JP2781775B2 - Method and apparatus for measuring seafloor acoustic characteristics - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and an apparatus for measuring seafloor acoustic characteristics, and more particularly to a method and apparatus for measuring submarine acoustic characteristics based on a correlation between measured data of sound waves propagating in the sea and replica data obtained by propagation calculation. Method and apparatus.

[0001]

2. Description of the Related Art Conventionally, this type of undersea acoustic characteristic measuring apparatus accurately and efficiently estimates multidimensional unknown parameters relating to undersea acoustic characteristics based on correlation values between measurement data of sound waves propagating in the sea and replica data. For example, for example, IEE Journal of Oceanic Engineering, July 18, 1993, Vol. 18, No. 3 (IEEE JOURNAL OF
OCEANIC ENGINEERING, VOL. 18, No. 3, JULY, 1993), pages 224 to 231. Here, this conventional submarine acoustic characteristic measuring apparatus will be described with reference to FIG.

FIG. 6 is a flowchart showing a processing procedure for outputting an estimated value of the submarine acoustic characteristic in the conventional submarine acoustic characteristic measuring apparatus, and this processing algorithm forms a loop of an iterative process. Each time this loop is repeated, an N-dimensional parameter set {X (1) to X (N)} relating to the ocean floor acoustic characteristics is set by random number calculation (S201). For the set parameter set, propagation calculation of replica data of each receiver is performed (S202). Here, the receiver receives a sound wave transmitted from a sound source and propagated in the sea. A correlation value R between the replica data and the measurement data input by each receiver at the time of the first loop is calculated (S203, S2).
04) Further, a correlation value change ΔR from the correlation value calculated in the previous loop is calculated (S205). Then, it is determined whether or not the correlation value R has increased based on the correlation value change ΔR (S206). If the correlation value change ΔR ≧ 0, that is, the correlation value R has increased, the value is set in the previous loop. The value of the parameter set is replaced with the value of the newly set parameter set (S207). On the other hand, when the correlation value change △ R <0, that is, when the correlation value R decreases, the probability distribution P (△ R) is further calculated, and the value φ obtained by the random number calculation becomes the probability distribution P (△ R). It is determined whether or not the condition is satisfied (S208). Note that the probability distribution P (△
R) is a value of 0 <P (△ R) <１ ／, and the value φ is
This is an arbitrary number selected with equal probability in the range [0, 1]. Here, when P (△ R) ≧ φ, that is, when the value φ obtained by the random number calculation satisfies the probability distribution P (△ R), the value of the parameter set set in the previous loop is newly set. It is replaced with the value of the parameter set (S207). On the other hand, when P (△ R) <φ, that is, when the value φ does not satisfy the probability distribution P (△ R), the value of the set parameter set is excluded, and the parameter set set in the previous loop remains ( S210). The magnitude of this probability distribution is proportional to the moving average value of the correlation value change ΔR. The moving average value of the correlation value change ΔR is ΔR <0 and the value φ
Is satisfied when the probability distribution is satisfied (S209), and from the next loop, the newly calculated moving average value is used for the calculation of the probability distribution.

The above processing loop is repeated until a predetermined number of times is reached, and when the number of repetitions reaches the set number, the value of the parameter set at that time is output as an estimated value of the seafloor acoustic characteristic (S211, S21
2).

[0004]

In this conventional seafloor acoustic characteristic measuring apparatus, the condition for calculating the moving average value of the correlation value change .DELTA.R is that the correlation value R between the replica data and the measured data is reduced and the random number is calculated. Is the probability condition P (△
R), it is difficult to reduce the moving average value, and accordingly, the probability distribution that allows the correlation value to decrease is also slow. Therefore, there is a problem that a large number of repetition processes are required until the overall correlation maximum point is approached.

Further, since the end point of the repetitive processing is assumed to be the case where the number of loops has reached a predetermined number, the value of the parameter set having a large error when the total correlation does not reach the maximum point is estimated. May be output, and accuracy problems remain.

