CN110365375B - Beam alignment and tracking method in millimeter wave communication system and computer equipment - Google Patents

Beam alignment and tracking method in millimeter wave communication system and computer equipment Download PDF

Info

Publication number
CN110365375B
CN110365375B CN201910560646.8A CN201910560646A CN110365375B CN 110365375 B CN110365375 B CN 110365375B CN 201910560646 A CN201910560646 A CN 201910560646A CN 110365375 B CN110365375 B CN 110365375B
Authority
CN
China
Prior art keywords
option
alignment
training
time
tracking
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
CN201910560646.8A
Other languages
Chinese (zh)
Other versions
CN110365375A (en
Inventor
黄永明
刘婷薇
章建军
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.)
Southeast University
Original Assignee
Southeast 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 Southeast University filed Critical Southeast University
Priority to CN201910560646.8A priority Critical patent/CN110365375B/en
Publication of CN110365375A publication Critical patent/CN110365375A/en
Application granted granted Critical
Publication of CN110365375B publication Critical patent/CN110365375B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B17/00—Monitoring; Testing
    • H04B17/30—Monitoring; Testing of propagation channels
    • H04B17/391—Modelling the propagation channel
    • 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
    • H04B7/0408—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas using two or more beams, i.e. beam diversity
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L5/00—Arrangements affording multiple use of the transmission path
    • H04L5/003—Arrangements for allocating sub-channels of the transmission path
    • H04L5/0048—Allocation of pilot signals, i.e. of signals known to the receiver
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W16/00—Network planning, e.g. coverage or traffic planning tools; Network deployment, e.g. resource partitioning or cells structures
    • H04W16/24—Cell structures
    • H04W16/28—Cell structures using beam steering

Landscapes

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

Abstract

The invention discloses a beam alignment and tracking method in a millimeter wave communication system. The proposed method models the beam alignment and tracking problem as a random selection optimization problem, where each option is defined by an offset and the number of scanned beams. In the initial stage of each time slot, an option is determined by using a confidence interval upper bound algorithm, a beam set needing training is determined according to the option, then each beam in the training beam set is used for transmitting a pilot signal in turn, and the optimal transmitting beam is determined based on the received signal. And performing data transmission by using the found optimal sending beam in the data transmission stage of the corresponding time slot, and simultaneously updating the related information of the option and determining the option of the next time slot. The beam alignment and tracking method designed by the invention can sense the change rate of the environment, adaptively adjust the beam training strategy, improve the successful probability of beam alignment, reduce the beam training overhead and effectively improve the throughput of the system.

