CN101129007B - Channel estimation method for TDD-CDMA system - Google Patents

Abstract

A method of channel estimate in TDD-CDMA system, including the following steps: finishing the simultaneity estimate of the channel impulse response in one time slot of each user using Steiner estimator, getting the originality time domain disperse channel impulse response of all the users hn,k=[hn, k, 1, hn,k, 2, hn,k, W], k=1, 2,..., K; calculating every tap adjudgement measurement of the time slot n of each user tn, k, i=atn-1, k, i+(l-a) hn, k, I|2, k=1, 2,..., K; all the taps getting tn,k=[tn, k, 1, tn, k, 2, tn, k, w]; carrying out the variable length multi time slots sliding average to the each tap of the time slot n of each user, getting the average value Fomula (I), i=1, 2,..., W, all the taps get hn,k=[hn, k, 1, hn, k, 2,..., h n, k, W], k =1, 2,..., K; filtering the noise taps to the time slot n of each user, getting h'n, k, i; outputting the channel estimate value h'n, k of a user in time slot n; repeating the above steps until gets the channel estimate values of all the users.

Description

The channel estimation methods of TDD-CDMA system

Technical field

The present invention relates to a kind of TDD-CDMA (Time Division Duplex-Code Division Multiple Access; Time division duplexing-CDMA) channel estimation methods of system; Relate in particular to a kind of channel estimation methods that adopts training sequence to carry out the TDD-CDMA system of channel estimating; Be applicable to and adopt training sequence to carry out all systems of channel estimating among the TDMA/CDMA, especially be suitable for TD-SCDMA and WCDMA TDD system in the 3-G (Generation Three mobile communication system).

Background technology

With TD-SCDMA (Time Division-Synchronous Code Division Multiple Access) system is example; As a kind of cell mobile communication systems; Several kinds are disturbed and noise below existing in the TD-SCDMA system: 1, owing to possibly there be a plurality of Code Channel transmission in the same time slot simultaneously; Under the mobile channel condition; Orthogonality between the Code Channel can be formed multiple access and disturb, be i.e. MAI (Multiple Access Interference) by partial destruction between the receiving terminal Code Channel; 2, because the multipath transmisstion of signal can form intersymbol interference at receiving terminal, i.e. ISI (Inter-Symbol Interference); 3, the thermal noise of receiver; 4, from the interference of other sub-district; 5, other interference.

The demodulation performance of TD-SCDMA system is generally by satisfying certain QOS (Quality of Service; Service quality) needed SIR of receiver time the (Signal to Interference Ratio, signal interference ratio) or SNR (Signal to Noise Ratio, signal to noise ratio) expression; Because honeycomb CDMA (Code Division Multiple Access; Code division multiple access) in the system, disturbs to have the statistical property of similar white noise, therefore generally not strict differentiation interference and noise.In order to obtain good demodulation performance, will find the solution and transfer algorithm to have good interference rejection capability.Inhibition ability to disturbing is strong more, and systematic function is good more, and the utilization ratio of the frequency spectrum that can provide is just high more.Therefore, interference mitigation technology becomes the key factor that improves the TD-SCDMA performance.

The sign indicating number design of TD-SCDMA system, frame structure, spreading rate etc. provide feasibility for interference mitigation technology.Associated detection technique is one of key technology of TD-SCDMA; In the TD-SCDMA system, use associated detection technique; Can effectively suppress MAI and ISI; Associated detection technique is to be based upon on the basis of channel estimating, and the effect of channel estimating has very significant effects to combined detection performance, and good channel estimation technique becomes the necessary condition of decision joint-detection.Usually, wireless mobile environment is modeled as the finite length impulse response model, and the task of channel estimating is exactly amplitude and the phase place that estimates each time delay tap.The multipath transmisstion of signal is very high to the requirement of channel estimating under the wireless mobile environment, how under the dissemination channel condition of complicacy, to obtain the key that channel estimating accurately becomes TD-SCDMA technology.

