CN110380996B - Frequency-dependent IQ imbalance compensation method in SC-FDE system - Google Patents

Frequency-dependent IQ imbalance compensation method in SC-FDE system Download PDF

Info

Publication number
CN110380996B
CN110380996B CN201910628302.6A CN201910628302A CN110380996B CN 110380996 B CN110380996 B CN 110380996B CN 201910628302 A CN201910628302 A CN 201910628302A CN 110380996 B CN110380996 B CN 110380996B
Authority
CN
China
Prior art keywords
imbalance
channel
frequency
alpha
parameters
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.)
Expired - Fee Related
Application number
CN201910628302.6A
Other languages
Chinese (zh)
Other versions
CN110380996A (en
Inventor
杨莹
成先涛
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.)
University of Electronic Science and Technology of China
Original Assignee
University of Electronic Science and Technology of China
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 University of Electronic Science and Technology of China filed Critical University of Electronic Science and Technology of China
Priority to CN201910628302.6A priority Critical patent/CN110380996B/en
Publication of CN110380996A publication Critical patent/CN110380996A/en
Application granted granted Critical
Publication of CN110380996B publication Critical patent/CN110380996B/en
Expired - Fee Related 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
    • 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
    • H04L25/03—Shaping networks in transmitter or receiver, e.g. adaptive shaping networks
    • H04L25/03006—Arrangements for removing intersymbol interference
    • H04L25/03159—Arrangements for removing intersymbol interference operating in the frequency domain

Landscapes

  • Engineering & Computer Science (AREA)
  • Power Engineering (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Digital Transmission Methods That Use Modulated Carrier Waves (AREA)
  • Noise Elimination (AREA)

Abstract

The invention belongs to the technical field of wireless communication, and relates to a frequency-dependent IQ imbalance compensation method in an SD-FDE system. The invention respectively treats the multipath channel and IQ imbalance parameters as random variables and unknown deterministic parameters to convert the estimation problem into an Expectation Maximization (EM) framework, and updates the posterior probability of the multipath coefficient and the IQ imbalance parameters in an iterative manner by maximizing an objective function. When iteration is terminated, the estimation value of the IQ imbalance parameters can be obtained, and the posterior mean value of the multipath coefficient is used as the channel estimation value. The invention has the advantages that the IQ imbalance interference-free signal compensation method can effectively compensate IQ imbalance interference signals and obviously improve the bit error rate performance of a system.

