CN104468429A - Threshold time domain least-square channel estimation method based on noise variance estimation - Google Patents

Abstract

The invention discloses a channel estimation algorithm of an IQ imbalance OFDM system under a sparse environment. A pilot symbol passes through a channel under the IQ imbalance condition, and a receiving end signal is obtained. IQ imbalance is taken as a channel parameter to be brought in estimation, sparse characteristics are utilized, and a time domain channel impact response is estimated through a least square method. A proper threshold is determined by estimating noise variance, and noise within cyclic prefix length is filtered out. According to the threshold time domain least-square estimation method based on noise variance estimation, good compromise of performance and complexity is achieved, meanwhile, only one OFDM symbol is needed, the high spectrum utilization rate is achieved, and system error code performance can be effectively improved.

Description

Based on the thresholding time domain LS channel estimation method of Noise Variance Estimation

Technical field

The present invention relates to wireless communication technology field, the thresholding time domain LS channel estimation algorithm particularly based on Noise Variance Estimation in ofdm system under the uneven condition of sparse channel of IQ.

Background technology

OFDM (Orthogonal Frequency Division Multiplexing, OFDM) be a kind of special Multicarrier Transmission Technology, it flows through serioparallel exchange serial data at a high speed and is assigned to parallel transmission on several relatively low subchannels of transmission rate.As the core technology of 4G communication, it has the plurality of advantages such as anti-multipath jamming ability is strong, the availability of frequency spectrum is high.But, in an ofdm system, the not matched filtering of the imperfect frequency conversion of transmitter and receiver, homophase and quadrature branch, amplification and sampling all can cause the amplitude of inphase/orthogonal (In-phase and quadrature-phase, IQ) branch road and phase place not to mate, and namely IQ is uneven.IQ imbalance can destroy the orthogonality between subcarrier, is formed between subchannel and disturbs, have a strong impact on systematic function.

Owing to having, structure is simple, low-power consumption and low cost and other advantages are paid close attention to widely and applied obtaining for Direct Conversion transreceiver in recent years, but compared to traditional super heterodyne architectures, Direct Conversion structure transreceiver is more easily subject to the unbalanced impact of IQ.Therefore, high-performance channel estimating is realized to the IQ of ofdm system imbalance and fine compensation is the main path improving the uneven ofdm system error performance of IQ, there is very profound significance and important theoretical research value and application prospect.

Summary of the invention

Technical problem to be solved by this invention is for the uneven ofdm system of the inphase/orthogonal under sparse environment, provide a kind of thresholding time domain LS channel estimation method based on Noise Variance Estimation of low complex degree, effectively to improve system error performance.

The present invention is for solving the problems of the technologies described above by the following technical solutions: a kind of thresholding time domain LS channel estimation method based on Noise Variance Estimation, comprising:

Consider the IQ imbalance of transmitting, receiving terminal, be used as channel parameter and include estimation in the lump;

Utilize the sparse characteristic of channel, carry out time-domain filtering, and determine suitable thresholding by estimating noise variance, the noise in filtration cycle prefix length.

As a kind of embodiment, the IQ of described consideration transmitting, receiving terminal is uneven, is used as channel parameter and includes estimation in the lump, comprising:

Step 11), if OFDM sub-carrier number is N, circulating prefix-length is L, gets an OFDM frequency pilot sign $\stackrel{\‾}{S}={[s\left(1\right)s\left(2\right)......s\left(N\right)]}^T;$

Step 12), consider the IQ imbalance of transmitting, receiver, the multipath fading through channel obtains receiving terminal Z=(μ _rμ _tdiag{H}+v _rv _t ^*diag{H ^#) S+ μ _rv _tdiag{H}+v _rμ _t ^*diag{H ^#) S ^#+ ω,