The method of estimation of existing TD-SCDMA system mainly contains following several kinds:

1, stainer (Steiner) method of estimation

In existing TD-SCDMA system; All users' of same sub-district training sequence is all through a training sequence midamble basic, that the cycle is P; Mode through cyclic shift produces, and corresponding channel impulse method of estimation generally all is a utilization Steiner method of estimation.Adopt the channel estimation methods of this method to be commonly referred to as the Steiner estimator, its estimation principle is to adopt least square method through separating the channel impulse response that system of linear equations once calculates all users.Because the training sequence of TD-SCDMA system has the cyclic shift characteristic; Therefore when calculating each user's channel impulse response, can use a correlator; Can use efficiently during concrete the realization, implementation method such as FFT (Fast Fourier Transfer fast; Fast Fourier transform) and IFFT (Inverse Fast Fourier Transfer; Inverse fast Fourier transform), the linear convolution of complicacy is converted into simple circular convolution, can reduces the algorithm complex and the cost of channel estimating.But the performance of Steiner estimator receives the influence of receiving terminal additive noise, and its channel impulse response that obtains is compared with the ideal communication channel impulse response, comprises noise component(s), the reason of this channel estimation errors generation just; Can enlarge noise power simultaneously, cause the relative input of output signal to noise ratio that a certain amount of loss is arranged, make the channel impulse response estimation decreased performance.What the Steiner estimator calculated is the subscriber channel impulse response estimated value of a time slot.Receive the restriction of transport overhead, the interior training sequence of time slot can not be oversize, and its performance receives the influence of receiving terminal additive noise and causes precision of channel estimation loss, directly uses its estimated result will inevitably cause the decline of systematic function if be left intact.

2, cross the processing method of the noise reduction of threshold judgement

Referring to document " Impact of Channel Estimation on the 3GPP-TD-CDMA "; Work such as Oliver

; GLOBECOM ' 01 IEEE:3365-3369vol.6 and " improvement of low-cost channel estimation methods in the TD-SCDMA system " Kang Shaoli; Work such as Qiu Zhengding; " communication journal ", 2002,23 (10): 125-130.Its basic thought is earlier to estimate original channel impulse response with the Steiner method of estimation, uses a thresholding then, does not have signal in the channel impulse response tap that the filtering estimation obtains and is the tap of noise contribution fully.Only increased very little amount of calculation in this way, can make channel estimating performance obtain certain improvement.

Do not have signal in the channel impulse response tap that this method obtains through the thresholding filtering Steiner estimator that is provided with and be the tap of noise contribution fully on Steiner estimator basis.Each tap of the channel impulse response that the Steiner estimator obtains is corresponding is that signal passes through mobile radio channel and propagates back each multipath composition when arriving receiver; Do not have only noise if do not contain signal in the tap; Then the merging of multipath just should not be participated in this tap by filtering; Otherwise merged is noise, will produce interference to the detection of useful signal.But this method does not still overcome the defective that existing 1 of method has been used the instantaneous estimation of a time slot; Therefore there are two shortcomings: first; The noise that contains in the channel impulse response in the tap of signal is not inhibited, and is superimposed upon the deterioration that noise on the signal causes signal detection performance; The second, the filtering method of noise tap is too simple, and is invalid to instantaneous very noisy, can be mistaken for signal by the very noisy of filtering, thereby interferer signal detection causes mis-behave.

3, zero correlation window sign indicating number method

The midamble sign indicating number that the method employing zero correlation window sign indicating number that this method proposes replaces existing TD-SCDMA standard code is as training sequence; How much use the code character of different zero correlation length according to user's in the system number; Can carry out a spot of shifting function to the pilot tone in user's time slot; If use smart antenna, can also be according to user's position, the configuration of the angle of arrival of signal; Make the configuration cross-correlation of angle of arrival neighboring user have the sign indicating number sequence of broad zero correlation window, the interference of the channel estimating that the minimizing system is inherent.

