JP2009047535A - Angle measuring apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an angle measuring apparatus capable of performing an angle measuring process after removing multipath waves, even if complex gain patterns of all sensor elements are not necessarily known. <P>SOLUTION: The angle measuring apparatus includes: a plurality of sensor elements each receiving an incoming wave and outputting observation information; a plurality of combining means for combining observation information sets from a portion or all of the plurality of sensor elements and outputting it as combined information; and a VESPA angle measuring process means which performs a VESPA process using the combined information as a guiding sensor, and estimates an incident angle of the incoming wave. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to an angle measuring device that measures the direction of arrival of a plurality of incoming radio waves.

  In mobile communication, radar, sonar, etc., it is important to measure the angle of arrival with high accuracy by separating the directions of arrival of a plurality of radio waves interfering with the same frequency band. As a conventional angle measurement method, a MUSIC (Multiple Signal Classification) method and an ESPRIT (Estimation of Signal Parameters via Rotational Innovation Techniques) method have been proposed. Furthermore, a VESPA (Virtual ESPRIT Algorithm) system has been proposed that has a smaller computational load than the MUSIC system and less restrictions on characteristics such as the array antenna array than the ESPRIT system (see, for example, Patent Document 1 and Non-Patent Document 1).

  The angle measuring device adopting the VESPA system includes a sensor element having a known complex gain pattern, a spatial filter, an angle measurement processing means, and a means for outputting an angle measurement value. And an incoming wave is observed by a sensor element, and is filtered by a spatial filter (for example, refer nonpatent literature 2). The angle measurement processing means uses the output of the spatial filter to estimate the incident angle of the incoming wave by angle measurement processing that depends on the processing of the spatial filter.

JP 2003-222665 A M.M. Dogan, 1 other, “Application of Cumulants to Array Processing-Part I: Approach Extension and Array Calibration”, IEEE TRANSACTIONS ON SIGNAL. 43, NO. 5, pp. 1200-1216 H. B. By LEE, 1 other, "Resolution Threshold of Beamspace MUSIC For Two Closed Spaced Emitters", IEEE TRANSACTIONS ON ACOUSTICS. SPEECH. AND SIGNAL PROCESSING, IEEE, September 1990, VOL. 38, NO. 9, pp. 1545-1559

  However, in the conventional angle measuring device, in order to perform angle measurement processing after suppressing the interference wave, it is necessary to measure and know in advance the complex gain patterns of all the sensor elements.

  An object of the present invention is to provide an angle measuring device capable of performing angle measuring processing after removing a multipath wave even if complex gain patterns of all sensor elements are not necessarily known.

  The angle measuring device according to the present invention combines a plurality of sensor elements that receive incoming waves and output them as observation information, and a plurality of combinations that combine observation information from a part or all of the plurality of sensor elements and output as combined information Means, and VESPA angle measurement processing means for estimating the incident angle of the incoming wave by VESPA processing using the synthesized information as a guiding sensor.

  The effect of the angle measuring device according to the present invention is that the complex gain pattern needs to be known only for the sensor elements used in the processing of the synthesizing unit when the predetermined arrival wave is suppressed using the guiding sensor output from the synthesizing unit. Therefore, the complex gain pattern needs to be known for all sensor elements as in the conventional angle measuring device, whereas the number of sensor elements whose complex gain pattern is known can be reduced.

Embodiment 1 FIG.
FIG. 1 is an example of a configuration diagram of an angle measuring device according to Embodiment 1 of the present invention.
Angle measuring device according to the first embodiment of the invention, the complex gain pattern known as N sensors elements 1 ₁ to 1 _N, a complex gain pattern unknown L number of sensor elements 5 ₁ to 5 _L, combining means Q spatial filters 2 ₁ to 2 _Q , a VESPA angle measurement processing unit 6 as angle measurement processing means by the VESPA method, and a angle measurement value output unit 4 as angle measurement value output means for outputting angle measurement values. .

