,
Figure BDA0002150041800000133
a non-line-of-sight path for the kth user; gkFor complex gain,. psikIn the form of a spatial orientation,
Figure BDA0002150041800000134
is an array response vector.
Step S02, using a Discrete Lens Array (DLA), converting the conventional spatial channel into a beam spatial channel through a spatial fourier transform matrix, and redefining an expression formula of the reception signal of the kth user according to the beam spatial channel. The spatial fourier transform matrix is a given set of orthogonal bases:
Figure BDA0002150041800000135
in the formula (I), the compound is shown in the specification,
Figure BDA0002150041800000136
thus, in the millimeter wave massive MIMO system, the expression formula of the reception signal of the k-th user is defined again as:
Figure BDA0002150041800000137
wherein the content of the first and second substances,
Figure BDA0002150041800000138
is a beam space channel, and the calculation formula is:
Figure BDA0002150041800000139
Figure BDA00021500418000001310
is the beam space channel vector of the kth user, and K is 1. Each element here represents the channel gain provided by a predefined beam.
And step S03, only selecting partial wave beams according to sparsity of the MIMO channel of the wave beam space, and establishing an objective function and constraint conditions selected by the antenna. Due to the sparsity of the channel, when calculating the transmission rate
Figure BDA00021500418000001311
Only a few main elements exist in the MIMO system, and the dimensionality of the MIMO system can be reduced on the premise of not causing obvious performance loss by only selecting a small number of proper beams. Thus, the received signals for the K users downlink can be expressed as:
Figure BDA00021500418000001312
in the formula (I), the compound is shown in the specification,
Figure BDA0002150041800000141
for the channel formed by the selected beam, Wr∈CK×KAnd is a reduced dimensionality digital precoding matrix.
And step S04, establishing an objective function and a constraint condition for antenna selection. K antennas are selected from the N antennas without causing significant performance loss, and an objective function and a constraint condition need to be established for the K antennas. Since the beam selection is performed in the analog domain, zero-forcing (ZF) precoding is used, which can be expressed as:
Figure BDA0002150041800000142
wherein α is a scaling factor and satisfies
Figure BDA0002150041800000143
ρ is the total transmitted power.
Assuming equal power is allocated to each user at the base station, the achievable average rate for the kth user is:
Figure BDA0002150041800000144
in the formula, σ2Is the noise power.
Further, an objective function is determined, which is:
Figure BDA0002150041800000145
xnkselecting the state value of the nth antenna for the kth user; rkIs the achievable average rate for the k-th user.
Furthermore, the expression that each user selects only one beam is:
Figure BDA0002150041800000146
each beam is selected by at most one user, and may not be selected by any user, and the corresponding expression is:
Figure BDA0002150041800000147
xnkthe value range of (1) is 0 or 1, 0 indicates that the antenna is not selected, 1 indicates that the antenna is selected, and the specific expression formula is as follows:
xnk∈{0,1},n＝1,2,...,N k＝1,2,...,K
on the basis of the above, beam selection model optimization, i.e., the work of example 1, can be performed.
In this embodiment, a simulation experiment is performed, where the specific setting conditions of simulation experiment parameters are as follows: suppose that the spatial channel of user k has one LoS component and 2 NLoS components, where
Figure BDA0002150041800000151
Figure BDA0002150041800000152
And
Figure BDA0002150041800000153
following a distribution interval in
Figure BDA0002150041800000154
Is uniformly distributed. The present embodiment considers two millimeter wave massive MIMO systems, i.e., a sparse system and a dense system, wherein a base station of the sparse system is equipped with 256 antennas and 32 users, a base station of the dense system is equipped with 32 users with 64 antennas and 32 users with 32 antennas, a transmission power ρ is 32mW, a bird nest number M is 20, Pa＝0.75，Pr＝0.5。
The convergence of the algorithm proposed in this embodiment is analyzed first. Setting the maximum number of iterations TmaxReferring to fig. 3, it can be seen that the maximum sum rate has been obtained substantially for about 100 iterations based on the DCS beam selection algorithm in the present embodiment, and in order to reduce the amount of computation, the maximum iteration number T is used in the subsequent simulations to reduce the amount of computationmax＝100。
Referring to fig. 4, the DCS-based beam selection algorithm is compared with the all-digital precoding algorithm, the IA algorithm, and the MM algorithm under the sparse system condition. It can be seen from fig. 4 that the IA algorithm is substantially consistent with MM-2 beam selection and rate performance for 2 beams per user, much better than MM-1 beam selection and rate performance for 1 beam per user, because the MM-1 scheme does not take into account interference between users, and the MM-2 achieved sum rate is not improved a little but results in higher energy consumption. The DCS beam selection algorithm realizes better sum rate performance than the three algorithms and is closer to a full digital system.
