KR20060095597A - Apparatus and method for generating smart antenna beams in broadband wireless communication system - Google Patents

Abstract

The present invention relates to a smart antenna beam forming in a broadband wireless communication system, and calculates a channel coefficient for each subcarrier using the received preamble signal, and calculates and outputs an average noise power using the preamble signal or a pilot signal; A noise power calculator and a representative for dividing an entire frequency band into a plurality of subbands, and calculating and outputting a representative beam coefficient for each of the plurality of subbands by using a channel coefficient for each subcarrier and the average noise power. A beam coefficient calculator and a linear interpolator for generating beam coefficients for all subcarriers by linear interpolation between a plurality of representative beam coefficients from the representative beam coefficient calculator. As such, the present invention has an advantage in that the amount of computation (complexity) can be significantly reduced by performing the MMSE operation performed for each tone by sub-band.

Smart Antenna, SDMA, Beam Coefficient, MMSE, Linear Interpolation

Description

Smart antenna beam forming apparatus and method in broadband wireless communication system {APPARATUS AND METHOD FOR GENERATING SMART ANTENNA BEAMS IN BROADBAND WIRELESS COMMUNICATION SYSTEM}

1 illustrates a cellular system that supports a conventional Space Division Multiple Access (SDMA) scheme.

2 illustrates a downlink and uplink frame structure of an 802.16 OFDMA system.

FIG. 3 conceptually illustrates a beam coefficient calculation algorithm for performing MMSE operations every tone. FIG.

4 illustrates an SDMA full frequency band divided into a plurality of subbands in accordance with the present invention.

FIG. 5 is a diagram illustrating a configuration of a receiver in an OFDM communication system using a space division multiple access (SDMA) scheme according to an embodiment of the present invention. FIG.

FIG. 6 is a diagram illustrating a detailed operation of the representative beam coefficient calculator 519 described in FIG. 5.

7 shows a detailed operation of the linear interpolator 529 described in FIG.

8 illustrates a concept of a beam coefficient linear interpolation algorithm according to the present invention.

Figure 9 is a graph comparing the performance of the present invention (MMSE (per-Bin) and the prior art (MMSE (per-Tone)).

The present invention relates to a smart antenna beam forming apparatus and method in a broadband wireless communication system, and more particularly, to an apparatus and method for reducing the amount of calculation according to antenna beam weight calculation.

In general, in the 4th Generation communication system, which is the next generation communication system, active research is being conducted to provide users with services having various Quality of Service (QoS) having a transmission rate of about 100 Mbps. Currently, the 3rd Generation communication system generally supports a transmission rate of about 384 Kbps in the outdoor having a relatively poor channel environment, and a maximum transmission rate of about 2 Mbps in a room having a relatively good channel environment.

Meanwhile, wireless local area network (LAN) systems and wireless metropolitan area network (MAN) systems generally support transmission rates of 20 Mbps to 50 Mbps. The wireless MAN system is suitable for high-speed communication service support because of its wide coverage and high-speed transmission, but it does not consider the mobility of a subscriber station at all. Therefore, in the 4th generation communication system, research is being actively conducted in the form of ensuring mobility and QoS in a wireless LAN system and a wireless MAN system that guarantee a relatively high transmission speed.

The orthogonal frequency division multiple access (OFDM) scheme and the orthogonal frequency division multiple access (OFDMA) scheme are applied to the physical channel of the wireless MAN system. Electronics Engineers) 802.16 system. In other words, the IEEE 802.16 system realizes high-speed data transmission by transmitting a physical channel signal using a plurality of subcarriers.

Meanwhile, in the IEEE 802.16 system, standards for supporting an adaptive antenna system (AAS) mode and a non-AAS mode are provided in one frame. There are two main reasons for applying the AAS mode to the system. The first is an increase in cell capacity, and the second is an increase in cell coverage.

That is, the adaptive antenna system (AAS) is a system that continuously detects a cell coverage area using an array antenna and adaptively forms a beam pattern in a variable wireless channel environment. For example, if there is only one terminal and there is no interference, the adaptive antenna system adaptively responds to the movement of the terminal by generating an effective antenna pattern that follows the movement of the terminal. At this time, the antenna pattern becomes a pattern which always shows the best gain in the terminal direction. Spatial Division Multiple Access (SDMA) systems can be implemented using such an adaptive antenna system. That is, if there are N terminals, N beams may be generated in each terminal direction.