Sensor elements 1 ₁ to 1 _N with known complex gain patterns obtain observation information y _{L + 1 to} y _M (M = N + L) of incoming waves.
Moreover, the complex gain pattern unknown sensor element ₅ 1 to 5 _L obtains observation information _y 1 ~y _L incoming waves. Hereinafter, the observation information is represented by an observation information vector y represented by Expression (1).

Observation information y _{L + 1 to} y _M of the incoming wave obtained by the sensor elements 1 ₁ to 1 _N with known complex gain patterns is sent to the spatial filters 2 ₁ to 2 _Q.
The spatial filters 2 _{1 to} 2 _Q perform general filtering described in Non-Patent Document 2.

Either digital processing or analog processing may be adopted as processing in the spatial filters 2 _{1 to} 2 _Q. The filter coefficient vector is set so as to create a null of a composite complex gain pattern by a plurality of sensor elements with respect to a predetermined incident wave incident angle, and the filter coefficient has different characteristics in each of the spatial filters 2 ₁ to 2 _Q. Use things.
A specific example of such a filter coefficient vector is a noise eigenvector used in the MUSIC method or the like when only a predetermined incoming wave is observed. At this time, the filter coefficient vector w _q used by the spatial filter 2 _q (q is an integer of 1 to _Q ) is expressed by Expression (2).

Observation information y _{L + 1 to} y _{M of} arrival waves obtained by sensor elements 1 ₁ to 1 _N with known complex gain patterns and observation information y _{1 of} arrival waves obtained from sensor elements 5 _{1 to} 5 _L with unknown complex gain patterns ˜y _L is sent to the VESPA angle measurement processing unit 6.
Further, the output _x 1 ~x _Q spatial filter ₂ 1 to 2 _Q is sent to VESPA angle measuring unit 6.

Next, processing in the VESPA angle measurement processing unit 6 will be described.
VESPA angle measuring unit 6, an output _x 1 ~x _Q spatial filter ₂ 1 to 2 _Q as Guiding sensor performs angle measurement processing by VESPA manner for observation information _y 1 ~y _M.
VESPA angle measuring unit 6 uses the output _x 1 ~x _Q spatial filter ₂ 1 to 2 _Q as Guiding sensor matrix _{R i} represented by the formula _(3), calculates the _j. Note that the fourth-order cumulant cum [z ₁ , z ₂ , z ₃ , z ₄ ] is defined by Equation (4). E [] represents an average operation as shown in Equation (5), S is the number of samples z ₁ , z ₂ , z ₃ , and z ₄ , and z ₁ (i) is the i-th sample of z ₁ . is there. Each of i and j represents a guiding sensor number, * represents a complex conjugate, and H represents a complex conjugate transpose.

Since there are Q guiding sensors, the VESPA angle measurement processing unit 6 develops the matrix C represented by the equation (6) by the eigenvalue matrix Λ represented by the equation (7) and the eigenvector E represented by the equation (8). . Here, lambda _m, eigenvalue (m = 1,2, ···, ( Q × M)), e m is the eigenvector.

When the incoming wave that is not suppressed by the spatial filters 2 _{1 to} 2 _Q is a K wave, the eigenvalue λ _m has the property of Equation (9). Here, | λ _m | represents the absolute value of the eigenvalue λ _m .

The eigenvector e _k corresponding to the eigenvalue λ _k (k = 1, 2,..., K) is expressed by Q M × K matrices E (bars) _q (q = 1, 2,. • Divide equally into Q).

Output _{x i} and a spatial filter ₂ j of the spatial filter _{2 i} (i, j is an integer of 1 to Q) when the angle measuring an output _{x j} as Guiding sensor matrix from the relationship of formula (11) Ψ _{i, j} Is estimated.

Here, for estimation of the matrix Ψ _{i, j,} a least square method, a total least square method, or the like can be used.
The matrix Ψ _{i, j} is expanded into an eigenvalue matrix Φ _{i, j} represented by the expression (13) and an eigenvector T _{i, j} represented by the expression (14) as in the expression (12). Here, φ _{i, j, k} (k = 1, 2,..., K) are eigenvalues, and t _{i, j, k} are eigenvectors.

