CN115149988A - Self-adaptive segmented matrix inverse tracking MIMO detection method based on SOR iterative method - Google Patents

Self-adaptive segmented matrix inverse tracking MIMO detection method based on SOR iterative method Download PDF

Info

Publication number
CN115149988A
CN115149988A CN202210769221.XA CN202210769221A CN115149988A CN 115149988 A CN115149988 A CN 115149988A CN 202210769221 A CN202210769221 A CN 202210769221A CN 115149988 A CN115149988 A CN 115149988A
Authority
CN
China
Prior art keywords
matrix
tracking
inverse
iteration
sor
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
CN202210769221.XA
Other languages
Chinese (zh)
Other versions
CN115149988B (en
Inventor
张华�
匡鹏翔
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 CN202210769221.XA priority Critical patent/CN115149988B/en
Publication of CN115149988A publication Critical patent/CN115149988A/en
Application granted granted Critical
Publication of CN115149988B publication Critical patent/CN115149988B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B7/00—Radio transmission systems, i.e. using radiation field
    • H04B7/02—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas
    • H04B7/04—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas
    • H04B7/0413—MIMO systems
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B17/00—Monitoring; Testing
    • H04B17/30—Monitoring; Testing of propagation channels
    • 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
    • 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
    • H04L25/0256—Channel estimation using minimum mean square error criteria
    • 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

Landscapes

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

Abstract

The invention discloses a self-adaptive segmented matrix inverse tracking MIMO detection method based on an SOR iterative method. Massive MIMO, one of the key technologies of 5G, has the advantages of high information transmission rate, high reliability, and high spectrum utilization rate compared to the conventional single antenna system. In a large-scale MIMO system, signal detection is a key technology for determining the reliability of the system, and is one of difficulties, and the traditional MIMO detection algorithm generally has the defects of high computational complexity and low convergence rate. The invention provides a self-adaptive segmented matrix inverse tracking detection method aiming at the change rate of a channel on the basis of an SOR iterative algorithm and in consideration of the time-varying characteristic of the channel. The invention tracks the MMSE filter matrix by utilizing the correlation characteristic of the channel in the time domain, adaptively updates the tracking step length according to the time domain change rate of the channel, and only carries out one-time SOR iterative operation at each tracking updating moment, thereby not only improving the convergence speed of the algorithm, but also reducing the calculation complexity.

