CN102655491A - Frequency shift estimation method and system for coherent demodulation frequency shift keying modulating signals - Google Patents

Abstract

The invention relates to the field of digital wireless communication and discloses a frequency shift estimation method and system for coherent demodulation frequency shift keying modulating signals. According to the method and the system, prefix signals of GFSK (Gaussian Frequency Shift Keying)/FSK (Frequency Shift Keying) signals are designed to a string of fixed-length codes with altering 0 and 1, and according to the prefix signals, an ideal receiving signal c(i) can be completely eliminated from an x(i) under the condition that a receiver does not need to copy the ideal receiving signal c(i), thereby solving the problem that as the frequency modulation index of the GFSK/FSK signal can not be obtained accurately and can change along with time and temperatures, a receiving terminal can not accurately copy the ideal receiving signal c(i), so as to improve the accuracy of the frequency shift evaluation. Therefore, by virtue of the iteration-combination of frequency evaluation results of R(N), R(2N),..., R (KN), the phase ambiguousness caused by the large frequency shift can be eliminated, and the precision of the frequency shift evaluation can be improved further.

Description

The frequency offset estimation methods and the system of coherent demodulation frequency shift keying modulation signal

Technical field

The present invention relates to digital wireless communication field, signal processing field especially is specifically related to a kind of frequency offset estimation methods and system of coherent demodulation frequency shift keying modulation signal.

Background technology

In digital wireless communication field, Frequency offset estimation is one of signal processing method of using always.

Because base station and travelling carriage, or travelling carriage can not definitely equate with carrier frequency between the travelling carriage, so frequency shift (FS) of comparatively fixing of existence between the reception signal of receiver actual reception signal and expectation.This frequency shift (FS) is harmful in digital wireless communication field, it can reduce channel estimating accuracy, cause the error rate to rise.Therefore in the digital radio communication system, receiver needs to carry out Frequency offset estimation according to up reception signal usually, carries out frequency offset compensation then, to eliminate all adverse effects that frequency shift (FS) was brought.

For the signal modulation, be divided into amplitude-shift keying ASK, frequency shift keying FSK/GFSK, phase-shift keying PSK.

Basic principle all is to act on (different modulation modes with carrier signal mutually with digital baseband signal; Computing formula is different); Output sending has obtained modulated high-frequency signal, and will recover original digital baseband signal at receiving terminal, just needs demodulation; Two kinds of demodulation modes are arranged, coherent demodulation and non-coherent demodulation mode.

The essential difference of two kinds of demodulation modes is; Coherent demodulation must recover coherent carrier; Utilize this coherent carrier and modulated signal effect, obtain initial digital baseband signal, and this coherent carrier is and be that same frequency is synchronous in the carrier signal of this baseband signal of transmitting terminal modulation originally; And non-coherent demodulation need not recover coherent carrier, so simpler than coherent demodulation mode.But the coherent demodulation mode in most of the cases, and the demodulation effect is quite a lot of.

So-called relevant, loosely says to be exactly the phase mutually synchronization, coherent demodulation is meant and utilizes multiplier, imports one the tunnel and multiplies each other with the reference signal and the carrier frequency of carrier frequency relevant (with the frequency homophase).

Such as obtaining signal Acos (ω t+ θ) after primary signal A and carrier frequency cos (the ω t+ θ) modulation;

Separate timing and introduce the reference signal cos (ω t+ θ) of relevant (with the frequency homophase), then obtain:

Acos(ωt+θ)cos(ωt+θ)

Utilize product to sum formula to obtain

A*1/2*[cos(ωt+θ+ωt+θ)+cos(ωt+θ-ωt-θ)]

=A*1/2*[cos(2ωt+2θ)+cos(0)]

=A/2*[cos(2ωt+2θ)+1]

=A/2+A/2cos(2ωt+2θ)

Utilize low pass filter with high-frequency signal cos (2 ω t+2 θ) filtering, promptly get primary signal A.

Therefore coherent demodulation needs receiver and carrier synchronization.

And non-coherent demodulation is not used multiplier, does not need receiver and carrier synchronization.

Demodulation mode is under the situation of non-coherent demodulation, need not accurate estimated frequency skew.For coherent demodulation, can significantly improve the performance of receiver, but accurately estimated frequency skew.

We with the GFSK/FSK signal indication that receives are:

$x\left(i\right)=c\left(i\right)e^{j(2\mathrm{\π}{f}_d\mathrm{iT}+\mathrm{\θ})}+n\left(i\right),i=\mathrm{1,2},...,Q$

Here c (i) represents desirable signal, the f of receiving _dRepresent frequency shift (FS), θ represents phase deviation, and T represents sampling time interval, and n (i) represents noise.

Because the frequency modulation(FM) index of GFSK/FSK signal can be to depart from its standard value (10% or higher) than mistake, this error can change the phase place of GFSK/FSK signal, and then changes the value of c (i), and the prior art scheme is difficult to accurate estimated frequency skew.