Referring to fig. 5, a comparison of DCS-based beam selection algorithm with existing algorithms for achieving sum rates under dense system conditions is shown. In the IA algorithm, there is a high probability that different users select the same antenna in a dense system, and the more users are interfered, the higher the complexity of the IA algorithm is. The DCS-based beam selection algorithm does not consider the interference among users, directly aims at optimizing the sum rate, and therefore achieves the best sum rate performance. Fig. 6 is a comparison between the number of users and the rate under the conditions of SNR 30dB and N100, and it can be seen from the figure that the best performance is obtained based on the DCS beam selection algorithm.
Example 3
The present embodiment provides a beam selection apparatus applied to a millimeter wave massive MIMO system, to which the beam selection method applied to the millimeter wave massive MIMO system of embodiment 1 or embodiment 2 is applied. The beam selection device comprises a fitness calculation module, a bird nest screening module, a bird nest position replacement module and an iteration number judgment module.
The fitness calculation module is used for firstly defining the number of bird nests, the bird nest discovery probability, the binary coding control parameters, the maximum iteration times, the number of antennas and the number of users of a millimeter wave large-scale MIMO system, then initializing a plurality of bird nests, enabling each bird nest to select a wave beam with the maximum channel amplitude and no repetition, finally calculating the fitness of the plurality of bird nests, and taking the maximum fitness of the current bird nest as a global optimal solution; the calculation formula of the fitness of the bird nests is as follows:
Figure BDA0002150041800000161
in the formula, xnkSelecting the state value of the nth antenna for the kth user; rkThe achievable average rate for the kth user; whereinThe average rate is calculated as:
Figure BDA0002150041800000162
in the formula, σ2Is the noise power; alpha is a scaling factor and satisfies
Figure BDA0002150041800000163
Rho is total transmitting power, and K is the number of users; h is the channel matrix, HkIs the channel vector between the kth user and the base station;
Figure BDA0002150041800000164
forming channels for the selected beams;
Figure BDA0002150041800000165
Wr∈CK×Kand is a reduced dimensionality digital precoding matrix.
The bird nest screening module is used for carrying out binary code mixed updating on a plurality of bird nests firstly after the fitness of the plurality of bird nests is calculated by the fitness calculating module, repairing abnormal codes, calculating the fitness of the newly generated plurality of bird nests, and finally screening the bird nests by retaining the bird nests with high fitness.
The bird nest position replacing module is used for comparing the bird nest discovery probability with random numbers obeying uniform distribution after the bird nest screening module screens all bird nests; if the random number is greater than the bird nest discovery probability, the bird nest position replacement module copies the bird nest of the global optimal solution to replace one found bird nest, and randomly changes the positions of the rest found bird nests to determine the bird nest position and the optimal value of the current optimal solution.
The iteration frequency judging module is used for judging whether the iteration frequency reaches the maximum iteration frequency after the bird nest position replacing module determines the bird nest position and the optimal value, if so, outputting a global optimal solution, and otherwise, driving the bird nest screening module to work.
Example 4
The present embodiments provide a computer terminal comprising a memory, a processor, and a computer program stored on the memory and executable on the processor. The processor implements the steps of the beam selection method applied to the millimeter wave massive MIMO system of embodiment 1 or embodiment 2 when executing the program.
When the beam selection method applied to the millimeter wave large-scale MIMO system in embodiment 1 or embodiment 2 is applied, the method may be applied in a software form, for example, a program designed to run independently is installed on a computer terminal, and the computer terminal may be a computer, a smart phone, a control system, other internet of things devices, and the like. The beam selection method applied to the millimeter wave massive MIMO system in embodiment 1 or embodiment 2 may also be designed as a program for embedded operation, and installed on a computer terminal, such as a single chip microcomputer.
Example 5
The present embodiment provides a computer-readable storage medium having a computer program stored thereon. The program, when executed by a processor, implements the steps of the beam selection method applied to the millimeter wave massive MIMO system of embodiment 1 or embodiment 2.
The beam selection method applied to the millimeter wave massive MIMO system in embodiment 1 or embodiment 2 may be applied in the form of software, for example, a program designed to be independently run by a computer-readable storage medium, which may be a usb disk designed as a usb shield, and a program designed to start the whole method by external triggering through the usb disk.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents and improvements made within the spirit and principle of the present invention are intended to be included within the scope of the present invention.

