JP3100563B2 - Received signal processing device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a received signal processing device provided with a digital beam forming circuit.

[0002]

2. Description of the Related Art In an apparatus for detecting a wide band signal having a wide relative bandwidth using an array sensor in which sensors such as antenna elements are arranged in an array, an arrival angle ( Direction of Arriva
l, hereinafter referred to as DOA. The following method has been conventionally proposed as an algorithm for detecting).

[0003] For example, in prior art document 1 "H. Wang et al.,
“Coherent Signal-Subspace Processing for the Dete
ction and Estimation of Angles of Arrival of Multi
pleWide-Band Sources ", IEEE Transactions on ASSP, Vo
l.33, No. 4, August 1985, the first prior art method is disclosed in prior art document 2 “ROSchmidt,“ Multiple
emitter location and signal parameter estimatio
n ", Proceddings of RADCSpectral Estimation Worksho
p. pp. 243-258, 1979, MUSIC (Multiple Signal Clas) for identifying a plurality of signals.
sification) is a noise-specific subspace method close to the algorithm technology, but is applied when a wideband signal is received using an array antenna. The MUSIC algorithm performs the discrete Fourier transform of the input data in the time direction instead of the original correlation matrix, and then executes the correlation matrix focused on a certain frequency in the frequency domain.

[0004] In the first prior art method, the prior art document 1 is used.
Only shows the detection of two received signals. In fact, as the number of signals increases, the performance of the method degrades significantly, in which case the number of signal eigenvectors is not the same as the number of signals as in the narrowband case, but rather is larger. Also, the rate of increase depends on the number of interfering signals and the increase in signal bandwidth and other variables. Due to this increase, the noise subspace to be detected may be easily erased, or signal detection by the MUSIC algorithm may not be possible. Even when the number of signals is small, DOA is directly applied to the antenna element output signal.
The element space process for performing the estimation process requires a large-scale eigen decomposition of a correlation matrix, which may cause high computational cost and existing numerical uncertainty and instability.

[0005] Prior art document 3 "KMBuckley et a
l., “Broad-Band Signal-SubspaceSpatial-Spectrum (B
ASS-ALE) Estimation ", IEEE Transactions on ASSP, Vo
l.36, No. 7, July 1988 ", the BASS-ALE method, which is a second conventional example, takes a large observation space, that is, snapshots from separate sensors or antenna elements. This is to solve the above-described problem in the first conventional example by taking not only (or signal samples) but also some signal samples in a time tap manner inside the sensor. Although the detection performance is improved for a small observation space (only two signals are verified in the related art document 3), it is presumed that such a case does not occur when the number of interferences is large. . Furthermore, due to the increase in the observation space, the calculation cost becomes higher than the coherent signal subspace method, which is the first conventional method.

SUMMARY OF THE INVENTION An object of the present invention is to solve the above-mentioned problems, and it is possible to detect even a relatively large number of received signals, for example, three or more, received by an array antenna. An object of the present invention is to provide a received signal processing device capable of greatly reducing the calculation cost.

[0007]

[0008]

[0009]

[0010]

[0011]

[0012]

According to a first aspect of the present invention, there is provided a received signal processing apparatus comprising: a plurality of N antenna elements which are received by an array antenna including a plurality of N antenna elements arranged in a line at a predetermined element interval; Conversion means for A / D-converting the N received signals into a plurality of N digital received signals and outputting the digital received signals, and detecting the angle of arrival of the desired received signal and detecting and outputting the angle of arrival signal indicating the angle of arrival; A reception signal processing device comprising: a plurality of M digital reception signals different from each other among the sub-array signals from a plurality of N digital reception signals output from the conversion means. Selecting means for selecting and outputting a plurality of B sets of subarray signals, a plurality of M FIR digital filters each having a predetermined filter coefficient, An adder for adding and outputting a plurality of M signals output from the FIR type digital filter, each having the same transfer function, and based on a plurality of B sets of sub-array signals output from the selection means; A plurality of B forming means for forming and outputting digital beam signals so that the main beam of the array antenna is directed in the direction of arrival of the desired received signal and the received signal in the direction of arrival of the interference wave signal is made zero. Wherein the detecting means detects a desired received signal based on the plurality of B digital signals output from the plurality of B forming means, and detects an angle of arrival of the desired received signal. The received signal processing device detects and outputs an angle of arrival signal indicating the angle of arrival, and the received signal processing device, based on the angle of arrival signal output from the detection means, detects the main beam of the array antenna. Arithmetic means for calculating and setting and controlling the filter coefficients of the plurality of M FIR digital filters of each of the above-mentioned forming means so as to make the received signal in the arrival direction of the desired received signal and in the arrival direction of the interference wave signal zero. Is further provided.

[0013]

[0014]

[0015]

[0016]

[0017]

[0018]

[0019]

[0020]

According to a second aspect of the present invention, there is provided a reception signal processing apparatus comprising: a plurality of N reception elements received by an array antenna including a plurality of N antenna elements arranged in a line at a predetermined element interval; Conversion means for A / D-converting a signal into a plurality of N digital reception signals and outputting the same; and arrival angle detection means for detecting an arrival angle of a desired reception signal and detecting and outputting an arrival angle signal indicating the arrival angle A reception signal processing device comprising: a plurality of M digital reception signals set different from each other among the sub-array signals from a plurality of N digital reception signals output from the conversion means; Selecting means for selecting and outputting a plurality of B sets of sub-array signals, and u signal processing means, each of which has a predetermined filter coefficient A plurality of M FIR digital filters and an adder for adding and outputting a plurality of M signals output from the plurality of M FIR digital filters; And the main beam of the array antenna is shifted in the same direction of arrival of the same desired reception signal by the same signal processing means based on a plurality of B sets of sub-array signals output from the selection means. A plurality of B forming means for respectively forming and outputting digital beam signals so as to make the received signal in the direction of arrival of the interference wave signal zero, and a plurality of B forming means output from the plurality of B forming means
Reception signal detection means for detecting a desired reception signal based on the digital signals, the arrival angle detection means,
The arrival angle of at least one desired reception signal is determined based on a plurality of N digital reception signals output from the conversion means or one set of a plurality of B sets of sub-array signals output from the selection means. Detects and outputs an angle of arrival signal indicating the angle of arrival, and the received signal processing device, based on the angle of arrival signal output from the angle of arrival detection means,
A plurality of M FIR digital devices of each of the above-mentioned forming means for directing a plurality of u main beams of the array antenna in a direction of arrival of at least one received signal and for reducing a received signal in a direction of arrival of an interference wave signal to zero. It is characterized by further comprising a calculating means for calculating and controlling the filter coefficient of the filter.

[0022]

[0023]

[0024]

[0025]

[0026]

[0027]

[0028]

[0029]

[0030]

[0031]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, a first embodiment according to the present invention will be described with reference to the drawings. First according to the present invention
The method of the embodiment solves the problem of resolution for obtaining a reliable noise subspace spectrum by operating in a smaller observation space. Here, by performing beam forming processing and selecting a specific direction to be detected and performing beam space processing (beam space processing), it is possible to filter the interference wave signal and suppress its influence. . The number of beams to be processed is generally much less than the number of elements in the antenna, so the computational cost is typically lower, especially in real radio communications where a larger antenna array is needed to resolve more received signals. It can be significantly reduced compared to the example method. further,
The method of this embodiment was performed without performance degradation, and in fact, our inventor's simulation shows that the beamspace processing has improved performance.

FIG. 1 is a block diagram showing a configuration of a received signal processing device according to a first embodiment of the present invention. In the reception signal processing apparatus of the present embodiment, the N / N received signals received by the array antenna 100 including the N antenna elements 1-0 to 1- (N-1) are converted by the A / D converter 3 After A / D conversion by −0 to 3- (N−1),
(A) A set of a plurality of M digital in-phase / quadrature signals x _n (m) (hereinafter referred to as “sub-array signals”) SA _{0 to} SA _B−1 different from each other in each sub-array signal is selected by B sets of the number of beams. A sub-array signal selection circuit 7 for output;
(B) Each digital beam forming circuit 8-0 to 8- (B
-1) is a plurality of M FIR digital filters 9-0.
9- (M-1) and one adder 12 and have the same transfer function (therefore, each of the digital beam forming circuits 8-0 to 8- (B-1) has the same beam transfer function). Is formed), and each of the filters 9-0 to 9- (M-1)
Filter coefficient w _{k, s} (k = 0,1,..., M−1; s =
0, 1,..., Q-1) are calculated and set by the FIR digital filter coefficient calculator 13, and based on the sub-array signals SA _{0 to} SA _B-1 , the main beam of the array antenna 100 is changed to a desired one. The digital beam signals y ₀ (m) to y are set so that the reception signal in the direction of arrival of the received signal and in the direction of arrival of the interference wave signal is made zero.
A plurality of B (B <N) digital beam forming circuits (hereinafter referred to as DBF circuits) for generating and outputting _B-1 (m).
8-0 to 8- (B-1) and (c) DBF circuit 8-0
MUSIC (Multiple Signal Cl) proposed in the prior art document 1 based on the digital beam signals y ₀ (m) to y _B-1 (m) output from 8- (B-1).
assification) using a modified version of the algorithm
A reception signal detection circuit 10 for detecting an arrival angle θ of a desired reception signal and outputting an arrival angle signal indicating the arrival angle, and detecting and outputting a reception baseband signal of the desired reception signal;
And characterized in that:

Here, in the received signal processing apparatus of the present embodiment, the FIR digital filter coefficient calculator 13
Is an arrival angle θ which is detected by the reception signal detection circuit 10 and is a direction in which a beam should be formed, as described in detail later.
, The frequency response G (F ₁ ,
F ₂ ), and the calculated frequency response G (F ₁ ,
F ₂ ) is subjected to an inverse Fourier transform to calculate an impulse response g (m ₁ , m ₂ ).
(m ₁ , m ₂ ) based on the load coefficient (or filter coefficient) w
_{k and s} are calculated and output, and the FIR digital filter 9
−0 to 9- (M−1) are the respective load coefficients w
_The input digital in-phase / quadrature signal x _k (m) is filtered based on _{k and s} , and output to the adder 12. The adder 12 outputs a plurality of B filtered digital in-phase signals.
The quadrature signal x _k ′ (m) is added, and the digital beam signal after the addition is converted into a digital beam signal y from which a wideband desired signal is extracted and a wideband interference signal is suppressed.
₀ (m) to y _B-1 (m). Where n =
, N-1, k = 0, 1,..., M-1, b =
0, 1,..., B-1, and the same applies hereinafter unless otherwise specified in this specification. Therefore, N is the number of antenna elements, M is the number of signals in one set of sub-array signals, and B is the number of digital beams to be formed (however, the same beam). The digital in-phase / quadrature signal x
m in parentheses () in _k (m) is the time (time number)
Represents Further, the digital in-phase / quadrature signal x
_k (m) and _xk '(m) represent two signals that are so-called I-channel signals and Q-channel signals that are orthogonal to each other.

Hereinafter, the received signal processing apparatus according to the first embodiment will be described in detail with reference to FIG. In the received signal processing device of the first embodiment, the array antenna 10
Numeral 0 is a linear array antenna in which N antenna elements 1-n are juxtaposed at a predetermined element interval d on a straight line. The high-frequency signals received by the plurality of N antenna elements 1-n of the array antenna 100 are converted to intermediate frequency signals by a receiver 2-n including a high-frequency amplifier, a frequency converter, an intermediate-frequency amplifier, and the like. , A / D converters 3-n are converted into digital intermediate frequency signals. These digital intermediate frequency signals are supplied to the digital local oscillator 6.
0 are each mixed with a digital sine-wave signal of the generated intermediate frequency f _IF and the mixer 4-n by high-frequency components are cut by the low pass digital filter 5-n, a low pass digital filter 5-n A digital in-phase / quadrature signal x _n (m), which is a complex number, is output.
Here, the digital in-phase / quadrature signal x _n (m) includes two signals, that is, a so-called I-channel signal and a Q-channel signal, which are orthogonal to each other, in a complex number format. Each digital in-phase / quadrature signal x _n (m) is input to the distributor 6-n. Each distributor 6-n, for example, distributes the digital in-phase / quadrature signal x _n (m) into predetermined P digital in-phase / quadrature signals x _n (m) and outputs the same to the sub-array signal selection circuit 7. The sub-array signal selection circuit 7 sets two or more digital in-phase / quadrature signals x
_{When n} (m) is not required, the distributor 6-n may not be provided.