Wherein, $\mathrm{\μ}=\cos \left(\frac{\mathrm{\θ}}{2}\right)+\mathrm{j\α}\sin \left(\frac{\mathrm{\θ}}{2}\right),v=\mathrm{\α}\cos \left(\frac{\mathrm{\θ}}{2}\right)-j\sin \left(\frac{\mathrm{\θ}}{2}\right),$ Subscript r, t represent receiving terminal, transmitting terminal respectively, and α, θ are amplitude and the phase distortion parameter of in-phase branch and quadrature branch, and H is channel impulse response h=[h ₀h ₁h _l] ^tdFT conversion, subscript # represent complex conjugate vector DFT conversion;

Step 13), note H _a=μ _rμ _tdiag{H}+v _rv _t ^*diag{H ^#, H _b=μ _rv _tdiag{H}+v _rμ _t ^*diag{H ^#, H _a=μ _rμ _tdiag{H}+v _rv _t ^*diag{H ^#, H _b=μ _rv _tdiag{H}+v _rμ _t ^*diag{H ^#, h _a, h _bbe respectively H _a, H _biDFT conversion, Z is expressed as Z=AX+ ω, wherein A=[diag{S}F _{n × N/2}diag{S ^#f _{n × N/2}], $X=\left[\begin{array}{c}{h}_a\\ 0_{(N/2-L)\×1}\\ h_b\\ 0_{(N/2-L)\×1}\end{array}\right],$ So X can be expressed as under time domain least-squares estimation

As a kind of embodiment, the described coefficient feature utilizing channel, carries out time-domain filtering, and determines suitable thresholding by estimating noise variance, and the noise in filtration cycle prefix length, comprising:

Step 21), according to time domain channel shock response concentration of energy in circulating prefix-length, after Cyclic Prefix individual value zero setting;

Step 22) definition thresholding wherein, after Cyclic Prefix the noise variance mean value of individual point, the span of coefficient gamma is $1\≤\mathrm{\γ}{\≤}_{n=1:L}^{\max }\left|{\hat{X}}_{\mathrm{LS}}\left(n\right)\right|;$

Step 23) incremented γ with 0.01, the value zero setting of λ will be less than in circulating prefix-length, obtain and with this restoring signal S, make the S error rate that estimates minimum, determine optimum thresholding λ ₀, obtain channel impulse response $\widehat{h_a}={\hat{X}}_{\mathrm{TB}-\mathrm{LS}}[1:L],\widehat{h_b}={\hat{X}}_{\mathrm{TB}-\mathrm{LS}}[\frac{N}{2}+1:\frac{N}{2}+L].$

The present invention adopts above technical scheme compared with prior art, has following technique effect:

The thresholding time domain least-squares algorithm performance based on Noise Variance Estimation proposed is better than existing frequency-domain and time-domain LS channel estimation method, approach the time domain iterative shrinkage algorithm based on compressed sensing, and computation complexity is starkly lower than the latter, achieve the good folding of performance and complexity, only need use an OFDM symbol simultaneously, there is the higher availability of frequency spectrum, can effectively improve system error performance.

The aspect that the present invention adds and advantage will part provide in the following description, and these will become obvious from the following description, or be recognized by practice of the present invention.

Accompanying drawing explanation

Fig. 1 is principle of the invention schematic diagram.

Embodiment

Below in conjunction with the drawings and specific embodiments, illustrate the present invention further, these embodiments should be understood only be not used in for illustration of the present invention and limit the scope of the invention, after having read the present invention, the amendment of those skilled in the art to the various equivalent form of value of the present invention has all fallen within the application's claims limited range.

Under a kind of sparse environment of the present invention, the uneven ofdm system of IQ comprises based on the time domain LS channel estimation method of noise variance thresholding: estimate channel impulse response by LS under IQ imbalance, then estimating noise variance, determines that in suitable thresholding filtration cycle prefix length, noise is to improve estimated accuracy.