The patent of ArrayComm company: " Method and apparatus for decision directed demodulation usingantenna arrays and spatial processing ", inventor: Barratt; CraigH. (Redwood City, CA); Farzaneh; Farhad (San Francisco, CA); Parish; David M (Los Altos, CA), open day: on 04 23rd, 1998; Publication number: WO 98/17037; Described a kind of method of Frequency offset estimation, this method is slightly searched with large span earlier in possible frequency deviation range, and the order that then little span essence is searched is carried out Frequency offset estimation.

Said method uses one section known signal to carry out Frequency offset estimation, and its cost function is the actual reception signal and adds the error power between the ideal signal of look-in frequency skew back.When error power hour, the look-in frequency deviant of then using is the valuation of this Frequency offset estimation.

During this method practical engineering application; Need at first in receiving signal, to estimate the position of this section known signal, increased the complexity that this method realizes, also increased the probability of abnormal conditions; That is: during known signal location estimation mistake, the Frequency offset estimation result of this method is wrong.

The present invention then provides a kind of new frequency offset estimation methods and system in order to improve or to solve the above problems.

Summary of the invention

The technical problem that the present invention will solve is to provide a kind of frequency offset estimation methods of coherent demodulation frequency shift keying modulation signal; It can need not to duplicate under the desirable signal c (i) of the reception situation; C (i) is removed from x (i) fully; Thereby avoided because the frequency modulation(FM) index of GFSK/FSK signal can't accurately obtain and in time temperature change and, improved the accuracy of frequency offset estimating in the difficult problem that receiving terminal can't accurately duplicate the desirable signal c of reception (i).

The present invention solves above-mentioned technical problem through such technical scheme:

A kind of frequency offset estimation methods of coherent demodulation frequency shift keying modulation signal is provided, and this method may further comprise the steps:

GFSK/FSK is provided signal;

Prefix at the GFSK/FSK signal that provides is set to the sign indicating number that 0 and 1 of regular length replaces, and becomes prefix signal x (i);

One unlimited response low pass filter is provided, and above-mentioned prefix signal x (i) is through should unlimited response low pass filter, through the signal behind the unlimited response low pass filter be filtered signal x ' (i)=f _IIR-LPF[x (i)];

Filtered prefix signal is carried out complex conjugate multiplication computing and accumulating operation, obtain one group and eliminated in the prefix signal ' 0 ' and ' 1 ' phase change that produced, only stay the signal R (N) of the phase change that frequency shift (FS) produces, R (2N) ..., R (KN);

To above-mentioned R (N); R (2N); R (KN) carries out Frequency Estimation, obtains Frequency offset estimation value

As a kind of improvement of said method, said that filtered prefix signal is carried out the calculating formula of complex conjugate multiplication computing is following::

r _N(i _N)=x′(i ₀+i _N+N)[x′(i ₀+i _N)] ^*

r _2N(i _2N)=x′(i ₀+i _2N+2N)[x′(i ₀+i _2N)] ^*

…

r _KN(i _KN)=x′(i ₀+i _KN+KN)[x′(i ₀+i _KN)] ^*

Wherein, L is the multiple that sampling rate is compared the GFSK/FSK chip rate, and N is the factor of L, i ₀Be the starting point of available prefix signal, i _KN=0,1 ..., 2M _KNL-1 is decided by the length of available prefix signal, and K is system complexity and estimated accuracy, gets the maximal accuracy value from 1, and * representes complex conjugate; Wherein, M is the coefficient relevant with system complexity and estimated accuracy, can be constant, also can be the correction factor that changes according to system complexity, 2M _KNL-1 can be regarded as the restriction of the length of prefix signal, and this length is relevant with the multiple of system complexity, chip rate;

To the r that obtains through above-mentioned algorithm _N(i _N), r _2N(i _2N) ..., r _KN(i _KN) carry out following add operation:

$R\left(N\right)=\underset{i_N=0}{\overset{{2M}_{N}L-1}{\mathrm{\Σ}}}{r}_N\left(i_N\right)$

$R\left(2N\right)=\underset{i_{2N}=0}{\overset{{2M}_{2N}L-1}{\mathrm{\Σ}}}{r}_{2N}\left(i_{2N}\right)$

…

$R\left(\mathrm{KN}\right)=\underset{i_{\mathrm{KN}}=0}{\overset{{2M}_{\mathrm{KN}}L-1}{\mathrm{\Σ}}}{r}_{\mathrm{KN}}\left(i_{\mathrm{KN}}\right).$

Through the R with certain intervals (N) that aforementioned calculation is obtained, R (2N) ..., R (KN) is combined into the line frequency skew and estimates, can reach higher Frequency Estimation precision with less K, thereby greatly reduce system complexity.

As a kind of improvement of said method, the cut-off frequency that infinitely responds low pass filter can be lower than the bandwidth of GFSK/FSK signal.