FIG. 1 illustrates a cellular system supporting a conventional SDMA scheme.

As illustrated, the base station 101 forms a spatial channel 1 transmitted to the beam 102 and a spatial channel 2 transmitted to the beam 103. At this time, the first spatial channel and the second spatial channel uses the same frequency and time resources. As such, uplink channel information is required to form a plurality of beams using the same resource in downlink. Accordingly, the existing IEEE 802.16 OFDMA system adds a preamble symbol to downlink and uplink for measuring a channel state of a corresponding beam.

2 illustrates a downlink and uplink frame structure of an 802.16 OFDMA system.

As shown, it can be seen that the AAS preamble is transmitted in front of each downlink burst and the uplink burst. The base station estimates a channel through an uplink AAS preamble and forms a beam using the estimated channel information for downlink beamforming. In this case, since the spatial channels formed by the beams are spatially separated from each other as shown in FIG. 1, even if they share the same frequency and time resources, the interference is not large. Therefore, every spatial channel can utilize all the frequency and time resources allowed by the system.

On the other hand, beam weights for beam shaping should be calculated for each subcarrier, for each transmit antenna and for each user. Hereinafter, a beamforming algorithm according to the prior art will be described.

First, the signal for the k-th tone received by the m-th sensor (or antenna) is represented by Equation 1 below.

Where U denotes the number of users assigned to the same frequency, M denotes the number of individual antennas constituting the array antenna, and H _{u, m} (f _k ) denotes the k-th of the u-th user received by the m-th antenna. Represents the channel coefficient (channel coefficient or channel signature) for the tone, s _u (f _k ) represents the data signal carried on the k-th tone of the u-th user, and z _m (f _k ) is the k-th at the m-th antenna Represents the thermal noise for a tone.

In the above Equation 1, when the number of users U is 1, the reception signal Y _m (f _k ) may be a reception signal in the non-SDMA mode, and when the number of users using the same band in the SDMA operation is 1 Can be a received signal.

Meanwhile, the vector of the received signal received through the M antennas may be represented by Equation 2 below.

Where Y _m (f _k ) represents a signal for the k th tone received at the m th antenna, and H _{m, u} (f _k ) represents the channel coefficient for the k th tone of the u th user received at the m th antenna (channel signature), s _u (fk) represents the data signal on the k-th tone of the u-th user.

As is well known, the optimal solution of OFDMA-SDMA detection is the Maximum Likelihood (ML) algorithm. However, the computational complexity is so large that implementation is virtually impossible. Therefore, we use a minimum mean squared error (MMSE) algorithm known as a sub-optimum algorithm. The MMSE algorithm is as follows.

First, when array combining is performed on a u-th user, a received SDMA signal is expressed by Equation 3 below.

Here, W _{u, m} (f _k ) represents a beam weight applied to the k th tone of the u th user received by the m th antenna.

On the other hand, when performing array combining (or antenna combining) for all users, the received SDMA vector is expressed by Equation 4 below.

Here, W _{u, m} (f _k ) represents the beam coefficient applied to the k-th tone of the u-th user received by the m-th antenna.

을 사용한다.The u-th column of the OFDMA-SDMA filter matrix W (f _k ) defined in Equation 4 corresponds to the SDMA filter for the u-th user. Meanwhile, the channel coefficient matrix H (f _k ) is a channel value estimated from an AAS preamble.

Use

In this case, the MMSE beaming matrix applied to the array combining is calculated as shown in Equation 5 below.

는 허미시안(Hermitian) 매트릭스 연산을 나타내며,

는 평균 잡음전력(noise variation)을 나타내고, I는 M×M 항등 행렬(identity matrix)을 나타내며,

은 역(inverse) 매트릭스 연산을 나타낸다.Here, H (f _k ) is the M ㅧ U channel signature matrix for the k th tone, as shown in Equation 2 above.

Represents a Hermitian matrix operation,

Is the average noise power, I is the M × M identity matrix,

Denotes an inverse matrix operation.

