CN1968030A

CN1968030A - Channel estimation method of frequency-domain receiver

Info

Publication number: CN1968030A
Application number: CNA2006100780598A
Authority: CN
Inventors: 曹爱军
Original assignee: Huawei Technologies Co Ltd
Current assignee: Huawei Technologies Co Ltd
Priority date: 2006-04-29
Filing date: 2006-04-29
Publication date: 2007-05-23

Abstract

The invention relates to a signal channel evaluating method of frequency-domain receiver, wherein it converts receiving signal to frequency domain to obtain frequency receiving signal; via iterate method, repeatedly processes spread-spectrum factor removing, signal channel removing, constant-gain combination, to obtain signal evaluation. The invention can improve the accuracy of signal channel evaluation.

Description

Channel estimation method of frequency domain receiver

Technical Field

The invention relates to the field of mobile communication, in particular to a channel estimation method of a DS-CDMA (Direct Sequence code division Multiple Access) signal frequency domain receiver.

Background

The direct sequence code division multiple access DS-CDMA method IS adopted in the second generation wireless cellular communication system IS-95 and the third generation wireless cellular communication systems CDMA2000, WCDMA, etc. which are currently being developed. In DS-CDMA systems, the low-speed signal of each user is multiplied by a high-speed signal, called the spread-spectrum signal, associated with the user as the user's transmitted signal; while all users in the system use the same bandwidth and frequency and can transmit simultaneously. Each user has its own spread spectrum signal that remains orthogonal to the spread spectrum signals of the other users. In the receiver, correlation operation is carried out on the transmitted signal of a specific user, the current wireless channel is estimated, and then the correlated result is corrected to recover the information of the user.

In general, in a DS-CDMA receiver, the estimation of the channel is often performed by temporally inserting a known pilot signal in the transmitted signal, and accordingly, the channel estimation is performed in the time domain. If the delay spread of the wireless channel is large, and the orthogonal characteristic of the spread spectrum signal after wireless multipath propagation is damaged to a certain extent, additional interference occurs, and the accuracy of channel estimation is affected.

Generally, within one Symbol (Symbol), the time domain expression of the transmitted signal is:

s = [\begin{matrix} s_{1} \\ . . . \\ s_{N} \end{matrix}] = \sqrt{E_{s}} [\begin{matrix} c_{1}^{d} \\ . . . \\ c_{N}^{d} \end{matrix}] b + \sqrt{E_{p}} [\begin{matrix} c_{1}^{p} \\ . . . \\ c_{N}^{p} \end{matrix}] - - - (1)

wherein E is_sFor user data energy, E_pFor the energy of the pilot frequency,the spreading code for the user of the dedicated channel,

is a pilot spreading code; b is a transmission symbol, N represents a spreading code length, P represents a pilot, and d represents a dedicated channel.

The transmitted signal is FFT transformed by M points, where M > ═ N, and the corresponding frequency domain FFT is expressed as:

S = [\begin{matrix} S_{1} \\ . . . \\ S_{M} \end{matrix}] = \sqrt{E_{s}} [\begin{matrix} C_{1}^{d} \\ . . . \\ C_{M}^{d} \end{matrix}] b + \sqrt{E_{p}} [\begin{matrix} C_{1}^{p} \\ . . . \\ C_{M}^{P} \end{matrix}] - - - (2)

wherein,for the FFT transformation of the spreading code of the user,

is an FFT transformation of the pilot spreading code.

Then, the FFT of the received signal can be expressed as:

R = [\begin{matrix} R_{1} \\ . . . \\ R_{M} \end{matrix}] = [\begin{matrix} H_{1} (\sqrt{E_{s}} C_{1}^{d} b + \sqrt{E_{p}} C_{1}^{p}) \\ . . . \\ H_{M} (\sqrt{E_{s}} C_{M}^{d} b + \sqrt{E_{p}} C_{M}^{P}) \end{matrix}] + [\begin{matrix} O_{1} \\ . . . \\ O_{M} \end{matrix}] - - - (3)

wherein,H₁…H_Mrepresenting the value of the impulse response of the channel after FFT, i.e. the channel spectrum, O₁…O_MIt refers to the value of the noise after FFT, i.e. the noise spectrum. If M > N, oversampling, common zero padding or interpolation, etc. may be applied to the input signal, so that the number of input signal points required for FFT is increased from N to M.