As a kind of improvement of said method, compare the value of a factor N of the multiple L of GFSK/FSK chip rate as sampling rate and pass through

Decide, that is, the value of N can satisfy this conditional, wherein f _{D max}Be the maximum possible frequency shift (FS) of GFSK/FSK signal, T is a sampling time interval.Because the value of N decides through

; Therefore; As long as R (N) does not exist phase place indistinct, just can estimate bigger frequency shift (FS).

As a kind of improvement of said method, the N value when the N value is greater than low signal-to-noise ratio when high s/n ratio.

As a kind of improvement of said method, merge R (N) through following iteration, R (2N) ..., the Frequency Estimation result of R (KN):

${\hat{f}}_d\left(1\right)=\frac{1}{2\mathrm{\πNT}}\mathrm{angle}\left[R\left(N\right)\right];$

Begin to i=N from i=2:

$p_i=\mathrm{round}\left\{\frac{\mathrm{angle}\left[R\left(\mathrm{iN}\right)\right]-\mathrm{iN}\×\mathrm{angle}\left[R\left(N\right)\right]}{2\mathrm{\π}}\right\},$

${\hat{f}}_d\left(i\right)=a{\hat{f}}_d(i-1)+(1-a)\frac{\mathrm{angle}\left[R\left(\mathrm{iN}\right)\right]-{2\mathrm{\πp}}_i}{\mathrm{iN}}.$

Wherein,

Be the i Frequency offset estimation result in step, p _iBe an intermediate variable that uses in i step estimation procedure, a be one greater than 0 but less than 1 constant.Merge through this iteration that to eliminate the phase place that big frequency shift (FS) causes indistinct, can further improve the precision of Frequency offset estimation.

The present invention provides a kind of Frequency offset estimation system that uses the frequency offset estimation methods of above-mentioned coherent demodulation frequency shift keying modulation signal in addition, and this system comprises: signal generator, signal add code element, infinitely respond low pass filter, prefix signal computing unit, frequency estimator;

Said signal generator is used to provide the GFSK/FSK signal;

Said signal adds code element and is set to the sign indicating number that 0 and 1 of regular length replaces in the prefix of the GFSK/FSK signal that provides and becomes prefix signal x (i);

Said unlimited response low pass filter receives said prefix signal x (i), and carry out filtering obtain filtered signal x ' (i)=f _IIR-LPF[x (i)];

Said prefix signal computing unit carries out complex conjugate multiplication computing and accumulating operation to filtered prefix signal; Obtain one group and eliminated in the prefix signal ' 0 ' and ' 1 ' phase change that produced, only stay the signal R (N) of the phase change that frequency shift (FS) produces, R (2N);, R (KN);

Said frequency estimator is to above-mentioned R (N); R (2N); R (KN) carries out Frequency Estimation, obtains Frequency offset estimation value

Compared with prior art, the present invention has the following advantages: the present invention is designed to { 0,1 of a string regular length with the prefix signal of GFSK/FSK signal; 0,1 ...; 0; 1} sign indicating number, the use of this prefix code let receiver c (i) removed from x (i) fully need not to duplicate under the desirable signal c (i) of the reception situation; Thereby avoided because the frequency modulation(FM) index of GFSK/FSK signal can't accurately obtain and in time temperature change and, improved the accuracy of frequency offset estimating in the difficult problem that receiving terminal can't accurately duplicate the desirable signal c of reception (i).Since calculating R (N), R (2N) ... R (KN) uses unlimited response low pass filter to reduce the noise among the x (i) before, and uses the R (N) with certain intervals, R (2N); R (KN) combination comes the estimated frequency skew, can reach higher Frequency Estimation precision with less K, thereby greatly reduce system complexity.The present invention can also estimate big frequency shift (FS), can be issued to high estimation accuracy in the situation of low signal-to-noise ratio, estimate the optimum sampling point of coherent demodulation.

Description of drawings

Fig. 1 is the flow chart according to the frequency offset estimation methods of the coherent demodulation frequency shift keying modulation signal of first embodiment of the invention;

Fig. 2 is the logical schematic according to the Frequency offset estimation system of the coherent demodulation frequency shift keying modulation signal of third embodiment of the invention.

Embodiment

For making the object of the invention, technical scheme and advantage clearer, will combine accompanying drawing that each execution mode of the present invention is carried out detailed elaboration below.Yet, persons of ordinary skill in the art may appreciate that in each execution mode of the present invention, in order to make the reader understand the application better many ins and outs have been proposed.But,, also can realize each claim of the application technical scheme required for protection even without these ins and outs with based on the many variations and the modification of following each execution mode.

First execution mode of the present invention relates to a kind of frequency offset estimation methods of coherent demodulation frequency shift keying modulation signal, and is as shown in Figure 1, specifically comprises following steps:

Step S101 provides GFSK/FSK signal;

In this execution mode, be with the GFSK/FSK signal indication that receives:

$x\left(i\right)=c\left(i\right)e^{j(2\mathrm{\π}{f}_d\mathrm{iT}+\mathrm{\θ})}+n\left(i\right),i=\mathrm{1,2},...,Q$

Here c (i) represents desirable signal, the f of receiving _dRepresent frequency shift (FS), θ represents phase deviation, and T represents sampling time interval, and n (i) represents noise.

