Summary of the invention
In order to solve the problems of the technologies described above, the invention provides a kind of beam form-endowing method of space-time unite, can be applicable to the various antenna systems that are under the adverse circumstances, solve effectively owing to the problem that reaches the transmission signals deterioration that causes than big intersymbol interference than the long time delay expansion.
Space-time united beam form-endowing method of the present invention comprises:
A, the signal that receives according to every antenna in the smart antenna array carry out channel estimating, obtain corresponding each user's of every antenna channel impulse response sequence;
B, according to corresponding each user's of described every antenna channel impulse response sequence, calculates every each user's of antenna correspondence weight vectors;
Corresponding each user's of every antenna that C, use calculate weight vectors carries out time domain preequalization Filtering Processing to each user data to be sent, for each chip in described each user data to be sent generates one or more wave beam according to different multidiameter delays;
The wave beam that D, described smart antenna array generate according to step C sends each chip in each user data to be sent.
The described channel estimating of steps A may further comprise the steps:
A1, from the signal that every antenna of smart antenna array receives, extract training sequence respectively, and calculate the channel impulse response of every antenna according to the training sequence that extracts and known basic training sequences;
A2, from corresponding each user's of the channel impulse response of every antenna window, extract corresponding each user's of every antenna channel impulse response sequence.
The channel impulse response of every antenna of steps A 1 described calculating comprises:
A11, respectively training sequence and the described basic training sequences extracted are carried out fast fourier transform from an antenna receiving signal;
A12, will be divided by by corresponding sampled point through the training sequence and the basic training sequences of fast fourier transform;
A13, the merchant that steps A 12 is calculated carry out inverse fast fourier transform, and the sequence that obtains is the channel impulse response of this antenna.
Corresponding each user's of every antenna of the described calculating of step B weight vectors comprises:
B1, according to corresponding each user's of described every antenna channel impulse response sequence, corresponding each user's of computational intelligence antenna array channel impulse response power sequence;
B2, from the corresponding user's of described smart antenna array channel impulse response power sequence, find out N sampled point of performance number maximum, and according to every antenna to channel impulse response that should the user at the value on this N sampled point and smart antenna array weight vectors to each user of every antenna correspondence of performance number calculating of channel impulse response power sequence on this N sampled point that should the user;
Wherein, described N is the integer more than or equal to 1.
Corresponding each user's of the described computational intelligence antenna array of step B1 channel impulse response power sequence comprises:
B11, according to corresponding each user's of described every antenna channel impulse response sequence, calculate the performance number of each sampled point of channel impulse response sequence of corresponding each user of every antenna, obtain corresponding each user's of every antenna channel impulse response power sequence;
B12, with the corresponding same users' of all antennas in the smart antenna array channel impulse response power sequence by corresponding sampled point addition, obtain this smart antenna array to channel impulse response power sequence that should the user.
Corresponding each user's of every antenna of the described calculating of step B2 weight vectors is: the weight vectors of establishing corresponding k the user of m root antenna is
Wherein, p weighted factor W
p M, kAdopt following formula to calculate:
Wherein, h
M, k(p) be the value of channel impulse response sequence on p the sampled point corresponding with p weighted factor of m root antenna respective user k, span 1~P of p, P are the channel impulse response sequence length of m root antenna respective user k; H
k(x
Maxj) be that the channel impulse response power sequence of smart antenna array respective user k is at sampled point x
MaxjValue; x
Maxj, x
MaxnAll represent one of N sampled point of performance number maximum in the channel impulse response power sequence of smart antenna array respective user k; Operator conj[] the expression conjugate operation.
The method of the invention further comprises behind the weight vectors that calculates corresponding each user of every antenna: the weighted factor sequence of intercepting from corresponding each user's of described every antenna weight vectors respectively from first non-vanishing weighted factor to a last non-vanishing weighted factor, and with the weighted factor sequence of described intercepting weight vectors as corresponding each user of every antenna.
The described preequalization Filtering Processing of step C is specially: the data sequence that user is to be sent is carried out convolution with the every antenna that calculates to weight vectors that should the user respectively, obtains the data sequence that this user will send on every antenna.
Convolution of the present invention realizes with digital signal processing chip or digital filter.