Next, based on the distributed digital in-phase / quadrature signals x _n (m), the sub-array signal selection circuit 7 sets a set of a plurality of M different digital in-phase / quadrature signals x _n (m) (hereinafter, referred to as a set). This is called a sub-array signal.) SA ₀
To SA _B-1 for each of the B beams, and
Output to the BF circuit 8-b. Here, M is the DBF circuit 8
-B is a constant determined according to the desired characteristic. The physical significance of the processing of the sub-array signal selection circuit 7 is to reduce the amount of signal processing calculation in the reception signal detection circuit 10. Specifically, when signals are selected by shifting one signal at a time, the results are as shown in Table 1. In the case of Table 1, N = B
+ M-1, for example, B = 10, M = 17, N = 2
6.

[0036]

[Table 1] ─────────────────────────────────── Sub-array signal Digital in-phase / quadrature signal ──── ─────────────────────────────── SA ₀ x ₀ (m), x ₁ (m), ..., x _{M- 1 (m) SA 1 x 1} (m), x 2 (m), ..., x M (m) ... ... SA n x n (m), x n + 1 (m), ..., x n + M- ₁ (m) ... SA _B-1 x _B-1 (m), x _B (m), ..., x _{B + M-2} (m) ─────────────── ────────────────────

As another specific example, when signals are selected by shifting two signals at a time, the results are as shown in Table 2. In the case of Table 2, N = 2 (B-1) + M-1. For example, B = 10, M = 17, and N = 34.

[0038]

[Table 2] ─────────────────────────────────── Sub-array signal Digital in-phase / quadrature signal ──── ─────────────────────────────── SA ₀ x ₀ (m), x ₁ (m), ..., x _{M- 1} (m) SA ₁ x ₂ (m), x ₃ (m), ..., x _{M + 1} (m) ... SA _n x _2n-1 (m), x _2n (m), ..., x _{2n + M-2} (m) ... SA _B-1 x2 _{(B-1) -1} (m), x2 _(B-1) (m), ..., x2 _{(B-1) + M-2} ( m) ───────────────────────────────────

Digital in-phase / quadrature signal x of sub-array signal SA _n output from sub-array signal selection circuit 7
_n (m), x _{n + 1} (m),..., x _{n + M-1} (m) (hereinafter collectively referred to as x _k (m)) are respectively included in each DBF circuit 8-b. FIR digital filters 9-0 to 9-
(M-1) (hereinafter referred to as 9-k). Each of the DBF circuits 8-0 to 8- (B-1) (hereinafter referred to as
They are collectively referred to as 8-b. ) Includes M FIR digital filters 9-k and an adder 12. The FIR digital filter coefficient calculator 13 indicates an arrival angle θ which is a direction in which a beam is to be formed, as described in detail later.
Based on the angle of arrival signal from the reception signal detection circuit 10, D
In the digital beam signal y _b (m), which is an output signal of the BF circuit 8-b, weighting factors w _{k, 0} , w _{k, 1,.}
, Wk _{, q-1} are calculated to obtain the weighting factors _{wk, 0} , _{wk, 1} ,.
w _{k, q -1} is output to the FIR digital filter 9-k.

Then, in the DBF circuit 8-b, the FI
R digital filter 9-k is weighting factor input from the FIR type digital filter coefficient calculator _{13 w k, 0, w k} , 1, ..., w k, based on the _q-1, digital in-phase input Digitally filtering the quadrature signal x _k (m), and filtering the digital in-phase and quadrature signal x _k '(m)
Is output to the adder 10. The adder 10 adds a plurality of M filtered digital in-phase / quadrature signals x _k ′ (m) input from each of the FIR digital filters 9-k, and converts the added signal to a wideband desired signal. Is extracted and output to the reception signal detection circuit 10 as a digital beam signal y _b (m) which is an output signal in which a wideband interference signal is suppressed.

Next, the configuration of the FIR digital filter 9-k will be described with reference to FIG. As shown in FIG. 2, the FIR digital filter 9-k includes (q-1) delay units 91-1 to 91- (q-1) and q multipliers 92-1 to 92-q And (q-1) adders 93
-1 to 93- (q-1). Where FIR
In the type digital filter 9-k, q is called a tap length, and is set to an odd number in the first embodiment. And
The digital in-phase / quadrature signal x _k (m) input to the FIR digital filter 9-k is input to the delay unit 91-1 and the multiplier 92-1. The weighting coefficient w _{k, s-1} input to the FIR digital filter 9-k is input to a multiplier 92-s (s = 1, 2,..., Q).

In the FIR digital filter 9-k, the delay unit 91-s (s = 1, 2,..., Q-1) receives the input digital in-phase / quadrature signal x _k (m-s + 1).
Is delayed by one sample period, and a signal x _k (ms) delayed by one sample period is output to the delay unit 91- (s + 1) and the multiplier 92- (s + 1). The multiplier 92-1 is
The input digital in-phase / quadrature signal x _k (m) is multiplied by the weight coefficient w _{k, 0} and output to the adder 93-1.
The multiplier 92-s (s = 2, 3,..., Q) multiplies the input signal x _k (m−s + 1) by the weighting coefficient w _{k, s−1,} and adds the result to the adder 93- (s). Output to -1). Adder 93-
1 is a signal input from the multiplier 92-1 and the multiplier 92-1
-2 is added to the signal inputted from the adder 93-2.
Output to The adder 93-s (s = 2, 3,..., Q-
2) adds the signal input from the adder 93- (s-1) and the signal input from the multiplier 92- (s + 1), and outputs the result to the adder 93- (s + 1). Adder 93-
(Q-1) adds the signal input from the adder 93- (q-2) and the signal input from the multiplier 92-q, and outputs the result to the adder 12. Here, in the first embodiment, (q-1) adders 93-1 to 93- (q-
1) is provided, but the present invention is not limited to this, and (q)
Instead of the -1) adders 93-1 to 93- (q-1), q output signals output from the multipliers 92-1 to 92-q are collectively added and the adder 12 is added. May be configured using a single adder that outputs the data to Further, q output signals output from the multipliers 92-1 to 92-q may be directly output to the adder 12 without using such an adder.

The FIR digital filter 9-k configured as described above includes the weighting coefficients w _{k, 0} , w _{k, 1} , input from the FIR digital filter coefficient calculator 13,
.., W _{k, q−1} , the digital in-phase / quadrature signal x _k (m) input from the sub-array signal selection circuit 7 is digitally filtered, and the filtered digital beam signal y
_b (m) is output to the reception signal detection circuit 10. The reception signal detection circuit 10 detects the reception baseband signal based on the input plurality of B digital beam signals y _b (m) using the method of the known prior art document 1 and sends it to the demodulation circuit 11. At the same time, the angle of arrival θ of each received baseband signal is detected, and the angle of arrival signal indicating the angle of arrival θ is output to the FIR digital filter coefficient calculator 13. Further, the demodulation circuit 11 demodulates the input reception baseband signal by a predetermined demodulation method and outputs the demodulated signal.

The M FIR digital filters 9
The DBF circuit 8-b having -k can be considered as a two-dimensional digital filter. Although the structure of the FIR digital filter 9-k in FIG. 2 is a direct type, it may be another type, for example, a cascade type or a lattice type. However,
The calculation method of the weight coefficient w _{k, s} input to the FIR digital filter 9-k is based on the FIR digital filter 9-k.
It depends on the structure of k.

Next, a method of calculating the weighting coefficient w _{k, s in} the FIR digital filter coefficient calculator 13 will be described. First, as shown in FIG. 1, the arrival angle θ of a desired signal is defined as the angle between the perpendicular of a line in which N antenna elements 1-n are juxtaposed and the direction in which the desired signal is incident. Also, a normalized time frequencies F ₁ by using the sampling frequency f _s of the carrier frequency f _c and A / D converter 3-n of the frequency f ₁ and the carrier signal of the desired signal is a high frequency signal by the following equation 1 Define. Further, using the angle of arrival θ, the speed of light c, and the frequency f _{1 of the} desired signal, the non-normalized spatial frequency f ₂ is defined as in Equation _2, and the unnormalized spatial frequency f ₂ is expressed by Equation 3. normalized to the inverse of the element spacing d of the antenna elements 1-n, we define the normalized spatial frequency F _2. Here, the reciprocal 1 / d of the element interval d of the antenna element is called a spatial sampling frequency. That is, the normalized time frequency F
₁ is a frequency obtained by normalizing the difference by the sampling frequency f _s of the frequency f ₁ and the carrier frequency f _c, the juxtaposition of the antenna elements 1-n at a certain time normalized spatial frequency F ₂ is in consideration of the angle of arrival θ Spatial frequency normalized on the line.

[0046]

[Number 1] _{F 1 = (f 1 -f c} ) / f s

F ₂ = (f ₁ sin θ) / c

## EQU3 ## F ₂ = df ₂ = (df ₁ sin θ) / c = ｛(dsin
_{θ) / c} (f s} F 1 + f c)

When the DBF circuit 8-b is regarded as a two-dimensional digital filter, the transfer characteristic H (f ₁ , θ) can be expressed by the following equation 4, and the transfer characteristic H (f ₁ , θ) )
Is a frequency response G shown in Equation 5 using the normalized time frequency F ₁ and the normalized spatial frequency F ₂ defined as described above.
(F ₁ , F ₂ ). That is, the frequency response G (F ₁ , F ₂ ) is a frequency response of the normalized time frequency F ₁ and the normalized spatial frequency F ₂ on a two-dimensional frequency plane.

[0048]

(Equation 4)

(Equation 5)

Where T _s is the sampling interval in the DBF circuit 8-b, and
b, the reciprocal of the sampling frequency f _s , ie, T _s
= Represented by 1 / f _s. In Equations 4 and 5, q
Is the tap length of the FIR digital filter 9-k as described above. Equations (4) and (5) above are equations that hold when the element spacing d is assumed to be constant, and there is a restriction that this method can be applied only to a linear array where the element spacing d is constant. However, when the N antenna elements 1-n are arranged at regular intervals in the vertical and horizontal directions in a grid pattern on a plane, it can be dealt with by expanding Equations 4 and 5. That is, the present invention can be applied to a case where the digital filter is three-dimensional (two-dimensional space-time).

The DBF circuit 8-b converts a desired characteristic on a frequency-angle plane for forming a broadband beam in a desired direction into a two-dimensional frequency plane of a normalized time frequency F ₁ and a normalized spatial frequency F ₂ plane. This is realized by converting the characteristics of the two-dimensional digital filter to the desired characteristics on the two-dimensional frequency plane by converting the characteristics into the desired characteristics described above. In the above-described processing, when information is transmitted using a carrier wave, the broadband DBF circuit 8-b for processing a signal having a wide fractional bandwidth has the frequency shown in FIG. -It has been found that the desired characteristic in the angle plane is represented by the desired passband shown in FIG. 3B on the two-dimensional frequency plane.

Therefore, first, the present inventor sets the DBF circuit 8-b
As a result of a detailed study of the digital in-phase / quadrature signal x _k (m) that is input to the digital in-phase / quadrature signal x _k (m), the DBF
It has been found that a two-dimensional spectrum relating to time and space at the time of input to the circuit 8-b is on a straight line represented by the following equation 6 on a two-dimensional frequency plane. FIG. 4 is a diagram schematically showing the two-dimensional spectrum Sp.