As a kind of embodiment, described method comprises: the IQ imbalance considering transmitting, receiving terminal, it is used as channel parameter and includes estimation in the lump, estimate to obtain channel impulse response by LS.Utilize the sparse characteristic of channel, carry out time-domain filtering, and determine suitable thresholding by estimating noise variance, the noise in filtration cycle prefix length, improve estimated accuracy.

The first step: set OFDM sub-carrier number as N, circulating prefix-length is L.Get OFDM frequency pilot sign S=[s (1) s (2) ... s (N)] ^t.

Second step: the IQ imbalance considering transmitting, receiver, the multipath fading through channel obtains receiving terminal Z=(μ _rμ _tdiag{H}+v _rv _t ^*diag{H ^#) S+ μ _rv _tdiag{H}+v _rμ _t ^*diag{H ^#) S ^#+ ω.Wherein, $\mathrm{\μ}=\cos \left(\frac{\mathrm{\θ}}{2}\right)+\mathrm{j\α}\sin \left(\frac{\mathrm{\θ}}{2}\right),v=\mathrm{\α}\cos \left(\frac{\mathrm{\θ}}{2}\right)-j\sin \left(\frac{\mathrm{\θ}}{2}\right),$ Subscript r, t represent receiving terminal, transmitting terminal respectively.α, θ are amplitude and the phase distortion parameter of in-phase branch and quadrature branch.H is channel impulse response h=[h ₀h ₁h _l] ^tdFT conversion.Subscript # represents the DFT conversion of complex conjugate vector, is defined as: $X=\left[\begin{array}{c}X\left(1\right)\\ X\left(2\right)\\ .\\ .\\ .\\ X(N/2)\\ X(N/2+1)\\ X(N/2+2)\\ .\\ .\\ .\\ X\left(N\right)\end{array}\right]\→X^{\#}=\left[\begin{array}{c}{X}^{*}\left(1\right)\\ X^{*}\left(N\right)\\ .\\ .\\ .\\ X^{*}(N/2+2)\\ X^{*}(N/2+1)\\ X^{*}(N/2)\\ .\\ .\\ .\\ X^{*}\left(2\right)\end{array}\right]$

3rd step: note H _a=μ _rμ _tdiag{H}+v _rv _t ^*diag{H ^#, H _b=μ _rv _tdiag{H}+v _rμ _t ^*diag{H ^#, H _a=μ _rμ _tdiag{H}+v _rv _t ^*diag{H ^#, H _b=μ _rv _tdiag{H}+v _rμ _t ^*diag{H ^#, h _a, h _bbe respectively H _a, H _biDFT conversion.Z is expressed as Z=AX+ ω, wherein A=[diag{S}F _{n × N/2}diag{S ^#f _{n × N/2}], $X=\left[\begin{array}{c}{h}_a\\ 0_{(N/2-L)\×1}\\ h_b\\ 0_{(N/2-L)\×1}\end{array}\right],$ So X can be expressed as under time domain least-squares estimation

4th step: because time domain impulse channel energy concentrates in circulating prefix-length, the point after Cyclic Prefix only comprises noise, so after circulating prefix-length individual value zero setting.

5th step: for suppressing the noise in Cyclic Prefix further, introduces noise variance threshold judgement, will be less than the value filtering of thresholding.Definition thresholding wherein, after Cyclic Prefix the noise variance mean value of individual point, the span of coefficient gamma is $1\≤\mathrm{\γ}{\≤}_{n=1:L}^{\max }\left|{\hat{X}}_{\mathrm{LS}}\left(n\right)\right|.$

6th step: increase progressively in units of 1 in the span of γ, search optimum thresholding coefficient makes the S error rate that estimates minimum.Note optimum thresholding is λ ₀, with λ in circulating prefix-length ₀for threshold filter noise, obtain channel impulse response $h_a={\hat{X}}_{\mathrm{TB}-\mathrm{LS}}[1:L],h_b={\hat{X}}_{\mathrm{TB}-\mathrm{LS}}[\frac{N}{2}+1:\frac{N}{2}+L].$

Under the sparse environment that the present invention proposes, the uneven ofdm system of IQ is estimated based on the time domain LS channel of noise variance thresholding individualized training symbol and IQ compensating parameter, performance is better than existing frequency-domain and time-domain LS channel estimation method, approach the time domain iterative shrinkage algorithm based on compressed sensing, and computation complexity is starkly lower than the latter, achieve the good folding of performance and complexity, only need use an OFDM symbol simultaneously, there is the higher availability of frequency spectrum, can effectively improve system error performance.