This shows, space-time united beam form-endowing method of the present invention replaces weighted factor of the prior art that user's information is carried out pre-equalization process on the time domain by using the weight vectors that generates according to the characteristic of channel, the footpath that realization is bigger with receive time delay sends ahead of time, and the footpath delayed delivery that receive time delay is less, can guarantee the signal that sent through different paths the time delay, substantially simultaneously arrive receiving terminal, make the main concentration of energy of transmission signals receive on the chip at one, thereby can improve the signal to noise ratio of receiving end signal, reduce intersymbol interference, improve the reliability of signal transmission.
Embodiment
For the purpose, technical scheme and the advantage that make invention is clearer, below with reference to the accompanying drawing embodiment that develops simultaneously, the present invention is described in further detail.
Space-time united beam form-endowing method of the present invention is when wave beam forming, generate corresponding each user's of every antenna weight vectors according to the channel impulse response sequence of every corresponding each user of antenna in the smart antenna array, and use the weight vectors that generates according to the characteristic of channel to replace weighted factor of the prior art that user's information to be sent is carried out preequalization Filtering Processing on the time domain, improve the signal to noise ratio of receiving end signal, reduce intersymbol interference, thereby improve the reliability of transmission signals.
Fig. 2 has shown the basic procedure of space-time united beam form-endowing method of the present invention.As shown in Figure 2, method of the present invention mainly may further comprise the steps:
A, the signal that receives according to every antenna in the smart antenna array carry out channel estimating, obtain corresponding each user's of every antenna channel impulse response.
This step further comprises following two steps:
A1, the signal extraction training sequence that every antenna receives from smart antenna array respectively, and according to the channel impulse response of training sequence that extracts and known every antenna of basic training sequences estimation.
Suppose that smart antenna array comprises M root antenna altogether, the training sequence that extracts from the received signal of m root antenna wherein is Midamble
m(1:128), the span of m from 1 to M.
This training sequence Midamble
m(1:128) comprise 128 sampled points, produce by basic training sequences Basic Midamble (1:128) cyclic shift, the training sequence of different user is with respect to the displacement difference of basic training sequences, even therefore in same time slot, have a plurality of users to transmit described training sequence simultaneously, also can distinguish mutually by the different displacements of training sequence.
Here, the channel impulse response of supposing m root antenna is h
m(1:128), have 128 sampled points, its computing formula is as follows:
Wherein, operator FFT and IFFT represent fast fourier transform and inverse fast fourier transform operation respectively.
In addition, because the training sequence that each user uses is all produced through cyclic shift by basic training sequences, and the training sequence of different user is with respect to the displacement difference of basic training sequences, this just is equivalent to the time delay difference of the channel impulse response correspondence of different user, therefore, can from the channel impulse response of every antenna, isolate corresponding each user's of every antenna channel impulse response sequence according to the time delay of different user channel impulse response.Concrete grammar is: the channel impulse response of every antenna evenly is divided into several windows according to number of users, and the corresponding user of window, this corresponding relation is determined by the displacement of training sequence that this user uses.Channel impulse response sequence in certain window is exactly the channel impulse response sequence of this window respective user, the length of this sequence be 128 with the merchant of number of users.
Fig. 3 has shown when having 8 users to communicate by letter simultaneously, the channel impulse response of certain root antenna in the smart antenna array.As shown in Figure 3, the channel impulse response of this antenna has comprised 128 sampled points, 8 crests altogether, the position respectively 0,16,32, near the 48......112.This channel impulse response on average is divided into 8 windows according to number of users, i.e. 8 intervals: [7,8], [9,24], [25,40] ... [89,104], [105,120], sequence in each window is exactly the channel impulse response sequence of certain user among corresponding these 8 users that communicate by letter simultaneously of this antenna, and this sequence length is 16.
A2, from corresponding each user's of every antenna channel impulse response window, extract corresponding each user's of every antenna channel impulse response.
In this step, the channel impulse response sequence of establishing corresponding k the user of m root antenna is h
M, k(1:P), wherein, P is the length of each subscriber channel impulse response sequence, its value be 128 with the merchant of active user's number.
B, according to corresponding each user's of every antenna in the smart antenna array channel impulse response sequence, calculates every each user's of antenna correspondence weight vectors.
