JP2024004717A - Abnormality information output device and abnormality information output method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an abnormality information output device capable of detecting an abnormality by selectively using a microphone arranged at an appropriate position around a vehicle.

SOLUTION: An abnormality information output device 10a comprises: a vehicle circumference information acquisition section 11 that acquires vehicle circumference information including a plurality of voices collected; a vehicle travel information acquisition section 12 that acquires information on a traveling state of a vehicle; a voice selection section 13 that selects at least one voice of the plurality of voices on the basis of the information on the traveling state of the vehicle; a voice recognition section 14a that recognizes the selected voice and determines whether or not a voice indicating an abnormality occurring to the vehicle is included; and an abnormality information output section 15 that outputs the abnormality occurring to the vehicle as abnormality information on the basis of the determination. The voice recognition section 14a includes a voice recognition dictionary selection section 141 that selects a voice recognition dictionary used for recognition on the basis of the information on the traveling state of the vehicle or the selected voice.

SELECTED DRAWING: Figure 1

COPYRIGHT: (C)2024,JPO&INPIT

Description

The present invention relates to an abnormality information output device and an abnormality information output method.

A microphone placed in a predetermined position around the driver's seat in the vehicle detects the voice, detects that it contains a pre-registered keyword, and sends emergency information and location information to the set emergency call destination. , techniques for transmitting audio and moving images are known. (For example, Patent Document 1)

Japanese Patent Application Publication No. 2016-133963

The communication type drive recorder described in Patent Document 1 transmits information including a moving image to an emergency contact when it is detected that a pre-registered keyword is included in the detected voice inside the vehicle. It is. A microphone that collects sound is placed at a predetermined position around the driver's seat, and the vehicle occupant can make an emergency call by uttering a specific keyword. However, the optimal location for a microphone for detecting abnormal situations occurring in a vehicle is not necessarily around the driver's seat. When detecting an abnormal situation occurring in a vehicle using audio, it is desirable to detect the abnormality by selectively using a microphone placed at an appropriate position for each assumed abnormal situation.

The present invention has been made in view of the above, and collects a plurality of voices from microphones placed at a plurality of positions around the vehicle, selects an appropriate voice, and performs voice recognition processing. An object of the present invention is to provide an abnormality information output device and an abnormality information output method that can detect an abnormality that has occurred and perform control to record an image when the abnormality occurs.

In order to achieve the above object, an abnormality information output device according to the present invention includes a vehicle surrounding information acquisition section that acquires vehicle surrounding information including a plurality of collected sounds, and a vehicle running information that acquires information regarding the running state of the vehicle. an acquisition unit; a voice selection unit that selects at least one voice from the plurality of voices based on information regarding the running state of the vehicle; and a voice selection unit that recognizes the selected voice and determines whether the selected voice includes a voice indicating an abnormality occurring in the vehicle. and an abnormality information output unit that outputs an abnormality that occurs in the vehicle as abnormality information based on the determination. The present invention is characterized by comprising a speech recognition dictionary selection unit that selects a speech recognition dictionary to be used for recognition based on the following.

In order to achieve the above object, an abnormality information output method according to the present invention includes a step in which a computer acquires vehicle surrounding information including a plurality of collected sounds, a step in which a computer acquires information regarding the running state of the vehicle, a step of selecting at least one voice from among the plurality of voices based on information regarding the driving condition of the vehicle; and selecting a voice recognition dictionary to be used for the recognition based on the information regarding the traveling condition of the vehicle or the selected voice; Recognizing the selected voice using the selected voice recognition dictionary and determining whether or not the voice indicating an abnormality occurring in the vehicle is included, and based on the determination, determining the abnormality occurring in the vehicle as abnormality information. Execute the step to print the output.

According to the present invention, it is possible to provide an abnormality information output device and an abnormality information output method that can detect an abnormality by selectively using microphones placed at appropriate positions around a vehicle.

1 is a block diagram showing an example of the configuration of an abnormality information output device according to a first embodiment. FIG. 5 is a flowchart illustrating an example of a processing procedure for outputting abnormality information according to the first embodiment. 3 is a block diagram showing an example of the configuration of an abnormality information output device according to a second embodiment. FIG.