Description

Frequency-dependent IQ imbalance compensation method in SC-FDE system
Technical Field
The invention belongs to the technical field of wireless communication, and relates to a frequency-dependent IQ imbalance compensation method in an SC-FDE system.
Background
In order to meet the increasing demand for high data rate transmission, Orthogonal Frequency Division Multiplexing (OFDM) and single carrier frequency domain equalization (SC-FDE) are widely used. Compared to OFDM, SC-FDE has a much lower peak-to-average power ratio (PAPR) of the modulated signal. This means that for SC-FDE the requirements on power amplifier linearity can be relaxed, which simplifies the transmitter design. On the other hand, SC-FDE is robust to amplifier non-linearity due to low PAPR, and thus can more efficiently use output power of the amplifier. Based on the above advantages, SC-FDE has been adopted by many wireless communication standards, such as IEEE 802.15.3c and IEEE 802.11 ad.
Meanwhile, in order to realize a low-cost, small-sized and power-efficient transceiver, it is necessary to solve the problem of radio frequency damage caused by non-idealities of analog devices in the transceiver, one of which is mismatching of amplitude, phase and frequency responses between I and Q branches in an analog front end, i.e., IQ imbalance. To describe IQ imbalance, the following two models are generally distinguished: frequency independent and frequency dependent. Frequency independent IQ imbalance is caused by non-ideal mixers. For an ideal mixer, its IQ branches should have the same amplitude and an exact 90-degree phase shift. However, there are some drawbacks in practical mixers, resulting in frequency independent IQ imbalance. In contrast, the frequency-dependent model considers not only frequency-independent IQ imbalance, but also frequency-dependent IQ imbalance. Frequency-dependent IQ imbalance is caused by filter mismatch and corresponds to an unbalanced frequency response of the IQ branch and varies with the system bandwidth. It is clear that the frequency dependent model is more comprehensive in characterizing IQ imbalance than the frequency independent model. IQ imbalance breaks the orthogonality between the I-branch and Q-branch signals and significantly degrades system performance, so it is necessary to compensate for IQ imbalance. For the frequency-dependent model in the SD-FDE system, the common literature compensation scheme only considers the IQ imbalance at the transmitting end or the receiving end, or estimates the channel and IQ imbalance parameters simultaneously to compensate the IQ imbalance at the transmitting end and the receiving end but has higher computational complexity.
Disclosure of Invention
The invention aims to provide a method for compensating for frequency-dependent IQ imbalance in an SD-FDE system under the condition that IQ imbalance exists at a transmitting end and a receiving end, so that the system performance is improved.
The invention respectively treats the multipath channel and IQ imbalance parameters as random variables and unknown deterministic parameters to convert the estimation problem into an Expectation Maximization (EM) framework, and updates the posterior probability of the multipath coefficient and the IQ imbalance parameters in an iterative manner by maximizing an objective function. When iteration is terminated, the estimation value of the IQ imbalance parameters can be obtained, and the posterior mean value of the multipath coefficient is used as the channel estimation value.
In order to facilitate the understanding of the technical solution of the present invention by those skilled in the art, a system model adopted by the present invention will be described first.
Considering the frequency-dependent IQ imbalance model of the transmitting and receiving ends, the parameter epsilonT、ΔφTRespectively representing the amplitude imbalance and the phase imbalance introduced at the transmitting end, the introduction of frequency dependence being passed through an equivalent low-pass filter hIT(t)、hQT(t). Using e at the receiving end respectivelyR、ΔφR、hIR(t)、hQR(t) represents the introduction of amplitude imbalance, phase imbalance and frequency dependence. The ideal complex baseband signal transmitted by the transmitting end is assumed to be x (t) ═ xI(t)+jxQ(t), due to the influence of IQ imbalance of the transmitting end, the rf transmission signal in the practical system is:
Figure BDA0002127861900000021