Description

Self-adaptive segmented matrix inverse tracking MIMO detection method based on SOR iterative method
Technical Field
The invention relates to a self-adaptive segmented matrix inverse tracking MIMO (multiple input multiple output) detection method based on an SOR (sequential Over-Relay) iteration method, and belongs to the technical field of wireless communication.
Background
Massive MIMO technology has received sufficient attention in recent years as a key technology of 5G. The MIMO detection is a technical difficulty in a large-scale MIMO scene, MIMO detection algorithms are mainly divided into two categories, namely linear detection and nonlinear detection, and the method is improved aiming at the linear detection algorithm.
The linear detection method mainly has the technical difficulties of solving the inverse of a filter matrix, extremely high precision calculation complexity of the matrix inverse, very difficult hardware realization, and generally adopts an iterative algorithm to calculate the matrix approximate inverse in engineering, and the commonly used iterative algorithms include Newton (Newton), gauss-Seidel, successive Over-Relaxation (SOR), richardson, jacobi and the like. For the above iterative algorithm, as the number of iterations increases, the error between the approximate inverse and the exact inverse of the filter matrix is continuously reduced, but the corresponding computational complexity increases, and generally, when the number of iterations exceeds 3, the computational complexity of the iterative algorithm exceeds the exact computation. Therefore, the current research mainly focuses on how to improve the convergence rate of the iterative algorithm, such as the research on the coefficient value taking problem and the iterative initial value selection problem in the Richardson algorithm under the non-stationary condition. However, these studies do not consider the correlation characteristic of the channel in time, but perform independent iterative operation at each sampling point, and actually, the channel matrix is a continuous change process in time, especially in an indoor MIMO scenario, the channel is slowly changed in time.
Disclosure of Invention
The technical problem is as follows: the main technical problem to be solved by the invention is to avoid the defects in the background art, and provide a self-adaptive segmented matrix inverse tracking MIMO detection method based on an SOR iteration method.
The technical scheme is as follows: the invention provides an adaptive segmented matrix inverse tracking method based on an SOR iterative method according to the time-varying characteristic of a channel, which comprises the following steps:
taking the inverse of a diagonal dominance matrix diagonal matrix as an iteration initial value of an SOR iteration algorithm, and performing iteration for a plurality of times to obtain an approximate inverse of the diagonal dominance matrix;
tracking the inverse of the diagonal dominant matrix, adaptively determining a tracking step length according to the time domain change rate of a channel, performing one-time SOR iteration by using the iteration result of the previous tracking update time as the iteration initial value of the next tracking update time, namely performing adaptive step length segmented matrix inverse tracking;
and restoring the transmitted signal according to the iteration result.
Further, in general, a zero matrix may be used as the initial iteration value, but this is not favorable for algorithm convergence, because the main diagonal dominance of the matrix G may be the inverse D of the diagonal matrix of the filter matrix W of the MMSE detection algorithm -1 As an iteration initial value, further, the value of a second-order Neumann series expansion of the matrix G can be calculated, the convergence of the algorithm can be accelerated more quickly, the SOR iteration algorithm is used for carrying out 4 times of iteration, and the initial value of the algorithm is calculated.
Further, a Gram matrix G = H is defined H H, and
Figure BDA0003723308700000021
where H is the channel matrix and y is the received signal. Then in accordance with the MMSE detection algorithm,
Figure BDA0003723308700000022
wherein sigma 2 Is the noise variance, W is the filter matrix, I is the unit matrix,
Figure BDA0003723308700000023
is the recovered signal. Considering the channel hardening characteristics of the Massive-MIMO channel, the filter matrix has diagonal dominance, the algorithm can therefore be iterated according to the SOR using the inverse of the diagonal matrix D of W as the initial value of the iteration
W -1(k) ＝(D+wL) -1 (wI+((1-w)D-wU)W -1(k-1) )k＝1,2,…
And (5) performing iteration, and performing 4 times of iterative operation to improve the accuracy of the initial value to obtain an iteration result of the algorithm at the first sampling point.
Further, the channel matrix H is a continuous variation process in time, and accordingly, the MMSE filtering matrix W is also continuous in time, which results in a small variation rate in the channel time domainIn the case of (3), since the inverses of the filter matrices corresponding to the adjacent and nearer sampling points are relatively close to each other, it is possible to use the same matrix inversion in one sampling point section and use the approximate inversion W of the previous sampling point section inv,t As initial value of next iteration
Figure BDA0003723308700000024
Lambda is the length of each segment, namely the inverse tracking step length of the matrix, and is determined in a self-adaptive manner according to the change rate of the channel. Adaptive compensation segment tracking procedure from the above-obtained filter matrix inverse
Figure BDA0003723308700000025
The matrix inverse tracking method combines the adaptive step size and the SOR iterative algorithm, makes full use of the correlation characteristic of the channel in time, and can obtain a convergence speed which is higher than that of the traditional algorithm for multiple iterations under the condition of only one iteration.
Further, the tracking step λ is determined by the following method, because the filter matrix W is dominant diagonally and the characteristic of the matrix is concentrated on the diagonal of the matrix, so the change rate v of the channel can be determined by the value D of the diagonal matrix of the current segment of W k (1, 1) and the value D of the diagonal matrix of the previous segment k-1 The difference between (1, 1) can be represented as:
Figure BDA0003723308700000026
λ =1 when the v value is 20% or less; λ =2 when the v value is greater than 20% and 40% or less; λ =3 when the v value is greater than 40%.
Further, based on the channel estimation result and the received signal
Figure BDA0003723308700000027
Recovering a transmitted signal
Figure BDA0003723308700000028