Because this method has adopted the zero correlation window sign indicating number that is different from existing TD-SCDMA standard as training sequence, the influence that the their cross correlation that exists between the reduction user's training sequence causes channel estimating, thus reach the purpose that improves channel estimating performance.But this method has two shortcomings: the first, need complicated training sequence allocation algorithm, a plurality of parallel correlator, cost is very high, and can't use efficiently, algorithm is realized fast; The second, this technical scheme and existing TD-SCDMA standard are incompatible.

Summary of the invention

The channel estimation methods that the purpose of this invention is to provide a kind of TDD-CDMA system; Can reduce in the TDD-CDMA system noise to the interference of signal; Can effectively suppress the interference of instantaneous very noisy, can realize the high performance channel estimator of TDD-CDMA efficiently, fast and cheaply, and strengthen the inhibition deviation capability of TDD-CDMA system; And compatible with current standard, need not existing standard is revised.

Be to realize above-mentioned purpose, the invention provides the channel estimation methods of a kind of TDD-CDMA system, it is characterized in that may further comprise the steps:

Step 1, estimate when utilizing the Steiner estimator to accomplish each user's the channel impulse response of a time slot, obtain all users' original time domain discrete channel impulse response h _{N, k}=[h _{N, k, 1}, h _{N, k, 2}... h _{N, k, W}], k=1,2 ..., K, n are time-gap numbers, and k is a Customs Assigned Number, and K is a number of users, and W is a channel impulse response length;

Step 2, each user calculated each tap decision metric t of time slot n _{N, k, i}=α t _{N-1, k, i}+ (1-α) | h _{N, k, i}| ², k=1,2 ..., K obtains the decision metric t of all taps _{N, k}=[t _{N, k, 1}, t _{N, k, 2}... t _{N, k, W}], wherein, 0＜α＜1; Step 3, elongated degree multi-slot moving average is carried out in each tap of the time slot n of each user k, obtain mean value I=1,2 ..., W obtains the mean value of all taps K=1,2 ..., K; Step 4, each user's time slot n is carried out the filtering noise tap handle, obtain h ' _{N, k, i} Step 5, user k of output are at time slot letter n road estimated value h ' _{N, k}, wherein, h ' _{N, k}=[h ' _{N, k, 1}, h ' _{N, k, 2}..., h ' _{N, k, W}];

Step 6, repeating step 1-5 are up to the channel estimation value that obtains all users.

In the said step 3 $c_j=\left\{\begin{array}{cc}\mathrm{\β},& j=0\\ \frac{1-\mathrm{\β}}{\underset{j=1}{\overset{J}{\mathrm{\Σ}}}j}j,& j=\mathrm{1,2},...,J\end{array}\right.,$ J is the number of time slots of carrying out the weighting moving average, the value of J between 1-16,0＜β＜1.When static or walking communication, the β value is 0.2～0.5; During high-speed mobile communications, the β value is 0.5～0.9.

Said step 4 is specially: step 4.1, calculating noise thresholding T; Step 4.2, each user's time slot is carried out the filtering noise tap handle, keep this tap if 10 times of the common logarithm of the decision metric of a tap are higher than noise filtering thresholding T otherwise this tap is changed to 0, promptly $h_{n,k,i}^{\′}=\left\{\begin{array}{cc}{\hat{h}}_{n,k,i}& ,10{\mathrm{Log}}_{10}\left(t_{n,k,i}\right)>T\\ 0& ,10{\mathrm{Log}}_{10}\left(t_{n,k,i}\right)\≤T\end{array}\right..$ Said noise gate T=10log ₁₀(P _Noise)+offset, P _NoiseBe the power of making an uproar at the end, offset is with respect to the make an uproar side-play amount of power of the end, and wherein, the power of making an uproar of the said end is to measure when not receiving signal through measuring receiver.Preferably, the power P of making an uproar of the described end _Noise=KT ' B+NF, K are Boltzmann constants, and T ' is an absolute temperature, and B is a signal bandwidth, and NF is a receiver noise factor.