The present invention will be described below through embodiments of the invention, but the claimed invention is not limited to the following embodiments. Furthermore, not all of the configurations described in the embodiments are essential as means for solving the problem. For clarity of explanation, the following description and drawings are omitted and simplified as appropriate.

<Embodiment 1>
Embodiment 1 of the present invention will be described using FIG. 1. FIG. 1 is a block diagram showing an example of the configuration of an abnormality information output system 1 according to the first embodiment. The abnormality information output system 1 includes an abnormality information output device 10, a sound collection section 20, an imaging section 30, sensors 40, and a recording section 50.

The sound collection unit 20 is a microphone that is arranged at various positions such as inside the vehicle, outside the vehicle, and around the vehicle, and collects sounds around the vehicle and voices of people. The sound collection unit 20 may be placed in various locations inside the vehicle, such as near the driver's seat, near the rear seat, inside the luggage compartment, and in the engine room. The sound collection unit 20 may be placed in various locations outside the vehicle, such as at the bottom of the vehicle, near the vehicle bonnet or door mirror, or at the front or rear of the vehicle. The sound collection unit 20 may be manually placed around the vehicle by a vehicle occupant when the vehicle is parked, or may be fixedly placed in a parking lot, for example, at the same time as a surveillance camera. It may also be placed in contact with the surrounding road surface.

The imaging unit 30 is a camera that is disposed inside or outside the vehicle and captures images of the surroundings of the vehicle. An image of a range including the inside of the vehicle may be captured. At this time, the image may be either a still image or a moving image. The imaging unit 30 may include a plurality of cameras. The imaging unit 30 may be a surveillance camera fixedly placed in the parking lot.

The sensors 40 are sensors that sense the running state of the vehicle. The sensors 40 may be a single sensor such as a vehicle speed pulse sensor, a vehicle speed sensor, an acceleration sensor, a door opening/closing sensor, or a seating sensor, or may be configured as a sensor group including a plurality of these sensors. . Furthermore, the above-mentioned sound collection section 20 and imaging section 30 may be used as an audio sensor or an image sensor, and the vehicle running information acquisition section described below may determine the stationary state of the vehicle from the acquired images and audio.

The recording unit 50 records information regarding the vehicle under the control of a recording control unit 16, which will be described later. The recording unit 50 is, for example, a storage medium such as a memory card, a nonvolatile memory such as a flash memory built into the abnormality information output device 10, or a storage device such as a hard disk drive or SSD (Solid State Drive). Note that the recording unit 50 is not necessarily an essential component of the abnormality information output device 10, and may be omitted.

The abnormality information output device 10 includes a vehicle surrounding information acquisition section 11 , a vehicle running information acquisition section 12 , a voice selection section 13 , a voice recognition section 14 , and an abnormality information output section 15 .

The vehicle surrounding information acquisition unit 11 acquires vehicle surrounding information including a plurality of sounds collected by the sound collection unit 20. More specifically, the vehicle surrounding information acquisition unit 11 acquires a plurality of sounds collected by a plurality of sound collection units 20 arranged at various positions such as inside the vehicle, outside the vehicle, and around the vehicle. In addition to acquiring audio, the vehicle surrounding information acquisition unit 11 acquires images captured by the imaging unit 30 placed on or around the vehicle, and information regarding the current position of the vehicle acquired by a position information acquisition unit (not shown). You may. The vehicle surrounding information acquisition section 11 may be connected to a plurality of sound collection sections 20 or an imaging section 30 by wire or wirelessly, or the vehicle surrounding information acquisition section 11 may include a plurality of sound collection sections 20 or an imaging section 30. It may also be configured.
The vehicle surrounding information acquisition unit 11 outputs the acquired vehicle surrounding information including a plurality of sounds to the sound selection unit 13.

