CN113938360A - Distributed MIMO system covariance matrix estimation method based on fingerprint positioning - Google Patents

Distributed MIMO system covariance matrix estimation method based on fingerprint positioning Download PDF

Info

Publication number
CN113938360A
CN113938360A CN202111186420.XA CN202111186420A CN113938360A CN 113938360 A CN113938360 A CN 113938360A CN 202111186420 A CN202111186420 A CN 202111186420A CN 113938360 A CN113938360 A CN 113938360A
Authority
CN
China
Prior art keywords
covariance matrix
user
pilot
representing
channel
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.)
Granted
Application number
CN202111186420.XA
Other languages
Chinese (zh)
Other versions
CN113938360B (en
Inventor
李佳珉
叶枫
朱鹏程
王东明
尤肖虎
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 CN202111186420.XA priority Critical patent/CN113938360B/en
Publication of CN113938360A publication Critical patent/CN113938360A/en
Application granted granted Critical
Publication of CN113938360B publication Critical patent/CN113938360B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L25/00—Baseband systems
    • H04L25/02—Details ; arrangements for supplying electrical power along data transmission lines
    • H04L25/0202—Channel estimation
    • H04L25/0204—Channel estimation of multiple channels
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G06F18/232—Non-hierarchical techniques
    • G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
    • G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L25/00—Baseband systems
    • H04L25/02—Details ; arrangements for supplying electrical power along data transmission lines
    • H04L25/0202—Channel estimation
    • H04L25/021—Estimation of channel covariance
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L25/00—Baseband systems
    • H04L25/02—Details ; arrangements for supplying electrical power along data transmission lines
    • H04L25/0202—Channel estimation
    • H04L25/0224—Channel estimation using sounding signals
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L25/00—Baseband systems
    • H04L25/02—Details ; arrangements for supplying electrical power along data transmission lines
    • H04L25/0202—Channel estimation
    • H04L25/024—Channel estimation channel estimation algorithms
    • H04L25/0242—Channel estimation channel estimation algorithms using matrix methods
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D30/00—Reducing energy consumption in communication networks
    • Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks

Abstract

The invention discloses a distributed MIMO system covariance matrix estimation method based on fingerprint positioning, which aims at the situation that a covariance matrix changes along with time due to rapid environmental change, eliminates interference and noise by introducing random phase shift, and obtains more accurate covariance matrix estimation by changing a pilot frequency distribution mode by utilizing fingerprint positioning so as to solve the problems that the prior art mostly does not consider unknown covariance matrix, the covariance matrix estimation is inaccurate and the calculation complexity is high. Compared with the prior art, the method does not additionally increase pilot frequency, has low calculation complexity on the basis of high estimation accuracy, can further effectively improve the accuracy of covariance matrix estimation by the proposed pilot frequency distribution updating algorithm based on fingerprint positioning, is suitable for various wireless communication systems, and has very important significance for researching non-ideal factors of mobile scenes, so the method has certain practical value.

