CN101001235A - Time synchronous and frequency synchronous method of weakly energy parallel PN sequence - Google Patents

Abstract

This invention provides a time and frequency synchronization method for weak energy parallel PN sequences, which overlaps weak energy PN sequences on a base band time domain signal modulated by OFDM and applies the PN sequences to demodulate the received signal at the receiving end combining related information protecting intervals to generate synchronous signals to get time and frequency synchronization, in which, the energy distributed to the PN sequence can be strong or weak and is realized easily.

Description

A kind of weak energy parallel PN sequence time synchronized and frequency synchronization method

Technical field:

The present invention relates to a kind of weak energy parallel PN sequence time synchronized and frequency synchronization method, belong to digital information transmission technical field, particularly digital television broadcasting system, single carrier ofdm communication system, multi-carrier OFDM communication system, wireless lan (wlan) etc. adopt the simultaneous techniques in the OFDM modulated digital communication system.Especially in digital terrestrial television (DVB-T) system, be used to improve the travelling performance of system.

Background technology:

OFDM modulation (OFDM) is very high because of its channel utilization, have good anti-fading ability to be widely used.Yet ofdm system is very responsive to timing error, and therefore, simultaneous techniques is one of key technology of this system.

Conventional method for synchronous can be divided into: protection is method for synchronous, pilot synchronization method, training sequence method for synchronous and blind synchronization method four classes at interval.

It is simple to carry out at interval synchronous algorithm by protection, and operand is little, and better performances under the awgn channel condition still, is difficult for guaranteeing that precise time is synchronous in the high-speed mobile multidiameter fading channel.If the time error of time synchronized surpasses protection at interval, also can destroy the orthogonality between subcarrier.And Frequency Synchronization and time synchronized are carried out simultaneously, though realize simply being subject to the influence of time synchronization error.

Utilizing pilot tone to obtain synchronizing signal is a good method, but these pilot tones in use are restricted because having taken valuable system bandwidth resource.And, both made in the DVB-T system, adopt pilot tone and protection to unite at interval and carry out synchronously, its performance is also undesirable.Generally when translational speed surpassed 100km/h, DVB-T will go wrong synchronously.Though the mobile receptivity of DVB-T system is also attempted to improve in Europe always,, still there is not disclosed correlation technique report so far.So the net synchronization capability that how to improve the DVB-T system under high-speed mobile environment is still the difficult point and the emphasis of academia's research.

Utilizing the PN sequence to carry out in the synchronous method, the applied PN sequence of most of method all with the signal serial arrangement, the method for synchronous that PN sequence and signal parallel are arranged is seldom seen.Wherein, all there is a bigger weakness in the serial PN sequence synchronized algorithm of better performances, that is: synchronizing signal platform occurred near the zero time offset place, rather than spike, even document Fredrik is Tufvesson, Mike Faulkner, Peter Hoeher and Ove Edfors, OFDM time and frequencysynchronization by spread spectrum pilot technique, IEEE Conference, 1999. method more excellent relatively, and document Hlaing Minn, Vijay K.Bhargava, and Khaled Ben Letaief, A robust timing and frequencysynchronization for OFDM systems, IEEE Trans.on wireless communications, vol.2, No.4, July2003, pp:822-838. and document Byungjoon Park, Hyunsoo Cheon, Changeon Kang and Daesik Hong, Anovel timing estimation method for OFDM systems, IEEE Communications letters, vol.7, No.5,, May2003, pp:239-241. also done deep discussion, but final effect is all unsatisfactory at this problem.And the PN sequence has taken valuable bandwidth resources in the serial PN sequence synchronized algorithm, only could be suitable in having systems such as the 4G of special-purpose synchronizing symbol, WLAN, then cannot use for the DVB-T system.For the parallel PN sequence synchronized algorithm, because the PN sequence is superimposed upon the interference that has caused on the signal signal, though Liying Song, Youxi Tang, Hongzhi Zhao, ShaoqianLi, etc, Optimum partial interference canceling factor pilot assistant coherent OFDM system withpartial power training sequences synchronization, The 14 ^ThIEEE 2003 international symposium onPersonal, Indoor and Mobile Radio Communication Proceedings, pp:1906-1910. proposed a kind ofly to eliminate the method that part is disturbed, but this method has implied a condition: promptly to the change of receiver structure; And document FredrikTufvesson, Mike Faulkner, Peter Hoeher and Ove Edfors, OFDM time and frequencysynchronization by spread spectrum pilot technique, IEEE Conference, 1999. and document SteingaB.A, etc, Frame synchronization using superimposed sequence, Proc.IEEE ISIT ' 97, ULM, Germany, 1997, pp:489. adopting at different synchronous phases selects for use the different-energy factor can realize really synchronously with the method for adjusting stack PN sequence energy, but caused the influence to systematic function, especially needed under high-speed mobile environment fast synchronously, this method more is not suitable for.