The vehicle running information acquisition unit 12 acquires information regarding the running state of the vehicle from sensors 40 provided in the vehicle. The information regarding the vehicle running state obtained here refers to information indicating the state of the vehicle, such as whether the vehicle is running, stationary, or parked, or materials for determining the state of the vehicle. Examples of the information include a vehicle speed pulse signal, acceleration sensor output, engine ON/OFF information, door opening/closing information, and seating sensor information. In addition to this, various information regarding the state of the vehicle may be acquired, such as the traveling speed of the vehicle and whether the vehicle is accelerating or decelerating. Further, the vehicle running information acquisition unit 12 may use the acquired information regarding the vehicle running state as it is to indicate the running state of the vehicle, or may determine the running state of the vehicle by combining a plurality of pieces of acquired information. It may also be used to indicate the running state.
In addition to acquiring information about the running state of the vehicle from sensors 40, the vehicle running information acquisition unit 12 receives information about the running state of the vehicle or materials for determining the state of the vehicle from the vehicle via a CAN (Controller Area Network). You may also obtain information that becomes . The information acquired from the vehicle via CAN includes, for example, shift position information, handbrake status information, engine ON/OFF information, vehicle speed information, door lock information, and the like.
The vehicle running information acquisition unit 12 outputs the acquired or determined information regarding the running state of the vehicle to the audio selection unit 13 .

The vehicle driving information acquisition unit 12 acquires information such as that the vehicle's shift position is in the parking position, that the handbrake is pulled, and that the vehicle's engine is in the OFF state, and the vehicle is in the OFF state. Determine whether the vehicle is in a parked state. In addition, information such as whether there are no occupants in the vehicle, whether the vehicle doors are locked from outside the vehicle, and whether the vehicle speed is zero for a predetermined period of time or more is obtained. It may also be used to determine whether the vehicle is in a parked state.

The audio selection unit 13 receives the information regarding the driving state of the vehicle output by the vehicle driving information acquisition unit 12, and selects at least one of the plurality of sounds acquired by the vehicle surrounding information acquisition unit 11 based on the driving state of the vehicle. Select one voice.
For example, when the vehicle is parked, the audio selection unit 13 selects the audio collected by the sound collection unit 20 disposed outside the vehicle from among the plurality of sound collection units 20. More preferably, among the plurality of sound collection units 20 disposed outside the vehicle, the sound collected by the sound collection unit 20 disposed closer to the lower part of the vehicle is selected.
The voice selection unit 13 outputs the selected voice to the voice recognition unit 14.

The audio selection unit 13 selects the audio collected by the sound collection unit 20 disposed inside the vehicle from among the plurality of sound collection units 20 when the vehicle is traveling at a vehicle speed equal to or higher than a predetermined value. However, it is preferable not to select the sound collected by the sound collection unit 20 disposed outside the vehicle. When the vehicle is traveling at a speed higher than a predetermined value, wind noise is generated outside the vehicle, and there is a risk that the sound indicating an abnormality occurring in the vehicle will be buried in the wind noise, making it impossible to make an accurate determination.
When the vehicle is stopped in a non-parked state or is traveling at a speed lower than a predetermined value (for example, 10 km/h), the audio selection unit 13 selects a sound pickup device located outside the vehicle. Both the sound collected by the section 20 and the sound collected by the sound collection section 20 placed indoors may be selected.

If the vehicle is stopped or traveling at a speed less than a predetermined value, if the vehicle is in the process of parking the vehicle, if the vehicle is avoiding pedestrians, or if the vehicle is on a narrow road. It is assumed that this is a case where slow driving is required, such as when passing each other at a traffic stop. In such a situation, there is a high possibility that abnormalities such as contact between vehicles or objects installed on the road will occur, and if the vehicle speed is low, the acceleration sensor installed on the vehicle may Since the output value of the collision detection signal tends to be low, the detection of an abnormality occurring in the vehicle by the collision detection signal may be insufficient. For this reason, abnormal sounds such as squeaks and contact noises generated outside the vehicle, screams and warning utterances made by occupants inside the vehicle, etc. are selected, and the voice recognition unit 14, which will be described later, performs voice recognition. It is effective to judge abnormalities that occur in