Because eigenvalues phi _{i, j,} the incident angle theta _k of _k and arriving wave k is that there is relationship of equation (15), the eigenvalues phi _{i, j,} incident angle theta _k incoming waves k from _k can be estimated. Note that F _i (θ _k ) is obtained by Expression (16).

Here, arg (z) is the phase value of a complex number z, g m _(θ) is the complex gain for the incident angle theta of the sensor element corresponding to the monitoring information y _m, lambda is the wavelength of the incoming waves, ξ (θ) of incidence unit vector corresponding to the angle theta, p _m denotes the coordinate vector of the sensor element corresponding to the monitoring information y _m.
When three or more guiding sensors are used, angle values estimated between the guiding sensors can be integrated.

Since the angle measuring device according to the first embodiment of the present invention uses the outputs x _{1 to} x _Q of the spatial filters 2 ₁ to 2 _Q as a guidance sensor, Reference Document 1 (by E. Gonon, 1 other person, “ There is a clear difference from the angle measurement method by the beam space VESPA shown in “Beamspace virtual-ESPRIT”, the Twenty-Ninth Asimar Conference 1995, October 1995, vol.1, pp.454-457).

It is also conceivable to use different filter coefficient vectors for different sensor elements to which the same filter coefficient is applied. For example, when the number M of all sensor elements is 4, filter coefficient vectors such as Expression (17) and Expression (18) can be considered. Here, γ ₁ , γ ₂ , and γ ₃ are arbitrary coefficients that are not zero.

  Filter coefficient vectors such as Expression (17) and Expression (18) are easy to use when the sensor elements exist in a straight line at equal intervals.

  By the way, a combination of guiding sensors that cannot be used by the VESPA angle measurement processing unit 6 is a case where the filter coefficient is a complex number z times as shown in the equation (19).

However, when the incident angle of a predetermined incoming wave is not uniquely determined, it may not be effective to form a null at a specific incident angle in a combined complex gain pattern by a plurality of sensor elements. This is a case where incident angles of incoming waves are defined by being distributed in an angle band, for example, when a source of a predetermined incoming wave exists in the vicinity of the sensor element.
However, it is well known that such incoming waves can be suppressed by a spatial filter. As a specific example, a noise eigenvector used in the MUSIC method or the like is set as a filter coefficient vector.
Also, when suppressing two completely correlated waves, a filter coefficient vector may be used in which the received power of the two waves is made equal and the phases are inverted and combined. This also shows that the spatial filter that suppresses the incoming wave is not limited to the process of forming nulls in the complex gain pattern.

In order to suppress a predetermined incoming wave by the spatial filters 2 _{1 to} 2 _Q , it is effective to use only the observation information y _{L + 1 to} y _M of the sensor elements 1 ₁ to 1 _N whose known complex gain patterns are known.
However, the observation information used in the spatial filters 2 _{1 to} 2 _Q is not limited to the sensor elements 1 ₁ to 1 _N having a known complex gain pattern. For example, even if the complex gain pattern is an unknown sensor element, the complex gain pattern If an error occurring in the signal is small, a predetermined incoming wave can be sufficiently suppressed. A small error of the complex gain pattern that can be suppressed includes a minute change due to the temperature characteristic of the amplifier.
It can also be expected that the influence of errors occurring in the complex gain pattern is averaged by combining the observation information of the plurality of sensor elements by the spatial filters 2 _{1 to} 2 _Q.
Even if the complex gain patterns of all the sensor elements are unknown, Reference Document 4 (E. Gonnen, two others, “Application of Cumulants to Array Processing-Part IV: Direction Finding in Coherent Signals Case”, IEEE NON TRANS ACT LON PROCESSING, IEEE, September 1997, VOL.45, NO.9, pp. 2265-2276), etc., is used to expand a predetermined gain by using a complex gain pattern estimated by VESPA (Extended VESPA) processing that is an extension of VESPA processing shown in FIG. Incoming waves can be suppressed. Information used in the EVESPA process is also calculated in the VESPA process.