It is poor that blind synchronization method performance Billy carries out synchronization performance with the sequence of priori.And compare with regard to adopting protection interval, pilot tone and PN sequence to carry out synchronization performance, it is better that employing PN sequence is carried out net synchronization capability.So, in actual engineering, generally select for use net synchronization capability preferably, algorithm performance quick PN sequence method for synchronous more stable, that computation complexity is lower.But the synchronized algorithm that will be satisfied these conditions simultaneously is very difficult.

Summary of the invention:

The objective of the invention is:, propose a kind of computation complexity quick time synchronized and frequency synchronization method low, that algorithm performance is stable at the deficiencies in the prior art.This method not only can satisfy all orthogonal FDM communication systems in the synchronous requirement of high-speed mobile environment, and the sequence of being utilized does not take valuable system bandwidth resource.Especially when realizing time synchronized and Frequency Synchronization, the training sequence of being introduced can be ignored the influence of systematic function.

For achieving the above object, technical scheme of the present invention is: at transmitting terminal, and for the subcarrier base band time domain signal after the OFDM modulation, the PN sequence of the weak energy of stack on information symbol therein, treated formation transmits; Produce synchronizing signal at receiving terminal by adopting the PN sequence to carry out despreading to received signal, obtain time synchronized and Frequency Synchronization.

Concrete operations step of the present invention is as follows:

Step 1PN sequence is superimposed upon on the ofdm signal in time domain, is assigned with extremely weak energy in this process.The gross power of PN sequence and signal is carried out normalized, and the power that the PN sequence is distributed is ρ.For sub-carrier number is N _cOfdm system, symbol lengths is T _sIn awgn channel, baseband receiving signals can be expressed as:

$r[k-\mathrm{\τ}]=s[k-\mathrm{\τ}]+n[k-\mathrm{\τ}]$

$=(\sqrt{\mathrm{\ρ}}{\mathrm{\σ}}_{s}m[k-\mathrm{\τ}]+\sqrt{1-\mathrm{\ρ}}{\mathrm{\σ}}_{s}d[k-\mathrm{\τ}])e^{j(\frac{2\mathrm{\π\ϵk}}{N_c}+\mathrm{\θ})}+n[k-\mathrm{\τ}]---\left(1\right)$

Wherein, m[k] be the PN sequence of transmission, d[k] be the OFDM data sequence, n[k] be white complex gaussian noise, θ is a carrier phase, and ε is a normalization frequency deviation to be estimated, and τ is a timing position to be estimated.PN sequence c[k] the energy ρ that distributed a little less than.Transmit and the power of noise is respectively σ _s ²And σ _n ²

What is particularly worth mentioning is that, in this patent, the energy ρ that the PN sequence is distributed can be very a little less than, though weak be to be added on the signal to this parallel sequence, can ignore the influence of systematic function; Simultaneously, the energy ρ that the PN sequence is distributed also can be more intense, by force to utilizing this parallel PN sequence can obtain good channel estimating performance, thereby can eliminate the influence of parallel PN sequence to systematic function well.This is to a core concept of PN sequence energy distribution in this patent.

