JPH1069594A - Vehicle accident detection device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To detect a car accident, that occurs on the road, without requiring any monitor work of a monitoring person by discrimination the collision of a vehicle to an object installed aside the road based on the state of detected variation. SOLUTION: A signal processing part 21(i) [(i) expresses any one from 1 to (n)] has a filtering circuit, etc., and removes a noise from a detecting signal from a G sensor 10(i). An accident discrimination part 22(i) has a level comparator circuit and a discriminating processing circuit or the like and when the amplitude level of the detecting signal from the G sensor 10(i) exceeds a prescribed reference level, the discriminate signal of accident occurrence is outputted through the signal processing part 21(i). When the vehicle collides with a guard rail, the peak level of the detecting signal from the G sensor 10(i) closest to the position of collision is maximized. A control system 50, which receives the discriminate signal of accident occurrence outputted from the accident discriminating part 22(i), recognizes the collision of the vehicle at the position of the guard rail near the G sensor 10(i).

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an apparatus for detecting a vehicle accident occurring on a road, and more particularly, to a collision of a vehicle with a roadside object, a collision between vehicles on the road, and a collision of a vehicle with a falling object on the road. The present invention relates to a vehicle accident detection device for detecting a vehicle accident such as a collision.

[0002]

2. Description of the Related Art In order to detect a vehicle accident occurring on a road, a monitoring camera is generally installed at a point to be monitored on the road, and an image obtained from the monitoring camera is displayed on a monitor. Conceivable. That is, the observer looks at the image displayed on the monitor and determines whether or not a vehicle accident has actually occurred at that point.

[0003]

However, since it is not known when a vehicle accident will occur on the road, the method of detecting a vehicle accident that occurs on the road using a surveillance camera as described above requires the observer to work day and night. Regardless of the vehicle, vehicle accidents that may occur on the road must be constantly monitored. Therefore, it may be possible to detect a vehicle accident at a major point to some extent, but it is difficult to detect a vehicle accident that may occur on a road over a long distance.

Accordingly, an object of the present invention is to provide a vehicle accident detection device capable of detecting a vehicle accident occurring on a road without requiring a monitoring operation by a supervisor.

[0005]

According to a first aspect of the present invention, there is provided a vehicle accident detecting apparatus for detecting a physical fluctuation of a roadside object. And an accident occurrence judging means for judging a collision of the vehicle with the roadside object based on the state of the fluctuation detected by the fluctuation detecting means.

When a vehicle collides with a roadside object, the roadside object physically and rapidly fluctuates greatly. Such a sudden and large fluctuation of the roadside object is detected by the fluctuation detecting means, and the accident occurrence judging means judges a collision of the vehicle with the roadside object based on the detected state of the fluctuation.

[0007] The roadside object is an object installed on the roadside such as a guardrail. In the above vehicle accident detection device, the present invention is described in claim 2 from the viewpoint that a collision accident of a vehicle with a roadside installed object installed over a long distance on the roadside such as a guardrail can be reliably detected. As described above, the fluctuation detecting means is set on the roadside object and has a plurality of sensors for detecting physical vibration of the roadside object, and the accident occurrence determining means is detected by the plurality of sensors. The collision of the vehicle with the roadside object is determined based on the vibration state.

In such a vehicle accident detection device, even if a roadside object such as a guardrail is installed over a long distance on the roadside, a suddenly large roadside object that is partially generated by a collision of the vehicle. Vibration can be detected by any of a plurality of sensors. Then, the accident occurrence determination means determines the collision of the vehicle with the roadside object based on the detected vibration state.

The sensor for detecting the vibration can be constituted by, for example, a so-called G sensor for detecting acceleration.
Furthermore, in the vehicle accident detection device in which a plurality of sensors are set on a roadside object in this manner, from the viewpoint that the collision position of the vehicle with respect to the roadside object can be specified without particularly increasing the number of sensors. According to a third aspect of the present invention, in the vehicle accident detection device in which the plurality of sensors are set on a roadside object, further based on a difference in timing of a change in the state of vibration detected by the plurality of sensors. The vehicle has a position estimating means for estimating a position at which the vehicle has collided with the roadside object.

When a vehicle collides with a roadside object, the vibration caused by the collision propagates sequentially from the collision position to the inside of the roadside object. For this reason, the timing of the state change due to the collision of the vibration detected by the plurality of sensors differs depending on the distance from the collision position. Therefore, the position estimating means estimates the collision position based on the difference in the timing of the change in the state of the vibration detected by the plurality of sensors.

