CN101359926A - Receiver, channel estimation method and apparatus - Google Patents

Abstract

The invention provides a receiver, a channel estimation method and a device, wherein, the channel estimation method comprises: multi-path search is carried out to a pilot channel and a path the energy of which is larger than precondition in a channel window is selected; channel estimation is carried out to the path; raised cosine interference of the channel estimation value obtained by the channel estimation is eliminated, and the channel estimation value without raised cosine interference is obtained; and raised cosine convolution is carried out to the channel estimation value without raised cosine interference and path energy of all locations in the pilot channel is obtained. When the invention is used for channel estimation, new noise can not be introduced, and the balanced performance of an LMMSE equalizer is effectively ensured; in addition, interference among introduced paths to occur with windage between the channel estimation value and actual channel value is avoided when RC convolution is carried out because RC influence exists, and the balanced effect of the LMMSE equalizer is further ensured.

Description

Receiver, channel estimation methods and device

Technical field

The present invention relates to the communication technology, especially a kind of channel estimating apparatus of eliminating the receiver that disturbs between the footpath, carrying out being used to carry out in the channel estimation methods of channel estimating and this receiver channel estimating with this receiver.

Background technology

Wideband Code Division Multiple Access (WCDMA) (Wideband Code division multiple access, hereinafter to be referred as: WCDMA) in the system, the many coded signals at the transmitter place are mutually orthogonals.Yet when these signals passed through the multipath fading dissemination channel, this orthogonality weakened or disappears, and for correct identification signal, introduced the orthogonality that RAKE receiver is come restoring signal at receiving terminal usually.Owing to adopted wideband spread-spectrum signal, the WCDMA system can tell different multipath components in the received signal with RAKE receiver, realizes diversity reception, overcomes the influence of multipath fading effectively, has improved the receptivity of system.But, in the descending reception of WCDMA system, because intersymbol interference and multi-user interference that multipath causes, limited the coherent reception performance of RAKE receiver, particularly under the high data rate situation, for example: high speed downlink packet inserts (High Speed Downlink Packet Access, hereinafter to be referred as: in the time of HSDPA), because spreading factor is less, not only the antijamming capability of RAKE receiver reduces, and merges the also reduction of introducing greatly of diversity gain by multipath.At this moment, in order to obtain comparatively ideal performance, prior art is introduced linear minimum mean-squared error (Linear Minimum Mean Square Error, hereinafter to be referred as: LMMSE) technology has realized the LMMSE receiver, in the LMMSE receiver, adopt the LMMSE equalizer to carry out equilibrium to received signal, the homeostatic principle of LMMSE equalizer is the mean square error minimum between equalizing signal and the transmission signal, owing to eliminated intersymbol interference and multi-user interference that multipath causes, the receptivity of LMMSE equalizer is better than original RAKE receiver.

As shown in Figure 1, be the structural representation of prior art WCDMA signal receiving and transmitting system.At transmitting terminal, the transmitting terminal root raised cosine (Root Raised Cosine, hereinafter to be referred as: RRC) filter is launched after to signal x (k) filtering; Transmit through being superimposed with the multidiameter fading channel transmission of white Gaussian noise, arrive the LMMSE receiver of receiving terminal afterwards, obtained signal y (k) behind the receiving terminal RRC filter filtering in the LMMSE receiver, this signal y (k) is sent to LMMSE equalizer and channel estimating apparatus respectively; Channel estimating apparatus carries out channel estimating based on signal y (k), estimate the energy in each footpath in the fading channel, and will estimate that the channel estimation value h channel estimate matrix that obtains sends to calculation element, when a plurality of footpath, this channel estimation value h shows as a channel estimate matrix H, simultaneously, noise function estimation unit estimates the white Gaussian noise energy in the channel, and with the noise energy σ that estimates _n ²Send to calculation element; Calculation element is according to channel estimate matrix estimated value h and white Gaussian noise energy σ _n ²Calculate, obtain the tap weight value w of LMMSE equalizer _d, and send to the LMMSE equalizer; The LMMSE equalizer is according to homeostatic principle, to y (k) and w _dCarry out equilibrium, obtain signal x ' (k); The descrambling and de-spreading device (k) carries out descrambling, despreading to this signal x ', obtains to send signal x (k).

By LMMSE receiver processing procedure to received signal among Fig. 1 as can be known, the tap weight value w of LMMSE equalizer _dBy channel estimation value h and white Gaussian noise energy σ _n ²Decision, the difference of channel estimation value h and actual channel value has directly influenced the performance of LMMSE equalizer, and channel estimating is approached actual channel conditions more, and the portfolio effect of LMMSE equalizer is good more.Therefore, the channel estimation value that how to carry out channel estimating, obtains to approach the actual channel value is very crucial.