The voice recognition unit 14 recognizes the voice selected by the voice selection unit 13 and determines whether the voice indicating an abnormality occurring in the vehicle is included. The speech recognition unit 14 includes a trained model for speech recognition as a speech recognition dictionary, and performs speech recognition processing to recognize speech to be recognized.
The voice recognition unit 14 inputs the voice selected by the voice selection unit 13 into a voice recognition dictionary that recognizes voices indicating abnormalities occurring in the vehicle. The voice recognition dictionary can detect that the input voice may contain sounds that indicate abnormalities that occur in the vehicle, such as collision sounds, destruction sounds, abnormal noises made by the engine, and sounds that occur when the car body is damaged. output as a value between 1 and 0. The closer the output value is to 1, the more likely the input audio includes audio indicating an abnormality occurring in the vehicle, and the closer the output value is to 0, the more likely it is that the input audio includes audio indicating an abnormality occurring in the vehicle. There is a low possibility that the audio shown is included. If this output value is a predetermined threshold value, for example, 0.8 or more, the voice recognition unit 14 determines that the voice selected by the voice selection unit 13 includes a voice indicating an abnormality in the vehicle. do.
The voice recognition unit 14 outputs the determination result to the abnormality information output unit 15.

A speech recognition dictionary is a trained model obtained by inputting a large amount of pre-generated input data into a neural network and performing machine learning. In the case of a speech recognition dictionary that recognizes sounds that indicate an abnormality that occurs in a vehicle, a pair of collected sounds and training data that indicates whether or not the sound indicates an abnormality that occurs in a vehicle are used as a set. Use the data as input data for learning. In this case, the training data is data that is labeled as 1 if the sound indicates an abnormality occurring in the vehicle, and 0 if the sound does not indicate an abnormality occurring in the vehicle. be.

The voice recognition unit 14 may recognize sounds such as footsteps of a person approaching the vehicle, noises occurring around the vehicle, and a scream of a person as sounds indicating an abnormality occurring in the vehicle. In other words, the sound that is recognized by the voice recognition unit 14 and indicating an abnormality occurring in the vehicle is not limited to sounds such as noises that occur when an abnormality occurs in the vehicle, but is also a suspicious sound or a sign that an abnormality is occurring in the vehicle. Contains audio that may become. In this case, a set of data that includes the collected sound and training data indicating that the sound is the footsteps of a person approaching the vehicle, noises occurring around the vehicle, screams of a person, etc. is used as input data for learning. In addition, a speech recognition dictionary is generated to recognize speech indicating an abnormality occurring in the vehicle, and is used for speech recognition processing.

The speech recognition unit 14 may recognize the speech to be recognized using a known technique such as pattern matching, in addition to the above-described method. The speech recognition section 14 may perform preprocessing using known techniques, such as noise reduction processing and speaker separation, on the speech selected by the speech selection section 13 in order to improve the accuracy of speech recognition.

The abnormality information output unit 15 outputs information regarding whether or not an abnormality has occurred in the vehicle, based on the voice recognition determination result output by the voice recognition unit 14.
The abnormality information output unit 15 receives, from the voice recognition unit 14, a signal indicating the result of determining whether or not the voice selected by the voice selection unit 13 includes a voice indicating an abnormality in the vehicle, and indicates that there is an abnormality in the vehicle. The abnormality information indicating that the abnormality has occurred is outputted to the recording control section 16, communication section 17, notification control section (not shown), etc., which will be described later. The output destination may be arbitrary and is not limited.

The recording control unit 16 controls recording of vehicle surrounding information including at least one of audio and images acquired by the vehicle surrounding information acquisition unit 11 in the recording unit 50 based on the abnormality information output by the abnormality information output unit 15. I do. The recording control unit 16 records information such as vehicle acceleration and vehicle speed regarding the vehicle running state acquired by the vehicle running information acquisition unit 12 in the recording unit 50 based on the abnormality information output by the abnormality information output unit 15. Control may also be performed. Note that the recording control unit 16 is not necessarily an essential component of the abnormality information output device 10, and may be omitted.