Description

Distributed MIMO system covariance matrix estimation method based on fingerprint positioning
Technical Field
The invention relates to the technical field of distributed MIMO systems, in particular to a distributed MIMO system covariance matrix estimation method based on fingerprint positioning.
Background
Distributed MIMO systems are a practical, scalable network MIMO scenario, where a large number of access points are geographically distributed and connected to one central processing unit, serving all users with the same time-frequency resources. In addition to the advantages of good propagation and channel reinforcement, MIMO has the characteristics of a distributed architecture, such as the benefits of macro diversity, no handover, and higher coverage. Therefore, the research of channel estimation techniques based on the case where the covariance matrix is unknown becomes more and more important.
However, all of these desirable advantages rely heavily on accurate channel state information. The covariance matrix of the channel vector characterizes the spatial correlation and is generally considered to be completely known in most techniques. But due to the mobility of the user and the linear relationship between the covariance matrix dimension and the number of antennas, the covariance matrix is not known in practical situations. Especially in the scene of fast channel change such as high-speed movement, the estimation of the covariance matrix and thus the accurate estimation of the channel are very beneficial to the analysis and optimization of the system performance. Therefore, the research of high-precision low-complexity covariance matrix estimation technology becomes more and more important.
Currently, wireless positioning is receiving increasing attention, especially in next generation communication systems, because location information can provide context-aware communication services. Fingerprint positioning is a fingerprint map based on channel state statistics, and is widely developed due to higher positioning accuracy in a rich scattering environment. Due to the limitation of the channel coherence interval length and the access of a large number of users, it cannot be guaranteed that the users are allocated to mutually orthogonal pilots, which will cause pilot pollution and seriously degrade the system performance. Pilot pollution is related to the user location of the shared pilots, and thus pilot pollution can be mitigated by optimizing pilot allocation based on location information. Therefore, the estimation accuracy of the covariance matrix can be improved by using the position information, which has an important role in dealing with non-ideal factors in the communication system, however, no solution for fusing the covariance matrix estimation with the positioning information has appeared in the prior art.
Disclosure of Invention
In view of the above, an object of the present invention is to provide a method for estimating a covariance matrix of a distributed MIMO system based on fingerprint positioning, so as to solve the problem in the prior art that the accuracy of channel estimation is not high in a real situation due to the fact that the covariance matrix is not considered to be unknown. The method and the device take the estimated covariance matrix as the fingerprint for positioning, perform clustering according to the positioning information, further change the pilot frequency distribution to obtain more accurate covariance matrix estimation, and improve the precision of covariance matrix estimation.
In order to achieve the purpose, the invention adopts the following technical scheme:
a distributed MIMO system covariance matrix estimation method based on fingerprint positioning, in the distributed MIMO system, a total of M access points are connected to a central processing unit, each access point has N antennas, jointly serves K single-antenna users in an area, and reference points of Z single antennas are uniformly distributed in a coverage area, the method specifically comprises the following steps:
step S1, establishing a channel model of distributed MIMO and an uplink signal transmission model to obtain a channel estimation expression of the covariance matrix under an unknown condition;
step S2, firstly, introducing random phase shift by using a channel block fading model, alternately using the distributed pilot frequency and the random phase shift of the pilot frequency in adjacent coherent blocks to obtain a received signal expression of the adjacent coherent blocks, then estimating sample covariance matrix estimation by using the spaced pilot frequency, and finally obtaining individual covariance matrix estimation by using the alternate received signal and considering Hermite symmetry;
step S3, firstly, the estimated covariance matrix is used as a fingerprint, the covariance matrix of a reference point is estimated as a fingerprint database, the user fingerprint is compared with the fingerprint database by using Euclidean distance, the reference point corresponding to the minimum Euclidean distance is used as the initial position of the user, then the user is clustered by using a K-means algorithm according to the position of the user, finally, orthogonal pilot frequency is adopted in the cluster according to the clustering result, then the pilot frequency is redistributed, and the covariance matrix is estimated again.
Further, the step S1 specifically includes:
step S101, considering a block fading channel, modeling a channel vector from user k to all access points as:
Figure BDA0003299410050000021
in the formula (1), the first and second groups,
Figure BDA0003299410050000022
λk,mrepresenting the large scale fading, I, from the m-th access point to the k-th userNRepresenting an N-dimensional identity matrix, hkRepresenting small-scale fading whose elements obey a standard Rayleigh distribution
Figure BDA0003299410050000023
Step S102, the number P of orthogonal pilot frequencies with length tau in the estimation of the assumed uplink channel<K, use of
Figure BDA0003299410050000024
Representing the user subset using pilot p, after associating the received training signal with the conjugate of the pilot sequence, the base station estimates the channel:
Figure BDA0003299410050000025
in the formula (2), ypIndicating the received signal of all users transmitting pilot p, r is the transmit power,
Figure BDA0003299410050000026
representing the noise, σ, of the transmitted pilot p2Representing the noise power;
due to gkSatisfy the requirement of
Figure BDA0003299410050000031
Then to
Figure BDA0003299410050000032
Channel gkThe expression for the minimum mean square error, MMSE, estimate of (c) is:
Figure BDA0003299410050000033
in the formula (3), IMNRepresents MN dimension unit matrix, let sigmapFor the covariance matrix of the received signal, ΛiChannel g representing user iiThe covariance matrix of (2).
Step S103, because ΛkSum ΣpAre all unknowns, so equation (3) is converted to:
Figure BDA0003299410050000034
in the formula (4), the first and second groups,
Figure BDA0003299410050000035
and
Figure BDA0003299410050000036
respectively represent ΛkSum ΣpIs estimated.
Further, the step S2 specifically includes:
step S201, firstly, assuming that the channel is a block fading model, specifically, the channel is at tauc＝BcTcIs unchanged within a symbol, wherein BcExpressed as coherence Bandwidth, TcExpressed as time, then assume that the covariance matrix is at τsThe coherent blocks are not changed, finally, the irrelevance of random phase shift is utilized to eliminate interference and noise, therefore, each user alternately transmits the distributed pilot frequency X in the adjacent coherent blockspAnd phase shifted pilot ΦkComprises the following steps:
Figure BDA0003299410050000037
in the formula (5), phik[2n]Denotes the pilot, X, sent by the 2n coherent block user kp[2n-1]The pilot representing the 2n-1 th coherent block is the pilot p allocated to user k, where n is 1,2k,2nRandom phase shift of 2n coherent block of k users, independent of channel vector and noise, and satisfying
Figure BDA0003299410050000038
Step S202, selecting N∑≤τsPer 2-interval coherent block estimation of covariance matrix sigma of received signalpThe expression is:
Figure BDA0003299410050000039
in the formula (6), yp[2n-1]Representing the received signal of the 2N-1 coherent block, N∑Representation for estimating ΣpNumber of coherent blocks, superscript (x)HWhich represents the transpose of the conjugate,
Figure BDA00032994100500000310
expressed as the covariance matrix sigma of the received signalp(ii) an estimate of (d);
step S203, in order to estimate the individual covariance matrix LambdakAn adjacent coherent block is required for decorrelation, and therefore, the pilot is sent to the adjacent coherent block, and the received signal observation is:
Figure BDA00032994100500000311
Figure BDA0003299410050000041
in the formula (7) and the formula (8),
Figure BDA0003299410050000042
representing the received signal observations when user k transmits the 2n-1 coherent block pilot signal,
Figure BDA0003299410050000043
represents the received signal observation, θ, of user k when it transmits the 2 n-th coherent block pilot signali,2nRepresenting the random phase generated by the pilot sequence generator when user i is in the 2n th coherence block.
Considering the signal is free from noise and random phase thetak,2nBy observing the neighboring coherent blocks, for
Figure BDA0003299410050000044
The covariance matrix of the channel obtained from the correlation of the adjacent staggered pilot received signal observations is:
Figure BDA0003299410050000045
considering hermitian symmetry, the individual covariance matrix
Figure BDA0003299410050000046
The expression of (a) is:
Figure BDA0003299410050000047
in the formula (10), NΛRepresentation for estimating ΛkThe required number of coherent block logarithms.
Further, the step S3 specifically includes:
step S301, calculating covariance matrix estimation of the reference point and the user respectively, wherein the covariance matrix estimation of the reference point approximately adopts orthogonal pilot frequency as a fingerprint database
Figure BDA0003299410050000048
Wherein the content of the first and second substances,
Figure BDA0003299410050000049
……,
Figure BDA00032994100500000410
representing a covariance matrix of reference point 1 to reference point Z estimates as fingerprints in a fingerprint database; estimating covariance matrix for user k
Figure BDA00032994100500000411
Comparing the fingerprints with a fingerprint database as fingerprints, and measuring the correlation degree between two different fingerprints by using Euclidean distance, wherein the expression is as follows:
Figure BDA00032994100500000412
in the formula (11), the reaction mixture,
Figure BDA00032994100500000413
denoted as the ith row and jth column of the covariance matrix for user k,
