CN100483984C - Automatic frequency control method in CDMA mobile communication system - Google Patents

Abstract

This invention refers to CDMA communication method and discloses a automatic frequency control method in CDMA mobile communication system to raise frequency deviation estimation precision and improve multi-route search and judgment property, which contains calculating sub composite time delay spectrum, dividing calculating coherent accumulated value sequence and power time delay point, searching roughly alternate route phase point, calculating frequency deviation estimation value of alternate route phase position.

Description

Auto frequency control method in the CDMA mobile communication systems

Technical field

The present invention relates to the CDMA mobile telecommunication method, the particularly control of the automatic frequency in the CDMA mobile communication systems.

Background technology

3-G (Generation Three mobile communication system) is to satisfy international mobile communication (the International Mobile Telecommunication 2000 that International Telecommunications Union proposes, be called for short " IMT-2000 ")/following public land mobile (the Future Public Land Mobile Telephone Systems of system, abbreviation " FPLMTS ") third generation mobile communication system of standard, requirement has good network compatibility, can realize the roaming between a plurality of different systems in the global range, not only will be for the mobile subscriber provide speech and low-rate data business, and multimedia service widely will be provided.According to this standard, Wideband Code Division Multiple Access (WCDMA) (Wide-band Code Division Multiple Access has been proposed at present in the world, abbreviation " WCDMA "), Cdma2000, TD SDMA 3G (Third Generation) Moblie system schemas such as (Time Division SynchronousCode Division Multiple Access are called for short " TD-SCDMA ").Though these schemes are very not identical, but because CDMA mobile communication systems has advantages such as high power capacity, high quality-of-service and good confidentiality, the whole world adopts code division multiple access (Code Division Multiple Access is called for short " CDMA ") technology to reach common understanding in 3-G (Generation Three mobile communication system).

Be important practical problem in the communication system synchronously.When adopting synchronous demodulation or coherent detection, receiving terminal need provide one with the coherent carrier of transmitting terminal modulated carrier with the frequency homophase.Obtaining of this coherent carrier is exactly carrier synchronization.It is particularly important that carrier synchronization seems in mobile communication system, because in the mobile communication environment, has various interference, such as multiple access interference, the diffusion of multipath frequency, Doppler frequency shift etc., makes channel performance reduce, and signal quality is not as wire communication.These problems have influenced the performance and the implementation complexity of carrier synchronization equally, such as when the relative transmitter of receiver is in the motion state of given pace, the signal that receives has Doppler frequency shift, amounts of frequency offset and relative velocity are directly proportional, if frequency deviation surpasses certain degree, will cause the pernicious decline of communication quality, and the carrier extract meeting inaccuracy that becomes.The problems referred to above can appear in the CDMA mobile communication system equally.

Therefore, in CDMA mobile communication system, how to improve the performance of carrier synchronization, become one of key technology of CDMA.(the User Equipment of subscriber terminal equipment in the cdma system, be called for short " UE ") and base station system (Base Station, abbreviation " BS ") up, the down channel between all adopts the detection method of coherent demodulation, and this requires the demodulation carrier wave of receiving terminal and the same homophase frequently of modulated carrier of transmitting terminal.Because volume and cost limit, it is very high that the long-time stability of the local crystal oscillator of UE can not be done.Add the Doppler frequency shift of the mobility generation of UE, this all makes and has differing and frequency difference of modulated carrier and demodulation carrier wave between UE and the BS.To such an extent as to become one of necessary key technology in UE end carrier synchronization, directly influence the quality and the service quality of received signal.

Especially in the WCDMA mobile communication system, according to standard, the crystal oscillator output jitter of UE must be less than 3ppm, the translational speed of UE but can reach 500km/h, the carrier wave shake of calculating UE thus can make the frequency deviation that produces between the modulated carrier of the demodulation carrier wave of UE and BS up to 6kHZ, add that frequency deviation can reach 7kHZ because the Doppler frequency shift that the mobility of UE produces is 1kHZ.The automatic frequency control of adopting at present (Automatic Frequency Control is called for short " AFC ") technology is the carrier frequency that is used for following the tracks of and catch the signal that receives from BS, with reach send consistent with receive frequency.According to the WCDMA standard, UE is after using the AFC technology, and its carrier wave shake should also be 200HZ less than 0.1ppm.This explanation AFC technology tool in the carrier synchronization problem that solves CDMA mobile communication system has an enormous advantage.

Only use the AFC technology, can not solve the carrier synchronization problem of BS at the UE end.Though the stability of the crystal oscillator of BS is obviously much higher than UE, but the Doppler frequency shift that the high-speed mobile of UE is brought, and UE communicates with a plurality of base stations in the soft handover between the sub-district simultaneously, but only and one of them base station implementation carrier synchronization, this will cause all that frequency deviation between UE and the BS is the highest may to reach 400-600HZ.So in BS, also need to adopt the AFC technology to carry out carrier synchronization.