[0006]

Means for Solving the Problems In order to solve the above-mentioned problems, a receiver for receiving a sound wave transmitted from a sound source and propagated in the sea, and measurement data of the seafloor acoustic characteristics obtained by the receiver are provided. And an analysis means for estimating the seafloor acoustic characteristic based on a correlation value with the replica data of the seafloor acoustic characteristic obtained by the propagation calculation and the seafloor acoustic characteristic. A replica data calculating means for calculating replica data, and a correlation calculating means for calculating a correlation value between the measured data obtained by the receiver and the replica data, according to a multidimensional parameter set relating to the seafloor acoustic characteristics set in A correlation change calculating means for calculating a difference between the correlation value and the correlation value calculated in the previous loop as a correlation value change, and a parameter set according to the correlation value change. A correlation change evaluation means for selecting whether to update or exclude the value of the parameter set; a moving average calculation means for calculating a moving average of the correlation value change when the parameter set update is selected; and a parameter set update selected. If the parameter set is updated, the parameter updating means updates the value of the parameter set to the value of the parameter set set in the current loop, and the loop is terminated based on predetermined conditions, and the parameter set at that time is terminated. Is provided as an estimated value.

Further, a receiver for receiving a sound wave transmitted from a sound source and propagated in the sea, measurement data of the seafloor acoustic characteristic obtained by the receiver, and replica data of the seafloor acoustic characteristic obtained by propagation calculation, The seafloor acoustic characteristic measuring method of the present invention for estimating the submarine acoustic characteristic based on the correlation value of the submarine acoustic characteristic is performed by calculating replica data according to a multidimensional parameter set related to the submarine acoustic characteristic set for each repetition loop, and A correlation value between the measurement data obtained by the measuring device and the replica data is calculated. The difference between the correlation value and the correlation value calculated in the previous loop is calculated as a correlation value change, and whether to update or exclude the value of the parameter set is selected according to the correlation value change. When updating the parameter set is selected, a moving average of the correlation value change is calculated. When the update of the parameter set is selected, the value of the parameter set is updated to the value of the parameter set set in the current loop. Based on predetermined conditions,
The repetition loop is terminated, and the value of the parameter set at that time is output as an estimated value.

[0008] In particular, the correlation change evaluation means in the seafloor acoustic characteristic measuring apparatus according to the present invention is arranged such that when the correlation value changes so as to increase, or when the correlation value changes so as to decrease and the value obtained by random number calculation is correlated. When the probability distribution that allows a decrease in the value is satisfied, it is selected to update the value of the parameter set.

The end determination means in the seafloor acoustic characteristic measuring apparatus of the present invention evaluates a variation range of the set value of the parameter set with respect to the number of repetitions, and the variation of the set values of all the parameter sets is a predetermined value. When it becomes stable below, the repetition loop is ended and the set value of the parameter set at that time is output as an estimated value of the seafloor acoustic characteristic.

[0010]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Next, an embodiment of the apparatus for measuring acoustic characteristics of the seabed of the present invention will be described in detail with reference to the drawings.

FIG. 1 is a diagram showing a configuration of an embodiment of the present invention. A sound wave transmitted from a sound source 1 propagates in the entire underwater sound field including under the seabed, and each of the receivers 2 Received by the waver. In the analysis device 3, the measurement data of each receiver is analyzed by the analysis unit 4, and the seafloor acoustic characteristics are output.