Via receiving end demodulation moduleThe method comprises the steps of obtaining information bits transmitted by a transmitting end, transmitting known information bits at the transmitting end, comparing data demodulated by a receiving end to obtain the error rate of the method, and evaluating the algorithm performance of the method according to the error rate.
Has the advantages that: the invention provides a self-adaptive segmented matrix inverse tracking detection method aiming at the channel change rate on the basis of an SOR iterative algorithm by considering the time-varying characteristic of a channel.
Drawings
Fig. 1 is a flowchart of an adaptive segmented matrix inverse tracking MIMO detection method based on an SOR iterative method according to an embodiment of the present invention.
Fig. 2 is a schematic structural diagram of an adaptive segmented matrix inverse tracking MIMO detection system based on an SOR iteration method according to an embodiment of the present invention.
FIG. 3 is a graph comparing the performance of the method of the present invention with that of a conventional method.
Detailed Description
The invention provides a self-adaptive segmented matrix inverse tracking MIMO (multiple input multiple output) detection method based on an SOR (sequential Over-Relay) iteration method, which is improved on the basis of the traditional SOR iteration algorithm aiming at a scene with low channel time-varying rate in a large-scale MIMO system, provides a self-adaptive step length segmented matrix inverse tracking detection algorithm, and improves the convergence speed while keeping low calculation complexity.
System model
Setting the number of transmitting antennas to be N t The number of receiving antennas is N r Consider an N r ×N t MIMO system, as shown in fig. 2, a transmission signal is modulated by a transmitter, transmitted through multiple antennas, and a signal received by a receiver is demodulated, which may be represented as y = Hx + z, where H is a channel matrix:
Figure BDA0003723308700000031
element h ji Represents the channel gain from the ith transmitting antenna to the jth receiving antenna, j =1,2, \8230, N r ；i＝1,2,…,N t 。x＝[x 1 ,x 2 ,…,x t ] T To transmit a signal, wherein x i The signal sent by the ith transmitting antenna; y = [ y 1 ,y 2 ,…,y r ] T To transmit a signal, wherein y j For the signal received by the jth receive antenna. z = [ z ] 1 ,z 2 ,…,z r ] T Is a noise vector, z j Represents the additive white noise of the j-th antenna with a variance of
Figure BDA0003723308700000041
In an actual scene, the influence of fading also needs to be considered, and for the indoor environment where the system is located, a rice fading model is adopted due to the existence of a strong direct path. The Rice channel matrix can represent
Figure BDA0003723308700000042
Wherein H LOS To determine the channel matrix components, H Rayleigh For the random channel matrix component, K is the rice factor.
When an algorithm performance simulation environment is built, in order to be more consistent with the actual channel of the system, the main diagonal is strengthened on the basis, namely, the main diagonal is increased
Figure BDA0003723308700000043
And (4) components. Meanwhile, in order to compare the system performances of different channels conveniently, the channel matrix needs to be normalized
Figure BDA0003723308700000044
Wherein | · | purple sweet Frobenius Representing the Frobenius norm of the matrix.
(II) SOR iterative algorithm
Defining Gram matrix G = H H H, and
Figure BDA0003723308700000045
where H is the channel matrix and y is the received signal. Then in accordance with the MMSE detection algorithm,
Figure BDA0003723308700000046
wherein σ 2 Is the noise variance, W is the filter matrix, I is the unit matrix,
Figure BDA0003723308700000047
is the recovered signal. Considering the channel hardening characteristics of Massive-MIMO channel, W is approximately a Hermition positive definite matrix, so W = D + L + U, where D, L, U are a diagonal matrix, a lower triangular matrix and an upper triangular matrix of W, respectively, and the SOR iterative algorithm of MMSE linear detection can be expressed as:
x (k) ＝(D+wL) -1 (wy MF +((1-w)D-wU)x (k-1) )k＝1,2,…
where k is the number of iterations and w is the relaxation factor. The iteration result of the above formula is the estimated transmission signal
Figure BDA0003723308700000048
However, the matrix inverse tracking algorithm proposed by the present invention needs to track the inverse of the filter matrix, and therefore needs to perform the deformation W on the SOR iterative algorithm -1(k) ＝(D+wL) -1 (wI+((1-w)D-wU)W -1(k-1) )k＝1,2,…
Wherein I is a unit array. Because W has diagonal dominance characteristic, the inverse of diagonal matrix D of W can be used as iteration initial value to iterate, and 4 times of iterative operation are performed to improve the accuracy of the initial value, so that the iterative result of the algorithm at the first sampling point is obtained.
(III) adaptive piecewise matrix inverse tracking
In practice, the channel matrix H is in timeCorrespondingly, the MMSE filter matrix W also has continuity in time, which results in that the inverses of the filter matrices corresponding to the adjacent and closer sampling points are relatively close to each other under the condition of a relatively small channel time domain change rate, so that the same matrix inversion can be used in a section of sampling points, and the approximate inverse W of the previous section can be used inv,t As initial value for next iteration
Figure BDA0003723308700000051
And lambda is the length of each segment, namely the inverse tracking step length of the matrix, and is determined in a self-adaptive mode according to the change rate of the channel. Adaptive compensation and segmented tracking process with filter matrix inverse obtained from above
Figure BDA0003723308700000052
The matrix inverse tracking method combines the adaptive step size and the SOR iterative algorithm, makes full use of the correlation characteristic of the channel in time, and can obtain a convergence speed which is higher than that of the traditional algorithm for multiple iterations under the condition of only one iteration. As shown in fig. 3, in the 64QAM modulation, channel rice factor K =6, and 8 × 8MIMO scenario, when the snr is greater than 20dB, the performance of the method of the present invention is significantly improved compared to the performance of the conventional SOR iterative algorithm (4 iterations), and is already close to the performance of the MMSE algorithm.
The embodiments of the present invention will be described in detail with reference to the drawings, but the present invention is not limited to the above embodiments, and various changes can be made within the knowledge of those skilled in the art without departing from the gist of the present invention.