Therefore, the present invention has the following advantages:

1, reduced the interference of noise, can effectively suppress the interference of instantaneous very noisy signal.

2, can use efficient, fast, low cost method realizes high performance channel estimator.

3, the inhibition deviation capability is strong.

4, compatible current standard need not existing standard is revised.

Below in conjunction with accompanying drawing and embodiment, technical scheme of the present invention is done further detailed description.

Description of drawings

Fig. 1 is the method flow diagram of the channel estimation methods of TDD-CDMA of the present invention system.

Fig. 2 handles sketch map for the user k of the channel estimation methods of TDD-CDMA of the present invention system in each tap calculation decision metric of original channel impulse response of time slot n.

Fig. 3 carries out the sketch map that elongated degree multi-slot average is handled for the user k of the channel estimation methods of TDD-CDMA of the present invention system in the original instantaneous channel estimating of time slot n.

Fig. 4 is the calculating user k of the channel estimation methods of the TDD-CDMA of the present invention system channel impulse response flow chart at time slot n.

Embodiment

Invention thought of the present invention provides a kind of novel TDD-CDMA channel estimation methods; From the digital spread spectrum signal that receives, extract training sequence symbols from a plurality of subscriber signals of being superimposed with of a plurality of travelling carriages or base station and noise and interference; Utilize local each user's who produces of base station receiver or mobile station receiver training sequence symbols; Estimate the channel impulse response of each user at a time slot; Then the channel impulse response of a plurality of time slots of same user is handled,, improved the precision of channel impulse response estimation to reduce The noise.

As shown in Figure 1, be method flow diagram of the present invention, the method for channel impulse response estimation according to the invention is following:

Step 1, estimate when using the Steiner estimator to accomplish each user's the channel impulse response of a time slot, obtain K user's original time domain discrete channel impulse response:

h _n，k＝[h _n，k，1，h _n，k，2，...h _n，k，W]，k＝1，2，...，K (1)

In the formula (1), n is a time-gap number, and k is a Customs Assigned Number, and K is a number of users, and W is a channel impulse response length, and W depends on the multidiameter delay expansion of mobile radio channel, TD-SCDMA system for example, and the typical span of W is 4～32 chips.

Step 2, each user calculated each tap decision metric of time slot n:

t _n，k＝[t _n，k，1，t _n，k，2，...t _n，k，W]，k＝1，2，...，K (2)

In the formula (2),

t _n，k，i＝αt _n-1，k，i+(1-α)|h _n，k，i| ²?i＝1，2，...，W，k＝1，2，..，K (3)

t _n，k，i＝0，n≤0，i＝1，2，...，W，k＝1，2，...，K (4)

In the formula (3), α is a forgetting factor, 0＜α＜1, and general value is 0.1～0.95.Because this step has adopted the processing method of noncoherent accumulation, can improve the ability of the anti-frequency deviation of channel estimating.

Referring to shown in Figure 2, for the user k of the channel estimation methods of system of the present invention handles sketch map in each tap calculation decision metric of original channel impulse response of time slot n.Comprise: W delivery also asked square operation device 202-1～202-W; W multiplier 203-1～203-W; W multiplier 204-1～204-W; W adder 205-1～205-W; W delayer 206-1～206-W, the output of delayer is than time slot of input time delay.Input 201-1～201-W h _{N, k, 1}, h _{N, k, 2}... h _{N, k, W}It is the original channel impulse response of the user k that obtains by the Steiner estimator at time slot n.Containing W identical processing procedure among this figure, is that example is described its processing procedure with first processing procedure, and other process is identical with it.Delivery and first tap h that asks square operation device 202-1 that the original channel of time slot n is estimated _{N, k, 1}Carry out delivery and ask square operation to obtain | h _{N, k, 1}| ², multiplier 203-1 is right | h _{N, k, 1}| ²Take advantage of a coefficient (1-α) to obtain (1-α) | h _{N, k, 1}| ², on the other hand, the t of the previous time slot of delayer 206-1 output _{N-1, k, 1}, multiplier 204-1 is to t _{N-1, k, 1}Take advantage of an alpha to obtain α t _{N-1, k, 1}, adder 205-1 obtains the output addition of the output of multiplier 203-1 and multiplier 204-1 first tap decision metric t at time slot n of user k _{N, k, 1}=α t _{N-1, k, 1}+ (1-α) | h _{N, k, 1}| ², last, above-mentioned W identical processing procedure exported W the tap decision metric t of user k at the channel impulse response of time slot n _{N, k}=[t _{N, k, 1}, t _{N, k, 2}... t _{N, k, W}].

