CN103825850A - Upstream channel estimation method and upstream channel estimation system suitable for LTE (Long Term Evolution)-Advanced system - Google Patents

Abstract

The invention relates to an upstream channel estimation method and an upstream channel estimation system suitable for an LTE (Long Term Evolution)-Advanced system. The upstream channel estimation method comprises the steps of extracting pilot frequency subcarrier information on a receiving antenna, and carrying out conjugate multiplying on the pilot frequency subcarrier information and a local pilot frequency of a first layer to obtain a frequency domain channel coefficient of all transport layers subjected to aliasing; according to the length of a subcarrier, adding a virtual subcarrier with a corresponding length at the tail of a primary channel coefficient; carrying out IDFT (Inverse Discrete Fourier Transform) calculation, and transforming to a time domain; adding a window in the time domain, because dislocation of a time domain impulse response of each transport layer occurs on the time domain, separating time domain impulse responses of different transport layers, and carrying out time domain cyclic shift to obtain a time domain impulse response corresponding to each transport layer; and carrying out DFT (Discrete Fourier Transformation) calculation on the time domain impulse response of each transport layer, transforming to the frequency domain, and removing the virtual subcarrier at the tail to obtain a frequency domain equivalent channel estimation matrix of each transport layer corresponding to the receiving antenna. With the adoption of the upstream channel estimation method and the upstream channel estimation system, the channel estimation problem after MIMO (Multiple Input Multiple Output) is introduced into the LTE-A upstream can be effectively solved, and the channel estimation accuracy is also improved; and the method is also suitable for an LTE system.

Description

A kind of uplink channel estimation method of applicable LTE-Advanced system and system

Technical field

The present invention relates to wireless communication system, more specifically, particularly for LTE-Advanced system uplink channel method of estimation and system.

Background technology

Advanced Long Term Evolution (Long Term Evolution-Advanced, LTE-Advanced) be the smooth evolution on LTE basis, the 3rd generation partner program (3rd Generation Partnership Project, 3GPP) in order to meet (the International Telecommunication Union of International Telecommunications Union, ITU) requirement of the IMT-Advanced proposing and 4G (the4th Generation) standard that proposes, it further promotes the performance of system on the basis of assurance and LTE backward compatibility, finally meet or exceed the requirement of IMT-Advanced.

LTE-A system has proposed descending peak rate and has exceeded the performance requirement that 1Gbit/s, up peak rate exceed 500Mbit/s.And in order to reach this requirement, LTE-A has introduced again the more advanced communication technology with respect to LTE: such as further expanding of carrier aggregation, cooperative multipoint transmission, relaying technique, MIMO (Multiple Input Multiple Output) technology etc.

The performance of wireless communication system is subject to the impact of wireless channel to a great extent, as shadow fading and frequency selective fading etc., makes the transmission path between transmitter and receiver very complicated.In order to recover exactly at receiving terminal the transmitted signal of transmitting terminal, need to carry out channel estimating accurately to wireless channel.Can obtain comparatively in detail and channel information accurately, thereby correctly demodulate and transmit at receiving terminal, be the important indicator of weighing a performance in wireless communication systems.

Before introducing MIMO, the channel estimation methods of up-downgoing is basic identical, and corresponding algorithm has LS, MMSE etc.After descending introducing MIMO, the pilot signal of different transmit antennas port is put and is staggered and diverse location, the various channel estimation methods therefore still can continue to adopt single antenna in the time of channel estimating time; And in the situation of Uplink MIMO, the pilot positions of different transmit antennas port is identical, take identical time-frequency domain resources, and be decided by the characteristic of uplink pre-coding matrix, reference signal in different transport layers is multiplied by after pre-coding matrix, and the pilot signal that different transmit antennas is sent also may be identical.If the various channel estimation methods while now re-using single antenna, can introduce the interference in another root antenna similar frequency bands, because what detect is two transmit antennas signal sums while receiving, and the pilot frequency sequence of two antennas is in time domain or frequency domain all mixes.Visible, traditional channel estimation in frequency domain technology is not suitable for the channel estimating of Uplink MIMO.