The following is the demodulation process for the received signal R:

a. removing the influence of the spreading factor:

R 1 = [\begin{matrix} R_{1} C_{1}^{d^{*}} \\ . . . \\ R_{N} C_{N}^{d^{*}} \end{matrix}] = [\begin{matrix} (H_{1} (\sqrt{E_{s}} C_{1}^{d} b + \sqrt{E_{p}} C_{1}^{p}) + O_{1}) C_{1}^{d^{*}} \\ . . . \\ (H_{N} (\sqrt{E_{s}} C_{N}^{d} b + \sqrt{E_{p}} C_{N}^{p}) + O_{N}) C_{N}^{d^{*}} \end{matrix}] - - - (4)

b. removing channel influence:

Y = [\begin{matrix} Y_{1} \\ . . . \\ Y_{N} \end{matrix}] = [\begin{matrix} R_{1} C_{1}^{d^{*}} H_{1}^{*} \\ . . . \\ R_{N} C_{N}^{d^{*}} H_{N}^{*} \end{matrix}] = [\begin{matrix} (H_{1} (\sqrt{E_{s}} C_{1}^{d} b + \sqrt{E_{p}} C_{1}^{p}) + O_{1}) C_{1}^{d^{*}} H_{1}^{*} \\ . . . \\ (H_{N} (\sqrt{E_{S}} C_{N}^{d} b + \sqrt{E_{p}} C_{N}^{p}) + O_{N}) H_{N}^{*} \end{matrix}]

= [\begin{matrix} \sqrt{E_{s}} {| H_{1} |}^{2} {| C_{1}^{d} |}^{2} b + {| H_{1} |}^{2} \sqrt{E_{p}} C_{1}^{p} C_{1}^{d^{*}} + O_{1} C_{1}^{d^{*}} H_{1}^{*} \\ . . . \\ \sqrt{E_{s}} {| H_{N} |}^{2} {| C_{N}^{d} |}^{2} b + {| H_{N} |}^{2} \sqrt{E_{p}} C_{N}^{p} C_{N}^{d^{*}} + O_{N} C_{N}^{d^{*}} H_{N}^{*} \end{matrix}] - - - (5)

c. the columns are summed to obtain the decision variable (equal gain combining):

the first term in the above equation (6) represents signal power, the third term is related to noise, the second term is interference caused by pilot, and the pilot spreading code and the user data spreading code are orthogonal in the time domain and also orthogonal in the frequency domain after FFT, but the second term is not 0 due to the channel weight, and therefore, interference cancellation is required. In the prior art, a frequency domain receiver for DS-CDMA is disclosed, in which pilot interference cancellation is performed before despreading in the downlink case, and this method uses a simple averaging method to obtain an estimate of the channel. However, when the channel estimation is performed by using this method, due to the presence of non-orthogonal terms, the estimation accuracy is low, which makes the channel estimation difficult, and it is completely impossible to perform simple channel estimation by using the time-averaging method as in the time domain, and thus a more effective channel estimation method is required.

Disclosure of Invention

The invention aims to provide a channel estimation method capable of improving the accuracy of channel estimation in a frequency domain receiver.

In order to achieve the above object, the present invention provides a channel estimation method for a frequency domain receiver, which performs channel estimation on a received signal of the frequency domain receiver, and converts the received signal to a frequency domain to obtain a frequency domain received signal; and repeatedly carrying out spreading factor influence removal, channel influence removal and equal gain combination processing on the channel frequency domain received signals for multiple times through an iteration method to obtain channel estimation.

In the above method of the present invention, the following operations are performed within the tth symbol, where t ≧ 1:

1) performing initial estimation on a channel to obtain initial channel estimation;

2) sequentially removing spread spectrum factor influence, removing channel influence and equal gain combination on the frequency domain received signals in the step 1) by using the current channel estimation to obtain initial judgment;

3) updating channel estimation according to the obtained judgment;

4) removing the influence of a spreading factor, removing the influence of a channel, carrying out equal gain combination on the frequency domain receiving signal by using the updated channel estimation, and then carrying out judgment to obtain a new judgment of the symbol;

5) and repeating the steps 3) and 4) for multiple iterations to obtain the final channel estimation.

In the above method of the present invention, before performing step 1), the following steps are further performed in the tth symbol:

1A) carrying out FFT operation on the received signal to obtain a frequency domain received signal on a frequency domain;

1B) for the ratio of user data power to pilot power in frequency domain received signal

Q = \frac{\sqrt{E_{s}}}{\sqrt{E_{p}}}

Estimating in the time domain;

in the step 3), channel estimation is updated according to the obtained decision and the Q parameter.