In cdma system, the AFC technology has two kinds, and difference is that the correction of frequency deviation is to carry out before Multipath searching or after the Multipath searching.If correcting frequency deviation before Multipath searching, will can not be subjected to the influence of frequency deviation in the Multipath searching process, can improve the accuracy of Multipath searching; Otherwise, after Multipath searching, carry out the complexity that correcting frequency deviation can reduce correcting frequency deviation, because can directly estimate partially and correction at the enterprising line frequency of phase point in these footpaths according to multipath information.Consider the requirement of systematic function, generally adopt the way of carrying out correcting frequency deviation before Multipath searching, what the present invention relates to is exactly the AFC method of not knowing to carry out before the multipath information correcting frequency deviation.

The AFC method of estimation that adopts comprises following steps at present:

It is relevant to slide on a segment signal of input by correlated series, obtains the multiple delay profile of son (Complex Delay Profile is called for short " CDP ");

Carry out AFC at all phase points and estimate, obtain frequency offset estimation result;

All phase points are rectified a deviation obtains complete CDP, can rectify a deviation to this phase point according to the frequency offset estimation result of phase point here, and the sub-CDP stack that will obtain after will rectifying a deviation again calculates complete CDP;

Rated output delay profile (Power Delay Profile is called for short " PDP "), this PDP has been the PDP that removes frequency deviation, can be used for Multipath searching and judgement, improves the performance of Multipath searching.

The AFC technology of current employing is estimated partially at the enterprising line frequency of all phase points, and multipath signal is not to be distributed on all phase points, just has the footpath signal on some phase point far away of being separated by in general, and this has just caused some problems.When having the footpath signal on this phase point, because the complete correlation of received signal, the frequency offset estimation result that obtains is more accurate; Otherwise, when not having the footpath signal on this phase point, because The noise makes the CDP that calculates can not accurately reflect frequency deviation, the situation of mistake correction can appear on the contrary, have a strong impact on the performance of Multipath searching.

In actual applications, there is following problem in such scheme: the inaccurate AFC technology that will cause of the frequency offset estimating on some phase point is unreasonable, influences the accuracy of Multipath searching and judgement, reduces the multipath handling property of system.

Cause a main cause of this situation to be, all phase points have all been carried out AFC frequency offset estimating and correction.

Summary of the invention

The technical problem to be solved in the present invention provides the auto frequency control method in a kind of CDMA mobile communication systems, makes the precision of frequency offset estimating be improved, thereby improves the performance of Multipath searching and judgement.

In order to solve the problems of the technologies described above, the invention provides the auto frequency control method in a kind of CDMA mobile communication systems, comprise following steps:

It is relevant that coherent signal sequence and input signal sequence are slided, and calculates the multiple delay profile of son;

According to the multiple delay profile of described son, coherent accumulation value sequence and power delay spectrum are calculated in segmentation;

Phase point according to described several candidate's path positions of power delay spectrum coarse search;

According to the numerical value of described coherent accumulation value and described candidate's path position place phase point, the frequency offset estimating value of the phase point of calculated candidate path position correspondence.

Wherein, the described step of calculating the multiple delay profile of son is: with length is N _cCoherent signal sequence c (n), n=1,2,3 ..., N _c, be N in length _xInput signal sequence x (n), n=1,2,3 ..., N _xLast slip N _yLength, then correlation is divided into the I section and superposes, every segment length is $N_a=\frac{N_c}{I},$ Obtaining I length is N _yThe multiple delay profile of son, use y _k(i) expression, its computing formula is as follows:

$y_k\left(i\right)=\underset{n=1+(i-1)N_a}{\overset{i{N}_a}{\mathrm{\Σ}}}c\left(n\right)x(n+k),$ i＝1，2，3，...，I，k＝1，2，3，...，N _y，

Wherein ∑ is a summation operation.

The formula that the coherent accumulation value sequence is calculated in described segmentation is:

$z_k\left(j\right)=\underset{i=1+(j-1)M}{\overset{\mathrm{jM}}{\mathrm{\Σ}}}{y}_k\left(i\right),$ k＝1，2，3，...，N _y $j=\mathrm{1,2},...,\frac{I}{M};$

The formula of rated output delay profile is:

$P_k={\left|\right|\underset{j=1}{\overset{\frac{I}{M}}{\mathrm{\Σ}}}{z}_k\left(j\right)\left|\right|}^2,k=\mathrm{1,2,3},...,N_y;$

Wherein, z _k(j) be j accumulated value on the k phase point, y _k(i) be i the multiple value of delay profile on the k phase point of son, N _yBe sliding length, the hop count of coherent accumulation value is calculated in the I segmentation, the number of samples of M for add up at every turn, P _kBe the power delay spectrum value, ∑ is a summation operation, and ‖ ‖ is a modulo operation.

The method of described coarse search comprises step:

A is changed to all phase points effectively, puts search sequence number initial value;

B searches out the phase point of the peaked phase point of power delay spectrum as candidate's path position from effective phase point, with this phase point adjacent about each 3 point be changed to invalidly, and will search for sequence number and increase one;

C judges whether that according to the search sequence number search finishes, if, promptly search for sequence number and count above the phase place that will search for, then finish this search; Otherwise, return step B and continue search.