For the GFSK/FSK signal, demodulation mode commonly used is a non-coherent demodulation, therefore need not accurate estimated frequency skew.Coherent demodulation can significantly improve the performance of receiver, but accurately estimated frequency skew.Because the frequency modulation(FM) index of GFSK/FSK signal can be to depart from its standard value (10% or higher) than mistake, this error can change the phase place of GFSK/FSK signal, and then changes the value of c (i), and the prior art scheme is difficult to accurate estimated frequency skew.

Because the frequency modulation(FM) index of GFSK signal can't accurately obtain and in time temperature change, can't accurately duplicate the desirable signal c (i) of reception at receiving terminal, also just can't c (i) be removed from x (i) fully, thereby influence the accuracy of frequency offset estimating.

Step S102 is provided with the sign indicating number that 0 and 1 of a string regular length replaces in the prefix of the GFSK/FSK signal that provides, such as, 0,1,0,1 ..., and 0, the 1} sign indicating number becomes prefix signal x (i).The use let us of this prefix code can be removed c (i) from x (i) need not to duplicate under the desirable signal c (i) of the reception situation fully, thereby has improved the accuracy of frequency offset estimating.

In order to reduce the noise in the prefix signal, this prefix signal is at first through a unlimited response low pass filter, obtain filtered signal x ' (i)=f _IIR-LPF[x (i)], i.e. execution in step S103.

Wherein, this cut-off frequency that infinitely responds low pass filter can suitably be lower than the bandwidth of GFSK/FSK signal.Using unlimited response low pass filter is because the complexity of unlimited response filter is low.In addition, the employed frequency offset estimating algorithm of this execution mode can not receive the influence of the phase distortion that unlimited response filter causes.

Then, in step S104, filtered prefix signal is carried out complex conjugate multiplication computing and accumulating operation; Obtain one group and eliminated in the prefix signal ' 0 ' and ' 1 ' phase change that produced, only stay the signal R (N) of the phase change that frequency shift (FS) produces, R (2N);, R (KN).

Specifically, earlier to the computing of filtered prefix signal complex conjugate multiplication, calculating formula is following:

r _N(i _N)=x′(i ₀+i _N+N)[x′(i ₀+i _N)] ^*

r _2N(i _2N)=x′(i ₀+i _2N+2N)[x′(i ₀+i _2N)] ^*

…

r _KN(i _KN)=x′(i ₀+i _KN+KN)[x′(i ₀+i _KN)] ^*

In this execution mode, suppose that sampling rate is L a times of GFSK/FSK chip rate, N is the factor of L.In the aforementioned calculation formula, i ₀Be the starting point of available prefix signal, i _KN=0,1 ..., 2M _KNL-1 is decided by the length of available prefix signal, and K is system complexity and estimated accuracy, gets the maximal accuracy value from 1, and * representes complex conjugate; M is the coefficient relevant with system complexity and estimated accuracy, can be constant, also can be the correction factor that changes according to system complexity, 2M _KNL-1 can be regarded as the restriction of the length of prefix signal, and this length is relevant with the multiple of system complexity, chip rate.

Then, the r that aforementioned calculation is obtained _N(i _N), r _2N(i _2N) ..., r _KN(i _KN) carry out following add operation:

…

$R\left(N\right)=\underset{i_N=0}{\overset{{2M}_{N}L-1}{\mathrm{\Σ}}}{r}_N\left(i_N\right)$

$R\left(2N\right)=\underset{i_{2N}=0}{\overset{{2M}_{2N}L-1}{\mathrm{\Σ}}}{r}_{2N}\left(i_{2N}\right)$

$R\left(\mathrm{KN}\right)=\underset{i_{\mathrm{KN}}=0}{\overset{{2M}_{\mathrm{KN}}L-1}{\mathrm{\Σ}}}{r}_{\mathrm{KN}}\left(i_{\mathrm{KN}}\right).$

Because i _KNContain M _KNIndividual " 0 " code element and " 1 " code element; And the phase change that " 0 " and " 1 " is produced in the GFSK/FSK signal is just in time opposite; Therefore, above-mentioned add operation will be eliminated the phase change that " 0 " and " 1 " is produced, and only stays the phase change that frequency shift (FS) produces.Like this, no matter how the frequency modulation(FM) index of GFSK/FSK signal changes, and need not duplicate the desirable signal c (i) of reception at receiving terminal, just can c (i) be removed from x (i) fully, thereby accurately estimate frequency deviation.

Obtain one group and eliminated in the prefix signal ' 0 ' and ' 1 ' phase change that produced; Only stay the signal R (N) of the phase change of frequency shift (FS) generation, R (2N) ... R (KN) afterwards; To the R (N) that obtains, R (2N) ... R (KN) carries out Frequency Estimation, and obtaining Frequency offset estimation value

is execution in step S105.