의 연산량이 필요하다. 즉, 종래기술은 어레이 사이즈가 증가할수록 빔계수 산출 연산량은 급수적으로 증가하는 문제점이 있다. 따라서, 상기 수학식 5에 따른 빔계수 산출 연산량을 줄이기 위한 방안이 요구되고 있다.As described above, according to the prior art, an MMSE operation is performed for every tone (subcarrier). 3 conceptually illustrates a beam coefficient calculation algorithm for performing an MMSE operation for every tone. Equation 5 is for calculating beam coefficients of the f _k th tones for all antennas, and in order to calculate beam coefficients for all 864 x M (number of M antennas) tones, it is necessary to perform 864 MMSE operations. . In this case, 864 direct matrix inverse (DMI) operations are required. As is known, to DMI the M × M matrix

The amount of computation is required. That is, the prior art has a problem in that the calculation coefficient of the beam coefficient increases in number as the array size increases. Accordingly, a method for reducing the beam coefficient calculation amount according to Equation 5 is required.

Accordingly, an object of the present invention is to provide an apparatus and method for reducing the amount of calculation according to antenna beam coefficient calculation in a broadband wireless communication system using a smart antenna.

Another object of the present invention is to provide an apparatus and method for reducing the amount of calculation according to antenna beam coefficient calculation in a broadband wireless communication system supporting a plurality of spatial channels.

It is still another object of the present invention to provide an apparatus and method for calculating a representative beam coefficient for each subband in a broadband wireless communication system supporting a plurality of spatial channels and calculating the beam coefficients for the remaining subcarriers through linear interpolation. have.

According to a first aspect of the present invention for achieving the above object, in the broadband wireless communication system using at least one spatial channel, the smart antenna beam forming apparatus, divides the entire frequency band into a plurality of subbands, A representative beam coefficient calculator for calculating and outputting a representative beam coefficient for each of the subbands of?, And linearly interpolating between a plurality of representative beam coefficients from the representative beam coefficient calculator to beam for all subcarriers And a linear interpolator for generating the coefficients.

According to a second aspect of the present invention, in a broadband wireless communication system supporting at least one spatial channel, the smart antenna beamforming apparatus calculates a channel coefficient for each subcarrier by using a received preamble signal. A channel and noise power calculator for calculating and outputting an average noise power using a preamble signal or a pilot signal, and dividing an entire frequency band into a plurality of subbands, and using the channel coefficient and the average noise power for each subcarrier. A representative beam coefficient calculator for calculating and outputting a representative beam coefficient for each of the plurality of subbands, and linear interpolation between the plurality of representative beam coefficients from the representative beam coefficient calculator to all subcarriers It characterized in that it comprises a linear interpolator for generating the beam coefficients for.

According to a third aspect of the present invention, in a wideband wireless communication system using at least one spatial channel, the smart antenna beam forming method divides an entire frequency band into a plurality of subbands and assigns each of the plurality of subbands. And a step of generating beam coefficients for all subcarriers by linear interpolation between the calculated plurality of representative beam coefficients.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings. In describing the present invention, if it is determined that the detailed description of the related known function or configuration may unnecessarily obscure the subject matter of the present invention, the detailed description thereof will be omitted.

Hereinafter, a method for calculating a representative beam coefficient for each subband in a broadband wireless communication system supporting a plurality of spatial channels and calculating the beam coefficients for the remaining subcarriers through linear interpolation will be described. In this case, since one direct matrix inverse (DMI) operation is required per subband, the computation amount due to beam coefficient calculation can be significantly reduced.

Next, the algorithm for reducing the amount of calculation according to the beam coefficient calculation will be described in detail.

4 illustrates an SDMA total frequency band divided into a plurality of subbands according to the present invention.

As shown, the entire SDMA band according to the present invention is divided into G subbands, each subband consisting of K tones. In this case, the number K of tones constituting each group (or subband) is determined in consideration of the channel coherence bandwidth. For example, if the frequency band constituting the data burst channel consists of consecutive AMC bins or Partial Usage of SubCarrier (PUSC) tiles, the subband may be an AMC bin (or PUSC tile). Can be determined in units. In the following description, a case where the subband is defined in units of AMC bins will be described as an example.

First, the cost function of the MMSE solution for each bin (or tile) is represented by Equation 6 below.

