CN101902435B - Downlink frequency deviation estimation method based on IEEE (Institute of Electrical and Electronics Engineers) 802.16e communication standard - Google Patents

Abstract

The invention discloses a downlink frequency deviation estimation method based on an IEEE 802.16e communication standard, wherein a frequency deviation value can be more accurately obtained by comparing the phase position of a receiving-end pseudo-repetitive sequence correlation value with the phase position of a transmitting-end pseudo-repetitive sequence correlation value; meanwhile, the invention is insensitive to timing synchronism deviation and prevents the performance decrease of a frequency deviation estimation algorithm caused by imprecise timing synchronism. In the invention, the influence of a channel response (multidiameter delay) on the frequency deviation estimation algorithm is eliminated by adopting an iterative algorithm, and under a general condition, a stable and accurate frequency deviation estimation value can be obtained by only adopting iteration for one time, and the complexity is not high.

Description

Down link frequency deviation estimating method based on IEEE 802.16e communication standard

Technical field

The present invention relates to a kind of down link frequency deviation estimation approach based on IEEE 802.16e communication standard.

Background technology

In today of mobile communication develop rapidly; WiMAX (Worldwide Interoperability for Microwave Access; Micro-wave access global inter communication) technology is just receiving the increasing concern of industry as a kind of wireless access wide band technology towards metropolitan area network.

The WiMAX technology is the basis with 802.16 series standards of IEEE; Difference according to frequency range has different physical-layer techniques corresponding with it respectively; Be SC (Signal Carrier, single carrier), OFDM (Orthogonal Frequency Division Multiplex; OFDM) and OFDMA (Orthogonal Frequency Division Multiple Access, OFDM).Because OFDM and OFDMA have the higher availability of frequency spectrum, and have the ability of the weak and narrow band interference of good opposing multipath effect, frequency selectivity, they are core technologies of WiMAX physical layer.In the 802.16e standard of formally passing through by IEEE in December, 2005; OFDMA is gradable; Support 2048 points, 1024 points, and 128 FFT (fast Fourier transform) at 512, can adapt to channel width difference, make mobility become possibility from 1.25～20MHz.Fixedly single user system before being different from exists a lot of difficult problems in mobile environment, for example switching, presence of intercell interference and problem such as synchronous.

Ofdm system relies on mutually orthogonal subcarrier transmission data, and its orthogonality is to depend on adjacent sub-carrier frequencies interval delta f=1/T _sGuarantee.If there is frequency shift (FS), then separating timing, at T _s(OFDM symbol period) at interval in any certain number of sub-carrier of demodulator will non-orthogonal with other subcarriers (inner product be 0), thereby can produce the interference between subcarrier.Therefore ofdm system is very responsive to frequency shift (FS), and the frequency shift (FS) meeting is destroyed the orthogonality between the subcarrier, causes the degradation of system.Frequency shift (FS) is to follow the deviation between the local carrier of mobile terminal by the base station, and the Doppler frequency shift of channel etc. causes.The carrier frequency synchronization error of IEEE 802.16e standard-required mobile terminal and base station subcarrier spacing 2% in, this has just proposed higher requirement to the precision of frequency offset estimating.

In the WiMAX of reality system; Need carry out synchronous work, its detailed process is as shown in Figure 1: the signal of receiving terminal at first carries out the synchronous of frame head through after the analog to digital conversion; The original position of specified data; Carry out fractional part of frequency offset then and estimate and compensation, then carry out OFDM demodulation (being the FFT conversion), after FFT, carry out integer frequency offset and estimate and compensation.

The down-link frequencies stationary problem of IEEE 802.16e standard.

The method of the existing WiMAX of being used for frequency offset estimating has the blind estimating method based on CP (Cyclic Prefix) structure.Cyclic Prefix is the last T with the signal waveform of each OFDM symbol _gWaveform copy in time is on the idle protection position at interval to the front originally.Realize for IFFT (inverse fast Fourier transform), exactly several last sample values are copied to the front, form prefix.If r (n) is for receiving signal; N is a FFT length; L is the length of CP, for the estimated value of fractional part of frequency offset is:

$\mathrm{\λ}\left(\mathrm{\θ}\right)=\underset{k=\mathrm{\θ}}{\overset{\mathrm{\θ}+L-1}{\mathrm{\Σ}}}r\left(k\right)r*(k+N)---\left(1\right)$