Summary of the invention

In order to solve traditional channel estimation methods to LTE-Advanced system limitation problem, the present invention proposes uplink channel estimation method and the system of a kind of LTE-Advanced of being applicable to system precoding MIMO, can effectively solve the channel estimation problems after the up introducing of LTE-A MIMO, improve the accuracy of channel estimating, the method is equally also applicable to LTE system simultaneously.

Utilize the pilot signal of reception antenna to be multiplied by the conjugation of the local pilot signal of the 1st transport layer, the local pilot signal producing does not need to carry out pre-encode operation, obtain the channel estimation coefficient of each transport layer aliasing on frequency domain, and separate different transport layers by the difference of different layers cyclic shift in time domain, obtain the equivalent channel estimation coefficient of transport layer and reception antenna.While utilizing this equivalence channel coefficients to carry out the recovery of upstream data information, without separating pre-encode operation.The method is utilized the feature of LTE-Advanced uplink reference signals, estimates the equivalent channel coefficient of transport layer and reception antenna, rather than the channel coefficients of traditional transmitting antenna and reception antenna.

A uplink channel estimation method for applicable LTE-Advanced system, comprises the following steps:

Step 1, obtains reception antenna pilot signal sequence Y.

Step 2, produces the local pilot frequency sequence X of the 1st layer ₁, X ₁it is the demodulated reference signal after cyclic shift 0≤n < M _sc, M _scfor demodulated reference signal length,

be basic sequence, the cyclic shift of distributing to υ layer transport layer is respectively α ₁, α ₂... α _υ,

wherein n _{cs, λ}(λ=1 ..., υ) be the cyclic shift value on λ layer, distributed by high level; The local reference signal X producing ₁do not need to carry out the pre-encode operation of transmitting terminal; The pilot frequency sequence receiving is multiplied by the complex conjugate of distributing to the 1st layer of reference signal, obtains the equivalent channel coefficient H=conj (X of all transport layers and reception antenna ₁) Y, rather than the channel matrix of traditional transmitting antenna and reception antenna (the transmission number of plies≤number of transmit antennas); H is the frequency domain channel coefficient that has mixed each transport layer.

Step 3, according to the subcarrier length of step 2 gained frequency domain channel coefficient, obtains corresponding virtual subnet carrier number L

for the subcarrier number on each RB, for the size of the RB of this sub-distribution,

the size of RB of distributing for the next one), and carry out the expansion of virtual subnet carrier channel coefficient, the channel coefficients length after expansion is N=N _sc+ L; Frequency domain channel coefficient after expansion is carried out to Inverse Discrete Fourier Transform to time domain, obtain the channel impulse response (channel coefficients) of each transport layer aliasing in time domain;

Preferential, described virtual subnet carrier wave is added on the afterbody of former frequency domain channel coefficient, for keeping the level and smooth of whole waveform, adds a cosine wave, H _vCH(n)=a × F (n)+b × G (n), wherein $F\left(n\right)=\frac{1}{2}\&times;(1+\cos (\mathrm{\&pi;}\&times;\frac{n}{L})),G\left(n\right)=\frac{1}{2}\&times;(1+\cos (\mathrm{\&pi;}\&times;\frac{-L+1+n}{L})),$ N ∈ [0, L-1], a and b are the degree of amplitude modulation factor;

Carry out Inverse Discrete Fourier Transform to time domain to having added the frequency domain channel coefficient of virtual subnet carrier wave, although the demodulated reference signal of different transport layers is all superimposed on time-frequency domain, but after discrete inversefouriertransform, the time domain impulse response of different transport layers can be by the n of different layers _{cs, λ}difference distinguish.Step 4, then time-domain windowed, filter noise, separates different transport layers, obtains the time domain channel coefficient of different transport layers;

According to the difference of each transport layer cyclic shift, calculate the position of each transport layer impulse response in time domain, adopt the mode of time-domain windowed filtering, dynamically retain time domain impulse response length corresponding to this transport layer and obtain the time domain impulse response of each transport layer; Due to each transport layer aliasing in time domain, when design windowed function, windowing length depends on noise level, channel latency spectrum and the RB number distributing;

This step is carried out windowing operation, is the time domain impulse response in order to separate different transport layers on the one hand, is in order to suppress noise in obtaining every layer signal energy on the other hand; After Inverse Discrete Fourier Transform, the n of each transport layer _{cs, λ}difference while having determined to carry out channel estimating, the separation degree of the time domain impulse response of different layers; Time-domain windowed length is crossed conference and is caused transport layer separation not thorough, and the too small meeting of time-domain windowed length causes the loss of signal energy, here according to the RB number of noise level, channel latency spectrum and distribution, dynamically retains the length of the time domain impulse response of λ transport layer;