FIG. 2 is a block diagram showing the configuration of the analysis unit 4 in FIG. 1. The parameter setting unit 5 sets a multidimensional parameter set relating to the seafloor acoustic characteristics. Here, the multi-dimensional parameter set includes, for example, a sound velocity profile, density, attenuation rate, layer structure, and the like below the seabed indicating seafloor acoustic characteristics. The replica calculator 6 calculates replica data corresponding to the set parameter set using a predetermined propagation model, and outputs the replica data to the correlation calculator 7. Correlation calculator 6
Calculates the correlation value between the measurement data obtained by each receiver and the replica data. At the time of the first loop,
The measurement data obtained by each receiver is also used in the correlation calculation in the subsequent loop. The correlation change calculator 8 calculates
The difference between the correlation value supplied from the correlation calculator 7 and the correlation value obtained in the previous loop, that is, the amount of change in the correlation value is calculated. The correlation change evaluation unit 9 evaluates a change amount of the correlation value supplied from the correlation change calculation unit 8. Specifically, the correlation change evaluation unit 9 determines whether or not the correlation value has changed so as to increase based on the amount of change in the correlation value. Further, when it is determined that the correlation value has decreased, it is determined whether or not the value obtained by the random number calculation satisfies a probability distribution that determines an allowable probability of allowing the correlation value to decrease. The moving average calculation unit 10 determines the correlation value when the correlation change evaluation unit 9 determines that the correlation value has increased, or when the above value satisfies the probability distribution even when the correlation value decreases. Calculate the moving average of the change. The parameter update unit 11
After the moving average calculation unit 10 calculates the moving average, the value of the parameter set is updated to the value newly set in the current loop. The end determination unit 12 determines whether the updated value of the parameter set satisfies a preset end condition. If the end condition is satisfied, the value of the parameter set at that time is used as the seafloor acoustic characteristic. Output.
If the termination condition is not satisfied, the parameter setting unit 5 starts the processing of the next loop. Note that the magnitude of the probability distribution described above is proportional to the moving average value of the change in the correlation value, and the moving average calculation unit 10 calculates the moving average value of the change in the correlation value. The average value is used for setting the probability distribution.

Next, the operation of this embodiment will be described.

FIG. 3 is a flowchart showing the operation of the present embodiment. This processing algorithm forms a loop of a repetitive processing. Each time this loop is repeated, the parameter setting unit 5 calculates the N-dimensional parameter set {X (1) to
X (N)} is set (S101). For the set parameter set, replica data of each receiver is calculated by the replica calculation unit 6 using a predetermined propagation model (S102). The correlation value R between this replica data and the measurement data input by each receiver at the time of the first loop
Is calculated by the correlation calculator 7 (S103, S1)
04). The measurement data obtained by each of the receivers is input during the first loop, and the data is continuously used as measurement data in the subsequent loops.

As a method of calculating the correlation value R, for example, B
The ARTLETT correlation equation can be used.

The measurement data vector of each receiver is defined as a vector M = {M (1),..., M (N)}, and the replica data vector of each receiver is defined as a vector X = {X (1).・
.., X (N)}, the correlation value R is calculated by the following equation.

[0017]

Here, T is an operator of complex conjugate transpose, and [] is a time average of measured values. The calculated correlation value R is normalized and becomes a MAX correlation value 1 if the vector M and the vector X completely match. Note that the calculation of the correlation value is not limited to this method.
Other calculation methods, such as a method using a correlation equation or the like, may be applied.

Next, the difference between the correlation value obtained by the correlation calculator 7 and the correlation value calculated in the previous loop is calculated by the correlation change calculator 8 as a correlation value change ΔR (S
105). Here, the equation for calculating the correlation value change ΔR is as follows: ΔR = R (i) −R (i−1) (i = 1, 2, 3,...)
・) The correlation change evaluation unit 9 determines whether the correlation value R has increased based on the correlation value change ΔR (S10).
6) When the correlation value change ΔR ≧ 0, that is, when the correlation value R increases, the moving average value of the correlation value change ΔR is calculated (S107), and then the value of the parameter set set in the previous loop Is replaced with the newly set parameter set value (S108). On the other hand, the correlation value change △ R
<0, that is, when the correlation value R decreases, the correlation change evaluation unit 9 further determines whether the value φ obtained by the random number calculation satisfies the probability distribution P (△ R) (S1).
09). Note that the probability distribution P (△ R) is calculated when △ R <0, has a value of 0 <P (△ R) <＜, and has a value φ
Is an arbitrary number selected with equal probability in the range [0, 1]. Here, when P (△ R) ≧ φ, that is, when the value φ obtained by the random number calculation satisfies the probability distribution P (△ R), the moving average value of the correlation value change is calculated (S107).
Thereafter, the value of the parameter set set in the previous loop is replaced with the value of the newly set parameter set (S108).