Described candidate's path position phase point frequency offset calculation method is, the coherent accumulation value of the time delay of will being separated by is carried out multiplying and added up, and computing formula is:

$f_{k_l}=\frac{1}{N_d}A\tan \left(\underset{j=1}{\overset{\frac{I}{M}-N_d}{\mathrm{\Σ}}}{z}_{k_l}^{*}\left(j\right)z_{k_l}(j+N_d)\right),l=\mathrm{1,2},...,L.$

Wherein, l represents the sequence number in candidate footpath, k _lBe l the residing phase point numerical value in candidate footpath,

Be k _lAlso the frequency offset estimating value in l candidate footpath on the individual phase point, N _dFor time delay value can be got

In arbitrary value, Atan represents arctan function, I is the hop count that the coherent accumulation value is calculated in segmentation, the number of samples of M for add up at every turn,

Be k _lAlso be j accumulated value of candidate's path position on the individual phase point, L is a candidate footpath number, z ^*The conjugate complex number of expression z, ∑ is a summation operation.

Described candidate's path position phase point frequency offset calculation method is:

$f_{k_l}=\frac{1}{N_d}A\tan \left(\underset{m=-N_1}{\overset{N_2}{\mathrm{\Σ}}}\underset{j=1}{\overset{\frac{I}{M}-N_d}{\mathrm{\Σ}}}{z}_{k_l+m}^{*}\left(j\right)z_{k_l+m}(j+N_d)\right),l=\mathrm{1,2},...,L.$

Wherein, k _lBe l the residing phase point numerical value in candidate footpath,

Be k _lAlso the frequency offset estimating value of candidate's path position on the individual phase point, N _dFor time delay value can be got

In arbitrary value, Atan represents arctan function, I is the hop count that the coherent accumulation value is calculated in segmentation, the number of samples of M for add up at every turn, N ₁, N ₂Respectively get the arbitrary value in 1,2,3, z _k(j) be j accumulated value on k the phase point, L is the number in candidate footpath, z ^*The conjugate complex number of expression Z, ∑ is a summation operation.

Described candidate's path position phase point frequency offset calculation method is, the coherent accumulation value of the time delay of will being separated by is carried out multiplying and added up, and the value of all candidate's path position phase points is averaged as the frequency offset estimating value, and formula is:

$f_{k_l}=f=\frac{1}{N_d}A\tan \left(\underset{l=1}{\overset{L}{\mathrm{\Σ}}}\underset{j=1}{\overset{\frac{I}{M}-N_d}{\mathrm{\Σ}}}{z}_{k_l}^{*}\left(j\right)z_{k_l}(j+N_d)\right),l=\mathrm{1,2},...,L.$

Wherein, k _lBe l the residing phase point numerical value in candidate footpath,

Be k _lAlso be the frequency offset estimating value of candidate's path position on the individual phase point, f is the i.e. frequency offset estimating value of all candidate's path positions of the mean value of trying to achieve, N _dFor time delay value can be got

In arbitrary value, Atan represents arctan function, I is the hop count that the coherent accumulation value is calculated in segmentation, the number of samples of M for add up at every turn,

Be k _lAlso be j accumulated value of candidate's path position on the individual phase point, L is the number in candidate footpath, z ^*The conjugate complex number of expression z, ∑ is a summation operation.

Described candidate's path position phase point frequency offset calculation method is:

$f_{k_l}=f=\frac{1}{N_d}A\tan \left(\underset{l=1}{\overset{L}{\mathrm{\Σ}}}\underset{m=-N_1}{\overset{N_2}{\mathrm{\Σ}}}\underset{j=1}{\overset{\frac{I}{M}-N_d}{\mathrm{\Σ}}}{z}_{k_l+m}^{*}\left(j\right)z_{k_l+m}(j+N_d)\right),l=\mathrm{1,2},...,L.$

Wherein, k _lBe l the residing phase point numerical value in candidate footpath,

Be k _lAlso be the frequency offset estimating value of candidate's path position on the individual phase point, f is the mean value of frequency offset estimating value of the phase point of candidate's path position, N _dFor time delay value can be got In arbitrary value, Atan represents arctan function, I is the hop count that the coherent accumulation value is calculated in segmentation, the number of samples of M for add up at every turn, N ₁, N ₂Respectively get the arbitrary value in 1,2,3, z _k(j) be j accumulated value on k the phase point, L is a candidate footpath number, z ^*The conjugate complex number of expression z, ∑ is a summation operation.

When described method is applied to calculating the phase point of non-candidate's path position,, calculate the frequency offset estimating value of the phase point of non-candidate's path position according to the frequency offset estimating of the phase point of candidate's path position.

The method of frequency offset estimating value that is applied to calculate the phase point of non-candidate's path position is directly to represent with the mean value of the frequency offset estimating value of the phase value of all candidate's path positions.

The method of frequency offset estimating value that is applied to calculate the phase point of non-candidate's path position is, directly get adjacent with this non-candidate's path position phase point nearest also be the phase point frequency offset estimating value at the phase point place of minimum candidate's path position at interval.

