CN101938762B - Downlink characteristic wave beam transmission method based on spatial spectrum estimation - Google Patents

Abstract

The invention discloses a downlink characteristic wave beam transmission method based on spatial spectrum estimation. The method comprises the following steps: (1) a user sends an access request to a base station; (2) the base station carries out combined spatial spectrum estimation by using C a p o n and M U S I C methods to find out L directions of channel energy concentrated in a space and corresponding energy magnitude in each direction; (3) the base station establishes an airspace covariance matrix according to the results of the combined spatial spectrum estimation; (4) the base station carries out eigenvalue decomposition on the airspace covariance matrix, selects P eigenvectors corresponding to P maximum eigenvalues as weighting vectors of an array, and form P independent wave beams in the space; (5) the base station selects the space time coding mode of an output channel number P, and emits the space time coding mode subjected to weighed by the characteristic wave beams to the user through a wireless channel; and (6) the user restores original data after carrying out the space time decoding on received signals; and thus the downlink transmission is completed. The downlink characteristic wave beam transmission method has good real-time, low complexity, better stability and easy realization.

Description

Descending eigen beam transmission method based on Estimation of Spatial Spectrum

Technical field

The invention belongs to the wireless communication signals process field, relate to a kind of descending eigen beam transmission method, improved linkage communication quality between base station in wireless communications and user based on Estimation of Spatial Spectrum.

Background technology

In radio communication, be called downlink communication by base station to user's communications.In descending multi-antenna transmitting is defeated, adopt transmit diversity or space multiplexing technique can effectively improve signal transfer quality or transmission rate.Applied environment is at independent rayleigh fading channel, and each element that promptly requires the transmission channel matrix of equivalence is independent identically distributed multiple Gaussian random variable.If with transmit diversity or space multiplexing technique directly applies in the correlation channel, can cause the remarkable decline of systematic function.In order to address this problem, eigen beam Methods for Coding when having proposed based on sky comprises based on the channel condition information method of feedback with based on the spatial beams method of estimation of not having feedback.

For example; The water-filling that utilizes that people such as Shengli Zhou propose in " Optimal Transmitter Eigen-Beamforming and Space-Time Block Coding Based on Channel Correlations.IEEE Trans.Inform.Theory; vol.42, No.7.July 2003 " realizes that transmit diversity and wave beam form the technology of combining and be applied to the down channel method for communicating.This method is that channel condition information is changed into the spatial domain covariance matrix; Obtain characteristic vector and characteristic value through characteristic value decomposition; Utilize the weighing vector of the maximum several characteristic value characteristic of correspondence vector of water-filling screening, thereby realize the optimum precoding processing of signal as array.The weak point of the method is that because the use of water-filling still depends on the accuracy of partial channel-state information, when there was error in channel condition information, when making in this way, communication performance will be had a strong impact on.And the base station receives tediously long channel state information matrix can produce bigger time delay, and the real-time that directly causes system is poor.Because each element of channel state information matrix all can receive quantization error influence in the transmission course, is very easily disturbed, make and contain transmission error in the channel condition information that the communication robustness is poor.In addition, when the user is carrying out channel estimating when containing error, the power efficiency that this method produced also can be under some influence.When channel condition information was inaccurate, error rate of system was than higher, and generally speaking this method can not the inaccurate communication environment of well adapted channel condition information.

And for example; The utilization that people such as Kevin H.Lin propose in " Adaptive Transmit Eigenbeamforming With Orthogonal Space-Time Block Coding in Correlated Space-Time Channels; 2004.Proceedings. (ICASSP'04) .IEEE International Conference on Volume:4 " does not have the down channel communication means of feedback technique; This method is to estimate to obtain several main direction of arrival through BTS channel; Combine the array steering vector to set up a spatial domain covariance matrix then; After characteristic value decomposition, still obtain the weighing vector of array, to realize the optimum precoding processing of signal through water-filling.This method effectively avoided channel condition information inaccurate with channel estimation errors to the pretreated influence of transmitting terminal; And proposed direction of arrival information is used in the structure of spatial domain covariance matrix; Make the signal emitting performance obtain remarkable lifting, thereby promoted communication quality.The weak point of the method is that the prerequisite of the acquisition of its spatial domain covariance matrix is known every paths direction of arrival, and the direction in all paths is difficult to obtain in practical application, causes accurately obtaining the spatial domain covariance matrix.Because this method causes the inaccurate of spatial domain covariance matrix at direction of arrival, thereby causes the decline of the communication quality of whole down channel.For confirming that all path directions cause the complexity of this method too high, hindered this method in actual application in engineering, be difficult for realizing.

