JPH0350442B2 - - Google Patents

Description

Detailed Description of the Invention The present invention provides a device using an array antenna with a so-called learning function that automatically determines the weight of each antenna based on a certain evaluation standard using the input signal of each antenna element. The present invention relates to an adaptive antenna device.

When noise or interference waves other than the desired signal are present in the received electric field, the signal-to-noise (SN) ratio can be improved by creating directivity with a zero point in the direction of the interference wave by directional synthesis. The adaptive antenna device realizes this directivity in real time using the system's learning function.

FIG. 1 shows a conventional adaptive antenna device, in which a plurality of antenna elements 11, 12...1
Antenna element group 1 forming an array antenna by l...1k receives external radio waves. The signal received by this antenna element group 1 is branched, and one is divided into digital phase shifters 21, 22, . . . 2l, . ...2k is supplied.
In addition, each of the above antenna elements 11, 12...1l...
...1k can also be constructed by a subarray composed of a plurality of antenna elements.

Each digital phase shifter 21, 22...2l...
2k controls the phase of each signal from each antenna element, and the phase-controlled signal is supplied to the combiner 3 and output. Now, the output signal of the combiner 3 is branched, one is taken out as an output, and the other is supplied to the signal processor 4. This signal processor 4 is a digital phase shifter 21, 22...2l
...In order to control 2k, a constraint is imposed such that the size of the weight amount is always 1. That is, the weight amount Wl of the phase shifter for an arbitrary l-th antenna element is expressed as Wl=e ^{j l} (1) (l: phase amount in the l-th antenna element). Therefore, when the complex quantity signal of the signal truly desired in the l-th antenna element is expressed as Sl, and the complex quantity signal of unnecessary signals such as jamming and interference is expressed as Zl, each composite output component of all k antenna elements S(1~k), Z(1~k)
are expressed by the following equations (2) and (3), respectively.

S (1 ~ k) = _k 〓 ^l=1 WlSl = [W] ^T [S] = [S] ^T [W] ...(2) Z (1 ~ k) = _k 〓 ^l=1 Wl (Zl + nl) = [W] ^T [N]...(3) However, in the above equation (3), nl represents the internal noise of the l-th channel.

In equations (2) and (3), [W] ^T and [S] ^T represent the transposition of vectors (meaning one-dimensional matrices) [W] and [S], respectively, and [W] ^T = [ W ₁ , W ₂ ...W _l ...W _k ] ...(4) [S] ^T = [S ₁ , S ₂ ...S _l ...S _k ] ...(5). Similarly, the transposition of [N] [N] ^T is expressed as [N] ^T = [Z ₁ +n ₁ , Z ₂ +n ₂ ...Z _l +n _l ...Z _k +n _k ] ...(6).

Therefore, now the SN at the output of the signal synthesizer 3 is
Consider the ratio γ. As is known, the SN ratio γ can be expressed as the ratio between the peak value of the truly desired signal and the collective averaged value of the unwanted wave signal on the time axis.

When the peak value of the signal that is truly sought is expressed by power P _S , it becomes P _S | [W] ^T [S] | ² ...(7). Similarly, if the collective averaged power of the unnecessary wave signal on the time axis is expressed as P _Z , the electric field composite output of the unnecessary wave signal is given by equation (3), so P _Z = 1/T ₀ ∫ ^T0 ₀ Z ^* Zdt=1/T ₀ ∫ ^T0 ₀ {[
W] ^T [N]｝ ^* {[W] ^T [N]}dt = [W ^* ] ^T {1/T ₀ ∫ ^T0 ₀ [N ^* ] [N] ^T d
t}[W]…(8) Note that the * mark represents a complex conjugate, and [N] ^* =[N ^* ]. In this equation (8), 1/T ₀ ∫ ^T0 ₀ [N ^* ] [N] ^T dt can be replaced with the collective average on the time axis, so equation (8) becomes P _Z = [W ^* ] ^T It can be expressed as E{[N ^* [N] ^T }[W]...(9). In this equation (9), E{ } is { }
In other words, it means taking the collective average on the time axis of the matrix [N ^* ] [N] ^T (diagonal elements are scalars, off-diagonal elements are generally complex numbers), and E { [N ^* ] [ By setting N] ^T } = [M] ... (10), equation (9) becomes P _Z = [W ^* ] [M] [W] ... (11). The matrix [M] corresponds to a covariance (correlation) matrix of unnecessary wave signal components. Therefore, the SN ratio γ=P _S /P _Z is as follows.

