JP4869142B2 - Adaptive antenna - Google Patents

Description

  The present invention relates to an adaptive antenna, and more particularly to an adaptive antenna applied to a multicarrier transmission scheme.

  There is multi-carrier transmission as a high-speed data transmission system. In particular, OFDM (Orthogonal Frequency Division Multiplexing) in which subcarriers are arranged at orthogonal frequency intervals has been adopted in many wireless systems such as wireless LAN (Local Area Network) and terrestrial digital broadcasting.

  In this OFDM, it is well known that waveform distortion due to a delayed wave can be reduced by inserting a guard interval (hereinafter referred to as GI). However, for delayed waves exceeding the guard interval, the characteristics are significantly degraded.

  Therefore, application of an adaptive antenna that spatially removes such a delayed wave that exceeds the guard interval using an array antenna is being studied, and a method for realizing an adaptive antenna for OFDM is also being studied (for example, Non-Patent Document 1, Patent Document 1, or Patent Document 2).

Imai, Ogawa, Ogane, “Study on Adaptive Array in OFDM Communication System,” IEICE Technical Report AP2001-115, 2001 JP-A-11-289213 Japanese Patent Laid-Open No. 10-2110099

However, the prior art has the following problems.
In the conventional technique described above, the method of Non-Patent Document 1 can capture delay waves in the GI well by combining each subcarrier separated by FFT (Fast Fourier Transform), but the number of subcarriers There was a problem that the amount of calculation increased in proportion to the increase.

  On the other hand, Patent Document 1 or Patent Document 2 uses a subcarrier after FFT, but is configured to give the same weighting coefficient (amplitude phase adjustment) to each subcarrier. This is to reduce the amount of calculation by performing control in the frequency direction by using a reference signal periodically inserted in each subcarrier.

  However, it can be said that such a calculation amount reduction method is difficult to operate in an actual environment because no countermeasure is taken against the following problems. The reason for this will be described below.

  In multicarrier transmission, a demultiplexer that separates each subcarrier is required. In the case of OFDM, FFT calculation corresponds to this. FIG. 5 is a diagram showing a configuration example of an OFDM symbol in conventional multicarrier transmission. In this configuration, it is necessary to remove the guard interval (GI) 101 in advance before performing the FFT process.

  The GI is obtained by copying a part of the latter half of the data symbol 102 on the transmission side and adding it to the head, and has a characteristic that it is resistant to multipath with delay. On the receiving side, it is ideal to perform FFT on the data symbol 102 from which the GI portion has been removed. However, in actuality, it is difficult to completely specify the position by multipath fading, and the FFT window position 103 for cutting out the data string input to the FFT is in a state including the GI 101 portion as shown in FIG. .

  For this reason, the signal after the FFT causes a phase rotation corresponding to the window position shift amount d between the subcarriers. Usually, when demodulation is performed, the amplitude and phase fluctuations in the transmission path are corrected, so this phase rotation is not a problem. However, when determining the weighting factor of the adaptive antenna, this phase rotation greatly affects the performance.

  The present invention has been made to solve the above-described problems, and an object thereof is to obtain an adaptive antenna that enables stable operation even when there is a phase rotation corresponding to the window position shift amount d. To do.

  The adaptive antenna in the present invention is an adaptive antenna applied to a multicarrier transmission system, and separates an array antenna composed of a plurality of antenna elements and each received signal received by the plurality of antenna elements into subcarrier signals. Subcarrier signal having the same frequency after subcarrier extraction means and adjusting the amplitude phase by applying a constant weight to each of the plurality of antenna elements for each subcarrier signal separated for each of the plurality of antenna elements Based on the combining means for combining them to generate a combined subcarrier signal, the subcarrier signal separated by the subcarrier extracting means, and the combined subcarrier signal generated by the combining means, the combining means generates an amplitude phase. Weight coefficient calculation means for determining weights for adjustment And the weighting factor calculating means further includes a phase estimator that estimates the phase rotation amount between adjacent subcarriers based on the combined subcarrier signal, and further determines the weight in consideration of the phase rotation amount. .

  According to the present invention, the phase rotation amount between subcarriers is estimated based on the combined subcarrier signal, the weighting factor is determined so as to correct the estimated phase rotation amount, and the amplitude phase is feedback-adjusted, Even when there is a phase rotation corresponding to the window position shift amount d, an adaptive antenna capable of stable operation can be obtained.

  Hereinafter, preferred embodiments of the adaptive antenna of the present invention will be described with reference to the drawings.