The above probability distribution P (△ R) is proportional to the moving average value of the correlation value change △ R, and the calculation formula is: P (△ R) = 1 / [1 + exp ｛-｛R / (<T > / L
n (K))}]. Here, <T> is a moving average value of the correlation value change ΔR, and K is a constant larger than 1. Correlation value change △
The moving average value of R is calculated each time the value of the parameter set is replaced, and in the next loop, the newly calculated moving average value is calculated using <T <T in the formula of the probability distribution P (△ R).
> Value. By doing so, it is possible to gradually allow the correlation value to converge while probabilistically allowing the case where the correlation value decreases, that is, the situation worsens.

When the parameter set is replaced, the convergence of each parameter, that is, the range of variation of the parameter set value with respect to the number of repetitions is evaluated (S111), and the variation of the set values of all parameters is stabilized below a predetermined value. Then, the loop of the repetitive processing ends, and the value of the newly replaced parameter set is output as the value of the seafloor acoustic characteristic (S112). If the variation of the set values of all the parameters does not become less than the predetermined value, the parameter setting unit 5 newly sets a parameter set and starts the processing of the next loop.

On the other hand, when the correlation change evaluating unit 9 determines that the correlation value change ΔR is reduced and P (△ R) <φ, that is, the value φ does not satisfy the probability distribution P (△ R), , The value of the set parameter set is excluded, and the parameter set set in the previous loop remains (S1).
10). Then, a new parameter set is set by the parameter setting unit 5, and the processing of the next loop is repeated.

FIG. 4 is a diagram showing an example of a change in the correlation value in the present embodiment and a change in the correlation value in the prior art. In any case, the parameter setting value is accepted while stochastically permitting the transition in the direction in which the correlation value decreases, that is, in the direction in which the situation worsens, and the permissible probability decreases as the moving average value decreases. This makes it possible to approach the overall correlation maximum point without staying at the local correlation increase point. Here, the maximum correlation point is a condition for the magnitude of the correlation value to be finally reached.

In this embodiment, as shown in FIG. 5, the calculation of the moving average value of the correlation value change is performed not only when the correlation value is reduced and the probability distribution is satisfied as in the related art, but also when the correlation value is changed. By performing this even when it increases, the moving average value is more easily reduced. Therefore, the decrease in the allowable probability in the direction of deterioration of the correlation value is quickened. As a result, as shown in FIG. 4, the correlation maximum point can be approached with a smaller number of repetitions as compared with the case of the related art.

Further, in the present embodiment, the output of a more accurate estimated value can be performed by setting the variation of the set values of all the parameters to be equal to or less than a predetermined value as an end condition of the repetitive processing. Making it possible.

[0026]

As described above, in the method and the apparatus for measuring the seafloor acoustic characteristics of the present invention, the moving average value of the change calculated for each repetition loop of the correlation values of the measurement data of the seafloor acoustic characteristics and the replica data is used for correlation. Not only when the value decreases and the value obtained by the random number calculation satisfies the probability distribution, but also when the correlation value increases, it is calculated and updated. You can get closer.

Further, according to the present invention, the fluctuation range of the set value of each parameter of the seafloor acoustic characteristic is evaluated for each repetition loop, and when the fluctuation of the set values of all parameters is stabilized to a predetermined value or less, Since the repetition loop is completed and the estimated value of the seafloor acoustic characteristic is output, the accuracy of the estimated value can be improved.

[Brief description of the drawings]

FIG. 1 is a schematic diagram showing a configuration of an embodiment of a submarine acoustic characteristic measuring apparatus according to the present invention.

FIG. 2 is a block diagram illustrating a configuration of an analysis unit in FIG. 1;

FIG. 3 is a flowchart showing an operation of one embodiment of the present invention.

FIG. 4 is a diagram showing a change in a correlation value according to an embodiment of the present invention and a conventional technique.

FIG. 5 is a diagram illustrating a difference between a moving average value calculation method according to an embodiment of the present invention and a moving average value calculation method according to the related art.

FIG. 6 is a flowchart showing the operation of a conventional submarine acoustic characteristic measuring apparatus.

