JP5063416B2 - Angle measuring device - Google Patents

Description

  The present invention relates to an angle measuring device that measures an incident angle of an incoming signal.

  In mobile communication, radar, sonar, and the like, if there is one incoming signal, the incident angle of the signal can be estimated from the reception characteristic difference between the two sensors. However, this method cannot handle a case where a plurality of signals arrive. On the other hand, by using an angle measurement method called VESPA (Virtual ESPRIT Algorithm), the incident angles of a plurality of signals are estimated from the reception characteristic difference between two sensors (referred to as guiding sensors) whose reception characteristics are known. There is an angle measuring device that can be used (for example, see Non-Patent Document 1).

  In the angle measurement method using VESPA, in order to improve the angle measurement accuracy, the interval between the guiding sensors may be increased. However, if the guiding sensor interval exceeds a half wavelength, angular ambiguity occurs in the angle measurement result. Therefore, a highly accurate angle measurement method that eliminates angular ambiguity by performing a VESPA process twice or more using a guiding sensor with a sensor interval larger than a half wavelength or a guiding sensor with a sensor interval less than a half wavelength. (For example, refer to Patent Document 1).

JP 2003-222666 A M. Dogan and J. Mendel, "Application of Cumulants to Array Processing Part I: Aperture Extension and Array Calibration," IEEE Trans. Signal Processing, vol.43, no.5, pp.1200-1216, May 1995.

However, the prior art has the following problems.
In the angle measuring device of Patent Document 1, it is necessary to execute the VESPA process twice or more, and there is a problem that the amount of calculation increases by the number of times of performing the VESPA process, and the calculation load becomes heavy.

  The present invention has been made in order to solve the above-described problems. By using one VESPA process or one ESPRIT process, the angle measurement accuracy can be improved and the amount of calculation can be reduced. The purpose is to obtain an angle measuring device.

The angle measuring device according to the present invention receives a plurality of incoming signals by a plurality of sensor elements, and calculates a first reception characteristic difference corresponding to each incoming signal based on one VESPA process or one ESPRIT process. And calculating the incident angle of the incoming signal as an estimated candidate value based on the calculated first reception characteristic difference and calculating the steering vector estimated by the one VESPA process or the one ESPRIT process. A second reception characteristic difference corresponding to each incoming signal is calculated based on the calculated second reception characteristic difference, and an incident angle of the incoming signal is calculated as an estimated reference value based on the calculated second reception characteristic difference. The final estimated incident angle of the incoming signal is obtained by selecting a value closest to the estimated reference value from

  According to the present invention, by calculating the incident angle estimated value θ based on the estimated value of the array mode vector corresponding to the incident angle θ of each incoming signal, one VESPA process or one ESPRIT process is used. Thus, it is possible to obtain an angle measuring device capable of improving the angle measuring accuracy and reducing the amount of calculation.

  Hereinafter, preferred embodiments of the angle measuring device of the present invention will be described with reference to the drawings.

Embodiment 1 FIG.
The present invention provides a highly accurate angle measuring device that does not produce angular ambiguity by performing angle measuring processing with a small sensor interval in VESPA processing using a guiding sensor with a sensor interval larger than a half wavelength. It is. In VESPA processing, as in Non-Patent Document 1, it is possible not only to perform angle measurement processing using a guiding sensor but also to estimate an array mode vector a (θ) corresponding to the incident angle θ of each incoming signal. It is.

  In the present invention, by using an element corresponding to a sensor with a known element antenna radiation pattern in the estimated array mode vector a (θ), angle measurement processing with a small sensor interval is also executed. That is, angle measurement processing with a small sensor interval is also performed by estimating the reception characteristic difference between sensors corresponding to each incoming signal from the mutual statistics between the sensor output signals.

  When the guiding sensor is larger than the half wavelength, the incident angle estimation result by the VESPA process is highly accurate, but if it is not changed, the angle ambiguity is generated and cannot be determined uniquely. However, the angle ambiguity can be eliminated by referring to another incident angle estimation result estimated from the array mode vector a (θ).

  As a result, a highly accurate angle measurement process similar to that disclosed in Patent Document 1 can be performed with one VESPA process, and an angle measurement process that does not cause angular ambiguity can be performed. That is, while the angle measuring device provided by Patent Document 1 requires two or more VESPA processes, the angle measuring apparatus provided by the present invention only requires one VESPA process, so that high accuracy is maintained. The amount of computation can be reduced as it is.

