WO2000065690A1 - Adaptive array antenna - Google Patents

Abstract

An adaptive array antenna operates using the method of least squares with no need for a known reference signal on the receiver side. An array antenna (1) includes a plurality of antenna elements (11 to 14), and down-converters (D/C) (21 to 24) are connected with a plurality of antenna elements (11 to 14) to convert the signals received through the array antenna (1) to IF signals. A/D converters (31 to 34) convert the IF signals to digital signals, and a signal processor (4) weights the antenna elements (11 to 14) based on the signals received through one of the antenna elements (11 to 14). A reference signal is generated from a received signal, and thus an adaptive array antenna using the method of least squares is realized without preparing a known reference signal on the receiver side.

Description

Specification Adaputibuare first antenna present invention, _c prior art relating to § Da Petit-flop array antenna using the least square error method

 A conventional least-squares-error adaptive array antenna compares the received signal of the array antenna with a known reference signal, and performs beam control to minimize the square error.

 As such a conventional technique, for example, Japanese Unexamined Patent Application Publication No. 9-219615 discloses weighting amplitude and phase of transmission and reception signals of a plurality of arranged antenna elements, and assigns weights to the antenna elements. The transmission signal is distributed and the signal received from the antenna element is synthesized. According to the invention described in this publication, in a wireless communication system that performs communication between a base station having an adaptive array transmitting / receiving apparatus and a plurality of terminals, the frequency of handoffs is reduced, and deterioration of communication quality due to inter-station interference is prevented. It is intended to Problems to be solved by the invention

 However, in the above prior art, when performing beam control so as to minimize the square error, it is necessary to prepare a known reference signal on the receiving side in order to compare the received signal of the array antenna with a known reference signal. There is.

 An object of the present invention is to provide an adaptive array antenna based on a least squares error method using a received signal as a reference signal without preparing a known reference signal on a receiving side. Disclosure of the invention

In order to achieve such an object, the invention according to claim 1 provides an array antenna composed of a plurality of element antennas, and an antenna connected independently to the plurality of element antennas. And a plurality of down converters (D / C) for converting signals received by the array antenna into IF signals, and a plurality of A / Cs for converting each IF signal from the plurality of down converters to digital signals. A D converter, and a signal processing device that performs a weighting process for each element antenna based on a received signal of any of the plurality of element antennas.

 The invention according to claim 2 is the invention according to claim 1, wherein the signal processing device is configured to detect the unnecessary wave when a signal arriving from different directions and combining a desired wave and an unnecessary wave is input. It is characterized by suppressing and extracting only the desired wave. According to a third aspect of the present invention, in the second aspect of the present invention, the desired wave in the digital signal input to the signal processing device has a return portion of the same signal sequence.

 The invention according to claim 4 is the invention according to claim 3, wherein any one of the element antennas among the digital signals extracted at the timing of the first signal sequence of the repeated portion of the same signal sequence is received. The obtained signal sequence is used as a reference signal.

 The invention according to claim 5, wherein, in the invention according to claim 4, a correlation vector comprising a correlation between the reference signal and a second signal sequence of a repeated portion of the same signal sequence; Using the covariance matrix calculated from the signal sequence of the second time, an approximate binar solution is obtained, each element antenna is weighted and combined.

 The invention according to claim 6 is the invention according to any one of claims 1 to 5, wherein the down converter operates by a common or individual local oscillator. BRIEF DESCRIPTION OF THE FIGURES

 FIG. 1 is a diagram showing a signal sequence applied to an embodiment of the adaptive array antenna of the present invention.

FIG. 2 is a block diagram showing an embodiment of an adaptive array antenna according to the present invention. FIG. 2A shows a configuration in which a local oscillator is independently connected to each element antenna circuit. B is a diagram in which a local oscillator is commonly connected to each element antenna circuit.

 FIG. 3 is a time-series vector showing a configuration example of a digital signal from each element antenna.

 FIG. 4 is a diagram showing an antenna pattern as a result of simulation of an adaptive array antenna.

 FIG. 5 shows a flowchart when the algorithm according to the first embodiment is applied.

 FIG. 6 shows a flowchart when the LMS algorithm in the second embodiment is applied.

 FIG. 7 shows a flowchart when the RLS algorithm in the second embodiment is applied. BEST MODE FOR CARRYING OUT THE INVENTION

 Next, an embodiment of the adaptive array antenna according to the present invention will be described in detail with reference to the accompanying drawings.