Corresponding each user's of every antenna of the described calculating of this step weight vectors further may further comprise the steps:
B1, according to corresponding each user's of described every antenna channel impulse response sequence, calculate the performance number of each sampled point on corresponding each user's of every antenna the channel impulse response sequence, obtain corresponding each user's of every antenna channel impulse response power sequence, and with the corresponding same users' of all antennas in the smart antenna array channel impulse response power sequence by corresponding sampled point addition, obtain corresponding each user's of this smart antenna array channel impulse response power sequence.
Be example with k user still, smart antenna array is as follows to computing formula that should subscriber channel impulse response power sequence:
Wherein, operator ‖ ‖ represents the modulo operation to plural number.
B2, from corresponding certain user's of described smart antenna array channel impulse response power sequence, find out N sampled point of performance number maximum, and according to every antenna to channel impulse response that should the user at the value on this N sampled point and smart antenna array weight vectors to each user of every antenna correspondence of performance number calculating of channel impulse response power sequence on this N sampled point that should the user.
N most powerful path of this N the corresponding received signal of sampled point difference.Here, N is the integer more than or equal to 1, and its value can be set according to the environmental characteristics of wireless channel.Wherein, the N value is big more, and the multipath signal that need consider when being illustrated in wave beam forming is many more, therefore, improves the signal to noise ratio of receiving end signal, and the effect that reduces intersymbol interference is just good more, but corresponding, and the amount of calculation of wave beam forming is also just big more.
Be example still below, this step is elaborated with k user.To finding out N sampled point of performance number maximum should user's channel impulse response power sequence, the position of establishing this N sampled point is respectively x from smart antenna array
Max1, x
Max2..., x
MaxNThe position of above-mentioned this N sampled point represents that respectively user k is at time-delay x
Maxn* T
cThe power of the channel impulse response at place is one of N point of maximum intensity, wherein, and T
cBe the chip width; The span of n is 1~N, be used for representing an above-mentioned N sampled point more arbitrarily.The position of above-mentioned this N sampled point represents that also the receive time delay of N most powerful path is respectively x in received signal
Max1* T
c, x
Max2* T
c..., x
MaxN* T
c
According to the sampling point position of above-mentioned N performance number maximum, the power of channel impulse response power sequence on this N sampled point that can obtain this smart antenna array respective user k is followed successively by: H
k(x
Max1), H
k(x
Max2) ..., H
k(x
MaxN); The value of channel impulse response on this N sampled point of m root antenna respective user k is followed successively by: h
M, k(x
Max1), h
M, k(x
Max2) ..., h
M, k(x
MaxN).
According to H
k(x
Max1), H
k(x
Max2) ..., H
k(x
MaxN) and h
M, k(x
Max1), h
M, k(x
Max2) ..., h
M, k(x
MaxN), utilize following computing formula, calculate the weight vectors of m root antenna respective user k
Wherein, p weighted factor W of described weight vectors
p M, kFor:
Wherein, operator conj[] expression asks conjugate operation; Parameter j span 1~N is used for the sampled point in N performance number maximum of summation process expression; x
MaxnBe in the sampled point of described N performance number maximum any.
Corresponding each user's of every antenna that C, use calculate weight vectors carries out time domain preequalization Filtering Processing to each user data to be sent, obtains the data sequence that each user will send on every antenna.Through the Filtering Processing of described time domain preequalization, smart antenna array is that each chip in described each user data to be sent generates a plurality of different wave beams according to different multidiameter delays.
Because corresponding each user's of described every antenna weight vectors is to determine according to N most powerful path in this user's the received signal, and the receive time delay of this N most powerful path is respectively x
Max1* T
c, x
Max2* T
c..., x
MaxN* T
cTherefore, symmetry according to channel, when signal sends, if big footpath shifts to an earlier date the corresponding time to delaying time, and the little footpath of will delaying time postpones the corresponding time, just can guarantee the signal that sent through different paths the time delay, can arrive receiving terminal substantially simultaneously, realize the stack of signal at receiving terminal.
Based on above-mentioned thought, preequalization Filtering Processing of the present invention is specially: the data sequence that each user is to be sent and the every antenna that calculates carry out convolution to weight vectors that should the user, and computing formula is as follows:
y
m,k(s)=x(s)
kW
m,k
Wherein, x
k(s) k data sequence that the user is to be sent of expression, operator represents convolution operation, y
M, k(s) k data sequence that the user sends on m root antenna after the expression process preequalization Filtering Processing.