When described method is applied to calculating the later power delay spectrum of correcting frequency deviation,, calculate the later power delay spectrum of correcting frequency deviation, be used for Multipath searching and judgement according to the frequency offset estimating value of the phase point that calculates gained candidate's path position and non-candidate's path position.

The formula that calculates the later power delay spectrum of correcting frequency deviation is:

${P\′}_k=\underset{i=1}{\overset{\frac{I}{M_f}}{\mathrm{\Σ}}}{\left|\right|\underset{n=1+(i-1)M_f}{\overset{{\mathrm{iM}}_f}{\mathrm{\Σ}}}{y}_k\left(n\right)e^{-j\frac{{\mathrm{nf}}_k}{M}}\left|\right|}^2,$ k＝1，2，3，...，N _y，

Wherein, P ' _kBe the power delay spectrum value behind the correcting frequency deviation, y _k(n) be n the multiple value of delay profile on the k phase point of son, N _yBe sliding length, the hop count of coherent accumulation value, M are calculated in the I segmentation _fBe the sampling point number of same phase point, the number of samples of M for add up at every turn, e is a natural constant, $j=\sqrt{-1}$ Expression imaginary unit, ∑ is a summation operation, ‖ ‖ is a modulo operation.

By relatively finding, technical scheme difference with the prior art of the present invention is, use three grades of coarse search methods of removing sampling point to search for the phase point of several candidate's path positions, and use the method for consecutive points stack to do frequency offset estimating to these phase points, but not the frequency offset estimating of the phase point of candidate's path position then obtains according to the frequency offset estimating of candidate's path position phase point, use the frequency offset estimating value to calculate PDP simultaneously, and be applied to Multipath searching and judgement.

Difference on this technical scheme, brought comparatively significantly beneficial effect, promptly make non-footpath signal and noise not influence frequency offset estimating, and the stack of footpath signal consecutive points has improved signal to noise ratio, thereby improve the accuracy of frequency offset estimating, PDP after the correction is applied to Multipath searching and judgement, makes system's multipath handling property be improved.

Description of drawings

Fig. 1 is an auto frequency control method flow chart according to an embodiment of the invention;

Fig. 2 is a coarse search method flow diagram according to an embodiment of the invention.

Embodiment

For making the purpose, technical solutions and advantages of the present invention clearer, the present invention is described in further detail below in conjunction with accompanying drawing.

The present invention provides the auto frequency control method in a kind of CDMA mobile communication systems, adopt the method for coarse search to determine several candidate footpath signal phase points from all phase points, and on these phase points, carry out AFC frequency offset estimating and correction, make non-footpath signal phase point can not influence the accuracy of estimation.

At first from mathematics, problem is carried out modeling, set forth idiographic flow of the present invention then.

If the discrete series of coherent signal can be expressed as c (n), n=1,2,3 ..., N _c, wherein n represents the sequence number of sequence, N _cSequence length for coherent signal; And the discrete series of input signal is represented with x (n), n=1, and 2,3 ..., N _x, wherein n represents the sequence number of sequence, N _xSequence length for input signal.

With the coherent signal sequence N that on input signal sequence, slides _yLength, then can obtain N _yIndividual phase point.Simultaneously the correlation of each phase point is divided into the I section and superposes, every segment length is $N_a=\frac{N_c}{I},$ Can obtain I length like this is N _ySub-CDP, use y _k(i) expression.Its computing formula is as follows:

Formula one $y_k\left(i\right)=\underset{n=1+(i-1)N_a}{\overset{i{N}_a}{\mathrm{\Σ}}}c\left(n\right)x(n+k),$ ?i＝1，2，3，...，I，k＝1，2，3，...，N _y。

The sampling point of same phase point is carried out coherent accumulation,, add up at twice here, so also can reduce memory space and computation complexity in order to preserve intermediate object program.For the first time, every M sampling point adds up, and obtains

Individual result, $z_k\left(j\right)=\underset{i=1+(j-1)M}{\overset{\mathrm{jM}}{\mathrm{\Σ}}}{y}_k\left(i\right),$ k＝1，2，3，...，N _y， $j=\mathrm{1,2},...,\frac{I}{M};$ For the second time, the result who obtains is for the first time added up, obtain ${z\′}_k=\underset{j=1}{\overset{\frac{I}{M}}{\mathrm{\Σ}}}{z}_k\left(j\right).$

Can calculate the PDP value according to the relevant accumulated value of each phase point is:

Formula two $P_k={\left|\right|{z\′}_k\left|\right|}^2={\left|\right|\underset{j=1}{\overset{\frac{I}{M}}{\mathrm{\Σ}}}{z}_k\left(j\right)\left|\right|}^2,$ k＝1，2，3，...，N _y

Wherein ‖ ‖ represents modulo operation.

