JPH0515971B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Field of Application] The present invention relates to an aircraft noise identification device, and in particular is capable of reducing the influence of noise components contained in aircraft noise.

[Summary of the invention]

According to the present invention, in an aircraft noise identification device using a cross-correlation calculation method, a cross-correlation calculation is executed after noise detection data is subjected to a difference calculation process, thereby making it possible to make the peak representing aircraft noise even steeper.

[Conventional technology]

As a standard for evaluating aircraft noise, the ``Environmental Standards for Aircraft Noise'' (December 27, 1971, Environment Agency Notification)
154), and automatic aircraft noise measurement equipment manufactured in accordance with this standard has been used for some time.

However, with this type of automatic aircraft noise measuring device, it is actually difficult to differentiate aircraft noise from ambient noise such as city noise and evaluate only aircraft noise. Conventionally, the first method is to evaluate the noise received by a microphone as aircraft noise when it exceeds a predetermined set level (Japanese Patent Publication No. 15143/1983), and the second method is to evaluate the noise received by microphones as aircraft noise when the noise exceeds a predetermined level. A method has been proposed (Japanese Patent Publication No. 13169/1983) in which a cross-correlation value is obtained for the output signal of a placed microphone, and aircraft noise is identified based on the position of the peak of the cross-correlation value.

[Problem that the invention seeks to solve]

However, with the first identification method, if the level of the noise other than aircraft noise is higher than the aircraft noise, there is a risk that the noise other than aircraft noise may be mistakenly identified as aircraft noise.This problem can be solved by the second identification method. Improvements are being made by using methods in combination. However, when the noise to be measured contains noise components with strong periodicity in the frequency spectrum, the second identification method may malfunction due to noise based on the periodicity being mixed into the cross-correlation value. be.

Incidentally, the second method for determining the cross-correlation value between two measurement signals is based on the position of the peak of the cross-correlation value for the total volume of the noise, assuming that the noise has the characteristics of white noise. Aircraft noise is identified by focusing on the fact that temporal changes in the noise represent the flight position of the aircraft.

However, in reality, for example, aircraft with propellers contain noise components with strong periodicity and high noise levels in the frequency range of about 100 to 200 [Hz], and street noises include sirens, horns, etc. As a result, a noise component having a frequency between 250 [Hz] and 1 [kHz] with strong periodicity and a high noise level is generated.

In this way, when a noise component with a high noise level for some frequencies in the frequency spectrum is mixed, if the fluctuation period of the noise level of the noise component is sufficiently longer than the cross-correlation calculation period, the cross-correlation value of the aircraft It changes based on movement, and at the same time changes with the period of the noise component.
This results in noise being mixed into the aircraft noise identification results.

It is extremely difficult in practice to determine whether or not it is aircraft noise based on identification results that include such noise, and it is desirable to improve this.

The present invention has been made in consideration of the above points, and
By making it possible to obtain cross-correlation values that are unaffected even when periodic noise components are mixed in the noise signal detected by the microphone, aircraft noise identification performance can be further improved. This paper attempts to propose a method for identifying aircraft noise.

[Means for solving problems]

In order to solve this problem, in the present invention, the noise detection signals S1 _A and S1 _B obtained from the first and second microphones 2A and 2B are sampled and are made up of sampling data that is sequentially arranged as time passes. Noise detection data forming means 9 for forming first and second noise detection data S4 _A and S4 _B , and first and second noise detection data S4
_A and _S4B , respectively, detecting the difference between the sequentially arranged sampling data to form first and second difference data _S5A and _S5B ; A cross-correlation determination means 20 is provided which calculates the cross-correlation of the difference data _S5A and _S5B and identifies aircraft noise based on the peak that occurs in the cross-correlation calculation signal S11.

[Effect]

Prior to performing the cross-correlation calculation by the cross-correlation determining means, the noise detection data S4 _A and S4 _B are subjected to differential calculation processing by the white noise conversion calculation means 15.