Figure BDA00032994100500000414
row i and column j of a covariance matrix representing an estimate of a reference point z in the fingerprint database;
selecting a reference point with the minimum Euclidean distance as an initial position of a user k;
step S302, dividing K users into K users by adopting a K-means clustering algorithm according to the user positioning result
Figure BDA00032994100500000415
A disjoint cluster, where P represents the number of orthogonal pilots, and the set of users in cluster t is denoted by Γt；
Step S303, for | Gammat|>P clusters, sorting the distance from each user to the corresponding gravity center, determining the P users with the minimum distance to directly form a cluster, and keeping the rest | gammatI-P usesThe user is saved in the set D; for | Γt|<P, and in addition, randomly selecting P- | Γ in the set Dt| users make up a cluster;
step S304, distributing orthogonal pilot frequency in the cluster, and estimating the covariance matrix of user k again
Figure BDA0003299410050000051
The invention has the beneficial effects that:
the invention considers the problem that the covariance matrix in the actual channel is unknown, carries out the estimation of the covariance matrix, the proposed method does not additionally increase pilot frequency, but only utilizes the special structure of sending the pilot frequency, effectively eliminates the interference and noise, further relieves the pilot frequency pollution by utilizing the fingerprint positioning mode, improves the estimation precision of the covariance matrix, has low calculation complexity, and is suitable for various wireless communication systems.
Drawings
Fig. 1 is a schematic diagram of a cellular-free distributed massive MIMO system provided in embodiment 1;
fig. 2 is a schematic flowchart of a method for estimating covariance matrix of a distributed MIMO system based on fingerprint positioning according to embodiment 1;
FIG. 3 is a simulation diagram of the relationship between the coherent block logarithm of individual covariance matrices and the normalized mean square error for channel estimation NMSE using different covariance matrix estimation methods, where random pilot represents the initial allocation of a random pilot estimation covariance matrix, fingerprint positioning represents the reallocation of a pilot estimation covariance matrix after fingerprint positioning, and via-Q is the estimation of the covariance matrix using a classical via-Q method;
fig. 4 is a simulation diagram of the relationship between the logarithm of coherent blocks used to compute the individual covariance matrices and the total spectral efficiency in MRC and ZF receivers using different covariance matrix estimation methods.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Example 1
Referring to fig. 1 to fig. 4, the present embodiment provides a method for estimating a covariance matrix of a distributed MIMO system based on fingerprint positioning, where the method is implemented based on the following large-scale MIMO system and an uplink received signal model of the system thereof, and specifically includes:
as shown in fig. 1, assume that a large-scale MIMO system without cellular distribution adopts a Time Division Duplex (TDD) mode, where M equals to 5 APs and each AP has N equals to 50 antennas. The reference points are uniformly distributed in the area at an interval of r-2 km, the area has K-12 users, the users are assumed to be equipped with single antennas and are independently distributed, and P-4 pilots are shared, so as to improve the utilization rate of the pilot resources.
The channel model consists of three parts: path loss, shadow fading, and small-scale fading, which can be expressed as
Figure BDA0003299410050000061
Wherein the content of the first and second substances,
Figure BDA0003299410050000062
let path fading factor α be 3.7 and reference distance be 1, sm,kTo satisfy the exponentially normally distributed shadow fading variables, hm,kRepresenting small-scale fading whose elements obey a standard Rayleigh distribution
Figure BDA0003299410050000063
Thus the uplink channel g of user k to all access pointskSatisfy the requirement of
Figure BDA0003299410050000064
Namely, it is
Figure BDA0003299410050000065
cov(gk,gk)＝Λk. The transmission power per user is ρ 1W, and the noise correlation σ is2-63 dBm. K users transmit information simultaneously, so the uplink receiving signal model of the system is
Figure BDA0003299410050000066
In equation (1), y indicates that the AP receives information from all users, and xkIs the information that is to be transmitted,
Figure BDA0003299410050000067
representing noise, let g bekIndependently of n.
Based on the non-cellular distributed large-scale MIMO system and the constructed uplink received signal model, the covariance matrix estimation method provided by the embodiment specifically includes the following steps:
step S1, establishing an uplink signal transmission model of the distributed MIMO system to obtain a channel estimation expression under the condition that a covariance matrix is unknown:
in this embodiment, step S1 specifically includes:
step S11, assuming the channel is a block fading model, in a certain phase bandwidth BcAnd a certain coherence time TcThe channel remains unchanged, i.e. the channel is at τc＝BcTcRemaining unchanged for 200 symbols. The covariance matrix is constant over the transmission bandwidth and changes slowly in time compared to the fast change of the channel vector, so it can be reasonably assumed that it is at τsRemaining unchanged in 20000 coherent blocks.