wherein, tBB(t)＝tI(t)+jtQ(t) is tRF(t) equivalent complex baseband signals. From (1) can be obtained:
Figure BDA0002127861900000022
wherein
Figure BDA0002127861900000023
αT＝cosΔφT+jεTsinΔφT，βT＝εTcosΔφT+jsinΔφT。
Radio frequency transmission signal passing through channel hRF(t) is propagated and subjected to Gaussian white noise wRFAfter the additive influence of (t), the band-pass signal sent to the receiving end is:
Figure BDA0002127861900000024
let h (t), w (t) be and h respectivelyRF(t)、wRF(t) the corresponding complex baseband impulse response, complex baseband Gaussian white noise, and rRF(t) corresponds toThe equivalent complex baseband signal r (t) can be expressed as:
Figure BDA0002127861900000031
rRF(t) first multiplying the local carrier interfered by IQ imbalance and down-converting the signal to baseband by a low-pass filter, wherein the complex baseband signal interfered by IQ imbalance at the receiving end is recorded as y (t) yI(t)+jyQ(t), wherein the I path and the Q path are respectively:
yI(t)＝LPF{2(1+εR)cos(2πfct+ΔφR)rRF(t)}
＝(1+εR)cos(ΔφR)rI(t)+(1+εR)sin(ΔφR)rQ(t) (5)
yQ(t)＝LPF{-2(1-εR)sin(2πfct-ΔφR)rRF(t)}
＝(1-εR)sin(ΔφR)rI(t)+(1-εR)cos(ΔφR)rQ(t) (6)
r (t) ═ rI(t)+jrQ(t) and bringing y (t) into yI(t)+jyQ(t), through a series of operations:
y(t)＝αRr(t)+βRr*(t) (7)
wherein alpha isR＝cosΔφR-jεRsinΔφR，βR＝εRcosΔφR+jsinΔφR。
Introduction of IQ unbalanced frequency correlation of receiving end is carried out through yI(t)、yQ(t) are respectively connected with an equivalent low-pass filter hIR(t)、hQR(t) and thus the final received equivalent complex baseband signal is:
Figure BDA0002127861900000032
wherein,
Figure BDA0002127861900000033
Figure BDA0002127861900000034
let x [ n ]]For a known training sequence of length N, x N]And its conjugate signal x*[n]N of (A)SPoint FFT is X respectivelykAnd
Figure BDA0002127861900000035
similarly, channel h [ n ]]And conjugation h thereof*[n]Through NSThe frequency domain responses obtained by the point FFT are respectively HkAnd
Figure BDA0002127861900000041
the equivalent low-pass FIR filters for introducing IQ two-path frequency response difference are h respectivelyIT＝[hIT1,hIT2,...,hITL']、hQT＝[hQT1,hQT2,...,hQTL']、hIR＝[hIR1,hIR2,...,hIRL']、hQR＝[hQR1,hQR2,...,hQRL']Where subscript I, Q denotes the I and Q signals, respectively, and subscript T, R denotes the transmit and receive ends, respectively, each of which contains L' taps. Therefore, when frequency-dependent IQ imbalance is introduced at the transmitting end and the receiving end simultaneously, the frequency domain expression of the final received signal is:
Figure BDA0002127861900000042
wherein,
Figure BDA0002127861900000043
Figure BDA0002127861900000044
Figure BDA0002127861900000045
Figure BDA0002127861900000046
is a frequency domain noise vector, HIT,k、HQT,k、HIR,k、HQR,kAre respectively hIT、hQT、hIR、hQRN of (A)SPoint FFT, derived from the definition of discrete fourier transform:
Figure BDA0002127861900000047
Figure BDA0002127861900000048
coefficient A in frequency domain expressionk、Bk、
Figure BDA0002127861900000049
IQ imbalance parameter alpha with transmitting endT、βT、hIT、hQTIs associated with Ck、DkIQ imbalance parameter alpha with receiving endR、βR、hIR、hQRCorrelation, considering the mutual independence between the parameters, and in most practical cases, there is α for the frequency-dependent IQ imbalance parameters of the transmitting and receiving endsT≈1，βT≈0，αR≈1，βR≈0，hIT≈ei，hQT≈ej，hIR≈ei，hQR≈ejWherein e ism(m ═ i, j) denotes a vector of length L', whose mth element is 1 and the remaining elements are 0. In view of this, an EM-based algorithm is proposed herein to perform joint estimation compensation on the frequency-dependent IQ imbalance parameters of the transmitting and receiving ends and the channel. Due to alphaT、βTAnd alphaR、βRThere is a conversion relationship between:
Figure BDA00021278619000000410
wherein
Figure BDA00021278619000000411
Representing taking the real and imaginary parts of the element, respectively. Therefore, when parameter estimation is carried out, the alpha does not need to be separately adjustedTAnd alphaREstimate by only dividing betaTAnd betaRSubstituting the estimated value into the formula to obtain alphaTAnd alphaRAn estimate of (d).
The invention is realized by the following steps:
s1, parameter initialization: initialization including IQ imbalance parameters and variance terms