As a result, difference data S from which periodic noise components included in noise detection data S4 _A and S4 _B are removed
5 _A and S5 _B , and by performing cross-correlation calculation processing on these, the aircraft noise identification device can effectively avoid the possibility of misjudgment based on noise components with periodicity and high noise level. It can be achieved.

〔Example〕

An embodiment of the present invention will be described in detail below with reference to the drawings.

[1] First Embodiment In FIG. 1, 1 indicates an aircraft noise identification device as a whole, in which noise detection signals S1 _A and S1 _B obtained from upper and lower microphones 2A and 2B are transmitted through amplifier circuits 3A and 3B. The signal is input to analog/digital conversion circuits 4A and 4B.

In this example upper and lower microphones 2
A and 2B are attached to each other by arms 7A and 7B at heights H _A and H _B from the ground surface 5, respectively, on a stand 6 that stands vertically on the ground surface 5, as shown in FIG. They are mounted so as to be arranged on the same vertical line L _V , so that the noise from the aircraft 8 passes through the paths WA and WB connecting the aircraft 8 and the upper and lower microphones 2A and 2B, respectively. 2A and 2B, with a time delay corresponding to the path difference between paths WA and WB (therefore, the difference in elevation angles θ _A and θ _B of paths WA and WB at the upper and lower microphone positions 2A and 2B). Noise detection signals S1 _A and S1 _B that change over time can be taken into the noise detection data forming means 9.

The analog/digital conversion circuits 4A and 4B sample the instantaneous values of the outputs S2 _A and S2 _B of the amplifier circuits 3A and 3B at the sampling period according to the sampling signal S3 given from the system controller 11, and generate noise detection data S4 _A.
and _S4B , the difference detection circuits 12A and 12B are output as the output of the noise detection data forming means 9.
supply to.

The difference detection circuit 12A (and 12B) has an immediately preceding data latch circuit 13A (and 13B) and a subtraction circuit 14A (and 14B). The noise detection data S4 _A (and S4 _B ) to be sent is subtracted by a subtraction circuit 14A.
(and 14B), and the immediately preceding data latch circuit 13A (and 13B) receives the immediately preceding data S4 _AX.
(and S4 _BX ), thus forming the white noise conversion calculation means 15.

The subtracting circuit 14A (and 14B) receives the immediately preceding data S of the immediately preceding data latch circuit 13A (and 13B).
4 _AX (and S4 _BX ) as subtraction input, the current point is calculated as shown in Figure 3A (and Figure 3B).
Difference data S5 _A (and S5 _B ) obtained by subtracting immediately preceding data S4 _AX (and S4 _BX ) from noise detection data S4 _A (and S4 _B ) of t _o is supplied to the cross-correlation determination circuit 20 .

Thus, the difference detection circuit 12A (and 12B)
Sequential sampling points... t _o-1 , t _o , t _o+1 ...
The noise change information extracted from the change in noise detection data S4 _A (and S4 _B ) in ... by a first-order difference calculation is supplied to the cross-correlation determination circuit 20 as difference data S5 _A (and S5 _B ). Become.

The cross-correlation determination circuit 20 includes a cross-correlation calculation circuit section 21 and a determination circuit section 2.

The cross-correlation calculation circuit section 21 calculates the difference data S5 _A
and a time data memory 25 that sequentially captures S5 _B.
A and 25B, and thus the differential data S5 _A and S5 _B held in the time data memories 25A and 25B are converted to correlation data S1 under the control of a control signal S11 given from the system controller 11.
2A and S12B are read out sequentially and multiplied in the multiplication circuit 26. Then, the multiplication data S13 is averaged in the average value circuit 27, and thus the cross-correlation value signal S14 is sent to the cross-correlation calculation circuit section 21.
Send as the output of

In this embodiment, the upper time data memory 25
As shown in FIG. 4A, A is the difference data S obtained at a predetermined number of sampling points, for example, 1064.
Consists of a shift register that sequentially stores 5 _A ,
For example, there is a simultaneous data area M _SIMA in the center that stores 1000 pieces of data, and a previous time data area M _BEFA that stores, for example, 32 pieces of data for the previous time, and 32 pieces of data for the later time. The rear time data area M _AFTA stores the data of
and has.