Based on the abnormality information output by the abnormality information output unit 15, the communication unit 17 transmits vehicle surrounding information, which is at least one of audio and images acquired by the vehicle surrounding information acquisition unit 11, to a predetermined external server (not shown). or to an external device such as a communication device maintained by a vehicle occupant or vehicle manager. Based on the abnormality information output by the abnormality information output unit 15, the communication unit 17 may similarly transmit information regarding the vehicle driving state acquired by the vehicle driving information acquisition unit 12 to an external device.
The communication unit 17 may transmit the abnormality information itself outputted by the abnormality information output unit 15 to a predetermined external device. Note that the communication unit 17 is not necessarily an essential component of the abnormality information output device 10, and may be omitted.

A notification control unit (not shown) may control abnormality notification by voice via a speaker (not shown), a buzzer, or a vehicle horn, based on the abnormality information output by the abnormality information output unit 15. Based on the abnormality information output by the abnormality information output unit 15, the notification control unit notifies the abnormality by display or light via a display device (not shown) such as a monitor or LED (Light Emitting Diode) or the headlights of the vehicle. May be controlled. Note that the notification control section is not necessarily an essential component of the abnormality information output device 10, and may be omitted.

Next, the processing procedure of the abnormality information output device 10 according to the first embodiment described above will be explained with reference to the flowchart shown in FIG. 2.

First, in step S<b>10 , the vehicle surrounding information acquisition unit 11 acquires vehicle surrounding information including a plurality of sounds, and outputs it to the sound selection unit 13 . After that, the process advances to step S11.

In step S<b>11 , the vehicle running information acquisition section 12 acquires information regarding the traveling state of the vehicle, for example, information regarding whether the vehicle is in a parked state, and outputs it to the audio selection section 13 . After that, the process advances to step S12.

In step S12, the voice selection unit 13 selects at least one voice from among the plurality of voices acquired by the vehicle surrounding information acquisition unit 11 based on the information regarding the driving state of the vehicle acquired by the vehicle running information acquisition unit 12. do. For example, when the vehicle is in a parked state, the sound collected by the sound collection unit 20 disposed outside the vehicle is selected and output to the voice recognition unit 14. After that, the process advances to step S13.

In step S13, the voice recognition unit 14 performs voice recognition processing to recognize the voice selected by the voice selection unit 13 and determine whether the voice indicating an abnormality that has occurred in the vehicle is included. The voice recognition unit 14 outputs the determined result to the abnormality information output unit 15. After that, the process advances to step S14.

In step S14, if the output of the voice recognition unit 14 indicates an abnormality that has occurred in the vehicle (Yes in S14), the abnormality information output unit 15 proceeds to step S15. If the output of the voice recognition unit 14 does not indicate any abnormality that has occurred in the vehicle (No in S14), the process returns to step S10 and the process is repeated.

In step S15, the abnormality information output unit 15 outputs abnormality information. The output destinations include the recording control section 16, the communication section 17, and the notification control section (not shown). Thereafter, the process returns to step S10 and the process is repeated.

Note that in this flow, there may be a step of terminating the process based on the information regarding vehicle travel acquired by the vehicle travel information acquisition unit 12. For example, the process may be terminated based on information related to vehicle running indicating that the vehicle has transitioned from a running state to a parked state.

As described above, according to the first embodiment, the abnormality information output device 10 selects at least one voice from a plurality of sound collection units 20 arranged around the vehicle based on the traveling information of the vehicle, and performs voice recognition processing. By doing so, abnormalities occurring in the vehicle can be more appropriately determined and output. For example, when the vehicle is parked, the sound of a collision or injury occurring outside the vehicle, or the footsteps of a person approaching the vehicle, is selected from the sound collection unit 20 located outside the vehicle, and the voice recognition is performed. By doing so, abnormalities occurring in the vehicle can be determined more accurately. In addition, by not selecting the sound collected from the sound collection unit 20 placed inside the vehicle when the vehicle is parked, the load on voice recognition processing can be reduced, and the vehicle can be driven with less power consumption. It is possible to determine abnormalities that occur in

<Embodiment 2>
Embodiment 2 of the present invention will be described using FIG. 3. The abnormality information output device 10a according to the second embodiment has a different configuration from the abnormality information output device 10 according to the first embodiment in the following points.
That is, the speech recognition unit 14a has a function of selecting a speech recognition dictionary to be used for speech recognition. In the abnormality information output device 10a, the vehicle surrounding information acquisition unit 11, the vehicle running information acquisition unit 12, the audio selection unit 13, and the abnormality information output unit 15 have the same configuration and functions as the abnormality information output device 10 according to the first embodiment. Therefore, detailed explanation thereof will be omitted.