Consider the distribution that multipath signal only can disperse on each phase point, thus adopt three grades to go the method for sampling point on all phase points, to carry out coarse search here, from N _yDetermine L phase point on the individual phase point as candidate footpath signal location, concrete coarse search process is as follows:

Search step one is with all N _yIndividual phase point is set to effectively, i.e. k ∈ V, and wherein V represents the set of effective phase point sequence number; Put search sequence number initial value a: l=1, what wherein search point sequence number l sign epicycle searched is l footpath signal phase point, l=1, and 2,3 ..., L, the subsequent L that carries out altogether takes turns search;

Search step two searches out the peaked phase point of PDP from effective phase point, maximum is expressed as

$P_{k_l}=\underset{k\∈V}{\max }\left\{P_k\right\},$ k _lBe the phase point sequence number of the PDP maximum correspondence that searches of epicycle (l wheel), max{} represents maximizing; This phase value adjacent about each 3 phase value be set to invalidly, promptly put k _l-3, k _l-2, k _l-1, k _l+ 1, k _l+ 2, k _l+ 3 is equal

, the phase value distribution that can guarantee two candidate footpath signals like this is at interval greater than 3; Increase progressively l=l+1 with the time point sequence number; Here because in actual applications, often because the diffusion of frequency deviation causes the adjacent phase point of phase place point value influence of footpath signal, and make that the value of adjacent phase point may be greater than the phase point of other footpath signals, if do not remove the influence of adjacent phase point, will cause the true directly illusion of signal crested, influence the performance of AFC;

Search step three, judging whether to finish search is l〉L, if then finish search, otherwise change search step two, continue search.

The sequence number of the candidate footpath signal phase point that search obtains is kept at k _l, l=1,2,3 ..., among the L.Frequency deviation that can calculated candidate path position phase point is as follows:

Formula three, $f_{k_l}=\frac{1}{N_d}A\tan \left(\underset{j=1}{\overset{\frac{I}{M}-N_d}{\mathrm{\Σ}}}{z}_{k_l}^{*}\left(j\right)z_{k_l}(j+N_d)\right),l=\mathrm{1,2},...,L.$

Wherein, N _dBe time delay value, can get following arbitrary value according to actual needs $N_d\∈\{\mathrm{1,2,3},...,\frac{I}{M}-1\};$ Atan represents arctan function, z ^*The conjugate complex number of expression z.

Those of ordinary skill in the art are appreciated that the formula of aforementioned calculation frequency deviation, and according to the needs of practical application, summation scope and mode can appropriate changes, and do not influence the spirit and scope of the invention.Calculating the formula of frequency deviation, also can be following several:

Formula five, $f_{k_l}=\frac{1}{N_d}A\tan \left(\underset{m=-N_1}{\overset{N_2}{\mathrm{\Σ}}}\underset{j=1}{\overset{\frac{I}{M}-N_d}{\mathrm{\Σ}}}{z}_{k_l+m}^{*}\left(j\right)z_{k_l+m}(j+N_d)\right),l=\mathrm{1,2},...,L.$

N ₁∈{1，2，3}，N ₂∈{1，2，3}；

Formula six $f_{k_l}=f=\frac{1}{N_d}A\tan \left(\underset{l=1}{\overset{L}{\mathrm{\Σ}}}\underset{j=1}{\overset{\frac{I}{M}-N_d}{\mathrm{\Σ}}}{z}_{k_l}^{*}\left(j\right)z_{k_l}(j+N_d)\right),l=\mathrm{1,2},...,L.$

Formula seven $f_{k_l}=f=\frac{1}{N_d}A\tan \left(\underset{l=1}{\overset{L}{\mathrm{\Σ}}}\underset{m=-N_1}{\overset{N_2}{\mathrm{\Σ}}}\underset{j=1}{\overset{\frac{I}{M}-N_d}{\mathrm{\Σ}}}{z}_{k_l+m}^{*}\left(j\right)z_{k_l+m}(j+N_d)\right),l=\mathrm{1,2},...,L.$

N _l∈{1，2，3}，N ₂∈{1，2，3}；

Wherein, formula five and formula seven are with three value coherent superposition about the phase point of candidate's path position, to improve signal to noise ratio, because consider in actual conditions, the phase point of candidate's path position can influence adjacent several phase points, so the value of consecutive points also can be attached candidate footpath signal message, so its stack can be improved the footpath signal energy, reduce the influence that noise brings, thereby improve the performance of frequency offset estimating.

In addition, formula six and formula seven are averaged the frequency offset estimating value of candidate's path position of obtaining, replace the frequency offset estimating value in candidate footpath with mean value, reduce multipath signal randomness like this, though can guarantee minimum estimated accuracy, the correcting frequency deviation performance does not have formula three and formula five good like that.

Like this, we have just drawn the frequency offset estimating value of candidate's path position, can be applied among the AFC of CDMA mobile communication system.

In a preferred embodiment of the present invention, phase value to non-candidate's path position is also estimated, the frequency deviation of other phase points can be represented with the mean value of the frequency offset estimating value of the phase value of candidate's path position, making does not like this have the frequency offset estimating value of the phase point of footpath signal to draw according to the footpath signal, shielded the influence that has the phase point of noise originally, when improving signal to noise ratio and reduce the number of the phase point of frequency offset estimating, its computing formula is as follows:

$f_{k\′}=\frac{1}{L}\underset{l=1}{\overset{L}{\mathrm{\Σ}}}{f}_{k_l},$ k′≠k _l，l＝1，2，...，L.