Step 2 received signal has by the despreading of local PN sequence:

$m^{*}\left[k\right]r[k-\mathrm{\τ}]=\sqrt{\mathrm{\ρ}}{\mathrm{\σ}}_s{m}^{*}\left[k\right]m[k-\mathrm{\τ}]e^{j(\frac{2\mathrm{\π\ϵk}}{N_c}+\mathrm{\θ})}+\sqrt{1-\mathrm{\ρ}}{\mathrm{\σ}}_s{m}^{*}\left[k\right]d[k-\mathrm{\τ}]e^{j(\frac{2\mathrm{\π\ϵk}}{N_c}+\mathrm{\θ})}+n^{\′}[k-\mathrm{\τ}]---\left(2\right)$

For time synchronized, suppose that two slide relative position τ between the sequence are integer.In order to carry out in formula (2), can regarding noise jamming as for last two synchronously.

At whole synchronous phase, the energy of PN sequence all only accounts for the part of gross energy.Be almost 1 at acquisition phase by the energy of PN sequence unlike other conventional method, and its energy is reduced to a degree that data-signal is not exerted an influence at tracking phase, this is the art of this patent and document Fredrik Tufvesson, Mike Faulkner, Peter Hoeher and Ove Edfors, OFDM time and frequency synchronization by spread spectrum pilot technique, IEEEConference, the difference of conventional method maximums such as 1999..

Step 3 is utilized the good autocorrelation of PN sequence, at receiving terminal the signal that receives and known PN sequence is done relevantly, can produce following signal:

${\mathrm{cor}}_1[k,\mathrm{\θ}]=\underset{l=0}{\overset{L-1}{\mathrm{\Σ}}}\left[\begin{array}{c}\left(\underset{n=0}{\overset{R-1}{\mathrm{\Σ}}}{m}^{*}[k-n-\mathrm{lR}-\mathrm{\θ}]r[k-n-\mathrm{lR}]\right)\\ \·\left(\underset{n=0}{\overset{R-1}{\mathrm{\Σ}}}{m}^{*}[k-n-\mathrm{lR}-\mathrm{\θ}]r[k-n-(l+R){{R]}^{*}}^{}\right)\end{array}\right]---\left(3\right)$

Wherein, m[k] be known PN sequence, R is a correlation length, P is relevant with estimated frequency deviation region and precision, the number of repetition of known PN sequence when L is the structure training sequence, m ^*[.] expression is to m[.] get conjugation.

Step 4 is in order to improve the performance of frame synchronization, and the existence of Cyclic Prefix has formed N _gOrganize the N of being separated by _cSampling right, it can obtain following signal at receiving terminal:

${\mathrm{cor}}_2[k,\mathrm{\θ}]=\underset{m=\mathrm{\θ}}{\overset{\mathrm{\θ}+N_g-1}{\mathrm{\Σ}}}{r}^{*}[m-\mathrm{\θ}]r[m-\mathrm{\θ}-N_c]---\left(4\right)$

Step 5 after merging the PN sequence and protecting correlation at interval, can obtain time synchronized in order to strengthen the synchronizing signal energy by (5):

$\widehat{\mathrm{\θ}}=\arg \left\{\underset{\mathrm{\θ}}{\max }[{\mathrm{cor}}_1[k,\mathrm{\θ}]+{\mathrm{cor}}_2[k,\mathrm{\θ}]]\right\}---\left(5\right)$

Step 6 is obtaining on the synchronizing signal basis, and according to the maximal possibility estimation principle, the phase place of the signal that is produced by formula (5) can be tried to achieve big frequency offset estimating, has:

${\widehat{\mathrm{\ϵ}}}_b=-\arg \left({\mathrm{Cor}}_1[k,0]\right)\·N_c/\left(2\mathrm{\πR}\right)---\left(6\right)$

Step 7 is after proofreading and correct big frequency deviation, and decimal frequency bias can be obtained by the phase place of formula (4) signal, has:

${\widehat{\mathrm{\ϵ}}}_s=\arg \left({\mathrm{Cor}}_2[k,0]\right)/\left(2\mathrm{\π}\right)---\left(7\right)$

The present invention has the PN sequence of only distributing less energy of best autocorrelation performance at the transmitting terminal of ofdm communication system by superposeing in time domain, can realize the synchronous and Frequency Synchronization of high-speed moving state DVB-T 2K system time, thereby improve the travelling performance of system.