Similarly, the lower time data memory 25B has a same time data area M _SIMB and a previous time data area.
It has M _BEFA and a later time data area M _AFTB .

Each data in the upper time data memory 25A and the lower time data memory 25B is read out as cross-correlation data S12A and S12B in the following procedure while maintaining a synchronized relationship with each other by the control signal S11 of the system controller 11. and output to the multiplication circuit 26.

That is, the 33rd and 34th data stored in the same time data area M _SIMB of the lower time data memory 25B
...The 1032nd data is read out repeatedly 32 times in response to the previous time data, then read out once in response to the same time data, and then read out repeatedly 32 times in response to the later time data. It will be done. Here, while the same time data area M _SIMB is being read once, 1000 data are read from the upper time data memory 25A.
The pieces of data are sequentially read out while being shifted one data at a time.

For example, the 1st to 1st of the previous time data area M _BEFA
At the timing when the 1000th data is read out in sequence, the 33rd to 1032nd data of the lower time data memory 25B are read out in sequence, multiplied by each other in the multiplication circuit 26, and then outputted to the average value circuit 27 as multiplication data S13. be done.

The average value circuit 27 calculates the average value from the 1000 multiplication data S13 that are sequentially output, and generates a cross-correlation value signal S representing the correlation degree of the first sampling point.
Output as 14.

When the cross-correlation calculation for the first sampling point of the previous time is completed in this way, the second to 1001th sampling points from the upper time data memory 25A are
The 33rd to 1032nd data (that is, data shifted by one data) are read out and repeatedly read out from the lower time data memory 25B.
The multiplied data S13 is sequentially multiplied with the th data.
It is output to the average value circuit 27 as a value. At this time, the average value circuit 27 calculates an average value from the 1000 multiplication data S13 that are sequentially output, and generates a cross-correlation value signal S14 representing the correlation degree of the second sampling point.
Send as.

Similarly, the upper time data memory 25A
32nd based on data from 32nd to 1031st
The cross-correlation value signal S14 for the th sampling point is sent out, and this causes the previous time data area to be
M Finish the cross-correlation operation related to _BEFA .

Thereafter, the upper time data memory 25A reads out the 33rd to 1032nd data stored in the simultaneous data area A _SIMA and supplies it to the multiplication circuit 26, as shown in FIG. 4B. At this time, the average value circuit 27 calculates the average value from the 1000 multiplication data S13 and sends out a cross-correlation value signal S14 representing the degree of correlation of _the 33rd sampling point, and thereby Finish the cross-correlation calculation.

Subsequently, the upper time data memory 25A
As shown in Figure 4C, the later time data area
In relation to M _AFTA , _the 34th, 32nd...
A cross-correlation value signal S14 representing the degree of correlation at the 66th sampling point is sent out.

The cross-correlation value signal S14 obtained from the average value circuit 27 in this way represents the correlation degree for 65 sampling points as shown in FIG. 5, but among these data, the 33rd data is This data represents the degree of correlation in a state where aircraft noise reaches the upper microphone 2A and the lower microphone 2B at the same time, and a sampling point number "0" is assigned to this data.

In addition, the 32nd, 31st...first data is data indicating that the time when the aircraft noise reaches the upper microphone 2A is before the time when it reaches the lower microphone 2B, and this data is sampled. Point number "+1", "+2"..."+32"
Assign.

Further, the 34th, 35th...65th data is data indicating that the time when the aircraft noise reaches the upper microphone 2A is later than the time when the aircraft noise reaches the lower microphone 2B, and this data is sampled. Point number "-1", "-2"... "-
32”.

Thus, as shown in FIG. 5, the sampling numbers "-32", ... "-1", "0", "+1" ...
If we form a cross-correlation value signal S14 in which the data of "+32" are arranged in that order and display this on the monitor graph, the correlation degree peak will be from the sampling point number "0" to the right sampling point number "+".
1", "+2"..."+32", this means that the aircraft
and 2B (therefore, the lengths of the paths WA and WB are almost equal (the elevation angles θ _A and θ _B are also almost equal)), and the state is approaching the position directly above (the following). This represents a state in which the length of the route WA becomes shorter than the length of the route WB.

On the other hand, the peak of the correlation degree is from the sampling point number "0" to the left sampling point number "-".
1", "-2"..."-32", this means that when the aircraft approaches the position directly above the upper and lower microphones 2A and 2B from a far away position, Aircraft noise is at ground level 5 (second
This indicates that the light was reflected by the light source (see the figure) and arrived from the lower microphone 2B side.

The cross-correlation value signal S14 is given to the peak detection circuit 28 of the determination circuit section 22, and the cross-correlation value signal S14
14 reaches the maximum value, the peak detection circuit 28 sends a peak detection signal S15 to the peak address memory 29.

The peak address memory 29 receives the peak detection signal S
Every time 15 is obtained, the sampling point number at that timing is stored and updated, and thus the sampling point number is changed from "+32" to "-32" (5th
The determination circuit 30 uses the sampling point number JDG with the maximum correlation value as the peak address detection signal S16 for the cross-correlation value signal S14 up to the point shown in FIG.
Send to.

Sampling point number of peak value of judgment circuit 30
Based on the JDG, an aircraft noise identification signal S17 representing the flight position of the aircraft is sent as an output of the cross-correlation determination circuit 20.

In this embodiment, the cross-correlation value signal S14 sent out from the cross-correlation calculation circuit section 21 is converted into an analog signal in the digital/analog conversion circuit 31, and then sent out from the cross-correlation determination circuit 20 as a monitor signal S18. By applying this monitor signal S18 to a monitor such as an oscilloscope, a correlation curve as shown in FIG. It is possible to confirm.

In the above configuration, when sampling noise detection data S4 A and S4 _B are obtained from the noise detection signals S1 _A and S1 _B obtained from the upper microphone 2A and the lower microphone 2B in the analog/digital conversion circuits _4A and 4B, this noise Detection data S4 _A and S4 _B are sent to difference detection circuits 12A and 1
2B, the cross-correlation determination circuit 20 calculates the cross-correlation and obtains the aircraft noise identification signal S17 representing the sampling point number corresponding to the peak position of the cross _- correlation value signal S14 _. can.

Therefore, the difference detection circuits 12A and 12
B can extract only the information representing the changes in the noise detection data S4 _A and S4 _B at each sampling time as the difference data S5 _A and S5 _B , so that even if the noise detection signals S1 _A and S
1 Even if _B contains a noise component with periodicity and a high noise level, by excluding the noise component, it can be given the characteristics of white noise with a nearly uniform frequency spectrum (this In this way, the cross-correlation value signal S14 having a stable signal level can be supplied to the cross-correlation determining circuit 20.

For example, the noise of an aircraft with a propeller
When periodic noise components of about 100 to 200 [Hz] are mixed, the period of the periodic noise components is sufficiently longer than the difference calculation period of the difference detection circuits 12A and 12B. Difference data S5
Since the signal levels of the aircraft noise components contained in _A and _S5B are almost equal, they are canceled out by the subtraction operations in the subtraction circuits 14A and 14B, so that they may not be included in the difference data _S5A and _S5B . .

In addition, in the same manner, for noise components of city noise other than aircraft noise, such as sirens and horns of about 250 [Hz] to 1 [kHz], the period of the noise component is determined by the difference detection circuits 12A and 12B.
This effect can be removed by making the period sufficiently longer than the period of the difference calculation in .

Therefore, according to the above configuration, even if the noise detection signals S1 of the upper and lower microphones 2A and 2B
Even if a noise component having periodicity and a high noise level is mixed into _A and _S1B , the flight position of the aircraft can be reliably detected without being substantially affected by the noise component.

Incidentally, regarding aircraft noise that has a noise component with periodicity of about 100 to 200 [Hz], the difference detection circuit 1
If the cross-correlation is determined in the cross-correlation determining circuit 20 without executing the difference calculation process of 2A and 12B, as shown by the curve S14X in FIG.
Based on the periodic noise components contained in aircraft noise, the degree of correlation changes greatly to correspond to the period of the noise components, resulting in multiple peaks or an ambiguous peak waveform. On the other hand, according to the configuration shown in Fig. 1, it is possible to obtain a cross-correlation value signal with a steep peak waveform as information representing the flight position of the aircraft, as shown by the maximum peak sampling point number JDG in Fig. 5. can.

According to experiments, the cross-correlation value signal S14 was measured for aircraft noise when an aircraft with a propeller passes overhead, as shown in Fig. 7B. At the start, the peak waveform representing the position was near sampling point number "0", but when the aircraft passed overhead 3.5 seconds later, it moved significantly in the direction of sampling point number "+32". We were able to obtain measurement information exhibiting a waveform change that returned to the method of sampling point number "0" as the aircraft moved away from the aircraft until 7.5 seconds later.

On the other hand, when similar measurements were made using a configuration in which the difference detection circuits 12A and 12B were omitted, as shown in FIG. 7A, the cross-correlation value signal S1
4X, periodic noise components based on the propeller noise are concentrated on the cross-correlation value signal waveform S14X from the time of start until 7.5 seconds have elapsed, so that the peak value that should represent the aircraft position is We obtained results in which not only the waveforms do not appear steeply, but also two peak waveforms occur at the time when the aircraft passes overhead (ie, after 3.5 seconds to 4.5 seconds).

Based on the experimental results, the difference detection circuits 12A and 1
2B was able to form a cross-correlation value signal S14 that is not affected by periodic noise components included in aircraft noise.

[2] Other Embodiments (1) In the above embodiments, an example was described in which a first-order difference calculation circuit was used as the white noise conversion calculation means 15, but instead of this, an n-order difference calculation circuit, a linear prediction circuit, etc. Various configurations such as filters, partial autocorrelation calculation circuits, and combinations thereof can be applied, and in short, white noise processing means may be applied.

(2) In the above embodiment, the case where the two microphones 2A and 2B are arranged vertically has been described, but the arrangement direction is not limited to this, and the above case can also be achieved by arranging them in various directions, such as horizontally. A similar effect can be obtained.

(3) In the above-mentioned embodiments, a case has been described in which the aircraft noise identification device according to the present invention is constructed using a hardware circuit means, but even if it is constructed using a software calculation means instead, the above-mentioned result will still be achieved. You can get the same effect as in the case.

〔Effect of the invention〕

As described above, according to the present invention, the noise detection signals obtained from a pair of microphones are converted into digital data, and then subjected to white noise processing and cross-correlation processing, so that periodic noise is not reflected in aircraft noise. To reliably identify aircraft noise without being affected even if noise components with strong noise levels are mixed in.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment of the aircraft noise identification device according to the present invention, FIG. 2 is a schematic side view showing the configuration of the microphone device, and FIG. 3 is a curve diagram for explaining the difference calculation. FIG. 4 is a schematic diagram for explaining the cross-correlation calculation, FIGS. 5 and 6 are signal waveform diagrams for explaining the cross-correlation calculation results, and FIG. 7 is a signal waveform diagram for showing the experimental results. DESCRIPTION OF SYMBOLS 1...Aircraft noise identification device, 2A, 2B...Upper and lower microphones, 9...Noise detection data formation circuit, 12A, 12B...Difference detection circuit, 13
A, 13B...Immediate data latch circuit, 14A,
14B... Subtraction circuit, 15... White noise conversion calculation means, 20... Cross-correlation determination circuit, 21... Cross-correlation calculation circuit section, 22... Determination circuit section.

Claims (1)

[Claims] 1. Noise forming first and second noise detection data consisting of sampling data that samples noise detection signals obtained from first and second microphones and sequentially arranges them over time. a detection data forming means; a white noise calculation means for detecting a difference between sequentially arranged sampling data of the first and second noise detection data to form first and second difference data; An aircraft noise identification device comprising: cross-correlation determining means for calculating a cross-correlation between the first and second difference data and identifying aircraft noise based on a peak occurring in the cross-correlation calculation signal.