Embodiment 1 FIG.
FIG. 1 is a configuration diagram of an adaptive antenna according to Embodiment 1 of the present invention. The adaptive antenna shown in FIG. 1 includes an array antenna 10, subcarrier extraction means 20, synthesis means 30, weight coefficient calculation means 40, demodulator 50, and parallel-serial converter 60.

  Here, the subcarrier extraction means 20 includes a series-parallel converter 21 and an FFT converter 22. The synthesizing unit 30 includes a multiplier 31 and a synthesizer 32. Further, the weight coefficient calculation means 40 includes a weight coefficient calculator 41 and a phase estimator 42.

  The array antenna 10 of FIG. 1 according to the first embodiment shows a case where four antenna elements # 1 to # 4 are configured. The number of elements is not limited to four, and can be similarly applied to any number of elements.

  In OFDM, which is multicarrier transmission, high-speed transmission is enabled by assigning different data to subcarriers having different frequencies. Therefore, the receiver needs a function of a duplexer that separates each subcarrier. In FIG. 1, the function of a duplexer is realized by the serial-parallel converter 21 and the FFT converter 22 in the subcarrier extraction means 20.

  The serial-parallel converter 21 has a function of cutting out data having a certain length, and also performs FFT window position control. Next, each FFT converter 22 performs frequency conversion by performing FFT on the extracted data, and divides the data into subcarrier signals in an orthogonal frequency arrangement. The output of each FFT converter 22 is an output having a size of several tens to several thousand in an actual system.

  Next, the synthesizing unit 30 synthesizes the subcarriers after adjusting the amplitude and phase. More specifically, each subcarrier is multiplied by a weighting coefficient for amplitude phase adjustment by a multiplier 31 in the synthesizing unit 30, and a signal in each antenna element is multiplied by a synthesizer 32 in the synthesizing unit 30. Subcarrier signals having the same frequency are synthesized.

  The combined subcarrier signals after being combined are demodulated by the demodulator 50 and rearranged to the original information sequence by the parallel-serial converter 60.

  A weighting coefficient for adjusting the amplitude phase in the multiplier 31 is calculated by a weighting coefficient calculator 41 in the weighting coefficient calculating means 40. Specifically, the weighting coefficient calculator 41 performs weighting calculation using information from the phase estimator 42 that estimates the phase difference between subcarriers described below. The technical feature of the present invention is that the weighting calculation is performed in consideration of the phase difference estimated by the phase estimator 42.

  Note that the phase difference between subcarriers estimated by the phase estimator 42 is not simply the phase difference between the combined subcarrier signals itself, but corresponds to the amount of phase rotation of the desired signal contained in the signal. This is an amount corresponding to the window position shift amount d in FIG.

  Next, a series of operations of the adaptive antenna in the first embodiment will be described more specifically. An RF (Radio Frequency) band signal X (t) received by the array antenna 10 is converted into a baseband digital signal by various receiving devices such as a low noise amplifier, a filter, a frequency converter or an A / D converter. The However, these receiving devices are omitted in FIG. 1 to simplify the description.

  The converted digital signal is then extracted as data corresponding to the FFT size by the serial-parallel converter 21 and separated into subcarriers by the FFT converter 22. The signal separated into subcarriers is further adjusted in amplitude phase by multiplying a constant weighting factor for each element by a multiplier 31.

  That is, for the subcarriers of the same element of the array antenna 10, the amplitude phase is adjusted by multiplying the same weighting factor for each element, and then the subcarrier signals having the same frequency are synthesized by the synthesizer 32. Then, a combined subcarrier signal is generated. The combined subcarrier signal generated in this way is then restored to the original data string by the demodulator 50 and the parallel-serial converter 60.

  The weighting factor calculation means 40 calculates the amplitude phase value (hereinafter referred to as a weighting factor) adjusted by the multiplier 31. The phase estimator 42 in the weighting factor calculation means 40 estimates the amount of phase rotation between subcarriers based on the combined subcarrier signal generated by the combiner 32.

  Furthermore, the weighting factor calculator 41 in the weighting factor calculation means 40 is based on each subcarrier signal after FFT, the combined subcarrier signal that is the output of the combiner 32, and the phase rotation amount estimated by the phase estimator 42. Thus, the multiplier 31 calculates a weighting coefficient used for adjusting the amplitude phase.

  Here, in order to explain the effectiveness of the phase estimator 42 which is a technical feature of the present invention, the influence on the performance of the data cut-out position in the serial-parallel converter 21 will be described using mathematical expressions. The transmission signal s (t) is expressed as the following formula (1).

  Here, N is the FFT size, and c (n) is the nth subcarrier signal. If it is assumed that the transmission path fluctuation is constant in the effective symbol period and the propagation coefficient is g, the received signal x (t) is expressed by the following equation (2).

