Vehicle accident detecting device

Abstract

A vehicle accident detecting device (2) comprises: a sound source position estimating unit (S210 to S250, S1110 to S1180); a vehicle body exterior determining unit (S260 to S330, S1190 to S1220); and a collision determining unit (S410, S450, S460, S480, S520 to S550). The sound source position estimating unit is configured to estimate an estimated sound source position, which is the position of a sound source, on the basis of the detection results of a plurality of sound detecting units (31 to 34) installed in the vehicle and configured to detect sound. The vehicle body exterior determining unit is configured to determine whether the estimated sound source position is on the vehicle body exterior of the vehicle. The collision determining unit is configured to determine that a collision accident has occurred if the vehicle body exterior determining unit determines that the estimated sound source position is on the vehicle body exterior.

Description

Vehicle Accident Detection Device CROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This international application claims the benefit of Japanese Patent Application No. 2023-127856, filed with the Japan Patent Office on August 4, 2023, the entire disclosure of which is incorporated herein by reference.

[0002] The present disclosure relates to a vehicle accident detection device that detects a collision accident in a vehicle.

[0003] Patent Document 1 describes a sound detection device that detects surrounding objects by calculating the phase correlation of sound signals collected by two or more sound collection units, and determines whether the surrounding object is approaching or moving away based on the degree of phase fluctuation.

[0004] Patent No. 5954429

[0005] As a result of detailed investigation by the inventors, it was found that the technology described in Patent Document 1 has the problem that it is not possible to calculate the distance to the source of the sound (i.e., the sound source), and therefore is not able to detect vehicle collision accidents.

[0006] The present disclosure detects crash events in vehicles.

[0007] One aspect of the present disclosure is a vehicle accident detection device including a sound source position estimation unit, a vehicle body exterior determination unit, and a collision determination unit.

[0008] The sound source position estimation unit is configured to estimate a sound source estimated position, which is the position of a sound source, based on detection results from a plurality of sound detection units that are installed in the vehicle and configured to detect sound.

[0009] The vehicle body exterior determining unit is configured to determine whether the estimated position of the sound source is on the vehicle body exterior.

[0010] The collision determination unit is configured to determine that a collision accident has occurred when the vehicle body exterior determination unit determines that the estimated position of the sound source is on the vehicle body exterior.

[0011] The vehicle accident detection device of the present disclosure configured in this manner can detect a collision accident in a vehicle using a plurality of sound detection units installed in the vehicle.

[0012] 1 is a block diagram showing the configuration of an accident information collection system. FIG. 1 is a block diagram showing the configuration of a vehicle accident detection device. FIG. 2 is a diagram explaining a sound source location estimation method of a first embodiment. FIG. 3 is a functional block diagram showing the functional configuration of the vehicle accident detection device of the first embodiment. FIG. 4 is a flowchart showing accident detection processing of the first embodiment. FIG. 5 is a flowchart showing sound source location estimation processing of the first embodiment. FIG. 6 is a diagram explaining first boundary determination conditions. FIG. 7 is a diagram explaining second boundary determination conditions. FIG. 8 is a flowchart showing collision determination processing of the first embodiment. FIG. 9 is a flowchart showing specific operation determination processing. FIG. 10 is a flowchart showing filter setting processing. FIG. 11 is a diagram explaining a sound source location estimation method of a second embodiment. FIG. 12 is a functional block diagram showing the functional configuration of the vehicle accident detection device of the second embodiment. FIG. 13 is a flowchart showing accident detection processing of the second embodiment. FIG. 14 is a flowchart showing sound source location estimation processing of the second embodiment. FIG. 15 is a diagram explaining a sound source location estimation method of a third embodiment. FIG. 16 is a functional block diagram showing the functional configuration of the vehicle accident detection device of the third embodiment. FIG. 17 is a flowchart showing accident detection processing of the third embodiment. FIG. 18 is a flowchart showing collision determination processing of the third embodiment. FIG. 19 is a diagram explaining a sound source location estimation method of another embodiment.

[0013] First Embodiment A first embodiment of the present disclosure will be described below with reference to the drawings.

[0014] As shown in FIG. 1, the accident information collection system 1 of this embodiment includes a plurality of vehicle accident detection devices 2 and an accident information server 3.

[0015] The vehicle accident detection device 2 is mounted on a vehicle and has a function of performing data communication with the accident information server 3 via a wide area wireless communication network NW.

[0016] The accident information server 3 acquires accident information, which will be described later, from the vehicle accident detection device 2 via the wide area wireless communication network NW. The accident information server 3 includes an accident information storage unit 4 that stores the accident information.