Since unlimited response low pass filter make x ' (i) in adjacent data institute noisy correlation appears, in this execution mode, the value of N can be by following criterion decision:

1.

The value that is N can satisfy this conditional, wherein f _{D max}Be the maximum possible frequency shift (FS) of GFSK/FSK signal, T is a sampling time interval;

2. the N value when the N value is greater than low signal-to-noise ratio when high s/n ratio; And the value of N is the bigger the better under the situation of high s/n ratio; The value of N should slightly reduce under the situation of low signal-to-noise ratio.

Compared with prior art, this execution mode with the prefix signal of GFSK/FSK signal be designed to a string regular length 0,1,0,1 ..., 0, the 1} sign indicating number.The use of this prefix code lets receiver need not to duplicate under the desirable signal c (i) of the reception situation; C (i) is removed from x (i) fully; Thereby avoided because the frequency modulation(FM) index of GFSK/FSK signal can't accurately obtain and in time temperature change and, improved the accuracy of frequency offset estimating in the difficult problem that receiving terminal can't accurately duplicate the desirable signal c of reception (i).

Since calculating R (N), R (2N) ... R (KN) uses unlimited response low pass filter to reduce the noise among the x (i) before, and uses the R (N) with certain intervals, R (2N); R (KN) combination comes the estimated frequency skew, can reach higher Frequency Estimation precision with less K, thereby greatly reduce system complexity.

In addition; Shown in formula ; Therefore this execution mode can estimate big frequency shift (FS) as long as R (N) does not exist phase place indistinct.To the r that obtains by the complex conjugate multiplication computing _N(i _N), r _2N(i _2N) ..., r _KN(i _KN) handle the optimum sampling point that can also estimate coherent demodulation.

Second execution mode of the present invention relates to a kind of frequency offset estimation methods of coherent demodulation frequency shift keying modulation signal.Second execution mode has been done further improvement on the first execution mode basis; Its improvements mainly are: in second embodiment of the invention, merge R (N) through iteration, R (2N); The Frequency offset estimation result of R (KN) to eliminate the phase place ambiguity that big frequency shift (FS) is caused, can further improve the precision of Frequency offset estimation.

Specifically, this execution mode uses following algorithm to eliminate the phase place ambiguity that big frequency shift (FS) causes, and merges R (N) through iteration, R (2N) ..., the Frequency Estimation result of R (KN) is to reach high estimation accuracy:

１. ${\hat{f}}_d\left(1\right)=\frac{1}{2\mathrm{\πNT}}\mathrm{angle}\left[R\left(N\right)\right].$

2. begin to i=N from i=2:

$p_i=\mathrm{round}\left\{\frac{\mathrm{angle}\left[R\left(\mathrm{iN}\right)\right]-\mathrm{iN}\×\mathrm{angle}\left[R\left(N\right)\right]}{2\mathrm{\π}}\right\},$

${\hat{f}}_d\left(i\right)=a{\hat{f}}_d(i-1)+(1-a)\frac{\mathrm{angle}\left[R\left(\mathrm{iN}\right)\right]-{2\mathrm{\πp}}_i}{\mathrm{iN}}.$

Wherein,

Be the i Frequency offset estimation result in step, p _iBe an intermediate variable that uses in i step estimation procedure, a be one greater than zero but less than one constant.The value of a in the formula is optimized the precision that can further increase Frequency Estimation according to the situation of signal to noise ratio.

Because this execution mode merges R (N) through iteration, R (2N) ..., the Frequency Estimation result of R (KN) can be issued to high estimation accuracy in the situation of low signal-to-noise ratio.

The step of top the whole bag of tricks is divided, and is just clear in order to describe, and can merge into a step during realization and perhaps some step split, and is decomposed into a plurality of steps, as long as comprise identical logical relation, all in the protection range of this patent; To adding inessential modification in the algorithm or in the flow process or introduce inessential design, but the core design that does not change its algorithm and flow process is all in the protection range of this patent.

Third embodiment of the invention relates to a kind of Frequency offset estimation system of frequency offset estimation methods of the coherent demodulation frequency shift keying modulation signal that uses first embodiment of the invention; As shown in Figure 2, this system comprises: signal generator, signal add code element, infinitely respond low pass filter, prefix signal computing unit, frequency estimator;

Wherein, signal generator is used to provide the GFSK/FSK signal;

Signal adds code element and is set to the sign indicating number that 0 and 1 of regular length replaces in the prefix of the GFSK/FSK signal that provides and becomes prefix signal x (i);

Unlimited response low pass filter receives said prefix signal x (i), and carry out filtering obtain filtered signal x ' (i)=f _IIR-LPF[x (i)]; Wherein, the cut-off frequency that infinitely responds low pass filter is lower than the bandwidth of GFSK/FSK signal.

