WO2008151468A1

WO2008151468A1 - Digital signal receiver and method for receiving digital signal

Info

Publication number: WO2008151468A1
Application number: PCT/CN2007/001882
Authority: WO
Inventors: Peng Liu; Li Zou
Original assignee: Thomson Licensing
Priority date: 2007-06-15
Filing date: 2007-06-15
Publication date: 2008-12-18

Abstract

The invention provides a digital signal receiver and method for receiving digital signal, the receiver comprises a first Common Phase Error removal unit (3401) configured to eliminate a first CPE of the digital signal obtained by calculating auto-correlation of the digital signal; a channel quality detector (301) configured to detect quality value of the CPE eliminated digital signal; and a second Common Phase Error removal unit (301-303, 3402) configured to selectively eliminate a second CPE of the CPE eliminated digital signal, according to the quality value.

Description

DIGITAL SIGNAL RECEIVER AND METHOD FOR RECEIVING DIGITAL SIGNAL

FIELD OF THE INVENTION

The present invention relates to a digital signal receiver and a method for receiving a digital signal, and more particularly, to a digital signal receiver and a method for receiving a digital signal used in a system based on multi- carrier modulation technology.

BACKGROUND OF THE INVENTION

Recently, multi-carrier modulation technology has been used widely, such as in electronic communications, optical communications, wired communications and wireless communications. Orthogonal Frequency Division Multiplexing (OFDM) is a typical multi-carrier modulation technology and a very promising access scheme for wideband wireless communication networks, which overcomes multi-path interference and becomes a focus of attention in a terrestrial digital broadcast. Therefore, OFDM has been adopted in a number of international standards such as DVB-T (Digital Video Broadcasting-Terrestrial) and wireless LAN (Local Area Network) . It is also a promising technique for the future wideband wireless communication systems, such as digital TV

(Television) broadcasting and 4th generation wireless networks

In practical communication systems, a modulator and a demodulator usually work either at baseband or at an intermediate frequency (IF) . As we have to transmit signal at some allocated radio frequency (RF) , it follows that we have to up-convert the modulated signal to RF channel in the transmitter, and down-convert RF signal to IF or baseband in the receiver. To do this, the practical local oscillators will be used, which bring phase noise (PHN) and interfere with the signal .

The phase noise is not a big problem for the conventional analog broadcasting systems, but the significance of the problem increases very strongly with the introduction of multi-carrier modulation systems, such as OFDM systems. The main difference between the OFDM and other digital modulation types is that the OFDM signal consists of multiple low rate sub-carriers that are orthogonal with each other, therefore the OFDM systems are very sensitive to phase noise. In addition, the low symbol rate makes the synchronization even more difficult when fast phase disturbances occur, so the phase noise degrades orthogonality of the sub-carriers. Especially, a consumer-oriented receiver based on OFDM technology has a more serious problem of degrading a bit error rate due to a phase noise of a local oscillator.

There are two influences of phase noise existed in a received OFDM signal. One is phase variation of sub-carriers generated by low-frequency components of the phase noise. This is called a common phase error (CPE) since all the sub- carriers are varied at the same angle. The other is an Inter- Carrier Interference (ICI) in which the SNR of carriers is degraded due to an interference of phase noise of other sub- carriers. They all interfere with demodulation of OFDM signals. In order to prevent efficiently the quality of OFDM signals from being degraded, the CPE has to be removed in the receiver.

Referring to Fig.l, which is a block diagram of an example of an OFDM signal receiver 100 in the prior art. The receiver 100 includes a demodulation module 110, a Fast Fourier Transform (FFT) module 120, a synchronization and timing (S&T) module 130, a Common Phase Error (CPE) removal module 140, a channel estimation (CE) module 150, and a channel decoding module 160. The demodulation module 110 demodulates OFDM signal received from an antenna to generate a complex signal including an in-phase signal and a quadrature-phase signal. The synchronization and timing module 130 estimates synchronization errors of the complex signal, and sends it to FFT module 120. The FFT module 120 determines a respective OFDM symbol according to the synchronization errors. Then the CPE removal module 140 estimates and corrects a Common Phase Error (CPE) contained in all sub-carriers of the OFDM digital signal. The CPE removal module 140 generally includes a CPE estimation module and a CPE removal module. Afterwards the corrected signal is equalized in the channel estimation module 150, and then transmitted to the channel decoding module 160 for the decoding process and output.