[0052]

[6] _{F 2 = {(dsinθ) /} c} (f s F 1 + f c)

That is, the straight line represented by Equation 6 is shown in FIG.
It is a straight line located at the center in the width direction of the desired pass band of the rectangle shown in FIG. Therefore, the desired rectangular passband shown in FIG. 3B can be expressed by the following inequality shown in Equation 7.
A / D and the first axis of the transducer 3-n normalized normalized time frequencies F ₁ at the sampling frequency of the second axis of the element normalized normalized spatial frequency F ₂ by the reciprocal of the interval And has a center on the second axis and has a predetermined width and extends from the negative normalized time frequency to the positive normalized time frequency in the two-dimensional frequency plane formed by

[0054]

Equation 7] {(dsinθ) / c} ( f s F 1 + f c) -ε ≦ F 2 ≦ {(ds
_{inθ) / c} (f s} F 1 + f c) + ε

Here, ε in Equation 7 is a predetermined small number corresponding to the width of the desired pass band of the rectangle shown in FIG. 3B. That is, the amplitude characteristic D in the desired pass band of the rectangular 3 shown by the number _{7 (b) (F 1,} F 2) is = 1, the desired passband in the external amplitude characteristic D (F _1, F ₂ ) = 0, if a DBF circuit 8-b, which is a two-dimensional digital filter capable of approximately realizing the amplitude characteristic, can be constructed,
DBF circuit 8 for forming a broadband beam in the direction of arrival angle θ
−b can be configured. Here, in the first embodiment, even in the vicinity of F ₁ = ± 0.5, the amplitude characteristic D (F
₁ , F ₂ ) = 0.

In the present embodiment, the frequency response of a one-dimensional zero-phase FIR type low-pass digital filter (hereinafter referred to as a one-dimensional original filter) is determined on a two-dimensional frequency plane according to a beam forming direction. By performing a variable transformation so as to be transformed into a straight line group of
The weight coefficient w _{k, s} is calculated by approximating the amplitude characteristic of the BF circuit 8-b to a desired characteristic. Hereinafter, the load coefficient calculation method of the present embodiment will be described in detail.

First, the variable conversion applied to the one-dimensional zero-phase FIR type low-pass digital filter will be described, and then the calculation method of the weight coefficient w _{k, s} of the DBF circuit 8-b will be described.

In the variable conversion in the first embodiment, cos (2πF) appearing in the frequency response equation of the one-dimensional zero-phase FIR type low-pass digital filter is converted into four appropriate functions.
cos (2πF ₁ ), sin (2πF ₁ ), cos (2πF ₂ ), sin (2πF ₂ )
This is performed by replacing the function S (F ₁ , F ₂ ) expressed by Here, the function S (F ₁ , F ₂ ) is a function approximating the function D _T (F ₁ , F ₂ ) shown in the following Expression 8, and will be described later in detail.

[0059]

Equation 8] _{D T (F 1, F 2} ) = cos [2πF 2 / {(f s / f c) F 1 +1}
−2πF _2shift ]

Here, F _2shift in _Expression 8 is a shift amount for performing parallel translation along the F ₂ axis, and is determined by an arrival angle θ corresponding to a direction in which a beam is to be formed. Function D _T of the number 8 (F _1, F ₂₎ is the number 6, it can be derived by using the definition formula of F _{2Shift "F 2shift = (df c / c} ) sinθ " 1-dimensional original filter Is mapped to a straight line on a two-dimensional frequency plane expressed by the following equation (9).

[0061]

Equation 9] _{F = F 2 / {(f} s / f c) F 1 +1} -F 2shift

For example, the one-dimensional frequency F = 0 of the one-dimensional original filter is represented by a straight line F ₂ = F _2shift {(f
_s / f _c ) F ₁ +1}. Therefore, Hodokose the mapping for one dimensional FIR type narrowband low-band pass digital filter used in the present embodiment, the straight line _{F 2} = F 2shift on passband 2-dimensional frequency plane {(f _s / f _c ) F ₁ +1}. In other words, in order to form a beam in the direction of the angle of arrival θ, the pass band on the two-dimensional frequency plane is F ₂ =
Since _{(df c sinθ / c) {} (f s / f c) F 1 +1} appear in the vicinity, the shift amount F _2Shift determined according to the definition formula of the F _2Shift,
For a one-dimensional FIR type narrow band low-pass digital filter, instead of the variable conversion method used in the first embodiment,
What is necessary is to perform the mapping of Formula 8 or Formula 9.

If the above-described mapping is ideally performed, each equal-amplitude line of the amplitude characteristic of the two-dimensional digital filter obtained by such a variable transformation follows the straight line of the formula 6 in the first embodiment. . On the other hand, the straight line represented by Expression 6 corresponds to a straight line parallel to the frequency axis at the angle θ on the frequency-angle plane. Therefore, the fact that the equal-amplitude line of the amplitude characteristic of the two-dimensional digital filter on the two-dimensional frequency plane follows Expression 6 means that the equal-amplitude line is parallel to the frequency axis on the corresponding frequency-angle plane. This means that the directional characteristics of the DBF circuit 8-b do not depend on the frequency. That is, 2 obtained by such a variable conversion
The directional characteristics of the DBF circuit 8-b as a dimensional digital filter depend only on the angle of arrival θ and not on the frequency. In other words, in the first embodiment, variable conversion is performed on the one-dimensional FIR narrow-band low-pass digital filter so that the directivity of the DBF circuit 8-b as a two-dimensional digital filter does not depend on frequency. ing.

Next, a method of obtaining a function S (F ₁ , F ₂ ) approximating the function D _T (F ₁ , F ₂ ) of Expression 8 will be described. This approximation is based on a window function method of terminating the impulse response to the frequency characteristic represented by Expression 8 with a two-dimensional window function (Prior Art Document 4, "Nishikawa et al.," Design of Fan Filter by Fourier Transform Method ", IEICE). Proceedings of the 6th Digital Signal Processing Symposium, Lecture No. B3-
2, pp. 287-292, 1991 ", which is called the Fourier series method. ) Is used. The specific procedure is shown in the following procedures SS1 to SS4.

(Procedure SS1) The normalized time frequency F ₁ and the normalized spatial frequency F ₂ are defined as −0.5 ≦ F ₁ <0.5, −
Sampling is performed at appropriate intervals in the range of 0.5 ≦ F ₂ <0.5, and each sampled normalized time frequency F ₁
And function D _T (F ₁ for each normalized spatial frequency F _2,
Calculate each frequency response value of F ₂ ). When fast Fourier transform is used later, the number of points to be sampled J (even number)
Is preferably set to a value that can be represented by a power of 2 such as 128 points. (Procedure SS2) Procedure SS1
In the obtained function _{D T (F 1, F 2} ) 2 -dimensional inverse discrete respective frequency response value (high-speed) and Fourier transform function D _T (F _1,
An impulse response (complex number) for the frequency response value of F ₂ ) is obtained. The impulse response is called an ideal impulse response and is represented by h _TI (m ₁ , m ₂ ). Here, (m ₁ , m
₂ ) = (0,0) is the ideal impulse response h _TI (m ₁ , m ₂ )
To be at the center of Arguments m ₁ and m ₂ are
-J / 2, (-J / 2) +1, ...,-1,0,1, ...,
(J / 2) -1, + J / 2. (Procedure SS3) The ideal impulse response h _TI (m ₁ , m ₂ ) is converted to a two-dimensional window function w
_{Cut off at 2D} (m ₁ , m ₂ ). That is, the product of the ideal impulse response h _TI (m ₁ , m ₂ ) and the two-dimensional window function w _2D (m ₁ , m ₂ ) is obtained. Here, the two-dimensional window function w _2D (m ₁ , m ₂ ) is represented by the following equation 10, for example.

[0066]

W _2D (m ₁ , m ₂ ) = w _1D (m ₁ ) w _1D (m ₂ )

[0067] In Equation 10, w _1D (m _i), to the truncation number of terms L _a is an odd number, an argument m _i is the time that satisfies the following equation, the center of the window function m _i = 0 _(i = 1, 2)
Is represented by an appropriate window function other than a rectangular window such as a Hamming window function. If the following expression is not satisfied, the function value is 0.
Is a window function.

[0068]

-(L _a -1) / 2 ≦ m _i ≦ (L _a −1) / 2

Here, the impulse response h _T ′ (m ₁ , m ₂ ) obtained by this operation is represented by the following equation.

[0070]

H ′ _T (m ₁ , m ₂ ) = h _TI (m ₁ , m ₂ ) w _2D (m ₁ , m ₂ ) where − (L _a −1) / 2 ≦ _mi ≦ ( L _a -1) / 2, i
= 1,2

(Procedure SS4) Scaling is performed. This means that the maximum value and the minimum value of the frequency response value of the impulse response h _T ′ (m ₁ , m ₂ ) obtained in step SS3 are set to 1 and −1, respectively, so as not to deteriorate the characteristics of the DBF circuit 8-b. Operation. Therefore, the frequency response value of the impulse response h _T ′ (m ₁ , m ₂ ) obtained in (step SS3) is calculated, and the maximum value h _T ′ _max and the minimum value h _T ′ _min are obtained. Then, the scaling constants c ₁ and c ₂ represented by Expression 13 are calculated, and the two-dimensional impulse response h _T (S _T after the scaling represented by Expression 14 or 15 is calculated using these scaling constants c ₁ and c _2. m ₁ , m ₂ ). In the two-dimensional impulse response h _T (m ₁ , m ₂ ), h _T (m ₁ , m ₂ ) = h _T * (−
m ₁ , −m ₂ ) and therefore its frequency response S
(F ₁ , F ₂ ) is a real number, and cos (2πF ₁ ), sin
(2πF ₁ ), cos (2πF ₂ ), sin (2πF ₂ )
Function. Here, * of h _T * (− m ₁ , −m ₂ )
Represents a complex conjugate of the two-dimensional impulse response h _T (-m ₁ , -m ₂ ).

[0072]

C ₁ = 2 / (h ′ _Tmax −h ′ _Tmin ), c ₂ = c ₁
h ' _Tmax -1

When (m ₁ , m ₂ ) ≠ (0, 0), h _T (m ₁ ,
m ₂ ) = c ₁ h ′ _T (m ₁ , m ₂ )

When (m ₁ , m ₂ ) = (0,0), h _T (m ₁ ,
m ₂ ) = c ₁ h ′ _T (0,0) −c ₂

Next, a method of calculating the weighting factor w _{k, s} in the DBF circuit 8-b using the two-dimensional impulse response h _T (m ₁ , m ₂ ) obtained by the above procedure will be described. In the following calculation method, the two-dimensional impulse response h _T (m ₁ ,
m ₂ ) is a function S (F ₁ , F ₂ ). The calculation method of the load coefficient in the DBF circuit 8-b is as follows.
It comprises steps 1 to S3.

<Step S1> First, a frequency response P (F) of a one-dimensional zero-phase FIR type low-pass digital filter whose impulse response length is M _c (odd number) is obtained. The frequency response P (F) of the one-dimensional zero-phase FIR type low-pass digital filter is expressed by the following equation. In the first embodiment, the value of the number of taps M _c in order to obtain good characteristics is set smaller than the array number M in the set of sub-arrays. p
(M) is an impulse response of the one-dimensional zero-phase FIR type low-pass digital filter. T _m (x) is an m-th order Chebyshev polynomial of the first kind.

[0075]

(Equation 16)

<Step S2> Next, variable conversion is performed to replace cos (2πF) on the right side of Equation 16 with a function S (F ₁ , F ₂ ) represented by Equation 18. Therefore, the frequency response G (F ₁ , F ₂ ) after replacement is expressed by the following equation (17).

[0077]

[Equation 17]

(Equation 18)

<Step S3> Next, similarly to the procedure SS2, the impulse response g (m ₁ , m ₂ ) of the frequency response G (F ₁ , F ₂ ) represented by Expression 17 is calculated, and the impulse is calculated. From the response g (m ₁ , m ₂ ), w _{k , s} is obtained.

[0079]

[Equation 19] w _{k, s} = g (s− (q−1) / 2, −k + (M−1) /
2) where k = 1, 2,..., M−1; s = 0, 1,.
−1, and q is the number of taps.

The iso-amplitude line of the amplitude characteristic on the frequency-angle plane of the DBF circuit 8-b obtained by the above method appears approximately parallel to the frequency axis on the frequency-angle plane. Therefore, the characteristics of the one-dimensional original filter can be set to a bandwidth corresponding to a desired beam width instead of a narrow band, and a side lobe level in a corresponding direction can be set by increasing an attenuation amount of a stop band in a specific frequency range. Can be reduced, thereby the DBF circuit 8-
As described in detail later, b can form a wide-angle and wide-angle beam, and can make the side lobe level lower than the side lobe level in other regions within a range of a predetermined direction. Note that, with the method of the related art document 4, the side lobe level in a range in a predetermined direction cannot be reduced below the side lobe level in another region.