Description

Beam alignment and tracking method in millimeter wave communication system and computer equipment
Technical Field
The invention belongs to the field of wireless communication, and particularly relates to a beam alignment and tracking method in a millimeter wave communication system and computer equipment.
Background
With the continuous development of wireless communication technology, high-speed data services and ubiquitous access demands are increasing explosively. The millimeter wave has a large available bandwidth, and can meet the exponentially-increased wireless transmission demand, thereby drawing wide attention of people. But the path loss of the millimeter wave signal is larger compared to the microwave communication. In order to solve the problem of large path loss of millimeter waves, a large-scale antenna array must be installed at a receiving/transmitting end. Fortunately, the shorter wavelength of the millimeter wave signal facilitates the placement of a large number of antennas in a small space. In wireless communication, channel state information plays an extremely important role in achieving high-speed transmission, but when the number of antennas is large and thus the beam width is narrow, it becomes more difficult to acquire Channel State Information (CSI), particularly in a dynamic environment such as a mobile environment.
Considering that the millimeter wave channel has sparsity, an effective method for obtaining CSI is through beam training/alignment, and accordingly obtains information such as an equivalent channel vector. The transmitting end/receiving end can determine the optimal transmitting/receiving beam through exhaustive search, adaptive search and other modes, but the beam search scheme is mainly suitable for single-user single-stream transmission, and the beam training overhead of a large-scale antenna array system is very large. More seriously, in dynamic or time-varying scenarios, the millimeter wave channel may change so quickly and with too short a coherence time that frequent and accurate beam alignment cannot be achieved. Although beam tracking techniques can be used to reduce training overhead for time-varying channels, beam tracking techniques are very demanding for channel modeling. Since channel modeling depends on the environment, factors influencing channel modeling include the positions of buildings, base stations and users, pedestrians, vehicles and the like, and it is difficult to provide a closed channel modeling expression for a practical complex scene, the conventional beam/channel tracking technology is only suitable for simpler or special occasions.
Frequent beam alignment in dynamic environments is unavoidable for better transmission performance and therefore more challenging. In order to reduce the overhead of beam alignment, effective information must be extracted from past beam alignment information to reduce the search space for subsequent beam alignment. With the rapid development of machine learning, low-overhead beam alignment is possible. The invention provides a beam alignment algorithm based on multi-arm gambling machine learning (bandit learning) from the perspective of sensing the rate of change of the environment by exploring/utilizing correlation techniques in machine/reinforcement learning. The proposed beam alignment method can sense the speed of environmental changes and learn the hidden probability structure information in the environment. Based on the sensed environment and probability structure information, the algorithm only needs to search a very small training beam set, and the corresponding training overhead can be ignored. Moreover, the proposed beam alignment algorithm can learn the required information from the environment, so that the prior knowledge such as the rigorous assumption required by time-varying channel modeling is not required, and the method has wider applicability.
Disclosure of Invention
The purpose of the invention is as follows: aiming at the problems in the prior art, the invention aims to provide a method, a device and equipment for aligning and tracking beams in a millimeter wave communication system, which are used for adaptively adjusting a beam training strategy, so that not only is the mismatching of the beams effectively avoided, but also the beam training overhead is reduced, and the throughput of the system is improved.
The technical scheme is as follows: in order to realize the purpose of the invention, the technical scheme adopted by the invention is as follows: a method for beam alignment and tracking in a millimeter wave communication system includes the following steps:
step (1): the beam alignment and tracking problem is modeled as a random optimization problem.
In this step, a randomly selected optimization model is constructed for a given millimeter wave communication system and a corresponding beam training codebook. For descriptive convenience, a mathematical model and a solving algorithm for random selection optimization are given, and then a beam alignment and tracking problem in communication is modeled into a corresponding random selection optimization problem.
Considering the most basic random selection optimization model, from a random variable family Xi,n(1 ≦ i ≦ K, n ≧ 1), where i represents an index of a random variable type and n represents an index of a random variable time. Successive selection/sampling of type i yields independent identically distributed returns, i.e., random variable Xi,1,Xi,2…. Recording random variable { Xi,nThe mathematical expectation of (u) } isiNote that the mathematics expects μiIs unknown. It should be noted that independence is also applicable to different types of rewards, namely Xi,sAnd Xj,tAre independent of one another (where s, t ≧ 1 and 1 ≦ i < j ≦ K) but generally obey different distributions. Given family of random variables or randomSelection model { Xi,nI is more than or equal to 1 and less than or equal to K, n is more than or equal to 1, and the optimization goal is to select the type of the random variable at each moment n according to a certain strategy so as to maximize the accumulated return.
One effective random variable type selection strategy is the Upper Confidence Bound (UCB) algorithm, which selects the type of random variable at the next time based on past operations and the obtained returns. The UCB algorithm first selects a random variable of each type once and obtains a corresponding reward, and simultaneously counts the usage of each type by niInitialisation to 1 and total counter n initialised to K. For the nth (n is more than or equal to K +1) operation, the selected types are as follows:
Figure BDA0002108165720000021
wherein the content of the first and second substances,
Figure BDA0002108165720000022
corresponding to the average return of type j after the first n-1 operations, njIndicating the number of times type j was selected in the first n-1 operations.
The following is a modeling of the beam alignment and tracking problem in the millimeter wave communication system as a random selection optimization problem, assuming the used beam training code notation:
Figure BDA0002108165720000023
wherein M represents the number of codebooks, and when the transmitting end uses the ith beam fiWhen sending signals, the receiving signal expression of the receiving end is as follows:
Figure BDA0002108165720000024
where P is signal transmission power, h is channel vector of transmitting and receiving end, s represents transmitted pilot symbol/sequence, niIs a complex gaussian noise vector.
As shown in fig. 2, each slot includes two phases, i.e., a beam training phase and a data transmission phase, a beam pairThe goal is to determine the best transmit and/or receive beams to improve the transmission efficiency of the system. Considering that the wave beam training occupies certain time resource, the practical reachable rate R is adoptedeffAs a performance index for measuring throughput of different beam alignment algorithms, the actual achievable rate is defined as:
Reff＝(1-tB/tS)log(1+P|hHfi|2)
wherein, tBAnd tSRespectively representing the beam training time and the total time of each slot.
In order to sense the environment and learn the hidden probability structure information in the environment, the selection type in the random selection optimization model is defined as an integer pair (a, b), wherein a represents the offset of the beam index used by the current time slot relative to the beam index used by the last time slot, b represents the number of beams used by the scanning beam space/subspace, and each type (a, b) corresponds to one beam scanning space
Figure BDA0002108165720000031
Where c represents the optimal beam index at the last time, and the corresponding single return when type (a, b) is selected is:
Figure BDA0002108165720000032
wherein, tBAnd tSRespectively representing the beam training time and the total time of each slot,
Figure BDA0002108165720000033
scanning a space corresponding to a beam
Figure BDA0002108165720000034
The best beam of (2) is formed by a set of all types
Figure BDA0002108165720000035
Size N, i.e.:
Figure BDA0002108165720000036
it should be noted that the value of N is related to the specific communication system and the usage environment, such as: n should be increased if the number of antennas at the transmitting end is large or the environment in which the antennas are used is changed drastically.
Step (2): the initialization operation, two steps, namely, determining an initial reference beam and initializing each type, namely, performing each type once respectively.
(2.1) finding an initial reference beam: using codebooks in sequence in the first time slot
Figure BDA0002108165720000037
The beam in (1) sends a pilot signal, and for the received signal of the ith beam mobile user terminal, the following is:
yi＝hHfis+ni
where h is the channel vector between the base station and the mobile subscriber,
Figure BDA0002108165720000038
representing a complex Gaussian noise vector, s representing a length LPPilot sequence of { y }iDetermine the best transmit beam as
Figure BDA0002108165720000039
Wherein i ∈ argmaxi||yi||/LPI.e. the index of the beam with the largest average received signal power.
(2.2) initialize each type: selecting each type once in each time slot of the next N time slots, and constructing corresponding beam scanning space for each type (a, b)
Figure BDA0002108165720000041
And transmitting a pilot signal using a beam in a beam scanning space and determining an initial backoff value
Figure BDA0002108165720000042
Figure BDA0002108165720000043
Wherein the content of the first and second substances,
Figure BDA0002108165720000044
to correspond to the beam space
Figure BDA0002108165720000045
The optimal transmission beam of (1) initializes each type of use count or the number of operations to 1, and makes the total number of operations N equal to N + 1.
And (3): the beam alignment and tracking operation is executed in the beginning stage or the beam training stage of each following time slot, and the related information is updated at the same time, which includes two steps:
(3.1) selecting a type using a confidence interval upper bound algorithm, i.e. as follows:
Figure BDA0002108165720000046
from which the arm (a) can be determined*,b*) Beam scanning space of
Figure BDA0002108165720000047
na,bIndicating the usage count of the arms (a, b).
(3.2) use of
Figure BDA0002108165720000048
The beam in (1) transmits a pilot signal, and the received signal can determine the space corresponding to the beam
Figure BDA0002108165720000049
The best transmit beam.
In particular, assume beam fiIs expressed as
Figure BDA00021081657200000410
The length of the pilot signal s being LpThen beam fiCorresponding average received power YiIs composed of
Figure BDA00021081657200000411
The beam having the largest average received power is regarded as the best transmission beam.
(3.3) transmitting data using the best beam obtained in step (3) for the remaining time of the slot, i.e., the effective data transmission stage, while updating n and type (a) as follows*,b*) The related information of (2):
n←n+1,
Figure BDA00021081657200000412
wherein the content of the first and second substances,
Figure BDA00021081657200000413
type of representation (a)*,b*) The count of the number of uses of (c),
Figure BDA00021081657200000414
type of representation (a)*,b*) The average return of (a) is,
Figure BDA00021081657200000415
type of representation (a)*,b*) The rate is effectively reached in this report.
Furthermore, the present invention also proposes a computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, which computer program, when loaded into the processor, implements the beam alignment and tracking method in a millimeter wave communication system as described.