There are mainly three type of conventional solutions for the CPE estimation in the CPE removal module 120, wherein the first one is a cross correlation method, the second one employs an auto-correlation method, and the third type of the conventional CPE estimation employs a two-steps method, as illustrated respectively through Figs.2A to 2C. The detailed algorithm of conventional CPE estimation method will be discussed in the following paragraph.

Assumed the useful length of an OFDM symbol is T", the number of sub-carriers is ^, the guard interval (GI) is L, then the modulated symbol in time domain is:

The signal received in the receiver is:

M"⁾-*,^^*00®*,00+/*,^(«) ₍2) Where ^'^ is the CPE, ''"' is the channel response,

P^^n> is the AWGN (Additive white Gaussian noise).

After FFT, the signal in frequency domain is:

Then assumed the local pilot sub-carriers in receiver is k ^ ^~ '"^■' ' , K is the number of pilots used in a symbol. For

_^t_uhe d_^i-scussi■on conveni•ence, we assume H₁'(A:)|' == 1 , w,hi.c,h can simplify the analysis.

As shown in Fig. 2A, the cross-correlation method is engaged in the first type of conventional CPE estimation method. The CPE estimate is carried out by delay circuit and cross-correlation circuit. The delay circuit is used to delay the OFDM symbol to synchronize the error angle calculated by the cross-correlation circuit, and an error angle for the OFDM symbol can be obtained as follows: φ=∞g(P_ke^J^^k) -HXk)-P^*)

=φXk)+φ_H(k) ₍₄₎

This method based on cross-correlation is the simplest. But it can not obtain good performance as it does not consider the channel effect. The CPE estimated results have a term of channel response. Under the multi-path channel or Doppler condition, the CPE estimation results will have great estimated errors since the channel response is not linearity.

In Fig.2B, the auto-correlation method is applied. In the second type of conventional CPE estimation method, the CPE estimate is carried out by delay circuit, RAM and autocorrelation circuit. The delay circuit is used to delay the OFDM symbol to synchronize the error angle calculated by the auto-correlation circuit, RAM is used to store pilot signals of the previous OFDM symbol, and after the auto-correlation calculation, an error angle for the OFDM symbol can be obtained as follows :

Aφ=_aτg(P_ke^^k) -H₁₊Xk)-(P_ke^J^^k) -Hχk)Y)

=axg(P_ke^J^ -H_l÷ι(k)-(P_ke^J^^k) -H₁(Jc)Y)

=φ_ι(k)-φ_i__ι(k) ₍₅₎

φ=φ_ι(k)=φ_i__ι(Jc)+Aφ ₍₆₎

In this auto-correlation method, the difference of PHN of the neighbor symbol is estimated, so the PHN result can be obtained by sum of estimated result.

After CPE removal, compensated OFDM symbol can be obtained as follows :

Ϋ_t(k) = Y₁(Jc) -e^~J*

= X_l(Jc)e^^k)-H_l(k)-e-^J^μ_ι(k)

= X_ι(k)-H_ι(k)+ μ_t(k) ₍₇₎

After CE (channel estimation) , compensated OFDM symbol can be obtained as follows:

X₁(k)=f,(k)-H:(k)

= (Xχk)-Hχk)+_μχk))-H,^*(k)

The auto-correlation method can obtain better performance than the cross-correlation method, since the channel effect can be almost eliminated by auto-correlation operation. But it has more implementation complexity than the cross-correlation method.

In the above mentioned two methods, because information obtained by CE has some difference with the real channel response, there will be some residual phase errors after CE. The two methods before CE are called pre-CE methods. As illustrated in Fig.2C, the conventional two steps method is shown, which can resolve above mentioned problem existing residual phase errors. Besides of CPE removal before CE, the fine CPE removal after CE is engaged. The part before CE is the same as the auto-correlation method (step 1) , and the fine CPE estimation is like to the cross-correlation method (step 2) . The fine removal process after CE will be not affected by channel conditions and it can eliminate the residual phase error

±_n equation (8) . So the two steps method is more accurate than the pre-CE methods.