The conventional channel method of estimation is directly carried out channel fading to pilot channel at path position according to the Multipath searching indication and is estimated.As shown in Figure 2, structural representation for the conventional channel estimation unit, its flow process of carrying out channel estimating is as follows: the Multipath searching module detects when the footpath is arranged in the pilot channel, to Common Pilot Channel (Common Pilot Channel, hereinafter to be referred as: CPICH) the descrambling and de-spreading module sends the Multipath searching indication information, this Multipath searching indication information comprises the initial path position and the channel length information of the pilot channel that need carry out channel estimating, because the position in initial footpath is exactly the original position of channel, just can know the channel window that need carry out channel estimating in conjunction with channel length, thus, the final position of channel window also can be known; CPICH descrambling and de-spreading module each path position in this channel window is successively carried out descrambling, despreading, obtains frequency pilot sign; The frequency pilot sign that the division module obtains descrambling, despreading is divided by the frequency pilot sign that sends, and obtains channel estimation value and send to limited long impulse response that (Finite Impulse Response is hereinafter to be referred as FIR) filter; The FIR filter carries out the front and back smoothing processing to the channel estimation value that receives; Owing to pilot signal has been taken advantage of a power weightings factor at transmitting terminal, current channel estimation value and actual power weightings factor of channel estimating value difference, the signal of power weightings factor processing module after to the FIR filter process obtains the channel estimation value CE of each path position in the channel window divided by this power weightings factor.

When adopting above-mentioned conventional channel method of estimation to carry out channel estimating, considered the footpath energy of all positions in the channel window, since the footpath energy of some position a little less than, and energy is all lower usually than the signal to noise ratio of weak position, therefore, introduce bigger noise when the conventional channel method of estimation is carried out channel estimating, thereby influenced the equalization performance of LMMSE equalizer, made its equalization performance descend.

In order to overcome in the conventional channel method of estimation owing to introduce the LMMSE equalizer performance decline that big noise causes, prior art is a precision with 1/4 chip, in the detected channel window of Multipath searching module, select signal to noise ratio to reach the single footpath of predetermined threshold value, adopt the conventional channel method of estimation that channel estimating is carried out in this single footpath then; Owing to the equivalent channel that needs to estimate is actual channel and the convolution that receives RRC filter, transmission RRC filter, and two RRC filter strings are equivalent to raised cosine filter together, represent this raised cosine filter parameter matrix with RC, the channel estimation value and the RC that estimate are carried out convolution, obtain the channel estimating of all phase points in the channel window, the precision of this phase point also is 1/4 chip.Consideration for the complexity aspect, tap number in chip of equalizer is generally 2 or 1, and needing chip-level is 1 or 1/2 channel estimating, therefore, carry out down-sampling at channel estimation value, obtain the channel estimation value of chip-level or 1/2 chip-level 1/4 chip precision after the convolution.

The inventor finds in realizing process of the present invention: above-mentioned to being that the method that channel estimating is carried out in single footpath that accuracy selection goes out is compared with the conventional channel method of estimation with 1/4 chip, though avoided owing to introduce the decline that noise causes the LMMSE equalizer performance, but, because the channel estimation value that estimates has been mixed with the influence of RC, this moment is again with the channel estimation value and the RC convolution that estimate, further introduced the interference between the footpath, this can cause producing deviation between the channel estimation value that obtains after the convolution and the actual channel value, thereby the portfolio effect of LMMSE equalizer is descended, particularly in channel, exist energy all stronger, and distance is during less than two of 1 chip footpaths, because footpath and the interference between the footpath that RC produces are bigger, the channel estimating effect is just poorer.

Summary of the invention

Embodiment of the invention technical problem to be solved is: eliminate channel estimation value because be mixed with that the RC influence causes and the footpath of introducing during RC system convolution between interference, and then avoid improving the portfolio effect of LMMSE equalizer because the interference between the footpath causes deviation occurring between channel estimation value after the convolution and the actual channel.

A kind of channel estimation methods that provides according to one embodiment of present invention comprises:

Pilot channel is carried out Multipath searching, select channel window self-energy greater than pre-conditioned footpath;

Channel estimating is carried out in described footpath;

The channel estimation value that channel estimating obtains is eliminated the raised cosine interference, obtain not having the channel estimation value that raised cosine disturbs;

The channel estimation value that described no raised cosine is disturbed carries out the raised cosine convolution, obtains all locational footpath energy in the pilot channel.