In the above method of the present invention, in step 1), initial estimation is performed on a channel according to a frequency domain received signal and an M-point FFT transform value of a local user spreading code, where M is the number of subsegments of the entire spectrum; in the step 3), updating the channel estimation according to the M-point FFT variation value of the local user spreading code, the M-point FFT variation value of the local pilot spreading code, the Q parameter obtained in the step 1B), the currently obtained decision, and the frequency domain received signal obtained in the step 1A).

In the above method of the present invention, in the step 1B), the Q parameter is estimated by a ratio of a signal-to-interference ratio, SIR, of a dedicated channel to a common channel.

In the above method of the present invention, further comprising step 6): and smoothing the iteration result by adopting a moving average method in time to obtain the channel estimation finally used for user data demodulation at present.

The invention can greatly improve the accuracy of channel estimation in a frequency domain receiver by iterative smoothing, and can be used in a DS-CDMA system or an MC-CDMA (multicarrier code division multiple access) system.

Drawings

Fig. 1 shows a flow chart of the channel estimation method of the frequency domain receiver of the present invention.

Detailed Description

The following detailed description of embodiments of the invention refers to the accompanying drawings. However, it should be noted that the drawings and the detailed description are only for illustrating the present invention and the present invention is not limited thereto.

The channel estimation method of the frequency domain receiver is used for carrying out channel estimation on a received signal of the frequency domain receiver and converting the received signal to a frequency domain to obtain a frequency domain received signal; and repeatedly carrying out spreading factor influence removal, channel influence removal and equal gain combination processing on the channel frequency domain received signals for multiple times through an iteration method to obtain channel estimation.

Specifically, in the above method of the present invention, the following operations are performed within the tth symbol, where t ≧ 1:

3) updating channel estimation according to the obtained judgment;

Wherein, in the tth symbol, before performing step 1), the following steps may be further performed:

Q = \frac{\sqrt{E_{s}}}{\sqrt{E_{p}}}

The estimation is done in the time domain where the Q parameter can be estimated as the ratio of the signal-to-interference ratio, SIR, of the dedicated channel to the common channel.

In addition, in the above method of the present invention, step 6): and smoothing the iteration result by adopting a moving average method in time to obtain the channel estimation finally used for user data demodulation at present.

More specifically, in the above method, in step 1), an initial estimation is performed on the channel according to the frequency domain received signal and the M-point FFT transform value of the local user spreading code, where M is the number of subsegments of the entire spectrum; in the step 3), updating the channel estimation according to the M-point FFT variation value of the local user spreading code, the M-point FFT variation value of the local pilot spreading code, the Q parameter obtained in the step 1B), the currently obtained decision, and the frequency domain received signal obtained in the step 1A).

In the following, for convenience of description, without loss of generality, the following assumptions are made:

● the pilot frequency has the same length as the user data spreading code, both are N, the following description and derivation are also applicable to the case of unequal spreading code lengths, but the corresponding signs need to be corrected;

● no scrambling code, perfect descrambling can be realized under the condition of synchronous transmitting and receiving signals;

● pilot symbols are all 1, because the pilot symbols are known, assuming all 1, without loss of generality, and meanwhile, the formula can be made simpler;

● other user interference is equivalent to noise, or a single user case.

Fig. 1 shows a channel estimation method of a frequency domain receiver of a communication system using DS-CDMA spreading under the above-mentioned assumption.

According to equation (3), the received signal of the k-th sub-band is subjected to M-point FFT (M ═ N), which can be expressed as: -

R_{k} = H_{k} (\sqrt{E_{S}} C_{k}^{d} b + \sqrt{E_{p}} C_{k}^{p}) + O_{k}

= \sqrt{E_{p}} H_{k} (\frac{\sqrt{E_{s}}}{\sqrt{E_{p}}} C_{k}^{d} b + C_{k}^{p}) + O_{k} - - - (7)

Wherein, C_k ^dM-point FFT transformation of spreading codes for local users, C_k ^pM-point FFT transform of a local pilot spreading code, H_MRepresenting the value of the channel impulse response after an M-point FFT, i.e. the channel spectrum, O_MThe value of the noise after M-point FFT, i.e. the noise spectrum, k 1.. M, indicates the sequence number of these spectrum subsegments, and M is the FFT length, and indicates that the whole spectrum is divided into M subsegments. Because the pilot frequency transmitting power does not change along with the time, the pilot frequency transmitting power can be changed along with the time

As estimated quantityHk represents the frequency domain response of the k-th sub-band spectrum, representing the spectral characteristics of the channel.