Next, a method of calculating the weighting factor of the DBF circuit 8-b so that the side lobe level in a predetermined direction range is lower than the side lobe levels in other regions will be described. In the following description, it is assumed that the direction in which a beam is formed is an angle θ _0, and the angle θ in a direction range in which a side lobe level is to be lowered is θ ₁ ≦ θ ≦ θ ₂ .
Here, it is assumed that θ ₀ > θ ₂ . That is, the direction range in which the side lobe level is desired to be reduced is smaller than θ ₀ . However, the present invention is not limited to this, and θ ₀ <θ ₁ may be set, and a plurality of direction ranges in which the side lobe level is desired to be reduced may be provided.

By using the above-described variable conversion,
The desired characteristics in the frequency-angle plane shown in (a) can be expressed on the two-dimensional frequency plane as shown in FIG. The one-dimensional digital filter to be designed is not the low-pass digital filter in the first embodiment, but a linear phase FIR type band-pass digital filter having complex coefficients (hereinafter referred to as a one-dimensional complex coefficient band-pass digital filter). It is. The one-dimensional complex coefficient band-pass digital filter has a narrow band pass band centered on the frequency F ₀ represented by the following equation 20, and a frequency range F ₀₁ ≦ F ≦ F represented by the following _equation 21 The side lobe level in ₀₂ is set to be lower than the side lobe levels in the other stop bands. This is described, for example, in Prior Art Document 5, "Akira Nishihara et al.," Design of Complex Digital Filter by Remez Algorithm ", IEICE Technical Report, CAS 88-20, pp. 37-42, 1988.
June, "can be used.

[0083]

F ₀ = (df _c / c) sin θ ₀

## _EQU21 ## F _0i = (df _c / c) sin θ _i , where i =
1,2

As for the frequency response, only the zero phase portion is considered. Frequency response P of one-dimensional zero-phase FIR complex coefficient digital filter with impulse response length M _c (odd)
_c (F) can be expressed as in Equation 22. Here, the impulse response _pc (m) is represented by _pc (m) = _pc * (-m).
(Where * represents a complex conjugate). Further, the number of taps M _c is chosen as in step S1 of the calculation method of the weight coefficients in the DBF circuit 8-b.

[0085]

(Equation 22)

Here, U _m (cos (2πF)) in Expression 22
Is a Chebyshev function of the second kind,
There is a relationship of Equation 24 and Equation 25.

[0087]

U ₀ (cos (2πF)) = 0

[Formula 24] U ₁ (cos (2πF)) = sin (2πF)

## EQU25 ## U _m (cos (2πF)) = 2 cos (2πF) U _m−1 (cos (2
πF)) − U _m−2 (cos (2πF)) where m = 2, 3,.

The feature in Equation 22 is that the frequency response P
_c (F) includes sin (2πF). Therefore, substituting the function S _c (F ₁ , F ₂ ) for cos (2πF) in equation 22 gives
In addition, a variable conversion is performed in which the function S _s (F ₁ , F ₂ ) is substituted for sin (2πF) in Expression 22. Here, the function S _c (F ₁ , F ₂ ) and the function S _s (F ₁ , F ₂ ) are respectively represented by D _TC (F ₁ , F ₂ ) of _Expression 26 and Expression 2
Four appropriate functions cos (2ππ) approximating _DTS (F ₁ , F ₂ )
F ₁ ), sin (2πF ₁ ), cos (2πF ₂ ), and sin (2πF ₂ ), which can be expressed by the following equation.

[0089]

S _c (F ₁ , F ₂ ) ≒ D _TC (F ₁ , F ₂ ) = cos [2πF _{2
/ {(f s / f c} ) F 1 +1}]

S _s (F ₁ , F ₂ ) ≒ D _TS (F ₁ , F ₂ ) = sin [2πF _{2
/ {(f s / f c} ) F 1 +1}]

Here, D in Expressions 26 and 27
_TC (F ₁ , F ₂ ) and D _TS (F ₁ , F ₂ ) are given by
F _2shift = 0 and F _2shift = 0.25, and can be obtained by approximation using the above-described methods (procedure SS1) to (procedure SS4).

The frequency response G _c (F ₁ , F ₂ ) of the two-dimensional digital filter obtained as described above can be expressed by the following equation. The impulse response is calculated using the frequency response G _c (F ₁ , F ₂ ) of the following equation 28 in the same manner as in the first embodiment, and the weight coefficient of the DBF circuit 8-b is obtained.

[0092]

[Equation 28]

As described above, since the variable conversion is performed using Equation 8, it is possible to lower the side lobe level in a specific direction without using the conventional optimization method.

For wide-angle beam forming, the range of the beam forming direction is set to θ ₁ ≦ θ ≦ θ _2, and the range of the pass band of the one-dimensional complex coefficient band-pass digital filter is expressed by the following equation.
The frequency range represented by 1 may be set to F ₀₁ ≦ F ≦ F ₀₂ . Therefore, in the first embodiment, a wide-angle beam can be formed, and the side lobe level in a specific direction can be reduced.

In the above description, θ ₀ > θ ₂ , that is, the direction range in which the side lobe level is desired to be reduced is smaller than θ _0. However, when θ ₀ <θ ₂ and the direction in which the side lobe level is required to be lowered. May be provided in plurality. That is, in the first embodiment, at least one direction in which the side lobe level should be reduced can be set, and the direction can be set at an arbitrary position.

In the first embodiment described above, Equation 8 or Equation 9
, The straight line corresponding to the angle of arrival θ on the frequency-angle plane is converted to a predetermined straight line represented by Expression 6 on the two-dimensional frequency plane. Therefore, on the frequency-angle plane, the equal-amplitude line of the DBF circuit 8-b as a two-dimensional digital filter is parallel to the frequency axis. Therefore, the directivity characteristic of the DBF circuit 8-b obtained by such a variable conversion depends only on the angle of arrival θ, and can be made independent of the frequency. In other words, in the first embodiment, variable conversion is performed on the FIR narrow-band low-pass digital filter so that the directional characteristics of the DBF circuit 8-b do not depend on frequency.

As described in detail above, in the present embodiment, an arbitrary one-dimensional frequency is expressed by the following equation (6) on a two-dimensional frequency plane.
By performing variable conversion on the frequency response of the one-dimensional linear phase FIR type digital filter so as to be converted on a predetermined straight line represented by
Calculating the frequency response of the two-dimensional digital filter with respect to the normalized time frequency F ₁ and the normalized spatial frequency F _2, and calculating the impulse response of the two-dimensional digital filter by performing an inverse Fourier transform on the calculated frequency response;
The weighting factors wk _{, s} of the FIR digital filters 9-k are calculated based on the calculated impulse responses so that the two-dimensional digital filter has an amplitude in a desired passband. Thereby, the DBF of the first embodiment is
The circuit 8-b can form a wide-angle and wide-angle beam, and can reduce a side lobe level in a predetermined direction to a side lobe level in another region.

Next, the DBF circuit 8-
The result of the simulation for b will be described. FIG. 6 is a diagram illustrating a DB when a broadband beam is formed in the θ = 56 ° direction using the DBF circuit 8-b of the first embodiment.
9 is a graph showing the directional characteristics of the F circuit 8-b. The graph of FIG. _{6, f = 0.95f c, f} = f c and f = 1.05
for three frequencies f _c, it shows the relative power for the angle theta. As is clear from FIG. 6, the directional characteristics at the above three frequencies are almost the same. That is, from the graph of FIG. 6, the DBF circuit 8 of the first embodiment
-B is found to be capable of forming a beam which does not depend on a relatively wide band of frequencies of ± 5% with respect to the center frequency f = f _c.

FIG. 7 shows a case in which a broadband beam is formed in the direction of θ = 60 ° using the DBF circuit 8-b of the first embodiment, and the side beam in the direction range of −50 ° ≦ θ ≦ −20 °. It is a graph which shows the directional characteristic when the lobe level is lower than the side lobe level of another area by about 20 dB.
In FIG. 7, similarly to FIG. 6, f = 0.95f _c, f
= The three frequencies f _c and f = 1.05f _c, shows the relative power for the angle theta. As is clear from FIG. 7, by using the DBF circuit 8-b of the first embodiment, a broadband beam can be formed in the direction of θ = 60 °, and −50 ° ≦ θ ≦ −20 ° It can be seen that the side lobe level in the area can be set to be about 20 dB lower than the side lobe level in other areas.

In the reception signal processing apparatus configured as described above, the FIR digital filter coefficient calculator 13
Calculates the frequency response G (F ₁ , F ₂ ) of the two-dimensional digital filter by frequency-converting the FIR type low-pass digital filter based on the arrival angle θ which is the direction in which a beam should be formed. , The frequency response G
An impulse response g (m ₁ , m ₂ ) is calculated by performing an inverse discrete Fourier transform on (F ₁ , F ₂ ), and a weighting factor w _k, based on the impulse response g (m ₁ , m ₂ ) _. Calculate _s . Then, in the DBF circuit 8-b, each of the FIR digital filters 9-k receives the input digital in-phase / quadrature signal x based on the weight coefficient w _{k, s.
k} (m) is digitally filtered, and a filtered digital in-phase / quadrature signal x _k ′ (m) is output. The adder 12 outputs a plurality of M filtered digital in-phase / quadrature signals x _k '
(M) is added and output. Thereby, DBF circuit 8-b, so operates as a two-dimensional digital filter having the above-mentioned amplitude characteristic _{D (F 1, F 2)} , when the angle of arrival of the desired signal θ is given, broadband in that direction And a digital beam signal y that is a wideband desired signal corresponding to the wideband beam.
_b (m) is output.

Further, the processing of the reception signal detection circuit 10 using a known MUSIC algorithm will be described. The reception signal detection circuit 10 is, for example, a CPU,
It is composed of a microprocessor or a digital signal processor (DSP) having a ROM and a RAM. First,
An FF is applied to the digital beam signal y _b (k) (b = 0, 1,..., B-1) output from each DBF circuit 8-b.
By performing a T operation, a normalized frequency j-th narrowband component signal matrix (vector) obtained from the k-th snapshot data of the digital beam signal y _b (k) is represented by Y
_k (f _j ). This snapshot data is
For example, data of the data of T ₀ seconds obtained by K divided. Then, the desired signal frequency band is changed from f _j1 to f _j2.
Then, the covariance matrix R in the frequency domain can be expressed by the following equation.

[0102]

(Equation 29)

Here, † indicates conjugate transpose, and T (f _j )
Is a transformation matrix of the array response that can be expressed by the following equation.

[0104]

T (f _j ) A (f _j ) = A (f ₀ )

Here, A (f ₀ ) and A (f _j ) are array responses at frequencies f ₀ and f _j , respectively. f ₀ is referred to as the focusing frequency. In the present embodiment, the array response refers to the antenna element 1
-N is a response determined at an arrangement interval of n. The transformation matrix T (f _j ) can be approximately expressed by the following equation.

[0106]

(Equation 31)

Here, a _l β (f _j ) (l = 0, 1,...,
B-1) is the l-th sensor response (in this embodiment, the antenna response) to the frequency f _j of the signal arriving from the direction β. In the present embodiment, predetermined DB (B <M) digital beam signals are generated by each DBF circuit 8-b. Since the beam space processing reduces the size of the observation space by spatial filtering, the calculation cost can be reduced as a result.
The detection performance does not deteriorate and is often improved. The physical model for the new method of the present invention can be formulated as follows. Assuming that the antenna element 1-0 is a reference antenna element, the beam b (b = 0, 1,..., B−
The output of the beam space from 1), that is, the spectrum Y _{b of the} digital beam signal y _b (m) after the Fourier transform.
(F) can be represented by the following equation in the frequency domain.

[0108]

(Equation 32)

Here, X (f) is the output spectrum of the received signal from antenna element 1-0, and H _i (f) (i
= 0, 1,..., M−1) is the i-th antenna element 1−
9 is a transfer function of an FIR digital filter for i. The relationship between spectrum Y _b (f) and spectrum Y _{b + 1} (f) can be easily expressed by the following equation.

[0110]

(33) Y _{b + 1} (f) = Y _b (f) exp ｛−j2πf
dsin (θ) / c｝

The array response of the beam space with reduced order, ie the digital beam signal y _b (m)
Can be expressed by the following equation.