  Here, Δt is a delay amount, and n (t) represents thermal noise. As for thermal noise, the property does not change before and after the FFT processing, and therefore, in the subsequent analysis, it is ignored.

  If the delay amount Δt is within the GI length, it does not affect the subcarrier separation itself, so it is ignored here and analyzed as the window position shift amount d as shown in FIG. The subcarrier signal X (f) after the FFT is expressed by the following equation (3).

  That is, as shown in the above equation (3), it can be seen that phase rotation proportional to the shift amount d occurs between the subcarrier signals. In OFDM transmission, since there is a reference signal periodically inserted in transmission data, each difference after FFT is detected by detecting the difference between the reference signal included in the received signal and the reference signal prepared in advance on the receiving side. The amplitude phase of the subcarrier signal can be corrected. For this reason, in general, there is no problem even if there is such a phase rotation. However, there may be a problem when a plurality of antenna elements are combined like an adaptive antenna.

  The least square error method (MMSE), which is the most general operating principle of an adaptive antenna, minimizes the difference between an array output signal and a reference signal, that is, an error signal. Therefore, as in the first embodiment of the present invention, when the same weighting factor is given to the subcarrier for each element, the weighting factor can be calculated using the data sample of the subcarrier direction, that is, the frequency axis.

  At this time, if the array output signal is y (f) and the reference signal is r (f), the evaluation function Q of the following equation (4) is minimized.

  Further, the optimal solution (Winner solution) W of MMSE is expressed by the following equation (5). However, E [] represents an expected value operation. The subscript H represents a complex conjugate transpose, and * represents a complex conjugate. Further, the following equation (6) is a correlation matrix Rxx, and the following equation (7) is a correlation vector Pxr between the input signal X (f) and the reference signal r (f).

  Here, when the received signal model of the above equation (2) is considered and the delay amount Δt is the window position deviation amount d, the relationship of the following equation (8) is obtained.

  Here, g is a complex propagation coefficient, V is a signal direction vector, and n is a noise vector. Since the direction vector V is stored even after the FFT, the subcarrier signal vector X (f) after the FFT is expressed by the following equation (9).

  Therefore, the correlation vector Pxr of the above equation (7) is represented by the following equation (10).

  Originally, good reception characteristics can be obtained by extracting the direction vector V of the desired signal by calculating the correlation vector. However, when there is phase rotation due to window position shift, a solution cannot be obtained due to the influence of this phase rotation term.

  The above description has been described using a statistical parameter called an expected value. However, this problem can be solved by applying optimization algorithms such as LMS (Least Mean Square), RLS (Recursive Last Square), and SMI (Sample Matrix Inversion). This is also true when actually controlling the weighting factors sequentially.

  Thus, it is a feature of the first embodiment that a good convergence characteristic is always obtained by estimating the phase rotation amount between subcarriers and correcting it. That is, the weighting coefficient is obtained based on the evaluation function Q (W) as in the following expression (11).

  Here, the estimated α represents the amount of phase rotation between adjacent subcarriers. By controlling the weighting factor based on such a criterion, the above problem can be overcome. As for the phase correction, in the above equation (11), the phase rotation amount of α is given to the reference signal r (f), but even if the phase rotation amount of −α is applied to the term of the array output y (f). Is equivalent.

  Regarding the phase correction amount α, it is possible to obtain the phase difference between adjacent subcarriers using the array output y (f) in which the reference signal r (f) is inserted. It is also possible to improve accuracy by averaging a plurality of phase differences.

  For example, the modulation signal component included in the array output y (f) is removed by the reference signal r (f), and only the propagation coefficient and the phase rotation term are extracted as b (f) as shown in the following equation (12).

  Next, an average value c of products between adjacent subcarrier signals is calculated according to the following equation (13), and a phase difference of the average value c is obtained according to the following equation (14). Can be obtained.

  Here, M is the number of averaged samples, Re {} is a real part, and Im {} is an imaginary part. Note that the multiplications in equations (12) and (13) may be replaced with division. Further, the averaging may be performed not by the equation (13) but by the equation (14) after obtaining the phase rotation amount α.

  As described above, according to the first embodiment, after adjusting the amplitude and phase by performing constant weighting for each antenna element, the subcarrier signals having the same frequency are combined to generate a combined subcarrier signal, It is possible to estimate the phase rotation amount between subcarriers based on the combined subcarrier signal, determine the weighting coefficient so as to correct the estimated phase rotation amount, and feedback adjust the amplitude phase. As a result, the adaptive antenna can always be normally converged even with respect to the FFT window position shift that varies depending on the reception environment, and a significant improvement in performance can be expected as compared with the conventional method.