Preferably, channel without time inclined to one side perfect condition under, the central point d of each transport layer time domain impulse response _λdepend on each layer of n _{cs, λ}difference, Δ n _{cs, λ}=n _{cs, 1}-n _{cs, λ}, N carries out counting of Inverse Discrete Fourier Transform;

With d _λas central point, dynamically retain the time domain impulse response length of λ transport layer in time domain, this length comprises left and right two parts:

Right half part:

Boundary retains β CP sampled point, i.e. d on the right _λ+ β CP-1

Wherein, CP is SC-FDMA title cyclic shift length, and β is an adjustable parameter, 0≤β < 1;

Left-half:

On the left side circle retains

individual sampled point, $(d_{\mathrm{\&lambda;}}-\mathrm{\&mu;}\frac{N}{M_{\mathrm{sc}}}+N)\mathrm{mod}\left(N\right)$

Retain

individual sampled point is to consider that actual allocated only accounts for a part for total bandwidth to user's subcarrier, and exists protection bandwidth, and some subcarrier is not used in transfer of data, therefore, in the time being transformed into time domain, existing time domain energy and reveals.

Parameter beta and μ, according to RB Resource Block (Resource Block) number of signal to noise ratio, channel latency spectrum and distribution, dynamically adjust.

Step 5, removes the time-domain cyclic shift of each transport layer, obtains the time domain impulse response of each transport layer;

Particularly, remove the time domain impulse response of λ transport layer after cyclic shift

for:

${\hat{h}}_{\mathrm{\&lambda;}}(1:N-d_{\mathrm{\&lambda;}})=h_{\mathrm{\&lambda;}}(d_{\mathrm{\&lambda;}}+1:N)$

${\hat{h}}_{\mathrm{\&lambda;}}(N-d_{\mathrm{\&lambda;}}+1:N)=h_{\mathrm{\&lambda;}}(1:d_{\mathrm{\&lambda;}})$

Wherein, h _λfor removing the time domain channel coefficient of λ transport layer before cyclic shift.

Step 6, is fourier transformed into frequency domain by the time domain channel coefficient of each transport layer of step 5 gained, and removes the virtual subnet carrier wave length that step 3 afterbody adds, and obtains all transport layers channel estimate matrix corresponding with this reception antenna.

The present invention is the corresponding uplink channel estimation system that a kind of applicable LTE-Advanced system is provided also, comprises and extracts pilot sub-carrier module, ZC compensating module, channel coefficients expansion module, time-domain windowed module on reception antenna, removes circular shift module and remove virtual subnet carrier module;

Extract pilot sub-carrier module on reception antenna, for extracting the information on the ascending pilot frequency subcarrier on certain root reception antenna, and input ZC compensating module;

ZC compensating module, for generation of the local pilot frequency sequence of the 1st layer, carries out conjugate multiplication with receiving terminal pilot frequency sequence, obtains the channel coefficients of each transport layer aliasing on frequency domain, and input channel coefficient expansion module;

Channel coefficients expansion module, for according to the subcarrier length of frequency domain channel coefficient, obtains corresponding virtual subnet carrier number, and expansion virtual subnet carrier channel coefficient, carries out Inverse Discrete Fourier Transform to time domain, and be input to time-domain windowed module;

Time-domain windowed module, for extracting the time domain impulse response of corresponding transport layer, dynamically retains this transport layer time domain impulse response length, completes the separation of each transport layer, and is input to and removes circular shift module;

Remove circular shift module, remove the time-domain cyclic shift of each transport layer, obtain the time domain impulse response of each transport layer, and be input to removal virtual subnet carrier module;

Remove virtual subnet carrier module, for all transport layer time domain channel coefficients are carried out to Fourier transform, remove the virtual subnet carrier number that afterbody adds, obtain the equivalent channel estimation in frequency domain matrix of transport layer and reception antenna.