[0112]

[Y ₀ (f), Y ₁ (f), ..., Y _B-1 (f)] ^T = [1 exp {-j2πfdsin (θ) / c} ... exp {-j2π (B-1 ) fdsin (θ) / c] ^T Y (f)

Here, Y (f) is the Fourier transform of the output spectrum of the beam space when b = 0, that is, the digital beam signal y ₀ (m).

In the reception signal detection circuit 10,
After calculating the covariance matrix R of the spatial spectrum using
In order to calculate a spatial spectrum showing a peak in the direction of arrival of a received signal, an MU proposed in prior art document 2 has been proposed.
The SIC algorithm is used. The MUSIC algorithm is based on the calculation of a noise subspace that collects eigenvectors corresponding to the smallest eigenvalue of the covariance matrix R of the spatial spectrum. These eigenvectors correspond to the vectors v _{Ns + 1} , v _{Ns + 2} ,..., V _B. Where Ns
(<B) is the size of the signal subspace, which is usually the number of incoming signals in the case of narrowband signals, while it is a larger number in the case of wideband signals than can be found by trial, eg Starting from 1 and increasing until the best spatial spectrum is obtained. The above vectors v _{Ns + 1} , v _{Ns + 2} ,..., V _B satisfy the following eigen decomposition equation.

[0115]

[Number 35] _{Rv i = R n v i λ} i, i = Ns + 1, Ns +
2,…, B

Here, R _n is a spatial covariance matrix that is averaged and converged at a frequency that depends only on the internal noise when there is no effect of the external signal, and λ _i is a generalization considering the existence of R _n . Eigenvalue. Here, R _n can be measured in advance during the calibration process of the array antenna 100 when no signal is incident. If only white noise is handled, R _n is close to a constant multiple of the unit matrix. If Rn is a constant multiple of the unit matrix, the calibration process can be simplified. However, according to the inventor's experience, calculating including the noise correlation matrix,
It has been found that calculating the generalized eigenvalues can provide improved resolution. In the present embodiment, the noise subspace V _n can be expressed by the following equation.

[0117]

V _n = [v _{Ns + 1} , v _{Ns + 2} ,..., V _B ]

Here, the beam space response a (θ) in Equation 34 can be expressed by the following equation.

[0119]

A (θ) = [1 exp {−j2πf ₀ dsin (θ) / c}
… Exp {−j2π (B-1) f ₀ dsin (θ) / c] ^T

Here, T represents transposition of a matrix, d is an interval between antenna elements, and f ₀ is a focusing frequency. Then, the spatial spectrum S (θ) by the method of the MUSIC algorithm can be expressed by the following equation.

[0121]

S (θ) = {(a (θ)) † · a (θ)} / {(a
(θ)) † ・ Vn ・ (Vn) † ・ a (θ)}

Here, S (θ) is a positive real function related to θ having a maximum value at a correct arrival angle θ. Accordingly, the reception signal detection circuit 10 calculates θ, which gives the maximum value of the function S (θ), as the arrival angle θ by calculating Expression 38, and converts the arrival angle signal indicating the arrival angle θ into the FIR digital filter coefficient. The signal is output to the arithmetic unit 13 and a signal having a peak of the function S (θ) is output to the demodulation circuit 11 as a reception baseband signal.

Next, it will be described below that not only the signal-to-interference signal power ratio but also the signal-to-noise ratio SNR are improved.

<Improvement of Signal-to-Interference Signal Power Ratio> From the characteristics of the wideband digital beam forming circuit shown in FIG. 8, the interference wave is far away from the current beam direction (the separation distance is larger than the width of the main lobe). In such a case, it will be easily understood that the reduction of the interference wave power depends on the band rejection capability of the original one-dimensional zero-phase FIR type low-pass digital filter. In the special case illustrated in FIG. 8, a reduction of the interference of at least 20 dB is seen.

<Contribution of Noise in Output of Broadband DBF Circuit> The noises having no correlation with each other in the plurality of N antenna elements 1-n spatially disperse the power σ ² over the effective bandwidth which is the same as the signal bandwidth. and assuming to have the time, the output noise power P _{nois e} in the digital beam DBF circuit 8-b can be expressed by the following equation.

[0126]

[Equation 39]

<Contribution of Received Signal at Output of Broadband DBF Circuit> Here, the received signal at reference antenna element 1-0 has a constant power spectral density (PSD) P _S (F ₁ ).
The output power spectral density P _y (F ₁ ) for one digital beam can be expressed by the following equation.

[0128]

(Equation 40)

Here, F ₂ θ indicates the spatial frequency corresponding to the angle of arrival θ of the received signal. Further, the integral term in Equation 40 can be expressed by the following equation.

[0130]

[Equation 41]

[0131] Here, when the B _w and the bandwidth of the received signal, the [Delta] F ₂ theta can be expressed by the following equation.

[0132]

ΔF ₂ θ ≒ B _W ｛dsin (θ) / c｝ <1

Therefore, the signal-to-noise ratio SNR can be increased to (ΔF ₂ θ) ⁻¹ by using the wideband DBF circuit 8-b.

In the reception signal processing apparatus of the first embodiment, the FIR digital filter coefficient calculator 13
Calculates the frequency response G (F ₁ , F ₂ ) without using the optimization method or solving the simultaneous equations, and performs the inverse Fourier transform on the frequency response G (F ₁ , F ₂ ) to obtain the weighting factor. Since w _{k, s} is calculated, the weight coefficient w _{k, s} can be calculated at high speed.

In the first embodiment described above, the DBF circuit 8-b includes the FIR digital filter coefficient calculator 1
The M FIR digital filters 9-k, which filter and output the digital in-phase / quadrature signal x _k (m) input based on the weighting factor w _{k, s} calculated by 3 and the adder 10, The amplitude characteristic D (F ₁ ,
It operates as a two-dimensional digital filter having F ₂ ). Thereby, the DBF circuit 8-b can form a broadband beam in the direction of the arrival angle θ of the desired signal,
It is possible to output a digital beam signal y _b (m), which is a wideband desired signal corresponding to the wideband beam.

The reception signal processing apparatus of the first embodiment includes the above-mentioned DBF circuit 8-b and the FIR digital filter coefficient calculator 13, so that when transmitting information using a carrier wave, A wide-band beam can be formed in the arrival direction of the desired signal, a wide-band desired signal corresponding to the wide-band beam is received, and a desired digital beam signal y _b (m) as a baseband signal is output. Thereafter, based on this, the arrival angle θ can be detected by the reception signal detection circuit 10 and the reception baseband signal can be detected.

<< Examples of the First Embodiment >>
In order to verify the operation of the reception signal processing device configured as described above, an operation simulation by a computer was performed. Hereinafter, the results will be described. In the simulation, 26 only contributes to phase distortion.
It is assumed that the linear array antenna 100 is composed of a plurality of isotropic antenna elements 1-n at equal intervals. It is assumed that three plane wave signals are incident at azimuth angles of -44 °, 42 °, and 45 °, respectively. Further, the amplitude of the wideband received signal is constant, the carrier frequency is 100 Hz, and the deviation of the frequency modulation of the received signal is 20 Hz from 90 Hz to 110 Hz.
Bandwidth. Furthermore, the sampling frequency is
Delay devices 91-1 to 91- (q-
Fs = 30 Hz, which is equal to 1) and the FFT frequency. Note that these frequency values themselves have no special meaning.

As the number K of snapshots and the size (size) of the FFT for calculating the covariance matrix R in the frequency domain, either 32 or 64 was selected. 11 if FFT size is 32 to cover signal bandwidth
Frequency bins from to are selected, and when the FFT size is 64, frequency bins from 22 to 44 are selected. Here, the frequency bin is such that 1 is −0.5 of the sampling frequency.
The double number 33 (65) is the frequency number representing 0.5 times the sampling frequency.

First, the digital IQ signal x _n (m) is directly input to the reception signal detection circuit 10 without using the DBF circuit 8-b, ie, the element space processing which is the method of the first conventional example. The result when detection was performed is shown. The initial angle in the coherent signal subspace method is 43.5.
実 行, where the signal-to-noise ratio SNR = 0
FIG. 9 shows spatial spectra in the case where no interference wave exists at the arrival angle θ = −44 ° in dB.
And 10. Here, RP is the number of trials and is obtained by randomly changing the seed of random numbers for simulating noise and the initial phase of a signal. 64-FFT indicates that the size (size) of the FFT for calculating the covariance matrix R in the frequency domain is 64. The number of noise eigenvectors contributing to the formation of the spatial spectrum is 24 and 2 for the case where there is no interference wave at −44 °, respectively.
It was one. As apparent from FIGS. 9 and 10, as expected, the spatial spectrum could be detected much more accurately when there was no interference wave at −44 °. If an interference wave at −44 ° is present, the signal subspace cannot be set any lower. Thus, the noise subspace is degraded by increasing signal contributions.

The detection performance can be determined, for example, by the method of detecting BASS-ALE (for example, see Prior Art Document 6, “KMBuckley et al.
l., “Broad-Band Signal-Subspace Spatial-Spectrum
(BASS-ALE) Estimation ", IEEE Transactions on ASSP, V
ol. 36, No. 7, July 1988 ". ) Can be improved by expanding the observation space. However, this can involve a high computational load. Instead, the present embodiment uses the beam space method.

All beams are in the same direction (45 ° in this case)
The broadband beam is directed according to the method described above. DBF circuit 8-b of the first preferred embodiment
Has a stop band attenuation of about 20 dB as shown in FIG. Similarly, FIGS. 11 and 12 show the simulation results when the signal-to-noise ratio SNR is 0 dB and the number of beams B is 4 and 6 in the presence of the same three interference wave signals as described above. In this case, the number of noise eigenvectors was 1 and 2, respectively. As is apparent from FIGS. 11 and 12, the detection of the angles of arrival θ = 42 ° and 45 ° is accurate even though the size of the noise subspace is much lower, and the detection performance is particularly 6 °. In the case of a single beam, it turns out to be acceptable. FIG.
= 6, the same result is shown when the FFT size is changed from 64 to 32. Again, the performance is F
Although the FT size is slightly lower than 64,
The results are still useful.

FIGS. 14 and 15 show the power spectrum densities of the interference wave and the noise before and after the beam space processing, respectively. The noise power is attenuated by about 10 dB corresponding to the (ΔF ₂ θ) ⁻¹ coefficient described above, and the DBF
It is recognized that the interference wave is attenuated by about 20 dB due to the band rejection characteristic in the circuit 8-b.

<Calculation Cost> Next, the received signal detection circuit 1
An approximate calculation of the number of operations required for the element space processing (element space processing) at 0 and the signal processing according to the first embodiment of the present invention will be described.

In the element space processing, when the number of FFT snapshots each having a length of 2 ^m is K and the number of antenna elements in the array antenna 100 is N, the number of operations of the complex number operation is as follows. It is as follows. (A) NKm2m for FFT processing. (B) About N ³ for eigenvalue decomposition processing.

In the beam space detection process, N
Is replaced by B (in this embodiment, B≪N),
And the number of operations is the same as in the element space processing, except that predetermined processing must be added to the DBF circuit 8-b. The number of operations of the complex number operation is BM ² 2 ^m , where M is 1
This is the number of arrays in the set of subarrays (M = 17 in the simulation example). The FFT processing for obtaining a snapshot of the spatial spectrum data has the highest computational cost. In the simulation example, when the FFT sizes are 32, 64, 128, and 256, respectively, the element space as shown in FIG. Detection processing and beam space detection processing (N = 26, B = 6 and B
= 4). In this example, the calculation cost in the beam space detection processing is about 1/5 of the element space detection processing for the number of beams B = 4, and about 1/6 for the FFT size of 256.

As described above, by combining the sub-array signal selection processing, the wideband DBF processing, and the reception signal detection processing using the MUSIC algorithm, the calculation cost can be greatly reduced as compared with the conventional example. A received signal processing device having high resolution can be realized. In the present embodiment, in the reception signal detection processing using the MUSIC algorithm, a broadband D
By changing the FIR digital filter coefficient in the BF circuit 8-b, adjustment can be made so that the case of an unsteady environment can be handled. The advantage of performing beamspace processing is that the size of the observation space can be significantly reduced. As a result, high-speed processing and real-time application become possible. Even if the detection rate of the optimal angle of arrival detection in the element space processing decreases, the beam space detection processing operates normally by properly selecting the number M of arrays in a set of sub-arrays, and the execution speed is reduced. The degree of improvement in terms also compensates for this. According to the calculation simulation comparing the element space processing and the beam space beam forming processing, the performance of the arrival angle detection method of the present embodiment according to the present invention is smaller than the conventional example in the spatial spectrum detection characteristics and the calculation cost. It has been found that has been greatly improved.