Embodiment 2. FIG.
In the first embodiment, the estimation of the phase rotation amount and the calculation of the weighting factor are basically performed independently. In the second embodiment, the case where the estimation of the phase rotation amount and the calculation of the weighting coefficient are performed simultaneously will be described in consideration of calculation efficiency.

  FIG. 2 is a configuration diagram of the adaptive antenna according to the second embodiment of the present invention, and is an example of a configuration in the case of using an LMS algorithm that sequentially updates weighting coefficients by iterative calculation. Compared with the configuration in FIG. 1 in the first embodiment, the configuration in FIG. 2 in the second embodiment is different in the components in the weight coefficient calculating means 40.

  The weight coefficient calculation means 40 in the second embodiment includes a phase estimator 42, an error signal generator 43, an update vector generator 44, a multiplier 45, a delay unit 46, and a combiner 47. The update procedure in the adaptive antenna of the second embodiment is performed simultaneously according to the following equations (15) to (18).

  Here, W (m) represents a weighting coefficient after being updated m times, and μ is a constant called a step size. The following description describes the procedure for obtaining the m-th weighting coefficient, and the subscripts of each variable are unified with m. In actual processing, averaging processing may be performed using a plurality of subcarrier signals X (f) and y (f), or may be sequentially updated for each subcarrier signal.

The above equation (15) represents the combined subcarrier signal y (m) that is the output of the combiner 32, and the phase estimator 42 obtains the phase amount α based on the above equation (16). It is to be noted that the same effect can be obtained even if the equation (16) is obtained by performing division of r (m) instead of multiplication of r ^* (m) as in the following equation (19).

Next, the error signal generator 43 generates an error signal e (m) considering phase rotation based on the above equation (17). Thereafter, the update vector generator 44 calculates the update vector component x (m) e ^* (m) of the LMS, and the multiplier 45 multiplies the step size μ. By adding this to the weight coefficient W (m) at the time of updating m times in the synthesizer 47, the m + 1th weight coefficient W (m + 1) is calculated as shown in the above equation (18).

  Based on this W (m + 1), the value of the multiplier 31 is controlled. Further, the delay unit 46 gives a certain time delay and is used as a component to be added in the next weighting coefficient calculation.

  Next, calculation examples for clarifying the effectiveness of the present invention will be shown. The antenna arrangement is a four-element equally spaced linear array, and the incoming signals are two waves, a preceding wave and a delayed wave exceeding the GI. Further, the SNR (signal to noise power ratio) per element is 20 dB.

  FIG. 3 is a graph showing a simulation result in the second embodiment of the present invention. The horizontal axis represents the window position shift d during FFT, and the vertical axis represents the output SINR (Signal to Interference plus Noise Ratio) after array synthesis. Higher output SINR means better characteristics.

  71 in the figure is the result when the weighting factor is updated by the conventional LMS algorithm, and 72 is the result when the weighting factor is updated by the new algorithm according to the present invention. As mentioned earlier, when there is a shift in the FFT window position, phase rotation between subcarriers occurs, and it is confirmed that the conventional LMS cannot fully follow this phase change and the characteristics deteriorate. it can.

  On the other hand, in the algorithm according to the present invention, as a result of overcoming this problem by correcting the phase difference estimated from the reference signal prepared on the receiving side and sequentially updating the weighting factor, very excellent characteristics are obtained. I understand that.

  As described above, according to the second embodiment, in addition to the effects of the first embodiment, the phase difference estimation is fused with the LMS algorithm, and it is possible to obtain further excellent followability by sequentially updating. .

  In the above description, LMS is taken as an example, but other optimization algorithms such as RLS algorithm and SMI algorithm may be used, and the same effect can be obtained.

Embodiment 3 FIG.
FIG. 4 is a configuration diagram of an adaptive antenna according to Embodiment 3 of the present invention. In the third embodiment, the initial value of the weighting factor is set and updated for the purpose of stable and fast convergence. Compared to the configuration of FIG. 1 in the first embodiment, the configuration of FIG. 4 in the third embodiment is different in that it further includes a power estimation unit 70.

  Not only in the first and second embodiments, but in adaptive antennas, it is known that how to give an initial value greatly affects the stability and high speed of the convergence characteristics when updating the weighting factor. Yes. The power estimation unit 70 according to the third embodiment obtains a signal-to-noise power ratio using the subcarrier signal of each antenna element after the output of the FFT converter 22. Furthermore, the power estimation unit 70 gives the multiplier 31 an initial value in which the weighting factor of the antenna element having the largest signal-to-noise power ratio is 1 and the others are 0, and starts updating the weighting factor.

