CN103441967B - Ofdm system channel estimation and signal detecting method based on basis expansion model - Google Patents

Abstract

Ofdm system channel estimation and signal detecting method based on basis expansion model, belongs to wireless and mobile communication technology field.Specifically include following steps：Transmitting terminal sends ofdm signal, receiving terminal Channel Modeling, initialization, estimation basis expansion model coefficient, channel equalization and signal detection, iteration, output.Receiving terminal Channel Modeling receiving terminal frequency-region signal is disturbed and useful information separation, initial state assumption non-cycle prefix is limited, estimate basis expansion model coefficient and detect and obtain currently transmitted data symbol, intersymbol interference frequency response and the frequency response of Cyclic Prefix reconstruct part are obtained by iteration, and then eliminates the impact that intersymbol interference brings.Method of estimation proposed by the present invention, on the one hand, can estimate the obvious channel of time variation, can effectively eliminate the intersymbol interference of multidiameter delay introducing again, complete channel estimation and signal detection to ofdm system under doubly selective channel, improve systematic function.

Description

OFDM system channel estimation and signal detection method based on basis expansion model

Technical Field

The invention belongs to the technical field of wireless and mobile communication, and particularly relates to an OFDM system combined channel estimation and signal detection algorithm based on cyclic prefix deletion of a base extension model under a double-channel selection.

Background

In a terrestrial ultra-high-speed mobile environment, Inter-carrier Interference (ICI) caused by fast time variation of a channel and Inter-Symbol Interference (ISI) caused by multipath effect destroy the transmission characteristics of an OFDM system, resulting in rapid deterioration of system performance. Therefore, how to utilize the comb pilots to perform effective channel estimation with such challenges, and improving the interference rejection capability of the system is one of the core technologies of the future wireless communication transmission systems.

In the face of the problem of Inter-Carrier Interference (ICI) caused by fast time variation of a channel, when the normalized maximum doppler frequency shift of the channel is greater than 20%, the parameter to be estimated on each separable path of the channel is the length of one OFDM symbol, and even if the channel has only one separable path, if a single-path fast-varying channel is to be estimated, all subcarriers of the transmitting end of the OFDM system need to be pilot carriers, so we must adopt a model to equate to the fast-varying channel. The basic extension model BEM (BasisExpansion model) approaches the channel by using mutually orthogonal basis functions, so that the parameter to be estimated on each separable path of the channel is reduced from the length of one OFDM symbol to the number of the basis functions, and the estimation parameter is greatly reduced. Technical document 1(k.a.d.teo and s.ohno, Optimal MMSE fine parameter model for double-selective channels [ C ], in IEEE Global characteristics Conference,2005. Global characteristics "05", 2005,6(5): 3502-3507) adopts DKL-BEM (Discrete karhun-Loeve BEM) to approximate a time-varying channel, which uses the statistical properties of the channel to construct BEM basis and uses the main feature vector of the channel autocorrelation function as basis function vector, which has better estimation effect but needs to know the statistical properties of the channel definitely.

In the face of the Inter-Symbol Interference (ISI) problem caused by the channel multipath effect, the technical document 2(Dukhyun Kim; stub, G.L.Residul ISI cancellation for OFDM with application to HDTV broadcasting, IEEE Trans.Commun, vol.16, No.8, pp.1590-1599, Oct.1998; and Chemol-Jin Park; Gi-Hong Im; effective Cyclic Prefix reconstruction for Coded OFDM Systems, IEEE Commun.let, vol.8, No.5, pp.274-276, May.2004) converts it into an OFDM system structure without Cyclic Prefix protection. Residual inter-symbol interference cancellation (RISIC) and Cyclic Prefix Reconstruction (CPR) algorithms can effectively cancel inter-symbol interference, and achieve effective detection of OFDM symbols. But this method is implemented assuming that the channel is slowly varying.