The speech recognition unit 14a selects an appropriate speech recognition dictionary from among a plurality of speech recognition dictionaries used for speech recognition based on the information regarding vehicle running acquired by the vehicle running information acquisition unit 12. A dictionary selection section 141 is provided. The speech recognition dictionary selection unit 141 selects a speech recognition dictionary that has been trained with different speech for each driving state of the vehicle.
Since the type of abnormality that occurs in the vehicle or the type of abnormality that is likely to occur in the vehicle differs depending on the driving state of the vehicle, the types of sounds that occur when an abnormality occurs and sounds that are a sign of an abnormality also differ. For this reason, the plurality of speech recognition dictionaries included in the speech recognition dictionary selection unit 141 perform machine learning on different types of speech, respectively. In other words, multiple voice recognition dictionaries register and learn many voices that can more appropriately determine abnormalities depending on each driving condition, and also learn voices that can more accurately determine abnormalities that are not appropriate for each driving condition. By not learning the speech recognition dictionary, the speech recognition dictionary is made appropriate for each different driving state of the vehicle.
The speech recognition dictionary selection unit 141 has a function of appropriately selecting the type of speech to be recognized for each driving state of the vehicle, and the speech recognition unit 14a uses the selected speech recognition dictionary to select a type of speech that is more appropriate. determines abnormalities that occur in the vehicle.

For example, when the information regarding vehicle running indicates a parking state of the vehicle, the voice recognition unit 14a registers and learns a voice for recognizing a voice indicating an abnormality that is likely to occur in a parked vehicle. A recognition dictionary is selected and used to determine abnormalities occurring in the vehicle. Here, the sounds indicating an abnormality that is likely to occur in a parked vehicle include, for example, the sound of a person scratching the vehicle body, the sound of footsteps of a person approaching the vehicle, and the sound of conversation near the vehicle.
For example, when the information regarding vehicle running indicates a running state of the vehicle at a running speed equal to or higher than a predetermined value, the voice recognition unit 14a generates a voice indicating an abnormality that is likely to occur in a parked vehicle as described above. The system selects and uses a voice recognition dictionary that has been learned by excluding voices that are difficult to recognize when the vehicle is running to determine abnormalities that occur in the vehicle.

The speech recognition dictionary selection unit 141 may select an appropriate speech recognition dictionary from among a plurality of speech recognition dictionaries used for speech recognition, based on the speech selected by the speech selection unit 13. More specifically, the speech recognition dictionary selection section 141 selects the position where the sound collection section 20 is located based on where around the vehicle the sound collection section 20 is located for the sound selected by the sound selection section 13. It may also be possible to register a larger number of voices that are appropriate when determining vehicle abnormalities that are likely to occur in each case, and to select a voice recognition dictionary that eliminates or reduces inappropriate voices.

For example, if the voice selected by the voice selection unit 13 is the voice collected by the sound collection unit 20 disposed outside the vehicle at the bottom of the vehicle, the voice recognition unit 14a registers more footsteps of a person approaching the vehicle. Then, the speech recognition dictionary that has been trained is selected and speech recognition processing is performed. In addition, if the sound selected by the sound selection unit 13 is the sound collected by the sound collection unit 20 located outside the vehicle near the hood or door mirror, it may be the sound of a person scratching the car body or the sound of a conversation near the vehicle. A speech recognition dictionary that has been trained by registering more words is selected and speech recognition processing is performed.

As described above, according to the second embodiment, the abnormality information output device 10a selects a speech recognition dictionary to be used in the speech recognition process based on vehicle travel information or the selected speech. With this configuration, a more appropriate speech recognition dictionary can be used based on the driving state of the car and the position of the sound collection unit 20 located in the vehicle to collect the sound. Abnormalities that occur in the vehicle can be determined and output.