Step 3, elongated degree multi-slot moving average is carried out in each tap of the time slot n of each user k:

${\hat{h}}_{n,k,i}=\underset{j=0}{\overset{J}{\mathrm{\Σ}}}{c}_j{h}_{n-j,k,i},$ i＝1，2，...，W，k＝1，2，...，K (5)

h _n-j，k，i＝0，n-j≤0，i＝1，2，...，W，k＝1，2，...，K (6)

${\hat{h}}_{n,k}=[{\hat{h}}_{n,k,1},{\hat{h}}_{n,k,2},...{\hat{h}}_{n,k,W}],$ k＝1，2，...，K (7)

In the formula (5), moving average length J is the number of time slots of carrying out the weighting moving average, c _jIt is weight coefficient.Moving average length J has comprised two kinds of possible situation: single time slot services and multislot services.For single time slot services, J just equals to participate in the moving average number of sub-frames, and for multislot services, J is relevant with multislot services institute's distributed time slot number in a sub-frame with the moving average number of sub-frames.The value of J is decided according to the speed of channel variation, for the user that low speed moves, communicates by letter during for example static or walking, and the J value can obtain good inhibition noise ability more greatly; For the user of high-speed mobile, the communication in the vehicle that for example goes, the J value is less can be in the quick variation of the simultaneous adaptation channel that obtains satisfied inhibition noise ability.The span of J can be unrestricted, but generally its value is 1～16.c _jValue also relevant with the speed of channel variation, for example a kind of possible value is following:

$c_j=\left\{\begin{array}{cc}\mathrm{\β},& j=0\\ \frac{1-\mathrm{\β}}{\underset{j=1}{\overset{J}{\mathrm{\Σ}}}j}j,& j=\mathrm{1,2},...,J\end{array}\right.---\left(8\right)$

In the formula (8), 0＜β＜1.For the user that low speed moves, to communicate by letter during for example static or walking, the β value is 0.2～0.5; For the user of high-speed mobile, the communication in the vehicle of for example running at high speed, the β value is 0.5～0.9.

Referring to shown in Figure 3, for user k of the present invention carries out the sketch map that elongated degree multi-slot average is handled in the original instantaneous channel estimating of time slot n.User k is carried out elongated degree multi-slot average in the original instantaneous channel estimating of time slot n handle, comprising: J+1 multiplier 302-0～302-J; J delayer 303-1～303-J, the output of delayer is than time slot of input time delay; 1 adder 304.Input 101h _{N, k}=[h _{N, k, 1}, h _{N, k, 2}... h _{N, k, W}] be the original channel impulse response of each user k of obtaining by the Steiner estimator at a time slot n, the original channel of time slot n is estimated h _{N, k}All tap h _{N, k, 1}, h _{N, k, 2}... h _{N, k, W}, multiplier 302-0～302-J each tap i (i=1,2 ..., channel estimating h W) _{N, k, i}And the original channel of preceding J the time slot of the tap i of J delayer 303-1～303-J output is estimated h _{N-1, k, i}, h _{N-2, k, i}... h _{N-J, k, i}With corresponding weight coefficient c _jMultiply each other, obtain c _jh _{N-j, k, i}, j=1,2 ..., J, adder 304 obtains the output addition of J+1 multiplier 302-0～302-J then