As can be seen from the above, the time-varying channel can be effectively tracked by modeling the fast time-varying channel using the basis extension model, and ISI caused by channel multipath can be overcome by reconstructing based on effective inter-symbol interference cancellation and cyclic prefix, but when the channel exhibits double-selection characteristics, that is, when both of them occur, how to perform channel estimation and signal detection of the OFDM system is still rarely discussed in documents so far, and the present invention is developed based on this.

Disclosure of Invention

The invention aims to solve the problems of the existing technology for estimating a fast-changing channel by using a basic extension model when applied to an OFDM system with a missing cyclic prefix, and provides a method for estimating the channel and detecting the signal of the OFDM system under the conditions of double channel selection and signal prefix missing based on the basic extension model.

In order to achieve the purpose, the technical scheme of the invention is as follows:

the OFDM system channel estimation and signal detection method based on the basis of the base extension model, as shown in fig. 3 and 4, includes the following steps:

the method comprises the following steps: setting the pilot frequency and data mixture sent by the sending end of the current OFDM systemHas a frequency domain OFDM symbol of XⁱIs marked asWhere N represents the length of an OFDM symbol, frequency domain OFDM symbol XⁱSome of the subcarriers with determined positions are allocated with pilot frequency, and the others are data subcarriers; the corresponding time domain OFDM symbol x currently transmitted by the transmitting endⁱCan be described asAnd xⁱ＝F^HXⁱIn which F is^HAn IFFT transformation matrix representing N points, wherein i represents the ith OFDM symbol, namely the index number of the OFDM symbol;

step two: assume time domain OFDM symbol xⁱAfter the channel selection by the superposition of fast time-varying and multipath effects, the receiving end receives the time domain OFDM symbol of

Double-selection channel h ═ h [ h ] using complex exponential basis expansion model to describe fast time varying and multipath effect superposition₀,h₁,...,h_l,...,h_L]Wherein h is_l＝[h_0,l,h_1,l,…,h_n,l,…,h_N-1,l]Indicating the impulse response of the l path, if h is used_q(l) Representing the BEM coefficient corresponding to the channel's first path, b_qRepresenting the basis functions, then:

wherein,denotes rounding up, f_dRepresents the maximum Doppler shift, and f_d＝f_cv/c, and f_cRepresents the center frequency of the carrier wave of the OFDM system, v represents the relative movement speed between the transmitting end and the receiving end, c represents the light speed, T_sRepresents a sampling period;

time domain OFDM symbol xⁱAfter the channel selection by the superposition of fast time varying and multipath effect, the receiving end receives the time domain OFDM symbol yⁱCan be expressed as:

wherein,the receiving end receives the ith OFDM symbol when the maximum multipath time delay which represents that the length of the cyclic prefix in the OFDM system is larger than the channel,representing the previous OFDM transmission symbol x^i-1For the current OFDM received symbol yⁱThe interference of (a) with the other,a reconstructed portion representing a cyclic prefix;

transforming equation (20) to the frequency domain, then:

step three: initializing;

setting the iteration number I to be 0 and enablingThus, the device is provided with

Step four: estimating corresponding channel base expansion model coefficients;

observed quantity Y of pilot frequency position in receiving end signal Y^pCan be expressed as:

wherein, Y^pIndicating the frequency response, Ph, at the pilot location of the receiving end_eqvIndicating the frequency response of the pilot generation, D^dS^dh_eqvRepresenting the interference part, W, of the data subcarriers of the OFDM symbol on the pilot subcarriers^pRepresenting the frequency domain noise at the pilot frequency position of the receiving end;

using LS estimation criterion for h_eqvAnd estimating, namely:

considering the accuracy of the inverse operation on the matrix P, a small perturbation is often added, of which α is the perturbation factorBy substituting into equation (15),the channel time domain impulse response can be obtained

Step five: channel equalization and signal detection;

channel time domain impulse response obtained by estimationConstruction matrixDetermining a channel frequency response matrixFrequency domain equalizer for constructing I frame OFDM data unit by using MMSE criterionWherein σ²Is the variance of the noise, I_NIs an identity matrix of N × N, and then the current frequency domain symbol vector of the transmitting end is detected and obtained