From above technical scheme provided by the invention, after receiving the uplink reference signals of certain root reception antenna, carry out conjugate multiplication with the 1st layer of reference signal after cyclic shift that this locality produces, obtain the channel coefficients of each transport layer aliasing on frequency domain, according to the length of this frequency domain channel coefficient, calculate virtual subnet carrier number, and the method for adding cosine wave by afterbody is expanded the frequency domain band channel of making an uproar, channel coefficients after being expanded, then obtains time domain channel coefficient through IDFT.By carry out windowing according to each transport layer time domain impulse response center position in time domain, and remove cyclic shift, take out the time domain impulse response of corresponding transport layer, transform to frequency domain and remove the virtual subnet carrier wave that afterbody adds finally by crossing DFT, obtain all transport layer frequency domain equivalent channel estimated matrix.

Compared with prior art, application technical solution of the present invention is estimated channel coefficients, can effectively solve the problem of many antennas aliasing after the up introducing of LTE-Advanced MIMO.Meanwhile, in time domain, carry out dynamic windowing according to the RB number of signal to noise ratio, channel time delay spectrum and distribution, windowing mode has neatly guaranteed channel effective diameter ground selecting properly, has improved the accuracy of channel estimating; The method that obtains pseudo channel coefficient in technical solution of the present invention is very simple, therefore can improve hardware handles efficiency; Technical solution of the present invention is not only applicable to LTE-Advanced system, is applicable in LTE system simultaneously yet.

Accompanying drawing explanation

Fig. 1 is the flow chart of method in the embodiment of the present invention.

Fig. 2 is the structure chart of system in the embodiment of the present invention.

Embodiment

In following several embodiments of the present invention, an embodiment provides a kind of uplink channel estimation method of applicable LTE-Advanced system, and an embodiment provides a kind of uplink channel estimation system of applicable LTE-Advanced system.This method embodiment relates generally to generation, the interpolation of virtual subnet carrier wave, the time-domain windowed of local reference signal when channel coefficients is estimated.

In order to make those skilled in the art understand better the technical scheme in the embodiment of the present invention, and make the above-mentioned purpose of the embodiment of the present invention, feature and advantage can be more obvious and understandable, below in conjunction with accompanying drawing, technical scheme in the embodiment of the present invention is described in further detail.

Referring to Fig. 1, embodiment is configured to the system bandwidth of up 20MHz, and the resource of distributing to alone family is i.e. 50 Resource Block of 50RB(), the antenna configuration of Uplink MIMO is 22 to be received, and code word number is 2, and transport layer is 2 layers, and detailed channel estimation steps is as follows:

Step 1, obtains second reception antenna processing procedure of first reception antenna pilot signal sequence Y(identical);

The present embodiment, receiving after the upstream data of antenna 1, obtains pilot signal according to demapping, as uplink reference signals Y, Y=HWX+N, H is the channel response matrix that transmitting antenna is corresponding with reception antenna, and W is the pre-coding matrix of Uplink MIMO, and N is that average is 0, variance is σ ²additive white Gaussian noise, be the cubic metric that guarantees LTE-Advanced up channel, the every a line of pre-coding matrix W has and only has a non-zero element, in this embodiment, $W=\frac{1}{2}\left[\begin{array}{cc}1& 0\\ 0& 1\end{array}\right],$ X represents the reference signal producing in transport layer, and N represents that average is 0, variance is σ ²white Gaussian noise.

Step 2, produces the local pilot frequency sequence X of the 1st layer ₁, X ₁it is the demodulated reference signal after cyclic shift

0≤n < M _sc, M _scfor demodulated reference signal length,

be basic sequence, the cyclic shift of distributing to 2 layers of transport layer is respectively α ₁, α ₂, wherein n _{cs, λ}(λ=1,2) are distributed by high level.The local reference signal X producing ₁do not need to carry out the pre-encode operation of transmitting terminal.

Be multiplied by local pilot frequency sequence X with the pilot frequency sequence Y receiving ₁conjugation, obtain frequency domain channel coefficient and be $\hat{H}=\mathrm{conj}\left(X_1\right)\&CenterDot;Y=H_1+e^{j2\mathrm{\&pi;n}\frac{n_{\mathrm{cs},2}-n_{\mathrm{cs},1}}{12}}{H}_2+N^{\&prime;},$

the frequency domain channel coefficient that has mixed two transport layers, H ₁and H ₂it is respectively the frequency domain channel coefficient of layers 1 and 2.