Calculate the phase point frequency deviation value of candidate's path position of gained for formula five and formula seven, formula eight also can be expressed as f _{K '}=f, k ' ≠ k _l, l=1,2,3 ..., L.

Perhaps can be according to nearby principle, get the frequency offset estimating value of the phase value of the adjacent nearest candidate's path position of sequence number, be formulated as follows:

Formula nine $f_{k\′}=f_{{k\′}_l},$ K ' l satisfies

Wherein min|| represents the minimum that takes absolute value.

In a preferred embodiment of the present invention, also calculated the later PDP of correcting frequency deviation: earlier each phase point is carried out correcting frequency deviation, sampling point to same phase point carries out coherent accumulation then, at last to the accumulation result delivery square after carry out irrelevant adding up again, computing formula is as follows:

Formula ten ${P\′}_k=\underset{i=1}{\overset{\frac{I}{M_f}}{\mathrm{\Σ}}}{\left|\right|\underset{n=1+(i-1)M_f}{\overset{{\mathrm{iM}}_f}{\mathrm{\Σ}}}{y}_k\left(n\right)e^{-j\frac{{\mathrm{nf}}_k}{M}}\left|\right|}^2,k=\mathrm{1,2,3},...,N_y,$

Wherein, P ' _kThe later PDP value of expression correcting frequency deviation, M _fBe the sampling point number of same phase point, f _kFrequency offset estimating value for obtaining previously particularly points out here $j=\sqrt{-1}$ Expression imaginary unit.

Again the PDP value behind the correcting frequency deviation that calculates is applied in Multipath searching and the judgement, will improves the Multipath searching performance greatly.

In conjunction with top described, can summarize AFC method of the present invention, Fig. 1 shows the flow process of AFC method.

Step 11, it is relevant that coherent signal sequence and input signal sequence are slided, and obtains sub-CDP, and computing formula is a formula one, and the sub-CDP that calculates can be used to calculate PDP and coherent accumulation value

Then enter step 12, coherent accumulation value sequence and PDP are calculated in segmentation, and the PDP expression formula is a formula two, and wherein the coherent accumulation segmentation is carried out, and obtains the relevant sequence of certain-length, can reduce complexity;

Then enter step 13, according to the phase point of several candidate's path positions of PDP coarse search, be used for the phase point of candidate's path position is carried out frequency offset estimating, wherein the method for coarse search provides hereinafter;

Then enter step 14, the frequency offset estimating value of the phase point of calculated candidate path position, computing formula can be any in formula three, formula five, formula six, the formula seven;

Then enter step 15, calculate the frequency offset estimating value of other phase points, because the phase point frequency offset estimating of non-candidate's path position determines that according to candidate footpath signal the The noise that can avoid non-candidate's path position to bring, computing formula can be any in formula eight, the formula nine;

Then enter step 16, calculate the later PDP of correcting frequency deviation, can be used for Multipath searching and judgement, computing formula is a formula ten.

Describe the method for carrying out coarse search in the step 13 for all phase points below in detail, i.e. three grades of methods of removing sampling point, Fig. 2 shows the flow process of this method:

Step 21 is changed to all phase points effectively, puts search sequence number initial value;

Then enter step 22, search out the phase point of the peaked phase point of PDP as candidate's path position from effective phase point, with each 3 point about this phase point, it is invalid that promptly adjacent 6 points are changed to, and will search for sequence number and increase one; Here the consecutive points of the phase point of the candidate's path position that searches being removed, make two candidate's path positions can not be too near to, also is to consider that candidate footpath signal often has certain width and can cause misjudgement in the reality, and also therefore being called three grades goes the sampling point method;

Then enter step 23, judge whether that according to the search sequence number search finishes, if, promptly search for sequence number and count above the phase place that will search for, then finish this search; Otherwise, return step 22 and continue search.

Those of ordinary skill in the art are appreciated that has related parameter and computational methods to revise in announcement scope of the present invention in the foregoing description.The frequency deviation estimating method of fragment sequence length value M, search candidate footpath number L, non-candidate's path position phase point etc. for example.

Description to embodiments of the invention makes any technical staff in the art can both use the present invention above.It is apparent to one skilled in the art that these embodiment are done various modifications, and total principle as described herein can be applied to other embodiment and not need creativity.Therefore the present invention embodiment of not being subject to here to be enumerated, and have the maximum magnitude that is consistent with disclosed principle and novel characteristics here.

Though by reference some preferred embodiment of the present invention, the present invention is illustrated and describes, but those of ordinary skill in the art should be understood that, can do various changes to it in the form and details, and the spirit and scope of the present invention that do not depart from appended claims and limited.