By amplitude imbalance epsilon of the transmitting endTAnd phase imbalance delta phiTCo-determined transmit IQ imbalance parameter alphaT、βTAre respectively initialized to 1 and 0, and the amplitude of the receiving end is unbalanced by epsilonRAnd phase imbalance delta phiRJointly determined IQ imbalance parameter alpha of receiving endR、βRRespectively initialized to 1 and 0, and equivalent low-pass FIR filters of an I path and a Q path of a transmitting end are respectively initialized to hIT＝eiAnd hQT＝ejRespectively initializing h by equivalent low-pass FIR filters of I path and Q path of receiving endIR＝eiAnd hQR＝ejWherein e ism(m ═ i, j) denotes a vector of length L', whose mth element is 1 and the remaining elements are 0. Channel with a plurality of channels
Figure BDA0002127861900000051
A priori variance of
Figure BDA0002127861900000052
Is initialized to
Figure BDA0002127861900000053
Wherein L represents the effective length of the channel time domain impulse response h, and L represents the channel
Figure BDA0002127861900000054
While letting the noise variance β be-1＝Pn(PnRepresenting the variance of the noise power resulting from the signal-to-noise ratio).
S2, realizing the iteration of the EM algorithm through the following steps:
s21, updating the posterior mean value of the channel h:
in order to facilitate the use of the sparse nature of the channel, according to
Figure BDA0002127861900000055
And
Figure BDA0002127861900000056
Hkand
Figure BDA0002127861900000057
respectively representing channels h and h*N of (A)SPoint FFT, and
Figure BDA0002127861900000058
and an FFT column vector corresponding to the k-th subcarrier is represented, wherein N represents the length of the training sequence. Therefore, the frequency domain received signal is denoted as
Figure BDA0002127861900000059
Wherein
Figure BDA00021278619000000510
Consider NSThe sub-carriers include:
Figure BDA00021278619000000511
expressing the above formula in matrix form: S-Mh + Nh*+ W, then conversion to the real number domain yields:
Figure BDA0002127861900000061
wherein,
Figure BDA0002127861900000062
wherein
Figure BDA0002127861900000063
Representing taking the real and imaginary parts of the element, respectively. For channel h is sparse
Figure BDA0002127861900000064
Is also sparse, and
Figure BDA0002127861900000065
and
Figure BDA0002127861900000066
sharing the same sparsity. Suppose that
Figure BDA0002127861900000067
Each element in (1) obeys a mean of 0 and a variance of
Figure BDA0002127861900000068
And α ═ α1,α2,...,α2L]To represent
Figure BDA0002127861900000069
A set of all element variances;
Figure BDA00021278619000000610
each element in the series obeys a mean value of 0 and a variance of beta-1The same gaussian distribution. Then
Figure BDA00021278619000000611
The prior probability of (d) can be written as:
Figure BDA00021278619000000612
conditional probability
Figure BDA00021278619000000613
Comprises the following steps:
Figure BDA00021278619000000614
based on
Figure BDA00021278619000000615
The channel can be obtained by using the formulas (12) and (13)
Figure BDA00021278619000000616
A posteriori probability of
Figure BDA00021278619000000617
Figure BDA00021278619000000618
Wherein,
Figure BDA00021278619000000619
t＝(βΦHΦ+D)-1and D ═ diag (alpha)1,α2,…αL,α1,α2,…αL). The mean posterior probability of the complex channel h is thus μ1μ (1: L,1) + i μ (L +1:2L,1), i representing the imaginary part of the complex number. The posterior mean value at this time is the estimated value of the channel h. The estimation of the h posterior mean requires simultaneous estimation of hyper-parameters α, β, taking into account the full likelihood function with respect to α, β using Sparse Bayes (SBL) criterion:
Figure BDA0002127861900000071
by maximizing the full likelihood function with respect to the parameters alpha, beta and separately for the parameter alphalBeta, obtaining a derivative and a value 0 to obtain a parameter alphalThe update formula of beta is:
Figure BDA0002127861900000072
Figure BDA0002127861900000073
s22, updating IQ imbalance parameters
Figure BDA0002127861900000079
Is estimated value of
Consider the objective function Q (θ) and retain terms related only to θ:
Figure BDA0002127861900000074
since in practical cases tr (Φ)HPhi t) is much smaller than
Figure BDA0002127861900000075
Therefore, only consider here
Figure BDA0002127861900000076
Where tr (-) denotes the trace of the matrix. Will channel H and HkIs recorded as V and VkEquation (18) can be expressed as:
Figure BDA0002127861900000077