According to another aspect of the present invention, a kind of channel estimating apparatus that provides comprises:

The Multipath searching module is used for pilot channel is carried out Multipath searching;

Select module, select energy greater than pre-conditioned footpath in the channel window that is used for going out from Multipath searching;

Channel estimation module is used for described energy is carried out channel estimating greater than pre-conditioned footpath;

Recovery module, the channel estimation value that is used for that channel estimating is obtained is eliminated raised cosine and is disturbed;

Convolution module is used for the channel estimation value of eliminating the no raised cosine interference that obtains after raised cosine disturbs is carried out the raised cosine convolution.

A kind of receiver that provides according to still another embodiment of the invention comprises root raised cosine filter, channel estimating apparatus, noise power estimation apparatus, weights calculation element, equalizer and descrambling and de-spreading device, and described channel estimating apparatus comprises:

The Multipath searching module is used for pilot channel is carried out Multipath searching;

Select module, select energy greater than pre-conditioned footpath in the channel window that is used for going out from Multipath searching;

Channel estimation module is used for described energy is carried out channel estimating greater than pre-conditioned footpath;

Recovery module, the channel estimation value that is used for that channel estimating is obtained is eliminated raised cosine and is disturbed;

Convolution module is used for the channel estimation value of eliminating the no raised cosine interference that obtains after raised cosine disturbs is carried out the raised cosine convolution.

When the embodiment of the invention is carried out channel estimating, only the channel window self-energy of selecting is carried out channel estimating greater than pre-conditioned footpath, can not introduce new noise, effectively guaranteed the equalization performance of LMMSE equalizer; In addition, after the channel estimation value that the conventional channel method of estimation is estimated is eliminated the RC influence, again it is carried out convolution, avoided owing to exist RC to influence the interference of introducing when carrying out the RC convolution again between the footpath, make channel estimation value more near the actual channel value, effectively guaranteed the portfolio effect of LMMSE equalizer.

Below by drawings and Examples, the technical scheme of the embodiment of the invention is described in further detail.

Description of drawings

Fig. 1 is the structural representation of prior art WCDMA signal receiving and transmitting system.

Fig. 2 is the structural representation of prior art conventional channel estimation unit.

Fig. 3 is the flow chart of channel estimation methods embodiment of the present invention.

Fig. 4 is the flow chart of another embodiment of channel estimation methods of the present invention.

Fig. 5 is the present invention's middle energy intercepting schematic diagram embodiment illustrated in fig. 4.

Fig. 6 is the structural representation of channel estimating apparatus embodiment of the present invention.

Fig. 7 is the structural representation of receiver embodiment of the present invention.

Embodiment

The embodiment of the invention is carried out channel estimating to channel window self-energy greater than pre-conditioned footpath, and carries out the RC convolution again after channel estimation value is eliminated the RC interference, obtains all locational footpath energy in the pilot channel.

As shown in Figure 3, be the flow chart of channel estimation methods embodiment of the present invention, it may further comprise the steps:

Step 101 is carried out Multipath searching to pilot channel, searches out channel window self-energy and satisfies pre-conditioned footpath;

Step 102 satisfies pre-conditioned footpath to energy and carries out channel estimating.

Only the energy of selecting is satisfied pre-conditioned footpath and carry out channel estimating, can not introduce new noise, effectively guaranteed the equalization performance of LMMSE equalizer;

Step 103 is eliminated the RC influence to the channel estimation value that channel estimating obtains, and obtaining does not have the corresponding channel estimation value directly that RC disturbs;

Step 104, corresponding channel estimation value and the RC filter parameter matrix directly that no RC is disturbed carries out convolution, obtains all locational footpath energy in the pilot channel.

After the channel estimation value that channel estimating is gone out is eliminated the RC influence, again it is carried out convolution, avoided owing to there is the RC influence, thereby the interference of introducing between the footpath when carrying out the RC convolution again produces deviation between channel estimation value and actual channel value, make channel estimation value more near the actual channel value, further guaranteed the portfolio effect of LMMSE equalizer.

As shown in Figure 4, be the flow chart of another embodiment of channel estimation methods of the present invention, it may further comprise the steps:

Step 201 is carried out Multipath searching to pilot channel, measures the energy in each footpath in the channel window that searches out, and selects the strongest footpath of channel window self-energy, and the footpath that energy is the strongest is also referred to as main footpath, and deletes each two phase point about this master footpath.