Has the advantages that: compared with the prior art, the technical scheme of the invention has the following beneficial technical effects:
(1) the beam alignment and tracking method disclosed by the invention can sense the change rate of the environment, adaptively adjust the beam training strategy, effectively avoid beam mismatch and improve the success rate of beam alignment.
(2) The beam alignment and tracking method disclosed by the invention has low beam training overhead and reserves more time for data transmission, thereby effectively improving the throughput of the system.
Drawings
FIG. 1 is an algorithmic flow chart of a beam alignment and tracking method of an embodiment of the present invention;
FIG. 2 is a diagram illustrating beam training and data transmission time allocation for each slot in an embodiment of the present invention;
FIG. 3 is a diagram illustrating a dynamic communication environment in an embodiment of the present invention;
FIG. 4 is an alignment success rate performance curve of the beam alignment and tracking method in an embodiment of the present invention;
fig. 5 is a plot of achievable rate performance for the beam alignment and tracking method in an embodiment of the invention.
Detailed Description
The invention is further elucidated with reference to the drawings and the embodiments.
As shown in fig. 1, the method for beam alignment and tracking in a millimeter wave communication system disclosed in the embodiment of the present invention mainly includes the following steps:
step (1): the beam alignment and tracking problem is modeled as a random optimization problem.
In this step, a randomly selected optimization model is constructed for a given millimeter wave communication system and a corresponding beam training codebook. For descriptive convenience, a mathematical model and a solving algorithm for random selection optimization are given, and then a beam alignment and tracking problem in communication is modeled into a corresponding random selection optimization problem.
Considering the most basic random selection optimization model, from a random variable family Xi,n(1 ≦ i ≦ K, n ≧ 1), where i represents an index of a random variable type and n represents an index of a random variable time. Successive selection/sampling of type i yields independent identically distributed returns, i.e., random variable Xi,1,Xi,2…. Recording random variable { Xi,nThe mathematical expectation of (u) } isiNote that the mathematics expects μiIs unknown. Finger for patientIt is noted that independence is also applicable to different types of rewards, namely Xi,sAnd Xj,tAre independent of one another (where s, t ≧ 1 and 1 ≦ i < j ≦ K) but generally obey different distributions. Given a family of random variables or a randomly selected model { Xi,nI is more than or equal to 1 and less than or equal to K, n is more than or equal to 1, and the optimization goal is to select the type of the random variable at each moment n according to a certain strategy so as to maximize the accumulated return.
One effective random variable type selection strategy is the Upper Confidence Bound (UCB) algorithm, which selects the type of random variable at the next time based on past operations and the obtained returns. The UCB algorithm first selects a random variable of each type once and obtains a corresponding reward, and simultaneously counts the usage of each type by niInitialisation to 1 and total counter n initialised to K. For the nth (n is more than or equal to K +1) operation, the selected types are as follows:
Figure BDA0002108165720000061
wherein the content of the first and second substances,
Figure BDA0002108165720000062
corresponding to the average return of type j after the first n-1 operations, njIndicating the number of times type j was selected in the first n-1 operations.
The following is a modeling of the beam alignment and tracking problem in the millimeter wave communication system as a random selection optimization problem, assuming the used beam training code notation:
Figure BDA0002108165720000063
wherein M represents the number of codebooks, and when the transmitting end uses the ith beam fiWhen sending signals, the receiving signal expression of the receiving end is as follows:
Figure BDA0002108165720000064
wherein, P is signal transmission power, h is channel vector of transmitting and receiving end, s represents transmissionSending pilot symbols/sequences, niIs a complex gaussian noise vector.
As shown in fig. 2, each slot includes two phases, i.e., a beam training phase and a data transmission phase, and the aim of beam alignment is to determine an optimal transmission and/or reception beam to improve the transmission efficiency of the system. Considering that the wave beam training occupies certain time resource, the practical reachable rate R is adoptedeffAs a performance index for measuring throughput of different beam alignment algorithms, the actual achievable rate is defined as:
Reff＝(1-tB/tS)log(1+P|hHfi|2)
wherein, tBAnd tSRespectively representing the beam training time and the total time of each slot.
In order to sense the environment and learn the hidden probability structure information in the environment, the selection type in the random selection optimization model is defined as an integer pair (a, b), wherein a represents the offset of the beam index used by the current time slot relative to the beam index used by the last time slot, b represents the number of beams used by the scanning beam space/subspace, and each type (a, b) corresponds to one beam scanning space
Figure BDA0002108165720000065
Where c represents the optimal beam index at the last time, and the corresponding single return when type (a, b) is selected is:
Figure BDA0002108165720000066
wherein, tBAnd tSRespectively representing the beam training time and the total time of each slot,
Figure BDA0002108165720000067
scanning a space corresponding to a beam
Figure BDA0002108165720000068
The best beam of (2) is formed by a set of all types
Figure BDA0002108165720000069
Size N, i.e.:
Figure BDA00021081657200000610
it should be noted that the value of N is related to the specific communication system and the usage environment, such as: n should be increased if the number of antennas at the transmitting end is large or the environment in which the antennas are used is changed drastically.