[0017] 2, the vehicle accident detection device 2 includes a control unit 11, a CAN communication unit 12, a storage unit 13, and a communication unit 14. CAN is an abbreviation for Controller Area Network. CAN is a registered trademark.

[0018] The control unit 11 is an electronic control device mainly composed of a microcomputer including a CPU 21, a ROM 22, a RAM 23, etc. The various functions of the microcomputer are realized by the CPU 21 executing a program stored in a non-transitory tangible recording medium. In this example, the ROM 22 corresponds to the non-transitory tangible recording medium storing the program. Furthermore, the execution of this program results in the execution of a method corresponding to the program. Note that some or all of the functions executed by the CPU 21 may be configured as hardware using one or more ICs, etc. Furthermore, the number of microcomputers constituting the control unit 11 may be one or more.

[0019] The CAN communication unit 12 is connected to a plurality of ECUs via communication lines so as to be able to communicate data with them, and transmits and receives data according to the CAN communication protocol. Specifically, the plurality of ECUs connected to the CAN communication unit 12 include an engine ECU that controls the engine, a brake ECU that controls the brakes, a steering ECU that controls the steering, a suspension ECU that controls the suspension, an ECU that controls the on / off of lights, etc. In Fig. 2, only ECUs 111, 112, and 113 are shown as ECUs connected to the CAN communication unit 12.

[0020] The storage unit 13 is a storage device for storing various data.

[0021] The communication unit 14 performs data communication with the accident information server 3 via the wide area wireless communication network NW.

[0022] 3, the vehicle is equipped with microphones 31, 32, 33, and 34. The microphones 31, 32, 33, and 34 receive sound waves as input, convert them into electrical signals (hereinafter referred to as sonic electrical signals) that indicate the change in amplitude of the input sound waves over time, and output the electrical signals.

[0023] In this embodiment, microphone 31 is installed, for example, on the left side of the dashboard, microphone 32 is installed, for example, on the right side of the dashboard, microphone 33 is installed, for example, on the rear side of the luggage compartment, and microphone 34 is installed, for example, in the center of the ceiling. Microphones 31, 32, and 33 are installed so as to be located on the same horizontal plane.

[0024] The microphones 31 to 34 are installed at different positions. Therefore, differences occur in the phase of the sound waves detected by the microphones 31 to 34 that detect the sound waves generated by the sound source SS1. The difference in the phase of the sound waves can be considered to be the difference in distance from the sound source SS1 to the microphones 31 to 34. The acoustic electrical signals ES1, ES2, ES3, and ES4 in Figure 3 are signals output by the microphones 31, 32, 33, and 34, respectively, when they detect the sound waves generated by the sound source SS1.

[0025] Therefore, for the pair of microphones 31 and 32, a curve (hereinafter referred to as a distance difference curve) L1 that passes through multiple points where the difference in distance to the sound source SS1 is equal can be calculated based on the difference in distance to the sound source SS1. Similarly, for the pair of microphones 32 and 33, a distance difference curve L2 can be calculated based on the difference in distance to the sound source SS1. Furthermore, for the pair of microphones 31 and 33, a distance difference curve L3 can be calculated based on the difference in distance to the sound source SS1. Note that for one pair of two microphones, typically two distance difference curves that pass through multiple points where the difference in distance to the sound source SS1 is equal are calculated. Which of the two distance difference curves for one pair is the distance difference curve that is estimated to pass through the sound source SS1 can be selected by identifying the microphone that detected sound waves earlier out of the two microphones.

[0026] The vehicle accident detection device 2 then estimates the point where two pairs of distance difference curves (e.g., distance difference curves L1 and L2) intersect, or the area surrounded by three pairs of distance difference curves (i.e., distance difference curves L1, L2, and L3), as the two-dimensional position of the sound source SS1 (hereinafter referred to as the two-dimensional position).

[0027] If there is no measurement error in the distance difference curves L1, L2, and L3, the distance difference curves L1, L2, and L3 intersect at the position of the sound source SS1. In this case, the vehicle accident detection device 2 can estimate the intersection of the three distance difference curves L1, L2, and L3 as the two-dimensional position of the sound source SS1.

[0028] Based on the same concept, by using the microphone 34 in addition to the microphones 31, 32, and 33, the vehicle accident detection device 2 can estimate the three-dimensional position of the sound source SS1 (hereinafter referred to as the three-dimensional position).

[0029] Furthermore, if the position of the sound source SS1 can be estimated, the distance from the microphones 31 to 34 to the sound source SS1 can also be estimated, and therefore the vehicle accident detection device 2 can also estimate the timing of the occurrence of the detected sound waves.