The prefix signal computing unit carries out complex conjugate multiplication computing and accumulating operation to filtered prefix signal; Obtain one group and eliminated in the prefix signal ' 0 ' and ' 1 ' phase change that produced, only stay the signal R (N) of the phase change that frequency shift (FS) produces, R (2N);, R (KN);

Specifically, the prefix signal computing unit comprises complex conjugate multiplication operator unit and accumulating operation subelement; Complex conjugate multiplication operator unit carries out the complex conjugate multiplication computing to filtered prefix signal:

r _N(i _N)=x′(i ₀+i _N+N)[x′(i ₀+i _N)] ^*

r _2N(i _2N)=x′(i ₀+i _2N+2N)[x′(i ₀+i _2N)] ^*

…

r _KN(i _KN)=x′(i ₀+i _KN+KN)[x′(i ₀+i _KN)] ^*

Wherein, L is the multiple that sampling rate is compared the GFSK/FSK chip rate, and N is the factor of L; i ₀Be the starting point of available prefix signal, i _KN=0,1 ..., 2M _KNL-1 is decided by the length of available prefix signal, and K is system complexity and estimated accuracy, gets the maximal accuracy value from 1, and * representes complex conjugate, and M is the correction factor relevant with system complexity and estimated accuracy;

The r of accumulating operation subelement to calculating through complex conjugate multiplication operator unit _N(i _N), r _2N(i _2N) ..., r _KN(i _KN) carry out add operation:

$R\left(N\right)=\underset{i_N=0}{\overset{{2M}_{N}L-1}{\mathrm{\Σ}}}{r}_N\left(i_N\right)$

$R\left(2N\right)=\underset{i_{2N}=0}{\overset{{2M}_{2N}L-1}{\mathrm{\Σ}}}{r}_{2N}\left(i_{2N}\right)$

$R\left(\mathrm{KN}\right)=\underset{i_{\mathrm{KN}}=0}{\overset{{2M}_{\mathrm{KN}}L-1}{\mathrm{\Σ}}}{r}_{\mathrm{KN}}\left(i_{\mathrm{KN}}\right).$

Frequency estimator is to above-mentioned R (N); R (2N); R (KN) carries out Frequency Estimation, obtains Frequency offset estimation value

In addition, the value of a factor N of comparing the multiple L of GFSK/FSK chip rate as sampling rate in the complex conjugate multiplication operator unit is passed through

Decide, wherein f _{D max}Be the maximum possible frequency shift (FS) of GFSK/FSK signal, T is a sampling time interval.And the N value when the N value is greater than low signal-to-noise ratio when high s/n ratio.

In addition; What deserves to be mentioned is; In practical application, the prefix signal computing unit can combine through complex conjugate unit as shown in Figure 2, N rank serial register, complex multiplier, adder, multiple register etc. and realize complex conjugate multiplication computing and accumulating operation, but the present invention is not as limit; Any ability realizes above-mentioned computing; Obtain one group and eliminated in the prefix signal ' 0 ' and ' 1 ' phase change that produced, only stay the signal of the phase change that frequency shift (FS) produces, supply module or combination of devices that frequency estimator carries out Frequency Estimation all within protection scope of the present invention.

Be not difficult to find that this execution mode is and the corresponding system embodiment of first execution mode, this execution mode can with the enforcement of working in coordination of first execution mode.The correlation technique details of mentioning in first execution mode is still effective in this execution mode, in order to reduce repetition, repeats no more here.Correspondingly, the correlation technique details of mentioning in this execution mode also can be applicable in first execution mode.

What deserves to be mentioned is that each involved in this execution mode module is logic module, in practical application, a logical block can be a physical location, also can be the part of a physical location, can also realize with the combination of a plurality of physical locations.In addition, for outstanding innovation part of the present invention, will not introduce in this execution mode, but this does not show the unit that does not have other in this execution mode with solving the not too close unit of technical problem relation proposed by the invention.

Four embodiment of the invention relates to a kind of Frequency offset estimation system.The 4th execution mode has been done further improvement on the 3rd execution mode basis, its improvements mainly are: in four embodiment of the invention, this system also comprises eliminates the indistinct unit of phase place; This unit merges R (N) through iteration; R (2N) ..., the Frequency Estimation result of R (KN):

${\hat{f}}_d\left(1\right)=\frac{1}{2\mathrm{\πNT}}\mathrm{angle}\left[R\left(N\right)\right];$

Begin to i=N from i=2:

$p_i=\mathrm{round}\left\{\frac{\mathrm{angle}\left[R\left(\mathrm{iN}\right)\right]-\mathrm{iN}\×\mathrm{angle}\left[R\left(N\right)\right]}{2\mathrm{\π}}\right\},$

${\hat{f}}_d\left(i\right)=a{\hat{f}}_d(i-1)+(1-a)\frac{\mathrm{angle}\left[R\left(\mathrm{iN}\right)\right]-{2\mathrm{\πp}}_i}{\mathrm{iN}}.$

Wherein,

Be the i Frequency offset estimation result in step, p _iBe an intermediate variable that uses in i step estimation procedure, a be one greater than 0 but less than 1 constant.