Step (2): the initialization operation, two steps, namely, determining an initial reference beam and initializing each type, namely, performing each type once respectively.
(2.1) finding an initial reference beam: using codebooks in sequence in the first time slot
Figure BDA0002108165720000071
The beam in (1) sends a pilot signal, and for the received signal of the ith beam mobile user terminal, the following is:
yi＝hHfis+ni
where h is the channel vector between the base station and the mobile subscriber,
Figure BDA0002108165720000072
representing a complex Gaussian noise vector, s representing a length LPPilot sequence of { y }iDetermine the best transmit beam as
Figure BDA00021081657200000714
Wherein i ∈ argmaxi||yi||/LPI.e. the index of the beam with the largest average received signal power.
(2.2) initialize each type: selecting each type once in each time slot of the next N time slots, and constructing corresponding beam scanning space for each type (a, b)
Figure BDA0002108165720000073
And using beam sweepingPilot signals are transmitted for beams in space and an initial return value is determined
Figure BDA0002108165720000074
Figure BDA0002108165720000075
Wherein the content of the first and second substances,
Figure BDA0002108165720000076
to correspond to the beam space
Figure BDA0002108165720000077
The optimal transmission beam of (1) initializes each type of use count or the number of operations to 1, and makes the total number of operations N equal to N + 1.
And (3): the beam alignment and tracking operation is executed in the beginning stage or the beam training stage of each following time slot, and the related information is updated at the same time, which includes two steps:
(3.1) selecting a type using a confidence interval upper bound algorithm, i.e. as follows:
Figure BDA0002108165720000078
from which the arm (a) can be determined*,b*) Beam scanning space of
Figure BDA0002108165720000079
na,bIndicating the usage count of the arms (a, b).
(3.2) use of
Figure BDA00021081657200000710
The beam in (1) transmits a pilot signal, and the received signal can determine the space corresponding to the beam
Figure BDA00021081657200000711
The best transmit beam.
In particular, assume beam fiIs expressed as
Figure BDA00021081657200000712
The length of the pilot signal s being LpThen beam fiCorresponding average received power YiIs composed of
Figure BDA00021081657200000713
The beam having the largest average received power is regarded as the best transmission beam.
(3.3) transmitting data using the best beam obtained in step (3) for the remaining time of the slot, i.e., the effective data transmission stage, while updating n and type (a) as follows*,b*) The related information of (2):
n←n+1,
Figure BDA0002108165720000081
wherein the content of the first and second substances,
Figure BDA0002108165720000082
type of representation (a)*,b*) The count of the number of uses of (c),
Figure BDA0002108165720000083
type of representation (a)*,b*) The average return of (a) is,
Figure BDA0002108165720000084
type of representation (a)*,b*) The rate is effectively reached in this report.
The method for aligning and tracking the wave beam in the millimeter wave communication system provided by the embodiment of the invention can learn the hidden probability structure information in the environment from the perspective of sensing the change rate of the communication environment. Based on the learned probability structure information, the interval of the optimal beam of the next time slot can be effectively determined, and because the interval is usually very small, the corresponding training overhead can be ignored, and the throughput of the system is effectively improved. Moreover, the proposed algorithm can learn the required information from the environment, does not need prior knowledge related to channel modeling, and can be suitable for more complex dynamic scenes.
It should be noted that the beam alignment and tracking method in the millimeter wave communication system disclosed by the invention is suitable for various complex dynamic scenes. To facilitate an understanding of the aspects and effects of the present invention, the present invention provides another exemplary embodiment considering a communication system environment as shown in fig. 3. The motion trail of the mobile user is a circle, and the base station is positioned at the center of the circle. The speed of motion of the mobile user is assumed to be random, but obeys a certain probability distribution. Due to the randomness of the motion of the mobile user, the beams switched in the next time slot are also random and obey a certain probability distribution as well. In order to capture the random behavior of the mobile user, a Markov random process is introduced to model the beam switching process of different time slots. Specifically, let fcThe beam used by the current time slot is represented, and the base station of the next time slot is switched to { fc+1,fc+2,...,fc+SHas a probability p of one ofsStill using the current beam fcHas a probability of 1-ps. Is switching to { fc+1,fc+2,...,fc+SUnder the condition of a certain wave beam in the wave beam f, the base station is switched to the wave beam fc+kThe conditional probability of (2) decays exponentially with k, i.e.
Figure BDA0002108165720000085
Where β represents the decay rate.
Assuming that the antenna array used by the base station is a uniform linear array, and the number of antennas is N-64, the mobile subscriber only installs a single antenna. Codebook used by base station
Figure BDA0002108165720000086
By uniformly sampling the beam space configuration, namely:
Figure BDA0002108165720000087
where a (-) denotes an array response vector of the antenna array, and the codebook size is M ═ 64. Let d and λ denote the spacing between adjacent antennas and the signal wavelength, respectively, then the mathematical expression for a (-) is:
Figure BDA0002108165720000088
the channel model is an extended Saleh-Vallenzuela geometric channel model, and the channel vector h between the base station and the mobile user can be expressed as
Figure BDA0002108165720000091
Where L and β are the number of paths and the average path loss, respectively (β is set to 1 in the simulation), αlComplex path gain, phi, representing the first pathl＝cos(θl) Is the cosine of the angle of departure (where θlIndicating the exit angle of the first diameter). Assuming that the time consumed for transmitting a pilot symbol is a unit time tUWhen the length of the pilot signal is LpAnd the time consumed for selecting type (a, b) for beam training is tB＝bLptU。