는 기대 값(expectation value)을 나타내고, k₀는 해당 빈의 시작을 지정하기 위한 부반송파 인덱스를 나타내며,

는 빈(bin)내 k번째 톤에 대한 인덱스를 나타내고, K는 빈(bin)을 구성하는 톤의 개수를 나타낸다. 한편,

와

는 수학식 2에서 설명한 바와 같으며,

는 상기 수학식 4에서 설명한 바와 같다. here,

Represents an expectation value, k ₀ represents a subcarrier index to specify the start of the bin,

Denotes the index of the k-th tone in the bin, and K denotes the number of tones constituting the bin. Meanwhile,

Wow

Is as described in Equation 2,

Is as described in Equation 4 above.

The solution to the cost function defined in Equation 6 is shown in Equation 7 below. That is, a solution for calculating the representative beam coefficient matrix for each bin is shown in Equation 7 below.

는 빈(타일) 내 k번째 톤에 대한 인덱스를 나타내며, K는 빈을 구성하는 톤들의 개수를 나타내고, H_m,n(f_k)는 m번째 안테나로 수신된 u번째 사용자의 k번째 톤에 대한 채널계수(channel signature)를 나타낸다.Where k ₀ represents a subcarrier offset for specifying the start of the bin,

Is the index of the kth tone in the bin (tile), K is the number of tones that make up the bin, and H _{m, n} (f _k ) is the kth tone of the uth user received by the mth antenna. It represents the channel signature for.

Meanwhile, when the representative beam coefficient for each bin is calculated through Equation 7, beam coefficients for the remaining subcarriers are calculated through linear interpolation. As described above, a system configuration and operation of calculating a representative beam coefficient for each subband and calculating the remaining beam coefficients through linear interpolation are as follows.

FIG. 5 illustrates a configuration of a receiver in an OFDM communication system using a space division multiple access (SDMA) scheme according to an embodiment of the present invention.

As shown, the receiver according to the present invention includes an array antenna consisting of a plurality of antennas 501-1 through 503-M, a plurality of FFT operators 503-1 through 503-M, and a subcarrier demapping machine. 505, a channel and noise power estimator 507, an array combiner 509, a demodulator 511, and a decoder 513. Here, the array combiner 509 includes a representative beam coefficient calculator 519, a linear interpolator 529, and a beamformer 539.

Referring to FIG. 5, although not shown, a signal received through the first antenna 501-1 is converted into baseband sample data and input to the first FFT operator 503-1. Similarly, the signal received through the M-th antenna 501 -M is converted into baseband sample data and input to the M-FFT operator 503 -M. The first FFT operator 503-1 to the M th FFT operator 503 -M respectively output the sample data of the input time domain by performing Fast Fourier Transform (FFT). That is, the FFT operators 503-1 to 503-m demodulate the received signal to the value of each subcarrier and output the demodulated signal.

The subcarrier demapper 505 outputs the subcarrier values carrying data among the subcarrier values from the FFT operators 503-1 to 503m to the array combiner 509, and the subcarrier values carrying the preamble and pilot. Output to the channel and noise power estimator 507. The figure exemplarily shows that 864 subcarrier values carrying actual data among the 1024 subcarriers corresponding to each FFT operator are provided to the array combiner 509. In this case, the received signal Y (f _k ) input to the array combiner 509 is represented by the equation (2). In Equation 2, Y ₁ (f _k ) represents a signal received through the first antenna 501-1, and Y _M (f _k ) represents a signal received through the Mth antenna 501 -M. .

)을 산출하여 상기 어레이 결합기(509) 및 상기 복조기(511)로 제공한다.The channel and noise power estimator 507 calculates a channel coefficient H (f _k ) by using a preamble signal (preamble subcarrier values) from the subcarrier demapper 505 and an array combiner 509 and a demodulator 511. To provide. Here, the channel coefficient H (f _k ) is as shown in Equation (7). In addition, the channel and noise power estimator 507 uses the preamble signal or the pilot signal to generate the noise power (

) Is provided to the array combiner 509 and the demodulator 511.

를 추출한다. The array combiner 509 calculates a representative beam coefficient (W _g ) for each subband (bin or tile), and linearly interpolates between the calculated representative beam coefficients to obtain beam coefficients for all subcarriers. Calculate The user signal is obtained by using the calculated beam coefficients and the received signal Y (f _k ) from the subcarrier demapper 505.