The present invention performs channel estimation based on the above formula (7), and the adopted channel estimation algorithm is as follows:

1) according to the formula (4), to receiveSignal R_kRemoving the effect of spreading factors, i.e. making R_kFFT transform C by local user spreading code_k ^dIs conjugated with (C)_k ^d*To obtain R1.

2) User data power to pilot power ratio in time domain

Q = \frac{\sqrt{E_{s}}}{\sqrt{E_{p}}}

Is estimated. For example, the ratio of the SIR of the dedicated channel to the common channel can be used for estimation.

3) Within the t-th Symbol (Symbol) the following is done:

a. the initial estimation of the channel is performed according to the following formula:

{\hat{H}}_{k}^{(0)} = \frac{1}{C_{k}^{p}} R_{k}

the formula is given by formula (7) in the condition that b and O are ignored_kIs obtained byRepresents an initial estimate of the frequency domain response for the kth sub-band, k 1.

b. Receiving signal R according to formula (5)_kRemoving the channel effect, i.e. multiplying R1 obtained in step 1) above by

c. The first equal gain combination is carried out according to the formula (6) to obtain Z⁽⁰⁾Wherein the channel H_iUsing the estimated value in step a

Instead, a decision is made to obtain the first decision b of the t Symbol_t ⁽⁰⁾；

d. Performing a second channel estimation according to the following formula:

{\hat{H}}_{k}^{(1)} = \frac{1}{\hat{Q} C_{k}^{d} b_{t}^{(0)} + C_{k}^{p}} R_{k}

e. the second removal of the channel effects is performed according to equation (5), where the channel spectral response H_iUsing the estimated value H in step d_k ⁽¹⁾Replacing;

f. second equal gain combining according to equation (6) where the channel spectral response H_iUsing the estimated value H in step d_k ⁽¹⁾Instead, the decision is made to obtain the second decision b of the t Symbol_t ⁽¹⁾：

g. Repeating the steps b to f, carrying out W times of iteration to obtain the W-th estimation of the channel

{\hat{H}}_{k}^{(W)} = \frac{1}{\hat{Q} C_{k}^{d} b_{t}^{(W - 1)} + C_{k}^{p}} R_{k}

(3) And smoothing the iteration result by adopting a moving average method in time to obtain a channel estimation value finally used for user data demodulation at present.

The above method is not only suitable for frequency domain reception of single carrier DS-CDMA signals, but also suitable for use in MC-CDMA receivers.

The above specific embodiments illustrate the present invention, but these embodiments are exemplary, and the present invention is not limited to the specific embodiments. Those skilled in the art may make modifications, changes or substitutions to the present invention without departing from the spirit and scope of the present invention.

Claims

1. A channel estimation method of a frequency domain receiver, which performs channel estimation on a received signal of the frequency domain receiver,

converting the received signal to a frequency domain to obtain a frequency domain received signal; and repeatedly carrying out spreading factor influence removal, channel influence removal and equal gain combination processing on the channel frequency domain received signals for multiple times through an iteration method to obtain channel estimation.

2. The method of claim 1,

within the t-th symbol, the following operations are carried out, wherein t is more than or equal to 1:

3) updating channel estimation according to the obtained judgment;

3. The method of claim 2,

in the t-th symbol, before performing step 1), the following steps are also performed:

Q = \frac{\sqrt{E_{s}}}{\sqrt{E_{p}}}

Estimating in the time domain;

4. The method of claim 3,

in the step 1), performing initial estimation on a channel according to a frequency domain receiving signal and an M-point FFT (fast Fourier transform) value of a local user spread spectrum code, wherein M is the number of subsegments of the whole frequency spectrum;

in the step 3), updating the channel estimation according to the M-point FFT variation value of the local user spreading code, the M-point FFT variation value of the local pilot spreading code, the Q parameter obtained in the step 1B), the currently obtained decision, and the frequency domain received signal obtained in the step 1A).

5. The method of claim 3,

in said step 1B), said Q parameter is estimated by the ratio of the signal-to-interference ratio, SIR, of the dedicated channel to the common channel.

6. The method according to one of claims 2 to 5,

further comprising step 6): and smoothing the iteration result by adopting a moving average method in time to obtain a channel estimation value finally used for user data demodulation at present.

7. Method according to one of the claims 2-5, characterized in that it is used in a DS-CDMA system or a MC-CDMA system.