γ=｜[W] ^T [S]｜ ² / [W ^* ] [M] [W]…
(12) Next, find the phase vector that maximizes the S/N ratio γ.
^T = 1, 2...l...k] are sequentially determined by a so-called gradient search method.

That is, when the phase vector determined by the (X+1)th repetition is now represented by [(X+1)], this phase vector [(X+1)] is the phase vector determined by the Xth repetition. (X)] can be expressed as in the following equation.

[(X+1)] = [(X)] +μ∂γ/∂[] | [] = [(X)]…(13) However, in the second term on the right side of equation (13), μ is the convergence of the solution and represents a coefficient that controls stability.

Also, , and the subscript is [] after differentiating equation (14).
obtained by the Xth iteration for [
(X)].

Now, in the above equation (13), the gradient ∂γ/∂ [] of the output SN ratio γ with respect to the vector []
As also noted, ∂γ/∂l=1/[W ^* ] ^T [M] [W] ∂
/∂l([W ^* ] ^T [S ^* ] [S] ^T [W]) − [W ^* ] ^T [S ^* ] [S] ^T [W]/([W]
^T [M] [W] ² ∂／∂l([W ^* ] ^T [M] [W]…(1
5). In the first term on the right side of equation (15), ∂/∂l ([W ^* ] ^T [S ^* ] [S] ^T [W]) = ∂/∂l ( _k 〓 ⁱ⁼¹ Si ^* e ^{-j i} _k 〓 ^m=1 Sme ^{j m} ) = −jSl ^* Wl ^* [S] ^T [W] +jSlWl [S ^* ] ^T [W ^* ] =2Im(Sl ^* Wl ^* [S] ^T [W], Also, in the second term on the right side of equation (15), ∂/∂l([W ^* ] ^T [M][W]) = ∂/∂l( _k 〓 ⁱ⁼¹ e ^{j i} _k 〓 ^m=1 Mime ^{j m} ) = −jWl ^* [WW]l + jWl[M ^* W ^* ]l = 2Im (Wl ^* [MW]l).Therefore, equation (15) is ∂γ/∂l=1/[ W ^* 〕 ^T 〔M〕〔W〕2Im
(Sl ^* Wl ^* [S] ^T [W]) - [W ^* ] ^T [S ^* ] [S] ^T [W]/([W ^* ]
^T [M] [W] ² 2Im (Wl ^* [MW]l)…(16). Now, by defining the vector [] ^T = [ ₁ , ₂ ...l...k]...(17), the slope of the SN ratio γ with respect to the phase vector becomes as follows.

∂γ／∂〔〕＝2Im{〔Ω ^* 〕(a〔S ^* 〕
]−｜a｜ ² [M] [W])}…(18) However Coefficient a = [S] ^T [W] / [W ^* ] ^T [M] [W] ...(20) Note that the coefficient a is a value derived mathematically, and in practice it is a solution of equation (27). You can freely choose between speed and stability for convergence. Also, Im means taking the imaginary part of each element of the matrix in { }. Therefore, the output SN expressed by equation (18)
The complex weight vector Wo of size 1 to maximize the ratio γ is Im {[Ω ^* ] (a [S ^* ] − | a | ² [M]
[W]}=O is satisfied.

Here, from equation (10), [M] = E { [N ^* ] [N] ^T }, and especially if the number of samples for the ensemble average is one, [M] ≒ [N ^* ] [N] ^T ... It can be approximated as (21). Also, when considering the actual transmission and reception status of pulse radar signals, in the time period until the emitted radar pulse signal is received, the signal [S]
= 0, it is considered that antenna output [Y] satisfies [S] + [N] = [Y]≒[N]. (However, [Y] ^T = [Y ₁ , Y ₂ ...Yk],
(Yk is the output signal of the k-th antenna element) Therefore, equation (21) can be set as [M]≒[Y ^* ][Y] ^T ...(22).