$\widehat{\mathrm{\θ}}=\underset{\mathrm{\θ}}{\arg }\max \left\{\right|\mathrm{\λ}\left(\mathrm{\θ}\right)\left|\right\}---\left(2\right)$

$\widehat{\mathrm{\ϵ}}=\frac{1}{2\mathrm{\π}}\mathrm{angle}\left(\mathrm{\λ}\left(\widehat{\mathrm{\θ}}\right)\right)---\left(3\right)$

The advantage of this method is simple and convenient, and can in each OFDM symbol, carry out frequency offset estimating.But it has an important disadvantages, because the influence of multipath channel, CP can receive the pollution of previous OFDM symbol; And under the very big situation of multidiameter delay; Might all can be polluted by whole C P, in this case, will be no longer suitable based on the frequency deviation estimating method of CP structure.

Also have reliable associating time synchronized and Algorithm of Carrier Frequency Offset Estimation by a kind of synchronizing symbol based on two sections repetitions of Timothy M.Schmidl and Donald C.Cox proposition about the classic algorithm of frequency offset estimating, more existing frequency deviation estimating methods have been used their thought.

The previous version IEEE 802.16-2004 of WiMAX for example, the frame head of employing value in frequency domain can be used following formulate:

$P_k=\left\{\begin{array}{cc}\sqrt{2}{P}_{\mathrm{ALL}}\left(k\right)& k\mathrm{<figure-callout\; id="2"\; label="mod"\; filenames="HDA0000024470630000021.png,HDA0000024470630000032.png"\; state="\{\{state\}\}">mod</figure-callout>}2=0\\ 0& k\mathrm{<figure-callout\; id="2"\; label="mod"\; filenames="HDA0000024470630000021.png,HDA0000024470630000032.png"\; state="\{\{state\}\}">mod</figure-callout>}2\&NotEqual;0\end{array}\right.---\left(4\right)$

Wherein Relevant with the 3dB gain, P _ALLFrequency domain sequence is stipulated by agreement.

The header signal that sends like this will be two sections of repetition in the duration at a symbol of time domain, because the existence of CP, as long as the multidiameter delay of channel is less than CP length, the header signal that then receives at receiving terminal still is two sections of repetition in time domain.Just can obtain frequency deviation value at receiving terminal like this through following method:

If N is a FFT length; is symbol timing estimation value,

be by the normalized frequency offset estimating value of subcarrier spacing

$\mathrm{\&lambda;}\left(\mathrm{\&theta;}\right)=\underset{k=\mathrm{\&theta;}}{\overset{\mathrm{\&theta;}+\frac{N}{2}-1}{\mathrm{\&Sigma;}}}r\left(k\right)r*(k+\frac{N}{2})---\left(5\right)$

$\widehat{\mathrm{\&theta;}}=\underset{\mathrm{\&theta;}}{\arg }\max \left\{\right|\mathrm{\&lambda;}\left(\mathrm{\&theta;}\right)\left|\right\}---\left(6\right)$

$\widehat{\mathrm{\&epsiv;}}=\frac{1}{\mathrm{\&pi;}}\mathrm{angle}\left(\mathrm{\&lambda;}\left(\widehat{\mathrm{\&theta;}}\right)\right)---\left(7\right)$

Can be known that by top formula the scope of the frequency deviation that this method is estimated is [Δ f, Δ f], wherein, Δ f is meant subcarrier spacing.

Yet change has taken place in the frame head structure of IEEE 802.16e standard.In mobile environment, each sub-district is divided into 3 sectors, and different Preamble sub-carrier set is used according to formula (5) in each sector:

Preamble_Carrier_Set _n＝n+3k (8)

Wherein n is a sector number, n ∈ 0,1,2},

N _uBe useful sub-carrier number.

Because FFT count (sub-carrier number) can not be divided exactly by sector number 3, so the time domain sequences of frame head do not have the characteristic of cycle repetition of stricti jurise, and just approximate to a certain extent have three sections repeatability (characteristic pseudoperiod), as shown in Figure 2.

Because the frame head time-domain signal that transmitting terminal sends is not to repeat in the strict cycle; And through after the channel; The signal that receives at receiving terminal does not just more possess strict characteristic of cycle repetition; Therefore conventional frequency offset estimating algorithm application will cause bigger error on IEEE 802.16e standard, have a strong impact on the performance of system.