In the present embodiment, M _sc=50 × 12=600.According to 3GPP36.211 agreement, uplink reference signals is one group of ZC sequence (sequence that auto-correlation is strong), therefore just can obtain with the local reference signal conjugate multiplication producing the frequency domain channel coefficient that band is made an uproar by the reference signal receiving.

Step 3, according to the length of the subcarrier of step 2 gained frequency domain channel coefficient, obtains corresponding virtual subnet carrier number, is designated as L, and the cosine wave that generation length is L, as virtual subnet carrier wave, is added on former frequency domain channel coefficient

afterbody, expansion after frequency domain channel coefficient be respectively H'.Carry out Inverse Discrete Fourier Transform to time domain to having added the frequency domain channel coefficient of virtual subnet carrier wave.

Distribute RB size need to meet certain requirement according to the ascending resource of LTE agreement regulation, i.e. the size of the RB of this sub-distribution meets

α, β, γ are non-negative integer.Under the system broad of corresponding 20MHz, the RB allocation set that meets described condition has amounted to 34 elements, only has 34 kinds of up RB resources to distribute.For this reason, at the subcarrier in frequency domain number having added after virtual subnet carrier wave, need to meet the subcarrier length of next RB size.The virtual subnet carrier number of supposing required interpolation is L, has according to described method

for the subcarrier number on each RB,

for the size of the RB of this sub-distribution,

the size of RB of distributing for the next one.The virtual subnet carrier wave number of adding in the present embodiment is L=4 × 12=48, and the spectral channel coefficient length after expansion is N=600+48=648.

In the present embodiment, produce a cosine wave and be added on as virtual subnet carrier wave the afterbody of former channel coefficients, the cosine wave of interpolation is H _{vCH, λ}(n)=a × F (n)+b × G (n), wherein

$G\left(n\right)=\frac{1}{2}\&times;(1+\cos (\mathrm{\&pi;}\&times;\frac{-L+1+n}{L}))$ N ∈ [0, L-1], a and b are the degree of amplitude modulation factor.

H' is carried out obtaining after discrete inversefouriertransform:

n (i) is noise in time domain, wherein i=0, and 1 ..., N, h ' ₁and h ' ₂be respectively the time domain impulse response of layers 1 and 2 after expansion.Can see, the impulse response of two-layer pilot signal has dislocation in time domain, by two layer n _{cs, λ}difference, the time domain impulse response of different layers can be taken out.

Step 4, in time-domain windowed filtering, takes out the time domain impulse response of two transport layers.

Channel without time inclined to one side perfect condition under, the central point d of each transport layer time domain impulse response _λdepend on each layer of n _{cs, λ}difference, $d_{\mathrm{\&lambda;}}=\left(\frac{{\mathrm{\&Delta;n}}_{\mathrm{cs},\mathrm{\&lambda;}}\&times;N}{12}\right)\mathrm{mod}\left(N\right),{\mathrm{\&Delta;n}}_{\mathrm{cs},\mathrm{\&lambda;}}=n_{\mathrm{cs},1}-n_{\mathrm{cs},\mathrm{\&lambda;},}$ N carries out counting of Inverse Discrete Fourier Transform.Thereby $d_1=0,d_2=\left(\frac{(n_{\mathrm{cs},1}-n_{\mathrm{cs},2})\&times;N}{12}\right)\mathrm{mod}\left(N\right).$

With d _λas central point, dynamically retain the time domain impulse response length of λ transport layer in time domain, this length comprises left and right two parts:

Right half part:

Boundary retains β CP sampled point, i.e. d on the right _λ+ β CP-1

Wherein, CP is SC-FDMA title cyclic shift length, and β is an adjustable parameter, 0≤β < 1;

Left-half:

On the left side circle retains

individual sampled point, $(d_{\mathrm{\&lambda;}}-\mathrm{\&mu;}\frac{N}{M_{\mathrm{sc}}}+N)\mathrm{mod}\left(N\right)$

Retain

individual sampled point is to consider that actual allocated only accounts for a part for total bandwidth to user's subcarrier, and exists protection bandwidth, and some subcarrier is not used in transfer of data, therefore, in the time being transformed into time domain, existing time domain energy and reveals.Parameter beta and μ, according to the RB number of signal to noise ratio, channel latency spectrum and distribution, dynamically adjust.