Step 202, except that the strongest footpath of energy, the strongest footpath of energy in all the other footpaths, that is: the stronger footpath of energy, the footpath that energy is stronger are also referred to as time footpath in the selective channel window, and each three phase point about the stronger footpath of deletion energy.By identical mode, in all the other each footpaths, to select the strongest time footpath of a plurality of energy successively and delete each three phase point about it, the concrete time footpath of selecting can be 8 with the quantity altogether in main footpath, also can be predefined other quantity.

Because the RC system is to the influence of multipath, the multipath in the reality is not the impulse response in single footpath, and the stronger footpath of energy still has bigger energy apart from this footpath less than the path position of a chip.When carrying out Multipath searching, reach the path position of predetermined threshold value for the signal to noise ratio of selecting, about respectively delete the phase point at 3/4 chip place.Multipath searching to satisfy in the pre-conditioned footpath because the interference that the subsequent step of this programme can be removed the RC system, the energy of reduction Dan Jing, therefore, the channel estimation value that keeps both sides, main footpath helps analog channel more accurately.Owing to when removing the interference of RC system, must satisfy conditions of diagonal dominance,, cause matrix morbid state to invert easily if select the phase point in each 1/2 chip about main footpath is directly removed interference to inverting of RC matrix.For this reason, in the present embodiment, select each two phase point of the main footpath of deletion the right and left, that is: the phase point in each 1/2 chip about the main footpath of deletion, each three phase point of deletion time footpath the right and left, a phase point is 1/4 chip, and four phase points are arranged in a chip.

In addition, also can delete each three phase point about this master footpath, each two phase point about deletion time footpath.But, because main footpath has the strongest energy, for each three phase point about deletion, each two phase point can keep the energy in main footpath morely about deletion, helps obtaining channel estimation value more accurately.

Step 203 from the strongest footpath of the energy selected and the stronger footpath of energy, is selected energy greater than pre-conditioned footpath.

Wherein, energy greater than pre-conditioned specifically can be: energy or signal to noise ratio be greater than predetermined threshold value, and the ratio of the energy on perhaps should the footpath and path energy is greater than preset value.

Step 204 satisfies each pre-conditioned footpath to energy and carries out channel estimating, obtains channel estimation value CE.Particularly, adopt conventional channel estimation unit as shown in Figure 2 to carry out channel estimating.

Step 205, because the RC between footpath and the footpath disturbs, have of the interference of other footpath on the channel estimation value in each footpath of obtaining by step 204, owing to the RC parameter is known to this footpath, adopt the method for least square that channel estimation value CE is eliminated the RC influence, obtain not having the estimation CE ' of the multipath of RC influence.

Suppose Multipath searching to path position be: P1, P2, P3, P4, P5, P6, P7 and P8; Delay inequality between above-mentioned footpath and the footpath is: ${\mathrm{\τ}}_{i,j}=\underset{i=1}{\overset{8}{\mathrm{\Σ}}}\underset{j=1;\mathrm{ji}}{\overset{8}{\mathrm{\Σ}}}(P_i-P_j);$ Channel estimation value on each footpath that is obtained by step 204 is: CE ₁, CE ₂, CE ₃, CE ₃, CE ₄, CE ₅, CE ₆, CE ₇, CE ₈After eliminating the RC influence, the channel estimation value of the no RC influence on each footpath that obtains is: CE ' ₁, CE ' ₂, CE ' ₃, CE ' ₃, CE ' ₄, CE ' ₅, CE ' ₆, CE ' ₇, CE ' ₈

RC is the filter parameter matrix of RRC convolution RRC, then can eliminate the RC influence by following equation:

$\left[\begin{array}{cccccccc}1& \mathrm{RC}\left({\mathrm{\τ}}_{\mathrm{1,2}}\right)& \mathrm{RC}\left({\mathrm{\τ}}_{\mathrm{1,3}}\right)& \mathrm{RC}\left({\mathrm{\τ}}_{\mathrm{1,4}}\right)& \mathrm{RC}\left({\mathrm{\τ}}_{\mathrm{1,5}}\right)& \mathrm{RC}\left({\mathrm{\τ}}_{\mathrm{1,6}}\right)& \mathrm{RC}\left({\mathrm{\τ}}_{\mathrm{1,7}}\right)& \mathrm{RC}\left({\mathrm{\τ}}_{\mathrm{1,8}}\right)\\ \mathrm{RC}\left({\mathrm{\τ}}_{\mathrm{2,1}}\right)& 1& \mathrm{RC}\left({\mathrm{\τ}}_{\mathrm{2,3}}\right)& ...& ...& ...& ...& ...\\ ...& ...& 1& ...& ...& ...& ...& ...\\ ...& ...& ...& 1& ...& ...& ...& ...\\ ...& ...& ...& ...& 1& ...& ...& ...\\ ...& ...& ...& ...& ...& 1& ...& ...\\ ...& ...& ...& ...& ...& ...& 1& \mathrm{RC}\left({\mathrm{\τ}}_{\mathrm{7,8}}\right)\\ \mathrm{RC}\left({\mathrm{\τ}}_{\mathrm{8,1}}\right)& ...& ...& ...& ...& ...& \mathrm{RC}\left({\mathrm{\τ}}_{\mathrm{8,7}}\right)& 1\end{array}\right]\left[\begin{array}{c}{{\mathrm{CE}}^{\′}}_1\\ {{\mathrm{CE}}^{\′}}_2\\ {{\mathrm{CE}}^{\′}}_3\\ {{\mathrm{CE}}^{\′}}_4\\ {{\mathrm{CE}}^{\′}}_5\\ {{\mathrm{CE}}^{\′}}_6\\ {{\mathrm{CE}}^{\′}}_7\\ {{\mathrm{CE}}^{\text{\′}}}_8\end{array}\right]+\left[\begin{array}{c}{n}_1\\ n_2\\ n_3\\ n_4\\ n_5\\ n_6\\ n_7\\ n_8\end{array}\right]=\left[\begin{array}{c}{\mathrm{CE}}_1\\ {\mathrm{CE}}_2\\ {\mathrm{CE}}_3\\ {\mathrm{CE}}_4\\ {\mathrm{CE}}_5\\ {\mathrm{CE}}_6\\ {\mathrm{CE}}_7\\ {\mathrm{CE}}_8\end{array}\right];$

Above-mentioned matrix equation can be abbreviated as: RCCE '+n=CE, wherein, n is the noise factor of above-mentioned each path position.

When each component of error vector has identical variance and each component when uncorrelated, can adopt the available mean square error minimum of method of least square.The equation that disturbs for RC between above-mentioned elimination footpath makes the quadratic sum minimum of error by the criterion of least square, that is:

Minimum.Wherein the number of the unknown parameter CE ' that separates of demand equates with the number of equation, and method by least square can obtain it and separate and be CE '=RC ^-1CE;

Further, obtain the channel estimation value CE ' in the footpath of preceding each position of RC influence to the RC matrix inversion.Because the RC matrix is a real symmetric matrix, can adopt invert method, Gaussian elimination of Gauss-Saden you (Gauss-Siedel) to decompose the method for inverting or Qiao Lisiji (cholesky) and decompose the method for inverting and wait matrix inversion, thereby try to achieve the footpath channel estimation value CE ' that eliminates the correspondence position that RC disturbs RC.

Step 206, the channel estimation value CE ' in the footpath that no RC is disturbed carries out convolution with RC filter parameter matrix, obtains all locational footpath energy in the pilot channel.

Step 207, because pilot channel adopts 1/4 sample rate, what eliminate that the channel estimation value CE ' that obtains after RC disturbs and RC filter parameter matrix carry out that convolution obtains is the channel estimation value of 1/4 chip-level

Channel estimation value CE ' and the convolution of RC filter parameter matrix can cause the hangover of energy, according to the descending order of energy, obtain energy sequence after convolution

In, intercept the identical portion of energy of the length on channel with the preceding CE ' of convolution.

As shown in Figure 5, be energy intercepting schematic diagram of the present invention.Wherein, the energy of the CE ' before Fig. 5-1 expression convolution on each path position supposes that its channel length is L; The energy of Fig. 5-2 expression RC filter on corresponding path position supposes that this channel length is M; Energy length after Fig. 5-3 expression channel estimation value CE ' and the RC filter parameter matrix convolution, its channel length becomes L+M, and in order to keep whole energy as far as possible, therefrom intercepting channel length is the bigger energy part of L.

Step 208, the channel estimated accuracy that step 207 obtains is 1/4 chip, because equalizer tap is generally one or two sampled point in the chip, therefore needed channel estimating also should be one or two sampled point in the chip, for this reason, all locational channel estimation values in the pilot channel are carried out down-sampling, obtain the channel estimation value CE ' of a chip or 1/2 chip.

Step 209 is by channel estimation value CE ' generation channel estimate matrix.