Further, in order to solve the above-mentioned problems, a vehicle accident detection apparatus which can detect not only collision of a vehicle with a roadside object, but also collision between vehicles, collision of a vehicle with a falling object on the road, and the like. According to a fourth aspect of the present invention, there is provided a sound detecting means for detecting a sound generated in a road, and a vehicle detecting a sound on the road based on a state of the sound detected by the sound detecting means. An accident occurrence judging means for judging a collision with an object is provided.

When a vehicle collides with any object, an impact sound is generated due to the collision. In such a situation, the state of the sound detected by the sound detecting means is a state corresponding to the impact sound (for example, a sudden and large change in sound pressure with a special frequency characteristic). Therefore, based on the state of the detected sound, the accident occurrence determination means determines that the vehicle has collided with any object on the road.

From the viewpoint that such a vehicle accident detection device can detect a vehicle collision accident over a long distance on a road, the present invention provides a vehicle
The sound detection means has a plurality of microphones arranged along the road and installed toward the inside of the road, and the accident occurrence judging means detects the road based on the state of output signals from the plurality of microphones. It is determined that the vehicle has collided with any object within the vehicle.

The output signal from the microphone closest to the point where the collision accident occurred has the largest fluctuation. Therefore,
The accident occurrence determination means determines that the vehicle has collided with any object on the road based on at least the state of the output signal from the microphone having the largest fluctuation.

Further, in a vehicle accident detection device in which a plurality of microphones are arranged along a road, from the viewpoint that a collision accident occurrence position on a long distance road can be specified without increasing the number of microphones. According to a sixth aspect of the present invention, there is provided a vehicle accident detection device in which the plurality of microphones are arranged along a road, further comprising a vehicle based on a difference in a state change of an output signal from the plurality of microphones. Has a position estimating means for estimating the position where it has collided with any object.

When the vehicle collides with any object on the road, the output signal of the microphone closest to the collision position has the largest fluctuation (for example, the fluctuation of the sound pressure level). Then, as the distance from the collision position increases, the fluctuation of the output signal of the microphone decreases. Therefore, the position estimating means estimates the collision position of the vehicle based on the difference in the state change of the output signals from the plurality of microphones.

[0017]

Embodiments of the present invention will be described below with reference to the drawings. In the first embodiment, a device for detecting a collision accident of a vehicle with a roadside object, for example, a guardrail is realized.

As shown in FIG. 1, a guardrail 200 having a predetermined length is continuously provided on the road side of a road 100. Each guardrail 200 has an acceleration sensor (hereinafter, referred to as an acceleration sensor).
G sensor) 10 (1), 10 (2),. . . , 1
0 (n) are attached at predetermined intervals. Each G sensor detects a vibration caused by an impact applied to the card rail 200, and outputs a detection signal according to the magnitude of the vibration.

Each G attached to the guardrail 200
The sensors 10 (1), 10 (2),. . . , 10 (n)
Are connected to the processing unit 20, as shown in FIG. The processing unit 20 determines whether there is a collision accident of the vehicle with the guardrail 200 based on the detection signal from each G sensor. The processing unit 20 is housed in, for example, an on-road machine installed on the roadside, and an accident determination signal output from the processing unit 20 is transmitted to the control system 50. The control system 50 transmits, for example, control information based on the accident occurrence determination signal to a vehicle having a subsequent communication function.

The processing unit 20 includes signal processing units 21 (1), 21 provided for each G sensor.
(2),. . . , 21 (n) and the accident determination unit 22
(1), 22 (2),. . . , 22 (n). The signal processing unit 21 (i) (i represents any one of 1 to n; the same applies hereinafter) has a filter circuit and the like, and removes noise from the detection signal from the G sensor 10 (i).
The accident determination unit 22 (i) has a level comparison circuit, a determination processing circuit, and the like, and the G sensor 1 via the signal processing unit 21 (i).
When the amplitude level of the detection signal from 0 (i) exceeds a predetermined reference level, an accident occurrence determination signal is output.

When a vehicle collides with the guardrail 200,
The guard rail fluctuates (vibrates) due to the impact. Then, the peak level of the detection signal from the G sensor 10 (i) closest to the collision position of the vehicle becomes the largest. For example, as shown in FIG. 3A, the waveform of the detection signal that has been processed by the signal processing unit 21 (i) has an amplitude that increases instantaneously when a collision occurs, and then rapidly attenuates. I will be. When the peak level of the instantaneously increased amplitude of the detection signal exceeds the predetermined reference values Th1 and Th2, the level comparison circuit of the accident determination unit 22 (i) outputs the signal from FIG.
A level comparison signal as shown in FIG. Then, based on the level comparison signal, the judgment processing circuit generates and outputs an accident occurrence judgment signal.