Step 5, removes the time-domain cyclic shift of each transport layer, obtains the time domain impulse response of each transport layer;

Particularly, remove after cyclic shift the time domain impulse response of the 1st layer for:

Remove after cyclic shift the time domain impulse response of the 2nd layer

for:

${\hat{h}}_2(1:N-d_2)=h_2^{\&prime;}(d_2+1:N)$

${\hat{h}}_2(N-d_2+1:N)=h_2^{\&prime;}(1:d_2)$

Wherein, h ' ₁and h ' ₂be respectively the time domain channel coefficient that removes the front layers 1 and 2 of cyclic shift.

Step 6, carries out Fourier transform to the time domain channel coefficient of obtained each transport layer, and removes the virtual subnet carrier wave length that afterbody adds, and obtains all transport layers equivalent channel estimation in frequency domain matrix corresponding with this reception antenna.

In this embodiment, through DFT, conversion obtains layers 1 and 2 frequency domain channel coefficient

with

remove L the virtual subnet carrier wave that afterbody adds, obtain the equivalent channel coefficient of layers 1 and 2 with respect to the 1st reception antenna $H_1={\hat{H}}_1(1:M_{\mathrm{sc}})$ With $H_2={\hat{H}}_2(1:M_{\mathrm{sc}}).$

Corresponding with the embodiment of the uplink channel estimation method of a kind of applicable LTE-Advanced system of the present invention, the present invention also provides a kind of embodiment of uplink channel estimation system of applicable LTE-Advanced system.

Referring to Fig. 2, be the block diagram of the uplink channel estimation system of a kind of applicable LTE-Advanced system in embodiment bis-, comprising:

Extract pilot sub-carrier module M1 on reception antenna, for extracting the information on the ascending pilot frequency subcarrier on certain root reception antenna, and input ZC compensating module M2;

ZC compensating module M2, for generation of the local pilot frequency sequence of the 1st layer, carries out conjugate multiplication with receiving terminal pilot frequency sequence, obtains the channel coefficients of each transport layer aliasing on frequency domain, and input channel coefficient expansion module M3;

Channel coefficients expansion module M3, for according to the subcarrier length of frequency domain channel coefficient, obtains corresponding virtual subnet carrier number, and expansion virtual subnet carrier channel coefficient, carries out Inverse Discrete Fourier Transform to time domain, and be input to time-domain windowed module M4;

Time-domain windowed module M4, for extracting the time domain impulse response of corresponding transport layer, dynamically retains this transport layer time domain impulse response length, completes the separation of each transport layer, and is input to and removes circular shift module M5;

Remove circular shift module M5, remove the time-domain cyclic shift of each transport layer, obtain the time domain impulse response of each transport layer, and be input to removal virtual subnet carrier module M6;

Remove virtual subnet carrier module M6, for all transport layer time domain channel coefficients are carried out to Fourier transform, remove the virtual subnet carrier number that afterbody adds, obtain the equivalent channel estimation in frequency domain matrix of transport layer and reception antenna

Extract pilot sub-carrier module M1, ZC compensating module M2, channel coefficients expansion module M3, time-domain windowed module M4 on reception antenna, remove circular shift module M5 and remove circular shift module M5 and be connected successively.

Each module can adopt software module piece method for designing to realize, and specific works is referring to the each step of method.

Those skilled in the art can very clearly recognize that the technology in the embodiment of the present invention can add essential general hardware platform by software and realize.Protection scope of the present invention is not limited to this example for this reason, but is encompassed in the protection range being limited by claims.

Claims (4)

1. a uplink channel estimation method for applicable LTE-Advanced system, is characterized in that: comprise the following steps:

Step 1, extracts the pilot sub-carrier information on reception antenna;

Step 2, produces the local pilot frequency sequence in the 1st transport layer, and the receiving terminal pilot frequency sequence obtaining with step 1 carries out conjugate multiplication, obtains the channel coefficients of each transport layer aliasing on frequency domain;