For this reason, when the complex gain patterns of all sensor elements are unknown sensor elements 5 _{1 to} 5 _L as shown in FIG. 2, or the complex gain patterns of all sensor elements are known as shown in FIG. The present invention can also be applied to a case where the sensor elements are 1 ₁ to 1 _N.

In angle measuring device according to the first embodiment of the present invention is used in the processing of the spatial filter 2 ₁ to 2 _Q when suppressing a predetermined arrival wave by using the guiding sensor output from the spatial filter 2 ₁ to 2 _Q Since only the sensor element needs to know the complex gain pattern, the complex gain pattern needs to be known for all sensor elements as in the case of the conventional angle measuring device, whereas the sensor that has the known complex gain pattern. The number of elements can be reduced.

  In the angle measuring device according to the first embodiment of the present invention, it is also possible to employ a process for suppressing the signal frequency or frequency band of a predetermined incoming wave with an FIR filter or the like. It is also possible to separate a signal of a predetermined incoming wave using ICA (Independent Component Analysis) or the like.

Embodiment 2. FIG.
In the VESPA angle measurement processing unit 6 according to Embodiment 1 of the present invention, the matrix C is expanded into the eigenvalue matrix Λ and the eigenvector E from the fourth-order cumulant matrix R _{i, j} using Q guiding sensors. In the VESPA angle measurement processing unit 6 according to Embodiment 2 of the present invention, the matrix C ′ shown in the equation (20) from the fourth-order cumulant matrix R _{i, 1 of one} guiding sensor is expressed by the equation (21). The singular value matrix Σ to be represented, the left singular vector U represented by the equation (22), and the right singular vector V represented by the equation (23) are expanded. Here, σ _m (m = 1 to M) is a singular value, u _m is a left singular vector, and v _m is a right singular vector. H represents complex conjugate transposition.

In the equation (20), the matrix _{R i,} a _j are shown as being limited to _{R i, 1,} Guiding sensor corresponding to the j selects which of _x 1, ···, _{x Q} It doesn't matter.
Note that although the matrix C ″ as in the equation (24) can be used for the angle measurement process, the description is omitted because it is obvious. U ′ is the left singular vector represented by the equation (25), V ′. Is a right singular vector represented by Equation (26).

When the incoming wave that is not suppressed by the spatial filter is a K wave, the singular value σ _m has the property of Equation (27). Here, | σ _m | represents the absolute value of the singular value σ _m .

The left singular vector u _k (k = 1, 2,..., K) corresponding to the singular value σ _k is expressed by Q M × K matrices U (bars) _q (q = 1, 2, ..., equally divided into Q).

When you angle measuring an output _{x j} of the output _{x i} and a spatial filter _{2 j} of the spatial filter _{2 i} as guiding sensor matrix from the relationship of formula (29) Ψ _'i, estimates the _j.

Here, the estimation method of the matrix Ψ ′ _{i, j} is the same as that of the matrix Ψ _{i, j in} the first embodiment, so that the description thereof is omitted.
The matrix Ψ ′ _{i, j} is expanded into an eigenvalue matrix Φ _{i, j} represented by the expression (13) and an eigenvector T ′ _{i, j} represented by the expression (31) as in the expression (30). Here, t ′ _{i, j, k} (k = 1, 2,..., K) is an eigenvector.

Since the relationship of the equation (15) exists between the eigenvalue φ _{i, j, k} (k = 1, 2,..., K) and the incident angle θ _k of the incoming wave _k , the incoming wave is _derived from the eigenvalue φ _{i, j, k.} The incident angle θ _{k of k} can be estimated.