 1 to 4, there is shown an embodiment of an adaptive array antenna according to the present invention.

As shown in FIG. 1, prior to communication, the transmitting side continuously transmits the same signal trains Dl and D2 composed of n data such as a desired wave 100. The receiving side receives a signal in which the desired wave 100 and an unnecessary wave composed of the reflected wave 200 and the interference wave 300 are combined. The number n is not limited to a specific number. On the receiving side, as shown in FIG. 2, the combined signal of the desired wave 100, the reflected wave 200, and the interference wave 300 is combined with the element antennas 11, 1, 1, and 14 constituting the array antenna 1. , ..., etc. The received signal is converted to an IF signal by a down-converter (DZC) 21, 1, '' 24, ..., etc. connected to each element antenna 11, 1, ..., 14, ..., etc., and converted. The IF signal is then converted into digital signals 61, '... 64, ..., etc. by A / D converters 31, ..., 34, ..., etc., and input to the signal processing device 4. Is done. FIG. 3 shows an example of the configuration of digital signals 61,... 64,..., Etc. received by the element antennas 11, 1,. . In these digital signals 61, ... 64, ..., etc., the received signal sequence X1 received at the timing of the first signal sequence D1 in the repetition part of the desired signal 100

One of X1, X21, X31, and X41 is stored in a memory and used as a reference signal. Then, the reference signal held in the memory and the received signal sequence X 12, X 22, X 32, X 32 received at the timing of the second signal sequence D 2 of the repeated portion of the desired signal 100 4 Correlation vector consisting of correlation with 2 and received signal sequence XI 2, X 2

2. Using the covariance matrix calculated from X32 and X42, an adaptive array antenna that approximates the Wiener solution, weights each element antenna, and combines them is realized. Here, the signal series XII, X21, X31, X41, X12, X22, X32, and X42 are represented by complex numbers consisting of the real part I component and the imaginary part Q component. Is done. First embodiment

 The adaptive array antenna according to the present invention includes the return signals Dl and D2 transmitted prior to the communication shown in FIG. 1 and an array having the configuration shown in FIG. 2 for receiving and processing these signals. This is realized by antenna 1. In the present embodiment, as shown in FIG. 2, the description will be made assuming that the set of the element antenna, the down converter, and the A / D converter is 4. However, the number of these sets is not particularly limited as long as it is 2 or more as in the case of n described above.

The combined signal of repetitive signals D 1 and D 2, reflected waves M 1 and M 2, and interference waves U 1 and U 2 shown in Fig. 1 is received by each element antenna 11 to 14 of array antenna 1. , As shown in Fig. 2 (1), or a local oscillator shared by each set as shown in Fig. 2 (2). The digital signals are converted into digital signals 61 to 64 by the down converters 21 to 24 and the AZD converters 31 to 34 that operate on 5, and sent to the signal processor (DSP) 4. The local oscillator may be used in common with any of the sets. The signal processing device 4 separates the received signal sequence into an I component of a real part and a Q component of an imaginary part, and converts the signal into a complex number having these I and Q components. The correlation vector obtained by holding the column X 11 in memory and calculating the correlation with the received signal sequence XI 2, X22, X32, X42 using this as a reference signal, Using the received signal sequence X 12, X 22, X 32, and X 42 covariance matrices, approximate the Wiener solution to determine the weight of each element antenna, and use the least squares error method. To realize an adaptive array antenna. Note that the memory can be provided in the signal processing device or in another portion (for example, outside), and is not particularly limited. motion

 Hereinafter, the operation of the first embodiment will be described. The signal processing performed by the signal processing device 4 will be described using digital signals from each element antenna in FIG. The vector X 11 of the digital time-series signal from each element antenna in FIG. 3 and the received signal series X 12 to X 42 are compared with the desired wave 100, reflected wave 200, interference Expressed using wave 300, it is as follows.