Claims (9)

1. A wave beam selection method applied to a millimeter wave massive MIMO system is characterized by comprising the following steps:
step S1, first defining the bird nest number, bird nest discovery probability, binary coding control parameters, maximum iteration times, antenna number and user number of the millimeter wave large-scale MIMO system, then initializing a plurality of bird nests, enabling each bird nest to select a wave beam with the largest channel amplitude value and no repetition for a user, finally calculating the fitness of the plurality of bird nests, and taking the maximum fitness of the current bird nest as a global optimal solution;
step S2, after calculating the fitness of a plurality of bird nests, firstly, performing binary code mixed updating on the plurality of bird nests, repairing abnormal codes, then calculating the fitness of the newly generated plurality of bird nests, and finally screening the bird nests by reserving the bird nests with larger fitness;
the abnormal code repairing method comprises the following steps:
step S26, sequentially calculating the state value x of the nth antenna of the kth user of the mth bird nestnkThat is, calculating the state value of the n row at the k column of the m matrix and determining the state value xnkWhether or not it is 1;
at the state value xnkIf it is 1, step S27 is executed to calculate the corresponding channel gain gnk；
All state values x in each column of the binary coded matrix are judgednkThen, step S28 is executed to select the beam and the beam set of the user by the following formula:
Figure FDA0003492479970000011
Figure FDA0003492479970000012
wherein, betakIs the beam of the k-th user,
Figure FDA0003492479970000013
a selected set of beams for the user;
step S3, after all bird nests are screened, comparing the bird nest discovery probability with random numbers obeying uniform distribution, if the random numbers are greater than the bird nest discovery probability, copying the bird nest of the global optimal solution to replace one of the found bird nests, randomly changing the positions of the rest of the found bird nests, reserving the global optimal solution to accelerate the convergence speed of the algorithm, and determining the position and the optimal value of the bird nest of the current optimal solution;
and step S4, after determining the position and the optimal value of the bird nest, judging whether the iteration times reach the maximum iteration times, if so, outputting the global optimal solution, otherwise, executing step S2.
2. The beam selection method applied to the mmwave massive MIMO system of claim 1, wherein in step S2, the binary coding hybrid update method comprises the steps of:
step S21, judging whether the random number of the system is not larger than the binary coding control parameter;
when the random number is not greater than the binary coding control parameter, executing step S22, mapping a real number into discrete binary data by using a sigmoid function, and calculating a state value of the nth antenna of the kth user in the (m + 1) th iteration through the following formula
Figure FDA0003492479970000021
Figure FDA0003492479970000022
Figure FDA0003492479970000023
Wherein Step is the Step length of the Levy flight jump path;
when the random number is larger than the binary coding control parameter, executing a Step S23, and judging whether the Step size Step of the Levy flight jump path is a positive number;
when the Step size Step of the Levy flight jump path is not a positive number, Step S24 is executed to calculate the (m + 1) th iteration by the following formulaState value of nth antenna of kth user
Figure FDA0003492479970000024
Figure FDA0003492479970000025
Figure FDA0003492479970000026
When the Step size Step of the Levy flight jump path is a positive number, Step S25 is executed, and the state value of the nth antenna of the kth user in the (m + 1) th iteration is calculated by the following formula
Figure FDA0003492479970000027
Figure FDA0003492479970000028
Figure FDA0003492479970000029
Wherein rand is the random number.
3. The beam selection method for the mmwave massive MIMO system of claim 1, wherein the beam selection method further comprises the steps of:
step S0, constructing the millimeter wave large-scale MIMO system; the construction method of the millimeter wave large-scale MIMO system comprises the following steps:
step S01, preliminarily defining an expression formula of the receiving signal of the kth user;
step S02, adopting a discrete lens array, converting the traditional space channel into a beam space channel through a space Fourier transform matrix, and redefining an expression formula of a receiving signal of the kth user according to the beam space channel;
and step S03, only selecting partial wave beams according to sparsity of the MIMO channel of the wave beam space, and establishing an objective function and constraint conditions selected by the antenna.
4. The method as claimed in claim 3, wherein in the mmwave massive MIMO system of the present embodiment, the expression formula of the received signal of the kth user is initially defined as:
y＝HHWs+n
where H is the channel matrix, H ═ H1,h2,...,hK]，hkIs the channel vector between the kth user and the base station; s is the original signal vector, s belongs to CK×1And normalizing the power E (ss)H)＝IK(ii) a W is a precoding matrix of size NxK, and tr (WW) is satisfiedH) Rho is not more than rho, and rho is total emission power; n is additive white Gaussian noise having a size of K × 1, and n to CN (0, σ)2IK)。