  The noise power here includes the power of the interference signal, and the signal-to-noise power ratio corresponds to the above-mentioned SINR (Signal to Interference plus Noise Ratio).

  In this way, by setting the initial value of the weighting factor using the power estimation unit 70, the antenna element with the best reception state can be selected at the initial stage. As a result, a reliable and fast convergence characteristic can be obtained in the subsequent updating of the weighting factor. SINR measurement can be easily realized by using a known reference signal in the received signal.

  As described above, according to Embodiment 3, by obtaining the signal-to-noise power ratio based on the subcarrier signal, it is possible to perform the initial setting of the weighting factor of the antenna element according to the reception state. This makes it possible to obtain a reliable and high-speed convergence characteristic in the subsequent update of the weight coefficient.

  Note that the initial value setting by the power estimation unit 70, which is a feature of the third embodiment, can be applied to the second embodiment as in the first embodiment.

It is a block diagram of the adaptive antenna in Embodiment 1 of this invention. It is a block diagram of the adaptive antenna in Embodiment 2 of this invention. It is a graph which shows the simulation result in Embodiment 2 of this invention. It is a block diagram of the adaptive antenna in Embodiment 3 of this invention. It is the figure which showed the structural example of the OFDM symbol in the conventional multicarrier transmission.

Explanation of symbols

  10 array antennas, 20 subcarrier extraction means, 21 serial-parallel converter, 22 FFT converter, 30 combining means, 31 multiplier, 32 combiner, 40 weight coefficient calculating means, 41 weight coefficient computing unit, 42 phase estimator, 43 error signal generator, 44 update vector generator, 45 multiplier, 46 delay unit, 47 synthesizer, 50 demodulator, 60 parallel-serial converter, 70 power estimation unit.

Claims (7)

An adaptive antenna applied to a multicarrier transmission system,
An array antenna composed of a plurality of antenna elements;
Subcarrier extraction means for separating each received signal received by the plurality of antenna elements into subcarrier signals;
For each subcarrier signal separated for each of the plurality of antenna elements, after adjusting the amplitude phase by applying a constant weight to each of the plurality of antenna elements, the subcarrier signals having the same frequency are combined. Combining means for generating a combined subcarrier signal;
Based on the subcarrier signal separated by the subcarrier extracting means and the synthesized subcarrier signal generated by the synthesizing means, a weight coefficient calculation for determining weights for adjusting the amplitude phase by the synthesizing means Means and
The weighting factor calculating means further comprises a phase estimator that estimates a phase rotation amount between adjacent subcarriers based on the combined subcarrier signal, and determines the weighting further considering the phase rotation amount. A characteristic adaptive antenna. The adaptive antenna according to claim 1,
The weight coefficient calculating means includes
Based on the subcarrier signal separated by the subcarrier extracting means, the synthesized subcarrier signal generated by the synthesizing means, and the phase rotation amount estimated by the phase estimator, an amplitude phase is obtained by the synthesizing means. An adaptive antenna comprising a weighting factor calculator for determining a weight for adjusting the frequency. The adaptive antenna according to claim 2,
The weighting factor calculator calculates a phase relative to the reference signal based on the phase rotation amount when minimizing a difference between a reference signal periodically inserted into transmission data and a reference signal prepared in advance on the reception side. An adaptive antenna, wherein the weighting value is determined by performing correction. The adaptive antenna according to claim 3,
The adaptive antenna according to claim 1, wherein the phase estimator estimates the amount of phase rotation using the reference signal. The adaptive antenna according to claim 1,
The weight coefficient calculating means includes
An error signal generator that generates an error signal based on the combined subcarrier signal generated by the combining means and the amount of phase rotation estimated by the phase estimator;
Based on the subcarrier signal separated by the subcarrier extraction means and the error signal generated by the error signal generator, a weighting factor calculator for determining weights for adjusting the amplitude phase by the synthesis means And an adaptive antenna. The adaptive antenna according to any one of claims 1 to 5,
The adaptive antenna according to claim 1, wherein the weighting factor calculating means sequentially updates the weighting using an optimization algorithm. The adaptive antenna according to any one of claims 1 to 6,
A signal-to-noise power ratio is calculated for each of the plurality of antenna elements based on the sub-carrier signal separated by the sub-carrier extracting means, and the combining means is calculated according to the calculated magnitude of the signal-to-noise power ratio A power estimation unit that initially sets the weighting in
The weighting factor calculating means sequentially updates the weighting initially set by the power estimation unit.

Images