Claims (6)

1. A self-adaptive segmented matrix inverse tracking MIMO detection method based on an SOR iterative method is characterized in that:
taking the inverse of a diagonal dominance matrix diagonal matrix as an iteration initial value of an SOR iteration algorithm, and performing iteration for a plurality of times to obtain an approximate inverse of the diagonal dominance matrix;
tracking the inverse of the diagonal dominant matrix, adaptively determining a tracking step length according to the time domain change rate of a channel, performing one-time SOR iteration by using the iteration result of the previous tracking update time as the iteration initial value of the next tracking update time, namely performing adaptive step length segmented matrix inverse tracking;
and restoring the transmitted signal according to the iteration result.
2. The adaptive segmented matrix inverse tracking MIMO detection method based on the SOR iterative method as claimed in claim 1, wherein: the diagonal dominant matrix is a filter matrix W of an MMSE detection algorithm.
3. The adaptive segmented matrix inverse tracking MIMO detection method based on the SOR iterative method as claimed in claim 2, wherein: using the inverse of the diagonal matrix D of the filter matrix W as an initial value of the iteration, according to the SOR iterative algorithm
W -1(k) ＝(D+wL) -1 (wI+((1-w)D-wU)W -1(k-1) )k＝1,2,…
Performing an iteration, wherein W is a relaxation factor, L is a lower triangular matrix of the filter matrix W, U is an upper triangular matrix of the filter matrix W, W -1(k-1) As a result of the last iteration, W -1(k) And I is a unit array for the current iteration result.
4. The MIMO detection method based on the inverse tracking of the adaptive segmented matrix of the SOR iterative method as claimed in claim 3, wherein: under the condition of small time domain change rate of the channel, the same diagonal matrix inverse is used in a section of sampling points, and the approximate inverse W of the previous section is used inv,t As initial value for next iteration
Figure FDA0003723308690000011
Lambda is the length of each segment, namely the inverse tracking step length of the matrix, and is determined in a self-adaptive manner according to the change rate of a channel; adaptive compensation and segmented tracking process with filter matrix inverse obtained from above
Figure FDA0003723308690000012
5. The adaptive piecewise matrix inverse tracking MIMO detection method based on the SOR iterative method according to claim 4, characterized in that: the tracking step λ is determined by the following method: the rate of change v of the channel is determined by the value D of the diagonal matrix of the current segment of W k (1, 1) and the value D of the diagonal matrix of the previous stage k-1 (1, 1) is characterized by the difference between:
Figure FDA0003723308690000013
λ =1 when the v value is 20% or less; λ =2 when the v value is greater than 20% and 40% or less; λ =3 when the v value is greater than 40%.
6. The adaptive piecewise matrix inverse tracking MIMO detection method based on the SOR iterative method as claimed in claim 1, characterized in that: further comprising the steps of: and (4) verifying whether the algorithm can accelerate iterative convergence by evaluating the hard-decision bit error rate.
CN202210769221.XA 2022-06-30 2022-06-30 Self-adaptive segmented matrix inverse tracking MIMO detection method based on SOR iteration method Active CN115149988B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210769221.XA CN115149988B (en) 2022-06-30 2022-06-30 Self-adaptive segmented matrix inverse tracking MIMO detection method based on SOR iteration method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210769221.XA CN115149988B (en) 2022-06-30 2022-06-30 Self-adaptive segmented matrix inverse tracking MIMO detection method based on SOR iteration method