In angle measuring device according to the second embodiment of the present invention is used in the processing of the spatial filter 2 ₁ to 2 _Q when suppressing a predetermined arrival wave by using the guiding sensor output from the spatial filter 2 ₁ to 2 _Q Since only the sensor element needs to know the complex gain pattern, the complex gain pattern needs to be known for all sensor elements as in the case of the conventional angle measuring device, whereas the sensor that has the known complex gain pattern. The number of elements can be reduced.

Embodiment 3 FIG.
FIG. 4 is an example of a configuration diagram of an angle measuring device according to Embodiment 3 of the present invention.
In the spatial filters according to the first and second embodiments, the incident angle of a predetermined incoming wave is known, or prior information obtained by observing only the predetermined incoming wave is obtained. Angle measurement using such a prior method There may be a case where the processing cannot be performed.
The angle measuring device according to the third embodiment of the present invention is the same as the angle measuring device according to the first embodiment except that a spatial filter update processing unit 7 is added to the angle measuring device according to the first embodiment. Reference numerals are added and description is omitted.
The spatial filter update processing unit 7 receives the output of the angle measurement value output unit 4 and updates the filter coefficient vectors of the spatial filters 2 ₁ to 2 _Q.
The spatial filter update processing unit 7 sets a filter coefficient vector w _q having 1 as the initial value and only 0 as the other vector element in the spatial filters 2 ₁ to 2 _Q as initial values.

When the filter coefficient vector shown in Expression (32) is set, the angle measuring device of FIG. 4 operates as a conventional VESPA angle measuring device. Then, the spatial filter update processing unit 7 updates the filter coefficient vectors of the spatial filters 2 ₁ to 2 _Q based on the angle measurement values of the conventional VESPA angle measurement process. The spatial filter update processing unit 7 sets a filter coefficient vector having a null at a specific incident angle of a combined complex gain pattern by a plurality of sensor elements, or sets a filter coefficient vector that separates incoming waves one by one. Then, processing such as changing the number Q of the spatial filters 2 _{1 to} 2 _Q is performed.

  Reference 2 (K. Takao, et al., “An Adaptive Antenna Array Directional Constraint”, IEEE TRANSACTION ON ANTENNAS AND PROPAGATION, 1976, September, 1976. 5, pp. 662-669) and the like described in DCMP (Directly Constrained Minimization of Power).

Embodiment 4 FIG.
FIG. 5 is a block diagram of an angle measuring device according to Embodiment 4 of the present invention.
The angle measuring device according to the fourth embodiment of the present invention has a null in the complex gain pattern instead of the sensor elements 1 ₁ to 1 _N whose complex gain pattern of the angle measuring device according to the first embodiment of the present invention is known. a second sensor element 8 ₁ to 8 _N, and add the sensor element control unit 9 for directing nulls present in the complex gain pattern of the second sensor element 8 ₁ to 8 _N to a predetermined arrival wave, the spatial filter 2 ₁ ˜2 _Q is different except that it is the same, and the others are the same.

The second sensor elements 8 ₁ to 8 _N are mechanical sensor elements that can change the direction in which the null of the fixed complex gain pattern is electrically directed by mechanically swinging the second sensor elements 8 ₁ to 8 _N. It is.
The sensor element control processing unit 9 mechanically swings the second sensor elements 8 ₁ to 8 _N to direct the null existing in the complex gain pattern in the direction of a predetermined incoming wave.
In the second sensor elements 8 ₁ to 8 _N , the observation information is obtained by suppressing the incoming wave to which the null existing in the complex gain pattern is directed. Then, since the obtained observation information is sent to the VESPA angle measurement processing unit 6, it can be used as a guiding sensor in the angle measurement processing in the VESPA angle measurement processing unit 6.

The second sensor elements 8 ₁ to 8 _N may be electrical sensor elements that can form an arbitrary null in a complex gain pattern such as an antenna having a parasitic element with a variable varactor diode. In this case, the sensor element control The processing unit 9 changes the electrical state of the second sensor elements 8 ₁ to 8 _N to direct the null existing in the complex gain pattern in the direction of a predetermined incoming wave.