X D 1 + U 1 + M 1 + N (1)

X I 2 = D 2 + U 2 + M 2 + N 1 2 (2)

X 2 2 = D 2 e x p (j k d s i n Θ d)

 4- U 2 e x p (j k d s i n Θ u)

 + M2 e x p (j k d s i n Θ m)

 + N 2 2 (3)

X 3 2 = D 2 e x p (j 2 k d s i n Θ d)

 + U 2 e x p (j 2 k d s i n Θ u)

+ M 2 exp (j 2 kdsin Θ m) + N 3 2 (4)

X 4 2 = D 2 e x p (j 3 k d s i n Θ d)

 4- U 2 e x p (j 3 k d s i n Θ u)

 + M 2 e x p (j 3 k d s i n Θ m)

 + N 4 2 (5) where, symbol d indicates the array spacing of element antennas 11 to 14, and symbols Θd, Θu, and 0m indicate when the front direction of the array antenna is “0” degrees. The angles of arrival of the desired wave 100, the reflected wave 200, and the interference wave 300 are shown. The symbol k indicates the propagation constant (2π / wavelength). In Equations (3) to (5), Χ 2 2 and the like can be expressed as Xm 2. In this case, the exponential function part is represented by j (m−1) k and the following equation ( It can be expressed in the same way as the exponential function part of 13). Specifically, for example, taking the first term on the right-hand side of X 22 in equation (3) “D 2 e X p (jkdsin 0 d)” as an example, the first term on the right-hand side of Xm 2 (m = 2) Is D 2 e X p (j (m−1) kdsin Θ d). Here, since m = 2, it is the one described in the first term on the right side of the above (3). As described above, in the present embodiment, the number of pairs of the element antenna, the down converter, and the AZD converter is limited to four, but the number of pairs is extended to an unlimited integer m. Is represented according to the above example.

 The desired waves D l and D 2, the reflected waves M l and M 2, the tides U l and U 2, and the thermal noise N ij are represented by the following column vectors using the signals at each time in Fig. 1. is there.

N i j = [n ij 1 n ii2… n iin] ··· (6)

D 1 = D 2 = [d 1 d 2

U 1 = [u 1 u 2 · ■ u n] (8)

U 2 = [u n + 1 u n + 2 d 1… d n-2] (9)

M 1 = [v 1 v 2 ... m n] (10)

M 2 = [dn-1 dnd 1--dn 2] ( Calculating the correlation vector r X d between the signal sequence X 11 and the received signal sequences X 12 to X 42 gives the following equation. rxd = (X 1 2-X ll ^H X 2 2X 1 1 ^H

X 3 2 · X 1 1 ^Η Χ 4 2 · Χ 1 1 ^Η ] ^τ / η

 … (1 2) Here, the sign Τ indicates transpose, and the sign Η indicates a Hermitian (conjugate transpose) matrix.

 When each element of the correlation vector is clearly displayed, the result is as follows. In the following equation (13), the symbol 符号 indicates conjugate transpose.

2 mX1 1 H-

D 2 · D 1 He x p (one j (m— 1) k d s i n Θ d)

+ D 2 · U 1 e x p (one j (m— 1) k d s i n Θ u)

+ D 2 · M 1 ^H exp (— j (m— 1) kdsin Θ m)

 + D 2 · N 1 1 He x p (— j (m— 1) k d s i n Θ m)

+ U 2 · U 1 ^H exp (-j (m— 1) kdsin Θ u)

+ U 2D 1 ^H exp (-j (m-- 1) kdsin Θ d)

+ U 2 · 1 ^H exp (-j (m— 1) kdsin Θ m)

 + U 2N 1 1 He x p (-j (m-- 1) k d s i n Θ m)

+ M 2M 1 e x p (-j (m-- 1) k d s i n Θ m)

+ M 2D 1 ^H ex (-j (m-- 1) kdsin Θ d)

 + M 2U 1 e x p (-j (m-- 1) k d s i n Θ d)

 + M 2N 1 1 He x (-j (m-- 1) k d s i n Θ d)

+ N 2 m1H Ex P (.m) k ds i n Θ m)

+ N 2 mD 1 H) k d s i n Θ d)

+ N 2 m · u 1 H. M) k d s i n Θ d)

+ N 2 mN 1 1 p (one j (m— 1) kdsin Θ d) 3) In the above formula, m represents an integer of 2 to 4 in the first embodiment, and j represents an imaginary unit.

 In the above equation, since D 2 is a repetitive signal, the correlation is high, and the others have no correlation and are considered to be “0”. Therefore, the above equation (13) eventually becomes the following equation (14) .