the parameter θ is updated by maximizing the objective function Q (θ), i.e., minimizing Q (θ). Definition of
Figure BDA0002127861900000078
Wherein t ═ βT,βR]TAnd estimating each subset theta of the parameter theta by Newton's methodi：
θi＝θi m-H-1(θi)δ(θi) (20)
Wherein the different parameters thetaiThe corresponding Hessian matrix H and gradient δ are:
Figure BDA0002127861900000081
Figure BDA0002127861900000082
Figure BDA0002127861900000083
Figure BDA0002127861900000084
Figure BDA0002127861900000085
Figure BDA0002127861900000086
Figure BDA0002127861900000087
Figure BDA0002127861900000088
Figure BDA0002127861900000089
Figure BDA00021278619000000810
Figure BDA00021278619000000811
Figure BDA00021278619000000812
Figure BDA0002127861900000091
Figure BDA0002127861900000092
Figure BDA0002127861900000093
Figure BDA0002127861900000094
Figure BDA0002127861900000095
Figure BDA0002127861900000096
Figure BDA0002127861900000097
Figure BDA0002127861900000098
Figure BDA0002127861900000099
Figure BDA00021278619000000910
a aboven,k(n＝1,2,3,4,5)，b1,k，c1,k，d1,k，e1,kThe intermediate variables introduced only for the convenience of representation of the Hessian matrix H and the gradient δ have no practical meaning.
S23, looping steps S21 and S22 until the maximum external iteration number N is reachedtAnd the posterior mean value mu at this time is1As an estimate of channel h.
S3, frequency-dependent IQ imbalance compensation and channel impact removal:
the received signal of equation (9) is additionally denoted as matrix form:
Figure BDA0002127861900000101
based on MMSE criterion, the recovered initial transmission signal is:
Figure BDA0002127861900000102
where η represents the signal-to-noise ratio and I is the identity matrix.
The invention has the advantages that the IQ imbalance interference-free signal compensation method can effectively compensate IQ imbalance interference signals and obviously improve the bit error rate performance of a system.
Drawings
FIG. 1 is a diagram of a transceiving end frequency-dependent IQ imbalance model used in the present invention;
FIG. 2 is a graph of BER performance of the algorithm of the present invention in LOS channel case 1;
FIG. 3 is a graph of the BER performance curve of the algorithm of the present invention in NLOS channel case 1;
FIG. 4 is a graph of BER performance of the algorithm of the present invention in LOS channel case 2;
fig. 5 is a graph of the BER performance curve of the algorithm of the present invention in NLOS channel case 2.
Detailed Description
The effectiveness of the invention is illustrated below with reference to the figures and simulation examples:
simulation of modulation based on SC-FDE system and using QPSKThe symbol rate is 1.76GHz, the up-sampling multiple of raised cosine roll-off filtering is 8, the roll-off factor is 0.25, the lead code and the channel adopt 802.11.ad standard, and the amplitude imbalance is set as epsilonT＝εRAt 1dB, the phase imbalance is set to Δ φT＝ΔφR5 ° (in a real system, IQ imbalance is usually better than the simulation setup). For frequency correlation, two three-tap cases are considered:
case 1: h isI＝[1.02,0.04,-0.03]T，hQ＝[1.03,-0.02,0.012]T
Case 2: h isI＝[0.01,1,0.01]T，hQ＝[0.01,1,0.2]T
Wherein the frequency correlation introduced by case 1 is small, similar to the case of frequency independent IQ imbalance, the frequency correlation introduced by case 2 is large.
Fig. 2 and fig. 3 are graphs of bit error rate simulations of LOS and NLOS channels in case 1, and it can be seen from these two graphs that the performance of the iterative estimation algorithm proposed herein for estimation and compensation of IQ imbalance at the transmitting and receiving ends is excellent, and is sufficiently close to the performance curve of MMSE compensation of an ideal channel, which shows that the estimation of channel state information by the algorithm proposed herein is sufficiently close to the ideal channel state information. Meanwhile, the error code performance after one iteration compensation is very close to the situation without IQI, so that the excellent performance effect can be achieved by only one iteration, thereby reducing the adverse effect of IQ imbalance.
Fig. 4 and 5 are simulation graphs of the bit error rate of LOS and NLOS channels in case 2, and it can be seen from these graphs that the iterative algorithm herein still has good performance for estimating and compensating IQ imbalance at the transmitting and receiving ends under the condition of severe frequency dependence.