Extract

)을 이용해 서브대역별 대표 빔계수(W_g)를 산출하여 출력한다. 여기서, 상기 서브대역별 대표 빔계수 산출 알고리즘은 상술한 수학식 7과 같다. 상기 대표 빔계수 산출기(519)의 동작은 후술되는 도 6의 참조와 함께 상세히 살펴보기로 한다. In more detail, the representative beam coefficient calculator 529 may calculate the channel coefficient H (f _k ) and the noise power for each subcarrier.

Representative beam coefficient (W _g ) for each subband is calculated and output. Here, the representative beam coefficient calculation algorithm for each subband is as shown in Equation (7). The operation of the representative beam coefficient calculator 519 will be described in detail with reference to FIG. 6 to be described later.

The linear interpolator 529 linearly interpolates the representative beam coefficients from the representative beam coefficient calculator 519 to calculate and output a beam coefficient W (f _k ) for each subcarrier. In this case, constant interpolation is performed on a region where linear interpolation is impossible. The operation of the linear interpolator 529 will be described in detail with reference to FIGS. 7 and 8 to be described later.

를 획득한다. 상기 사용자 신호

를 획득하기 위한 알고리즘은 상기 수학식 4와 같고, 이렇게 획득된 사용 자 신호

는 복조기(511)로 제공된다.The beamformer 539 is array-coupled by using the received signal Y (f _k ) input from the subcarrier demapper 505 and the beam coefficient W (f _k ) for each subcarrier from the linear interpolator 529. SDMA filtered user signal

Acquire. The user signal

Algorithm for acquiring is as shown in Equation 4, the user signal thus obtained

Is provided to the demodulator 511.

를 상기 채널 및 잡음전력추정기(507)로부터의 채널계수 및 잡음전력을 이용해 복조(demodulation)하여 출력한다. 복호기(513)는 상기 복조기(511)로부터의 복조심볼(또는 LLR(Log likelihood Ratio)값)들을 복호(decoding)하여 정보 비트열(bit stream)을 출력한다.The demodulator 511 is a user signal from the array combiner 509

Is demodulated using the channel coefficients and the noise power from the channel and the noise power estimator 507 and outputted. The decoder 513 decodes demodulation symbols (or log likelihood ratio values) from the demodulator 511 and outputs an information bit stream.

If the number of SDMA users is two, the number of user signals output from the array combiner 509 is two, and then the demodulator 511 and the decoder 513 perform demodulation and decoding on each user signal.

6 illustrates a detailed operation of the representative beam coefficient calculator 519 described in FIG.

)을 입력한다. 상기 채널계수 매트릭스 H(f_k)와 상기 평균 잡음전력(

)는 상기 채널 및 잡음전력 추정기(507)로부터 입력된다.Referring to FIG. 6, the representative beam coefficient calculator 519 first calculates the channel coefficient matrix H (f _k ) and the average noise power of each tone (subcarrier) in step 601.

Enter). The channel coefficient matrix H (f _k ) and the average noise power (

) Is input from the channel and noise power estimator 507.

In operation 603, the representative beam coefficient calculator 519 calculates a Hermitian matrix of the channel coefficient matrix of each tone. In operation 605, the representative beam coefficient calculator 519 multiplies the corresponding channel coefficient matrix and the hermithian matrix and adds the corresponding sub coefficients.

) 및 항등행렬(I)을 곱해서 산출된다. 한편, 상기 대표 빔계수 산출기(519)는 609단계에서 상기 입력된 채널계수 매트릭스들을 서브대역별로 합산한다. In operation 607, the representative beam coefficient calculator 519 adds the summation matrix for each subband and the noise component matrix, and calculates an inverse matrix of the added matrix. Here, the noise component matrix is represented by Equation 7, the number of tones (K) constituting the bin (K), the noise power (

) And the identity matrix (I). In operation 609, the representative beam coefficient calculator 519 sums the input channel coefficient matrices for each subband.

In operation 611, the representative beam coefficient calculator 519 calculates the representative beam coefficient matrix W _g (W _group-MMSE ) for each subband by multiplying the corresponding inverse matrix and the channel coefficient sum matrix for each subband. The channel coefficient matrix of each tone is calculated by linear interpolation using the representative beam coefficient matrices thus calculated.

