JP2003329539A - Method for creating vehicle crash prediction waveform - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To predict and create a crash waveform at the time of a vehicle crash on the other condition based on a crash test using an actual vehicle on a predetermined condition. <P>SOLUTION: The crash waveform at a first predetermined speed of the actual vehicle is found as a time axis waveform. The crash waveform is decomposed to a low frequency component and a high frequency component. With regards to the low frequency component, the crash waveform of the low frequency component at a second predetermined speed lower than the first predetermined speed is obtained from an energy absorption characteristic based on modelling by a predetermined vehicle body deformation model, and found as a time axis waveform based on an energy value corresponding to the low frequency component of the crash waveform at the first predetermined speed. With regards to the high frequency component, a phenomenon in a high frequency region is modelled based on a predetermined vehicle vibration model and decomposed to a plurality of frequency bands. An effective value is found in each frequency band. The frequency component is converted into strength. The crash waveform of the high frequency component at the second predetermined speed is found as a time axis waveform based on an energy value corresponding to the high frequency component of the crash waveform at the first predetermined speed in strength data. Then, the crash waveforms of the high and low frequency components at the second predetermined speed are combined and the crash waveform at the second predetermined speed is created as a time axis waveform. <P>COPYRIGHT: (C)2004,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a vehicle collision prediction waveform generation method for predicting a collision waveform during a vehicle collision and generating it as a time axis waveform.

[0002]

2. Description of the Related Art As is well known, in a vehicle equipped with an airbag system, an airbag sensor is provided to detect a collision of the vehicle and operate (inflate / deploy) the airbag. When developing various characteristics of the impact sensitivity of this airbag sensor, the impact sensitivity is evaluated by attaching the airbag sensor to be tested to an actual vehicle and repeating an actual crash test. This is done by collecting and analyzing data such as deceleration applied to the bag sensor as a collision waveform.

In particular, in recent years, it has been required to be able to suitably adjust the force acting on an occupant due to the operation and deployment of an airbag in accordance with various conditions. That is, it is required that the deployment strength of the airbag be controlled to be switched between three stages of non-deployment and low output and high output depending on the collision speed, the seatbelt wearing state, or the physique of the occupant. Are becoming available.

In order to determine the threshold value for switching the output of the airbag in this way, acceleration or deceleration data (G data) under various conditions is required for the airbag sensor. For example, regarding the vehicle speed, G data under various different vehicle speeds are required. As described above, regarding the airbag sensor, a new evaluation is required in addition to the conventional performance evaluation at the time of high output deployment, and further, it is necessary to perform such evaluation under various conditions. The types and amounts of are increasing dramatically.

[0005]

However, the collision test using the actual vehicle as described above is a destructive test of the vehicle, and the number of tests is naturally limited. Therefore, the number of data obtained is inevitably limited to some extent. This applies not only to the impact sensitivity of the airbag sensor, but also to other various types of collision data to be acquired based on the collision waveform obtained in the collision test on the actual vehicle.

In particular, when the airbag sensor is developed in cooperation with the body development of a new vehicle, the performance evaluation of the airbag sensor must rely on a collision test using an actual vehicle. Practical development activities can be started much later than vehicle body development activities, and there is a problem that the development schedule and period are greatly restricted.

In relation to such a problem, various methods such as a method of predicting the vehicle body deceleration at the time of a collision and a method of performing a collision simulation using a finite element analysis method have been tried, but all of them are put to practical use. The reality is that it has not arrived. still,
For example, Japanese Unexamined Patent Publication No. 7-271290 discloses a vehicle collision simulator and a simulation test method using the same.

As is well known, a collision waveform of a vehicle includes a low frequency component and a high frequency component, and an air bag sensor sets a threshold value for operating an air bag based on both of these components. However, it is difficult to predict and simulate high-frequency components, and the method has not been established.

The present invention has been made in view of the above technical problem, and based on a collision test under a predetermined condition using an actual vehicle, predicts a collision waveform at the time of a vehicle collision under other conditions, and a collision prediction waveform. It is intended to provide a method by which can be generated.

[0010]

Therefore, a method of generating a vehicle collision prediction waveform according to the present invention includes a step of obtaining a collision waveform when a vehicle collides at a first predetermined speed as a time axis waveform, and the collision. And a step of separating the waveform into a low frequency component and a high frequency component by a predetermined conversion means. The low frequency component is based on an energy value corresponding to the low frequency component of the collision waveform at the first predetermined speed, from the energy absorption characteristic of the vehicle body based on modeling by a predetermined vehicle body deformation model. The collision waveform of the low frequency component at the second predetermined speed lower than the first predetermined speed is obtained as the time axis waveform. On the other hand, with respect to the high frequency components, a phenomenon in the high frequency region is modeled by a predetermined vehicle body vibration model, and after being decomposed into a plurality of frequency bands, the effective value in each frequency band is obtained to strengthen the frequency components. And then
Based on the energy value corresponding to the high frequency component of the collision waveform at the first predetermined speed of the strengthened data, the collision waveform of the high frequency component at the second predetermined speed lower than the first predetermined speed is used as a time axis waveform. Ask. Then, a step of synthesizing the collision waveform of the high frequency component and the collision waveform of the low frequency component at the second predetermined speed and generating the collision waveform at the second predetermined speed as a time axis waveform is provided.

In this case, the vehicle body deformation model is preferably an analytical model composed of a spring body and a mass body, and the vehicle body vibration model is preferably a vibrator that generates a predetermined vibration. An analytical model corresponding to each predetermined part of the vehicle body is used. Further, as the converting means, more preferably, a fast Fourier transformer is used.

Further, in the above case, when the high frequency component is decomposed into a plurality of frequency bands, more preferably, the frequency division is performed at equal ratio intervals. Further, in the above case, it is preferable that the repulsion area in the collision is corrected by the correction coefficient based on the reduction rate of the acceleration according to the vehicle type. Furthermore, in the above case, in the low / medium speed region, it is more preferable to perform the correction by the correction coefficient based on the acceleration ratio according to the speed difference between the first and second predetermined speeds.

Furthermore, in the above case, when there is a difference in vehicle body weight, it is more preferable to perform correction using a correction coefficient based on the vehicle body weight ratio. Furthermore, in the above cases, the collision prediction waveform may be used for impact sensitivity evaluation of the airbag sensor.

[0014]

BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, an embodiment of a method for generating a vehicle collision prediction waveform according to the present invention will be described in detail with reference to the accompanying drawings in the case where the embodiment is applied to collision performance evaluation of an airbag sensor. explain. FIG. 1 is an explanatory diagram showing a basic algorithm of the collision prediction waveform generation method according to the present embodiment. Incidentally, a specific flow of steps (steps) based on this algorithm and description of the process contents will be described later.

As shown in FIG. 1, in the present embodiment, first, in order to obtain an actual collision waveform when the target vehicle collides, a predetermined speed (the speed at this time A collision test is performed at a predetermined speed), and pulse data (collision waveform data) of the vibration acceleration G applied to the airbag sensor at that time is sampled as a time axis waveform (see graph).

As described above, the collision waveform obtained in the collision test using the actual vehicle is decomposed by a predetermined conversion means and separated into a low frequency component (graph) and a high frequency component (graph). In the present embodiment, as will be described in detail later, a so-called fast Fourier transform (FFT) is used as a transforming unit used for the decomposition of the collision waveform.
rm) is adopted. This converter preferably performs a fast Fourier transform in a programmed process.

Next, the low frequency component is based on the energy value corresponding to the low frequency component of the collision waveform at the first predetermined speed, from the energy absorption characteristics of the vehicle body based on the simple modeling based on the predetermined vehicle body deformation model. Then, the collision waveform of the low frequency component at the second predetermined speed lower than the first predetermined speed is obtained as a time axis waveform (see graph).

In order to predict and calculate the low frequency component from the energy absorption characteristic of the vehicle body, as schematically shown in FIG.
A vehicle body deformation model including the mass body 3 and the spring body 4 was applied. That is, as shown in FIG. 2, the mass body 3 travels on the traveling path 1 at the first predetermined speed to absorb energy by deforming the vehicle body colliding with the wall surface at the first predetermined speed to absorb energy.
A simple model is created by the phenomenon in which the spring body 4 elastically deforms and absorbs energy when it collides with the wall surface 2 through the two spring bodies 4 at the first predetermined speed, and the energy absorption characteristics at that time are analyzed. The so-called collapsed line diagram is obtained. Note that FIG. 3 corresponds to FIG.
The deceleration (G) characteristic in the low frequency region was calculated as a time axis waveform by the vehicle body deformation model of the above, and the sensor pulse of the low frequency component obtained by decomposing the collision waveform of the actual vehicle with the fast Fourier transformer is obtained. It is equivalent.

The second speed lower than the first predetermined speed
For a predetermined speed, the energy absorption characteristics of the vehicle body are calculated based on the above-mentioned simple modeling by the vehicle body deformation model, and a collapsed diagram thereof is obtained. Next, the energy ratio at both velocities is calculated by the square ratio of the velocities, and based on the calculation result, as shown in the graph of FIG. The collision prediction waveform of the frequency component is obtained as the time axis waveform. That is, based on the energy absorption characteristic of the vehicle body based on the simple modeling by the predetermined vehicle body deformation model, it is lower than the first predetermined speed based on the energy value corresponding to the low frequency component of the collision waveform at the first predetermined speed. The collision prediction waveform of the low frequency component at the second predetermined speed is obtained as the time axis waveform.

On the other hand, with respect to the high frequency component, a phenomenon in a high frequency region is modeled by a predetermined vehicle body vibration model, decomposed into a plurality of frequency bands, and an effective value in each frequency band is obtained to obtain a frequency component of the frequency component. Strengthening is performed, and the collision waveform of the high frequency component at the second predetermined speed lower than the first predetermined speed is based on the energy value corresponding to the high frequency component of the collision waveform at the first predetermined speed of the strengthened data. Is obtained as a time axis waveform.

In order to model the vibration phenomenon of the vehicle body at high frequencies, as shown schematically in FIG. 4 and FIG.
A vibration model in which a plurality of vibrators 5 that generate specific vibrations are made to correspond to respective parts of the vehicle body is applied to the vehicle body deformation model including the spring body 4 and the spring body 4. FIG. 6 shows the deceleration (G) characteristic in the high frequency range obtained as a time axis waveform by the vehicle body vibration model of FIG. 4, and is obtained by decomposing the collision waveform of an actual vehicle with a fast Fourier transformer. It corresponds to a sensor pulse of a high frequency component. Based on this vehicle body vibration model, the high frequency component is decomposed into a plurality of frequency bands, and effective values (RMS values) of the respective components are taken to strengthen them.

Then, energy calculation is performed in the same manner as in the case of the low frequency component, and a collision prediction waveform of the high frequency component at the second predetermined speed lower than the first predetermined speed is obtained as a time axis waveform (FIG. 1). See the graph). That is, at the second predetermined speed lower than the first predetermined speed, based on the energy value corresponding to the high frequency component of the collision waveform at the first predetermined speed of the intensity data based on the modeling by the vehicle body vibration model. The collision prediction waveform of the high frequency component of is obtained as the time axis waveform.

As described above, for the second predetermined speed lower than the first predetermined speed for which the collision test with the actual vehicle is performed,
After obtaining the collision prediction waveforms of the low-frequency component and high-frequency component as time-axis waveforms, and then reconstructing both waveforms, a collision prediction waveform containing both low-frequency and high-frequency components can be obtained (see graph). .

That is, based on a collision test under a predetermined condition (first predetermined speed) using an actual vehicle, a collision waveform at the time of a vehicle collision under another condition (second predetermined speed) is predicted to generate a predicted collision waveform. can do. In other words, for the second predetermined speed, the collision waveform can be obtained as the prediction data without performing a collision test using an actual vehicle. In particular, it becomes possible to predict the high frequency component contained in the collision waveform of the vehicle.

As a result, the number of times of the collision test using the actual vehicle is limited, and the number of data obtained is inevitably limited to some extent, but the collision test cannot be performed. With respect to the condition (1), since it is possible to obtain the collision waveform as the prediction data, it becomes possible to perform a more detailed evaluation of the collision characteristics of the vehicle. In addition, especially when the airbag sensor is developed in cooperation with the development of a new vehicle body, the performance of the airbag sensor can be evaluated without being greatly restricted by the development schedule of the vehicle body. , Will be able to increase the degree of freedom about its development schedule and period.

In this embodiment, various improvements have been made in order to further improve the accuracy of the prediction data (collision prediction waveform) of the collision waveform obtained as described above. As one of the improvement points for improving accuracy, there is an influence due to the characteristics of the filter used when decomposing an actual collision waveform into low frequency components and high frequency components. In the case of performing a filtering process using a general moving average in the past when separating the frequency components, as shown by a broken line in FIG. Was adversely affecting the prediction accuracy of. Therefore, a filter based on the fast Fourier transform is used instead of the filtering process using the moving average. As a result, the phase delay is significantly improved as shown by the solid line in FIG.

Further, when performing prediction in a high frequency region, phase information is lost by strengthening the amplitude of vibration, so the high frequency component is divided into several frequency bands, and this division is performed after prediction. Each frequency band is represented by one frequency, and the original vibration waveform is restored. However, if this frequency band is divided at equal intervals, a particular frequency is emphasized when the waveform is reconstructed. However, there is a drawback that G at the frequency is largely expressed. Therefore, by dividing the frequency band at equal intervals instead of the equal intervals, a specific frequency is not emphasized, and a natural waveform is obtained.

Further, the waveform of the repulsion area after the vehicle body is crushed due to a collision has an influence when the prediction result according to the present invention is used in the occupant simulation. However, since it is a phenomenon after the vehicle body has been crushed, it cannot be dealt with by the above-described algorithm or the idea underlying it. Therefore, as indicated by the broken line in FIG. 8, the accuracy of the predicted waveform in the latter half of the collision phenomenon is not sufficient.

With respect to this point, as a result of conducting a large number of crash tests using various types of actual vehicles, it was found that the phenomenon in this repulsion region shows a similar tendency for the same vehicle type, although it varies considerably depending on the vehicle type. It was Then, in the case of the same vehicle type, it was found that the acceleration decreases in this repulsion region at substantially the same constant rate. Therefore, in the repulsion region, the rate of decrease in acceleration (that is, the slope of the G curve) can be arbitrarily input as the correction coefficient, and the required repulsion correction coefficient is input according to the vehicle type. As a result, the prediction accuracy in the latter half of the collision phenomenon can be significantly improved as shown by the solid line in FIG.

Further, in the low / medium speed range where the vehicle speed is lower than 40 km / h, for example, for the speed of 12.8 km / h, the prediction accuracy of the speed change is not sufficient as shown by the broken line in FIG. There was a difficulty. In this regard, as a result of repeated collision tests using an actual vehicle, as a result, in the speed range in which the vehicle speed is lower than 40 km / h, FIG.
As shown by the solid line, broken line, and alternate long and short dash line in Fig. 3, it was found that the collision energy absorption characteristics of the vehicle body have a considerable speed dependency.

In order to more accurately predict the peak level of the initial G, which is indispensable for determining the deployment / non-deployment of the airbag in the low / medium speed range, a correction coefficient based on the speed dependence coefficient is used. So that the required speed dependence coefficient can be input according to the speed. As a result, as shown by the solid line in FIG. 9, the accuracy of prediction of speed change can be significantly improved.

Regarding the method for obtaining the above velocity dependence coefficient,
A more detailed description will be given with reference to FIGS. 11 to 14.
Crash tests at various set speeds are piled up using actual vehicles of the same vehicle type, and acceleration data in each test is collected as time axis data (acceleration-time curve) (see FIG. 11). This Figure 1
An acceleration-displacement curve is obtained as shown in FIG.

Then, as shown in FIG. 13, the initial displacement region corresponding to the initial G is enlarged, and the acceleration ratio immediately after the collision of the two velocities is calculated. For example, taking Cases 1, 2, and 3 shown in FIG. 13 as an example, the speed difference (ΔV) between two speeds and the acceleration ratio (CSD) are as follows.・ Case 1: ΔV = 40.0-12.6 = 27.4 km
/ H CSD = C / A = 150/300 = 0.5 ・ Case 2: ΔV = 40.0-22.5 = 17.5 km
/ H CSD = C / A = 240/300 = 0.8 ・ Case 3: ΔV = 22.5-12.6 = 9.9 km /
h CSD = C / A = 150/240 = 0.63

Speed difference (ΔV) between the two speeds as described above
The calculation of the acceleration ratio (CSD) and the acceleration ratio (CSD) for various vehicle types resulted in the graph shown in FIG. As can be seen from this graph, the acceleration ratio (CSD) changes so as to decrease as the speed difference (ΔV) increases, and there is a negative correlation between the acceleration ratio (CSD) and the speed difference (ΔV). I knew it was. When the predicted waveform at the second predetermined speed is obtained based on the first predetermined speed, in the low / medium speed region, the acceleration ratio (CS
D) is used as a correction coefficient (speed-dependent coefficient).

Further, when there is a difference in vehicle body weight, as shown by the broken line in FIG. 10, the speed prediction accuracy is not sufficient, and there is a considerable difference in energy absorption characteristics at the time of collision. . Therefore, in order to predict the energy absorption characteristics with higher accuracy, a correction coefficient based on the vehicle body weight ratio can be input, and a required correction coefficient is input according to the weight difference. As a result, as shown by the solid line in FIG. 10, the prediction accuracy can be significantly improved.

The results of prediction using the method of generating a collision prediction waveform whose prediction accuracy has been improved by various corrections as described above are compared with the results of a collision test using an actual vehicle, and the prediction accuracy for each data of acceleration and velocity is compared. Was evaluated. In FIG. 15, a broken line shows prediction data before improvement by correction, and a solid line shows prediction data after improvement by correction. Further, in FIG. 16, the solid line indicates the prediction data after the improvement by the correction.

The deployment timing of the airbag is set to, for example, about 30 ms after a collision in the case of a head-on collision. As is clear from these figures, the prediction accuracy up to the airbag deployment timing is the above-mentioned corrections. The predicted speed from the collision to the airbag deployment timing is close to the experimental value, and the variation of the airbag deployment timing is within a small range (±
Within 10%), a satisfactory prediction accuracy can be achieved.

Next, specific steps (steps) and the flow of the method of generating a collision prediction waveform according to this embodiment will be described with reference to the flowcharts of FIGS. 17 to 19. To start execution of the method according to the present embodiment, first, in step # 1, a collision test using an actual vehicle is performed, and a collision waveform (airbag sensor) when the vehicle collides at a first predetermined speed. Sensor pulse of time axis data G
Get as (t).

Then, in step # 2, the sensor pulse is decomposed by the fast Fourier transform. At this time, for the above-mentioned reason, the sensor pulse is divided into a plurality of (n + 1) frequency bands at equal ratio intervals instead of equal difference intervals. For example, the ratio is divided at an equal ratio of 1.5, and f ₀ = 100 Hz, f
₁ = 150 Hz, f ₂ = 225 Hz, ... In the present embodiment, frequencies below 100 Hz are low frequency components, and frequencies above 100 Hz are high frequency components.
That is, the sensor pulse is decomposed into low frequency components ( _{0 to} f ₀ ) and high frequency components (f _{1 to} f _n ) by the process of step # 2.

Then, the low frequency component subroutine (step # 3L) or the high frequency component subroutine (step # 3H) is executed according to the frequency band. That is, the low frequency components ( _{0 to} f ₀ : G0 (t)) are shown in FIG.
The subroutine shown in is executed. First, step # 1
The low frequency component G0 (t) of the sensor pulse is called by 1 and double integrated to obtain the displacement X (step # 1).
2). Then, the G0 (t) -X diagram is obtained (step # 13).

Next, a predetermined speed lower than the collision speed (first predetermined speed) in the actual vehicle test is set as the second predetermined speed, and the energy ratio at both speeds is set by the square ratio of the speeds as described above. And G01 at the second predetermined speed
The (t) -X diagram is calculated (step # 14).

Then, based on the calculation result, a collision prediction waveform of a low frequency component at a low speed (second predetermined speed) is created as a time axis waveform (G01 (t)) (step #
15). After that, as described above, the repulsion area correction and the speed dependence correction (steps # 16 and # 17) are performed, and the collision predicted waveform (G02 ′) after the correction for the low frequency component is corrected.
(T)) is obtained as a time axis waveform.

On the other hand, high frequency components (f _{1 to} f ₂ : G1)
_{(T), f 2 ~f 3} : G2 (t), f 3 ~f 4: G3
_{(T), ..., f n} ~f n + 1: Gn (t), for), the subroutine shown in FIG. 19 for each frequency band is executed. First, in step # 21, the frequency component Gm (t) of the frequency band of the sensor pulse is called (however,
1 ≦ m ≦ n), the effective value (RMS) in the frequency band is obtained to strengthen the frequency component (step # 22).
~ # 24).

That is, first, Gm (t) is squared (2
Multiplied value Gm1 (t): step # 22), and this is averaged (average value Gm2 (t): step # 23). In addition, in this averaging, the difference dt of the time t is, for example, 0.1 m.
s (dt = 0.1 ms) and k = 100. Then, the square root of the average value is calculated (Gm3
(T): Step # 24).

For Gm3 (t) thus obtained, Gm3 'at a low speed (second predetermined speed) is calculated based on the same energy calculation as in the case of the low frequency component.
Calculate the (t) -X diagram (steps # 25- # 27),
The collision prediction waveform of the high frequency component at the second predetermined speed is created as a time axis waveform (Gm3 ′ (t)) (step #
28). After that, as described above, the waveform is inversely converted (step # 29).

As described above, after obtaining the collision prediction waveform at the second predetermined speed for all the frequency bands, these pulses (waveforms) are reconstructed (FIG. 17: step #).
4). And for the waveform obtained by this reconstruction,
The above-mentioned weight difference correction is performed (step # 5), and finally,
A collision prediction waveform (prediction sensor pulse) at the second predetermined speed including both low frequency and high frequency components is obtained (step # 6). That is, based on a collision test under a predetermined condition (first predetermined speed) using an actual vehicle, it is possible to predict a collision waveform at the time of a vehicle collision under another condition (second predetermined speed) and generate a collision prediction waveform. You can do it.

The present invention is not limited to the above-mentioned embodiments, and within the scope of the gist of the present invention,
It goes without saying that various improvements and design changes are possible.

[0048]

According to the vehicle collision prediction waveform generating method of the present invention, another condition (second predetermined speed) is determined based on a collision test using an actual vehicle under a predetermined condition (first predetermined speed). A collision prediction waveform can be generated by predicting a collision waveform at the time of a vehicle collision in. In other words, for this second predetermined speed, without carrying out a collision test using an actual vehicle,
The collision waveform can be obtained as prediction data. In particular,
It becomes possible to predict the high frequency component included in the collision waveform of the vehicle. As a result, the number of collision tests using the actual vehicle is limited, and the number of data obtained is inevitably limited to some extent. Since it is possible to obtain the collision waveform as the prediction data, it becomes possible to perform a more detailed evaluation of the collision characteristics of the vehicle.

In this case, an analytical model composed of a spring body and a mass body is used as the vehicle body deformation model for the low frequency component, and the vibrator that generates a predetermined vibration is used as the vehicle body vibration model for the high frequency component. Since an analytical model that corresponds to each predetermined part of the vehicle body is used, a well-established analysis method can be applied when predicting the energy characteristics of the vehicle body.

Further, in the above case, when the vehicle collides at the first predetermined speed, the fast Fourier transformer is used as the transforming means used when decomposing the collision waveform into the low frequency component and the high frequency component. Compared to the case where the filtering process using the moving average is performed, the phase delay in the low frequency region can be suppressed, and the prediction accuracy of the low frequency component can be further enhanced.

Furthermore, in the above case, when the high frequency component is decomposed into a plurality of frequency bands, frequency division is performed at equal ratio intervals, so that the frequency bands are divided at equal difference intervals. As described above, when the waveform is reconstructed, a specific frequency is not emphasized and G at that frequency is not significantly displayed, and a natural waveform can be obtained.

Furthermore, in the above case, the repulsion area in the collision is corrected by the correction coefficient based on the reduction rate of the acceleration according to the vehicle type, so that the prediction accuracy in the latter half of the collision phenomenon is improved. Can be increased.

Furthermore, in the above case, in the low / medium speed range, the correction coefficient based on the acceleration ratio corresponding to the speed difference between the first and second predetermined speeds is used to perform correction. The accuracy of speed change prediction in the low / medium speed range can be improved.

Furthermore, in the above case, when there is a difference in vehicle body weight, correction is made by a correction coefficient based on the vehicle body weight ratio, so that even if there is a vehicle body weight difference, energy absorption is possible. The characteristics can be predicted with higher accuracy, and the prediction accuracy of the collision waveform can be improved.

Furthermore, in the above case, by using the above-mentioned collision prediction waveform for the impact sensitivity evaluation of the airbag sensor, even when the airbag sensor is developed in cooperation with the development of the body of the new vehicle. The performance of the airbag sensor can be evaluated without being greatly restricted by the development schedule of the vehicle body, and the degree of freedom regarding the development schedule and period can be increased.

[Brief description of drawings]

FIG. 1 is an explanatory diagram showing a basic algorithm of a collision prediction waveform generation method according to an embodiment of the present invention.

FIG. 2 is an explanatory diagram schematically showing modeling of a low frequency component by a vehicle body deformation model.

FIG. 3 is an explanatory diagram showing a G characteristic for a low frequency component according to the vehicle body deformation model.

FIG. 4 is an illustration schematically showing modeling of a high frequency component by a vehicle body vibration model.

FIG. 5 is an illustration schematically showing a collision state of the vehicle body vibration model.

FIG. 6 is an explanatory diagram showing a G characteristic for a low frequency component according to the vehicle body vibration model.

FIG. 7 is an explanatory diagram showing an example of improving a phase delay in a low frequency region by optimizing a filter when a collision waveform is decomposed.

FIG. 8 is an explanatory diagram showing an example of improvement in prediction accuracy in a repulsive area.

FIG. 9 is an explanatory diagram showing an example of improvement in prediction accuracy by correction based on a speed dependence coefficient.

FIG. 10 is an explanatory diagram showing an example of improving the prediction accuracy by correction based on the weight difference of the vehicle body.

FIG. 11 is an explanatory diagram showing an example of acceleration-time data used for calculation of a velocity dependence coefficient.

FIG. 12 is an explanatory diagram showing an example of acceleration-displacement data based on the acceleration-time data.

FIG. 13 is an explanatory diagram showing an enlarged main part of the acceleration-displacement data.

FIG. 14 is an explanatory diagram showing a correlation between a speed difference and a speed dependence coefficient.

FIG. 15 is an explanatory diagram showing an example of acceleration-time data before and after prediction accuracy improvement by correction.

FIG. 16 is an explanatory diagram showing an example of speed-time data before and after prediction accuracy improvement by correction.

FIG. 17 is a flowchart for explaining specific steps of the collision prediction waveform generation method according to the present embodiment.

FIG. 18 is a flowchart illustrating a low frequency component subroutine.

FIG. 19 is a flowchart illustrating a high frequency component subroutine.

[Explanation of symbols]

3 ... Mass 4 ... Spring body 5 ... Transducer

   ─────────────────────────────────────────────────── ─── Continued front page    (72) Inventor Takehiro Narukawa             3-1, Shinchi, Fuchu-cho, Aki-gun, Hiroshima Prefecture Mazda             Within the corporation (72) Inventor Hirotake Shibasaki             3-1, Shinchi, Fuchu-cho, Aki-gun, Hiroshima Prefecture Mazda             Within the corporation

Claims (8)

[Claims] 1. A step of obtaining a collision waveform as a time axis waveform when a vehicle collides at a first predetermined speed; a step of separating the collision waveform into a low frequency component and a high frequency component by a predetermined conversion means; Regarding the low frequency component, based on the energy absorption characteristic of the vehicle body based on modeling by a predetermined vehicle body deformation model, based on the energy value corresponding to the low frequency component of the collision waveform at the first predetermined velocity, from the first predetermined velocity The step of obtaining the collision waveform of the low frequency component at the second predetermined speed which is also low as a time axis waveform, and modeling the phenomenon in the high frequency region by the predetermined vehicle body vibration model for the above high frequency component, and After decomposing, the effective value in each frequency band is obtained to strengthen the frequency component, and corresponds to the high frequency component of the collision waveform at the first predetermined speed of the strengthened data. A step of obtaining a collision waveform of a high frequency component at a second predetermined speed lower than the first predetermined speed as a time axis waveform based on the energy value; and a collision waveform of the high frequency component at the second predetermined speed and a low frequency component. And a step of synthesizing a collision waveform and generating the collision waveform at the second predetermined speed as a time axis waveform. 2. The vehicle collision prediction waveform generating method according to claim 1, wherein the vehicle body deformation model is an analytical model composed of a spring body and a mass body, and the vehicle body vibration model produces a predetermined vibration. A method for generating a vehicle collision prediction waveform, which is an analytical model in which a generated oscillator is associated with each predetermined part of a vehicle body. 3. The method for generating a vehicle collision prediction waveform according to claim 1 or 2, wherein the converting means is a fast Fourier transformer. 4. The vehicle collision according to claim 1, wherein when the high frequency component is decomposed into a plurality of frequency bands, frequency division is performed at equal ratio intervals. Predicted waveform generation method. 5. The vehicle according to claim 1, wherein the rebound area in the collision is corrected by a correction coefficient based on a reduction rate of acceleration according to a vehicle type. Generation method of collision prediction waveforms in. 6. In the low / medium speed range, correction is performed by a correction coefficient based on an acceleration ratio corresponding to a speed difference between the first and second predetermined speeds. A method of generating a vehicle collision prediction waveform according to any one of 1. 7. The vehicle collision prediction according to claim 1, wherein when there is a difference in vehicle body weight, a correction coefficient based on a vehicle body weight ratio is used for correction. Waveform generation method. 8. The collision prediction waveform is used for impact sensitivity evaluation of an airbag sensor.
A method of generating a vehicle collision prediction waveform according to claim 7.