Step 3, according to the subcarrier length N of step 2 gained frequency domain channel coefficient _sc, obtaining corresponding virtual subnet carrier number L, generating virtual subcarrier is also added on the afterbody of former frequency domain channel coefficient, and the channel coefficients length after expansion is N=N _sc+ L; Frequency domain channel coefficient after expansion is carried out to Inverse Discrete Fourier Transform to time domain, obtain the channel impulse response of each transport layer aliasing in time domain;

Step 4, according to the difference of each transport layer cyclic shift, calculate the position of each transport layer impulse response in time domain, adopt the mode of time-domain windowed filtering, by time domain impulse response length corresponding to this transport layer of dynamic reservation, thereby obtain the time domain impulse response of each transport layer; Due to each transport layer aliasing in time domain, when design window function, windowing length depends on noise level, channel latency spectrum and the RB number distributing;

Step 5, after time-domain filtering, removes the time-domain cyclic shift of each transport layer, obtains the time domain impulse response of each transport layer;

Step 6, is fourier transformed into frequency domain by the time domain channel coefficient of each transport layer of step 5 gained, and removes the virtual subnet carrier wave length that step 3 afterbody adds, and obtains all transport layers equivalent channel estimated matrix corresponding with this reception antenna.

2. the uplink channel estimation method of a kind of applicable LTE-Advanced system according to claim 1, is characterized in that: in described step 2, produce the local pilot frequency sequence X in the 1st transport layer ₁, be the demodulated reference signal after cyclic shift

0≤n < M _sc, M _scfor demodulated reference signal length,

be basic sequence, the cyclic shift of distributing to υ layer transport layer is respectively α ₁, α ₂... α _λ... α _υ,

wherein n _{cs, λ}(λ=1 ..., υ) be the cyclic shift value on λ layer, distributed by high level; The pilot frequency sequence receiving is multiplied by the complex conjugate of distributing to the 1st layer of reference signal, obtains the equivalent channel coefficient of all transport layers and reception antenna.

3. the uplink channel estimation method of a kind of applicable LTE-Advanced system according to claim 1, it is characterized in that: in described step 4, according to the RB number of noise level, channel latency spectrum and distribution, dynamically retain the length of the time domain impulse response of λ transport layer; With the time domain impulse response position d of λ layer _λas central point, dynamically retain the time domain impulse response length of λ transport layer in time domain, this length comprises left and right two parts:

Right half part:

Boundary retains β CP sampled point, i.e. d on the right _λ+ β CP-1

Wherein, CP is SC-FDMA title cyclic shift length, and β is an adjustable parameter, 0≤β < 1;

Left-half:

On the left side circle retains

individual sampled point, $(d_{\mathrm{\&lambda;}}-\mathrm{\&mu;}\frac{N}{M_{\mathrm{sc}}}+N)\mathrm{mod}\left(N\right);$ N carries out counting of Inverse Discrete Fourier Transform.

4. a uplink channel estimation system for the applicable LTE-Advanced system based on said method design, is characterized in that: comprise and extract pilot sub-carrier module, ZC compensating module, channel coefficients expansion module, time-domain windowed module on reception antenna, remove circular shift module and remove virtual subnet carrier module;

Extract pilot sub-carrier module on reception antenna, for extracting the information on the ascending pilot frequency subcarrier on certain root reception antenna, and input ZC compensating module;

ZC compensating module, for generation of the local pilot frequency sequence of the 1st layer, carries out conjugate multiplication with receiving terminal pilot frequency sequence, obtains the channel coefficients of each transport layer aliasing on frequency domain, and input channel coefficient expansion module;

Channel coefficients expansion module, for according to the subcarrier length of frequency domain channel coefficient, obtains corresponding virtual subnet carrier number, and expansion virtual subnet carrier channel coefficient, carries out Inverse Discrete Fourier Transform to time domain, and be input to time-domain windowed module;

Time-domain windowed module, for extracting the time domain impulse response of corresponding transport layer, dynamically retains this transport layer time domain impulse response length, completes the separation of each transport layer, and is input to and removes circular shift module;

Remove circular shift module, remove the time-domain cyclic shift of each transport layer, obtain the time domain impulse response of each transport layer, and be input to removal virtual subnet carrier module;

Remove virtual subnet carrier module, for all transport layer time domain channel coefficients are carried out to Fourier transform, remove the virtual subnet carrier number that afterbody adds, obtain the equivalent channel estimation in frequency domain matrix of transport layer and reception antenna.

Images