Substituting this equation (22) into equation (18), ∂γ/∂[]=2Im{[Ω ^* ](a[S ^* ] −|a| ² [Y ^* ][Y] ^T [W]) } …(23) becomes. Here, since the output y(X) of the signal synthesizer 3 is y(X)=[Y] ^T [W]...(24), equation (23) is ∂γ/∂[]=2Im{[Ω ^* 〕(a〔S ^* 〕−|a| ₂ 〔Y ^* 〕
y)}=2|a| ² Im{[Ω ^* ]([S ^* ]/a ^* −[Y ^* ]y)
}...(25) becomes. Therefore, ∂γ／∂〔〕｜〕＝〔(X)〕=2｜a(X)｜ ² Im
{[Ω ^* (X)] ([S ^* (X)]/a ^* (X)−[Y ^* (X)]y(X))}
…(26) is obtained. Substituting this equation (26) into equation (13), we get [(X+1)] = [(X)] + μsIm {[Ω ^* (X)
]([S ^* (X)]/a ^* (X))−[Y ^* (X)y(X)]}…(27) However, 0< μs, μs; The stability of the solution and the speed of convergence are Coefficient to control. That is, in the signal processor 3 (27)
This means that if an algorithm based on the formula is used, an adaptive algorithm using only the digital phase shifters 21, 22...2l...2k can be realized. The fact that the antenna directivity can be improved by the algorithm based on equation (27) is shown, for example, in the simulation results shown in FIG. 3 of Japanese Patent Application Laid-open No. 54-116868.

This simulation result is based on the assumption that the phase amount of the phase shifter can be controlled to the same value as the highly accurate calculation result of the computer that calculates equation (27).

However, in reality, the accuracy of the digital phase shifter is lower than that of a computer, and the digital phase shifter can usually be directly controlled by using the calculation result of the following equation.

[(X+1)]=[(X)]+360°/2 ⁿ Sg
n{Im([Ω ^* (X)] {[S ^* ](X)]/a ^* (X)−[Y ^* (X)]y(X)})} …(28) where n is the digital transfer The number of bits of the phaser is also
Sgn(δ) represents a sign function, and the sign function Sgn(δ) is determined as follows.

Sgn(δ)=1δ0 −1δ<0 However, when the digital phase shifter is directly controlled by equation (28), a sufficient convergence solution may not be obtained unless extremely fine bits are used. For example, Figure 2 shows the simulation results when the number of bits of the digital phase shifter is n = 4, the number of antenna elements is k = 10, and an unnecessary signal is incident from an azimuth angle of θ = 30°. It can be seen that no zero point of directivity is formed in the incident direction of the wave signal, and desired characteristics are not obtained. This is because the roughness of the calculation result due to the digital calculation of equation (28) hinders the improvement of the accuracy of the arrival direction characteristics of the unnecessary signal. In this case, higher accuracy can be achieved by increasing the number of bits n of the digital phase shifter, but in reality there are many difficulties in increasing n. Therefore, the algorithm based on equation (28) has the drawback that it is not necessarily possible to form a zero point of antenna directivity in the direction of incidence of unnecessary signals, and the S/N ratio cannot be improved.

The present invention eliminates the above-mentioned drawbacks, and controls only the phase of each phase shifter to provide directivity that matches the surrounding environment in which the antenna is placed. An object of the present invention is to provide an adaptive antenna device that has a significantly improved SN ratio and has no inherent gain.

FIG. 3 is a configuration diagram showing an embodiment of the adaptive antenna device according to the present invention. An antenna element group 1 that constitutes an array antenna by a plurality of antenna elements 11, 12...1l...1k receives external radio waves. do. The signals received by this antenna element group 1 are branched, one for each antenna element 11, 12...
...1l... Each digital phase shifter 21, 22 connected in series corresponds to each 1k...2l...
2k is supplied. In addition, each of the above antenna elements 1
1, 12...1l...1k can also be configured by subarrays. Each digital phase shifter 21,
22...2l...2k control the phase of the signals from each antenna element, and the phase-controlled signals are supplied to the combiner 3 and output. Now, the output signal of the combiner 3 is branched, one is taken out as an output, and the other is supplied to the signal processor 4. Based on the signal supplied to the signal processor 4, the arithmetic unit 41 calculates each digital phase shifter 21,
The optimum phase amount at 22...2l...2k is calculated by equation (27) instead of equation (28). After this,
The obtained optimum phase amount is quantized by a quantizer 42 in accordance with the number of bits of each phase shifter, and a phase control signal is supplied to each digital phase shifter 21, 22...2l...2k.

In this case, the equation for finding the optimal phase amount found in the calculator 41, that is, the phase amount that maximizes the SN ratio, is exactly the same as equation (27) and is expressed by the following equation.