However, the third method has worse performance than the other two methods under good channel condition as the fine CPE removal part introduces some degradation resulting from the finite word-length effect of fixed data. Therefore, the three methods are only used in some special application environments respectively.

In view of the above, it is needed to provide an auto- adaptive CPE removal method in OFDM system which can perform the CPE removal accurately under all the channel environments .

SUMMARY OF TH₁E INVENTION

An object of the invention is to provide an auto- adaptive CPE removal method based on the Channel Quality Information (CQI) so as to obtain a more accurate digital signal when the channel condition is bad. In addition, the invention can also obtain a better performance when the channel condition is good, so as to be used effectively at any channel condition.

According to one aspect of the invention, provide a digital signal receiver comprising: a first Common Phase Error removal unit (3401) configured to eliminate a first CPE of the digital signal obtained by calculating auto-correlation of the digital signal; a channel quality detector (301) configured to detect quality value of the CPE eliminated digital signal; and a second Common Phase Error removal unit (301-303, 3402) configured to selectively eliminate a second CPE of the CPE eliminated digital signal, according to the quality value.

In one embodiment, the second CPE is obtained by calculating cross-correlation of the CPE eliminated digital signal and pilot signal of the digital signal.

In one embodiment, the second Common Phase Error removal unit (301-303, 3402) further eliminate the second CPE if the quality value is lower than a predetermined value.

According to another aspect of the invention, provide a method for receiving digital signal comprising steps: eliminating a first CPE of the digital signal obtained by calculating auto-correlation of the digital signal; detecting quality value of the CPE eliminated digital signal; and selectively eliminating a second CPE of the CPE eliminated digital signal obtained by calculating cross-correlation of the CPE eliminated digital signal and pilot signal of the digital signal, according to the quality value.

BRIEF DESCRIPTION OF DRAWINGS

Fig. 1 is a block diagram of an example of a OFDM signal receiver in the prior art; Fig. 2A-2C are schematic diagrams illustrating three types of CPE removal method in the conventional OFDM digital signal receiver;

Fig. 3 is block diagram illustrating CPE removal modules of the OFDM digital signal receiver in accordance with the present invention;

Fig. 4 is block diagram illustrating a fine CPE estimation circuit in accordance with the present invention; Fig.5 is a flow chart showing the CPE removal method according to a preferred embodiment of the invention; and

Fig.6 is the experiment results of the CPE removal method and device according to the preferred embodiment of the invention.

DETAIL DESCRIPTION OF PREFERRED EMBODIMENTS

A description will now be given to illustrate many advantages/features of the present invention, according to the preferred embodiments of the present invention. In the attached figures, the like reference numbers indicate the similar elements.

Fig. 3 is block diagram illustrating CPE removal modules of the OFDM digital signal receiver in accordance with the present invention. In the preferred embodiment, CPE estimation is controlled according to the channel quality information, the channel quality information can be determined by the channel types, signal to noise ratio (SNR), and other interference generated before signal demodulation. A description will now be given to illustrate the circuit structure of CPE removal based on the invention, wherein the description for the same blocks as those of Fig.2A-2C will be omitted.

As shown in Fig.3, the coarse-CPE removal module 3401 corresponds to circuit for the step 1 in the two-step method of the prior art. The fine CPE removal is achieved by a delay circuit 302, channel quality detector 301, comparison and selection circuit 303, fine CPE estimation circuit 304 and fine CPE removal module 3402. The delay circuit 302 has the same function as those of the prior art, and used to delay the OFDM symbol to synchronize the error angle calculated by the fine CPE estimation circuit 304. The channel quality detector 301 obtains the quality information of the channel and output a quality value to the comparison and selection circuit 303. Meanwhile, the fine CPE estimation circuit 304 performs the fine CPE estimation by calculating the cross-correlation of the coarse-CPE eliminated OFDM symbol and pilot signal of the OFDM symbol, and output the estimated result to comparison and selection circuit 303. The comparison and selection circuit