I=1,2 ..., W, the final channel estimating 305 of user k behind the multi-slot average of time slot n that get ${\hat{h}}_{n,k}=[{\hat{h}}_{n,k,1},{\hat{h}}_{n,k,2},...{\hat{h}}_{n,k,W}],$ K=1,2 ..., K

Step 4, each user's time slot n is carried out the filtering noise tap handle

$h_{n,k,i}^{\′}=\left\{\begin{array}{cc}{\hat{h}}_{n,k,i}& ,10{\log }_{10}\left(t_{n,k,i}\right)>T\\ 0& ,10{\log }_{10}\left(t_{n,k,i}\right)\≤T\end{array}\right.---\left(9\right)$

In the formula (9), T is the noise filtering thresholding, and according to the operational environment setting of reality, for example a kind of possible obtaining value method is following:

T＝10log ₁₀(P _noise)+offset (10)

In the formula (10), P _NoiseBe the power of making an uproar at the end, when not receiving signal, measure, also can calculate according to following formula (11) through measuring receiver.P _NoiseOnce confirming generally not need adjustment.Offset is that its span is unrestricted, can obtain optimal value through Computer Simulation or experiment usually with respect to the make an uproar side-play amount of power of the end.

P _noise＝KT′B+NF (11)

In the formula (11), K is a Boltzmann constant, K=1.38 * 10 ²³, T ' is an absolute temperature, and B is a signal bandwidth, and NF is a receiver noise factor.The unit of K be burnt/open (J/K), the unit of T ' is out (K), the unit of B is hertz (Hz), the unit of NF is a decibel (dB).

Step 5, output user k are at time slot n channel estimation value h ' _{N, k}:

h′ _n，k＝[h′ _n，k，1，h′ _n，k，2，...，h′ _n，k，W] (12)

Referring to shown in Figure 4, for the present invention calculates the channel impulse response flow chart of user k at time slot n.Filter thresholding T according to formula (8) calculating noise, from tap i=1, whether judges k satisfies 10log at each tap i of the channel impulse response of time slot n ₁₀(t _{N, k, i})＞T if judge establishment, then keeps the channel estimating of this tap, promptly

Increase progressively otherwise tap is changed to 0, one tap back tap counter i that disposes, dispose, export user k at last at time slot n channel estimation value h ' up to all W tap _{N, k}=[h ' _{N, k, 1}, h ' _{N, k, 2}..., h ' _{N, k, W}]

Step 6, above step are the operations to a user, repeat to obtain for K time all users' channel estimating.

The present invention be a kind of suitable TDD-CDMA standard efficiently, fast, combination channel estimating method cheaply; On Steiner estimator basis; Adopted a kind of multi-slot to handle the processing method that combines with noncoherent accumulation, overcome the defective of prior art, reduced the interference of noise signal; Can effectively suppress the interference of instantaneous very noisy, also have good anti-frequency deviation performance simultaneously.Channel estimation methods of the present invention has increased very little amount of calculation on Steiner estimator basis, can use efficient, fast, low cost method realizes high performance channel estimator.

Therefore this method has following advantage: the first, reduced noise to the interference of signal and can effectively suppress the interference of instantaneous very noisy; The second, have good anti-frequency deviation performance; The 3rd, can use efficient, fast, low cost method realizes high performance channel estimator.The 4th, compatible current standard need not existing standard is revised.

The present invention also is applicable to and adopts training sequence to carry out all systems of channel estimating among the TDMA/CDMA; Especially be applicable to TD-SCDMA and WCDMA TDD system in the 3-G (Generation Three mobile communication system); Though technical scheme of the present invention is primarily aimed at the wireless communication system of TDD-CDMA; But be applicable to the CDMA, the tdma system that adopt similar transmission structure too, both be applicable to base station receiver, also be applicable to mobile station receiver.In conjunction with the measurement of correlation process of receiver, this aspect can be adaptively according to the operational environment dynamically-adjusting parameter.