Step six: judging; order:

I＝I+1 (10)

if I is less than or equal to M, M is the iteration frequency with the value range of natural numbers, executing the seventh step, otherwise, turning to the eighth step;

step seven: iteration is carried out;

channel time domain impulse response obtained by estimationConstruction matrixPlus the previous time estimateThe frequency domain symbol vector obtained by the calculationAnd a current frequency domain symbol vectorTime domain substitution equations (25) and (26) are obtainedAndupdating equation (24) after, and then going to step four;

step eight: detecting to obtain the current frequency domain symbol vector of the transmitting endCalculating a final estimate

The invention has the beneficial effects that:

the invention provides a channel estimation and signal detection method of an OFDM system under the condition of double channel selection and signal prefix deletion based on a basic extension model.

Drawings

Fig. 1 is a schematic diagram of a processing flow of a transmitting end according to the present invention.

FIG. 2 is a schematic diagram of a time domain data structure model adopted in the present invention.

FIG. 3 is a schematic diagram of a receiving end processing flow according to the present invention.

FIG. 4 is a flowchart illustrating the steps of the algorithm of the present invention.

Detailed Description

The OFDM system channel estimation and signal detection method based on the basis of the base extension model, as shown in fig. 3 and 4, includes the following steps:

the method comprises the following steps: setting a frequency domain OFDM symbol of pilot frequency and data mixture sent by a sending end of the current OFDM system as XⁱIs marked asWhere N represents the length of an OFDM symbol, frequency domain OFDM symbol XⁱSome of the subcarriers with determined positions are allocated with pilot frequency, and the others are data subcarriers; the corresponding time domain OFDM symbol x currently transmitted by the transmitting endⁱCan be described asAnd xⁱ＝F^HXⁱIn which F is^HAn IFFT transformation matrix of N points is represented, and i represents an ith OFDM symbol, i.e., an index number of the OFDM symbol.

Step two: assume time domain OFDM symbol xⁱAfter the channel selection by the superposition of fast time-varying and multipath effects, the receiving end receives the time domain OFDM symbol ofSince the time domain OFDM symbol received by the receiving end will be subjected to inter-symbol interference (ISI, shown as a gray shaded portion in fig. 2 (b)) from an adjacent OFDM symbol after the time domain OFDM symbol of the transmitting end passes through the double-channel selection, the following steps are performed:

whereinL denotes the number of paths of the multipath channel, L denotes the ith path of the channel,represents the impulse response of the ith path of the ith OFDM symbol at the nth sample time,<n-l>n denotes the remainder of (N-l) on N, x_<n-l>Representing the sampled output value at the n-th sampling instant,represents the variance of the nth sampling time as sigma²Zero-mean time-domain additive white gaussian noise.

Expression (11) is expressed in vector form, with:

whereinThe expression is as follows:

double-selection channel h ═ h [ h ] using complex exponential basis expansion model to describe fast time varying and multipath effect superposition₀,h₁,...,h_l,...,h_L]Wherein h is_l＝[h_0,l,h_1,l,…,h_n,l,…,h_N-1,l]Indicating the impulse response of the l path, if h is used_q(l) Representing the BEM coefficient corresponding to the channel's first path, b_qRepresenting the basis functions, then:

wherein,denotes rounding up, f_dRepresents the maximum Doppler shift, and f_d＝f_cv/c, and f_cRepresents the center frequency of the carrier wave of the OFDM system, v represents the relative movement speed between the transmitting end and the receiving end, c represents the light speed, T_sRepresenting the sampling period.

Now, assume that the length of the cyclic prefix in the OFDM system is greater than the maximum multipath delay of the channel, in this case, the receiving end receives the ith OFDM symbol asExpressed as:

here, theRepresenting time-domain channel matrices by time-varying channel impulse response coefficientsThe cyclic shift constitutes, which can be expressed as:

comparing equations (12) and (18) yields:

wherein,

here, ,representing the previous OFDM transmission symbol x^i-1For the current OFDM received symbol yⁱThe interference of (a) with the other,representing the reconstructed portion of the cyclic prefix.