Summary of the invention

Frame head time domain sequences to above-mentioned IEEE 802.16e standard does not possess strict repeat property; But have the problem of the characteristic that repeats three periods pseudoperiods to a certain extent, the invention provides a kind of down link frequency deviation estimation approach based on IEEE 802.16e standard.

For reaching above purpose; The present invention utilized frame head three sections pseudoperiod repeat property; And the algorithm that adopts iteration eliminates the influence of multipath channel to frequency offset estimation result, provides effectively a kind of and method easily for estimating frequency deviation accurately, comprises the steps:

(a) the pseudo-repetition of the frame head time domain sequences of base station end emission is s (n)=IFFT (X (n)); Wherein X (n) is the lead code frequency domain sequence, and IFFT counts and is N, gets T=pseudoperiod [N/3] of frame head; Wherein [] expression round off criterion rounds operation to the number in [], according to formula The correlation of calculating between the pseudo-repetitive sequence of the frame head of transmitting terminal is with this correlation R _tPhase place as transmitter, phase θ _t

(b) in the mobile terminal to received signal r (n) to carry out timing synchronous, regularly synchronous frame head position is n ₀, according to formula

Calculate the correlation of the pseudo-repetitive sequence of frame head of receiving terminal, with this correlation R _rPhase place as receiving phase θ _r

(c) through comparing receiving phase θ _rWith transmitter, phase θ _tBetween difference, according to formula

Calculate the frequency deviation value of little several times

Span be [1.5,1.5];

(d) according to the fractional part of frequency offset value calculated to received signal r (n) carry out the fractional part of frequency offset compensation; Then the signal after the fractional part of frequency offset compensation is carried out the FFT conversion; Convert frequency domain sequence Y (n) to, Y (n) and frame head lead code frequency domain sequence X (n) carried out computing cross-correlation:

$C\left(m\right)=\underset{n=0}{\overset{N-1}{\mathrm{\&Sigma;}}}X(n+m)*Y^{*}\left(n\right),$ -(N-1)≤m≤(N-1)

Peak m according to cross correlation value _MaxObtain the integer frequency offset value;

(e) signal after the fractional part of frequency offset compensation is carried out the integer frequency offset compensation; Then the signal after the compensate of frequency deviation is carried out channel estimating, obtain frequency domain channel estimated value

channel estimating and adopt the channel estimation methods of existing band pilot tone;

(f) according to channel estimation value

Revise transmitter, phase θ _t, make previous formula

Become Wherein

θ _tRevise correlation R for this reason _tPhase place;

(g) adopt iterative algorithm: repeat (c), (d), (e), step (f) is up to the frequency offset estimating value stabilization.

In the said method, in the said step (e), the channel estimation methods of existing band pilot tone is LS channel estimating, MMSE channel estimating or its improved method.

The invention has the beneficial effects as follows:

1. three sections characteristics that repeat pseudoperiod of frame head time domain sequences have been utilized fully.Because be not to repeat in the strict cycle, therefore the phase place of the correlation between the pseudo-repetitive sequence is not 0, therefore the frequency deviation estimating method by routine is to obtain frequency deviation value accurately.The present invention passes through to compare the phase place of the pseudo-repetitive sequence correlation of receiving terminal and the phase place of the pseudo-repetitive sequence correlation of transmitting terminal, can obtain frequency deviation value more exactly; Simultaneously, the present invention is insensitive to the timing synchronism deviation, prevents to cause owing to the synchronous inaccuracy of timing the decline of frequency offset estimating algorithm performance.

2. because the frame head time domain sequences is not strict the repetition, so channel response can change the correlation of pseudo-repetitive sequence, thereby the result of frequency offset estimating is impacted.The present invention takes the algorithm of iteration to eliminate the influence of channel response (multidiameter delay) to the frequency offset estimating algorithm.And, generally speaking, only need take an iteration just can obtain stable and accurate frequency offset estimating value, complexity is not high.

Description of drawings

Fig. 1 is the synchronous flow chart of prior art WiMAX system.

Domain structure when Fig. 2 is the frame head lead code of IEEE 802.16e standard.

Fig. 3 is the FB(flow block) of frequency deviation estimating method of the present invention.