Step S12, use
Figure BDA0003299410050000068
Indicating that, after the received training signal is associated with the conjugate of the pilot sequence using the subset of users of pilot p, the observed value of the base station is:
Figure BDA0003299410050000069
in the formula (2), ypIndicating the received signals of all users transmitting pilot p,
Figure BDA00032994100500000610
noise representing the transmitted pilot p, due to gkSatisfy the requirement of
Figure BDA00032994100500000611
Then to
Figure BDA00032994100500000612
Channel gkThe minimum mean square error MMSE estimate of (c) should be:
Figure BDA00032994100500000613
in the formula (3), IMNRepresents MN dimension unit matrix, let sigmapFor the covariance matrix of the received signal, ΛiChannel g representing user iiThe covariance matrix of (2).
Step S13, in practical applications, the covariance matrix is usually imperfect, which means ΛkSum ΣpAre all unknown numbers, and therefore the channel estimation expression is converted into
Figure BDA0003299410050000071
In the formula (4), the first and second groups,
Figure BDA0003299410050000072
and
Figure BDA0003299410050000073
respectively represent ΛkSum ΣpIs estimated.
Step S2, adopting channel block fading model, introducing random phase shift, and using distributed pilot frequency and the random phase shift of the pilot frequency alternately in adjacent coherent blocks to obtain the receiving signal expression of the adjacent coherent blocks. Estimating a sample covariance matrix estimate using the spaced pilots; individual covariance matrix estimates are obtained using the alternating received signals and taking into account hermitian symmetry.
In this embodiment, step S2 specifically includes:
step S21, interference and noise are eliminated by using the irrelevancy of random phase shift, so each user alternately sends the allocated pilot X in adjacent coherent blockspAnd phase shifted pilot ΦkThe expression is:
Figure BDA0003299410050000074
in the formula (5), phik[2n]Pilot, X, representing the 2n coherent block user kp[2n-1]The pilot representing the 2n-1 th coherent block is the pilot p assigned to user k, n is 1,2k,2nRandom phase shift of 2n coherent block of k users, independent of channel vector and noise, and satisfying
Figure BDA0003299410050000075
Step S22, selecting N∑＝5000<τsPer 2 coherent block interval coherent block estimation sigmapThe expression is:
Figure BDA0003299410050000076
in the formula (6), yp[2n-1]Indicating the received signal of the 2n-1 th coherent block, superscript (-) toHDenotes conjugate transpose when N∑In the case of the larger size,
Figure BDA0003299410050000077
the estimation of (c) will be more accurate.
Step S23, in order to estimate the individual covariance matrix LambdakRequires adjacent coherent blocks to decorrelateAlternatively, the pilot is sent to adjacent coherent blocks, with a received signal observation of
Figure BDA0003299410050000078
Figure BDA0003299410050000079
In the formula (7) and the formula (8),
Figure BDA0003299410050000081
representing the received signal observations when user k transmits the 2n-1 coherent block pilot signal,
Figure BDA0003299410050000082
represents the received signal observation, θ, of user k when it transmits the 2 n-th coherent block pilot signali,2nRepresenting the random phase generated by the pilot sequence generator when user i is in the 2n th coherence block.
Considering the signal is free from noise and random phase thetak,2nThe present embodiment eliminates interference by observing neighboring coherent blocks, for
Figure BDA0003299410050000083
The covariance matrix of the channel estimated from the adjacent staggered pilot received signal observations is:
Figure BDA0003299410050000084
in addition, in view of the hermitian symmetry,
Figure BDA0003299410050000085
should be that
Figure BDA0003299410050000086
In the formula (10), NΛRepresentation for estimating ΛkThe required number of coherent block logarithms.
The general flowchart of the covariance matrix estimation method based on fingerprint positioning in step S3 is shown in fig. 2. In the off-line phase, the uplink pilot signals may be sent by the single antenna test equipment at each different access point in turn. Therefore, there is no mutual interference between the access points. The present embodiment employs the estimated reference point covariance matrices as the fingerprint database because they can reflect the location information. In the online phase, each user estimates the covariance matrix as a fingerprint and compares it with a fingerprint database. And obtaining an initial position according to the matching result, and clustering the users by using a K-means algorithm. And reallocating pilot frequency according to the clustering result, and re-estimating the covariance matrix to obtain more accurate channel estimation. The following are specific implementation steps:
step S31, respectively calculating covariance matrix estimation of the reference point and the user, and using the covariance matrix estimation of the reference point as a fingerprint database
Figure BDA0003299410050000087
Wherein the content of the first and second substances,
Figure BDA0003299410050000088
……,
Figure BDA0003299410050000089
representing the covariance matrix of reference point 1 to reference point Z estimates as fingerprints in the fingerprint database. Estimating covariance matrix for user k