  First, the VESPA processing method disclosed in Non-Patent Document 1 will be described. In the VESPA process, for example, if the #i sensor and the #j sensor are selected as guiding sensors among L (L is an integer of 2 or more) omnidirectional sensors, the received data vector r (t ) To calculate a matrix V as shown in the following equation (1).

Here, r _i (t) is the #i element of the vector r (t), * is a complex conjugate, H is a complex conjugate transpose, Γ is a diagonal matrix of the fourth-order cumulant of the incident signal, K is the number of incident signals, s _k (t) is the complex amplitude of the #k incident signal at time t, a _i (θ) is the #i element of the vector a (θ), p _i is the coordinate vector of the #i element, T is the transpose, q (θ) Represents a unit coordinate vector corresponding to the incident angle θ, λ represents a signal wavelength, and n (t) represents a noise vector at time t.

In particular, the above equation (10) represents the relationship between the reception characteristic difference phi _{i, j} between #i sensor and #j sensor corresponding to the incident angle theta _{k, k} an array mode vector a (theta) . The conversion process from the reception characteristic difference φ _{i, j, k} to the incident angle θ _k is performed by the inverse calculation of the above equation (10), and the magnitude of the vector (p _j −p _i ) is the signal wavelength λ. An angle ambiguity is produced when it becomes larger than half of the angle.

Then, next, the improvement measure of this angle ambiguity by this invention is demonstrated below.
When the matrix V is subjected to singular value decomposition, the following equations (11) and (12) are obtained.

Here, among the left singular vectors u _l , a matrix U _{s in} which vectors corresponding to the signal subspace are arranged is divided into two L × K matrices U _X and U _Y as shown in the following equation (13).

Array mode matrix A by a linear combination of the signal subspace _{U S,} can be expressed by the following equation (14a) ~ (14c).

The regular matrix T and the matrix Φ _{i, j} can be determined from the above equations (14a) to (14c). Thereafter, in the VESPA process, as in Non-Patent Document 1, it is possible to estimate the reception phase difference estimated value φ (˜) _{i, j, k} between the # i- # j elements from the matrix Φ _{i, j} . Furthermore, A (￣), which is an estimated value of the array mode matrix, can be estimated from the above equation (14b) (for example, non-patent literature E. Gonen, JM Mendel, “Applications of cumulants to array processing. IV. Direction finding”). in coherent signals case, "IEEE Trans. Signal Processing, vol. 45, no. 9, pp. 2265-2276, Sep 1997.").

  Here, (˜) means an estimated value with “˜” attached to the sign described before (), and means an estimated value by VESPA using a guiding sensor whose sensor interval is larger than a half wavelength. is doing. Further, (￣) means an estimated value in which ￣ is added on the code described before (), and an estimated value or array mode by a VESPA process using a guiding sensor whose sensor interval is less than half wavelength. Means an estimate based on a vector.

In the present invention, the matrix A (￣) estimated in this way, that is, the array mode vector a (￣) (θ _k ) (k = 1,..., K) is estimated, and an arbitrary # l− The reception phase difference estimated value φ (￣) _{l, m, k} between #m elements is _obtained by the following equation (15).

Here, a _l (￣) (θ _k ) is the #l element of the vector a (￣) (θ _k ). From the relationship between a _l (￣) (θ _k ) and a _m (￣) (θ _k ), an incident angle estimated value θ _k (￣) is obtained. For example, consider a case represented by the following equations (16) and (17).

In such a case, the incident angle estimated value θ (￣) _k is _obtained from the received phase difference estimated value φ (￣) _{l, m, k} of the above equation (15) using the following equation (18).

  Here, Im [Z] is an arithmetic process for obtaining the imaginary part of the complex number Z, and ln () is a natural logarithmic function.

By the above equation (18), it is possible to calculate the incident angle estimated value θ (˜) _k of the k-th wave from the received phase difference estimated value φ (˜) _{l, m, k} . Here, a sensor arrangement as shown in FIG. 1 is considered. FIG. 1 is an explanatory diagram showing the arrangement of each sensor according to Embodiment 1 of the present invention, and the alphabets marked with # mean the respective sensor numbers.

As in the guiding sensor composed of the #i sensor and the #j sensor in FIG. 1, when the interval ‖p _j −p _iの between the two sensor elements exceeds a half wavelength, the received phase difference estimated value φ (˜) Ambiguity occurs when _{l, m, and k} are converted into angle values. As a result, estimated candidate values θ (˜) _1, k˜θ (˜) _{Q, k} for a plurality of incident angles are obtained.