[0030] As shown in Figure 4, the vehicle accident detection device 2 includes functional blocks that are realized by the CPU 21 executing a program stored in the ROM 22, such as a vehicle information acquisition unit 41, a sound acquisition unit 42, a sound analysis unit 43, a video acquisition unit 44, a video analysis unit 45, an accident detection control unit 46, a collision determination unit 47, and an upload unit 48.

[0031] As indicated by arrow A1, vehicle information acquisition unit 41 constantly acquires vehicle information from a plurality of ECUs via CAN communication unit 12. The vehicle information includes, for example, information such as vehicle speed, engine RPM, steering angle, brake opening, on / off status of the parking brake, shift position, on / off status of windows, lock / unlock status of doors, on / off status of blinkers, etc.

[0032] Furthermore, the vehicle information acquisition unit 41 constantly acquires acceleration information indicating the acceleration detected by the acceleration sensor 35 mounted on the vehicle, as indicated by an arrow A2.

[0033] As indicated by arrows A3, A4, A5, and A6, the sound acquisition unit 42 constantly acquires sonic electric signals from the microphones 31 to 34. Based on the sound waves indicated by the acquired sonic electric signals, the sound acquisition unit 42 detects sounds that may have been generated by an object colliding with the vehicle (hereinafter, referred to as candidate collision sounds).

[0034] When the sound acquisition unit 42 detects a candidate collision sound, it outputs sound wave data indicating the change over time in amplitude of the sound wave corresponding to the candidate collision sound to the sound analysis unit 43, as indicated by arrow A7.

[0035] The sound analysis unit 43 analyzes the sound waves indicated by the sound wave data acquired from the sound acquisition unit 42 to estimate the position of the sound source of the collision candidate sound and the occurrence timing of the collision candidate sound.

[0036] As indicated by an arrow A8, the sound analysis unit 43 notifies the sound acquisition unit 42 of the position of the sound source of the collision candidate sound and the occurrence timing of the collision candidate sound.

[0037] When the sound acquisition unit 42 detects a candidate collision sound, it notifies the accident detection control unit 46 of the position of the sound source of the candidate collision sound, the timing of occurrence of the candidate collision sound, and the waveform of the candidate collision sound, as shown by arrow A9.

[0038] As indicated by arrows A10 and A11, the image acquisition unit 44 constantly acquires image data generated by the cameras 36 and 37 mounted on the vehicle. The cameras 36 and 37 are installed so as to be able to capture at least the exterior of the vehicle.

[0039] The video acquisition unit 44 outputs the acquired video data to the video analysis unit 45, as indicated by an arrow A12.

[0040] The video analysis unit 45 analyzes the video data acquired from the video acquisition unit 44 and determines whether or not an object has come into contact with the vehicle. The video analysis unit 45 notifies the video acquisition unit 44 of the determination result as to whether or not an object has come into contact with the vehicle, as indicated by arrow A13.

[0041] When the accident detection control unit 46 is notified of the candidate collision sound by the sound acquisition unit 42, it acquires vehicle information, acceleration information, and video data at the timing when the candidate collision sound occurs, as indicated by arrows A14 and A15.

[0042] After acquiring the vehicle information, acceleration information, and video data, the accident detection control unit 46 outputs a collision determination request to the collision determination unit 47, as indicated by arrow A16.

[0043] When the collision determination unit 47 receives a collision determination request from the accident detection control unit 46, it obtains vehicle body information from the vehicle body information storage unit 49, as indicated by an arrow A17. The vehicle body information is information indicating the coordinate position of the vehicle body exterior.

[0044] The collision determination unit 47 determines whether or not an object has collided with the vehicle based on the vehicle information, acceleration information, video data, and vehicle body information.

[0045] The collision determination unit 47 notifies the accident detection control unit 46 of the determination result as to whether or not an object has collided with the vehicle, as indicated by an arrow A18.

[0046] When an object collides with the vehicle, the accident detection control unit 46 generates accident information including the acquired vehicle information and video data based on the determination result acquired from the collision determination unit 47. The accident detection control unit 46 stores the generated accident information in the temporary storage unit 50, as indicated by arrow A19.

[0047] After storing the accident information in the temporary storage unit 50, the accident detection control unit 46 outputs an upload instruction to the upload unit 48, as indicated by an arrow A20.

[0048] When the upload unit 48 receives an upload instruction from the accident detection control unit 46, it retrieves the accident information from the temporary storage unit 50, as indicated by arrow A21. The upload unit 48 uploads the retrieved accident information to the accident information server 3. As a result, the accident information server 3 stores the uploaded accident information in the accident information storage unit 4.