Those of ordinary skill in the art are appreciated that, characteristic according to convolution algorithm, in this step, the signal that each user is to be sent and the every antenna that calculates carry out convolution to weight vectors that should the user, just be equivalent to the receive time delay according to N the most powerful path signal that receives, for each chip in each user data to be sent generates a plurality of different wave beams, and timesharing sends signal, promptly the bigger footpath of time delay is sent ahead of time, and the footpath delayed delivery less to time delay.Can guarantee substantially the signal that sent through different paths the time delay, arrive receiving terminal simultaneously, the main concentration of energy of transmission signals is received on the chip at one, thereby improves the power that receives chip, also reduce intersymbol interference simultaneously.
In the application of reality, above-mentioned preequalization Filtering Processing operation can have multiple implementation, for example: can develop Digital Signal Processing (DSP) chip and directly realize above-mentioned convolution of function, the data sequence that each user is to be sent is carried out the preequalization Filtering Processing; Can also adopt digital filter to realize above-mentioned convolution of function.
Fig. 4 has shown the digital filter principle that realizes filtering pre-equalization process function of the present invention.As shown in Figure 4, described weight vectors shows as the tap coefficient of digital filter in digital filter.Z among Fig. 4
-1Be delay cell, be used for chip width of signal delay input.Digital filter from Fig. 4 can clearerly find out, when k user with data sequence x to be sent
kWhen first chip (s) was input in this digital filter, this chip can be input to delay cell Z on the one hand
-1, on the other hand by weighted factor W
P M, kWeighting, output signal y at this moment
M, k(s) be x
k(1) * W
P M, k, be equivalent to send weighted factor W
P M, kCorresponding multipath signal.When described user k with x
kWhen second chip (s) was input in this digital filter, this chip can be input to delay cell Z on the one hand
-1, on the other hand also will be by weighted factor W
P M, kWeighting, first chip of this user input this moment are input to next delay cell Z with one side after through the time-delay of a chip time
-1, on the other hand will be through weighted factor W
P-1 M, kWeighting, at this time, output signal y
M, k(s) be x
k(2) * W
P M, k+ x
k(1) * W
P-1 M, k, be equivalent to send simultaneously weighted factor W
P M, kAnd W
P-1 M, kCorresponding multipath signal.By that analogy, through preequalization Filtering Processing of the present invention, weighted factor W
P M, kCorresponding multipath signal, promptly the footpath of receive time delay maximum will send at first, and weighted factor W
1 M, kCorresponding multipath signal, promptly the footpath of receive time delay minimum will send at last.
From the computational process of corresponding each the user's weight vectors of above-mentioned every antenna as can be seen, because N P usually, and in actual applications, the prominent N of a channel impulse response point should be relatively to concentrate, and promptly puts shared length L to N from first o'clock
k=x
MaxN-x
Max1N≤L will satisfy condition
kP, like this, at corresponding each the user's weight vectors of every antenna at L
kOutside value will all be 0, therefore, in order to simplify calculating, can from the weight vectors that calculates, remove L
kOutside, value is 0 amount.Like this, reduce the length of described weight vectors, thereby reduced the amount of calculation of preequalization Filtering Processing convolution process.
D, will on being loaded into every antenna of smart antenna array, the data sequence that sends on the every antenna send through each user of preequalization Filtering Processing.In this step, described smart antenna array sends each chip in each user data to be sent according to being the one or more wave beam that each chip in each user data to be sent generates among the step C.
As can be seen from the above analysis, method of the present invention replaces weighted factor of the prior art by using according to characteristic of channel generation weight vectors, user's information is carried out pre-equalization process on the time domain, and the footpath bigger to time delay sends ahead of time, and to the later transmission in the less footpath of time delay.Like this, can guarantee the signal that sent through different paths the time delay, can arrive receiving terminal substantially simultaneously, make the main concentration of energy of transmission signals receive on the chip at one, thereby can improve the signal to noise ratio of receiving end signal, and reduce intersymbol interference, improve the reliability of transmission signals.