Summary of the invention

The object of the invention has proposed a kind of descending eigen beam transmission method based on Estimation of Spatial Spectrum to the deficiency of above-mentioned prior art.This method need not the channel condition information feedback, utilizes the Estimation of Spatial Spectrum result to make up the spatial domain covariance matrix, can farthest approach the performance based on the eigen beam transmission method of channel condition information feedback.

Performing step of the present invention is following:

Step 1, the user sends the request of networking to the base station.

Step 2, base station utilize Capon and MUSIC method to carry out the estimation of joint space spectrum, find out L the direction of channel energy in spatial concentration, and corresponding energy size on each direction.Wherein, setting the estimable arrival ripple of MUSIC direction number is M-1, and M is the array number of base station transmit antennas.

Step 3, estimated result is composed based on joint space in the base station, makes up spatial domain covariance matrix R _SCM

Step 4, the base station is to spatial domain covariance matrix R _SCMCarry out characteristic value decomposition, choose P the maximum pairing P of a characteristic value characteristic vector,, form P independently wave beam in the space as the weighing vector of array.

Step 5, base station are according to value and the quality of services for users demand of P, and selecting the output channel number is the Space Time Coding pattern of P.The P road of Space Time Coding module output signal through after the eigen beam weighting by base station, arrive the user through wireless channel.

Step 6, the user after the decoding, recovers initial data to the signal that receives when empty, and downlink transmission process is accomplished.

The present invention compared with prior art has following advantage:

The first, the present invention is real-time.The present invention adopts the joint space spectrum estimation approach of Capon and MUSIC in the process that makes up the spatial domain covariance matrix, avoided the base station to receive tediously long channel state information matrix, has promoted the real-time of system.

The second, robustness of the present invention is good.Constructed spatial domain covariance square can farthest approach real spatial domain covariance matrix; In addition, the eigen beam among the present invention forms the channel estimation results that does not rely on feedback, therefore do not receive the influence of channel estimation errors and feedback delay, so the present invention has robustness preferably.

The 3rd, complexity of the present invention is low.MUSIC of the present invention spectrum estimates that the estimated ripple direction number that reaches is M-1, and generally, actual space path number is counted M much larger than array antenna, therefore, and compares based on the algorithm for estimating of estimating every paths direction of arrival, and complexity of the present invention is low, is easy to realize.Result of study shows that the present invention can lower complexity approach the performance based on perfect channel state information feedback method.

Description of drawings

Fig. 1 is a flow chart of the present invention.

Fig. 2 is the Estimation of Spatial Spectrum figure of the present invention under 2 bunches of multipath situation.

Fig. 3 is the Estimation of Spatial Spectrum figure of the present invention under 5 bunches of multipath situation.

Fig. 4 be the present invention with the performance comparison diagram of class methods when the perfect channel condition information.

Fig. 5 be the present invention with class methods at the asynchronous performance comparison diagram of number of path.

Fig. 6 be the present invention with class methods in the space performance comparison diagram during 5 bunches of multipaths.

Embodiment

Ginseng Fig. 1, the concrete mode that the present invention implements comprises the steps:

Step 1, certain user sends the request of networking to the base station.

Step 2, the spatial spectrum that the base station obtains through Capon and MUSIC are united the angle information of judging in the space channel.Estimation of Spatial Spectrum is carried out with accepting state in the base station, supposes base station employing uniform line-array, and array number is M.

2a) setting the estimable arrival ripple of MUSIC direction number is M-1, and M is the array number of base station transmit antennas.Estimate M-1 distinguishable arrival ripple direction through the MUSIC algorithm.

2b) combine Capon space power spectrum result, filter out L distinguishable arrival ripple direction and corresponding power information, wherein, L≤M-1.In Capon space power spectrum drawing for estimate, confirm the spectrum peak-to-peak value E that power is the highest.Confirm the resulting M-1 of a MUSIC peak corresponding performance number in the Capon space power spectrum, as exist the corresponding performance number of certain angle to think then that less than 0.5E it does not constitute the actual spectrum peak and casts out.