[0049] Next, a procedure for the accident detection process executed by the control unit 11 will be described. The accident detection process is a process that is repeatedly executed during operation of the vehicle accident detection device 2. Note that, in this embodiment, for the sake of simplicity, a form in which an accident is detected using the microphones 31, 32, and 33 will be described.

[0050] When the accident detection process is executed, the CPU 21 of the control unit 11 detects a candidate collision sound in S10, as shown in Fig. 5. Specifically, the CPU 21 determines that a candidate collision sound has been detected when the amplitude of the sound wave indicated by the sonic electrical signal acquired from the microphones 31 to 33 is equal to or greater than a preset collision sound determination value. Note that the CPU 21 may perform filtering to remove sounds that are unlikely to be collision sounds from the waveform of the sound wave before determining whether the amplitude of the sound wave is equal to or greater than the collision sound determination value.

[0051] In S20, the CPU 21 determines whether or not a candidate collision sound has been detected based on the detection result of S 10. If a candidate collision sound has not been detected, the CPU 21 ends the accident detection process.

[0052] On the other hand, if a collision candidate sound is detected, the CPU 21 executes a sound source position estimation process, which will be described later, in S30. In the sound source position estimation process, the CPU 21 estimates the position of the sound source of the collision candidate sound (hereinafter referred to as the estimated sound source position) and the occurrence timing of the collision candidate sound (hereinafter referred to as the sound occurrence timing).

[0053] In S40, the CPU 21 acquires vehicle information around the sound generation timing. The vicinity of the sound generation timing is, for example, a time range from a predetermined acquisition time before the sound generation timing (e.g., one second before) to a predetermined acquisition time after the sound generation timing (e.g., one second after).

[0054] In S50, the CPU 21 acquires acceleration information around the sound generation timing.

[0055] In S60, the CPU 21 acquires video data around the sound generation timing.

[0056] In S70, the CPU 21 executes a collision determination process, which will be described later. In the collision determination process, the CPU 21 determines whether or not an object has collided with the vehicle.

[0057] In S80, the CPU 21 determines whether or not a collision has occurred based on the determination result of S70. If a collision has not occurred, the CPU 21 ends the accident detection process.

[0058] On the other hand, if a collision has occurred, the CPU 21 generates the above-mentioned accident information and stores the generated accident information in the temporary storage unit 50 in S90.

[0059] In S100 , the CPU 21 retrieves the accident information generated in S90 from the temporary storage unit 50 , and uploads the retrieved accident information to the accident information server 3 .

[0060] When the process of S100 ends, the CPU 21 ends the accident detection process.

[0061] Next, the procedure of the sound source position estimation process executed in S30 will be described.

[0062] 6, the CPU 21 determines in S210 whether the distance difference curves have been calculated for all pairs of microphones 31, 32, and 33. In this embodiment, all pairs are the pair of microphone 31 and microphone 32, the pair of microphone 32 and microphone 33, and the pair of microphone 31 and microphone 33.

[0063] Here, if there are pairs for which the distance difference curve has not been calculated, the CPU 21 selects one pair for which the distance difference curve has not been calculated in S220, and calculates the phase difference between the two sound waves detected by the two microphones that make up the selected pair.

[0064] In S230, the CPU 21 calculates the difference in distance from the two microphones constituting the selected pair to the sound source based on the phase difference calculated in S220.

[0065] In S240, the CPU 21 calculates an equation representing a distance difference curve that passes through multiple points where the distance difference from the two microphones that make up the selected pair to the sound source is equal, based on the distance difference calculated in S230, and then proceeds to S210.

[0066] If distance difference curves have been calculated for all pairs in S210, the CPU 21 estimates the coordinate position to which the distance difference curves of all pairs are closest as the estimated sound source position in S250.

[0067] In S260, the CPU 21 determines whether or not the vehicle body information is stored in the vehicle body information storage unit 49. If the vehicle body information is stored, the CPU 21 acquires the vehicle body information from the vehicle body information storage unit 49 in S270.

[0068] In S280, the CPU 21 determines whether the first boundary determination condition is met.

[0069] As shown in FIG. 7, the vehicle body information includes information on the coordinate positions of a plurality of points (hereinafter referred to as "surface points") P1, P2, . . . , P29, P30 set on the surface of the vehicle body contour VS.

[0070] The first boundary determination condition is that both the boundary point determination condition and the boundary line determination condition below are met.

[0071] The condition for determining a surface point is that the distance between the estimated sound source position estimated in S250 and the surface point closest to this estimated sound source position is equal to or less than a preset surface point determination distance.