FIG. 7 shows the detailed operation of the linear interpolator 529 described in FIG.

Referring to FIG. 7, in operation 701, the linear interpolator 529 initializes the subband index g to '1'. In operation 703, the linear interpolator 529 checks whether the value of the subband index g is smaller than the predetermined number of subbands G. If the value of the subband index g is less than the number of subbands G, the linear interpolator 529 proceeds to step 705, otherwise ending the present algorithm.

Then, the linear interpolator 529, in step 705, the difference between the representative beam coefficient matrix W _{g + 1 of the} ( _{g + 1} ) th subband and the representative beam coefficient matrix W _g of the g th subband. (??) After calculating the difference, the linear interpolator 529 initializes the variable k value to '1' in step 707.

The linear interpolator 529 proceeds to step 709 to check whether the variable k value is smaller than the predetermined number of subcarriers K (determined by the number of tones in the subband). If the value of the variable k is less than the value K, the linear interpolator 529 proceeds to step 711, otherwise proceeds to step 715 to increase the subband index g by '1' and then returns to step 703. Goes.

번째 부반송파에 대한 빔계수 매트릭스

을 하기 수학식 8과 같이 산출한다. Then, the linear interpolator 529 in step 711

Beam Coefficient Matrix for First Subcarrier

It is calculated as shown in Equation 8.

Here, k ₀ represents a subcarrier offset index and is determined as a subcarrier index corresponding to the center of the corresponding subband.

번째 부반송파에 대한 빔계수 매트릭스를 산출한후, 상기 선형 보간기(529)는 713단계로 진행하여 상기 변수 k 값을 '1'만큼 증가한후 상기 709단계로 되돌아간다.As above,

After calculating the beam coefficient matrix for the first subcarrier, the linear interpolator 529 proceeds to step 713 to increase the variable k by '1' and then returns to step 709.

8 shows a concept of a beam coefficient linear interpolation algorithm according to the present invention.

Referring to FIG. 8, the beam coefficients for the center subcarriers of each bin represent representative beam coefficients for each subband calculated through Equation (7). As shown, a linear interpolation scheme for beam coefficients for each of the subcarriers located between the center of the first bin and the center of the second bin using the representative beam coefficient of the first bin and the representative beam coefficient of the second bin. Calculate as Similarly, a beam coefficient for each of the subcarriers located between the center of the second bin and the center of the third bin is calculated using the linear beam coefficient of the second bin and the representative beam coefficient of the third bin by linear interpolation. . Thus, after calculating the beam coefficients for the subcarriers located between the center of the fourth bin and the center of the fifth bin, constant interpolation is performed on the region where the remaining linear interpolation is impossible, so that the entire SDMA band ( Obtain beam coefficients for bandwidth). In this case, since the direct matrix inverse (DMI) operation required to calculate the beam coefficient is performed only once per subband (or bin), the amount of computation can be significantly reduced.

Meanwhile, the beam coefficient linear interpolation algorithm described above may be summarized as Equation 9 below.

Here, G represents the number of subbands, K represents the number of tones constituting the subband, and k ₀ represents a subcarrier offset for determining the starting point of linear interpolation. As described above, k ₀ is determined as the center subcarrier index of the corresponding g th subband (empty).

As we have seen above. In calculating the MMSE-SDMA filtering matrix (beam coefficient matrix), the prior art (Equation 5) requires one indirect matrix (DMI) operation for each tone (subcarrier), but the present invention (Equation 7) ) Requires one inverse matrix operation per subband (bin or tile), which can significantly reduce the computational complexity. In particular, as the size of the array antenna increases, the calculation amount according to the present invention decreases exponentially.

Here, when comparing the calculation amount according to the present invention and the calculation amount of the prior art in detail as shown in Table 1.

U: the number of SDMA users using the same frequency

M: Smart Antenna Array Size

N _tone : Number of tones to which traffic data is allocated in subcarriers

G: number of subbands

K: number of tones constituting subband

P _r : Number of real multiplication operations

A _r : number of real addition operations

D _r : Number of real division operations

Table 1 shows the amount of computation required for 1 second for each of the prior art (MMSE) and the present invention (Group-MMSE) when N _tone = 864, M = 4, U = 2, K = 9, and G = 96. Is a comparison. DSP (Digital Processor) complexity is calculated based on the ADSP-TS201 chip.