2c) statistics is effectively composed the corresponding center angle θ in peak _iWith performance number σ _i, 1≤i≤L wherein.

Step 3, base station are utilized joint space spectrum estimated result, make up the spatial domain covariance matrix model of down channel:

3a) steering vector with respect to the base station array is obtained according to the center angle in L path of joint space spectrum estimated value acquisition in the base station:

Wherein d is an array element distance, and λ is a signal wavelength, θ _iFor effectively composing the corresponding center angle in peak.

3b) base station makes up the spatial domain covariance matrix according to the power information of L center angle of joint space spectrum estimation $R_{\mathrm{SCM}}=\underset{i=1}{\overset{L}{\mathrm{\Σ}}}{\mathrm{\σ}}_i^{2}a\left({\mathrm{\θ}}_i\right)a^H\left({\mathrm{\θ}}_i\right).$

Step 4, the base station is to R _SCMCarry out feature decomposition, select the weighing vector of P eigenvectors as array.

4a) spatial domain, base station covariance matrix carries out feature decomposition

R _SCMBe the Hermite matrix, it carried out feature decomposition can obtain individual features value and characteristic vector.

To R _SCMCarry out characteristic value decomposition: R _SCM=VDV ^H,

D=diag (λ wherein ₁..., λ _M), diagonal element λ _m(1≤m≤M) is R _SCMCharacteristic value, V=[v ₁..., v _M] ∈ C ^{M * N}Be the corresponding eigenvectors { v of these characteristic values _m∈ C ^{M * 1}, the unitary matrice that 1≤m≤M} formed.

4b) select P group vector

Among the D P biggest characteristic value just corresponding P bunch, with this P characteristic value { λ _p, the pairing P of a 1≤p≤P} characteristic vector { w _p, 1≤p≤P} is as the array weight vector of emission array: W=[ξ ₁..., ξ _P], be used for the emission of base station signal.

Step 5, base station are according to value and the quality of services for users demand of P, and selecting the output channel number is the Space Time Coding pattern of P, produce signal matrix X.The P road of Space Time Coding module output signal through after the eigen beam weighting by base station, arrive the user through wireless channel.

Step 6 is sent when requesting users need be carried out sky to received signal and is decoded, and decoding processing can obtain Space Time Coding matrix signal X during through sky, and descending radio communication finishes.Acknowledge(ment) signal Y=XW in this process ^HH+N, wherein X is the Space Time Coding matrix signal that obtains in the step 5, and W is the pre-coding matrix that obtains in the step 4, and H is a real channel, and N is a white Gaussian noise.

Effect of the present invention can further be explained through following experiment simulation.

Channel model and simulation parameter are set, and channel is quasistatic when supposing transmitting terminal and receiving terminal sky; Because it is higher that set up the base station, the scattering object around it has constituted P bunch in the space.Suppose that P bunch includes the Q paths, in each bunch, some paths are spacedly distributed in certain angle spread scope.In addition, suppose that adjacent two bunches do not have overlappingly in the space, and keep certain interval.Suppose that the user is positioned at by abundant scattering object and surround; And antenna distance is at least greater than 2 times wavelength; The received signal fadeout of different antennae is separate, and under above assumed condition, the channel of user side can be described with the rayleigh fading channel model of standard.Simulation parameter: adopt the BPSK modulation, the uniform line-array of 8 array elements is adopted in the base station, and the user adopts single antenna to receive.Select two groups of channel parameters: there are 2 bunches and 5 bunches of two kinds of situation in the space, and the angle spread of each bunch is 12 degree.Fig. 2, Fig. 4, Fig. 5 are 2 bunches situation, and under 2 bunches situation, the center angle is got (60 °, 30 °) or (0 °, 30 °); Fig. 3, Fig. 6 are 5 bunches situation, and under 5 bunches situation, the center angle is got (60 ° ,-30 °, 0 °, 30 °, 60 °).