Claims (1)

  1. A method for frequency-dependent IQ imbalance compensation in an SD-FDE system, comprising the steps of:
    s1, parameter initialization: by amplitude imbalance epsilon of the transmitting endTAnd phase imbalance Δ φTAre combined togetherDetermined transmitter IQ imbalance parameter alphaT、βTAre respectively initialized to 1 and 0, and the amplitude of the receiving end is unbalanced by epsilonRAnd phase imbalance Δ φRJointly determined IQ imbalance parameter alpha of receiving endR、βRRespectively initialized to 1 and 0, and equivalent low-pass FIR filters of an I path and a Q path of a transmitting end are respectively initialized to hIT＝eiAnd hQT＝ejRespectively initializing h by equivalent low-pass FIR filters of I path and Q path of receiving endIR＝eiAnd hQR＝ejWherein e ism(m ═ i, j) denotes a vector of length L', whose mth element is 1 and the remaining elements are 0; channel with a plurality of channels
    Figure FDA0002948295020000011
    A priori variance of
    Figure FDA0002948295020000012
    Is initialized to
    Figure FDA0002948295020000013
    L is more than or equal to 1 and less than or equal to 2L, wherein L represents the effective length of the channel time domain impulse response, and L represents the channel
    Figure FDA0002948295020000014
    While letting the noise variance β be-1＝Pn，PnRepresenting the variance of the noise power resulting from the signal-to-noise ratio;
    s2, realizing the iteration of the EM algorithm through the following steps:
    s21, updating the posterior mean value of the channel h:
    in order to facilitate the use of the sparse nature of the channel, according to
    Figure FDA0002948295020000015
    And
    Figure FDA0002948295020000016
    Hkand
    Figure FDA0002948295020000017
    respectively representing channels h and h*Through NSFrequency domain response obtained by point FFT, and FkRepresents the FFT column vector, N, corresponding to the k-th sub-carrierSIs the number of subcarriers, where N represents the length of the training sequence, and the frequency domain received signal is denoted as
    Figure FDA0002948295020000018
    Wherein
    Figure FDA0002948295020000019
    Figure FDA00029482950200000110
    Figure FDA00029482950200000111
    Figure FDA00029482950200000112
    Figure FDA00029482950200000113
    Figure FDA0002948295020000021
    Figure FDA0002948295020000022
    Is a noise vector, HIT,k、HQT,k、HIR,k、HQR,kAre respectively hIT、hQT、hIR、hQRN of (A)SPoint FFT, XkFor a known training sequence x N of length N]N of (A)SThe point FFT is a point FFT that is,
    Figure FDA0002948295020000023
    is x [ n ]]Is a conjugate signal x of*[n]N of (A)SThe point FFT is a point FFT that is,
    Figure FDA0002948295020000024
    then there are:
    Figure FDA0002948295020000025
    expressing the above formula in matrix form: S-Mh + Nh*+ W, then conversion to the real number domain yields:
    Figure FDA0002948295020000026
    wherein,
    Figure FDA0002948295020000027
    wherein
    Figure FDA0002948295020000028
    Representing the real and imaginary parts of the elements, respectively, since the channel h is sparse
    Figure FDA0002948295020000029
    Is also sparse, and
    Figure FDA00029482950200000210
    and
    Figure FDA00029482950200000211
    share the same sparsity, order
    Figure FDA00029482950200000212
    Each element in (1) obeys a mean of 0 and a variance of
    Figure FDA00029482950200000213
    The same gaussian distribution of (a);
    Figure FDA00029482950200000214
    each element in the series obeys a mean value of 0 and a variance of beta-1The same gaussian distribution of (a);
    Figure FDA00029482950200000215
    the prior probability of (a) is:
    Figure FDA00029482950200000216
    conditional probability
    Figure FDA00029482950200000217
    Comprises the following steps:
    Figure FDA00029482950200000218
    based on
    Figure FDA0002948295020000031
    The channel can be obtained by using the formulas (12) and (13)
    Figure FDA0002948295020000032
    A posteriori probability of
    Figure FDA0002948295020000033
    Figure FDA0002948295020000034
    Wherein,
    Figure FDA0002948295020000035
    τ＝(βΦHΦ+D)-1and D ═ diag (alpha)1,α2,…αL,α1,α2,…αL) So that the mean posterior probability of the complex channel h is μ1μ (1: L,1) + i μ (L +1:2L,1), i denotes the imaginary part of the complex number, and the posterior mean value at this time is the estimated value of channel h; the estimation of the h posterior mean value needs to estimate the hyperparameters alpha and beta simultaneously, and the full likelihood function of the alpha and the beta is considered by utilizing the sparse Bayesian criterion SBL:
    Figure FDA0002948295020000036
    by maximizing the full likelihood function with respect to the parameters alpha, beta and separately for the parameter alphalBeta, obtaining a derivative and a value 0 to obtain a parameter alphalThe update formula of beta is:
    Figure FDA0002948295020000037
    Figure FDA0002948295020000038
    s22, updating IQ imbalance parameters
    Figure FDA0002948295020000039
    Is estimated value of
    The objective function Q (θ) retains terms related only to θ:
    Figure FDA00029482950200000310
    since in practical cases tr (Φ)HΦ τ) are much smaller than
    Figure FDA00029482950200000311
    tr (-) denotes the trace of the matrix and therefore only consider
    Figure FDA00029482950200000312
    Will channel H and HkIs recorded as V and VkEquation (18) is expressed as:
    Figure FDA0002948295020000041
    updating the parameter θ by maximizing the objective function Q (θ), i.e., minimizing Q (θ); definition of
    Figure FDA0002948295020000042
    Wherein t ═ βT,βR]TAnd estimating each subset theta of the parameter theta by Newton's methodi：
    θi＝θi m-H-1(θi)δ(θi) (20)
    Wherein the different parameters thetaiThe corresponding Hessian matrix and gradient are:
    Figure FDA0002948295020000043
    Figure FDA0002948295020000044
    Figure FDA0002948295020000045
    Figure FDA0002948295020000046
    Figure FDA0002948295020000047
    Figure FDA0002948295020000048
    Figure FDA0002948295020000049
    Figure FDA00029482950200000410
    Figure FDA00029482950200000411
    Figure FDA0002948295020000051
    Figure FDA0002948295020000052
    Figure FDA0002948295020000053
    Figure FDA0002948295020000054
    Figure FDA0002948295020000055
    Figure FDA0002948295020000056
    Figure FDA0002948295020000057
    Figure FDA0002948295020000058
    Figure FDA0002948295020000059
    Figure FDA00029482950200000510
    Figure FDA00029482950200000511
    Figure FDA00029482950200000512
    Figure FDA00029482950200000513
    a aboven,k(n＝1,2,3,4,5)，b1,k，c1,k，d1,k，e1,kIntermediate variables are introduced only for the convenience of representation of the Hessian matrix H and the gradient δ, and have no actual physical meaning;
    s23, looping steps S21 and S22 until the maximum external iteration number N is reachedtAnd the posterior mean value mu at this time is1As an estimate of channel h;
    s3, frequency-dependent IQ imbalance compensation and channel impact removal:
    the received signal of equation (9) is additionally denoted as matrix form:
    Figure FDA0002948295020000061
    based on MMSE criterion, the recovered initial transmission signal is:
    Figure FDA0002948295020000062
    where η represents the signal-to-noise ratio and I is the identity matrix.
CN201910628302.6A 2019-07-12 2019-07-12 Frequency-dependent IQ imbalance compensation method in SC-FDE system Expired - Fee Related CN110380996B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910628302.6A CN110380996B (en) 2019-07-12 2019-07-12 Frequency-dependent IQ imbalance compensation method in SC-FDE system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910628302.6A CN110380996B (en) 2019-07-12 2019-07-12 Frequency-dependent IQ imbalance compensation method in SC-FDE system