Furthermore, if the wideband DBF circuit 8-b is properly selected in the proper direction and the characteristics of the spatial stopband, the spatial spectrum is greatly improved, and in some cases, two adjacent desired receptions are performed. When a signal, one interference wave signal, and white noise having a signal-to-noise ratio SNR = 0 dB are present, resolution of 2 ゜ or less can be obtained by performing signal processing by the reception signal processing device. Further, the calculation cost is greatly reduced depending on the size of the FFT used in the subspace processing of the interference wave signal, and the signal processing amount is only 1/5 as compared with the signal processing of the element space processing.

FIG. 17 is a block diagram showing a configuration of a received signal processing device according to a first modification of the first embodiment. A first modification of the first embodiment will be described with reference to FIG. In the first embodiment, the FIR digital filter coefficient calculator 13 calculates a weighting coefficient w _k, based on an arrival angle signal from the reception signal detection circuit 10 indicating an arrival angle θ which is a direction in which a beam should be formed _{. 0} , w _{k, 1} , ...,
w _{k, q} is output to the FIR digital filter 9-k. In the first modification of the first embodiment, an arrival angle detection circuit 14 is newly provided, and the FIR digital coefficient operation circuit 13 instead arrival angle signal from the received signal detection circuit 10, based on the arrival angle signal outputted from the arrival angle detector 14, a load factor _{w k, 0, w k,} 1, ..., w k, _q-calculated Output. Hereinafter, differences from the first embodiment will be described.

The angle-of-arrival detection circuit 14 comprises a DBF circuit 8
-B is a circuit for roughly detecting the direction of arrival in order to determine the beamforming direction of -b. Arrival angle detection circuit 14
, A digital in-phase / quadrature signal x _n (m) which is a complex number output from the low-pass digital filter 5-n is input. In DOA detection in arrival angle detecting circuit 14, the angle of arrival detection circuit 14 digital in-phase and quadrature signals x ₀ which is input to the _{(m), x 1 (m} ), ..., x N-1
(M) is subjected to discrete Fourier transform of N points in the spatial direction of the antenna element 1-n, and the angle of arrival is detected by obtaining the peak of the absolute value of the transform value. Here, for example, the calculated conversion value is compared with a predetermined threshold value, and an incident angle of a signal having a peak equal to or higher than the threshold value is detected as an arrival angle. In this example, the arrival direction is detected by executing the discrete Fourier transform while fixing the time m. However, the present invention is not limited to this. A time average value of the absolute value of the transform value obtained by the discrete Fourier transform with respect to is obtained, and the arrival direction may be detected from the peak. Further, the algorithm for detecting the angle of arrival may use the algorithm of the related art document 1 or the MUSIC algorithm of the related art document 2.
However, when using the MUSIC algorithm, it is necessary to narrow the signal band using a band-pass filter.

The arrival angle detection circuit 14 outputs an arrival angle signal indicating the detected arrival angle θ to the FIR digital filter coefficient calculator 13. The FIR type digital filter coefficient calculator 13 calculates the
Similar to the first embodiment, the load coefficients w _{k, 0} , w _{k, 1} ,.
By calculating _{k and q} , the calculated weighting factor is calculated by the DBF circuit 8-
Input to each FIR digital filter 9-k in b.

The first modified example of the first embodiment configured as described above has the same effects as the first embodiment. In the first embodiment, a coefficient is calculated in an FIR digital filter coefficient calculator 13 based on an angle-of-arrival signal obtained as a result of processing a received signal through a reception signal detection circuit 10. However, in the first modified example, the received signal is
An arrival angle signal is obtained in the arrival angle detection circuit 14 based on the digital in-phase / quadrature signal x _n (m) processed only through −n, and the coefficient is calculated in the FIR digital filter coefficient calculator 13 based on the signal. In order to calculate, the first embodiment is compared with the first modified example.
Is a real-time detection of the angle of arrival and the DBF
Subsequent processing becomes possible in the circuit 8-b.

FIG. 18 is a block diagram showing a configuration of a received signal processing circuit according to a second modification of the first embodiment. First
The second modification of the embodiment is an example in which the first modification of the first embodiment is further modified. In a first modification of the first embodiment, the digital in-phase / quadrature signal x ₀ filtered by the low-pass digital filter 5-n.
( _M ), x ₁ (m),..., X _N−1 (m) have been input to the angle-of-arrival detection circuit 14a, but in the second modification, the sub-array signal output from the sub-array signal selection circuit 7 SA
_{0, SA 1, ..., (} SA 0 in FIG. 18) for any pair of SA _B-1 is input to the arrival angle detector 14a. Other processes and circuit configurations are the same as those of the first modification of the first embodiment. Arrival angle detecting circuit 14a, the arrival angle by a the subarray signals SA _b inputted performs discrete Fourier transform of the point M in the space direction of the element of the antenna element 1-n, obtains the peak of the absolute value of the conversion value Is detected. Here, for example, the calculated conversion value is compared with a predetermined threshold value, and an incident angle of a signal having a peak equal to or higher than the threshold value is detected as an arrival angle. In this example, the arrival direction is detected by executing the discrete Fourier transform while fixing the time m. However, the present invention is not limited to this. A time average value of the absolute value of the transform value obtained by the discrete Fourier transform with respect to is obtained, and the arrival direction may be detected from the peak. Further, the algorithm for detecting the angle of arrival may use the algorithm of the related art document 1 or the MUSIC algorithm of the related art document 2. However, when the MUSIC algorithm is used, it is necessary to narrow the band using a band-pass filter.

In the sub-array signal selection circuit 7, the sub-array signals are selected so as not to overlap at all, and the sum of the total number of sub-array signals SA ₀ , SA ₁ ,..., SA _B-1 is smaller than N. Sub-array signals SA ₀ , S
A ₁ ,..., SA _B-1 at least one set
4a.

In a second modification of the first embodiment,
The signal input to the angle-of-arrival detection circuit 14a is a set of the signals selected and output by the sub-array signal selection circuit 7, and in other words, the angle-of-arrival detection circuit 14a according to the second modified example. The number M of signals input to the
Since the number of signals input to the angle-of-arrival detection circuit 14 in the modified example of the present embodiment is smaller than the number N of signals input to the angle-of-arrival detection circuit 14, the discrete Fourier transform executed in the angle-of-arrival detection circuit 14a is reduced from N-point discrete Fourier transform to M-point discrete Fourier transform (N> M), the resolution of the second modification is lower than that of the first modification, but the calculation cost in the arrival angle detection circuit 14a is also reduced.

<Second Embodiment> Next, a second embodiment will be described with reference to FIG. FIG. 19 is a block diagram illustrating a configuration of a reception signal processing device according to the second embodiment. In contrast to the single beam direction in the first embodiment, in the second embodiment, it is possible to use a multi-beam to direct the beam in a plurality of directions, and to exist around different directions. Provided is a reception signal processing device having a multi-beam space detection process for simultaneously detecting the arrival directions of a broadband interference wave and a non-interference wave with high resolution.

In the first embodiment, by processing a beam instead of an element signal, an interfering signal existing away from the main beam is removed and the size of the observation space is reduced, but the detection performance is not reduced. It was shown that the detection performance increased in some cases. The second embodiment includes significant improvements over the first embodiment, as well as savings in computational cost over prior art methods in prior art documents 1 and 3. Although the overall structure of the present embodiment will be described later in detail, a plurality u of received signal processing circuits 15-1, 1
,..., 15-u all operate in parallel, so that they operate at the same speed as the first embodiment having one directional direction. On the other hand, the required computational cost increases with the increase in hardware for detecting different desired directions of arrival. That is, u received signal processing circuits 15-1,
The computational cost required by the increase in the number of 15-2,..., 15-u increases.

The configuration and processing of the second embodiment will be described with reference to FIG. The reception signal processing device according to the present embodiment includes a plurality of u reception signal processing circuits 15-1 each including a plurality of B DBF circuits 8-b having the same configuration, a reception signal detection circuit 10a, and a demodulation circuit 11. , 15-2, ..., 15
-U, the received signal processing circuits 15-1, 15-
2, ..., 15-u is the received signal R ₀ when the directed respectively main beam in different directions, R _1, ..., perform beam space processing and outputs the R _u. Here, each DBF circuit 8-
b, the received signal detection circuit 10a and the demodulation circuit 11
Has the same processing and configuration as those used in the embodiment. Each of the received signal processing circuits 15-1, 15-2,.
5-u each of a plurality of B DBF circuits 8-0, 8-1,...
8- (B-1) includes sub-array signals SA ₀ , S
_{A 1, ..., SA B-} 1 is input, multiple u number of receiving signal processing circuit 15-1 and 15-2, ..., respectively each DBF circuits 8-b of 15-u, the incident angle theta _1, theta _2, ..., toward the theta _u to form a main beam, a plurality u number of receiving signal processing circuit 15
1, 15-2,..., 15-u
a and the demodulation circuit 11 incident angle θ _1, θ _2, ..., received signal by detecting a signal incident from the vicinity of θ _u R _1, R _2,
..., and outputs the R _u. Therefore, the received signal processing device is
Signals arriving from multiple directions can be simultaneously received, processed and detected.

In the second embodiment, each of the incident angles θ _{1 to} θ _u is determined when the number of main beams is small and the multi-beam does not cover the angle range of −90 ° to 90 °. 19, the arrival angle θ is detected based on the digital in-phase / quadrature signal x _n (m) input to the arrival direction detection circuit 14a, and the arrival angle signals DA1, DA2,. Is output to the FIR digital filter coefficient calculator 13a, and the weighting factors w _{k, s} (B × u) are calculated based on the angle-of-arrival signals DA1, DA2,..., DAv input to the FIR digital filter coefficient calculator 13a. Calculated and calculated load coefficient w _{k, s}
, A plurality of u reception signal processing circuits 15-1, 15-2,
, 15-u to each of a plurality of B DBF circuits 8-b, and a plurality of u received signal processing circuits 15-1, 15-2,
, 15-u in each DBF circuit 8-b, based on the input weighting factor w _{k, s} , each internal DB
The F circuit 8-b forms the main beam toward the incident angles θ ₁ , θ ₂ ,..., Θ _u , and a plurality of u reception signal processing circuits 15-1,
, 15-u detect signals in the directivity direction of the main beam formed in the plurality of B DBF circuits 8-b in each of them, and receive signals R ₁ , R ₂ ,.
And outputs the R _u. That is, the arrival angle detection circuit 14a and the FI
The R-type digital filter coefficient calculator 13a calculates the incident angles θ ₁ , θ of the main beams of the plurality of B DBF circuits 8-b inside the received signal processing circuits 15-1, 15-2,..., 15-u.
₂ ,…, greatly involved in determining θ _u

The reception signal processing circuits 15-1, 15-2,
.., 15-u is changed according to the width of the beam of the received signal, and is reduced as the beam width of the received signal is large. Specifically, the number u is at least two. If the width is small, it is necessary to install more. For example, the DBF of the broadband beam used in FIG.
If the beam width for 3 dB attenuation of the circuit 8-b is about 15 ° for a pointing angle of 42 ° (in fact, as the pointing angle increases, this beam width also increases), −90 ° to 90 °
In order to cover the angle range of °, for example, about 10 parallel received signal processing circuits 15 each performing the method of the known prior art document 1 for 10 different directional directions.
-U (that is, u = 10) is used.

Here, the arrival angle detection circuit 14a and the FIR digital filter coefficient calculator 13a will be described. The arrival angle detection circuit 14a is the same as that used in the two modified examples of the first embodiment, and the input digital in-phase / quadrature signals x ₀ (m), x ₁ (m),. _N-1 (m)
., DAv indicating the detected angles of arrival are input to the FIR digital filter coefficient calculator 13a. FIR digital filter coefficient calculator 13a also has the same function as two variants of the first embodiment, the input angle of arrival signals DA0, DA1, ..., on the basis of the DAV, load coefficient w _{k , s} to each of the received signal processing circuits 15-1, 15-2,
, 15-u DBF circuits 8-0, 8-1, ..., 8-
(B-1). Note that u ≦ v.