In addition, also can increase the equation number, make the number of equation number greater than unknown number, adopt least mean-square error (Minimum Mean Square Error, hereinafter to be referred as: MMSE) or other least square method, come channel estimation value CE is eliminated the RC influence, obtain not having the estimation CE ' of the multipath of RC influence.

Suppose by step 201 and step 202, Multipath searching to path position be followed successively by according to the energy size: P1, P2, P3, P4, P5, P6, P7 and P8, wherein, have only path position P1 and P2 to satisfy pre-conditioned in the step 203, from path position P3, P4, P5, P6, P7 and P8, elect a path position that carries out channel estimating additional member, suppose to select P3, with the channel estimation value CE of path position P1, P2 and P3 ₁, CE ₂With CE ₃Restore the channel estimation value CE ' of the elimination RC influence of path position P1 and P2 ₁With CE ' ₂Can eliminate the RC influence by following equation:

$\left[\begin{array}{cc}1& \mathrm{RC}\left({\mathrm{\τ}}_{\mathrm{1,2}}\right)\\ \mathrm{RC}\left({\mathrm{\τ}}_{\mathrm{2,1}}\right)& 1\\ \mathrm{RC}\left({\mathrm{\τ}}_{\mathrm{3,1}}\right)& \mathrm{RC}\left({\mathrm{\τ}}_{\mathrm{2,3}}\right)\end{array}\right]\left[\begin{array}{c}C{E^{\′}}_1\\ {{\mathrm{CE}}^{\′}}_2\end{array}\right]+\left[\begin{array}{c}{n}_1\\ n_2\\ n_3\end{array}\right]=\left[\begin{array}{c}{\mathrm{CE}}_1\\ {\mathrm{CE}}_2\\ {\mathrm{CE}}_3\end{array}\right];$

Adopt the method for least square to try to achieve

Channel estimate matrix H ' be:

Adopt the MMSE method to try to achieve

Channel estimate matrix H ' be:

Wherein, σ _n ²Be noise energy, I ₃Be unit matrix.

As shown in Figure 6, be the structural representation of channel estimating apparatus embodiment of the present invention, this channel estimating apparatus comprises Multipath searching module 301, selection module 302, channel estimation module 303, the recovery module 304 and convolution module 305 that connects successively.Wherein, Multipath searching module 301 is used for pilot channel is carried out Multipath searching; Select energy greater than pre-conditioned footpath in the channel window of selecting module 302 to be used for going out from Multipath searching; Channel estimation module 303 is used for energy is carried out channel estimating greater than pre-conditioned footpath; The channel estimation value that recovery module 304 is used for that channel estimating is obtained is eliminated RC and is disturbed; Convolution module 305 is used for the channel estimation value of eliminating the no RC interference that obtains after RC disturbs is carried out the RC convolution.

In the channel estimating apparatus shown in Figure 6, select module 302 specifically can constitute by measuring unit 3021 and selected cell 3022, wherein, measuring unit 3021 is connected with Multipath searching module 301, be used to measure the energy or the signal to noise ratio of the channel window internal diameter that Multipath searching module 301 searches out, further, measuring unit 3021 can also select successively the strongest footpath of channel window self-energy with all the other the footpath in the stronger footpath of energy; Selected cell 3022 is connected with channel estimation module 303, is used for the measurement result according to measuring unit 3021, selects channel window self-energy greater than pre-conditioned footpath and send to channel estimation module 303.The channel estimating apparatus of this embodiment can be used for realizing the channel estimation methods of embodiment as shown in Figure 3.

Further, select to comprise delete cells 3023 in the module, be arranged between measuring unit 3021 and the selected cell 3022, be used to delete each two phase point about the strongest main footpath of the energy of selecting measuring unit 3021 to select, each three phase point about the stronger time footpath of energy.

In addition, channel estimating apparatus shown in Figure 6 can also comprise information storage module 306, be connected with selected cell 3022, be used for the conditional information of stored energy greater than pre-conditioned footpath, for example: energy is greater than preset value A, perhaps signal to noise ratio is greater than preset value B, perhaps with the ratio of the footpath energy of minimum greater than preset value C; Particularly, selected cell 3022 is used for the main footpath of the transmission of measuring unit 3021 is compared with the energy in inferior footpath or the corresponding conditions information of signal to noise ratio and information storage module 306, selects channel window self-energy greater than pre-conditioned footpath.