Because second execution mode is corresponding each other with this execution mode, thus this execution mode can with the enforcement of working in coordination of second execution mode.The correlation technique details of mentioning in second execution mode is still effective in this execution mode, and the technique effect that in second execution mode, can reach can be realized in this execution mode too, in order to reduce repetition, repeats no more here.Correspondingly, the correlation technique details of mentioning in this execution mode also can be applicable in second execution mode.

The above is merely preferred embodiments of the present invention; Protection scope of the present invention is not exceeded with above-mentioned execution mode; As long as the equivalence that those of ordinary skills do according to disclosed content is modified or changed, all should include in the protection range of putting down in writing in claims.

Claims (12)

1. the frequency offset estimation methods of a coherent demodulation frequency shift keying modulation signal is characterized in that, this method may further comprise the steps:

GFSK/FSK is provided signal;

The prefix of the GFSK/FSK signal that provides is set to the sign indicating number that 0 and 1 of regular length replaces, and becomes prefix signal x (i);

One unlimited response low pass filter is provided, and above-mentioned prefix signal x (i) is through should unlimited response low pass filter, through the signal behind the unlimited response low pass filter be filtered signal x ' (i)=f _IIR-LPF[x (i)];

Filtered prefix signal is carried out complex conjugate multiplication computing and accumulating operation, obtain one group and eliminated in the prefix signal ' 0 ' and ' 1 ' phase change that produced, only stay the signal R (N) of the phase change that frequency shift (FS) produces, R (2N) ..., R (KN);

To above-mentioned R (N); R (2N); R (KN) carries out Frequency Estimation, obtains Frequency offset estimation value

2. the frequency offset estimation methods of coherent demodulation frequency shift keying modulation signal according to claim 1 is characterized in that: said that filtered prefix signal is carried out the calculating formula of complex conjugate multiplication computing is following:

r _N(i _N)=x′(i ₀+i _N+N)[x′(i ₀+i _N)] ^*

r _2N(i _2N)=x′(i ₀+i _2N+2N)[x′(i ₀+i _2N)] ^*

…

r _KN(i _KN)=x′(i ₀+i _KN+KN)[x′(i ₀+i _KN)] ^*

Wherein, L is the multiple that sampling rate is compared the GFSK/FSK chip rate, and N is the factor of L, i ₀Be the starting point of available prefix signal, i _KN=0,1 ..., 2M _KNL-1 is decided by the length of available prefix signal, and K is system complexity and estimated accuracy, gets the maximal accuracy value from 1, and * representes complex conjugate, and M is the correction factor relevant with system complexity and estimated accuracy;

To the r that obtains through aforementioned calculation _N(i _N), r _2N(i _2N) ..., r _KN(i _KN) carry out add operation:

$R\left(N\right)=\underset{i_N=0}{\overset{{2M}_{N}L-1}{\mathrm{\Σ}}}{r}_N\left(i_N\right)$

$R\left(2N\right)=\underset{i_{2N}=0}{\overset{{2M}_{2N}L-1}{\mathrm{\Σ}}}{r}_{2N}\left(i_{2N}\right)$

…

$R\left(\mathrm{KN}\right)=\underset{i_{\mathrm{KN}}=0}{\overset{{2M}_{\mathrm{KN}}L-1}{\mathrm{\Σ}}}{r}_{\mathrm{KN}}\left(i_{\mathrm{KN}}\right).$

3. the frequency offset estimation methods of coherent demodulation frequency shift keying modulation signal according to claim 1 is characterized in that: the cut-off frequency of said unlimited response low pass filter is lower than the bandwidth of GFSK/FSK signal.

4. the frequency offset estimation methods of coherent demodulation frequency shift keying modulation signal according to claim 2 is characterized in that: the value of a factor N of comparing the multiple L of GFSK/FSK chip rate as sampling rate is passed through

Decide, wherein f _{D max}Be the maximum possible frequency shift (FS) of GFSK/FSK signal, T is a sampling time interval.

5. the frequency offset estimation methods of coherent demodulation frequency shift keying modulation signal according to claim 4 is characterized in that: the N value when the N value is greater than low signal-to-noise ratio when high s/n ratio.

6. the frequency offset estimation methods of coherent demodulation frequency shift keying modulation signal according to claim 1 is characterized in that: merge R (N) through following iteration, and R (2N) ..., the Frequency offset estimation result of R (KN):

${\hat{f}}_d\left(1\right)=\frac{1}{2\mathrm{\πNT}}\mathrm{angle}\left[R\left(N\right)\right];$

Begin to i=N from i=2:

$p_i=\mathrm{round}\left\{\frac{\mathrm{angle}\left[R\left(\mathrm{iN}\right)\right]-\mathrm{iN}\×\mathrm{angle}\left[R\left(N\right)\right]}{2\mathrm{\π}}\right\},$

${\hat{f}}_d\left(i\right)=a{\hat{f}}_d(i-1)+(1-a)\frac{\mathrm{angle}\left[R\left(\mathrm{iN}\right)\right]-{2\mathrm{\πp}}_i}{\mathrm{iN}}.$

Wherein,

Be the i Frequency offset estimation result in step, p _iBe an intermediate variable that uses in i step estimation procedure, a be one greater than 0 but less than 1 constant.