Therefore, the estimated candidate value _{θ (~) 1, k ~θ} (~) Q, in order to solve the ambiguity between the _k, the reception phase known and sensor spacing ‖p _m -p _l ‖ less than half a wavelength #L sensor and #m sensor are selected. Then, using the above equation (18), an incident angle estimation reference value θ (￣) _k that does not cause an angle ambiguity is calculated from the received phase difference estimated value φ (￣) _{l, m, k} .

The angle value closest to the estimated reference value θ (￣) _k is selected from the estimated candidate values θ (˜) _1, k˜θ (˜) _{Q, k} obtained previously, and the final incident angle estimated value Output as θ (^) _k . Here, θ (^) means an estimated value in which ^ is added on θ, and means a final incident angle estimated value. By performing such a series of arithmetic processing, the angle measuring device according to the first embodiment can realize angle measuring processing with high accuracy and no angular ambiguity.

  Next, a specific configuration of the angle measuring device according to the first embodiment will be described. FIG. 2 is a configuration diagram of the angle measuring device according to the first embodiment of the present invention. The angle measuring device of FIG. 2 includes L (L is an integer of 2 or more) sensors 1 (1) to 1 (L), an observation data vector generation unit 2, a VESPA processing unit 3, and angle conversion processing units 4a and 4b. , A pairing processing unit 5, an angle measurement result output unit 6, an array mode vector estimation processing unit 7, and a reception phase difference calculation processing unit 8.

Here, of the L sensors 1 (1) to 1 (L), the sensor 1 (1) and the sensor 1 (2) have a sensor interval smaller than half of the signal wavelength, and the sensor 1 (1) Sensor 1 (L) has a sensor interval larger than half of the signal wavelength. Observation data r ₁ (t) to r _L (t) in each of the sensors 1 (1) to 1 (L) is generated as a vector r (t) by the observation data vector generation unit 2.

In the VESPA processing unit 3, the sensors 1 (1) and 1 (L) are guiding sensors, and the observation data r ₁ (t) and r _L (t) and the observation data vector r (t) are used as input values. The VESPA process is performed using the matrix V in the above equation (1). The VESPA processing unit 3 obtains the reception phase difference φ between the guiding sensors expressed by the above equation (10), and outputs the reception phase difference estimated value φ (˜). Furthermore, the VESPA processing unit 3 outputs a matrix U _S of the above equation (13) and a matrix T that can be determined from the above equation (14).

Array mode vector estimation processor 7, based on equation (14a) ~ (14c), to estimate the matrix A (¯) from the matrix _{U S} and the matrix T. Next, the reception phase difference calculation processing unit 8 calculates the reception phase difference φ (￣) between the sensor 1 (1) and the sensor 1 (2) from the matrix A (￣) based on the above equation (15). . The angle conversion processing unit 4a converts the reception phase difference φ (￣) into an angle θ (￣) as shown in the above equation (18). On the other hand, the angle conversion processing unit 4b converts the reception phase difference φ (˜) into the angle θ (˜) based on the above equation (10).

However, since the sensor interval between the sensor 1 (1) and the sensor 1 (L) is larger than half of the signal wavelength λ, the angle ambiguity is generated in the conversion based on the above equation (10). Therefore, the angle conversion processing unit 4b outputs a plurality of Q (Q is an integer of 2 or more and K or less) angles θ (˜) ₁ to θ (˜) _Q.

The pairing processing unit 5 selects one closest to the angle θ (￣) from a plurality of Q angles θ (˜) ₁ to θ (˜) _Q, and determines the incident angle estimated value θ (^). To do. Further, the angle measurement result output unit 6 outputs the incident angle estimated value θ (^) determined by the pairing processing unit 5.

  Here, the difference between the configuration of the angle measuring device in the first embodiment and the conventional angle measuring device will be described with reference to the drawings. FIG. 6 is a configuration diagram of a conventional angle measuring device based on Patent Document 1. In FIG. The conventional angle measuring device of FIG. 6 includes L sensors 1 (1) to 1 (L), an observation data vector generation unit 2, VESPA processing units 3a and 3b, angle conversion processing units 4a and 4b, and a pairing processing unit. 5 and an angle measurement result output unit 6.

Here, of the L sensors 1 (1) to 1 (L), the sensor 1 (1) and the sensor 1 (2) have a sensor interval smaller than half of the signal wavelength, and the sensor 1 (1) Sensor 1 (L) has a sensor interval larger than half of the signal wavelength. In the conventional angle measuring device, when the reception phase difference φ (￣) between the sensor 1 (1) and the sensor 1 (2) is estimated, the VESPA processing unit 3a uses the observation data r ₁ (t) and r ₂ ( t) and the observed data vector r (t) are used as input values, and VESPA processing is performed using the matrix V in the above equation (1).