Fig. 4 is transmitter, phase θ in the inventive method step (a) _tAnd carry out the relation between the starting position offset of sequence of related operation.

Fig. 5 is the correlation R of receiving terminal in the inventive method step (b) _rAmplitude and phase place and carry out the original position n of the sequence of related operation ₀Between relation.

Fig. 6 is the amplitude of the cross correlation value between the lead code frequency domain sequence of frame head frequency domain sequence and emission of receiving terminal in the inventive method step (d).

Fig. 7 is the relation between final frequency offset estimating value of the inventive method and the iterations.

Embodiment

The present invention is described in further detail below in conjunction with accompanying drawing and practical implementation example.

Fig. 3 is the FB(flow block) of frequency deviation estimating method of the present invention.

IEEE 802.16e standard is supported 2048 points, 1024 points, and 128 FFT at 512, will do detailed explanation to the present invention as an example with 512 FFT below.

(a) the frame head lead code frequency domain sequence of base station is X (n); Through obtaining frame head time domain sequences s (n) after the FFT conversion, s (n) has the characteristic that repeat three periods pseudoperiods, gets T=pseudoperiod [N/3] of frame head; In the present embodiment; When N=512, round off T=171 presses first section sequence and second section sequence Carry out related operation, the phase place of this correlation as transmitter, phase θ _tAs shown in Figure 4, provided among the figure by

The phase value θ that calculates _tAnd first section sequence start bit put the relation between the offset, and wherein the offset span is [0,170].As can be seen from the figure first section sequence start bit put offset to θ _tInfluence very little.

(b) as shown in Figure 3, be r (n) at the mobile-terminated signal of receiving, the one tunnel postpones to get conjugation behind the individual sampled point of T (171), multiplies each other with the original signal sequence then.Through the burst sum R after obtaining behind the correlator multiplying each other _r, in this instance, the length of window value of correlator is T=171, T=171 substitution formula Obtain

n ₀Represent the original position of sequence, get this correlation R _rPhase place as receiving phase θ _rAs shown in Figure 5, among the figure correlation R of the inventive method receiving terminal _rAmplitude and phase place and sequence start position n ₀Between relation.At first according to correlation R _rAmplitude confirm regularly sync bit, as can be seen from the figure, the peak value platform appears in the amplitude of correlation when detecting frame head, the length of peak value platform is L+T, confirms the timing sync bit n of frame head thus ₀From figure, can see receiving phase θ _rPosition in that the peak value platform is corresponding is very steady, and therefore, regularly synchronous deviation can be ignored the influence of this method basically, thereby prevents because the synchronous inaccuracy of timing causes this method performance decrease.Simultaneously, can also find receiving phase θ _rWith transmitter, phase θ _tBetween difference has appearred, this difference is caused by two parts factor: one is frequency shift (FS), one is multipath channel response.

(c) as shown in Figure 3, through comparing receiving phase θ _rWith transmitter, phase θ _tBetween difference according to formula

Calculate the frequency deviation value of little several times.Owing to the influence of multipath channel response, the fractional part of frequency offset value of estimating here is not accurate frequency deviation value, needs to adopt the algorithm of iteration to eliminate the influence that multipath channel responds.

(d) as shown in Figure 3, according to the fractional part of frequency offset value of calculating to received signal r (n) carry out fractional part of frequency offset compensation, then the signal after the fractional part of frequency offset compensation is carried out the FFT conversion, convert frequency domain sequence Y (n) to.Y (n) and frame head lead code frequency domain sequence X (n) are carried out computing cross-correlation,

$C\left(m\right)=\underset{n=0}{\overset{N-1}{\mathrm{\&Sigma;}}}X(n+m)*Y^{*}\left(n\right),$ -(N-1)≤m≤(N-1)

Because the influence that integer frequency offset causes the reception frequency domain sequence is equivalent to the emission frequency domain sequence is shifted.Therefore, according to the peak m of cross-correlation _MaxObtain the integer frequency offset value.As shown in Figure 6, cross correlation value can be obtained maximum in a certain position, this peak m _MaxIt is by the emission frequency domain sequence m that moves to left that representative receives frequency domain sequence _MaxNumber of sub-carrier obtains at interval, so the integer frequency offset value is m _Max

(e) as shown in Figure 3; Signal to after the fractional part of frequency offset compensation carries out the integer frequency offset compensation; Then the signal behind the compensate of frequency deviation is carried out channel estimating; Obtain frequency domain channel estimated value

(being the channel estimation value on each number of sub-carrier), channel estimating adopts the MMSE channel estimation methods of existing band pilot tone.