[(X+1)] = [(X)] + μsIm {[Ω ^* (X)
]
([S ^* (X)]/a ^* (X)−[Y ^* (X)]y(X))}…(27) where 0< μs, μs; controls the stability of the solution and the speed of convergence coefficient. In this case, the phase amount calculated using equation (27) has less error than the phase amount calculated using equation (28), and is a more optimal phase amount. In the present invention, the arithmetic unit 41 calculates the optimum phase amount based on the algorithm according to equation (27), and then the quantizer 42 converts the optimum phase amount into quantizers according to the number of bits of each digital phase shifter. It has become. Therefore, the quantized optimal phase amount in the present invention has less error and is more accurate than the phase amount shown by equation (28). FIG. 4 shows a device according to the present invention under the same conditions as FIG. 2, that is, the number of bits of the digital phase shifter n = 4, the number of antenna elements k = 10, and an unnecessary signal is incident from an azimuth angle of 30°. The results of the simulation are shown below. It can be seen in Fig. 4 that the antenna directivity at an azimuth angle of 30° has been greatly improved by more than 25 dB compared to Fig. 3. Therefore, if the device according to the present invention is used in an actual adaptive antenna, great effects will be brought about.

As described above, the adaptive antenna device according to the present invention improves the antenna directivity in the direction of arrival of unnecessary signals, that is, significantly improves the SN ratio, and brings about a large practical effect.

In the above embodiment described with reference to FIG. 3, a state in which the adaptive antenna device operates for reception has been described, but of course it can also be considered to be used for both transmission and reception. When considered as a shared transmission/reception configuration, the configuration consists of the point where the signal is taken into the arithmetic unit 41, that is, the antenna element group 1 and the digital phase shifter 2.
1, 22...2l...2k, and a second transmitting/receiving switching device is also provided on the output side of the signal synthesizer 3 to switch between the first and second transmitting/receiving. A signal is supplied to the arithmetic unit 41 through the receiver at the time of reception. It goes without saying that during transmission, the signal combiner 3 has the same configuration and functions as a transmitting signal distributor.

[Brief explanation of drawings]

FIG. 1 is a schematic configuration diagram showing a conventional adaptive antenna device, FIG. 2 is an antenna pattern diagram obtained by the device shown in FIG. 1, and FIG. 3 is an embodiment of the adaptive antenna device according to the present invention. FIG. 4 is a diagram showing an example of an antenna pattern diagram obtained by the apparatus shown in FIG. 3. 1... Antenna element group, 21, 22... 2l...
...2k...digital phase shifter, 3...signal synthesizer,
4... Signal processor, 41... Arithmetic unit, 42... Quantizer.

Claims (1)

[Claims] 1. An antenna element group for transmitting or receiving electromagnetic waves, a plurality of digital phase shifters connected in series to each antenna element of the antenna element group, and a plurality of digital phase shifters connected to the plurality of digital phase shifters. A circuit for distributing or combining signals, and a signal from the digital phase shifter obtained through this circuit and a signal from the antenna element group are introduced to maximize the signal-to-noise ratio of the combined output signal of the circuit. The phase amount [] to make is, [(X+1)] = [(X)] + μsIm {[Ω ^* (
)] ([S ^* (X)] / a ^* (X) - [Y ^* (X)] y (X)
)}, and a calculation means for deriving the phase amount [ ] by the calculation means, quantizes the phase amount [ ] corresponding to the number of bits of the digital phase shifter, and supplies a phase control signal to the digital phase shifter. An adaptive antenna device comprising: quantization means. [However, [] ^T = [ ₁ , ₂ ... _k ], (T means transposition), _k is the phase amount in the phase shifter for the k-th antenna element, μs and a are the stability of the solution a coefficient that controls the speed of convergence, W _k is the weight amount of the phase shifter for the k-th antenna element, W _k = e ^{j k} , [S] ^T = [S ₁ , S ₂ ...S _k ], S _k is the weight amount of the phase shifter for the k-th antenna element. Desired signal, [Y] ^T = [Y ₁ , Y ₂ ... Y _k ], Y _k is the output signal of the k-th antenna element, y is the output signal of the combining circuit, X is the number of repetitions of operation, * is The symbol Im{ } representing complex conjugate means taking the imaginary part of each element of the matrix in { }. ]