<Modified example>
When the voice selected by the voice selection unit 13 is the voice collected by the sound collection unit 20 arranged on the road surface around the vehicle, the voice recognition unit 14a uses a voice recognition dictionary that recognizes conductive voice transmitted on the road surface. may be selected.
The speech recognition dictionary selected in this case includes a conductive sound transmitted through the road surface collected by the sound collection unit 20 grounded on the road surface, and a teacher indicating whether or not the sound indicates an abnormality occurring in the vehicle. This is a speech recognition dictionary that is trained using a pair of data as input data. As described above, the sound indicating an abnormality occurring in the vehicle includes suspicious sounds occurring around the vehicle or sounds indicating an abnormality requiring attention.

In this modification, the sound collection section 20 is placed in advance in contact with the road surface around the vehicle. For example, it may be placed on the road surface in a parking space for a vehicle and connected to the abnormality information output device 10 by wireless communication such as Bluetooth (registered trademark). Further, the sound collection unit 20 may be manually placed on the road surface by an occupant of the vehicle or the like when the vehicle is parked. The sound collection unit 20 is automatically positioned by a position control mechanism (not shown) so that it is grounded on the road surface using a wire or the like when the vehicle is in a parked state, and is automatically wound up when the vehicle is in a running state. It may be stored at a predetermined location.

In this modification, when the vehicle running information acquisition unit 12 acquires information indicating that the vehicle is in a parked state, the audio selection unit 13 selects the audio collected by the sound collection unit 20 placed in contact with the road surface. is selected and output to the speech recognition section 14a.
The voice recognition unit 14a recognizes the selected voice when the vehicle running information acquisition unit 12 acquires information indicating that the vehicle is in a parked state, or when the voice recognition unit 14a recognizes that the selected voice is a sound pickup unit placed on the road surface around the vehicle. If 20 is the collected sound, select a speech recognition dictionary that recognizes conductive sounds that has been trained by registering conductive sounds that travel on the ground, and determine the abnormality that occurs in the vehicle. .

With this configuration, the abnormality information output device 10a in this modification recognizes the footsteps of a person approaching the vehicle and the sound of other vehicles running as conductive sounds, and more accurately identifies them as abnormalities occurring in the vehicle. can be determined. In addition, since the recognized sound is conductive sound that travels through the ground, it is possible to suppress the possibility that non-conductive sound such as wind noise, rain sound, animal cry, etc. will be mistakenly judged as an abnormality occurring in the vehicle. can.

<Other embodiments>
The specific examples shown in Embodiment 1, Embodiment 2, and modifications of Embodiment 2 described above in the present invention are merely illustrative to facilitate understanding of the invention, and unless otherwise specified, do not limit the present invention. It's not something you do.
Note that the present invention is not limited to Embodiment 1, Embodiment 2, and modifications of Embodiment 2, and can be modified as appropriate without departing from the spirit.

For example, in Embodiment 1, Embodiment 2, and a modification of Embodiment 2, based on the vehicle running information acquired by the vehicle running information acquisition unit 12, the audio control unit (not shown) controls the sound collected by the sound collection unit 20. There may be embodiments in which gain adjustments are made. For example, when the vehicle is parked, the audio control unit (not shown) changes the gain of the sound collected by the sound collection unit 20 located outside the vehicle under the vehicle body. By setting the gain higher than the gain of the sound picked up by the sound unit 20, the sound of the footsteps of a person approaching the vehicle can be recognized more accurately. Further, when the vehicle is in a parked state, an audio control unit (not shown) changes the gain of the sound collected by the sound collection unit 20 located outside the vehicle under the vehicle body to the sound collected by the sound collection unit 20 located outside the vehicle near the bonnet or door mirror. By setting the gain to be lower than the gain of the sound collected by the unit 20, it is possible to make the sound that can cause damage to the vehicle to be recognized more accurately. Preferably, these settings of the voice control unit (not shown) can be set arbitrarily by the vehicle occupant. Further, when the vehicle is traveling at a speed equal to or higher than a predetermined value, control may be performed to suppress the gain of the sound collected by the sound collection unit 20 disposed outside the vehicle.