7. Frequency offset estimation system that uses the frequency offset estimation methods of coherent demodulation frequency shift keying modulation signal as claimed in claim 1; It is characterized in that this system comprises: signal generator, signal add code element, infinitely respond low pass filter, prefix signal computing unit, frequency estimator;

Said signal generator is used to provide the GFSK/FSK signal;

Said signal adds code element and is set to the sign indicating number that 0 and 1 of regular length replaces in the prefix of the GFSK/FSK signal that provides and becomes prefix signal x (i);

Said unlimited response low pass filter receives said prefix signal x (i), and carry out filtering obtain filtered signal x ' (i)=f _IIR-LPF[x (i)];

Said prefix signal computing unit carries out complex conjugate multiplication computing and accumulating operation to filtered prefix signal; Obtain one group and eliminated in the prefix signal ' 0 ' and ' 1 ' phase change that produced; Only stay the signal R (N) of the phase change of frequency shift (FS) generation; R (2N) ..., R (KN);

Said frequency estimator is to above-mentioned R (N); R (2N); R (KN) carries out Frequency Estimation, obtains Frequency offset estimation value

8. Frequency offset estimation according to claim 7 system, it is characterized in that: said prefix signal computing unit comprises complex conjugate multiplication operator unit and accumulating operation subelement;

Said complex conjugate multiplication operator unit carries out the complex conjugate multiplication computing to filtered prefix signal:

r _N(i _N)=x′(i ₀+i _N+N)[x′(i ₀+i _N)] ^*

r _2N(i _2N)=x′(i ₀+i _2N+2N)[x′(i ₀+i _2N)] ^*

…

r _KN(i _KN)=x′(i ₀+i _KN+KN)[x′(i ₀+i _KN)] ^*

Wherein, L is the multiple that sampling rate is compared the GFSK/FSK chip rate, and N is the factor of L, i ₀Be the starting point of available prefix signal, i _KN=0,1 ..., 2M _KNL-1 is decided by the length of available prefix signal, and K is system complexity and estimated accuracy, gets the maximal accuracy value from 1, and * representes complex conjugate, and M is the correction factor relevant with system complexity and estimated accuracy;

The r of said accumulating operation subelement to calculating through complex conjugate multiplication operator unit _N(i _N), r _2N(i _2N) ..., r _KN(i _KN) carry out add operation:

…

$R\left(N\right)=\underset{i_N=0}{\overset{{2M}_{N}L-1}{\mathrm{\Σ}}}{r}_N\left(i_N\right)$

$R\left(2N\right)=\underset{i_{2N}=0}{\overset{{2M}_{2N}L-1}{\mathrm{\Σ}}}{r}_{2N}\left(i_{2N}\right)$

$R\left(\mathrm{KN}\right)=\underset{i_{\mathrm{KN}}=0}{\overset{{2M}_{\mathrm{KN}}L-1}{\mathrm{\Σ}}}{r}_{\mathrm{KN}}\left(i_{\mathrm{KN}}\right).$

9. Frequency offset estimation according to claim 7 system, it is characterized in that: the cut-off frequency of said unlimited response low pass filter is lower than the bandwidth of GFSK/FSK signal.

10. Frequency offset estimation according to claim 8 system is characterized in that: the value of a factor N of comparing the multiple L of GFSK/FSK chip rate as sampling rate in the said complex conjugate multiplication operator unit is passed through

Decide, wherein f _{D max}Be the maximum possible frequency shift (FS) of GFSK/FSK signal, T is a sampling time interval.

11. Frequency offset estimation according to claim 10 system is characterized in that: the N value when the N value is greater than low signal-to-noise ratio when high s/n ratio.

12. Frequency offset estimation according to claim 7 system is characterized in that: this system also comprises eliminates the indistinct unit of phase place, and this unit merges R (N) through iteration, R (2N) ..., the Frequency Estimation result of R (KN):

${\hat{f}}_d\left(1\right)=\frac{1}{2\mathrm{\πNT}}\mathrm{angle}\left[R\left(N\right)\right];$

Begin to i=N from i=2:

$p_i=\mathrm{round}\left\{\frac{\mathrm{angle}\left[R\left(\mathrm{iN}\right)\right]-\mathrm{iN}\×\mathrm{angle}\left[R\left(N\right)\right]}{2\mathrm{\π}}\right\},$

${\hat{f}}_d\left(i\right)=a{\hat{f}}_d(i-1)+(1-a)\frac{\mathrm{angle}\left[R\left(\mathrm{iN}\right)\right]-{2\mathrm{\πp}}_i}{\mathrm{iN}}.$

Wherein,

Be the i Frequency offset estimation result in step, p _iBe an intermediate variable that uses in i step estimation procedure, a be one greater than 0 but less than 1 constant.

Images