  In other words, even when the sensor interval is smaller than half the signal wavelength, the same processing as that for the sensor interval larger than the signal wavelength is performed, and it is necessary to execute the VESPA process twice. As a result, in the conventional angle measuring device, the amount of calculation increases as the number of times of performing the VESPA process increases, and the calculation scale is large and the processing load is heavy.

  On the other hand, the angle measuring apparatus having the configuration of FIG. 1 according to the first embodiment requires only one VESPA process, and therefore, the calculation scale is small and high-speed calculation processing can be realized.

When there are a plurality of incident signals, a plurality of reception phase differences φ (￣) _{1 to} φ (￣) _K and reception phase differences φ (˜) ₁ to φ (˜) _K can be obtained. The conventional angle measuring device of FIG. 6 in Patent Document 1 combines the reception phase difference φ (￣) _k and the reception phase difference φ (˜) _k corresponding to the same signal before the pairing processing unit 5 is implemented. I could not.

On the other hand, in the angle measuring device of FIG. 2 in the first embodiment, the reception phase difference φ (￣) _k and the reception phase difference φ (˜) _k corresponding to the same signal are expressed by the above equation (14). Based on this, it can be combined before the pairing processing unit 5 is implemented. As a result, the estimation accuracy of the array mode matrix A (￣) is improved, and the accuracy of the finally obtained incident angle estimation value θ (^) can be improved.

  In particular, it is possible to improve measurement accuracy by using a sensor pair having a maximum interval among sensors with known reception characteristics for VESPA processing. Furthermore, with respect to the angular ambiguity generated when the guiding sensor interval exceeds half wavelength, among the array mode vectors a (θ) estimated by the VESPA processing, the element antenna radiation pattern corresponds to a known sensor. By using elements, angle measurement processing with a small sensor interval can also be executed.

  In this way, the final incident angle estimated value can be obtained using the estimated value obtained by the VESPA process and the estimated value obtained based on the array mode vector a (θ) estimated by the VESPA process. . As a result, by performing one VESPA process, it is possible to improve both the angle measurement accuracy and reduce the amount of calculation.

  Further, by using a sensor having the same reception characteristics as a guiding sensor for performing VESPA processing, it is possible to further improve the angle measurement accuracy.

  As described above, according to the first embodiment, the incident angle estimated value θ (θ) is calculated based on the estimated value of the array mode vector a (θ) corresponding to the incident angle θ of each incoming signal. The VESPA process can be completed only once, and the problem of angle ambiguity can be solved. As a result, it is possible to obtain an angle measuring device capable of improving the angle measuring accuracy and reducing the amount of calculation.

In the first embodiment, the case where the angle measurement process is performed by assuming the omnidirectional sensor and estimating the reception phase difference between the sensors has been described. However, the angle measuring device in Embodiment 1 is not limited to this. Even when there is a gain difference in the element antenna radiation pattern of each sensor, the incident angle estimated value θ (￣) _k is calculated from the phase difference of the estimated values φ (￣) _{l, m, k} similarly to the above equation (18). It is possible to obtain.

Embodiment 2. FIG.
FIG. 3 is a configuration diagram of the angle measuring device according to the second embodiment of the present invention. In the first embodiment, the reception phase difference φ (˜) between the sensor 1 (1) and the sensor 1 (L) is obtained by the VESPA processing unit 3 as shown in FIG. On the other hand, in the second embodiment, as shown in FIG. 3, the reception phase difference calculation processing unit 8b is used to calculate the reception phase difference φ (˜) from the matrix A (￣) based on the above equation (15). ).

  As described above, the angle measuring device according to the second embodiment having the configuration shown in FIG. 3 has the reception phase difference calculation processing units 8a and 8b even if the reception phase of the guiding sensor used in the VESPA processing unit 3 is unknown. Thus, the reception phase difference between sensors with known reception phases can be calculated.

  As described above, according to the second embodiment, the incident angle estimated value θ (モ ー ド) and the incident angle estimated value are based on the estimated value of the array mode vector a (θ) corresponding to the incident angle θ of each incoming signal. By calculating θ (˜), the VESPA process can be performed only once. As a result, similarly to the first embodiment, it is possible to obtain an angle measuring device capable of improving the angle measuring accuracy and reducing the amount of calculation.