X 2 mX I 1 H

なお前記式において、 r x dは、 前記したとおりであり、 また共分散行列 R_xxは、 下記に示す式により、 計算できる。 D 2 D 1 ^H exp (−jmkdsin fl d) (14) This means that D 1 is a reference signal and X 2 n is a reception signal, as shown in the above equation (14). It matches the elements of the correlation vector when the least squares error method is applied. Therefore, the solution of the least squares error method is obtained from the correlation vector rxd of the signal sequence X11 and the received signal sequence X12 to X42 and the covariance matrix Rxx of the received signal sequence X12 to X42. By calculating and using it as the weight “W” of the element antenna, an adaptive array antenna operation by the least squares error method can be realized. The weight “W” of the element antenna is calculated as the solution of Wiener as follows.

In the above equation, rxd is as described above, and the covariance matrix R _xx can be calculated by the following equation.

R _xx = [X 1 2 X 2 2 X 3 2 X 4 2]

 [X 1 2 X 2 2 X 3 2 X 4 2] n

6) Note that, in the above equation, the present embodiment has been described by limiting the case where X of the covariance matrix is set to 4. However, in the present invention, the number of X is the same as that of m and is not particularly limited. In addition, in the above equation (16), the method of finding the inverse matrix of R _x is not particularly limited.

 FIG. 4 is an antenna pattern as a result of a simulation of an adaptive array antenna embodying the present invention. The solid line indicates the antenna pattern of the present invention. The dashed line is an antenna pattern based on the least squares error method using a reference signal prepared on the receiving side, and achieves the same characteristics. In Fig. 4, the angle is 1 2

From 0 degree (deg) to about 50 degrees, the simulation results obtained by the present invention and the conventional method are overlapped. Here, the desired wave is 2

From 0 °, it is assumed that unwanted waves are coming from 130 °, and a 4-element linear array antenna with 0.5 wavelength (λ / 2: λ is wavelength) interval is used. As shown in FIG. 4, in the present invention, when a desired signal arriving from different directions and a digital signal in which unnecessary waves composed of a reflected wave and an interference wave are combined, when the front direction of the array antenna is set to 0 degree, It is characterized in that when it is input to the antenna, unwanted waves consisting of reflected waves and interference waves are suppressed and only the desired waves are extracted.

 In addition, when the weight calculation is completed by the repetitive signal, the correlation between the signal sequence XI 1 and the received signal sequence X 12 is observed, and the value shows a peak (maximum value or maximum value). Can be detected.

 For example, as shown in FIG. 5, first, each signal sequence X11, X12 to Χ42 is stored (step S11). The stored signal sequences X 11 and X 12 to Χ 42 are read out, separated into a real part I and an imaginary part Q, and stored (step S 12).

Next, rXd and Fx- ¹ are calculated through the above equations (12) and (16), respectively (steps S13 and S14, respectively). The step S13 and the step S14 are calculated according to the above-mentioned formulas, and either of them may be calculated first, or they may be calculated simultaneously. these Is calculated and stored, and finally W is obtained according to the above formula (15). Next, the received W values (wlw 2 w 3 w 4) are multiplied by the received signals 61 to 64 of each element antenna, and these multiplied values are added and combined to obtain a value. As described above, according to the method of the present invention, the same signal sequence is repeatedly and continuously transmitted as a desired signal, and therefore, even in an environment where interference waves or unnecessary waves exist, continuous transmission is performed at each element antenna of the array antenna.繰 り 返 し One of the trains is stored in memory and used as a reference signal. An adaptive array antenna can be realized.

 Also, since a reference signal prepared on the receiver side is not used, an adaptive array antenna using the least squares error method that operates regardless of synchronization between transmission and reception can be realized. Second embodiment

 In the first embodiment, the weights of the element antennas are calculated by directly calculating the inverse matrix of the covariance matrix. For example, the LMS algorithm based on the steepest descent method, the RLS algorithm (recursive By using the least squares method, the weighting W can also be calculated. Others are the same as the first embodiment.

 For example, the case where the LMS algorithm is adopted will be described first. Similar to the first embodiment, the present embodiment will be described by limiting to an example in which X 11 is used as a reference signal and XI 2 to X 42 (m = 4). However, in the present invention, m is not limited to 4 as in the first embodiment.

W (i + 1) = W (i) + μ [4-'Z Xi4. E i *]… (17)

Where i is the number of updates of the recurrence equation, μ is the step size, 0 <μ <(e max) ¹

Here, A max is the maximum eigenvalue of the correlation matrix R _X x. Note that ma X is obtained as the maximum value of the eigen equation of the correlation matrix R _xx , (R ^ —λ) Z = 0 (where 0 is a null vector and Ζ is an eigen Represents a vector).