Claims (3)

1., based on a thresholding time domain LS channel estimation method for Noise Variance Estimation, it is characterized in that, comprising:

Consider the IQ imbalance of transmitting, receiving terminal, be used as channel parameter and include estimation in the lump;

Utilize the sparse characteristic of channel, carry out time-domain filtering, and determine suitable thresholding by estimating noise variance, the noise in filtration cycle prefix length.

2. method according to claim 1, is characterized in that, the described IQ imbalance considering transmitting, receiving terminal, is used as channel parameter and includes estimation in the lump, comprising:

Step 11), if OFDM sub-carrier number is N, circulating prefix-length is L, gets an OFDM frequency pilot sign $\stackrel{\‾}{S}={\left[\begin{array}{ccccc}s\left(1\right)& s\left(2\right)& ...& ...& s\left(N\right)\end{array}\right]}^T;$

Step 12), consider the IQ imbalance of transmitting, receiver, the multipath fading through channel obtains receiving terminal Z=(μ _rμ _tdiag{H}+v _rv _t ^*diag{H ^#) S+ μ _rv _tdiag{H}+v _rμ _t ^*diag{H ^#) S ^#+ ω,

Wherein, $\mathrm{\μ}=\cos \left(\frac{\mathrm{\θ}}{2}\right)+\mathrm{j\α}\sin \left(\frac{\mathrm{\θ}}{2}\right),v=\mathrm{\α}\cos \left(\frac{\mathrm{\θ}}{2}\right)-j\sin \left(\frac{\mathrm{\θ}}{2}\right),$ Subscript r, t represent receiving terminal, transmitting terminal respectively, and α, θ are amplitude and the phase distortion parameter of in-phase branch and quadrature branch, and H is channel impulse response h=[h ₀h ₁h _l] ^tdFT conversion, subscript # represent complex conjugate vector DFT conversion;

Step 13), note H _a=μ _rμ _tdiag{H}+v _rv _t ^*diag{H ^#, H _b=μ _rv _tdiag{H}+v _rμ _t ^*diag{H ^#, H _a=μ _rμ _tdiag{H}+v _rv _t ^*diag{H ^#, H _b=μ _rv _tdiag{H}+v _rμ _t ^*diag{H ^#, h _a, h _bbe respectively H _a, H _biDFT conversion, Z is expressed as Z=AX+ ω, wherein A=[diag{S}F _{n × N/2}diag{S ^#f _{n × N/2}], $X=\left[\begin{array}{c}{h}_a\\ 0_{(N/2-L)\×1}\\ h_b\\ 0_{(N/2-L)\×1}\end{array}\right],$ So X can be expressed as under time domain least-squares estimation

3. method according to claim 1, is characterized in that, the described coefficient feature utilizing channel, carries out time-domain filtering, and determines suitable thresholding by estimating noise variance, and the noise in filtration cycle prefix length, comprising:

Step 21), according to time domain channel shock response concentration of energy in circulating prefix-length, after Cyclic Prefix individual value zero setting;

Step 22) definition thresholding wherein, after Cyclic Prefix the noise variance mean value of individual point, the span of coefficient gamma is

Step 23) incremented γ with 0.01, the value zero setting of λ will be less than in circulating prefix-length, obtain and with this restoring signal S, make the S error rate that estimates minimum, determine optimum thresholding λ ₀, obtain channel impulse response