[0072] The contour line determination condition is that the distance between the sound source estimated position estimated in S250 and the contour line that is closest to the sound source estimated position among multiple contour lines connecting two adjacent contour points is less than or equal to a predetermined contour line determination distance.

[0073] For example, when the distance d between the sound source SS2 shown in FIG. 7 and the surface point P21 closest to the sound source SS2 is equal to or less than the surface point determination distance, the surface point determination condition is met.

[0074] Furthermore, when the distance h between the sound source SS2 shown in FIG. 7 and the outline SL1 connecting the outline point P20 and the outline point P21 is equal to or less than the outline determination distance, the outline determination condition is met.

[0075] As shown in FIG. 6, if the first outer shell determination condition is met in S280, the CPU 21 determines in S290 that the estimated sound source position estimated in S250 is near the vehicle outer shell, and ends the sound source position estimation process.

[0076] On the other hand, if the first outer shell determination condition is not met in S280, the CPU 21 determines in S300 that the estimated sound source position estimated in S250 is not near the vehicle outer shell, and ends the sound source position estimation process.

[0077] If it is determined in S260 that no vehicle body information has been stored, the CPU 21 proceeds to S310 to determine whether or not the second contour determination condition is met.

[0078] The second boundary determination condition is that the following first distance determination condition, second distance determination condition, and third distance determination condition are all met.

[0079] The first distance determination condition is that the distance ΔA is X1 to X2 times the distance AB, the distance ΔB is X3 to X4 times the distance AB, and the distance ΔA is X5 to X6 times the distance ΔB.

[0080] The second distance determination condition is that the distance ΔB is Y1 to Y2 times the distance BC, the distance ΔC is Y3 to Y4 times the distance BC, and the distance ΔB is Y5 to Y6 times the distance ΔC.

[0081] The third distance determination condition is that the distance ΔC is Z1 to Z2 times the distance CA, the distance ΔA is Z3 to Z4 times the distance CA, and the distance ΔC is Z5 to Z6 times the distance ΔA.

[0082] As shown in Fig. 8, distance ΔA is the distance between microphone 31 and sound source SS2. Distance ΔB is the distance between microphone 32 and sound source SS2. Distance ΔC is the distance between microphone 33 and sound source SS2. Distance AB is the distance between microphone 31 and microphone 32. Distance BC is the distance between microphone 32 and microphone 33. Distance CA is the distance between microphone 33 and microphone 31.

[0083] The first distance determination condition indicates that when a triangle is formed with the positions of microphone 31, microphone 32, and sound source SS2 as vertices, the lengths of the three sides of the triangle fall within a certain range. X1 to X6 are set so that the lengths of the three sides of the triangle fall within a certain range.

[0084] The second distance determination condition indicates that when a triangle is formed with the positions of microphone 32, microphone 33, and sound source SS2 as vertices, the lengths of the three sides of the triangle fall within a certain range. The above Y1 to Y6 are set so that the lengths of the three sides of the triangle fall within a certain range.

[0085] The third distance determination condition indicates that when a triangle is formed with the positions of microphone 33, microphone 31, and sound source SS2 as vertices, the lengths of the three sides of the triangle fall within a certain range. Z1 to Z6 are set so that the lengths of the three sides of the triangle fall within a certain range.

[0086] As shown in FIG. 6, if the second outer shell determination condition is met in S310, the CPU 21 determines in S320 that the estimated sound source position estimated in S250 is near the vehicle outer shell, and ends the sound source position estimation process.

[0087] On the other hand, if the second outer shell determination condition is not met in S310, the CPU 21 determines in S330 that the estimated sound source position estimated in S250 is not near the vehicle outer shell, and ends the sound source position estimation process.

[0088] Next, the procedure of the collision determination process executed in S70 will be described.

[0089] When the collision determination process is executed, the CPU 21 determines whether the estimated position of the sound source is near the outer hull of the vehicle in S410, as shown in Fig. 9. If the estimated position of the sound source is not near the outer hull of the vehicle, the CPU 21 proceeds to S490.

[0090] On the other hand, if the estimated sound source position is near the exterior of the vehicle body, the CPU 21 acquires acceleration information around the sound generation timing in S420.

[0091] At S430, the CPU 21 acquires video data around the sound generation timing, and analyzes the acquired video data to determine whether or not an object has come into contact with the vehicle near the estimated sound source position.

[0092] At S440, the CPU 21 acquires vehicle information around the sound generation timing.

[0093] At S450, the CPU 21 determines whether the acceleration indicated by the acceleration information acquired at S420 is equal to or greater than a preset collision determination acceleration. If the acceleration is less than the collision determination acceleration, the CPU 21 proceeds to S490.