It should be noted last that; Above embodiment is only unrestricted in order to technical scheme of the present invention to be described; Although with reference to preferred embodiment the present invention is specified, those of ordinary skill in the art should be appreciated that and can make amendment or be equal to replacement technical scheme of the present invention; And not breaking away from the spirit and the scope of technical scheme of the present invention, it all should be encompassed in the middle of the claim scope of the present invention.

Claims (4)

1. the channel estimation methods of a TDD-CDMA system is characterized in that may further comprise the steps:

Step 1, estimate when utilizing the Steiner estimator to accomplish each user's the channel impulse response of a time slot, obtain all users' original time domain discrete channel impulse response h _{N, k}=[h _{N, k, 1}, h _{N, k, 2}... h _{N, k, W}], k=1,2 ..., K, n are time-gap numbers, and k is a Customs Assigned Number, and K is a number of users, and W is a channel impulse response length;

Step 2, each user calculated each tap decision metric t of time slot n _{N, k, i}=α t _{N-1, k, i}+ (1-α) | h _{N, k, i}| ², k=1,2 .., K obtains the decision metric t of all taps _{N, k}=[t _{N, k, 1}, t _{N, k, 2}... t _{N, k, W}], wherein, 0＜α＜1;

Step 3, elongated degree multi-slot moving average is carried out in each tap of the time slot n of each user k, obtain mean value

I=1,2 ..., W obtains the mean value of all taps ${\hat{h}}_{n,k}=[{\hat{h}}_{n,k,1},{\hat{h}}_{n,k,2},...{\hat{h}}_{n,k,W}],$ K=1,2 ..., K, wherein, $c_j=\left\{\begin{array}{cc}\mathrm{\β},& j=0\\ \frac{1-\mathrm{\β}}{\underset{j=1}{\overset{J}{\mathrm{\Σ}}}j}j,& j=\mathrm{1,2},...,J\end{array}\right.,$ J is the number of time slots of carrying out the weighting moving average, the value of J between 1-16,0＜β＜1;

Step 4, each user's time slot n is carried out the filtering noise tap handle, obtain h ' _{N, k, i}, comprise the steps: step 4.1, calculating noise thresholding T; Step 4.2, each user's time slot is carried out the filtering noise tap handle, keep this tap if 10 times of the common logarithm of the decision metric of a tap are higher than noise filtering thresholding T otherwise this tap is changed to 0, promptly $h_{n,k,i}^{\′}=\left\{\begin{array}{cc}{\hat{h}}_{n,k,i}& ,10{\mathrm{Log}}_{10}\left(t_{n,k,i}\right)>T\\ 0& ,10{\mathrm{Log}}_{10}\left(t_{n,k,i}\right)\≤T\end{array}\right.;$

Step 5, user k of output are at time slot letter n road estimated value h ' _{N, k}, wherein, h ' _{N, k}=[h ' _{N, k, 1}, h ' _{N, k, 2}..., h ' _{N, k, W}];

Step 6, repeating step 1-5 are up to the channel estimation value that obtains all users.

2. the channel estimation methods of TDD-CDMA according to claim 1 system is characterized in that: when static or walking communication, the β value is 0.2～0.5; During high-speed mobile communications, the β value is 0.5～0.9.

3. the channel estimation methods of TDD-CDMA according to claim 1 system is characterized in that: said noise filtering thresholding T=10log ₁₀(P _Noise)+offset, P _NoiseBe the power of making an uproar at the end, offset is with respect to the make an uproar side-play amount of power of the end, and wherein, the power of making an uproar of the said end is to measure when not receiving signal through measuring receiver.

4. the channel estimation methods of TDD-CDMA according to claim 3 system is characterized in that: the power P of making an uproar of the described end _Noise=KT ' B+NF, K are Boltzmann constants, and T ' is an absolute temperature, and B is a signal bandwidth, and NF is a receiver noise factor.

Images