The control system 50 which has received the accident occurrence judgment signal output from the accident judgment unit 22 (i) as described above.
Recognizes that the vehicle has collided with the guardrail 200 at a position near the G sensor 10 (i). The accident determination unit 22 (i) directly compares the level of the detection signal from the G sensor 10 (i) that has passed through the signal processing unit 21 (i) with the reference level to determine whether an accident has occurred. The processing in the accident determination unit 22 (i) is not limited to this. For example, the detection signal can be differentiated, and the differentiated signal can be compared with a predetermined reference level to determine whether an accident has occurred.

The processing unit 20 can be configured, for example, as shown in FIG. In this example, when the vehicle collides with a position between the G sensors 10 (i) and 10 (i + 1), the vibration generated by the collision of the vehicle propagates to the G sensors 10 (i) and 10 (i + 1). And the peak level of the detection signal is reduced to the reference levels Th1 and Th2.
May not be reached. Processing unit 20 shown in FIG.
In this case, even in such a case, a vehicle collision accident can be accurately determined.

In FIG. 4, the processing unit 20 includes a signal processing unit 21 (i) and an accident determination unit 22 (i) corresponding to each G sensor 10 (i), as in the above example (see FIG. 2). Have. The processing unit 20 further has a waveform comparator 23 and a final accident determination unit 24. Each accident determination unit 22 (i) determines the presence or absence of a vehicle collision according to the same algorithm as in the above example. Each accident judgment unit 2
When 2 (i) determines that a vehicle collision accident has occurred, the determination signal is provided to the final accident determination unit 24 bypassing the waveform comparator 23. Then, the final accident determination unit 24 that has received the determination signal from one of the accident determination units,
The position of the G sensor that detected the collision and the accident determination result are transmitted to the control system 50.

The waveform comparator 23 is connected to each of the
Reference levels Th1 and Th2 used in the determination process of 2 (i)
A detection signal exceeding a smaller predetermined reference level is specified (for example, a detection signal from the G sensor 10 (i)). Further, in this waveform comparator 23, the adjacent G sensor 10
A detection signal having a waveform similar to the specified detection signal is extracted from the detection signals from (i-1) and 10 (i + 1). The waveform comparator 23 determines the occurrence of a collision accident with the guardrail 200 based on the peak level of the specified detection signal and the peak level of the extracted detection signal (for example, based on the sum thereof).

The result of the judgment is provided from the waveform comparator 23 to the final accident judging section 24. The final accident judging section 24 detects the position of the adjacent G sensor which has detected the collision obtained from the judgment result of the waveform comparator 23. Is transmitted to the control system 50. Further, the processing unit 2
0 can also be configured as shown in FIG.

When the vehicle collides with the guardrail 200 at a position between the two G sensors, the vibration based on the impact propagates through the guardrail 20 and reaches the G sensor closer to the collision position earlier. Therefore, this example shows that multiple G
This is to specify the collision position of the vehicle based on the difference between the accident determination times based on the detection signal from the sensor.

In FIG. 5, the processing unit 20 includes a signal processing unit 21 (i) and an accident determination unit 2 provided for each G sensor 10 (i), similarly to the above example (see FIG. 2).
2 (i). The processing unit 20 further includes an accident time recording unit 25 (i) for recording the judgment time in each accident judgment unit 22 (i), a position specifying unit 26, and an accident position reporting unit 27.

The relationship between the difference .DELTA.T between the accident determination times based on the detection signals from the G sensors 10 (i) and 10 (i + 1) and the position where the accident occurred is obtained experimentally in advance, for example, as shown in FIG. ing. The position identification processing unit 26 stores a map (see FIG. 7) indicating the relationship between the accident determination time difference ΔT and the accident occurrence position. This position specifying processing unit 26
Performs, for example, an accident location specifying process according to the procedure shown in FIG.

In FIG. 6, the position specifying processing unit 26
When it is recognized that an accident has been determined based on the detection signal from the G sensor 10 (i) (S1), an accident determination result based on the detection signal from the adjacent G sensor 10 (i + 1) is input. (S2). Then, the position specifying unit 26 determines whether or not the determination result indicates that an accident has occurred (S3).