In Embodiment 1, Embodiment 2, and the modification of Embodiment 2, voice recognition processing was performed in order to determine an abnormality occurring in the vehicle from the voice selected by the voice selection unit 13. Instead of performing voice recognition processing using the voice recognition process, abnormality occurring in the vehicle may be determined using means such as volume and frequency analysis of the selected voice. At this time, if the volume of the selected sound exceeds a predetermined value predetermined for each piece of information indicating which position of the vehicle the sound collection unit 20 at which the selected sound is located collects the sound, the vehicle It is determined that an abnormality has occurred. Alternatively, the volume of each frequency band of the selected audio may be predetermined for each audio frequency band based on information indicating where in the vehicle the sound collection section 20 located the selected audio collected the audio. If the value exceeds a predetermined value, it is determined that an abnormality has occurred in the vehicle. With such a configuration, the processing required for voice recognition processing can be significantly suppressed, and abnormalities occurring in the vehicle can be determined and output with a simple configuration and low power consumption.

In the embodiments described above, the present invention has been described as a hardware configuration, but the present invention is not limited to this. The present invention can also be implemented by causing a processor to execute a computer program.

In the examples described above, the program may be stored and provided to the computer using various types of non-transitory computer-readable media. Non-transitory computer-readable media include various types of tangible recording media, including, for example, magnetic recording media such as hard disk drives, optical recording media, magneto-optical recording media, and semiconductor memory. Semiconductor memory can be various types of rewritable ROM (Read Only Memory) and RAM (Random Access Memory), and also includes those provided as SSD (Solid State Drive). The program may also be delivered to the computer via various types of temporary computer-readable media, such as electrical signals, optical signals, , and electromagnetic waves.

The processes in the apparatus and method shown in the claims, specification, and drawings can be performed in any order. Even if the claims, specification, and operational flows in the drawings are explained using "first", "next", etc. for convenience, this does not mean that it is essential to perform them in this order. do not have.

10, 10a Abnormality information output device 11 Vehicle surrounding information acquisition unit 12 Vehicle running information acquisition unit 13 Voice selection unit 14, 14a Voice recognition unit 141 Voice recognition dictionary selection unit 15 Abnormality information output unit 20 Sound collection unit (microphone)
30 Imaging unit (camera)
40 Sensors 50 Recording section

Claims (3)

a vehicle surrounding information acquisition unit that obtains vehicle surrounding information including a plurality of collected sounds;
a vehicle running information acquisition unit that acquires information regarding the running state of the vehicle;
a voice selection unit that selects at least one voice from the plurality of voices based on information regarding the driving state of the vehicle;
a voice recognition unit that recognizes the selected voice and determines whether a voice indicating an abnormality occurring in the vehicle is included;
an abnormality information output unit that outputs an abnormality occurring in the vehicle as abnormality information based on the determination;
Equipped with
The voice recognition unit includes a voice recognition dictionary selection unit that selects a voice recognition dictionary to be used for the recognition based on information regarding the driving state of the vehicle or the selected voice.
Abnormality information output device. The plurality of sounds include sounds collected by a sound collection unit placed in contact with the road surface,
When the information regarding the running state of the vehicle indicates that the vehicle is in a parked state, the sound selection unit selects the sound collected by the sound collection unit disposed in contact with the road surface;
The voice recognition unit selects a voice recognition dictionary that recognizes conductive voice transmitted on the road surface as a voice recognition dictionary used for the recognition, recognizes the selected voice, and recognizes a voice indicating an abnormality occurring in the vehicle. Determine whether or not it is included,
The abnormality information output device according to claim 1. The computer is
acquiring vehicle surrounding information including a plurality of collected sounds;
obtaining information regarding the driving state of the vehicle;
selecting at least one voice from the plurality of voices based on information regarding the driving state of the vehicle;
A voice recognition dictionary to be used for the recognition is selected based on the information regarding the driving state of the vehicle or the selected voice, and the selected voice is recognized using the selected voice recognition dictionary and generated in the vehicle. a step of determining whether or not audio indicating an abnormality is included;
outputting an abnormality occurring in the vehicle as abnormality information based on the determination;
An error information output method that executes.

Images