(f) as shown in Figure 3, according to channel estimation value

Revise transmitter, phase θ _t, make previous formula

Become

Wherein

θ _tRevise correlation R for this reason _tPhase place;

(g) as shown in Figure 3, the method for employing iteration repeats above-mentioned (c), and (d), (e), step (f) is up to the frequency offset estimating value stabilization.Along with the increase of iterations will obtain more and more accurate frequency offset estimating value and channel estimation value.

Fig. 7 is the relation between final frequency offset estimating value of the inventive method and the iterations.The subcarrier spacing of WiMAX is 10.94KHz, and in emulation, true frequency deviation is 3KHz, can find do not having under the situation of iteration, and the frequency offset estimating value is 2.2KHz, the requirement of discontented football association view.And through the frequency offset estimating value of an iteration with through the frequency offset estimating value basically identical of iteration repeatedly, very near 3KHz.Therefore, generally speaking, the deviation that promptly can satisfy WiMAX agreement defined through the frequency offset estimating value that obtains after iteration is no more than the requirement of 2% subcarrier spacing, can think enough frequency offset estimating value accurately.

Claims (2)

1. the down link frequency deviation estimating method based on IEEE 802.16e communication standard is characterized in that, comprises the steps:

(a) the pseudo-repetition of the frame head time domain sequences of base station end emission is s (n)=IFFT (X (n)); Wherein X (n) is the lead code frequency domain sequence; IFFT counts and is that N, N are one of 2084,1024,512 and 128, get T=pseudoperiod [N/3] of frame head; Wherein [] expression round off criterion rounds operation to the number in [], according to formula

The correlation of calculating between the pseudo-repetitive sequence of the frame head of transmitting terminal is with this correlation R _tPhase place as transmitter, phase θ _t

(b) in the mobile terminal to received signal r (n) to carry out timing synchronous, regularly synchronous frame head position is n ₀, according to formula Calculate the correlation of the pseudo-repetitive sequence of frame head of receiving terminal, with this correlation R _rPhase place as receiving phase θ _r

(c) through comparing receiving phase θ _rWith transmitter, phase θ _tBetween difference, according to formula

Calculate the frequency deviation value of little several times

Span be [1.5,1.5];

(d) according to the fractional part of frequency offset value calculated to received signal r (n) carry out the fractional part of frequency offset compensation; Then the signal after the fractional part of frequency offset compensation is carried out the FFT conversion; Convert frequency domain sequence Y (n) to, Y (n) and frame head lead code frequency domain sequence X (n) carried out computing cross-correlation:

$C\left(m\right)=\underset{n=0}{\overset{N-1}{\mathrm{\&Sigma;}}}X(n+m)*Y*\left(n\right),-(N-1)\&le;m\&le;(N-1)$

Peak m according to cross correlation value _MaxObtain the integer frequency offset value;

(e) signal after the fractional part of frequency offset compensation is carried out the integer frequency offset compensation; Then the signal behind the compensate of frequency deviation is carried out channel estimating, obtain frequency domain channel estimated value

channel estimating and adopt the channel estimation methods of existing band pilot tone;

(f) according to channel estimation value

Revise transmitter, phase θ _t, make previous formula $R_t=\underset{i=0}{\overset{T-1}{\mathrm{\&Sigma;}}}s\left(i\right)s^{*}(i+T)$ Become $R_t=\underset{i=0}{\overset{T-1}{\mathrm{\&Sigma;}}}{s}^{\&prime;}\left(i\right)s^{\&prime;*}(i+T),$ Wherein $s^{\&prime;}\left(n\right)=\mathrm{IFFT}(X\left(n\right)*\hat{H}\left(n\right)),$ θ _tRevise correlation R for this reason _tPhase place;

(g) adopt iterative algorithm: repeat (c), (d), (e), step (f) is up to the frequency offset estimating value stabilization.

2. the down link frequency deviation estimating method based on IEEE 802.16e communication standard as claimed in claim 1 is characterized in that, in the said step (e), the channel estimation methods of existing band pilot tone is LS channel estimation methods or MMSE channel estimation methods.

Images