Be with the difference of prior art maximum: the first, PN sequence and signal parallel, that is: the PN sequence is superimposed upon on the signal; The second, Die Jia sequence (as: Barker code, chaos sequence etc.) is only distributed the energy (institute's distribute energy can be more weak energy, also can be more intense energy) of the part of gross energy, thereby different with conventional method.Three, the training sequence in this patented technology can be m sequence, chaos sequence etc.; Four, synchronized algorithm difference; Five, can be used for realizing that DVB-T 2K system realizes the time synchronized and the Frequency Synchronization of high-speed mobile environment, and conventional algorithm all can not be too high for DVB-T 2K system translational speed.

The present invention is applicable to that all adopt OFDM to carry out carrying out time synchronized and Frequency Synchronization in the communication system for modulation, especially are applicable in European digital terrestrial television (DVB-T) system.Though technology of the present invention is primarily aimed at the system of frequency division multiple access, but any engineer with knowledge background such as signal processing, communications, can design corresponding frame synchronization or time synchronism apparatus at code division multiple access and time division multiple access according to the present invention, it all should be included in thought of the present invention and the scope.

Description of drawings:

Fig. 1 is conventional ofdm system block diagram.

Fig. 2 is the ofdm system block diagram that comprises the art of this patent.As shown in the figure, after the position that transmitting terminal adds training sequence is OFDM modulation and IFFT, before up-sampling and the molding filtration; After the position that receiving terminal adds lock unit is down-sampling and molding filtration, before FFT and the OFDM demodulation.

Fig. 3 is the schematic diagram of stack PN sequence on information symbol.

Fig. 4 is a art of this patent synchronized algorithm block diagram.

Embodiment:

Below by concrete enforcement technical scheme of the present invention is further described.

Apply the present invention to ofdm system.This system adopts the QPSK modulation, and emulation is all based on DVB-T 2K system's awgn channel and Rayleigh fading channel.The major parameter of emulation has: suppose that frequency shift (FS) is 3.6, translational speed is 150km/h.Concrete steps are:

1, transmitting terminal is sent into the OFDM base band signal modulated, and stack PN sequence on information symbol therein, and formation transmits.

2, obtain received signal at receiving terminal, and carry out demodulation to received signal, obtain basic demodulation cell signal according to formula (2).

3, produce synchronizing signal according to formula (3) and formula (4), and detect, obtain time synchronized by formula (5) signal being carried out peak value.

4, obtaining on the synchronizing signal basis, according to the maximal possibility estimation principle, the phase place of the signal that produces by formula (5), through type (6) is tried to achieve big frequency deviation.

5, after big frequency deviation is proofreaied and correct, decimal frequency bias can be by the phase place of formula (4) signal, and through type (7) obtains.Then, little frequency deviation is proofreaied and correct.

Claims (4)

1, the present invention relates to a kind of weak energy parallel PN sequence time synchronized and frequency synchronization method, it is characterized in that comprising the steps:

Step 1PN sequence is superimposed upon on the ofdm signal in time domain, and the PN sequence is assigned with certain energy in this process.The gross power of PN sequence and signal is designated as 1, and the power of PN sequence allocation is ρ.For sub-carrier number is N _cOfdm system, symbol lengths is T _sIn awgn channel, baseband receiving signals can be expressed as:

$r[k-\mathrm{\τ}]=s[k-\mathrm{\τ}]+n[k-\mathrm{\τ}]$

$=\left(\sqrt{\mathrm{\ρ}}{\mathrm{\σ}}_{s}m\right[k-\mathrm{\τ}]+\sqrt{1-\mathrm{\ρ}}{\mathrm{\σ}}_{s}d[k-\mathrm{\τ}\left]\right)e^{j(\frac{2\mathrm{\π\ϵk}}{N_c}+\mathrm{\θ})}+n[k-\mathrm{\τ}]---\left(1\right)$

Wherein, m[k] be the PN sequence of transmission, d[k] be the OFDM data sequence, n ' [k] is a white complex gaussian noise, and θ is a carrier phase, and ε is a normalization frequency deviation to be estimated, and τ is a timing position to be estimated.PN sequence c[k] the energy ρ that distributed very a little less than.Transmit and the power of noise is respectively σ _s ²And σ _n ²

Step 2 received signal has by the despreading of local PN sequence:

$m^{*}\left[k\right]r[k-\mathrm{\τ}]=\sqrt{\mathrm{\ρ}}{\mathrm{\σ}}_s{m}^{*}\left[k\right]m[k-\mathrm{\τ}]e^{j(\frac{2\mathrm{\π\ϵk}}{N_c}+\mathrm{\θ})}+\sqrt{1-\mathrm{\ρ}}{\mathrm{\σ}}_s{m}^{*}\left[k\right]d[k-\mathrm{\τ}]e^{j(\frac{2\mathrm{\π\ϵk}}{N_c}+\mathrm{\θ})}+n^{\′}[k-\mathrm{\τ}]---\left(2\right)$

For time synchronized, suppose that two slide relative position τ between the sequence are integer.In order to carry out in formula (2), can regarding noise jamming as for last two synchronously.

Step 3 is utilized the good autocorrelation of PN sequence, and receiving terminal is done the signal that receives and known PN sequence relevant, can produce following signal:

${\mathrm{cor}}_1\left[k,\mathrm{\θ}\right]=\underset{l=0}{\overset{L-1}{\mathrm{\Σ}}}\left[\begin{array}{c}\left(\underset{n=0}{\overset{R-1}{\mathrm{\Σ}}}{m}^{*}\right[k-n-\mathrm{lR}-\mathrm{\θ}\left]r\right[k-n-\mathrm{lR}\left]\right)\\ \·{\left(\underset{n=0}{\overset{R-1}{\mathrm{\Σ}}}{m}^{*}\right[k-n-\mathrm{lR}-\mathrm{\θ}\left]r\right[k-n-(l+P)R\left]\right)}^{*}\end{array}\right]---\left(3\right)$

Wherein, m[k] be known PN sequence, R is a correlation length, P is relevant with estimated frequency deviation region and precision, the number of repetition of known PN sequence when L is the structure training sequence, m ^*[.] expression is to m[.] get conjugation.

Step 4 is in order to improve the performance of frame synchronization, and the existence of Cyclic Prefix has formed N _gOrganize the N of being separated by _cSampling right, it can obtain following signal at receiving terminal:

${\mathrm{cor}}_2[k,\mathrm{\θ}]=\underset{m=\mathrm{\θ}}{\overset{\mathrm{\θ}+N_g-1}{\mathrm{\Σ}}}{r}^{*}[m-\mathrm{\θ}]r[m-\mathrm{\θ}-N_c]---\left(4\right)$

Step 5 after merging the PN sequence and protecting correlation at interval, can obtain time synchronized in order to strengthen the synchronizing signal energy by (5):

$\widehat{\mathrm{\θ}}=\arg \left\{\underset{\mathrm{\θ}}{\max }\right[{\mathrm{cor}}_1[k,\mathrm{\θ}]+{\mathrm{cor}}_2[k,\mathrm{\θ}]\left]\right\}---\left(5\right)$

Step 6 is obtaining on the synchronizing signal basis, and according to the maximal possibility estimation principle, the phase place of the signal that is produced by formula (5) can be tried to achieve big frequency offset estimating, has:

${\widehat{\mathrm{\ϵ}}}_b=-\arg \left(\mathrm{Co}{r}_1\right[k,0\left]\right)\·N_c/\left(2\mathrm{\πR}\right)---\left(6\right)$

Step 7 is after proofreading and correct big frequency deviation, and remaining decimal frequency bias can be obtained by the phase place of formula (4) signal, has:

${\widehat{\mathrm{\ϵ}}}_s=\arg \left({\mathrm{Cor}}_2\right[k,0\left]\right)/\left(2\mathrm{\π}\right)---\left(7\right)$

2, said as claim 1, a kind of weak energy parallel PN sequence time synchronized and frequency synchronization method is characterized in that: to the energy that distributed of PN sequence of stack can be a little less than, also can be stronger.

3, said as claim 1, a kind of weak energy parallel PN sequence time synchronized and frequency synchronization method is characterized in that: selected sequence can be PN sequence with good autocorrelation, chaos sequence, multiple Barker code etc.

4, said as claim 1, a kind of weak energy parallel PN sequence time synchronized and frequency synchronization method is characterized in that: not only can be used for the OFDM communication system for modulation, and can be used in CDMA, the tdma system.Be particularly useful for improving the net synchronization capability of DVB-T system.

Images