The spatial spectrum that Estimation of Spatial Spectrum draws among the present invention is Fig. 2, and is shown in Figure 3, wherein comprised the Estimation of Spatial Spectrum figure of Capon and 2 kinds of methods of MUSIC.Through Fig. 2, Fig. 3 can find out the stronger M-1 of energy direction center angle, then through obtaining L effective angle after the screening of Capon half-power.Wherein the MUSIC spatial spectrum is for definite path direction, and the Capon method can be screened the bigger direction of energy and confirmed its corresponding power.Can find out that by Fig. 4 along with the increase of information source number, performance of the present invention promotes gradually; When using M-1 information source number in the present invention, its performance is with basic identical with the error rate of class methods when the perfect channel condition information.Can find out that by Fig. 5 the present invention still has stability preferably when number of path increases.Fig. 6 be the present invention with class methods in the space performance comparison diagram during 5 bunches of multipaths, can find out that when multipath quantity increased in the space, the present invention still had preferable performance.

Can find out that from above emulation performance of the present invention can reduce to approach conventional method under the situation of system complexity.In descending wireless communication system, systematic function and channel condition information are closely related, and the present invention can pass through L stronger direction of the accurate measured energy of Estimation of Spatial Spectrum on the one hand, thereby has guaranteed the accuracy of channel azimuth information; On the other hand, the influence that no feedback requirements has also avoided channel estimation errors that channel energy information is brought can guarantee that systematic function has stability preferably.

Claims (3)

1. descending eigen beam transmission method based on Estimation of Spatial Spectrum may further comprise the steps:

Step 1, the user sends the request of networking to the base station;

Step 2, base station utilize Capon and MUSIC method to carry out the estimation of joint space spectrum, find out L the direction of channel energy in spatial concentration; And corresponding energy size on each direction; Wherein, setting the estimable arrival ripple of MUSIC direction number is M-1, and M is the array number of base station transmit antennas;

Step 3, estimated result is composed based on joint space in the base station, makes up spatial domain covariance matrix R _SCM

Step 4, the base station is to spatial domain covariance matrix R _SCMCarry out characteristic value decomposition, choose P the maximum pairing P of a characteristic value characteristic vector,, form P independently wave beam in the space as the weighing vector of array;

Step 5, base station be according to value and the quality of services for users demand of P, and selecting the output channel number is the Space Time Coding pattern of P, the P road output signal of Space Time Coding module through after the eigen beam weighting by base station, arrive the user through wireless channel;

Step 6, the user after the decoding, recovers initial data to the signal that receives when empty, and downlink transmission process is accomplished.

2. the descending eigen beam transmission method based on Estimation of Spatial Spectrum according to claim 1; Wherein the described base station of step 2 utilizes Capon and MUSIC method to carry out the estimation of joint space spectrum; Find out L the direction in spatial concentration of channel energy; And corresponding energy size on each direction, carry out as follows:

2a) setting the estimable arrival ripple of MUSIC direction number is M-1, and M is the array number of base station transmit antennas, estimates M-1 distinguishable arrival ripple direction through the MUSIC algorithm;

2b) combine Capon space power spectrum result; Filter out L distinguishable arrival ripple direction and corresponding power information, wherein, L≤M-1; In Capon space power spectrum drawing for estimate; Confirm the spectrum peak-to-peak value E that power is the highest, confirm the resulting M-1 of a MUSIC peak corresponding performance number in the Capon space power spectrum, as exist the corresponding performance number of certain angle to think then that less than 0.5E it does not constitute the actual spectrum peak and casts out;

2c) statistics is effectively composed the corresponding center angle θ in peak _iWith performance number σ _i, 1≤i≤L wherein.

3. the descending eigen beam transmission method based on Estimation of Spatial Spectrum according to claim 1, wherein the described base station of step 3 utilizes joint space spectrum estimated result, makes up the spatial domain covariance matrix model R of down channel _SCM, carry out as follows:

3a) steering vector with respect to the base station array is obtained according to the center angle in L path of joint space spectrum estimated value acquisition in the base station:

$a\left({\mathrm{\θ}}_i\right)=e^{j2\mathrm{\π}(n-1)d\sin \left({\mathrm{\θ}}_i\right)/\mathrm{\λ}},1\≤i\≤L,1\≤n\≤M$

Wherein d is an array element distance, and λ is a signal wavelength, θ _iFor effectively composing the corresponding center angle in peak;

3b) base station makes up the spatial domain covariance matrix according to the power information of L center angle of joint space spectrum estimation $R_{\mathrm{SCM}}=\underset{i=1}{\overset{L}{\mathrm{\Σ}}}{\mathrm{\σ}}_i^{2}a\left({\mathrm{\θ}}_i\right)a^H\left({\mathrm{\θ}}_i\right).$

Images