Further, the above-mentioned channel estimating apparatus that the embodiment of the invention provides can also comprise parameter memory module 307, is connected with recovery module 304 and convolution module 305 respectively, is used to store RC filter parameter matrix; Particularly, recovery module 304 is used for obtaining RC filter parameter matrix from parameter memory module 307, to the channel estimation value quadrature that this RC filter parameter inverse of a matrix matrix and channel estimating obtain, the channel estimation value that comes channel estimating is obtained is eliminated RC and is disturbed; Convolution module 305 is used for obtaining RC filter parameter matrix from parameter memory module 307, and channel estimation value and RC filter parameter matrix that no RC is disturbed carry out the raised cosine convolution.

Referring to Fig. 6, the channel estimating apparatus of the embodiment of the invention can also comprise down sample module 308 again, is used for all the locational footpath energy of pilot channel that obtain after the RC convolution are carried out down-sampling.Further, can also comprise matrix generation module 309 in the above-mentioned channel estimating apparatus, the channel estimation value that is used for 1/2 chip that obtains according to down-sampling generates channel estimate matrix.The channel estimating apparatus of this embodiment can be used for realizing the channel estimation methods of embodiment as shown in Figure 4.

Referring to Fig. 7, structural representation for receiver embodiment of the present invention, the receiver of this embodiment comprises RRC filter, channel estimating apparatus, weights calculation element, equalizer and the descrambling and de-spreading device that connects successively, and the noise power estimation apparatus that is connected with the weights calculation element, the output of RRC filter also is connected with the input of equalizer.Wherein, channel estimating apparatus can adopt the channel estimating apparatus of arbitrary embodiment as shown in Figure 6, particularly, the output of RRC filter is connected with Multipath searching module 301 in the channel estimating apparatus, and the convolution module 305 in the channel estimating apparatus or the output of down sample module 308 are connected with the input of weights calculation element.

Particularly, the RRC filter is used to receive through the transmitting of the multidiameter fading channel transmission that is superimposed with white Gaussian noise, and carries out sending to respectively after the filtering Multipath searching module 301 in equalizer and the channel estimating apparatus to the received signal;

Channel estimating apparatus adopts and carry out channel estimating as Fig. 3 or method shown in Figure 4, and with the channel estimation value of estimating, that is: all locational footpath energy or channel estimate matrix sends to the weights calculation element in the pilot channel based on filtered signal;

Noise power estimation apparatus is estimated the white Gaussian noise energy in the channel, and the noise energy of estimating is also sent to calculation element;

Calculation element calculates according to the channel estimation value of channel estimating apparatus transmission and the noise energy of noise power estimation apparatus transmission, obtains the tap weight value of equalizer and sends to equalizer;

Equalizer carries out equilibrium to the filtered signal of RRC filter transmission and the tap weight value of equalizer, and the equalizing signal that obtains after the equilibrium is sent to the descrambling and de-spreading device according to homeostatic principle.This equalizer is the LMMSE equalizer;

The descrambling and de-spreading device carries out descrambling, despreading to equalizing signal, obtains the original transmission signal that transmitter sends.

When the embodiment of the invention is carried out channel estimating, only the channel window self-energy of selecting is carried out channel estimating greater than pre-conditioned footpath, can not introduce new noise, effectively guaranteed the equalization performance of LMMSE equalizer; In addition, after the channel estimation value that the conventional channel method of estimation is estimated is eliminated the RC influence, again it is carried out convolution, avoided owing to there is the RC influence, when carrying out the RC convolution, introduce the interference between the footpath, thereby between channel estimation value and actual channel value, produce deviation, further guaranteed the portfolio effect of LMMSE equalizer.

It should be noted last that: above embodiment is only in order to illustrating technical scheme of the present invention, but not the present invention is made restrictive sense.Although the present invention is had been described in detail with reference to above-mentioned preferred embodiment, those of ordinary skill in the art is to be understood that: it still can make amendment or be equal to replacement technical scheme of the present invention, and this modification or be equal to the spirit and scope that replacement does not break away from technical solution of the present invention.

Claims (11)

1, a kind of channel estimation methods is characterized in that, may further comprise the steps:

Pilot channel is carried out Multipath searching, select channel window self-energy greater than pre-conditioned footpath;

Channel estimating is carried out in described footpath;

The channel estimation value that channel estimating obtains is eliminated the raised cosine interference, obtain not having the channel estimation value that raised cosine disturbs;

The channel estimation value that described no raised cosine is disturbed carries out the raised cosine convolution, obtains all locational footpath energy in the pilot channel.