And (3) transforming the equations (18) and (20) to the frequency domain, wherein the expressions of the components in the frequency domain are as follows:

the simultaneous equations (15) and (19) yield:

wherein H_qIs formed byA circularly symmetric matrix of the form:

step three: and (5) initializing.

Setting the iteration number I to be 0 and enablingThus, the device is provided with

Step four: the corresponding channel basis extension model coefficients are estimated.

Substituting equation (27) into equation (23) includes:

wherein, F_LRepresenting the first L columns of the FFT transformation matrix F.

For convenience of representation, the superscript of i is omitted, and the observed quantity Y of the pilot frequency position in the receiving end signal Y is^pCan be expressed as:

wherein, Y^pIndicating the frequency response, Ph, at the pilot location of the receiving end_eqvIndicating the frequency response of the pilot generation, D^dS^dh_eqvRepresenting the interference part, W, of the data subcarriers of the OFDM symbol on the pilot subcarriers^pRepresenting the frequency domain noise at the pilot frequency position of the receiving end; here:

h_eqv＝[h₀ ^T,h₁ ^T,...,h_q ^T,...h_Q ^T]^T(34)

h_q＝[h_q(0),h_q(1)...，h_q(L)]^T(35)

similarly, D can be written out^d、S^dThe index p of the pilot subcarrier position is changed into the index d of the data subcarrier position.

Using LS estimation criterion for h_eqvAnd estimating, namely:

considering the accuracy of the inversion operation on the matrix P, a small perturbation is often addedα is a disturbance factor, and the BEM coefficient obtained by estimationSubstituting into equation (15) to obtain the channel time domain impulse response

Step five: channel equalization and signal detection.

Channel time domain impulse response obtained by estimationConstruction matrixDetermining a channel frequency response matrixFrequency domain equalizer for constructing I frame OFDM data unit by using MMSE criterionWherein σ²Is the variance of the noise, I_NIs an identity matrix of N × N, and then the current frequency domain symbol vector of the transmitting end is detected and obtained

Step six: and (6) judging.

I＝I+1 (37)

If I is less than or equal to M, executing the seventh step, otherwise, turning to the eighth step

Step seven: and (6) iteration.

Channel time domain impulse response obtained by estimationConstruction matrixPlus the frequency domain symbol vector estimated at the previous timeAnd a current frequency domain symbol vectorTime domain substitution equations (25) and (26) are obtainedAndequation (24) is post-updated and then go to step four.

Step eight: detecting to obtain the current frequency domain symbol vector of the transmitting endCalculating a final estimate

Claims (1)

1. The OFDM system channel estimation and signal detection method based on the basis of the basic extension model comprises the following steps:

the method comprises the following steps: setting a frequency domain OFDM symbol of pilot frequency and data mixture sent by a sending end of the current OFDM system as XⁱIs marked asWhere N represents the length of an OFDM symbol, frequency domain OFDM symbol XⁱSome of the subcarriers with determined positions are allocated with pilot frequency, and the others are data subcarriers; the sending end is currentlyCorresponding time domain OFDM symbol x transmittedⁱCan be described asAnd xⁱ＝F^HXⁱIn which F is^HAn IFFT transformation matrix representing N points, wherein i represents the ith OFDM symbol, namely the index number of the OFDM symbol;

step two: assume time domain OFDM symbol xⁱAfter the channel selection by the superposition of fast time-varying and multipath effects, the receiving end receives the time domain OFDM symbol of

Dual-selection channel for describing fast time-varying and multipath effect superposition by utilizing complex exponential basis expansion modelWhereinIs shown asRadial impulse response, if usedRepresents the channel ofBEM coefficient corresponding to bar diameter, b_qRepresenting the basis functions, then:

$h_l=\underset{q=0}{\overset{Q}{\&Sigma;}}{b}_q{h}_q\left(l\right)---\left(1\right)$

$b_q=\&lsqb;1,e^{\frac{j2\mathrm{\&pi;}}{N}(q-Q/2)},...e^{\frac{j2\mathrm{\&pi;}(N-1)}{N}(q-Q/2)}\&rsqb;---\left(2\right)$

wherein,denotes rounding up, f_dRepresents the maximumDoppler shift, and f_d＝f_cv/c, and f_cRepresents the center frequency of the carrier wave of the OFDM system, v represents the relative movement speed between the transmitting end and the receiving end, c represents the light speed, T_sRepresents a sampling period;

time domain OFDM symbol xⁱAfter the channel selection by the superposition of fast time varying and multipath effect, the receiving end receives the time domain OFDM symbol yⁱCan be expressed as:

${\tilde{y}}^i=y^i-y_{ISI}^i+y_{Add}^i---\left(4\right)$

wherein,the receiving end receives the ith OFDM symbol when the maximum multipath time delay which represents that the length of the cyclic prefix in the OFDM system is larger than the channel,representing the previous OFDM transmission symbol x^i-1For the current OFDM received symbol yⁱThe interference of (a) with the other,a reconstructed portion representing a cyclic prefix;

transforming equation (4) to the frequency domain, then:

${\tilde{Y}}^i=Y^i-Y_{ISI}^i+Y_{Add}^i---\left(5\right)$

$Y_{ISI}^i=F\left(H_1^i\right)F^H{X}^{i-1}---\left(6\right)$

$Y_{Add}^i=F\left(H_1^i\right)F^H{X}^i---\left(7\right)$

step three: initializing;

setting the iteration number I to be 0 and enablingThus, the device is provided with

Step four: estimating corresponding channel base expansion model coefficients;

observed quantity Y of pilot frequency position in receiving end signal Y^pCan be expressed as:

$\begin{array}{c}{Y}^p={\tilde{Y}}^p\\ =D^p{S}^p{h}_{eqv}+D^d{S}^d{h}_{eqv}+W^p\\ ={\mathrm{Ph}}_{eqv}+D^d{S}^d{h}_{eqv}+W^p\end{array}---\left(8\right)$

wherein, Y^pIndicating the frequency response, Ph, at the pilot location of the receiving end_eqvIndicating the frequency response of the pilot generation, D^dS^dh_eqvRepresenting the interference part, W, of the data subcarriers of the OFDM symbol on the pilot subcarriers^pRepresenting the frequency domain noise at the pilot frequency position of the receiving end;

using LS estimation criterion for h_eqvAnd estimating, namely:

${\hat{h}}_{eqv}={(P^{H}P+\mathrm{\&alpha;}I)}^{-1}{P}^H{Y}^p---\left(9\right)$

considering the accuracy of the inverse operation of the matrix P, adding a small disturbance, wherein α is a disturbance factor, and estimating the obtained BEM coefficientSubstituting into equation (1) to obtain the channel time domain impulse response

Step five: channel equalization and signal detection;

channel time domain impulse response obtained by estimationConstruction matrixDetermining a channel frequency response matrixFrequency domain equalizer for constructing I frame OFDM data unit by using MMSE criterionWherein σ²Is the variance of the noise, I_NIs an identity matrix of N × N, and then the current frequency domain symbol vector of the transmitting end is detected and obtained

Step six: judging; order:

I＝I+1 (10)

if I is less than or equal to M, M is the iteration frequency with the value range of natural numbers, executing the seventh step, otherwise, turning to the eighth step;

step seven: iteration is carried out;

channel time domain impulse response obtained by estimationConstruction matrixPlus the frequency domain symbol vector estimated at the previous timeAnd a current frequency domain symbol vectorTime domain substitution equations (6) and (7) are obtainedAndupdating equation (5) and then going to step four;

step eight: detecting to obtain the current frequency domain symbol vector of the transmitting endCalculating a final estimate