Claims (13)

1. the auto frequency control method in the CDMA mobile communication systems is characterized in that, comprises following steps:

It is relevant that coherent signal sequence and input signal sequence are slided, and calculates the multiple delay profile of son;

According to the multiple delay profile of described son, coherent accumulation value sequence and power delay spectrum are calculated in segmentation;

Phase point according to described several candidate's path positions of power delay spectrum coarse search;

According to the numerical value of described coherent accumulation value and described candidate's path position place phase point, the frequency offset estimating value of the phase point of calculated candidate path position correspondence.

2. according to the auto frequency control method in the described CDMA mobile communication systems of claim 1, it is characterized in that the described step of calculating the multiple delay profile of son is: with length is N _cCoherent signal sequence c (n), n=1,2,3 ..., N _c, be N in length _xInput signal sequence x (n), n=1,2,3 ..., N _xLast slip N _yLength, then correlation is divided into the I section and superposes, every segment length is $N_a=\frac{N_c}{I},$ Obtaining I length is N _yThe multiple delay profile of son, use y _k(i) expression, its computing formula is as follows:

$y_k\left(i\right)=\underset{n=1+(i-1)N_a}{\overset{i{N}_a}{\mathrm{\Σ}}}c\left(n\right)x(n+k),i=\mathrm{1,2,3},...,I,k=\mathrm{1,2,3},...,N_y,$

Wherein ∑ is a summation operation.

3. according to the auto frequency control method in the described CDMA mobile communication systems of claim 1, it is characterized in that the formula that the coherent accumulation value sequence is calculated in described segmentation is:

$z_k\left(j\right)=\underset{i=1+(j-1)M}{\overset{\mathrm{jM}}{\mathrm{\Σ}}}{y}_k\left(i\right),$ k＝1，2，3，...，N _y， $j=\mathrm{1,2,3},...,\frac{I}{M};$

The formula of rated output delay profile is:

$P_k={\left|\right|\underset{j=1}{\overset{\frac{I}{M}}{\mathrm{\Σ}}}{z}_k\left(j\right)\left|\right|}^2,k=\mathrm{1,2,3},...,N_y;$

Wherein, z _k(j) be j accumulated value on the k phase point, y _k(i) be i the multiple value of delay profile on the k phase point of son, N _yBe sliding length, the hop count of coherent accumulation value is calculated in the I segmentation, the number of samples of M for add up at every turn, P _kBe the power delay spectrum value, ∑ is a summation operation, || || be modulo operation.

4. according to the auto frequency control method in the described CDMA mobile communication systems of claim 3, it is characterized in that the method for described coarse search comprises step:

A is changed to all phase points effectively, puts search sequence number initial value;

B searches out the phase point of the peaked phase point of power delay spectrum as candidate's path position from effective phase point, with this phase point adjacent about each 3 point be changed to invalidly, and will search for sequence number and increase one;

C judges whether that according to the search sequence number search finishes, if, promptly search for sequence number and count above the phase place that will search for, then finish this search; Otherwise, return step B and continue search.

5. according to the auto frequency control method in the described CDMA mobile communication systems of claim 1, it is characterized in that described candidate's path position phase point frequency offset calculation method is, the coherent accumulation value of the time delay of will being separated by is carried out multiplying and is added up, and computing formula is:

$f_{k_l}=\frac{1}{N_d}A\tan \left(\underset{j=1}{\overset{\frac{1}{M}-N_d}{\mathrm{\Σ}}}{z}_{k_l}^{*}\left(j\right)z_{k_l}(j+N_d)\right),l=\mathrm{1,2},...,L.$

Wherein, l represents the sequence number in candidate footpath, k _lBe l the residing phase point numerical value in candidate footpath,

Be k _lAlso the frequency offset estimating value in l candidate footpath on the individual phase point, N _dFor time delay value can be got

In arbitrary value, Atan represents arctan function, I is the hop count that the coherent accumulation value is calculated in segmentation, the number of samples of M for add up at every turn,

Be k _lAlso be j accumulated value of candidate's path position on the individual phase point, L is a candidate footpath number, z ^*The conjugate complex number of expression z, ∑ is a summation operation.

6. according to the auto frequency control method in the described CDMA mobile communication systems of claim 1, it is characterized in that described candidate's path position phase point frequency offset calculation method is:

$f_{k_l}=\frac{1}{N_d}A\tan \left(\underset{m=-N_1}{\overset{N_2}{\mathrm{\Σ}}}\underset{j=1}{\overset{\frac{1}{M}-N_d}{\mathrm{\Σ}}}{z}_{k_l+m}^{*}\left(j\right)z_{k_l+m}(j+N_d)\right),l=\mathrm{1,2},...,L.$

Wherein, k _lBe l the residing phase point numerical value in candidate footpath,

Be k _lAlso the frequency offset estimating value of candidate's path position on the individual phase point, N _dFor time delay value can be got

In arbitrary value, Atan represents arctan function, I is the hop count that the coherent accumulation value is calculated in segmentation, the number of samples of M for add up at every turn, N ₁, N ₂Respectively get the arbitrary value in 1,2,3, z _k(j) be j accumulated value on k the phase point, L is the number in candidate footpath, z ^*The conjugate complex number of expression z, ∑ is a summation operation.