Publications (2)

Publication Number Publication Date
CN110380996A CN110380996A (en) 2019-10-25
CN110380996B true CN110380996B (en) 2021-06-08

Family

ID=68252895

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910628302.6A Expired - Fee Related CN110380996B (en) 2019-07-12 2019-07-12 Frequency-dependent IQ imbalance compensation method in SC-FDE system

Country Status (1)

Country Link
CN (1) CN110380996B (en)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113055032B (en) * 2019-12-27 2022-07-19 珠海全志科技股份有限公司 Method and device for acquiring compensation values of IQ mismatch and carrier leakage and computer equipment
CN111342850B (en) * 2020-03-01 2021-07-06 中国电子科技集团公司第五十四研究所 Receiving array broadband quadrature imbalance compensation device
CN111416783B (en) * 2020-03-31 2021-10-26 华南理工大学 Circulation-based IQ imbalance adaptive blind compensation method and system
CN111917672B (en) * 2020-07-18 2023-05-30 青岛鼎信通讯股份有限公司 IQ path orthogonal estimation and calibration method for carrier wireless dual-mode communication

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103312640A (en) * 2013-06-30 2013-09-18 电子科技大学 Channel estimation and IQ (In-phase Quadrature) imbalance united compensation method
CN104639490A (en) * 2015-01-27 2015-05-20 电子科技大学 Joint estimation and compensation method for frequency-dependent IQ (In-phase Quadrature) mismatch and channel
CN108616469A (en) * 2018-05-11 2018-10-02 东南大学 The estimation of receiving terminal IQ imbalances and compensation method of a kind of SC-FDE systems and device