The weighting factor w _{k, s} of the plurality M of FIR digital filters 9-k of the plurality B of DBF circuits 8-b is such that the pattern of the antenna array is substantially independent of frequency in a desired bandwidth. It is set to be. 21 and 22 show two different high bandwidth DBFs.
4 shows the resulting array pattern for the circuit.
Here, FIG. 21 and FIG. 22 show 41 ° and −35, respectively.
° broadband beam pattern, with 15 antenna elements, 15 tap delays, f _s = 35 Hz, f _c =
100 Hz, attenuation less than 40 dB.

An advantage of the second embodiment according to the present invention is that a plurality u of received signal processing circuits 15-1, 15-2,.
The object of the present invention is to simultaneously detect at least one desired incoming signal with high resolution by directing the main beam in a plurality of u directions using 5-u. Further, by performing the beam space processing, in the detection of the multi-beam space processing according to the present invention, the arrival angle in the antenna array processing is more accurately detected as compared with the processing of the element space in the conventional example. Further, in the embodiment, a signal separated from the main beam is removed. Therefore, the present multi-beam space processing method operates only on signals collected around the focusing direction. This makes it possible to remove unnecessary interference waves and detect a desired signal with higher resolution.

Next, a first modification of the second embodiment will be described. FIG. 20 is a diagram showing a modification of the second embodiment. Hereinafter, only differences from the second embodiment will be described. The second embodiment and a modification of the second embodiment are similar to the first embodiment.
This is the same as the relationship between the first modification of the embodiment and the second modification of the first embodiment. That is, in the second embodiment, the digital in-phase / quadrature signals x ₀ (m), x ₁ (m), filtered by the low-pass digital filter 5-n,
..., but x _N-1 (m) has been input to the arrival angle detector 14a, in a variant, the subarray signals SA _0, SA ₁ output from the subarray signal selecting circuit 7, ..., of the SA _B-1 An arbitrary set (SA _{0 in} FIG. 20) is input to the arrival angle detection circuit 14a. Other processing and circuit configuration
This is the same as the second embodiment. The arrival angle detection circuit 14
a performs the same processing as that of the arrival angle detection circuit 14a of the second embodiment described above, and executes a discrete Fourier transform based on an arbitrary set of input sub-array signals (SA _{0 in} FIG. 20). To detect the angle of arrival. The signal input to the angle-of-arrival detection circuit 14a in FIG. 20 is a set of other sub-array signals SA ₀ , SA ₁ ,..., SA _B-1 or the sub-array signals SA ₀ , SA ₁ ,. , SA _B-1 may be a plurality of sets of signals, and may be a plurality of signals obtained by combining duplicate signals into one.

The effect of this modification is that, similarly to the second modification of the first embodiment, the signal input to the arrival angle detection circuit 14a is selected and output by the sub-array signal selection circuit 7. In other words, the number of signals M input to the angle-of-arrival detection circuit 14a in the modification is the number N of signals input to the angle-of-arrival detection circuit 14 in the second embodiment. Due to the further decrease, the discrete Fourier transform executed in the arrival angle detection circuit 14a is reduced from the N-point discrete Fourier transform to the M-point discrete Fourier transform (N <M), and the modified example is described in the second embodiment. Although the resolution is lower than that of
The computational cost at a is to be reduced as well.

Next, a second modification of the second embodiment will be described. In the second embodiment, the multi-beam is −90 ° to
Although the mode is used when not covering the angle region of 90 °, when the multi-beam covers -90 ° to 90 °, the main beam is directed as in the second modification. The incident angles θ _{1 to} θ _u to be used may be set in advance so as to cover all directions. In this case, the arrival angle detection circuit 14ab and the FIR digital filter coefficient operation unit 13a are not used as compared with the second embodiment of FIG. Further, in this case, if the output signal power of DBF circuit 8-b is lower than a predetermined level, the reception signal detection processing need not be performed in that direction. Therefore, the processing in the angle-of-arrival detection circuit 14ab and the FIR digital filter coefficient operation unit 13a is not required, which is advantageous from the viewpoint of reduction in calculation cost.

<< Example of Second Embodiment >><ComputationCost> Next, the number of calculations required for the conventional element space processing in the received signal detection circuit 10a and the signal processing in the second embodiment according to the present invention are described. Estimation is shown below.

In the conventional element space processing, a complex number operation is performed when the number of FFT snapshots each having a length of 2 ^m is K and the number of antenna elements in the array antenna 100 is N. number is NKm2 ^m + N ^3.

In the multi-beam space detection processing, except that N is replaced with B and that predetermined processing must be added to the DBF circuit 8-b,
The number of calculations is the same as in the element space processing of the conventional example. The number of operations of the complex number operation is u (BKm2 ^m +
B ³ + BM ² 2 ^m ), where M is the number of arrays in a set of subarrays (M = 1 in the simulation example)
5). The FFT processing for obtaining the snapshot of the spatial spectrum data has the highest calculation cost. In the simulation example, when the FFT sizes (sizes) are 32, 64, 128 and 256, respectively, FIG.
, The element space detection processing and the multi-beam space detection processing (N = 24, B = 10, B =
6, B = 4). In the method of the present embodiment, the calculation cost is saved for B = 6 and B = 4.

<Computation Simulation> In order to verify the operation relating to the received signal processing device of the second embodiment, the present inventor performed an operation simulation using a computer as in the first embodiment. Hereinafter, the results will be described. First, an experimental result using the element space method, which is a conventional example, is shown. In the simulation, it is assumed that the linear array antenna 100 is arranged at equal intervals and includes 24 isotropic antenna elements 1-n that contribute only to phase distortion. It is assumed that six mutually coherent signals from far away are incident on each other three waves from near -35 ° and 41 °. Where u = 2,
The main beam is directed at 41 ° and -35 °, respectively, and the broadband D
The side lobe level of the BF circuit is about 40 dB. The number M of arrays in one set of sub-arrays is 15, and the number of tap delays is also 15. The three signals are azimuth -38 °,-
It is assumed that light is incident from 35 ° and −32 °. The carrier frequency is 100 Hz, and the shift frequency of the frequency modulation of the received signal is 30 Hz from 85 Hz to 115 Hz. Further, the sampling frequency is determined by delay units 91-0 to 91- (q-1), which are tap delay lines, and F
F _s = 30 Hz is set equal to the frequency of the FT. Note that the present invention is not limited to these frequency values themselves.

The number of snapshots K and the size of the FFT for calculating the covariance matrix R in the frequency domain are as follows:
64 was selected. Frequency bins from 18 to 48 were selected to cover the bandwidth for an FFT size of 64.

First, the detection processing in the element space processing of the conventional example is executed by using the array antenna 100 composed of 24 antenna elements 1-n and setting the initial angle in the coherent signal subspace method to -35 °. ,
The results of the spatial spectrum are shown in FIGS. 24 and 25 as an overall view and an enlarged view around −35 °, respectively. FIG. 25 is an enlarged view of FIG. Here, the number of trials is six, and the seeds of random numbers for simulating noise and the initial phase of the signal are randomly changed. Signal to noise ratio SN
The set value of R is 0 dB. 26 and 27 are diagrams showing a case where the initial angle in the coherent signal subspace method is set to 41 ° in a similar experiment. FIG.
FIG. 27 is an enlarged view of the vicinity of 41 ° in FIG. 26. The number of corresponding noise eigenvectors that form the spatial spectrum was chosen to be 17 (in computer simulation this is the "best" spatial spectrum, i.e. the sharpest peak).

Next, the same experiment as described above is performed using the beam space processing of the present invention. Figures 28 and 29 show -35
30 and 31 show the experimental results by the beam space processing directed to 41 °. FIG. 30 and FIG. 31 show the experimental results by the beam space processing directed to 41 °. All beams belonging to the same set are oriented at -35 ° or 41 °, as in the experiment in the element space processing described above. FIG. 28 and FIG.
9 and FIGS. 30 and 31 show the simulation results under the condition that the signal-to-noise ratio SNR = 0 dB and the presence of six identical interference signals, as in the above-described experiment. 29 and 31 are enlarged views of FIGS. 28 and 30, respectively. Here, the number of noise eigenvectors is 4 in each case.

In the experimental results of the beam space processing, although the dimension of the noise subspace is considerably small, six signals are correctly detected with respect to the detection of the angle of arrival. On the other hand, the experimental results obtained by the conventional element space processing are always inaccurate as shown in FIGS. As shown in FIG. 23, B =
The calculation cost of 10 is the FF used for the simulation.
When the size of T is 64, it is slightly higher. If the number of beams used is reduced, the computational cost can be reduced, but the number of signals that can be detected is also reduced by reducing the size of the correlation matrix.

[0174]

[0175]

[0176]

[0177]

[0178]

[0179]

As described above in detail, according to the reception signal processing apparatus of the first aspect of the present invention, an array antenna composed of a plurality of N antenna elements arranged in a straight line at a predetermined element interval is provided. Conversion means for A / D converting the received plurality of N received signals into a plurality of N digital received signals and outputting the digital received signals, and detecting an arrival angle of a desired received signal and detecting an arrival angle signal indicating the arrival angle And a detecting means for outputting a plurality of M digital signals which are different from each other among the sub-array signals from a plurality of N digital received signals output from the converting means. Selecting means for selecting and outputting a plurality of B sets of sub-array signals comprising a set of digital reception signals; a plurality of M FIR digital filters each having a predetermined filter coefficient; Each comprise a plurality of M FIR type adding a plurality of M signal output from the digital filter to an adder which outputs have the same transfer function with each other,
Based on a plurality of B sets of sub-array signals output from the selection means, to direct the main beam of the array antenna to the arrival direction of the desired reception signal and to zero the reception signal in the arrival direction of the interference wave signal, A plurality of B forming means for forming and outputting digital beam signals, respectively, wherein the detecting means is configured to output a desired reception signal based on the plurality of B digital signals output from the plurality of B forming means. While detecting the angle of arrival of the desired received signal to detect and output the angle of arrival signal indicating the angle of arrival, and the received signal processing device, based on the angle of arrival signal output from the detection means, A plurality of M FIR type digital signals of each of the forming means for directing a main beam of the array antenna in a direction of arrival of a desired received signal and making a received signal in the direction of arrival of an interference wave signal zero. Further comprising calculating means for setting control by calculating the filter coefficients of the filter. That is, by combining the sub-array signal selection processing, the wideband DBF processing, the arrival angle detection processing, and the reception signal detection processing, the calculation cost can be significantly reduced as compared with the conventional example, and the reception signal having high resolution A processing device can be realized. Here, it is possible to execute the arrival angle detection processing in advance to detect the arrival angle in more real time, and then to execute processing including a filter coefficient calculation processing and the like. An advantage of performing the beam space processing of the present invention is that the size of the observation space can be significantly reduced. As a result, high-speed processing and real-time application become possible. In addition, in the conventional element space processing, the detection of the angle of arrival is reduced, whereas in the present invention, the reception signal detection processing operates normally by appropriately selecting the digital beam bandwidth, and the execution speed is reduced. The degree of improvement in terms also compensates for this.