303 outputs the error angle obtained by the cross-correlation calculation as the fine CPE correction value, to the fine CPE removal circuit 3402 to eliminate the CPE from the OFDM symbol, when the quality value is lower than a predetermined value. Otherwise, the comparison and selection circuit 303 outputs 0 to the fine CPE removal circuit 3402, that is, there is no error angle to be eliminated from the OFDM symbol. Alternatively, another fine CPE estimation solution can be used. That is, according to the quality value inputted from the channel quality detector 301, the comparison and selection circuit 303 controls the fine CPE estimation circuit 304 to perform the cross-correlation calculation. When the quality value is lower than a predetermined value, the comparison and selection circuit 303 controls the fine CPE estimation circuit

304 to perform the fine CPE estimation by calculating the cross-correlation of the coarse-CPE eliminated OFDM symbol and pilot signal of the OFDM symbol, and output the estimated result as a fine CPE correlation value, to fine CPE removal circuit 3402. Otherwise, the fine CPE estimation circuit 304 does not perform CPE estimation and does not output its result, that is, the function of the fine CPE estimation circuit 304 is bypassed. It is well known that the quality value can be channel types, SNR, and other interference generated before signal demodulation, as long as the quality value can reflect the quality of the channel. In addition, according to the embodiments of the invention, it is easy for one skilled in the art to set a predetermined value of the channel quality, and compare it with the quality value obtained by the channel quality detector 301, as long as the error angle obtained by the fine CPE estimation circuit 304 can be eliminated from the OFDM symbol when the channel quality is bad.

A description will now be given using SNR as an example to illustrate the process for determining the CQI by the channel quality detector.

For example, it is known that SNR is a typical CQI, and the SNR estimation is performed as follows: The pilots in the received OFDM symbol are extracted firstly. Then, the average power P₄ of the pilots is calculated as follows:

Where X is the local pilot, and N_p is the number of the pilots in one OFDM symbol.

Then, The power P_n of the noise is calculated by the differences between the pilots in the received OFDM symbol and the local pilots:

Where Y_pι is the extracted pilot from the received OFDM symbol .

Finally, the SNR can be calculated as follows: S^ = IOlOg₁₀ (P₁ / P₁₁)

A description will now be given to illustrate the process for performing CPE estimation by cross-correlation calculation in the fine CPE estimation circuit. As shown in Fig.4, which is a block diagram illustrating a fine CPE estimation circuit 304 in accordance with the present invention. In the fine CPE estimation circuit 304, a multiplier is used to multiply the received OFDM symbol at the pilot location and the local pilot sequence, and then the products in an OFDM symbol are accumulated by an adder. At last, an arctangent circuit is used to calculate the phase of the sum of the accumulation. The phase (error angle) is the common phase error calculated by the fine CPE estimation circuit 304 and outputted to the fine CPE removal circuit 3402.

Referring to Fig.5, the method for estimating the CPE is described. Fig. 5 is a flow chart showing the method according to a preferred embodiment of the invention.

At step SlOl, OFDM symbol after coarse-CPE estimation and removal is received, that is, the OFDM symbol has been processed via the step 1 of the method in the prior art. Then, at step S102, the channel quality information of the OFDM symbol is determined by the channel quality detector. At this step, a channel quality value that can be used to compare with a predetermined threshold is obtained. At step S103, the obtained channel quality value is compared with the predetermined threshold, and then at step S104, when the channel quality value is lower than the predetermined threshold, the process goes ahead to step S105, the fine CPE estimation is performed. Otherwise, when the channel quality value is beyond the predetermined threshold, the process is end.

A variety of simulation experiments are made to assess the performance of the present invention, the conditions are set with reference to the Table 1 as follows.

Table 1. Conditions of simulations

The experiment results for the above channel models can be shown in Fig.6. From those simulation results, we can find that there are cross points between the two steps method and the coarse-CPE methods in the BER (bit error ratio) curves. When the SNR is smaller than that of the cross points, performance of the two steps method is better. The coarse-CPE methods achieve lower BER when the SNR is bigger than that of the cross points. Therefore, At the SNR below the cross point, the method based on the invention will operate complying with the curve of the two steps method, and it will comply with the curve of the auto-correlation method at the SNR above the cross point.