When it is determined that the result of the determination indicates that an accident has occurred, the position identification processing unit 26 determines whether or not an accident has occurred based on the detection signal from the G sensor 10 (i). ΔT from the time TGi + 1 at which the accident was determined based on the detection signal from the G sensor 10 (i + 1)
Is calculated (S4). Then, the position identification processing unit 26
The position M corresponding to the difference ΔT is specified with reference to the map shown in FIG. For example, when ΔT = 0, the G sensor 1
The center position between 0 (i) and 10 (i + 1) is ΔT>
When 0, the position closer to the G sensor 10 (i) by a distance corresponding to ΔT from the center position, and when ΔT <0, the G sensor 10 (i) by a distance corresponding to ΔT from the center position. The position near i + 1) is specified.

On the other hand, when the occurrence of an accident is not determined based on the detection signal from the adjacent G sensor 10 (i + 1) (S3, No), the position specifying processing unit 26 further proceeds to the opposite G sensor 10 (i + 1). An accident determination result based on the detection signal from (i-1) is input (S6), and it is determined whether the determination result indicates that an accident has occurred (S7). If it is determined that the result indicates that an accident has occurred, the position identification processing unit 26 determines the times TGi and G when the occurrence of the accident is determined based on the detection signal of the G sensor 10 (i).
Based on the detection signal of the sensor 10 (i-1), a difference ΔT from the time TGi-1 at which the occurrence of the accident is determined is calculated (S
8). Then, the position specifying processing unit 26 refers to the map shown in FIG. 7 and calculates the position M corresponding to the calculated difference ΔT.
Is specified (S9).

The G sensor 10 (i-1) on the opposite side is also
When the determination of the occurrence of the accident has not been made on the basis of the detection signal from (S7, No), the position specifying processor 26 determines that the accident has occurred at the position of the G sensor 10 (i) (S7).
10). When the position identification processing unit 26 identifies the position of the occurrence of the accident as described above, the accident position notification processing unit 27 that has received the information notifies the control system 50 of the position of the accident together with the information of the occurrence of the accident (transmission). ) Processing P1
Execute 0.

As described above, when the location of the accident occurrence is specified based on the difference between the accident occurrence determination times corresponding to the difference between the output timings of the detection signals from the two G sensors, the G sensor attached to the guardrail 200 can be identified. Even if the number is not increased, the position where the accident has occurred can be specified with high accuracy.

Next, a second embodiment of the present invention will be described. The second embodiment realizes a device that can detect not only a collision accident of a vehicle with a roadside object such as a guardrail, but also a collision accident between vehicles that has occurred on the road and a collision accident between a vehicle and a falling object on the road. I have.

As shown in FIG. 8, a plurality of microphones 11 (1), 11
(2),. . . , 11 (n), 11 (n + 1),. . .
Is installed. Each microphone has a predetermined directivity and faces the road 100. And road 1
The sound generated on 00 is detected by each microphone 11 (i), and each microphone 11 (i) outputs a detection signal corresponding to the sound.

Each microphone 1 thus installed
1 (1), 11 (2),. . . , 11 (n) are connected to a processing unit 30, as shown in FIG. The processing unit 30 determines the presence or absence of a vehicle collision accident occurring on the road based on the detection signal from each microphone. This processing unit 30 is also housed in a roadside machine installed on the roadside, as in the example described above, and an accident occurrence determination signal output from the processing unit 30 is transmitted to the traffic control system.

The processing unit 30 includes a signal processing unit 31 provided for each microphone 11 (i).
(I), sound pressure level recording unit 32 (i) and accident determination unit 3
3 (i). Each signal processing unit 31 (i) has a band-pass filter or an FFT processing circuit, and performs signal adjustment processing with frequency characteristics corresponding to a specific collision sound generated in a collision accident. Each sound pressure level recording unit 32 (i)
Records the sound pressure level of the detection signal adjusted by the corresponding signal processing unit 31 (i). Each accident determination unit 33
(I) includes a level comparison circuit, a judgment processing circuit, etc., and outputs an accident occurrence judgment signal when the sound pressure level recorded in the corresponding sound pressure level record 32 (i) exceeds a predetermined reference level. I do.

When the vehicle collides with the guardrail 200, another vehicle, a falling object, or the like, a specific impact sound is emitted. Then, the microphone 11 closest to the collision position of the vehicle
The sound pressure level of the detection signal from (i) becomes the largest.
The waveform of this detection signal that has been processed by the signal processing unit 31 (i) is similar to the waveform shown in FIG. When the sound pressure level obtained from the detection signal exceeds a predetermined reference level, a level comparison signal as shown in FIG. 3B is output from the level comparison circuit of the accident determination unit 33 (i) in the same manner as in the above-described example. Is output. Then, based on the level comparison signal, the judgment processing circuit generates and outputs an accident occurrence judgment signal.

In the accident judging section 33 (i), the sound pressure level of the detection signal from the microphone 11 (i) passed through the signal processing section 31 (i) is directly compared with the reference level to judge whether an accident has occurred. However, this accident determination unit 33
The processing in (i) is not limited to this. For example, the detection signal can be differentiated, and the differentiated signal can be compared with a predetermined reference level to determine whether an accident has occurred.

The processing unit 30 further includes a position specifying unit 34 and an accident position reporting unit 35.
When a vehicle collides with an object at a position between two microphones, the sound pressure levels of the detection signals from the microphones closer to the collision position are compared when the sound pressure levels of the detection signals from the microphones based on the impact are compared. Be larger. The position identification processing unit 34 identifies a collision position of the vehicle with respect to a road object based on a ratio of sound pressure levels of detection signals from the microphones. The relationship between the ratio ΔD of the sound pressure levels of the detection signals from the adjacent microphones 11 (i-1) and 11 (i) and the accident occurrence position is previously experimentally obtained as shown in FIG. 11, for example. I have. The position specifying processing unit 34 stores a map (see FIG. 11) indicating the relationship between the sound pressure level ratio ΔD and the accident occurrence position, and performs the accident position specifying process according to, for example, the procedure shown in FIG.

In FIG. 11, the position identification processing unit 34
Upon recognizing that an accident has been determined based on the sound pressure level of the detection signal from the i-th microphone 11 (i) (S11), the sound pressure level of the detection signal from the adjacent microphone 11 (i-1) is reduced. An accident determination result based on the accident is input (S12). Then, the position specifying processor 34 determines whether or not the determination result indicates that an accident has occurred (S
13).

When it is determined that the result of the determination indicates that an accident has occurred, the position specifying processor 34 determines the sound pressure level Dn of the detection signal from the microphone 11 (i).
Then, a ratio ΔD ΔD = Dn-1 / Dn of the detection signal from the adjacent microphone 11 (i-1) to the sound pressure level Dn-1 is calculated (S14). Then, the position specifying processing unit 34
Specifies the position corresponding to the ratio ΔD with reference to the map shown in FIG. For example, when ΔD = 1, the center position between the microphones 11 (i) and 11 (i−1) is set, and when ΔD> 1, the microphone 11 (i) is separated from the center position by a distance corresponding to ΔD. If ΔD <1, the position closer to the microphone 11 (i) by a distance corresponding to ΔD from the center position is specified.

On the other hand, the next microphone 11 (i-1)
If the occurrence of an accident has not been determined based on the sound pressure level of the detection signal from the (S13, No), the position identification processing unit 34 further proceeds to the microphone 11 (i +
An accident determination result based on the sound pressure level of the detection signal from 1) is input (S16), and it is determined whether the determination result indicates the occurrence of an accident (S17). If it is determined that the result indicates that an accident has occurred, the position identification processing unit 34 determines the sound pressure level Dn of the detection signal from the microphone 11 (i) and the sound pressure level Dn from the opposite microphone 11 (i + 1). The ratio ΔD of the detection signal to the sound pressure level Dn + 1 is calculated (S18). Then, the position specification processing unit 34 specifies a position corresponding to the calculated difference ΔD with reference to the map shown in FIG. 11 (S19).

Note that the microphone 11 (i
If the determination of the occurrence of the accident is not made based on the sound pressure level of the detection signal from (+1) (S17, No), the position identification processing unit 34 determines that the accident has occurred at the position of the microphone 11 (i). (S20).

As described above, when the position identification processing unit 34 identifies the position of the occurrence of the accident, the accident position notification processing unit 35 that has received the information notifies the control system 50 of the position of the occurrence of the accident together with the information of the occurrence of the accident. A process P10 for reporting (transmitting) is executed. As described above, when the location of the accident occurrence is specified based on the ratio of the sound pressure levels of the detection signals from the two microphones, the position of the accident occurrence can be determined without increasing the number of microphones to be installed on the roadside. It can be specified with high accuracy.

In the above-described apparatus according to the first embodiment of the present invention, each G sensor corresponds to a fluctuation detecting means (claim 1), and the processing unit 20 is an accident occurrence determining means (claim 1).
Corresponding to Further, the position specifying processing unit 26 in the processing unit 20 corresponds to a position estimating unit (claim 3).

In the above-described apparatus according to the second embodiment of the present invention, each microphone corresponds to sound detection means (claim 4), and the processing unit 30 corresponds to accident occurrence determination means (claim 4). Further, the position specifying processing unit 34 in the processing unit 30 corresponds to a position estimating unit (claim 6).

[0049]

As described above, according to the present invention described in the claims, the physical fluctuation of the roadside object detected by the fluctuation detecting means and the sound fluctuation detected by the sound detecting means are detected. Based on the sound generated on the road, it detects collisions of vehicles with roadside objects, collisions between vehicles, and collisions of vehicles with falling objects on the road. A vehicle accident occurring on a road can be detected without the need for a vehicle.

[Brief description of the drawings]

FIG. 1 is a diagram illustrating a guardrail installed on a road and a G sensor attached to the guardrail.

FIG. 2 is a block diagram showing a configuration example of a device for detecting a collision accident of a vehicle with a guard rail shown in FIG.

FIG. 3 is a waveform diagram showing a detection signal waveform and a level comparison signal generated in the processing unit shown in FIG. 2;

FIG. 4 is a block diagram showing another configuration example of the processing unit shown in FIG. 2;

FIG. 5 is a block diagram showing still another configuration example of the processing unit shown in FIG. 2;

FIG. 6 is a flowchart illustrating a procedure of processing in a position specifying processing unit of the processing unit illustrated in FIG. 5;

FIG. 7 is a diagram illustrating an example of a map showing a relationship between a difference between accident occurrence determination timings based on detection signals from adjacent G sensors and an accident occurrence position.

FIG. 8 is a diagram showing a microphone installed along the road side of the road.

9 is a block diagram illustrating a configuration example of a device that detects an accident on a road based on a detection signal from the microphone illustrated in FIG. 8;

FIG. 10 is a flowchart illustrating a procedure of a process in a position specification processing unit in the apparatus illustrated in FIG. 9;

FIG. 11 is a diagram showing an example of a map showing a relationship between a ratio of sound pressure levels of detection signals from adjacent microphones and an accident occurrence position.

[Explanation of symbols]

 10 (1) -10 (n) Acceleration sensor (G sensor) 11 (1) -11 (n) Microphone 20 Processing unit 21 (1) -21 (n) Signal processing unit 22 (1) -22 (n) Accident Judgment unit 23 Waveform comparator 24 Final accident judgment unit 25 (1) -25 (n) Accident time recording unit 26 Position specification processing unit 27 Accident position report processing unit 30 Processing unit 31 (1) -31 (n) Signal processing unit 32 (1) -32 (n) Sound pressure level recording unit 33 (1) -33 (n) Accident determination unit 34 Position identification processing unit 35 Accident position notification processing unit 50 Traffic control system

Claims (6)

[Claims] 1. A fluctuation detecting means for detecting a physical fluctuation of a roadside object, and an accident occurrence judging for judging a collision of a vehicle with the roadside object based on a state of the fluctuation detected by the fluctuation detecting means. And a vehicle accident detection device. 2. The vehicle accident detecting device according to claim 1, wherein said fluctuation detecting means includes a plurality of sensors set on said roadside object and detecting a physical vibration of said roadside object. A vehicle accident detection device, wherein the occurrence determination means determines a collision of the vehicle with the roadside object based on a state of vibration detected by the plurality of sensors. 3. The vehicle accident detection device according to claim 2, further comprising: a position for estimating a position at which the vehicle has collided with the roadside installation based on a difference in timing of a change in a state of vibration detected by the plurality of sensors. A vehicle accident detection device having estimating means. 4. A sound detecting means for detecting a sound generated in a road, and an accident which determines that a vehicle has collided with an object on the road based on a state of the sound detected by the sound detecting means. A vehicle accident detection device having a determination unit. 5. The vehicle accident detecting device according to claim 4, wherein said sound detecting means has a plurality of microphones arranged along the road and installed toward the inside of the road. Is a vehicle accident detection device that determines that a vehicle has collided with any object on a road based on the state of output signals from the plurality of microphones. 6. The vehicle accident detecting device according to claim 4, further comprising: position estimating means for estimating a position at which the vehicle has collided with any object based on a difference in state change of output signals from the plurality of microphones. Accident detection device.