2, channel estimation methods according to claim 1 is characterized in that, energy is specially greater than pre-conditioned: energy is greater than preset value, and perhaps signal to noise ratio is greater than preset value, perhaps with the ratio of the footpath energy of minimum greater than preset value.

3, channel estimation methods according to claim 1 and 2, it is characterized in that, described pilot channel is carried out Multipath searching, selecting channel window self-energy comprises greater than pre-conditioned footpath: pilot channel is carried out Multipath searching, select each two or three phase point about the strongest main footpath of energy and the main footpath of deletion from the footpath that the Multipath searching operation obtains; From all the other path positions, select the strongest time footpath of energy and delete each two or three phase point about this time footpath, select and delete each two or three phase point about a plurality of inferior footpaths successively; Directly with a plurality of the footpaths select energy from the master who selects greater than pre-conditioned footpath.

4, channel estimation methods according to claim 3, it is characterized in that, the described channel estimation value that channel estimating is obtained is eliminated raised cosine and is disturbed and be specially: adopt least square method to the channel estimation value quadrature that inverse matrix and the channel estimating of raised cosine filter parameter matrix obtains, eliminate described channel estimation value raised cosine interference.

5, according to any described channel estimation methods of claim 1 to 4, it is characterized in that, obtain in the pilot channel also comprising behind all locational footpath energy:

All locational footpath energy in the pilot channel are carried out down-sampling, obtain the channel estimation value of 1/2 chip;

Channel estimation value according to described 1/2 chip generates channel estimate matrix.

6, a kind of channel estimating apparatus is characterized in that, comprising:

The Multipath searching module is used for pilot channel is carried out Multipath searching;

Select module, select energy greater than pre-conditioned footpath in the channel window that is used for going out from Multipath searching;

Channel estimation module is used for described energy is carried out channel estimating greater than pre-conditioned footpath;

Recovery module, the channel estimation value that is used for that channel estimating is obtained is eliminated raised cosine and is disturbed;

Convolution module is used for the channel estimation value of eliminating the elimination raised cosine interference that obtains after raised cosine disturbs is carried out the raised cosine convolution.

7, channel estimating apparatus according to claim 6 is characterized in that, also comprises:

Parameter memory module is used to store the raised cosine filter parameter matrix;

Described recovery module is used for channel estimation value quadrature that the inverse matrix of raised cosine filter parameter matrix and channel estimating are obtained, and the channel estimation value that comes channel estimating is obtained is eliminated raised cosine and disturbed.

Described convolution module is used for channel estimation value and described raised cosine filter parameter matrix that described elimination raised cosine disturbs are carried out the raised cosine convolution.

8, according to claim 6 or 7 described channel estimating apparatus, it is characterized in that, also comprise:

Down sample module is used for all the locational footpath energy of pilot channel that obtain after the raised cosine convolution are carried out down-sampling;

The matrix generation module, the channel estimation value that is used for 1/2 chip that obtains according to down-sampling generates channel estimate matrix.

9, a kind of receiver comprises root raised cosine filter, channel estimating apparatus, noise power estimation apparatus, weights calculation element, equalizer and descrambling and de-spreading device, it is characterized in that described channel estimating apparatus comprises:

The Multipath searching module is used for pilot channel is carried out Multipath searching;

Select module, select energy greater than pre-conditioned footpath in the channel window that is used for going out from Multipath searching;

Channel estimation module is used for described energy is carried out channel estimating greater than pre-conditioned footpath;

Recovery module, the channel estimation value that is used for that channel estimating is obtained is eliminated raised cosine and is disturbed;

Convolution module is used for the channel estimation value of eliminating the no raised cosine interference that obtains after raised cosine disturbs is carried out the raised cosine convolution.

10, receiver according to claim 9 is characterized in that, described channel estimating apparatus also comprises:

Parameter memory module is used to store the raised cosine filter parameter matrix;

Described recovery module is used for channel estimation value quadrature that the inverse matrix of raised cosine filter parameter matrix and channel estimating are obtained, and the channel estimation value that comes channel estimating is obtained is eliminated raised cosine and disturbed.

Described convolution module is used for channel estimation value and described raised cosine filter parameter matrix that described elimination raised cosine disturbs are carried out the raised cosine convolution.

11, according to claim 9 or 10 described receivers, it is characterized in that described channel estimating apparatus also comprises:

Down sample module is used for all the locational footpath energy of pilot channel that obtain after the raised cosine convolution are carried out down-sampling;

The matrix generation module, the channel estimation value that is used for 1/2 chip that obtains according to down-sampling generates channel estimate matrix.

Images