The beam alignment and tracking method in the millimeter wave communication system provided by the embodiment comprises the following steps:
step (1): and (3) performing mathematical modeling on the beam alignment and tracking problem, and randomly selecting an optimization model by modeling, wherein the key of the mathematical modeling is a definition type. In the present embodiment, the number of types is K-9, i.e.
Figure BDA0002108165720000092
Where a represents an offset of a beam index used in a current slot with respect to a beam index used in a previous slot, and b represents a size of a beam scanning space.
Assuming a last time slotThe beam used for data transmission is fcThen the beam corresponding to type (a, b) scans the space
Figure BDA0002108165720000093
Is composed of
Figure BDA0002108165720000094
The corresponding single reward is:
Figure BDA0002108165720000095
wherein the content of the first and second substances,
Figure BDA0002108165720000096
to correspond to the beam space
Figure BDA0002108165720000097
The best transmit beam.
Step (2): the initialization operation mainly includes two steps, i.e., determining an initial reference beam and initializing each type (i.e., selecting and executing a beam training process corresponding to each type).
First, finding an initial reference beam: using codebooks in sequence in the first time slot
Figure BDA0002108165720000098
The beam in (1) transmits pilot signal, and for the received signal of the ith beam mobile user terminal is
yi＝hHfis+ni，
Where P is the transmit power, h is the channel vector between the base station and the mobile user,
Figure BDA0002108165720000099
representing a complex Gaussian noise vector, s representing a length LPThe pilot sequence of (1). By { yiDetermine the best transmit beam as
Figure BDA00021081657200000910
Wherein i ∈ argmaxi||yi||/LPI.e. the index of the beam with the largest average received signal power.
Second, initialize each type: performing a beam training operation for each type in the next time slot of K-9, and constructing a corresponding beam scanning space for each type (a, b)
Figure BDA0002108165720000101
And using the beam in the beam scanning space to transmit the pilot signal and determine the initial return value
Figure BDA0002108165720000102
Figure BDA0002108165720000103
Wherein the content of the first and second substances,
Figure BDA0002108165720000104
to correspond to the beam space
Figure BDA0002108165720000105
The best transmit beam. The usage count or the number of operations of each arm is initialized to 1, and the total number of operations is made n + K +1 to 10.
And (3): the beam alignment and tracking operation is executed in the beginning stage or the beam training stage of each following time slot, and the related information is updated at the same time, which includes two steps:
first, a confidence interval upper bound algorithm is used, i.e., a type is selected as follows:
Figure BDA0002108165720000106
from which the type (a) can be determined*,b*) Beam scanning space of
Figure BDA0002108165720000107
Second step, using in sequence
Figure BDA0002108165720000108
The beam in (1) transmits a pilot signal, and the received signal can determine the space corresponding to the beam
Figure BDA0002108165720000109
The best transmit beam. In particular, assume beam fiIs expressed as
Figure BDA00021081657200001010
The length of the pilot signal s being LpThen beam fiCorresponding average received power YiIs composed of
Figure BDA00021081657200001011
The beam having the largest average received power is regarded as the best transmission beam.
Thirdly, in the remaining time of the time slot, namely the effective data transmission stage, the data is transmitted by using the optimal beam obtained in the step (3), and the n and the type (a) are updated according to the following formula*,b*) The related information of (2):
n←n+1,
Figure BDA00021081657200001012
wherein the content of the first and second substances,
Figure BDA00021081657200001013
type of representation (a)*,b*) The count of the number of uses of (c),
Figure BDA00021081657200001014
type of representation (a)*,b*) The average return of (a) is,
Figure BDA00021081657200001015
type of representation (a)*,b*) The rate is effectively reached in this report.
The success rate of beam alignment for the beam alignment and tracking method in the millimeter wave communication system disclosed in this embodiment is shown in fig. 4. In order to show the superiority of the method disclosed by the present invention, fig. 4 also provides the beam alignment success rate performance of the other two algorithms (beam alignment algorithm ESBB based on exhaustive search and beam alignment algorithm DUBB directly based on UCB). The DUBB approach performs the worst in terms of alignment success rate among the three algorithms because the DUBB only chooses one beam at a time randomly, often with beam mismatch. The ESBB method performs best because the exhaustive search can find the best transmit beam. It can be seen from fig. 4 that the success rate of the beam alignment and tracking method disclosed in the present embodiment gradually approaches the ESBB as the signal-to-noise ratio increases.
The actual achievable rate performance of the beam alignment and tracking method disclosed in this embodiment is shown in fig. 5. In order to embody the superiority of the method disclosed by the present invention, fig. 5 also provides the actual achievable rate performance of the other three algorithms (beam alignment algorithm ESBB based on exhaustive search, beam alignment algorithm DUBB directly based on UCB, and algorithm-idealized algorithm based on Oracle, without training overhead and providing the best transmission beam each time). Compared with an ESBB algorithm and an algorithm directly based on UCB, the method for aligning and tracking the beam disclosed by the embodiment can obtain the optimal reachable rate performance, and is characterized in that the method disclosed by the invention can sense the change of the environment and adaptively adjust the beam training strategy, so that the training overhead is effectively reduced, more time is reserved for a data transmission stage, and the throughput of the system is improved.