As shown in Table 1, it can be seen that when using the method according to the present invention, the complexity of the DSP cycle aspect is reduced to approximately 1/6. In particular, the scheme according to the invention has the advantage of not accompanied by performance degradation.

9 is a graph comparing the performance of the present invention (MMSE (per-Bin) and the prior art (MMSE (per-Tone)). As a result of the simulation of the Wibro uplink smart antenna, the base station in the Urban Macro channel environment In case of receiving signals with these four antennas, the experimental results when QPSK and R = 1/2 are shown.

As shown, the performance of both per-tone algorithm (prior art) and per-bin algorithm (invention) performance of MMSE and successive interference cancellation (MMSE-SIC) when link performance is compared in terms of frame error rate (FER) This almost coincides. In other words, the present invention can be referred to as an algorithm that can significantly reduce the complexity without accompanying performance degradation.

Meanwhile, in the detailed description of the present invention, specific embodiments have been described, but various modifications are possible without departing from the scope of the present invention. Therefore, the scope of the present invention should not be limited to the described embodiments, but should be determined not only by the scope of the following claims, but also by those equivalent to the scope of the claims.

As described above, the present invention can significantly reduce the amount of computation (complexity) by performing the MMSE operation performed for each tone in sub-bands in the OFDM (A) system supporting the smart antenna function. . In particular, the present invention has the advantage of reducing the complexity but does not accompany performance degradation.

Claims (24)

An apparatus for forming a smart antenna beam in a broadband wireless communication system using at least one spatial channel, A representative beam coefficient calculator for dividing an entire frequency band into a plurality of sub bands, calculating and outputting a representative beam coefficient for each of the plurality of sub bands; And a linear interpolator for linearly interpolating between a plurality of representative beam coefficients from the representative beam coefficient calculator to generate beam coefficients for all subcarriers. The method of claim 1, The subband is determined in units of bins or tiles. The method of claim 1, The representative beam coefficient calculator is characterized in that for calculating a representative beam coefficient for each subband using the following equation (10).

Where g denotes an index of the subband, M denotes the number of sensors configured in the array antenna, U denotes the number of users using the same frequency, and k ₀ denotes a subcarrier for designating the start of the corresponding subband. Offset,

Denotes the index of the k-th subcarrier in the subband, K denotes the number of subcarriers constituting the subband, H (f _k ) denotes the channel coefficient of the k-th subcarrier, and H _{m, n} (f _k ) Denotes the channel signature of the k th subcarrier of the u th user received by the m th antenna,

Is the average noise power and I is the M × M identity matrix. The method of claim 3, Wherein the channel coefficient H (f _k ) is calculated using a preamble signal. The method of claim 3, The average noise power

Is calculated using a preamble signal or a pilot signal. The method of claim 1, The linear interpolator calculates a beam coefficient for each subcarrier using Equation 11 below.

Here, G represents the number of subbands, K represents the number of subcarriers constituting the subband, and k ₀ represents a subcarrier offset for determining the start of linear interpolation in the g-th subband. The method of claim 1, And each of the representative beam coefficients calculated by the representative beam coefficient calculator is determined as a beam coefficient for a center subcarrier in a corresponding subband. The method of claim 1, And a beamformer for extracting a spatial division multiple access (SDMA) user signal by array combining the beam coefficients and the received subcarrier values for all subcarriers from the linear interpolator. An apparatus for forming a smart antenna beam in a broadband wireless communication system supporting at least one spatial channel, A channel coefficient for each subcarrier using a received preamble signal, a channel and noise power calculator for calculating and outputting an average noise power using the preamble signal or a pilot signal; A representative beam coefficient calculator for dividing an entire frequency band into a plurality of subbands, calculating and outputting a representative beam coefficient for each of the plurality of subbands using a channel coefficient for each subcarrier and the average noise power; And a linear interpolator for linearly interpolating between a plurality of representative beam coefficients from the representative beam coefficient calculator to generate beam coefficients for all subcarriers. The method of claim 9, The subband is determined in units of bins or tiles. The method of claim 9, The representative beam coefficient calculator is characterized in that for calculating the representative beam coefficient for each subband using the following equation (12).