7. according to the auto frequency control method in the described CDMA mobile communication systems of claim 1, it is characterized in that, described candidate's path position phase point frequency offset calculation method is, the coherent accumulation value of the time delay of will being separated by is carried out multiplying and is added up, and the value of all candidate's path position phase points is averaged as the frequency offset estimating value, formula is:

$f_{k_l}=f=\frac{1}{N_d}A\tan \left(\underset{l=1}{\overset{L}{\mathrm{\Σ}}}\underset{j=1}{\overset{\frac{1}{M}-N_d}{\mathrm{\Σ}}}{z}_{k_l}^{*}\left(j\right)z_{k_l}(j+N_d)\right),l=\mathrm{1,2},...,L.$

Wherein, k _lBe l the residing phase point numerical value in candidate footpath,

Be k _lAlso be the frequency offset estimating value of candidate's path position on the individual phase point, f is the i.e. frequency offset estimating value of all candidate's path positions of the mean value of trying to achieve, N _dFor time delay value can be got

In arbitrary value, Atan represents arctan function, I is that the hop count of thousand accumulated values is mutually calculated in segmentation, the number of samples of M for add up at every turn,

Be k _lAlso be j accumulated value of candidate's path position on the individual phase point, L is the number in candidate footpath, z ^*The conjugate complex number of expression z, ∑ is a summation operation.

8. according to the auto frequency control method in the described CDMA mobile communication systems of claim 1, it is characterized in that described candidate's path position phase point frequency offset calculation method is:

$f_{k_l}=f=\frac{1}{N_d}A\tan \left(\underset{l=1}{\overset{L}{\mathrm{\Σ}}}\underset{m=-N_1}{\overset{N_2}{\mathrm{\Σ}}}\underset{j=1}{\overset{\frac{1}{M}-N_d}{\mathrm{\Σ}}}{z}_{k_l+m}^{*}\left(j\right)z_{k_l+m}(j+N_d)\right),l=\mathrm{1,2},...,L.$

Wherein, k _lBe l the residing phase point numerical value in candidate footpath, Be k _lAlso be the frequency offset estimating value of candidate's path position on the individual phase point, f is the mean value of frequency offset estimating value of the phase point of candidate's path position, N _dFor time delay value can be got

In arbitrary value, Atan represents arctan function, I is the hop count that the coherent accumulation value is calculated in segmentation, the number of samples of M for add up at every turn, N ₁, N ₂Respectively get the arbitrary value in 1,2,3, z _k(j) be j accumulated value on k the phase point, L is a candidate footpath number, z ^*The conjugate complex number of expression z, ∑ is a summation operation.

9. according to the auto frequency control method in the described CDMA mobile communication systems of claim 1, it is characterized in that, when described method is applied to calculating the phase point of non-candidate's path position, according to the frequency offset estimating of the phase point of candidate's path position, calculate the frequency offset estimating value of the phase point of non-candidate's path position.

10. according to the auto frequency control method in the described CDMA mobile communication systems of claim 9, it is characterized in that the method for frequency offset estimating value that is applied to calculate the phase point of non-candidate's path position is directly to represent with the mean value of the frequency offset estimating value of the phase value of all candidate's path positions.

11. according to the auto frequency control method in the described CDMA mobile communication systems of claim 9, it is characterized in that, the method of frequency offset estimating value that is applied to calculate the phase point of non-candidate's path position is, directly get adjacent with this non-candidate's path position phase point nearest also be the phase point frequency offset estimating value at the phase point place of minimum candidate's path position at interval.

12. according to the auto frequency control method in the described CDMA mobile communication systems of claim 9, it is characterized in that, when described method is applied to calculating the later power delay spectrum of correcting frequency deviation, frequency offset estimating value according to the phase point that calculates gained candidate's path position and non-candidate's path position, calculate the later power delay spectrum of correcting frequency deviation, be used for Multipath searching and judgement.

13. the auto frequency control method according in the described CDMA mobile communication systems of claim 12 is characterized in that, the formula that calculates the later power delay spectrum of correcting frequency deviation is:

${P\′}_k=\underset{i=1}{\overset{\frac{I}{M_f}}{\mathrm{\Σ}}}{\left|\right|\underset{n=1+(i-1)M_f}{\overset{{\mathrm{iM}}_f}{\mathrm{\Σ}}}{y}_k\left(n\right)e^{-j\frac{{\mathrm{nf}}_k}{M}}\left|\right|}^2,k=\mathrm{1,2,3},...,N_y,$

Wherein, P ' _kBe the power delay spectrum value behind the correcting frequency deviation, y _k(n) be n the multiple value of delay profile on the k phase point of son, N _yBe sliding length, the hop count of coherent accumulation value, M are calculated in the I segmentation _fBe the sampling point number of same phase point, the number of samples of M for add up at every turn, e is a natural constant, $j=\sqrt{-1}$ Expression imaginary unit, ∑ is a summation operation, || || be modulo operation.

Images