The cross point is decided by the channel type and the fixed-point precision of fine CPE estimation processing. Under the above simulation environment, the phase angle being the 13 bits fixed-point processing and the sine & cosine being the 10 bits fixed-point processing, the switch point is as figure 6, about 23.2~23.3dB. As described above, the method and device in the invention can obtain better effect whatever the channel quality is, and the accurate CPE correction value can be obtained. So the invention can adapt to much more application environments than the conventional methods.

The foregoing merely illustrates the principles of the invention and it will thus be appreciated that those skilled in the art will be able to devise numerous alternative arrangements which, although not explicitly described herein, embody the principles of the invention and are within its spirit and scope. It is to be understood that numerous modifications may be made to the illustrative embodiments and that other arrangements may be devised without departing from the spirit and scope of the present invention as defined by the appended claims .

Claims

1. A digital signal receiver comprising: a first Common Phase Error removal unit (3401) configured to eliminate a first CPE of the digital signal obtained by calculating auto-correlation of the digital signal; a channel quality detector (301) configured to detect quality value of the CPE eliminated digital signal; and a second Common Phase Error removal unit (301-303, 3402) configured to selectively eliminate a second CPE of the CPE eliminated digital signal, according to the quality value.

2. The digital signal receiver according to claim 1, wherein the second CPE is obtained by calculating cross-correlation of the CPE eliminated digital signal and pilot signal of the digital signal.

3. The digital signal receiver according to claim 1 or 2, wherein the second Common Phase Error removal unit (301-303,

3402) further eliminate the second CPE if the quality value is lower than a predetermined value.

4. The digital signal receiver according to claim 1 or 2, further comprises: a channel estimation unit (150) configured to equalize the first CPE eliminated digital signal and then provide the equalized and first CPE eliminated digital signal to the channel quality detector (301) and the second Common Phase Error removal unit (301-303, 3402) .

5. The digital signal receiver according to claim 1 or 2, wherein the second Common Phase Error removal unit (301-303, 3402) further comprises: a fine CPE estimation unit (304) configured to calculate cross-correlation of the first CPE eliminated digital signal and pilot signal of the digital signal, a compare selection unit (303) configured to output the calculated result as the second CPE if the quality value is lower than a predetermined value, and a fine CPE removal unit (3402) configured to further eliminate the second CPE from the first CPE eliminated digital signal .

6. The digital signal receiver according to claim 1 or 2, wherein the second Common Phase Error removal unit (301-303,

3402) further comprises: a comparison and selection unit (303) configured to output a control value for determining to calculate the second

CPE, if the quality value is lower than a predetermined value, a fine CPE estimation unit (304) configured to calculate cross-correlation of the first CPE eliminated digital signal and pilot signal of the digital signal, and a fine CPE removal unit (3402) configured to further eliminate the second CPE obtained from the fine CPE estimation unit (304) from the first CPE eliminated digital signal.

7. A method for receiving digital signal comprising steps: eliminating a first CPE of the digital signal obtained by calculating auto-correlation of the digital signal; detecting quality value of the CPE eliminated digital signal; and selectively eliminating a second CPE of the CPE eliminated digital signal, according to the quality value.

8. The method according to claim 7, wherein the second CPE is obtained by calculating cross-correlation of the CPE eliminated digital signal and pilot signal of the digital signal .

9. The method according to 7 or 8, wherein the step of selectively eliminating comprises further eliminating the second CPE if the quality value is lower than a predetermined value .

10. The method according to claim 7 or 8, further comprises equalizing the first CPE eliminated digital signal before the steps of detecting and selectively eliminating.

11. >The method according to claim 7 or 8, wherein the step of selectively eliminating comprises: calculating cross-correlation of the first CPE eliminated digital signal and pilot signal of the digital signal, outputting the calculated result as the second CPE if the quality value is lower than a predetermined value, and further eliminating the second CPE from the first CPE eliminated digital signal.

12. The method according to claim 7 or 8, further comprises

Outputting a control value for determining to calculate the second CPE, if the quality value is lower than a predetermined value, calculating cross-correlation of the first CPE eliminated digital signal and pilot signal of the digital signal, and further eliminating the second CPE obtained the step of calculating from the first CPE eliminated digital signal.