Publications (2)

Publication Number Publication Date
CN115149988A true CN115149988A (en) 2022-10-04
CN115149988B CN115149988B (en) 2024-07-09

Family

ID=83410557

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210769221.XA Active CN115149988B (en) 2022-06-30 2022-06-30 Self-adaptive segmented matrix inverse tracking MIMO detection method based on SOR iteration method

Country Status (1)

Country Link
CN (1) CN115149988B (en)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20180167237A1 (en) * 2016-12-14 2018-06-14 Intel IP Corporation Data detection in mimo systems with demodulation and tracking reference signals
CN109245804A (en) * 2018-08-27 2019-01-18 江南大学 Extensive MIMO signal detection method based on Jacobi iteration
CN114070354A (en) * 2021-12-10 2022-02-18 东南大学 Adaptive segmented matrix inverse tracking MIMO detection method based on GS iteration method
CN114142905A (en) * 2021-12-10 2022-03-04 东南大学 Improved method of MIMO detection algorithm based on Newton iteration method under time-varying channel

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20180167237A1 (en) * 2016-12-14 2018-06-14 Intel IP Corporation Data detection in mimo systems with demodulation and tracking reference signals
CN109245804A (en) * 2018-08-27 2019-01-18 江南大学 Extensive MIMO signal detection method based on Jacobi iteration
CN114070354A (en) * 2021-12-10 2022-02-18 东南大学 Adaptive segmented matrix inverse tracking MIMO detection method based on GS iteration method
CN114142905A (en) * 2021-12-10 2022-03-04 东南大学 Improved method of MIMO detection algorithm based on Newton iteration method under time-varying channel

Also Published As

Publication number Publication date
CN115149988B (en) 2024-07-09

Similar Documents

Publication Publication Date Title
CN108650201B (en) Neural network-based channel equalization method, decoding method and corresponding equipment
CN108566257B (en) Signal recovery method based on back propagation neural network
CN109951214B (en) Signal detection method suitable for large-scale MIMO system
US20060256888A1 (en) Multi input multi output wireless communication reception method and apparatus
CN114070354B (en) Adaptive segmented matrix inverse tracking MIMO (multiple input multiple output) detection method based on GS (generalized likelihood analysis) iterative method
US8295341B2 (en) Cognitive radio device
CN101499840B (en) Iteration detection method for MIMO system
CN114142905B (en) Improved method of MIMO detection algorithm based on Newton iteration method under time-varying channel
CN112953862A (en) High-speed railway scene channel estimation method based on deep learning
CN115250216A (en) Underwater sound OFDM combined channel estimation and signal detection method based on deep learning
CN114499601B (en) Large-scale MIMO signal detection method based on deep learning
CN108736934B (en) Large-scale MIMO system signal detection method
Gizzini et al. Adaptive channel estimation based on deep learning
CN101494625A (en) Linear equilibrium method and linear equalizer
CN115149988B (en) Self-adaptive segmented matrix inverse tracking MIMO detection method based on SOR iteration method
CN115173899B (en) Self-adaptive segmented matrix inverse tracking MIMO detection method based on Richardson iteration method
CN113660016B (en) EPA-based MIMO detection method, device, equipment and storage medium
CN113114423B (en) Adaptive sphere decoding detection method
CN112671419B (en) Wireless signal reconstruction method, device, system, equipment and storage medium
CN108282200B (en) Factor graph-based confidence propagation signal detection method in large-scale MIMO system
CN111682924B (en) Bidirectional frequency domain Turbo equalization method adopting expected propagation
CN109842581B (en) Channel estimation method based on three-level threshold variable-step-size adaptive compressed sensing technology
Niazadeh et al. ISI sparse channel estimation based on SL0 and its application in ML sequence-by-sequence equalization
CN110971546A (en) Channel tracking method for large-scale MIMO system
CN108449120B (en) Low-complexity QAM-MIMO detection method based on differential measurement

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