5. The method of claim 4, wherein the channel vector is calculated by the following formula:
Figure FDA0003492479970000031
in the formula (I), the compound is shown in the specification,
Figure FDA0003492479970000032
for the line-of-sight path of the kth user,
Figure FDA0003492479970000033
a non-line-of-sight path for the kth user; gkFor complex gain,. psikIn the form of a spatial orientation,
Figure FDA0003492479970000034
and
Figure FDA0003492479970000035
the gain and direction of the ith non-line-of-sight path for the kth user respectively,
Figure FDA0003492479970000036
for array response vectors, L is the total number of non-line-of-sight paths.
6. The method of claim 5, wherein the spatial Fourier transform matrix is:
Figure FDA0003492479970000037
in the formula (I), the compound is shown in the specification,
Figure FDA0003492479970000041
thus, in the millimeter wave massive MIMO system, the expression formula of the reception signal of the k-th user is defined again as:
Figure FDA0003492479970000042
wherein the content of the first and second substances,
Figure FDA0003492479970000043
is a beam space channel, and the calculation formula is:
Figure FDA0003492479970000044
Figure FDA0003492479970000045
is the beam space channel vector of the kth user, and K is 1.
7. The beam selection method applied to the mmwave massive MIMO system of claim 6, wherein the expression formula of the downlink reception signal is:
Figure FDA0003492479970000046
in the formula (I), the compound is shown in the specification,
Figure FDA0003492479970000047
for the channel formed by the selected beam, Wr∈CK×KAnd is a reduced dimensionality digital precoding matrix.
8. The method of claim 6, wherein the objective function is:
Figure FDA0003492479970000048
in the formula, xnkThe state value of the nth antenna of the kth user; rkThe achievable average rate for the kth user; wherein, the calculation formula of the average speed is as follows:
Figure FDA0003492479970000049
in the formula, σ2In order to be able to measure the power of the noise,
Figure FDA00034924799700000410
and Wr∈CK×K(ii) a Alpha is a scaling factor and satisfies
Figure FDA00034924799700000411
ρ is the total transmitted power.
9. A beam selection apparatus applied to a millimeter wave massive MIMO system, which applies the beam selection method applied to the millimeter wave massive MIMO system as claimed in any one of claims 1 to 8, characterized in that it comprises:
the fitness calculation module is used for firstly defining the number of bird nests, the bird nest discovery probability, the binary coding control parameters, the maximum iteration times, the number of antennas and the number of users of a millimeter wave large-scale MIMO system, then initializing a plurality of bird nests, enabling each user in the bird nests to select a wave beam with the maximum channel amplitude and no repetition, finally calculating the fitness of the plurality of bird nests, and taking the maximum fitness of the current bird nest as a global optimal solution; the calculation formula of the fitness of the bird nests is as follows:
Figure FDA0003492479970000051
in the formula, xnkThe state value of the nth antenna of the kth user in the mth bird nest; rkThe achievable average rate for the kth user; wherein, the calculation formula of the average speed is as follows:
Figure FDA0003492479970000052
in the formula, σ2For noise power, α is a scaling factor and satisfies
Figure FDA0003492479970000053
Rho is total transmitting power, and K is the number of users; h is the channel matrix, HkIs the channel vector between the kth user and the base station;
Figure FDA0003492479970000054
forming channels for the selected beams;
Figure FDA0003492479970000055
Wr∈CK×Kand is a reduced dimensionality digital precoding matrix;
the bird nest screening module is used for firstly carrying out binary code mixed updating on the plurality of bird nests after the fitness calculation module calculates the fitness of the plurality of bird nests, repairing abnormal codes, calculating the fitness of the plurality of newly generated bird nests, and finally screening the bird nests by reserving the bird nests with higher fitness;
a bird nest position replacement module for comparing the bird nest discovery probability with random numbers subject to uniform distribution after the bird nest screening module screens all bird nests; if the random number is greater than the bird nest discovery probability, the bird nest position replacement module copies the bird nest of the global optimal solution to replace one found bird nest, and randomly changes the positions of the rest found bird nests to determine the bird nest position and the optimal value of the current optimal solution; and the iteration frequency judging module is used for judging whether the iteration frequency reaches the maximum iteration frequency after the bird nest position replacing module determines the bird nest position and the optimal value, if so, outputting the global optimal solution, and otherwise, driving the bird nest screening module to work.
CN201910705982.7A 2019-07-31 2019-07-31 Beam selection method and device applied to millimeter wave large-scale MIMO system Active CN110311719B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910705982.7A CN110311719B (en) 2019-07-31 2019-07-31 Beam selection method and device applied to millimeter wave large-scale MIMO system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910705982.7A CN110311719B (en) 2019-07-31 2019-07-31 Beam selection method and device applied to millimeter wave large-scale MIMO system