Figure BDA00032994100500000810
As fingerprints, the fingerprints are compared with a fingerprint database, and Euclidean distances are used for measuring the correlation degree between two different fingerprints
Figure BDA00032994100500000811
In the formula (11), the reaction mixture,
Figure BDA00032994100500000812
denoted as the ith row and jth column of the covariance matrix for user k,
Figure BDA00032994100500000813
row i and column j of the covariance matrix representing the estimate of reference point z in the fingerprint database.
The reference point with the smallest euclidean distance is selected as the initial position of the user k.
Step S32, dividing K users into K users by adopting a K-means clustering algorithm according to the user positioning result
Figure BDA00032994100500000814
A plurality of disjoint clusters, each cluster having a center of gravity, and the number of iterations T being 20. Where P represents the number of orthogonal pilots, and the set of users in cluster t is denoted by Γt。
Step S33 for | Γt|>And P clusters, sequencing the distance from each user to the corresponding gravity center, and determining that the P users with the minimum distance directly form one cluster. The rest | Gammat-P users are kept in a set D; for | Γt|<P, and in addition, randomly selecting P- | Γ in the set DtThe | users together make up a cluster.
Step S34, distributing orthogonal pilot frequency in the cluster, estimating again the covariance matrix of user k
Figure BDA0003299410050000091
The whole process of estimating the covariance matrix based on fingerprint positioning by using the method provided by the embodiment is shown above.
Fig. 3 is a simulation diagram showing the relationship between the coherent block logarithm of the individual covariance matrix and the normalized mean square error of the channel estimate NMSE for different covariance matrix estimation methods, where the error represents the channel estimate error for known and unknown covariance matrices. The random pilot frequency represents the initial distribution of the covariance matrix of the random pilot frequency estimation, the fingerprint positioning represents the redistribution of the covariance matrix of the pilot frequency estimation after fingerprint positioning, and via-Q is a classic covariance matrix estimation method, which eliminates interference by sending the pilot frequency of other users sharing the same pilot frequency. Random pilot frequency distribution is adopted in the initial stage, and the positioning and pilot frequency redistribution algorithm based on the fingerprint is obtained by carrying out covariance matrix re-estimation after fingerprint positioning and pilot frequency redistribution. It can be seen that under the same random pilot allocation, the covariance matrix estimation performance of the method is better than via-Q and is close to the case that the covariance matrix is known. On the basis of fingerprint-based positioning, the covariance matrix estimation becomes more accurate due to pilot reallocation.
Fig. 4 is a simulation diagram showing the relationship between the logarithm of coherent blocks used to compute the individual covariance matrices and the total spectral efficiency in different covariance matrix estimation methods under MRC and ZF receivers. It can be seen that the covariance matrix estimation method provided by the embodiment has better total spectrum efficiency than the via-Q method, and the system performance is further improved after fingerprint positioning and pilot frequency reallocation. At the same time, the overall spectral efficiency is increased with the number of coherent block logarithms N used to compute the individual covariance matrix estimatesΛIs increased because the calculated covariance matrix estimate is more accurate. In addition, the system performance of the ZF receiver is superior to the MRC receiver.
In summary, the present invention provides a distributed antenna system covariance matrix estimation method for eliminating interference and noise by introducing random phase shift and using cross correlation of distributed pilot frequency and random phase shift pilot frequency, aiming at the covariance matrix estimation problem of a distributed MIMO system, so as to solve the problem that the channel estimation accuracy is not high in the real situation because the covariance matrix is not known in the prior art. In addition, the estimated covariance matrix is used as a fingerprint for positioning, clustering is carried out according to positioning information, and then pilot frequency distribution is changed to obtain more accurate covariance matrix estimation, so that the accuracy of covariance matrix estimation is improved, and the method has practical significance.
The invention is not described in detail, but is well known to those skilled in the art.
The foregoing detailed description of the preferred embodiments of the invention has been presented. It should be understood that numerous modifications and variations could be devised by those skilled in the art in light of the present teachings without departing from the inventive concepts. Therefore, the technical solutions available to those skilled in the art through logic analysis, reasoning and limited experiments based on the prior art according to the concept of the present invention should be within the scope of protection defined by the claims.