[0180]

[0181]

[0182]

[0183]

[0184]

[0185]

[0186]

[0187]

Further, according to the reception signal processing apparatus of the second aspect of the present invention, a plurality of N antenna elements received by an array antenna composed of a plurality of N antenna elements juxtaposed at a predetermined element interval on a straight line. Conversion means for A / D converting the received signal into a plurality of N digital received signals and outputting the digital signal, and an angle of arrival for detecting the angle of arrival of the desired received signal and detecting and outputting the angle of arrival signal indicating the angle of arrival In a reception signal processing device comprising a detection unit, the reception signal processing device,
Selecting means for selecting from a plurality of N digital received signals output from the converting means and selecting and outputting a plurality of B sets of sub-array signals composed of a set of a plurality of M digital received signals different from each other among the sub-array signals; A plurality of u signal processing means, wherein each of the signal processing means includes a plurality of M FIR digital filters each having a predetermined filter coefficient, and a plurality of M FIR digital filters output from the plurality of M FIR digital filters. And an adder for adding and outputting the signals of the same. Each of the same signal processing means has the same transfer function. Based on a plurality of B sets of sub-array signals output from the selection means, The main beam of the array antenna is directed in the same desired direction of arrival of the received signal by the same signal processing means and the direction of arrival of the interference wave signal. A plurality of B forming means for respectively forming and outputting digital beam signals so as to make the received signal zero, and a desired B based on the plurality of B digital signals output from the plurality of B forming means. Reception signal detection means for detecting a reception signal, wherein the angle-of-arrival detection means comprises a plurality of N digital reception signals output from the conversion means or a plurality of B sets of sub-array signals output from the selection means. Based on one of the sets, an arrival angle of at least one desired reception signal is detected, an arrival angle signal indicating the arrival angle is detected and output, and the reception signal processing device outputs On the basis of the output angle-of-arrival signal, the plurality of u main beams of the array antenna are directed to the direction of arrival of at least one received signal and the received signal in the direction of arrival of the interference wave signal is set to zero. Such further comprising a plurality of M calculation means for calculating a filter coefficient setting controls the FIR digital filter of the respective forming unit.
That is, by using a plurality of u signal processing means, u signals arriving from relatively distant directions can be simultaneously detected, and a received signal processing device having high resolution can be realized. According to a calculation simulation comparing the element space processing of the conventional example and the multi-beam space processing of the present invention, the performance of the arrival angle detection method of the present embodiment according to the present invention is smaller than that of the conventional example in spatial spectrum detection. It was found that the characteristics and the calculation cost were greatly improved. In addition, the signals separated from the respective main beams are removed. Therefore, it acts only on signals gathering around the beam direction. This makes it possible to remove unnecessary interference waves and detect a desired signal with higher resolution.

[0189]

[0190]

[0191]

[0192]

[0193]

[0194]

[0195]

[0196]

[0197]

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of a received signal processing device according to a first embodiment of the present invention.

FIG. 2 is a block diagram showing a configuration of an FIR digital filter of the reception signal processing device of FIG.

3A is a diagram illustrating a beam in a desired direction on a frequency-angle plane for explaining the operation principle of the DBF circuit 8-b in FIG. 1, and FIG. 3B is a diagram illustrating the beam in FIG. FIG. 7 is a diagram illustrating a desired passband when a beam in a desired direction on a frequency-angle plane is converted into a two-dimensional frequency plane.

FIG. 4 shows a two-dimensional spectrum Sp on a two-dimensional frequency plane.
It is the figure which showed typically.

FIG. 5A is a diagram illustrating a beam in a desired direction and a region in which a side lobe level is to be reduced in a frequency-angle plane, for explaining the operation principle of the DBF circuit 8-b according to the first embodiment; (B) is to lower the desired pass band and side lobe level when the beam in the desired direction and the area to lower the side lobe level in the frequency-angle plane of (a) are converted into a two-dimensional frequency plane. It is a figure which shows an area | region.

FIG. 6 shows an incident angle θ = 56 ° in the first embodiment.
7 is a graph showing directivity characteristics when a broadband beam is formed in the direction of.

FIG. 7 shows an incident angle θ = 60 ° in the first embodiment.
A broadband beam is formed in the direction of
It is a graph which shows the directivity characteristic when the side lobe level in the direction of -20 degrees is lowered.

FIG. 8 shows a broadband DB in an example of the first embodiment.
It is a graph which shows F characteristic.

FIG. 9 is a graph showing a spatial spectrum characteristic of two signals in a conventional example using the conventional element space processing.

FIG. 10 is a graph showing spatial spectrum characteristics of three signals in a conventional example using the conventional element space processing.

FIG. 11 is a graph showing first spatial spectral characteristics of three signals in the example of the first embodiment.

FIG. 12 is a graph showing second spatial spectral characteristics of three signals in the example of the first embodiment.

FIG. 13 is a graph showing third spatial spectral characteristics of three signals in the example of the first embodiment.

FIG. 14 is a graph showing a relative power spectral density characteristic of a DBF circuit input signal in the example of the first embodiment.

FIG. 15 is a graph showing a relative power spectral density characteristic of a DBF circuit output signal in the example of the first embodiment.

FIG. 16 is a graph showing a relative comparison between the signal processing operation amount required for the conventional element space processing and the signal processing operation amount according to the first embodiment of the present invention.

FIG. 17 is a block diagram illustrating a configuration of a reception signal processing device according to a first modification of the first embodiment of the present invention.

FIG. 18 is a block diagram illustrating a configuration of a reception signal processing device according to a second modification of the first embodiment according to the present invention.

FIG. 19 is a block diagram illustrating a configuration of a received signal processing device according to a second embodiment of the present invention.

FIG. 20 is a block diagram illustrating a configuration of a reception signal processing device according to a first modification of the second embodiment of the present invention.

FIG. 21 shows an incident angle θ = 41 in the second embodiment.
9 is a graph showing directional characteristics when a broadband beam is formed in the direction of °.

FIG. 22 shows an incident angle θ = −3 in the second embodiment.
It is a graph which shows the directivity characteristic at the time of forming a broadband beam in the direction of 5 degrees.

FIG. 23 is a graph showing a relative comparison between a signal processing operation amount required for conventional element space processing and a signal processing operation amount according to the second embodiment of the present invention.

FIG. 24 is a graph showing spatial spectrum characteristics of six signals when an initial angle is set in a direction of an incident angle θ = −35 ° in a conventional example using the conventional element space processing.

FIG. 25 shows a spatial spectrum characteristic of six signals when an initial angle is set in a direction of an incident angle θ = −35 ° in a conventional example using the conventional element space processing.
FIG. 4 is an enlarged view of FIG.

FIG. 26 is a graph showing spatial spectral characteristics of six signals when an initial angle is set in a direction of an incident angle θ = 41 ° in a conventional example using the conventional element space processing.

FIG. 27 is a diagram showing spatial spectrum characteristics of six signals when an initial angle is set in a direction of an incident angle θ = 41 ° in a conventional example using the conventional element space processing.
FIG.

FIG. 28 is a graph showing spatial spectral characteristics of three signals out of six signals when a broadband beam is formed in the direction of the incident angle θ = −35 ° in the example of the second embodiment.

FIG. 29 is a diagram showing the spatial spectrum characteristics of three of the six signals when a wideband beam is formed in the direction of the incident angle θ = −35 ° in the example of the second embodiment.
FIG. 8 is an enlarged view of FIG.

FIG. 30 is a diagram showing a case where a broadband beam is formed in the direction of the incident angle θ = 41 ° in the example of the second embodiment.
5 is a graph showing a spatial spectrum characteristic of a signal.

FIG. 31 shows an example in which a broadband beam is formed in the direction of the incident angle θ = 41 ° in the example of the second embodiment.
FIG. 30 showing the spatial spectrum characteristics of three of the signals
FIG.

[Explanation of symbols]

1-0 to 1- (N-1): antenna element; 2-0 to 2- (N-1): receiver; 3-0 to 3- (N-1): A / D converter; 0 to 4- (N-1): mixer; 5-0 to 5- (N-1): low-pass digital filter; 6-0 to 6- (N-1): distributor; 7: sub-array signal Selection circuit, 8-0 to 8- (B-1) ... DBF circuit, 9-0 to 9- (M-1) ... FIR type digital filter, 10, 10a ... reception signal detection circuit, 11 ... demodulation circuit, 12 .., Adders, 13, 13a, FIR digital filter coefficient calculators, 14, 14a, 14ab, arrival angle detection circuits, 15-1 to 15-u, reception signal detection circuits, 60, digital local oscillators, 91-1 to 91-1 91- (q-1): delay device, 92-1 to 92-q: multiplier, 93-1 Itaru 93- (q-1) ... adder, 100 ... array antenna.

Continuation of the front page (72) Inventor Yoshio Karasawa Kyoto, Soraku-gun, Seika-cho, 5F, Inani, 5F, Sanpiraya Inside the ATR Environmental Adaptive Communications Research Laboratories (56) References JP-A-3-105276 ( JP, A) Sekiguchi et al., “Design of Wideband Digital Beamformer by Combining Spectral Transformation and Window Function Method”, IEICE Technical Report, IE96 1-7 (IEICE Technical Report Vol. 96, No. 2). ), Pp. 9-16, April 1996. O. Schmidt, "Multiple Emitter Location and Signal Parameter Estimation", IEEE Trans. Antenas Propagat. , Vol. AP-34, no. 3, pp. 276-280, 1986. Kloche-Djedid et al., "High Resolution DOA Estimation Algorithm by Beam Space Processing for Broadband Sensor Array", IEICE Technical Report, A.P9666-80 (IEICE Technical Report Vol. 96, No. .3773); 51-55, November 21, 1996. Sekiguchi et al., "Beamspace Adaptive Array for Broadband Signals," IEICE Technical Report, A. P95 92-97 (IEICE Technical Report Vol. 95, No. 465); 1-8, January 1996. (58) Field surveyed (Int. Cl. ⁷ , DB name) G01S ⁷ /00-7/42 G01S 13/00-13/95 H01Q 3/00-3/46 H01Q 21/00-21/30 H01Q 23/00 H01Q 25/00-25/04

Claims (2)

(57) [Claims] An A / D converter converts a plurality of N received signals received by an array antenna composed of a plurality of N antenna elements arranged in parallel at a predetermined element interval into a plurality of N digital received signals. And a detecting means for detecting an angle of arrival of a desired received signal and detecting and outputting an angle of arrival signal indicating the angle of arrival, wherein the receiving signal processing apparatus comprises: Selecting means for selecting from a plurality of N digital received signals output from the converting means and selecting and outputting a plurality of B sets of sub-array signals composed of a set of a plurality of M digital received signals different from each other among the sub-array signals; Multiple M FIRs each having a predetermined filter coefficient
Type digital filter and an adder for adding and outputting a plurality of M signals output from the plurality of M FIR digital filters, have the same transfer function as each other, and are output from the selection means. Digital beam signals are formed on the basis of the plurality of B sets of sub-array signals so that the main beam of the array antenna is directed to the arrival direction of the desired reception signal and the reception signal in the arrival direction of the interference wave signal is made zero. And a plurality of B forming means for outputting the desired received signal based on the plurality of B digital signals output from the plurality of B forming means, The angle of arrival of the received signal is detected to detect and output an angle of arrival signal indicating the angle of arrival. And a filter coefficient of a plurality M of FIR digital filters of each of the forming means for directing a main beam of the array antenna in a direction of arrival of a desired reception signal and for reducing a reception signal in a direction of arrival of an interference wave signal to zero. A reception signal processing device further comprising a calculation means for calculating and controlling the setting. 2. A / D conversion of a plurality of N received signals received by an array antenna composed of a plurality of N antenna elements juxtaposed at a predetermined element interval on a straight line into a plurality of N digital received signals. A receiving means for detecting the angle of arrival of a desired received signal, and detecting and outputting an angle-of-arrival signal indicative of the angle of arrival. Selecting means for selecting and outputting a plurality of B sets of sub-array signals composed of a plurality of M sets of digital received signals different from each other among the plurality of N digital received signals output from the conversion means; And a plurality of u signal processing means, wherein each of the signal processing means includes a plurality of M FIRs each having a predetermined filter coefficient.
Type digital filter and an adder for adding and outputting the plurality of M signals output from the plurality of M FIR digital filters, and having the same transfer function in each of the same signal processing means. The main beam of the array antenna is directed in the same direction of arrival of the same desired reception signal by the same signal processing means based on the plurality of B sets of sub-array signals output from the selection means, and the interference wave signal A plurality of B forming means for forming and outputting digital beam signals, respectively, such that the received signal in the direction of arrival of the signal becomes zero, based on the plurality of B digital signals output from the plurality of B forming means. Receiving signal detecting means for detecting a desired receiving signal, and wherein the angle-of-arrival detecting means includes a plurality of N digital receiving means output from the converting means. Detecting an arrival angle of at least one desired reception signal based on the signal or one of a plurality of B sets of sub-array signals output from the selection means, and detecting an arrival angle signal indicating the arrival angle; The reception signal processing device directs a plurality of u main beams of the array antenna in the arrival direction of at least one reception signal based on the arrival angle signal output from the arrival angle detection means; The receiving means further comprises calculating means for calculating and setting and controlling the filter coefficients of the plurality of M FIR digital filters of the respective forming means so as to make the received signal in the direction of arrival of the interference wave signal zero. Signal processing device.