Embodiment 3 FIG.
In the first and second embodiments, the angle measuring device by performing the VESPA process once has been described. On the other hand, in this Embodiment 3, the angle measuring device by performing ESPRIT processing once instead of VESPA processing is demonstrated.

  In the sensors 1 (1) to 1 (L), the VESPA process can be replaced with the ESPRIT process depending on the conditions of the reception phase and the installation position. The VESPA process is based on the ESPRIT process, and can be returned to the ESPRIT process if the condition is satisfied. Such a condition is, for example, a case where the sensors 1 (1) to 1 (L) having a non-directional reception phase are installed on a straight line at equal intervals.

  FIG. 4 is a configuration diagram of the angle measuring device according to the third embodiment of the present invention, in which the VESPA processing unit 3 in FIG. 2 in the previous first embodiment is replaced with an ESPRIT processing unit 9. FIG. 5 is another configuration diagram of the angle measuring device according to the third embodiment of the present invention, in which the VESPA processing unit 3 in FIG. 3 in the previous second embodiment is replaced with an ESPRIT processing unit 9. .

  As described above, according to the third embodiment, the same effect as in the first and second embodiments can be obtained by replacing the VESPA process with the ESPRIT process.

  In the first to third embodiments, the case where the phase difference is used as the reception characteristic difference has been described. However, when the sensor element is a directional sensor, the measurement is also performed by using the reception power difference as the reception characteristic difference. Corner processing can be performed. In general, the angle measurement process is performed using both the reception phase difference and the reception power difference. However, in the VESPA process and ESPRIT process described above, the angle measurement method can be sufficiently implemented using only the reception phase difference. . However, when a plurality of directivity sensors are installed in different directions, angle measurement processing based on the received power difference is also possible.

It is explanatory drawing which showed arrangement | positioning of each sensor in Embodiment 1 of this invention. It is a block diagram of the angle measuring device in Embodiment 1 of this invention. It is a block diagram of the angle measuring apparatus in Embodiment 2 of this invention. It is a block diagram of the angle measuring apparatus in Embodiment 3 of this invention. It is another block diagram of the angle measuring device in Embodiment 3 of this invention. It is a block diagram of the conventional angle measuring device based on patent document 1. FIG.

Explanation of symbols

  1 sensor, 2 observation data vector generation unit, 3 VESPA processing unit, 4a, 4b angle conversion processing unit, 5 pairing processing unit, 6 angle measurement result output unit, 7 array mode vector estimation processing unit, 8, 8a, 8b reception Phase difference calculation processing unit, 9 ESPRIT processing unit.

Claims (8)

A plurality of incoming signals are received by a plurality of sensor elements, a first reception characteristic difference corresponding to each incoming signal is calculated based on one VESPA process or one ESPRIT process, and the calculated first reception The incident angle of the incoming signal is calculated as an estimated candidate value based on the characteristic difference, and the second corresponding to each incoming signal based on the steering vector estimated in the one VESPA process or the one ESPRIT process And calculating the incident angle of the incoming signal as an estimated reference value based on the calculated second received characteristic difference, and selecting a value closest to the estimated reference value from the estimated candidate values. An angle measuring device characterized in that a final incident angle estimation value of the incoming signal is obtained by selection . The angle measuring device according to claim 1 ,
The estimated candidate value is calculated based on the first reception characteristic difference calculated for a sensor pair whose sensor interval is wider than half wavelength, and is calculated for a sensor pair whose sensor interval is narrower than half wavelength. The angle measuring device characterized in that the estimated reference value is calculated based on the second reception characteristic difference. The angle measuring device according to claim 1 or 2 ,
The first reception characteristic difference and the second reception characteristic difference between the guiding sensors are calculated, and both are combined to obtain the final incident angle estimation value of the incoming signal. Corner device. The angle measuring device according to claim 3 ,
An angle measuring device using a pair of sensors having a maximum interval among sensors having a known reception characteristic as the guiding sensor. The angle measuring device according to claim 3 or 4 ,
An angle measuring device using a sensor having the same reception characteristic as the guiding sensor. The angle measuring device according to any one of claims 3 to 5 ,
An angle measuring device characterized in that an interval between the guiding sensor pair exceeds a half wavelength. The angle measuring device according to any one of claims 1 to 6 ,
The angle measuring device according to claim 1, wherein the first reception characteristic difference is calculated corresponding to each incoming signal based on a steering vector estimated by the one VESPA process or the one ESPRIT process . The angle measuring device according to any one of claims 1 to 7 ,
An angle measuring device using a phase difference as the first reception characteristic difference and the second reception characteristic difference .

Images