Embodiment 5 FIG.
FIG. 6 is a diagram illustrating a state in which the angle measuring device described in Non-Patent Document 1 is installed in the vicinity of an inanimate object whose influence cannot be ignored.
In the VESPA angle measuring device described in Non-Patent Document 1, the angle measurement accuracy may be deteriorated when performing the correlation estimation of the incoming waves. For example, when the angle measuring device 10 is installed, there is an inanimate object 11 whose influence cannot be ignored due to restrictions on installation conditions. In such a case, the multipath wave can be regarded as coming from the direction of the inanimate object 11 without depending on the incident angle of the incoming wave. In the conventional angle measuring device 10, even if VESPA angle measurement processing is adopted, the complex gain pattern of all sensor elements is known, or nulls existing in the complex gain pattern of all sensor elements are replaced with the orientation of the inanimate object 11. The effect of multipath waves could not be ignored unless it was directed to

  Reference 3 (E. Gonen, two others, “Applications of Cumulants to Array Processing-Part IV: Direction Finding in Coherent Signals Cases, 1997,” .9, pp. 2265-2276), there are significant restrictions on the sensor elements, such as arranging a plurality of guiding sensors with matching complex gain patterns in a straight line at regular intervals. is there.

  If the angle measuring device according to the present invention is installed in a place as shown in FIG. 6, if even some sensor elements have a known complex gain pattern, the interference wave suppression effect in the spatial filter can reduce the multipath wave. The influence can be neglected, and it is possible to separate and estimate the correlated incoming waves, which are not good at the VESPA angle measurement process.

It is an example of the block diagram of the angle measuring device which concerns on Embodiment 1 of this invention. It is another example of the block diagram of the angle measuring device which concerns on Embodiment 1 of this invention. It is another example of the block diagram of the angle measuring device which concerns on Embodiment 1 of this invention. It is an example of the block diagram of the angle measuring apparatus which concerns on Embodiment 3 of this invention. It is an example of the block diagram of the angle measuring apparatus which concerns on Embodiment 4 of this invention. It is a figure which shows a mode that the angle measuring apparatus of the nonpatent literature 1 was installed in the vicinity of the inanimate body whose influence cannot be disregarded.

Explanation of symbols

1 ₁ to 1 _N sensor element, 2 ₁ to 2 _Q spatial filter, 4 angle measurement value output unit, 5 _{1 to} 5 _L sensor element, 6 VESPA angle measurement processing unit, 7 spatial filter update processing unit, 8 ₁ to 8 _N Second sensor element, 9 sensor element control processing unit, 10 angle measuring device, 11 inanimate object, x _{1 to} x _Q (from spatial filter) output, y _{1 to} y _M observation information.

Claims (6)

A plurality of sensor elements that receive incoming waves and output them as observation information;
A plurality of combining means for combining observation information from a part or all of the plurality of sensor elements and outputting the combined information;
VESPA angle measurement processing means for estimating an incident angle of an incoming wave by VESPA processing using the composite information as a guiding sensor;
An angle measuring device comprising:   The angle measuring device according to claim 1, wherein the sensor element that sends observation information to the synthesizing means has a known complex gain pattern.   The angle measuring device according to claim 1, wherein the synthesizing unit suppresses a predetermined incoming wave.   4. The angle measuring device according to claim 3, further comprising means for updating a processing content for suppressing a predetermined incoming wave of the synthesizing means using a result of the VESPA angle measuring processing means. A plurality of first sensor elements that receive incoming waves and output them as observation information;
A plurality of second sensor elements that output observation information in which the null existing in the complex gain pattern is adjusted to be directed to a predetermined incoming wave and the predetermined incoming wave is suppressed;
VESPA angle measurement processing means for estimating an incident angle of an incoming wave by VESPA processing using observation information from the second sensor element as a guiding sensor;
An angle measuring device comprising:   6. The angle measuring device according to claim 5, further comprising means for adjusting an orientation of the second sensor element using a result of the VESPA angle measurement processing means.

Images