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2003101064A1 (en) * 2002-05-23 2003-12-04 Interuniversitair Microelektronica Centrum Vzw Method and device for estimating and compensating iq imbalance
TWI264193B (en) * 2005-03-04 2006-10-11 Mediatek Inc Apparatus and method for compensating IQ imbalance in OFDM system with carrier frequency offset
TWI309940B (en) * 2006-06-05 2009-05-11 Realtek Semiconductor Corp Method and device for compensating iq imbalance
CN104935537A (en) * 2015-06-17 2015-09-23 电子科技大学 Joint estimation compensation improving method of transmitter-receiver IQ imbalance and channel
CN106161304B (en) * 2016-04-01 2019-03-12 电子科技大学 A kind of transmitting terminal IQ imbalance compensation method of joint channel estimation

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103312640A (en) * 2013-06-30 2013-09-18 电子科技大学 Channel estimation and IQ (In-phase Quadrature) imbalance united compensation method
CN104639490A (en) * 2015-01-27 2015-05-20 电子科技大学 Joint estimation and compensation method for frequency-dependent IQ (In-phase Quadrature) mismatch and channel
CN108616469A (en) * 2018-05-11 2018-10-02 东南大学 The estimation of receiving terminal IQ imbalances and compensation method of a kind of SC-FDE systems and device

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
"Comparison of the sensitivity of OFDM and SC-FDE to CFO, SCO and IQ imbalance";F. Horlin,A. Bourdoux;《3rd International Symposium on Communications, Control and Signal》;20080606;111-116 *
"Three-Stage Treatment of TX/RX IQ Imbalance and Channel with CFO for SC-FDE Systems";C. Zhang, Z. Xiao, B. Gao, et al.;《IEEE Communications Letters》;20141231;297-300 *
"Training sequence design for joint channel and I/Q imbalance parameter estimation in mobile SC-FDE transceivers";S. Narayanan, B. Narasimhan, N. Al-Dhahir.;《2010 IEEE International Conference on Acoustics, Speech and Signal Processing》;20100628;1392-1395 *

Also Published As

Publication number Publication date
CN110380996A (en) 2019-10-25

Similar Documents

Publication Publication Date Title
CN110380996B (en) Frequency-dependent IQ imbalance compensation method in SC-FDE system
US10778498B2 (en) TX IQ mismatch pre-compensation
CN103312640B (en) A kind of method of joint channel estimation and IQ imbalance compensation
US8971465B2 (en) Receiver-side estimation of and compensation for signal impairments
Zhou et al. Finite-alphabet based channel estimation for OFDM and related multicarrier systems
CN104717172B (en) IQ imbalance compensations method and apparatus in a kind of emitter
CN104639490B (en) A kind of Combined estimator and the compensation method of frequency dependence IQ imbalance and channel
US8634496B2 (en) Method for determining and compensating transceiver non-idealities
CN105847198B (en) The IQ imbalances of OFDM-WLAN radio frequency test systems estimate and compensation method
CN102769599B (en) A kind of novel signal of orthogonal frequency division multiplexing system processing method and device
CN106101045B (en) A kind of OFDM full duplex polarization self-interference removing method based on phase noise Gauss albefaction
US8126086B2 (en) System and method for in-phase/quadrature-phase (I/Q) mismatch measurement and compensation
CA2724735A1 (en) System and method for iq imbalance estimation using loopback with frequency offset
CN105791182B (en) IQ imbalances and channel joint estimation method suitable for MIMO-OFDM systems
CN109104201A (en) Based on the frequency dependence IQ mismatch calibration of FFT operation and compensation method
CN108616469B (en) Method and device for estimating and compensating IQ imbalance of receiving end of SC-FDE system
CN106161304B (en) A kind of transmitting terminal IQ imbalance compensation method of joint channel estimation
CN113472712B (en) Phase noise suppression method
CN104935537A (en) Joint estimation compensation improving method of transmitter-receiver IQ imbalance and channel
CN111525946B (en) Anti-time-varying wave beam sequence number modulation method suitable for broadband large-scale millimeter wave system
WO2010017064A1 (en) Iq imbalance equalization system and method
CN107147606B (en) Lattice reduction assisted linear detection method in generalized spatial modulation
CN101582864B (en) SAGE channel estimation method based on partial interference cancellation
CN108540187B (en) Method and device for improving performance of nonlinear MU MIMO system based on MMSE (minimum mean square error) criterion
CN105227506A (en) The I/Q unbalance compensating method of frequency selectivity ofdm communication 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
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20210608

CF01 Termination of patent right due to non-payment of annual fee