Claims (4)

1. A method for beam alignment and tracking in a millimeter wave communication system, the method comprising the steps of:
(1) modeling a beam alignment and tracking problem in a millimeter wave communication system as a random selection optimization problem, wherein options in the optimization model are defined as integer pairs (a, b), and an optimization target is maximized cumulative return, namely a cumulative reachable rate;
the method of step (1) is specifically as follows:
(1.1) the beam training code used is assumed to be written as:
Figure FDA0003008313230000011
wherein M represents the number of codebooks, and when the transmitting end uses the ith beam fiWhen sending signals, the receiving signal expression of the receiving end is as follows:
Figure FDA0003008313230000012
where P is signal transmission power, h is channel vector of transmitting and receiving end, s represents transmitted pilot symbol/sequence, niIs a complex gaussian noise vector;
(1.2) each time slot comprises two stages, namely a beam training stage and a data transmission stage, the aim of beam alignment is to determine the optimal receiving and transmitting beam to improve the transmission efficiency of the system, and the actual achievable rate R is adopted in consideration of certain time resources occupied by beam trainingeffAs a performance index for measuring throughput of different beam alignment algorithms, the actual achievable rate is defined as:
Reff＝(1-tB/tS)log(1+P|hHfi|2)
wherein, tBAnd tSRespectively representing the beam training time and the total time of each time slot;
(1.3) defining each selectable item in the randomly selected optimization model as an integer pair (a, b), where a represents the offset of the beam index used by the current slot relative to the beam index used by the last slot, b represents the number of beams used when scanning the beam space/subspace, and each option (a, b) corresponds to one beam scanning space
Figure FDA0003008313230000018
Where c represents the optimal beam index at the last time, and the corresponding single return when (a, b) is selected is:
Figure FDA0003008313230000013
wherein, tBAnd tSRespectively representing the beam training time and the total time of each slot,
Figure FDA0003008313230000014
scanning a space corresponding to a beam
Figure FDA0003008313230000015
The best beam in the set of global options is
Figure FDA0003008313230000016
Its size is N, i.e.:
Figure FDA0003008313230000017
(2) initializing operation, namely determining an initial optimal reference beam and initializing each option;
(3) in the wave beam training stage of each time slot, a confidence interval upper bound algorithm is applied to determine an option, a corresponding training wave beam set is determined according to the option, wave beams in the set are used for sending pilot signals in sequence, and the best wave beam is found out according to the received signals.
2. The method for beam alignment and tracking in a millimeter wave communication system according to claim 1, wherein the specific method in step (2) is as follows:
(2.1) finding an initial reference beam: using codebooks in sequence in the first time slot
Figure FDA0003008313230000021
The beam in (1) sends a pilot signal, and for a received signal of the ith beam user terminal, the received signal is:
yi＝hHfis+ni
wherein the content of the first and second substances,h is the channel vector between the base station and the mobile user,
Figure FDA0003008313230000022
representing a complex Gaussian noise vector, s representing a length LPPilot sequence of { y }iDetermine the best transmission beam as
Figure FDA0003008313230000023
Wherein i*∈arg maxi||yi||/LPI is the index of the beam with the maximum average power of the received signal, and I is an identity matrix;
(2.2) initialize each option: selecting each option once in each of the next N time slots, and constructing corresponding beam scanning space for each option (a, b)
Figure FDA0003008313230000024
Transmitting a pilot signal using a beam in a beam sweep space and determining an initial reward value
Figure FDA0003008313230000025
Figure FDA0003008313230000026
Wherein the content of the first and second substances,
Figure FDA0003008313230000027
to correspond to the beam space
Figure FDA0003008313230000028
The number of times of operation or the use count of each option is initialized to 1, and the total number of operations is made N + 1.
3. The method for beam alignment and tracking in a millimeter wave communication system according to claim 2, wherein the specific method in step (3) is as follows:
(3.1) selecting an option using a confidence interval upper bound algorithm, i.e. as follows:
Figure FDA0003008313230000029
accordingly, determine (a)*,b*) Beam scanning space of
Figure FDA00030083132300000210
na,bRepresents the usage count of (a, b);
(3.2) use of
Figure FDA00030083132300000211
The beam in (1) transmits a pilot signal, and the beam space is determined from the received signal
Figure FDA00030083132300000212
The optimal transmit beam of (a);
suppose a beam fiIs expressed as
Figure FDA00030083132300000213
The length of the pilot signal s being LpThen beam fiCorresponding average received power YiComprises the following steps:
Figure FDA00030083132300000214
regarding the beam with the maximum average received power as the optimal transmission beam;
(3.3) transmitting data using the best beam obtained in step (3.2) for the remaining time of the slot, i.e., the effective data transmission stage, while updating n and option (a) as follows*,b*) The related information of (2):
Figure FDA0003008313230000031
wherein the content of the first and second substances,
Figure FDA0003008313230000032
presentation options (a)*,b*) The count of the number of uses of (c),
Figure FDA0003008313230000033
presentation options (a)*,b*) The average return of (a) is,
Figure FDA0003008313230000034
presentation options (a)*,b*) The rate is effectively reached in this report.
4. A computer device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, the computer program when loaded into the processor implementing the beam alignment and tracking method in a millimeter wave communication system of any of claims 1 to 3.
CN201910560646.8A 2019-06-26 2019-06-26 Beam alignment and tracking method in millimeter wave communication system and computer equipment Active CN110365375B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910560646.8A CN110365375B (en) 2019-06-26 2019-06-26 Beam alignment and tracking method in millimeter wave communication system and computer equipment