[0094] On the other hand, if the acceleration is equal to or greater than the collision determination acceleration, the CPU 21 determines in S460, based on the determination result in S430, whether or not the video data acquired in S430 contains video of an object that has come into contact with the vehicle near the estimated sound source position.

[0095] If there is no image of an object that has come into contact with the vehicle near the estimated position of the sound source, the CPU 21 proceeds to S490. On the other hand, if there is an image of an object that has come into contact with the vehicle near the estimated position of the sound source, the CPU 21 executes a specific operation determination process, which will be described later, at S470.

[0096] At S480, the CPU 21 determines whether or not a specific operation, which will be described later, is being performed based on the determination result at S470. If a specific operation is being performed, the CPU 21 proceeds to S490.

[0097] When the process proceeds to S490, the CPU 21 determines that the candidate collision sound detected in S10 is not caused by a collision, and ends the collision determination process.

[0098] If it is determined in S480 that a specific operation is not being performed, the CPU 21 proceeds to S500 to execute a filter setting process, which will be described later.

[0099] In S510, the CPU 21 analyzes the waveform of the sound wave for which the filter setting process in S500 has been executed, thereby determining whether or not the sound has characteristics of a collision sound or a non-collision sound.

[0100] At S520, the CPU 21 determines whether the sound waves around the sound generation timing have the characteristics of a non-collision sound based on the determination result of S510. If the sound waves have the characteristics of a non-collision sound, the CPU 21 proceeds to S550. Examples of non-collision sounds include vehicle operation sounds generated by operations performed by the driver inside the vehicle (e.g., the sound when operating the parking brake), and talking.

[0101] On the other hand, if the sound does not have the characteristics of a non-collision sound, the CPU 21 determines in S530 whether or not the sound waves around the sound generation timing have the characteristics of a collision sound based on the determination result of S510. If the sound does not have the characteristics of a collision sound, the CPU 21 proceeds to S550. If the sound does have the characteristics of a collision sound, the CPU 21 proceeds to S540.

[0102] When the process proceeds to S540, the CPU 21 determines that the collision candidate sound detected in S10 is caused by a collision, and ends the collision determination process.

[0103] When the process proceeds to S550, the CPU 21 determines that the candidate collision sound detected in S10 is not caused by a collision, and ends the collision determination process.

[0104] Next, the specific operation determination process executed in S470 will be described.

[0105] When the specific operation determination process is executed, the CPU 21 determines at S610 whether it is time to switch the parking brake based on vehicle information around the sound generation timing, as shown in Fig. 10. If it is time to switch the parking brake, the CPU 21 proceeds to S660. On the other hand, if it is not time to switch the parking brake, the CPU 21 determines at S620 whether it is time to switch gears based on vehicle information around the sound generation timing.

[0106] If it is time to change gears, the CPU 21 proceeds to S660. On the other hand, if it is not time to change gears, the CPU 21 proceeds to S630 and determines whether it is time to change windows based on vehicle information around the sound generation timing.

[0107] If it is time to switch the window, the CPU 21 proceeds to S660. On the other hand, if it is not time to switch the window, the CPU 21 determines in S640 whether it is time to switch the door lock based on the vehicle information around the sound generation timing.

[0108] If it is time to switch the door locks, the CPU 21 proceeds to S660. On the other hand, if it is not time to switch the door locks, the CPU 21 proceeds to S650.

[0109] When the process proceeds to S650, the CPU 21 determines that the driver is not performing a specific operation, and ends the specific operation determination process.

[0110] When the process proceeds to S660, the CPU 21 determines that the driver is performing a specific operation, and ends the specific operation determination process.

[0111] Next, the procedure of the filter setting process executed in S500 will be described.

[0112] When the filter setting process is executed, the CPU 21 determines whether the vehicle is backing up or not based on the vehicle information around the sound generation timing, as shown in Fig. 11 . If the vehicle is not backing up, the CPU 21 proceeds to S730. On the other hand, if the vehicle is backing up, the CPU 21 performs a filter process on the sound wave data around the sound generation timing in S720 to remove the backing warning sound, and then proceeds to S730.

[0113] In S730, the CPU 21 determines whether the vehicle's turn signals are on or not based on the vehicle information around the sound generation timing. If the vehicle's turn signals are not on, the CPU 21 proceeds to S750. On the other hand, if the vehicle's turn signals are on, the CPU 21 performs filtering on the sound wave data around the sound generation timing in S740 to remove the turn signal sound, and then proceeds to S750.