Publications (2)

Publication Number Publication Date
CN110311719A CN110311719A (en) 2019-10-08
CN110311719B true CN110311719B (en) 2022-03-29

Family

ID=68082847

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910705982.7A Active CN110311719B (en) 2019-07-31 2019-07-31 Beam selection method and device applied to millimeter wave large-scale MIMO system

Country Status (1)

Country Link
CN (1) CN110311719B (en)

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112261728A (en) * 2020-12-22 2021-01-22 之江实验室 Beam selection matrix design method based on lens array
CN113114319B (en) * 2021-04-20 2022-02-15 安徽大学 Joint optimization method based on beam selection and interference elimination and application thereof
CN113824478B (en) * 2021-10-11 2023-07-04 北京邮电大学 Broadband millimeter wave multi-user large-scale MIMO uplink spectrum efficiency optimization method assisted by discrete lens antenna array
CN114024589B (en) * 2021-11-11 2024-02-06 深圳大学 MISO communication system design method and device
CN115102586B (en) * 2022-06-18 2024-04-05 河南工业大学 Beam selection method and device based on cuckoo search and ant colony optimization

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109299570A (en) * 2018-10-25 2019-02-01 中国人民解放军陆军工程大学 Array aerial direction figure self-repair method

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8537737B2 (en) * 2010-03-11 2013-09-17 Nec Laboratories America, Inc. Scalable video multicast with non-overlapping beamforming antennas

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109299570A (en) * 2018-10-25 2019-02-01 中国人民解放军陆军工程大学 Array aerial direction figure self-repair method

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
Thinning of Concentric Circular Antenna Arrays Using Improved Discrete Cuckoo Search Algorithm;Geng Sun等;《2017 IEEE Wireless Communications and Networking Conference (WCNC)》;20170511;全文 *
二进制布谷鸟搜索算法;冯登科等;《计算机应用》;20130601(第06期);全文 *

Also Published As

Publication number Publication date
CN110311719A (en) 2019-10-08

Similar Documents

Publication Publication Date Title
CN110311719B (en) Beam selection method and device applied to millimeter wave large-scale MIMO system
US11381286B2 (en) Encoding and decoding of information for wireless transmission using multi-antenna transceivers
Gao et al. Machine learning inspired energy-efficient hybrid precoding for mmWave massive MIMO systems
Chen et al. Intelligent massive MIMO antenna selection using Monte Carlo tree search
CN109921882B (en) Deep learning-based MIMO decoding method, device and storage medium
Yang et al. Machine learning enabling analog beam selection for concurrent transmissions in millimeter-wave V2V communications
CN107294590B (en) Digital-analog hybrid beam forming method based on uplink training
CN113411110A (en) Millimeter wave communication beam training method based on deep reinforcement learning
US20210320709A1 (en) Apparatus and method for designing a grid-of-beams using machine learning
WO2022174418A1 (en) Wireless telecommunications network
US11871251B2 (en) Method of association of user equipment in a cellular network according to a transferable association policy
CN112564754B (en) Wave beam selection method based on self-adaptive cross entropy under millimeter wave Massive MIMO system
CN113261016A (en) Single-shot multi-user multiple-input multiple-output (MU-MIMO) resource pairing using Deep Q Network (DQN) based reinforcement learning
Zhang et al. Intelligent beam training for millimeter-wave communications via deep reinforcement learning
Fredj et al. Distributed uplink beamforming in cell-free networks using deep reinforcement learning
Elbir et al. Cognitive learning-aided multi-antenna communications
Yu et al. Deep Reinforcement Learning-Based NOMA-Aided Slotted ALOHA for LEO Satellite IoT Networks
EP3878114A1 (en) Processing of uplink data streams
KR102046033B1 (en) Method and apparatus for providing hybrid beamforming in large-scale antenna system
Huang et al. Self-attention reinforcement learning for multi-beam combining in mmWave 3D-MIMO systems
Yan et al. Machine learning-based beamforming design for millimeter wave IRS communications with discrete phase shifters
KR101482925B1 (en) Zero-forcing Beamformer Design Device and Method in MISO Broadcast Channel Based on k-regularity
JP2023502488A (en) Method for wireless X2X access and receiver for large-scale multi-dimensional wireless systems
Akbarpour-Kasgari et al. Deep Reinforcement Learning in mmW-NOMA: Joint Power Allocation and Hybrid Beamforming
Ding et al. Grouping optimization based hybrid beamforming for multiuser mmWave massive MIMO systems

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant