JP2025018792A - Emergency Vehicle Response System - Google Patents

Abstract

To provide an emergency vehicle response system capable of securely preventing an emergency vehicle from exiting a location and entering a roadway when an emergency vehicle is approaching, without impeding the driving of the emergency vehicle.SOLUTION: An emergency vehicle response system 1 includes: a sensor unit 2 that acquires surrounding situation information regarding a surrounding situation including a location C; an exit suppression unit 4 having a vehicle exit suppression function that suppresses the exit of a vehicle from the location C to a road A at an exit B; an emergency vehicle approach estimation unit 305 for estimating that an emergency vehicle D is approaching the location based on the surrounding situation information acquired by the sensor unit 2; and an exit suppression control unit 306 that causes the exit suppression unit 4 to perform the vehicle exit suppression function based on the estimation by the emergency vehicle approach estimation unit 305.SELECTED DRAWING: Figure 3

Description

The present invention relates to an emergency vehicle response system that is used in construction sites, parking lots, and other locations that are connected to roads via exits and that responds to emergency vehicles traveling on the roads.

Generally, a site such as a construction site is adjacent to a road, and vehicles such as construction-related vehicles enter and exit between the road and the site through a designated entrance/exit.
For example, as described in Patent Document 1, security guards are often stationed at entrances and exits.

On the other hand, when an emergency vehicle traveling on a road approaches the site, it is necessary to reliably stop the vehicle from exiting the site and entering the road so as not to impede the emergency vehicle's travel.
In particular, emergency vehicles frequently enter and exit hospitals and fire stations, and if a vehicle enters the road from a site close to such a facility and hits the emergency vehicle, it could be life-threatening, so it is necessary for vehicles to avoid exiting when an emergency vehicle approaches.

Security guards and drivers of exiting vehicles will make the necessary decisions to prevent approaching emergency vehicles from leaving the premises.

JP 2021-60893 A

Judgments made by security guards, drivers, or other people can easily run into problems, such as not noticing an approaching emergency vehicle due to construction noise, security guards and drivers having difficulty determining the distance of the emergency vehicle from the site, or there being individual differences in judgment among security guards and drivers.

The problem that this invention aims to solve is to provide an emergency vehicle response system that can reliably stop an approaching emergency vehicle from leaving a site and entering a road, without impeding the vehicle's travel.

Means 1 is an emergency vehicle response system used in a site adjacent to a road via an exit and capable of responding to emergency vehicles, characterized in that it includes a sensor unit that acquires surrounding situation information related to the surrounding situation including the site, an exit suppression unit having a vehicle exit suppression function that suppresses the exit of a vehicle from the site to the road at the exit, an emergency vehicle approach estimation unit that estimates that the emergency vehicle has approached the site based on the surrounding situation information acquired by the sensor unit, and an exit suppression control unit that causes the exit suppression unit to exert the vehicle exit suppression function based on the estimation by the emergency vehicle approach estimation unit.

Means 2 is the emergency vehicle response system described in claim 1, characterized in that the surrounding situation information includes sound data of the surroundings including the premises.

The emergency vehicle response system according to claim 2, characterized in that means 3 includes a learning unit that generates a first learning model that learns the relationship between the surrounding situation information and whether the surrounding situation information includes information related to the emergency vehicle, and the emergency vehicle approach estimation unit estimates that the emergency vehicle has approached the premises based on the first learning model generated by the learning unit.

Means 4 is the emergency vehicle response system described in claim 3, characterized in that the emergency vehicle approach estimation unit performs emergency vehicle sound presence estimation to estimate whether the sound data includes a siren sound of the emergency vehicle based on the first learning model generated by the learning unit, and performs emergency vehicle sound approach estimation to estimate whether an emergency vehicle is approaching the premises based on a temporal change in the frequency components of the sound data.

Means 5 is the emergency vehicle response system described in claim 3, characterized in that the learning unit generates a second learning model that learns the relationship between the surrounding situation information and the distance at which the emergency vehicle approaches the premises, and the emergency vehicle approach estimation unit performs an emergency vehicle sound presence estimation that estimates whether the sound data includes the siren sound of the emergency vehicle based on the first learning model generated by the learning unit, and performs an emergency vehicle sound approach estimation that estimates whether the emergency vehicle is approaching the premises based on the second learning model generated by the learning unit.

According to the emergency vehicle response system of the present invention, when it is estimated that an emergency vehicle is approaching the premises based on the surrounding situation information acquired by the sensor unit, the exit prevention unit automatically performs the vehicle exit prevention function, so that when an emergency vehicle approaches, it can reliably stop the vehicle from exiting the premises and prevent the emergency vehicle's travel from being impeded.

In addition, the emergency vehicle approach estimation unit estimates the approach of the emergency vehicle, making it possible to accurately estimate the approach of the emergency vehicle without variations in judgment due to individual differences.

1 is a schematic diagram of the periphery of a site according to an embodiment of the present invention. 1 is a schematic diagram of an emergency vehicle response system showing an embodiment of the present invention. 1 is a block diagram of an emergency vehicle response system showing an embodiment of the present invention. 1 is a flowchart of a machine learning process according to an embodiment of the present invention. 4 is a flowchart of an emergency vehicle response process according to an embodiment of the present invention.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will now be described with reference to FIGS.
It goes without saying that the present invention is not limited to the embodiments.

Figure 1 is a schematic diagram of the periphery of a site according to an embodiment, Figure 2 is a schematic diagram of an emergency vehicle response system according to an embodiment, Figure 3 is a block diagram of an emergency vehicle response system according to an embodiment, Figure 4 is a flowchart of machine learning processing according to an embodiment, and Figure 5 is a flowchart of emergency vehicle response processing according to an embodiment.

As shown in Figures 1 and 2, the emergency vehicle response system 1 is used in a site C adjacent to a road A via an exit B, and responds to an emergency vehicle D traveling on the road A.
In this embodiment, the site C is a construction site that is partitioned off by a temporary enclosure Ca. The exit B also serves as an entrance through which construction-related vehicles can enter and exit.

The emergency vehicle D is a vehicle that runs while emitting a siren, such as an ambulance, a fire engine, or a police vehicle.
In this embodiment, an example of application will be described in which the emergency vehicle D is an ambulance, and the site C is a construction site close to a hospital where the ambulance arrives.

The emergency vehicle response system includes a sensor unit 2, an information processing device 3, and an exit prevention unit 4.

The sensor unit 2 acquires (collects) surrounding situation information, which is information about the surrounding situation including the site C. In particular, the sensor unit 2 can collect surrounding situation information about the situation of vehicles traveling on the road A.

The information processing device 3 determines whether an emergency vehicle is approaching the site C based on the surrounding situation information collected by the sensor unit, and based on this determination, causes the exit prevention unit 4 to perform the vehicle exit prevention function.

The exit prevention unit 4 has a vehicle exit prevention function that prevents vehicles from exiting the site onto the road at the exit, and performs the vehicle exit prevention function based on the estimation and judgment of the information processing device 3.

The sensor unit 2 is installed on the top of a sensor pole 20 erected inside the site C near the exit B, and includes a camera 21, which is an imaging means for taking pictures in a predetermined direction from the exit B on the road A, and a microphone 22, which is a sound collection means for picking up sounds around the exit B.

The surrounding situation information includes video data such as still images and videos captured by the camera 21, and sound data collected by the microphone 22.
The surrounding situation information may include time information and position information of the sound and image pickup.

The microphone 22 and camera 21 are installed inside the site C, which is the construction site, so unlike roads, which require installation permission, there is no need to obtain permission to install them. Therefore, they can be installed for a long period of time, and surrounding situation information for generating the learning model described below can be steadily obtained.

The sound captured by microphone 22 will be affected by the installation conditions, such as reflection and absorption by the terrain and noise caused by construction work, but because the sound captured at the actual site is used, the learning model described below can be generated with high accuracy.

The information processing device 3 is housed in a housing attached to the sensor support 20 and is electrically connected to the sensor unit 2 and the exit prevention unit 4.

The information processing device 3 is a so-called computer that includes a processor that performs data transfer processing, arithmetic processing, etc., various storage units such as memory that stores data, and hardware such as various interfaces. The storage unit stores software data, and the hardware executes specific processing based on the software, thereby functioning as a functional unit described below.

The information processing device 3 is a single computer, but is not limited to this. Each function may be configured by a plurality of computers.
The information processing device 3 may be one or more virtual devices that operate at least some of the functions on a cloud, which is a collection of computer resources. For example, a transmission/reception unit may be provided in an information processing device located on the premises, and a learning unit and a storage unit may be operated on the cloud.

As shown in FIG. 3 , the information processing device 3 includes an information processing unit 30 and a storage unit 31 .
The information processing unit 30 and the storage unit 31 are configured to exchange data, and the information processing unit 30 stores data in the storage unit 31 and reads data from the storage unit 31 .

The information processing unit 30 is electrically connected to the sensor unit 2, the exit prevention unit 4, and the input/output device 5.

The information processing unit 30 includes an acquisition unit 301, a voice analysis unit 302, a learning unit 303, an input/output unit 304, an emergency vehicle approach estimation unit 305, and an exit prevention control unit 306.

The acquisition unit 301 acquires the surrounding situation information acquired by the sensor unit 2. Specifically, the acquisition unit 301 acquires the video data transmitted by the camera 21 and the audio data transmitted by the microphone 22.

The acquisition unit 301 may acquire surrounding situation information such as time information, location information, temperature information, and weather information when sound and image were collected. This information may be acquired from sensors other than the camera 21 and microphone 22, or from a connectable external network.

The acquisition unit 301 stores the acquired surrounding situation information other than the audio data in a linkable manner in the storage unit 31.

The audio analysis unit 302 separates the sound data acquired by the acquisition unit 301 into predetermined frequency bands, and converts the data into sound data of frequency components in which the frequency bands and amplitudes are quantified.

The audio analysis unit 302 stores the converted audio data in the storage unit 31 in a manner that allows the data to be linked to other data.
Incidentally, sound data that is not converted into frequency components may also be stored in the storage unit 31 .

The learning unit 303 acquires the surrounding situation information stored in the memory unit 31, performs machine learning based on the acquired information, generates a learning model M1, and stores the model in the memory unit 31.

The generated learning model M1 learns the relationship between surrounding situation information and whether the surrounding situation information includes information related to an emergency vehicle D, that is, by inputting sound data converted into frequency components, it outputs information on whether the sound includes the siren sound of an emergency vehicle. The learning model M1 corresponds to the first learning model of the present invention.

The learning unit 303 performs machine learning on training data that includes at least data such as sound data converted into frequency components with reference to video data and other information on the surrounding circumstances, and generates a training model M1. Note that the training data may further include video data, for example.

Video data is suitable as training data that can be used to confirm whether or not an emergency vehicle is included. For example, the training data video data can be used to learn while visually viewing the data, making it easier to confirm the presence of an emergency vehicle and enabling accurate learning.

The learning model M1 by the learning unit 303 is, for example, a neural network, and the learning process uses, for example, the backpropagation method, but is not limited to this.

The learning unit 303 may generate a learning model M1 using other surrounding situation information such as time information, position information, temperature information, and weather information of the sound and image pickup as teacher data from the sound data and video data converted into frequency components.

The machine learning process is carried out as shown in Figure 4.

In step 1, the acquisition unit 301 acquires the surrounding situation information acquired by the sensor unit 2. Specifically, the acquisition unit 301 acquires the video data transmitted by the camera 21 and the audio data transmitted by the microphone 22 (S1).

In step 2, the audio analysis unit 302 separates the sound data acquired by the acquisition unit 301 into predetermined frequency bands, and converts the data into sound data of frequency components in which the frequency bands and amplitudes are quantified (S2).

In step 3, the memory unit 31 stores the acquired surrounding situation information, particularly the video data other than the sound data, and the sound data converted into frequency components (S3).

In step 4, the learning unit 303 acquires the surrounding situation information stored and accumulated in the memory unit 31 and performs machine learning based on this (S4).

In step 5, the learning unit 303 generates a learning model M1 that receives the sound data converted into frequency components and outputs information on whether or not the sound includes a siren sound of an emergency vehicle (S5).

In step 6, the learning unit 303 stores the generated learning model M1 in the memory unit 31 (S6).

This machine learning by the learning unit 303 is performed as the surrounding situation information accumulates to a considerable extent in the storage unit 31, and a learning model M1 is generated and stored. In addition, machine learning is performed each time the surrounding situation information is appropriately recorded, and the learning model M1 is updated and stored.

Return to the explanation of Figure 3.

The input/output unit 304 controls the input/output device 5, outputs the processing results of each functional unit to the input/output device 5, and accepts various settings and operations input from the input/output device 5.

The emergency vehicle approach estimation unit 305 uses a learning model to estimate whether an emergency vehicle is approaching the exit of a construction site based on the currently acquired sound data.

The emergency vehicle approach estimation unit 305 acquires the learning model M1 that the learning unit 303 has generated through machine learning at the site C described above and stored in the memory unit 31.
Also, if there is an updated learning model M1, the updated one is obtained.

The emergency vehicle approach estimation unit 305 also acquires the learning model M2 that was generated by machine learning outside the site C and stored in the memory unit 31.

This learning model M2 is generated, for example, through machine learning at other construction sites or in environments with relatively little noise, such as a testing site, and outputs information on whether or not the sound includes the siren sound of an emergency vehicle by inputting sound data converted into frequency components.

The learning model M2 is input, for example, from the input/output device 5 via the input/output unit 304 and stored in the memory unit 31.

The emergency vehicle approach estimation unit 305 uses learning model M1 and learning model M2 to perform emergency vehicle sound presence estimation to estimate whether the currently acquired sound data contains the siren sound of emergency vehicle D.

If the emergency vehicle approach estimation unit 305 estimates that the currently acquired sound data contains the siren sound of an emergency vehicle D, for example, in the emergency vehicle sound presence estimation of either learning model M1 or learning model M2, the emergency vehicle approach estimation unit 305 estimates that the currently acquired sound data contains the siren sound of an emergency vehicle D.

The sound of an emergency vehicle siren contains certain frequency components, so converting it into frequency components is effective and can generate a learning model with a relatively high accuracy rate.

The use of learning model M2 is effective when there is little accumulated information on the surrounding conditions and the accuracy rate of estimations made by learning model M1 at the site C in question is low.

Depending on the situation, at least one of learning model M1 and learning model M2 may be used for estimation. For example, when there is little accumulated information on the surrounding conditions at the site C, only learning model M2 may be used, and then learning model M1 may be used in addition. When a considerable amount of information has been accumulated, estimation may be performed using only learning model M1.

The emergency vehicle approach estimation unit 305 acquires the currently acquired sound data that has been converted into frequency components.
The sound data converted into frequency components may be acquired from the storage unit 31 or may be acquired directly from the voice analysis unit 302 .

The emergency vehicle approach estimation unit 305 performs emergency vehicle sound approach estimation, which estimates whether emergency vehicle D is approaching the site C from the change over time in frequency components of the sound data currently acquired and converted into frequency components in the emergency vehicle sound presence estimation that are estimated to include the siren sound of emergency vehicle D.

For example, the emergency vehicle approach estimation unit 305 monitors the temporal change in a predetermined frequency component (e.g., a high frequency component similar to a siren sound) for sound data currently acquired and converted into frequency components in the emergency vehicle sound presence estimation that is estimated to include the siren sound of emergency vehicle D, and estimates that emergency vehicle D has approached a predetermined distance (e.g., 100 m) from exit B of site C when the increasing trend of the frequency component exceeds a set threshold value.

If the predetermined distance for estimating the approach of emergency vehicle D is set to about 100 m, it will take emergency vehicle D about 7 to 8 seconds to reach exit B at its normal speed, which is preferable since it ensures sufficient time for the exit prevention unit 4 to exert its vehicle exit prevention function and also eliminates the risk of vehicles attempting to exit site C having to wait too long.

In addition, the emergency vehicle approach estimation unit 305 may estimate the approach of an emergency vehicle sound based on a learning model generated by machine learning, similar to the emergency vehicle sound presence estimation.

In this case, too, the learning unit 303 acquires the surrounding situation information stored in the memory unit 31, performs machine learning based on this, generates a learning model M3, and stores it in the memory unit 31.

The generated learning model M3 learns the relationship between surrounding situation information and the distance at which emergency vehicle D approaches exit B of site C, i.e., by inputting sound data converted into frequency components, it outputs information indicating whether emergency vehicle D has approached a specified distance from exit B of site C, for example, the distance from exit B and its direction. Learning model M3 corresponds to the second learning model of the present invention.

The learning unit 303 performs machine learning using training data that includes at least sound data converted into frequency components based on video data and other information about the surrounding environment, and generates a learning model M3.

The video data is suitable as training data that can confirm whether emergency vehicle D is approaching a specified distance from exit B. For example, the system can learn by visually observing certain signs, houses, and other immovable objects that are captured in the video data of the training data and whose distance from exit B is known, making it easier to grasp the distance and enabling accurate learning.

In addition, a learning model M4 similar to learning model M2 generated by machine learning outside the site C where the emergency vehicle sound presence estimation is performed and stored in memory unit 31 may be obtained, and learning model M3 and learning model M4 may be used to estimate whether or not the vehicle is approaching a specified distance from exit B, thereby estimating the approach of the emergency vehicle sound.

When the emergency vehicle approach estimation unit 305 estimates that the emergency vehicle D is approaching the exit B of the construction site in the emergency vehicle sound approach estimation, the exit prevention control unit 306 sends an approach signal to the exit prevention unit 4 indicating that the emergency vehicle D is approaching the exit B of the construction site. In other words, the exit prevention control unit 306 issues a command to the exit prevention unit 4 to exert the vehicle exit prevention function based on the estimation by the emergency vehicle approach estimation unit 305.

Specifically, the exit prevention control unit 306 transmits an approach signal to the control panel 42 of the exit prevention unit 4 indicating that an emergency vehicle is approaching exit B of the construction site.
The exit prevention control unit 306 instructs the control panel 42 of the exit prevention unit 4 to cause the exit prevention unit 4 to perform the vehicle exit prevention function.

When the emergency vehicle approach estimation unit 305 estimates that the emergency vehicle D is approaching the exit B of the construction site in the emergency vehicle sound approach estimation, the exit prevention control unit 306 determines whether the emergency vehicle D has passed through the exit B, and when it determines that the emergency vehicle D has passed through the exit B, it sends to the exit prevention unit 4 a departure signal indicating that the emergency vehicle D has passed through the exit B of the construction site. In other words, the exit prevention control unit 306 issues a command to the exit prevention unit 4 to stop the vehicle exit prevention function.

The exit prevention control unit 306 may, for example, determine that the emergency vehicle D has passed exit B when a predetermined time has elapsed after transmitting an approach signal indicating that the emergency vehicle is approaching exit B of the construction site to the control panel 42 of the exit prevention unit 4. The predetermined time may, for example, be set based on the relationship between a predetermined distance from exit B set in the emergency vehicle sound approach estimation and the set speed of the emergency vehicle D.

The exit prevention control unit 306 may also determine that the emergency vehicle D has passed through the exit B based on the video data from the camera 21 of the sensor unit 2.

Furthermore, the exit prevention control unit 306 may determine that the emergency vehicle D has passed through the exit B when the sound picked up by the microphone of the sensor unit 2 exceeds a peak in intensity or when the frequency changes due to the Doppler effect.

As shown in Figures 1 and 2, the exit prevention unit 4 includes a barrier 40 that can open and close the exit B, an alarm 41 that issues a warning with light and sound near the exit B, and a control panel 42.
The alarm 41 is installed at the top of a support pillar 41a erected adjacent to one side of the exit B.

The control panel 42 is housed in a housing attached to the support 41a and is electrically connected to the circuit breaker 40 and the alarm 41.

The control panel 42 is electrically connected to the information processing device 3 (Figure 3), receives a signal to prevent the vehicle from exiting from the exit prevention control unit 306, and generates a command signal to close the barrier 40 and activate the alarm 41.

The control panel 42 receives a signal from the exit prevention control unit 306 to release the inhibition of the vehicle's exit, and generates a command signal to open the barrier 40 and stop the alarm 41.

The exit prevention unit 4 performs a vehicle exit prevention function that prevents vehicles from exiting through exit B by closing the barrier 40 and activating the alarm 41 to emit light and sound in response to a command signal generated by the control panel 42.

The exit prevention unit 4 opens the barrier 40 and stops the alarm 41 in response to a command signal generated by the control panel 42, thereby stopping the vehicle exit prevention function that prevents vehicles from exiting through exit B.

The input/output device 5 is a device for various inputs and outputs, such as a touch panel, keyboard, and USB memory, controlled by the input/output unit 304, and can input various settings and operations, display the results of calculations, etc.

As shown in FIG. 5, the emergency vehicle response system 1 performs emergency vehicle response processing.

When the emergency vehicle response process starts, in step 11, the emergency vehicle approach estimation unit 305 acquires the learning model M1 that the learning unit 303 has generated through machine learning at site C and stored in the memory unit 31. If there is an updated learning model M1, the updated one is acquired (S11).

In addition, in step 11, the emergency vehicle approach estimation unit 305 also acquires the learning model M2 that was generated by machine learning outside the site C and stored in the memory unit 31 (S11).

Next, in step 12, the acquisition unit 301 acquires surrounding situation information including that acquired by the sensor unit 2. Specifically, the acquisition unit 301 acquires surrounding situation information including video data transmitted by the camera 21 and audio data transmitted by the microphone 22 (S12).

Next, in step 13, the acquisition unit 301 stores the acquired surrounding situation information other than the sound data in the memory unit 31 in a linkable manner, and transfers the sound data of the acquired surrounding situation information to the voice analysis unit 302 (S13).
The surrounding situation information other than the sound data stored in the memory unit 31 is used as data for machine learning by the learning unit 303 (step 3 (S3) and step 4 (S4) in FIG. 4).

Next, in step 14, the audio analysis unit 302 divides the received sound data into predetermined frequency bands and converts the frequency bands and amplitudes into numerically quantified frequency component sound data (S14).

Next, in step 15, the audio analysis unit 302 passes the converted sound data to the storage unit 31 so that it can be linked to other data (S15).

In addition, in step 15, the voice analysis unit 302 passes the converted sound data to the emergency vehicle approach estimation unit 305 (S15).

Next, in step S16, the storage unit 31 stores the converted sound data passed from the voice analysis unit 302 in such a manner that the data can be linked to other data (S16).
The surrounding situation information of the sound data stored in the storage unit 31 is used as data for machine learning by the learning unit 303 (step 3 (S3) and step 4 (S4) in FIG. 4).

Next, in step 17, the emergency vehicle approach estimation unit 305 obtains the converted sound data from the voice analysis unit 302 as the subject of estimation (S17).

Next, in step 18, the emergency vehicle approach estimation unit 305 uses the acquired learning model M1 and learning model M2 to perform emergency vehicle sound presence estimation to estimate whether the currently acquired and converted sound data contains the siren sound of emergency vehicle D (S18).

In step 18, if it is estimated that the sound data does not contain the siren sound of emergency vehicle D (NO), the process ends. If it is estimated that the sound data contains the siren sound of emergency vehicle D (YES), the process proceeds to step 19.

Next, in step 19, the emergency vehicle approach estimation unit 305 performs an emergency vehicle sound approach estimation to estimate whether or not emergency vehicle D is approaching the site C from the change over time in frequency components of the sound data currently acquired and converted into frequency components in the emergency vehicle sound presence estimation that is estimated to include the siren sound of emergency vehicle D (the estimation in step 18 is YES) (S19).

In step 19, if it is estimated that emergency vehicle D is not approaching site C (NO) in the emergency vehicle sound approach estimation, the process ends, and if it is estimated that emergency vehicle D is approaching site C (YES), the process proceeds to step 20.

Next, in step 20, if the emergency vehicle approach estimation unit 305 estimates that emergency vehicle D is approaching exit B of the construction site in the emergency vehicle sound approach estimation (the estimation in step 19 is YES), the exit prevention control unit 306 sends an approach signal to the exit prevention unit 4 indicating that emergency vehicle D is approaching exit B of the construction site (S20).

Next, in step 21, the control panel 42 of the exit prevention unit 4, which has received a proximity signal from the exit prevention control unit 306, closes the barrier 40 and activates the alarm 41. In other words, the exit prevention unit 4 exerts the vehicle exit prevention function of preventing vehicles from exiting through exit B (S21).

Next, in step 22, the exit prevention control unit 306 performs an emergency vehicle passing determination to determine whether or not emergency vehicle D has passed through exit B (S22).

In the emergency vehicle passing judgment in step 22, if it is determined that emergency vehicle D has not passed through exit B of site C (NO), the process returns to before step 22, and if it is determined that emergency vehicle D has passed through exit B of site C (YES), the process proceeds to step 23.

Next, in step 23, if the exit prevention control unit 306 determines in the emergency vehicle passage judgment that emergency vehicle D has passed through exit B of site C (the judgment in step 22 is YES), the exit prevention control unit 306 sends a departure signal to the exit prevention unit 4 indicating that emergency vehicle D has passed through exit B of the construction site (S23).

Next, in step 24, the control panel 42 of the exit prevention unit 4, which has received the departure signal from the exit prevention control unit 306, opens the barrier 40 and stops the alarm 41. In other words, the exit prevention unit 4 releases the vehicle exit prevention function that prevents the vehicle from exiting from exit B (S24), and the process ends.

The emergency vehicle approach estimation unit 305 may determine whether an emergency vehicle has passed in step 22 .
In this case, when the emergency vehicle approach estimation unit 305 determines in the emergency vehicle passage judgment that emergency vehicle D has passed exit B of the site C, the exit prevention control unit 306 sends an exit signal to the exit prevention unit 4 indicating that emergency vehicle D has passed exit B of the construction site.

[Other Modifications]
The present invention is not limited to the above-mentioned embodiment, and may also include the following, for example.

In this embodiment, the barrier 40 and the alarm 41 are used in the exit prevention unit 4, but this is not limited to this and either one may be used.
In addition, other means may be used as long as they have the function of preventing vehicles from exiting from exit B.
For example, a sign that displays a stop sign using LEDs or a traffic light that switches between red and green may be provided at the exit.
For example, a suppression signal such as a warning sound may be sent to a wireless receiver carried by a security guard stationed at an entrance/exit, and based on this command, the security guard may give a signal to suppress the exit of the vehicle.

In this embodiment, the exit prevention unit 4 employs multiple means, namely a barrier 40 and an alarm 41, and when the estimation by the emergency vehicle approach estimation unit 305 is positive, the vehicle exit prevention function of the multiple means is simultaneously exerted, but this is not limited to this, and the vehicle exit prevention function may also be exerted in stages.
By gradually activating the vehicle exit prevention function, the number of notifications within the premises can be increased, and drivers and security guards who are about to exit the premises can be alerted and mentally prepared. In particular, if the vehicle exit prevention function is activated in order from weakest to strongest, for example, the alarm and then the barrier, the effect is remarkable.

For example, when the emergency vehicle approach estimation unit 305 estimates the approach of the emergency vehicle sound, the alarm 41 may be activated when it is estimated that the emergency vehicle D has approached a first predetermined distance (e.g., 150 m) from the exit B of the site C, and the barrier 40 may be closed when it is estimated that the emergency vehicle D has approached a second predetermined distance (e.g., 100 m) that is shorter than the first predetermined distance.

For example, when the emergency vehicle approach estimation unit 305 estimates the approach of the emergency vehicle sound, the alarm 41 may be weakly activated (e.g., illuminated in yellow) when it is estimated that the emergency vehicle D has approached a first predetermined distance (e.g., 150 m) from the exit B of the site C, and may be strongly activated (e.g., illuminated in red) when it is estimated that the emergency vehicle D has approached a second predetermined distance (e.g., 100 m) that is shorter than the first predetermined distance. In addition, the barrier 40 may be closed when it is estimated that the emergency vehicle D has approached a third predetermined distance (e.g., 50 m) that is shorter than the second predetermined distance.

In addition, for example, when the emergency vehicle approach estimation unit 305 estimates that the acquired sound data contains the siren sound of an emergency vehicle D, the alarm 41 may be activated, and when the emergency vehicle approach estimation unit 305 estimates that the emergency vehicle D has approached a predetermined distance from the exit B of the site C, the barrier 40 may be closed.

In this embodiment, the approach of the emergency vehicle sound is estimated based on the sound data acquired by collecting sounds as surrounding situation information, but this is not limited to this, and the approach of the emergency vehicle may also be estimated based on video data acquired as surrounding situation information.

The technical matters in any of the embodiments (including variations; the same applies below) may be applied to other embodiments to serve as examples.

A Road B Exit C Site D Emergency vehicle 1 Emergency vehicle response system 2 Sensor unit 20 Sensor pole 21 Camera 22 Microphone 3 Information processing device 30 Information processing unit 31 Memory unit 301 Acquisition unit 302 Voice analysis unit 303 Learning unit 304 Input/output unit 305 Emergency vehicle approach estimation unit 306 Exit prevention control unit 4 Exit prevention unit 40 Barrier 41 Alarm 42 Control panel 5 Input/output device

Claims (5)

An emergency vehicle response system for use in a site adjacent to a road via an exit, which responds to emergency vehicles,
A sensor unit for acquiring surrounding situation information regarding the surrounding situation including the site;
an exit prevention unit having a vehicle exit prevention function that prevents vehicles from exiting from the site to the road at the exit;
an emergency vehicle approach estimation unit that estimates that the emergency vehicle is approaching the site based on the surrounding situation information acquired by the sensor unit;
an exit prevention control unit that causes the exit prevention unit to exert the vehicle exit prevention function based on an estimation by the emergency vehicle approach estimation unit. The emergency vehicle response system according to claim 1, characterized in that the surrounding situation information includes sound data of the surrounding area including the premises. a learning unit that generates a first learning model that learns a relationship between the surrounding situation information and whether or not the surrounding situation information includes information related to the emergency vehicle;
The emergency vehicle response system according to claim 2 , wherein the emergency vehicle approach estimation unit estimates that the emergency vehicle is approaching the premises based on the first learning model generated by the learning unit. The emergency vehicle approach estimation unit
performing emergency vehicle sound presence estimation to estimate whether the sound data includes a siren sound of the emergency vehicle based on the first learning model generated by the learning unit;
The emergency vehicle response system according to claim 3, further comprising: an emergency vehicle sound approach estimation for estimating whether the emergency vehicle is approaching the premises based on a time change in a frequency component of the sound data. The learning unit generates a second learning model that learns a relationship between the surrounding situation information and a distance at which the emergency vehicle approaches the site,
The emergency vehicle approach estimation unit
performing emergency vehicle sound presence estimation to estimate whether the sound data includes a siren sound of the emergency vehicle based on the first learning model generated by the learning unit;
The emergency vehicle response system according to claim 3, further comprising an emergency vehicle sound approach estimation unit that estimates whether the emergency vehicle is approaching the premises based on the second learning model generated by the learning unit.

Images