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910560646.8A CN110365375B (en) 2019-06-26 2019-06-26 Beam alignment and tracking method in millimeter wave communication system and computer equipment

Publications (2)

Publication Number Publication Date
CN110365375A CN110365375A (en) 2019-10-22
CN110365375B true CN110365375B (en) 2021-06-08

Family

ID=68217535

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910560646.8A Active CN110365375B (en) 2019-06-26 2019-06-26 Beam alignment and tracking method in millimeter wave communication system and computer equipment

Country Status (1)

Country Link
CN (1) CN110365375B (en)

Families Citing this family (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111446999A (en) * 2020-03-26 2020-07-24 上海无线通信研究中心 Position-assisted beam alignment method and system based on multi-arm forced theft
CN111555786B (en) * 2020-04-24 2022-08-19 深圳清华大学研究院 Optimal beam pair searching method and device, computer equipment and storage medium
CN111934735A (en) * 2020-07-13 2020-11-13 东南大学 Adaptive beam scanning and tracking method and device for millimeter wave frequency band
CN113259033B (en) * 2021-05-13 2022-11-29 东南大学 High-speed wave beam control method in dynamic millimeter wave communication scene based on FPGA
CN113364505B (en) * 2021-06-02 2022-06-03 东南大学 Self-adaptive beam control FPGA implementation method for ultra-high-speed serial interface array antenna
CN113271569A (en) * 2021-06-09 2021-08-17 南京万般上品信息技术有限公司 Intelligent vehicle-mounted communication method based on 5G mmWaves
WO2022266868A1 (en) * 2021-06-22 2022-12-29 深圳大学 Beam tracking method and related device
CN113437999B (en) * 2021-06-23 2023-01-17 东南大学 Adaptive beam width modulation method for inhibiting beam drift effect in millimeter wave communication system
CN114189852B (en) * 2021-12-01 2024-02-02 浙江大学 Downlink multiuser beam alignment and data transmission method for millimeter wave hidden communication
CN114553284B (en) * 2022-04-27 2022-07-05 四川太赫兹通信有限公司 Beam alignment method, device, base station and computer readable storage medium
CN114785393B (en) * 2022-06-21 2022-09-02 四川太赫兹通信有限公司 Self-adaptive beam width determining method, system, base station and medium
CN115623595A (en) * 2022-11-03 2023-01-17 北京邮电大学 Transmission slot configuration method, beam tracking method, electronic device, and storage medium

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2016077127A1 (en) * 2014-11-11 2016-05-19 Massachusetts Institute Of Technology A distributed, multi-model, self-learning platform for machine learning
CN109600813A (en) * 2018-12-05 2019-04-09 广州大学 It is a kind of ensure information safety wirelessly take can relay system and method
CN109831236A (en) * 2018-11-13 2019-05-31 电子科技大学 A kind of beam selection method based on Monte Carlo tree search auxiliary

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2016077127A1 (en) * 2014-11-11 2016-05-19 Massachusetts Institute Of Technology A distributed, multi-model, self-learning platform for machine learning
CN109831236A (en) * 2018-11-13 2019-05-31 电子科技大学 A kind of beam selection method based on Monte Carlo tree search auxiliary
CN109600813A (en) * 2018-12-05 2019-04-09 广州大学 It is a kind of ensure information safety wirelessly take can relay system and method

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
Deep Learning Coordinated Beamforming for Highly-Mobile Millimeter Wave Systems;Ahmed Alkhateeb 等;《IEEE Access》;20180625;第6卷;全文 *
Efficient Beam Alignment in Millimeter Wave Systems Using Contextual Bandits;Morteza Hashemi 等;《百度学术URL：https://xueshu.***.com/usercenter/paper/show?paperid=86bf4fe16760e33d48c0e0981d662965&site=xueshu_se》;20171231;正文第III-V节 *
Multi-Armed Bandit Beam Alignment and Tracking for Mobile Millimeter Wave Communications;Matthew B.Booth 等;《IEEE Communications Letters》;20190527;第23卷(第7期);全文 *

Also Published As

Publication number Publication date
CN110365375A (en) 2019-10-22

Similar Documents

Publication Publication Date Title
CN110365375B (en) Beam alignment and tracking method in millimeter wave communication system and computer equipment
CN110971279B (en) Intelligent beam training method and precoding system in millimeter wave communication system
Wu et al. Fast mmwave beam alignment via correlated bandit learning
Gante et al. Deep learning architectures for accurate millimeter wave positioning in 5G
Booth et al. Multi-armed bandit beam alignment and tracking for mobile millimeter wave communications
JP3929639B2 (en) Wireless communication system with space-time architecture that employs multi-element antennas at both transmitter and receiver
CN101808341B (en) Arrangements for beam refinement in wireless network
US8457240B2 (en) Methods of selecting signal transmitting, receiving, and/or sensing devices with probabilistic evolutionary algorithms in information conveyance systems
KR100913940B1 (en) Multiple antenna communication system including adaptive updating and changing of codebooks
CN112564752A (en) Intelligent surface auxiliary communication method for optimizing activation and reconfiguration of sparse antenna
JPWO2007037418A1 (en) Antenna selection method and wireless communication apparatus
Zhang et al. Intelligent beam training for millimeter-wave communications via deep reinforcement learning
US20220368393A1 (en) Method of performing beam training based on reinforcement learning and wireless communication device performing the same
CN114844545A (en) Communication beam selection method based on sub6GHz channel and partial millimeter wave pilot frequency
US11743163B2 (en) Interactive beam alignment with delayed feedback
Narengerile et al. Deep reinforcement learning-based beam training for spatially consistent millimeter wave channels
Alata et al. Law recognitions by information criteria for the statistical modeling of small scale fading of the radio mobile channel
CN114726412A (en) Channel information acquisition method, device and related equipment
Gante et al. Enhancing beamformed fingerprint outdoor positioning with hierarchical convolutional neural networks
Nguyen et al. Universal beamforming: A deep RFML approach
Zou et al. Step-wisely refinement based beam searching scheme for 60 GHz communications
WO2023287769A2 (en) Reinforcement learning of beam codebooks for millimeter wave and terahertz mimo systems
Seo et al. Pilot beam sequence design for channel estimation in millimeter-wave MIMO systems: A POMDP framework
Amuru Beam Learning--Using Machine Learning for Finding Beam Directions
Domae et al. Machine learning prediction for phase-less millimeter-wave beam tracking

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