[0114] In S750, the CPU 21 determines whether the driver has operated the parking brake based on the vehicle information around the sound generation timing. If the driver has not operated the parking brake, the CPU 21 proceeds to S770. On the other hand, if the driver has operated the parking brake, the CPU 21 performs filtering on the sound wave data around the sound generation timing in S760 to remove the parking brake switching sound, and then proceeds to S770.

[0115] In S770, the CPU 21 determines whether the driver has opened or closed a window based on the vehicle information around the sound generation timing. If the driver has not opened or closed a window, the CPU 21 proceeds to S790. On the other hand, if the driver has opened or closed a window, the CPU 21 performs a filter process on the sound wave data around the sound generation timing in S780 to remove the window opening or closing sound, and then proceeds to S790.

[0116] In S790, the CPU 21 determines whether the driver has performed a door lock operation based on the vehicle information around the sound generation timing. If the driver has not performed a door lock operation, the CPU 21 proceeds to S810. On the other hand, if the driver has performed a door lock operation, the CPU 21 performs a filter process on the sound wave data around the sound generation timing in S800 to remove door lock opening / closing sounds, and then proceeds to S810.

[0117] In S810, the CPU 21 determines whether the estimated sound source position is near the driver's head or not, and whether the estimated sound source position is near the in-vehicle speaker or not. If the estimated sound source position is neither near the driver's head nor near the in-vehicle speaker, the CPU 21 ends the filter setting process.

[0118] On the other hand, if the estimated sound source position is near the driver's head or near the vehicle speaker, the CPU 21 performs a filter process to remove speaking voices on the sound wave data near the sound generation timing in S820, and then ends the filter setting process.

[0119] The order of the processes of S720, S740, S760, S780, S800, and S820 is not particularly limited.

[0120] The vehicle accident detection device 2 configured in this manner is configured to estimate the position of a sound source (i.e., an estimated sound source position) based on the detection results of the multiple microphones 31 to 34 that are installed in a vehicle and configured to detect sound. Specifically, the vehicle accident detection device 2 is configured to estimate the estimated sound source position based on the difference in phase of the sound waves detected by each of the multiple microphones 31 to 34 (i.e., the difference in sound detection timing). The vehicle accident detection device 2 is configured to determine whether the estimated sound source position is on the vehicle body exterior. The vehicle accident detection device 2 is configured to determine that a collision accident has occurred when it is determined that the estimated sound source position is on the vehicle body exterior.

[0121] Such a vehicle accident detection device 2 can detect a vehicle collision accident using a plurality of microphones 31 to 34 installed in the vehicle.

[0122] The vehicle accident detection device 2 is also configured to determine that the estimated sound source position is on the vehicle body exterior when a predetermined first exterior determination condition or a predetermined second exterior determination condition indicating that the estimated sound source position is near the vehicle body exterior is satisfied. The first exterior determination condition is set based on vehicle information to indicate that the estimated sound source position is near the vehicle body exterior. The second exterior determination condition is set based on a plurality of distances from each of the plurality of microphones 31 to 33 to the estimated sound source position.

[0123] The vehicle accident detection device 2 further includes a sound analysis unit 43 configured to analyze the waveform of the sound detected by the multiple microphones 31 to 33. When the vehicle accident detection device 2 determines that the estimated position of the sound source is on the vehicle body exterior, the vehicle accident detection device 2 is further configured to determine that a collision accident has occurred if the sound detected by the multiple microphones 31 to 33 has preset collision sound characteristics based on the analysis results by the sound analysis unit 43. Because the vehicle accident detection device 2 further makes a determination based on the collision sound characteristics, it is possible to improve the accuracy of collision accident determination.

[0124] Furthermore, when it is determined that the estimated sound source position is on the exterior of the vehicle body, the vehicle accident detection device 2 is further configured to determine that a collision accident has occurred if the sounds detected by the multiple microphones 31 to 33 do not have non-collision sound characteristics that are set in advance as characteristics of non-collision sounds, based on the analysis results by the sound analysis unit 43. Since such a vehicle accident detection device 2 further makes a determination based on the non-collision sound characteristics, it is possible to improve the accuracy of collision accident determination.

[0125] The vehicle accident detection device 2 is further configured to apply a filter process to remove non-collision sounds to the sounds detected by the multiple microphones 31 to 33. Such a vehicle accident detection device 2 can prevent the occurrence of a situation in which the occurrence of a non-collision sound is incorrectly determined to be a collision sound, thereby improving the accuracy of collision accident determination.