Claims (4)

1. A distributed MIMO system covariance matrix estimation method based on fingerprint positioning, in the distributed MIMO system, a total of M access points are connected to a central processing unit, each access point has N antennas, jointly serves K single-antenna users in an area, and reference points of Z single antennas are uniformly distributed in a coverage area, the method is characterized by specifically comprising the following steps:
step S1, establishing a channel model of distributed MIMO and an uplink signal transmission model to obtain a channel estimation expression of the covariance matrix under an unknown condition;
step S2, firstly, introducing random phase shift by using a channel block fading model, alternately using the distributed pilot frequency and the random phase shift of the pilot frequency in adjacent coherent blocks to obtain a received signal expression of the adjacent coherent blocks, then estimating sample covariance matrix estimation by using the spaced pilot frequency, and finally obtaining individual covariance matrix estimation by using the alternate received signal and considering Hermite symmetry;
step S3, firstly, the estimated covariance matrix is used as a fingerprint, the covariance matrix of a reference point is estimated as a fingerprint database, the user fingerprint is compared with the fingerprint database by using Euclidean distance, the reference point corresponding to the minimum Euclidean distance is used as the initial position of the user, then the user is clustered by using a K-means algorithm according to the position of the user, finally, orthogonal pilot frequency is adopted in the cluster according to the clustering result, then the pilot frequency is redistributed, and the covariance matrix is estimated again.
2. The method of claim 1, wherein the step S1 specifically includes:
step S101, considering a block fading channel, modeling a channel vector from user k to all access points as:
Figure FDA0003299410040000011
in the formula (1), the first and second groups,
Figure FDA0003299410040000012
lk,mrepresenting the large scale fading, I, from the m-th access point to the k-th userNRepresenting an N-dimensional identity matrix, hkRepresenting small-scale fading whose elements obey a standard Rayleigh distribution
Figure FDA0003299410040000013
Step S102, the number P of orthogonal pilot frequencies with length tau in the estimation of the assumed uplink channel<K, use of
Figure FDA0003299410040000014
Representing the user subset using pilot p, after associating the received training signal with the conjugate of the pilot sequence, the base station estimates the channel:
Figure FDA0003299410040000015
in the formula (2), ypRepresenting the received signals of all users transmitting pilot p, p is the transmit power,
Figure FDA0003299410040000016
representing the noise, σ, of the transmitted pilot p2Representing the noise power;
due to gkSatisfy the requirement of
Figure FDA0003299410040000021
Then to
Figure FDA0003299410040000022
Channel gkThe expression for the minimum mean square error, MMSE, estimate of (c) is:
Figure FDA0003299410040000023
in the formula (3), IMNRepresents MN dimension unit matrix, let sigmapFor the covariance matrix of the received signal, ΛiChannel g representing user iiThe covariance matrix of (a);
step S103, because ΛkSum ΣpAre all unknowns, so equation (3) is converted to:
Figure FDA0003299410040000024
in the formula (4), the first and second groups,
Figure FDA0003299410040000025
and
Figure FDA0003299410040000026
respectively represent ΛkSum ΣpIs estimated.
3. The method of claim 2, wherein the step S2 specifically includes:
step S201, firstly, assuming that the channel is a block fading model, specifically, the channel is at tauc＝BcTcIs unchanged within a symbol, wherein BcExpressed as coherence Bandwidth, TcExpressed as time, then assume that the covariance matrix is at τsAnd is unchanged in coherent block, finally utilizes irrelevance of random phase shift to eliminate interference and noise,so that each user alternately transmits the allocated pilot X in adjacent coherent blockspAnd phase shifted pilot ΦkComprises the following steps:
Figure FDA0003299410040000027
in the formula (5), phik[2n]Denotes the pilot, X, sent by the 2n coherent block user kp[2n-1]The pilot representing the 2n-1 th coherent block is the pilot p allocated to user k, where n is 1,2k,2nRandom phase shift of 2n coherent block of k users, independent of channel vector and noise, and satisfying
Figure FDA0003299410040000028
Step S202, selecting N∑≤τsPer 2-interval coherent block estimation of covariance matrix sigma of received signalpThe expression is:
Figure FDA0003299410040000029
in the formula (6), yp[2n-1]Representing the received signal of the 2N-1 coherent block, N∑Representation for estimating ΣpNumber of coherent blocks, superscript (x)HWhich represents the transpose of the conjugate,
Figure FDA00032994100400000210
expressed as the covariance matrix sigma of the received signalp(ii) an estimate of (d);
step S203, in order to estimate the individual covariance matrix LambdakAn adjacent coherent block is required for decorrelation, and therefore, the pilot is sent to the adjacent coherent block, and the received signal observation is:
Figure FDA0003299410040000031
Figure FDA0003299410040000032
in the formula (7) and the formula (8),
Figure FDA0003299410040000033
representing the received signal observations when user k transmits the 2n-1 coherent block pilot signal,
Figure FDA0003299410040000034
represents the received signal observation, θ, of user k when it transmits the 2 n-th coherent block pilot signali,2nRepresenting the random phase generated by the pilot sequence generator when the user i is in the 2n th coherent block;
considering the signal is free from noise and random phase thetak,2nBy observing the neighboring coherent blocks, for
Figure FDA0003299410040000035
The covariance matrix of the channel obtained according to the correlation of the observed values of the adjacent staggered pilot frequency receiving signals is:
Figure FDA0003299410040000036
considering Hermite symmetry and the individual covariance matrix
Figure FDA0003299410040000037
The expression of (a) is:
Figure FDA0003299410040000038
in the formula (10), NΛRepresentation for estimating ΛkThe required number of coherent block logarithms.
4. The method of claim 3, wherein the step S3 specifically includes:
step S301, calculating covariance matrix estimation of the reference point and the user respectively, wherein the covariance matrix estimation of the reference point approximately adopts orthogonal pilot frequency as a fingerprint database
Figure FDA0003299410040000039
Wherein the content of the first and second substances,
Figure FDA00032994100400000310
representing a covariance matrix of reference point 1 to reference point Z estimates as fingerprints in a fingerprint database; estimating covariance matrix for user k
Figure FDA00032994100400000311
Comparing the fingerprints with a fingerprint database as fingerprints, and measuring the correlation degree between two different fingerprints by using Euclidean distance, wherein the expression is as follows:
Figure FDA00032994100400000312
in the formula (11), the reaction mixture,
Figure FDA00032994100400000313
denoted as the ith row and jth column of the covariance matrix for user k,
Figure FDA00032994100400000314
row i and column j of a covariance matrix representing an estimate of a reference point z in the fingerprint database;
selecting a reference point with the minimum Euclidean distance as an initial position of a user k;
step S302, dividing K users into K users by adopting a K-means clustering algorithm according to the user positioning result
Figure FDA0003299410040000041
A disjoint cluster, where P represents the number of orthogonal pilots, and the set of users in cluster t is denoted by Γt；
Step S303, for | Gammat|>P clusters, sorting the distance from each user to the corresponding gravity center, determining the P users with the minimum distance to directly form a cluster, and keeping the rest | gammat-P users are kept in a set D; for | Γt|<P, and in addition, randomly selecting P- | Γ in the set Dt| users make up a cluster;
step S304, distributing orthogonal pilot frequency in the cluster, and estimating the covariance matrix of user k again
Figure FDA0003299410040000042
CN202111186420.XA 2021-10-12 2021-10-12 Distributed MIMO system covariance matrix estimation method based on fingerprint positioning Active CN113938360B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111186420.XA CN113938360B (en) 2021-10-12 2021-10-12 Distributed MIMO system covariance matrix estimation method based on fingerprint positioning