Further, in the above equation (17), e is an error signal, and is expressed by the difference between the reference signal X 11 and the actual array signal as in the following equation. em = X11—Y = X11— ^WH WXm2 (18) In the equation, XII represents a reference signal, and Y represents an array output signal.

 Further, the value of W (0) where i = 0 in equation (17) can be variously set by a known method. For example, one element of the W (0) vector can be set to 1 as the W (0) value, and the other element can be set to 0 (nu 1 1: null).

 Figure 6 shows a flowchart of such an LMS algorithm. Next, an example in which the RL algorithm is adopted will be described.

In the RLS algorithm, R _x ¹ is estimated.

 When the RLS algorithm is adopted, the following equation can be used.

W (m + 1) = W (m) + y R _xx ^l (m) X (m + 1) e * (m + 1)

… (19) In the above equation, Rxx ¹ (m) is obtained as the recurrence equation as follows.

R _xx ¹ (0) = δ ¹ I

R _xx ¹ (m) = Fei ^{- 1 Rxx 1 (m - 1} ) one

[R ^ ¹ (m-1) X (m) X ^H (m) R .. ¹ (m-1)] / [

a ² + a X ^H (m) R _xx ^l (m-1) X (m)] (20)

(In the above equation, δ is a positive constant, and I represents a unit matrix.) In this manner, in the RLS algorithm, R xx ¹ (m) is updated, and W is obtained using the updated value.

 Figure 7 shows a flowchart calculated using such an RLS algorithm. In the first embodiment, the weight is calculated in the element space used for the signal of each element antenna. However, the signal of each element antenna is converted into a beam space (frequency space) by performing a spatial FFT (Fast Fourier Transform), and the signal of each beam is used instead of the signal from the element antenna. Weighting can be calculated and combined by the method described above. According to the present invention described above, since a reference signal is generated from a received signal, an adaptive array antenna using the least square error method can be realized without preparing a known reference signal on the receiving side. In the second embodiment, by transmitting a repetition signal prior to data from the transmission side, the reception side extracts one of the repetition signals and generates a reference signal, and the reception side prepares a reference signal in advance. Thus, an adaptive array antenna using the least squares error method can be realized without performing the method.

 Each of the above embodiments is an example of a preferred embodiment of the present invention. However, the present invention is not limited to this. For example, as shown in FIG. 2, the number of pairs of the element antenna, the down converter, and the A / D converter, which is a circuit input to the signal processing device, is 4, It is also possible to use an independent local oscillator, or to connect all the local oscillators in common to the down converter, as appropriate.Otherwise, within the scope not departing from the gist of the present invention, Various modifications can be made. Industrial applicability

As is clear from the above description, the adaptive array antenna of the present invention has Thus, according to the adaptive array antenna of the present invention, one of the repetitive signals received by a plurality of element antennas is retained, and the reference signal is used as the reference signal to weight the antenna by the least square error method. Perform processing.

 Therefore, since the reference signal is generated from the received signal, it is possible to realize an adaptive array antenna by the least square error method without preparing a known reference signal on the receiving side.

Claims

The scope of the claims 1. An array antenna composed of a plurality of element antennas,  A plurality of downconverters (D / C), each independently connected to the plurality of element antennas, for converting a signal received by the array antenna into an IF signal, and each IF signal from the plurality of downconverters ^: Multiple AZD converters that convert to digital signals,  A signal processing device for performing a weighting process on each of the element antennas based on a reception signal of any of the plurality of element antennas; An adaptive array antenna comprising: 2. The signal processing apparatus is characterized in that when a signal arriving from different directions and in which a desired wave and an unnecessary wave are combined is input, the unnecessary wave is suppressed and only the desired wave is extracted. The adaptive array antenna according to claim 1. 3. The adaptive array antenna according to claim 2, wherein a desired wave in the digital signal input to the signal processing device has a repetitive portion of the same signal sequence. 4. A signal sequence received at any one of the element antennas among the digital signals extracted at a timing of a first signal sequence of a repeated portion of the same signal sequence is set as a reference signal. The adaptive array antenna according to item 3. 5. A correlation vector composed of a correlation between the reference signal, a second signal sequence of the repeated portion of the same signal sequence, and a covariance matrix calculated from the second signal sequence of the repeated portion. 5. The adaptive array antenna according to claim 4, wherein the antenna solution is approximately determined, the element antennas are weighted and combined. 6. The adaptive array antenna according to any one of claims 1 to 5, wherein the down converter is operated by a local oscillator.