[Explanation of symbols]

 REFERENCE SIGNS LIST 1 sound source 2 receiver array 3 analysis device 4 analysis unit 5 parameter setting unit 6 replica calculation unit 7 correlation calculation unit 8 correlation change calculation unit 9 correlation change evaluation unit 10 moving average calculation unit 11 parameter update unit 12 end determination unit

Continuation of the front page (56) References JP-A-9-127237 (JP, A) JP-A-8-339200 (JP, A) JP-A-8-179029 (JP, A) JP-A-6-347530 (JP, A) (A) JP-A-5-107329 (JP, A) (58) Fields investigated (Int. Cl. ⁶ , DB name) G01S 3/80-3/86 G01S 5/18-5/30 G01S 7/52 -7/64 G01S 15/00-15/96

Claims (7)

(57) [Claims] 1. A receiver for receiving a sound wave transmitted from a sound source and propagated in the sea, measurement data of the seafloor acoustic characteristics obtained by the receiver and replica data of the seafloor acoustic characteristics obtained by propagation calculation And an analysis means for estimating a seafloor acoustic characteristic based on a correlation value with the submarine acoustic characteristic, wherein the analyzing means comprises a multidimensional parameter set related to the submarine acoustic characteristic set for each repetition loop. Accordingly, replica data calculating means for calculating replica data, correlation calculating means for calculating a correlation value between the measurement data obtained by the receiver and the replica data, and the correlation value and calculated in a previous loop. Correlation change calculating means for calculating a difference from a correlation value as a change in the correlation value; and updating or eliminating the value of the parameter set according to the change in the correlation value. Correlation update evaluating means for selecting whether or not, when updating of a parameter set is selected, moving average calculating means for calculating a moving average of the correlation value change, and when updating of the parameter set is selected, Parameter updating means for updating the value of the parameter set to the value of the parameter set set at the time of this loop; and terminating the iterative loop based on a predetermined condition, and setting the value of the parameter set at that time. And an end determination unit that outputs the estimated value as the estimated value. 2. The method according to claim 1, wherein the correlation change evaluation means changes the correlation value so as to increase, or changes the correlation value so as to decrease, and a value obtained by random number calculation allows the correlation value to decrease. The apparatus according to claim 1, wherein when a probability distribution is satisfied, updating of the value of the parameter set is selected. 3. The moving average calculating means calculates a moving average of the correlation value change, and updates the probability distribution proportional to the moving average value according to the calculated moving average value. In the next loop, the correlation change evaluation means uses the updated probability distribution,
The apparatus according to claim 2, wherein whether to update the value of the parameter set is selected. 4. The end determination means evaluates a variation range of the set values of the parameter set with respect to the number of repetitions, and when the variation of the set values of all the parameter sets is stabilized to a predetermined value or less, The apparatus according to claim 1, wherein the repetition loop is terminated. 5. A method based on a correlation value between measured data of seafloor acoustic characteristics obtained by a receiver for receiving a sound wave transmitted from a sound source and propagated in the sea, and replica data of submarine acoustic characteristics obtained by propagation calculation. A submarine acoustic characteristic measuring method for estimating a submarine acoustic characteristic, the replica data calculating step of calculating replica data according to a multidimensional parameter set relating to the submarine acoustic characteristic set for each repetition loop; A correlation calculation step for calculating a correlation value between the measurement data obtained by the wave filter and the replica data; and a correlation change calculation step for calculating a difference between the correlation value and the correlation value calculated in the previous loop as a correlation value change. And a correlation change evaluation step of selecting whether to update or exclude the value of the parameter set according to the change in the correlation value. When the update of the parameter set is selected, a moving average calculation step of calculating a moving average of the change in the correlation value, and when the update of the parameter set is selected, the value of the parameter set is changed at the time of this loop. A parameter updating step of updating to the set parameter set value; and an end determining step of terminating the iterative loop based on a predetermined condition and outputting the parameter set value at that time as the estimated value. And a method for measuring seafloor acoustic characteristics. 6. The correlation change evaluation step, wherein the correlation value changes so as to increase, or the correlation value changes so as to decrease, and a value obtained by random number calculation allows the decrease in the correlation value. The method according to claim 5, wherein when the probability distribution is satisfied, updating the value of the parameter set is selected. 7. The end determination step evaluates a variation range of the set value of the parameter set with respect to the number of repetitions, and when the variation of the set values of all the parameter sets is stabilized to a predetermined value or less, The method according to claim 5, wherein the repetition loop is terminated.