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111186420.XA CN113938360B (en) 2021-10-12 2021-10-12 Distributed MIMO system covariance matrix estimation method based on fingerprint positioning

Publications (2)

Publication Number Publication Date
CN113938360A true CN113938360A (en) 2022-01-14
CN113938360B CN113938360B (en) 2024-02-27

Family

ID=79278906

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111186420.XA Active CN113938360B (en) 2021-10-12 2021-10-12 Distributed MIMO system covariance matrix estimation method based on fingerprint positioning

Country Status (1)

Country Link
CN (1) CN113938360B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116366405A (en) * 2023-03-17 2023-06-30 哈尔滨工业大学（深圳） Large-scale MIMO channel estimation method and base station for high mobility communication

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111556431A (en) * 2020-04-21 2020-08-18 西安交通大学 Fingerprint indoor positioning method based on signal subspace
US20210204241A1 (en) * 2017-10-12 2021-07-01 Southeast University 5g-signal-based doa fingerprint-based positioning method
CN113067783A (en) * 2021-03-18 2021-07-02 东南大学 Covariance matrix estimation method for distributed antenna system

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20210204241A1 (en) * 2017-10-12 2021-07-01 Southeast University 5g-signal-based doa fingerprint-based positioning method
CN111556431A (en) * 2020-04-21 2020-08-18 西安交通大学 Fingerprint indoor positioning method based on signal subspace
CN113067783A (en) * 2021-03-18 2021-07-02 东南大学 Covariance matrix estimation method for distributed antenna system

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116366405A (en) * 2023-03-17 2023-06-30 哈尔滨工业大学（深圳） Large-scale MIMO channel estimation method and base station for high mobility communication
CN116366405B (en) * 2023-03-17 2023-11-07 哈尔滨工业大学（深圳） Large-scale MIMO channel estimation method and base station for high mobility communication

Also Published As

Publication number Publication date
CN113938360B (en) 2024-02-27

Similar Documents

Publication Publication Date Title
CN109005551B (en) Multi-user NOMA downlink power distribution method of non-ideal channel state information
Decurninge et al. CSI-based outdoor localization for massive MIMO: Experiments with a learning approach
Shaham et al. Fast channel estimation and beam tracking for millimeter wave vehicular communications
CN106972907A (en) Extensive antenna system channel training and transmitting procedure inter-cell interference cancellation method
Hu et al. Cluster grouping and power control for angle-domain mmWave MIMO NOMA systems
Chen et al. Reconfigurable-intelligent-surface-assisted B5G/6G wireless communications: Challenges, solution, and future opportunities
CN111132264B (en) User clustering method for downlink of multi-user MIMO-NOMA system
CN109691050B (en) Method, device and system for estimating frequency offset in OFDM system
CN113067783B (en) Covariance matrix estimation method for distributed antenna system
CN101472286A (en) Virtual MIMO system and apparatus thereof
CN114337976B (en) Transmission method combining AP selection and pilot frequency distribution
Muppirisetty et al. Location-aided pilot contamination elimination for massive MIMO systems
Göttsch et al. Deep learning-based beamforming and blockage prediction for sub-6GHz/mm wave mobile networks
Li et al. Selective uplink training for massive MIMO systems
CN109412661B (en) User clustering method under large-scale MIMO system
Wang et al. Millimeter wave cell-free massive MIMO systems: Joint beamforming and AP-user association
CN113490219B (en) Dynamic resource allocation method for ultra-dense networking
CN113938360B (en) Distributed MIMO system covariance matrix estimation method based on fingerprint positioning
Ye et al. Fingerprint-based covariance matrix estimation for cell-free distributed massive MIMO systems
Liu et al. An improved adaptive beamforming-based machine learning method for positioning in massive mimo systems
Diao et al. A Scalable Deep Learning Based Active User Detection Approach for SEU-Aided Cell-Free Massive MIMO Systems
Elfiatoure et al. Cell-Free Massive MIMO for ISAC: Access Point Operation Mode Selection and Power Control
CN103117821A (en) Weighing cooperation spectrum sensing method based on rayleigh quotient
CN105847196A (en) DOA-based multi cell Massive MIMO system channel estimation method
Bartoli et al. Physical Resource Block clustering method for an OFDMA cognitive femtocell system

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