Where g denotes an index of the subband, M denotes the number of sensors configured in the array antenna, U denotes the number of users using the same frequency, and k ₀ denotes a subcarrier for designating the start of the corresponding subband. Offset,

Denotes the index of the k-th subcarrier in the subband, K denotes the number of subcarriers constituting the subband, H (f _k ) denotes the channel coefficient of the k-th subcarrier, and H _{m, n} (f _k ) Denotes the channel signature of the k th subcarrier of the u th user received by the m th antenna,

Is the average noise power and I is the M × M identity matrix. The method of claim 9, The linear interpolator calculates a beam coefficient for each subcarrier using Equation 13 below.

Here, G represents the number of subbands, K represents the number of subcarriers constituting the subband, and k ₀ represents a subcarrier offset for determining the start of linear interpolation in the g-th subband. The method of claim 9, And each of the representative beam coefficients calculated by the representative beam coefficient calculator is determined as a beam coefficient for a center subcarrier in a corresponding subband. The method of claim 9, And a beamformer for extracting a spatial division multiple access (SDMA) user signal by array combining the beam coefficients and the received subcarrier values for all subcarriers from the linear interpolator. A method for forming a smart antenna beam in a broadband wireless communication system using at least one spatial channel, Dividing an entire frequency band into a plurality of subbands and calculating a representative beam coefficient for each of the plurality of subbands; And linearly interpolating the calculated representative beam coefficients to generate beam coefficients for all subcarriers. The method of claim 15, The subbands are determined in units of bins or tiles. The method of claim 15, The representative beam coefficient is characterized in that it is calculated as in Equation 14.

Where g denotes an index of the subband, M denotes the number of sensors configured in the array antenna, U denotes the number of users using the same frequency, and k ₀ denotes a subcarrier for designating the start of the corresponding subband. Offset,

Is the index for the k-th subcarrier in the subband, K is the number of subcarriers constituting the subband, H (f _k ) is the channel signature for the kth subcarrier, and H _{m, n} (f _k ) represents the channel coefficient for the k th subcarrier of the u th user received by the m th antenna,

Is the average noise power and I is the M × M identity matrix. The method of claim 17, The channel coefficient H (f _k ) is calculated using a preamble signal. The method of claim 17, The average noise power

Is calculated using a preamble signal or a pilot signal. The method of claim 15, The beam coefficient for each of the total subcarriers is calculated as shown in Equation 15 below.

Here, G represents the number of subbands, K represents the number of subcarriers constituting the subband, k ₀ represents a subcarrier offset for determining the start of linear interpolation in the g-th subband, and W _g represents the gth Representative beam coefficient of sub band. The method of claim 15, Wherein each of the calculated representative beam coefficients is determined as a beam coefficient for a center subcarrier in a corresponding subband. The method of claim 15, Extracting a spatial division multiple access (SDMA) user signal by array combining using the beam coefficients and the received subcarrier values for the generated subcarriers. The method of claim 15, wherein the calculating of the representative beam coefficients for the subbands, Calculating a Hermitian matrix of M × U (M is the number of antennas and U is the number of users) channel coefficient matrix for each of all subcarriers, Multiplying corresponding channel coefficient matrices and hermithian matrices and summing them by subband; Calculating an inverse matrix after adding a noise component to each of the summation matrices for each subband; Summing channel coefficient matrices for the entire subcarriers for each subband; And calculating a representative beam coefficient matrix for each subband by multiplying a corresponding channel coefficient summing matrix and an inverse matrix for each subband. The method of claim 15, wherein the generating of the beam coefficients for the entire subcarriers comprises: Reading g (1 ≦ g ≦ G−1, where G is the number of subbands) representing a subband index, and calculating a difference Δ between a representative beam coefficient of the g + 1th subband and a representative beam coefficient of the gth subband; Calculating a beam coefficient by using Equation 16 for each of the subcarriers existing between the subcarrier to which the representative beam coefficient of the g-th subband is allocated and the subcarrier to which the representative beam coefficient of the g + 1th subband is allocated. Method comprising a.

Here, k ₀ represents a subcarrier index corresponding to the center of the g-th subband, k is 1 ≦ k ≦ K-1 (K is the number of subcarriers present in the subband), and Wg is a representative of the g-th subband. Indicates beam coefficients.

Images