[0126] The vehicle accident detection device 2 is further configured to determine whether the driver of the vehicle has performed a predetermined specific operation. When the vehicle accident detection device 2 determines that the estimated sound source position is on the vehicle body exterior, and further determines that the driver has performed a specific operation, the vehicle accident detection device 2 is configured to determine that a collision accident has not occurred. This vehicle accident detection device 2 can prevent the occurrence of a situation in which a collision sound is determined to have occurred due to the driver performing a specific operation, thereby improving the accuracy of collision accident determination.

[0127] Furthermore, when the vehicle accident detection device 2 determines that the estimated sound source position is on the vehicle body exterior, if the acceleration at the sound generation timing when the sound is generated from the sound source is less than a preset collision determination acceleration, the vehicle accident detection device 2 is configured to determine that a collision accident has not occurred. Since the vehicle accident detection device 2 further performs determination based on acceleration, it is possible to improve the accuracy of collision accident determination.

[0128] The vehicle accident detection device 2 further includes a video analysis unit 45 configured to estimate the position of a collision sound source, which is the position where an object collided with the vehicle, by analyzing video data generated by cameras 36, 37 mounted on the vehicle and configured to capture at least the exterior of the vehicle. The vehicle accident detection device 2 is further configured to determine that a collision accident has not occurred if, when it has determined that the estimated sound source position is on the vehicle body exterior, an object did not collide with the vehicle at the sound generation timing when the sound was generated by the sound source, based on the analysis result of the video analysis unit 45. Because the vehicle accident detection device 2 further makes a determination based on video data, it is possible to improve the accuracy of collision accident determination.

[0129] Furthermore, the microphones 31 to 34 are installed inside the vehicle, which eliminates the need for protective measures for the microphones 31 to 34 and also allows the microphones 31 to 34 to be used for other purposes.

[0130] In the embodiment described above, the microphones 31 to 34 correspond to the sound detection unit, S210 to S250 correspond to the processing of the sound source position estimation unit, S260 to S330 correspond to the processing of the vehicle body exterior determination unit, and S410, S450, S460, S480, and S520 to S550 correspond to the processing of the collision determination unit.

[0131] Furthermore, the cameras 36 and 37 correspond to the photographing unit, S710 to S820 correspond to the processing of the excluding unit, and S610 to S660 correspond to the processing of the specific operation determining unit.

[0132] Second Embodiment A second embodiment of the present disclosure will be described below with reference to the drawings. In the second embodiment, differences from the first embodiment will be described. The same reference numerals will be used to designate common components.

[0133] As shown in FIG. 12, the accident information collection system 1 of the second embodiment differs from the first embodiment in that the microphone 32 is omitted and the microphone 31 is installed in the center of the dashboard.

[0134] The vehicle accident detection device 2 of the second embodiment detects information other than sound (for example, acceleration in this embodiment) to estimate the timing at which sound is generated at the sound source (i.e., sound generation timing).

[0135] The microphones 31 and 33 detect the sound waves generated by the sound source SS1 with a delay from the timing at which the sound waves are generated by the sound source SS1. This time difference can be regarded as the distance from the sound source SS1.

[0136] Therefore, for each of the microphones 31 and 33, it is possible to calculate circles (hereinafter referred to as distance circles) CL1 and CL3 that pass through multiple points at equal distances to the sound source SS1 based on the distance from the sound source SS1.

[0137] Furthermore, for the pair of microphones 31 and 33, a distance difference curve L11 can be calculated that estimates the position of sound source SS1 based on the difference in distance from sound source SS1.

[0138] The vehicle accident detection device 2 then estimates the point where the distance circle CL1 and the distance difference curve L11 intersect, or the point where the distance circle CL2 and the distance difference curve L11 intersect, or the area surrounded by the distance circle CL1, the distance circle CL2, and the distance difference curve L11, as the two-dimensional position of the sound source SS1.

[0139] If there is no measurement error in the distance circles CL1, CL2 and the distance difference curve L11, the distance circles CL1, CL2 and the distance difference curve L11 intersect at the position of the sound source SS1. In this case, the vehicle accident detection device 2 can estimate the intersection of the distance circles CL1, CL2 and the distance difference curve L11 as the two-dimensional position of the sound source SS1.

[0140] Based on the same concept, when the microphone 34 is used in addition to the microphones 31 and 33, the vehicle accident detection device 2 can estimate the three-dimensional position of the sound source SS1.

[0141] As shown in Figure 13, the vehicle accident detection device 2 of the second embodiment includes functional blocks realized by the CPU 21 executing a program stored in the ROM 22, such as a vehicle information acquisition unit 41, a sound acquisition unit 42, a sound analysis unit 43, a video acquisition unit 44, a video analysis unit 45, an accident detection control unit 46, a collision determination unit 47, and an upload unit 48.

[0142] The vehicle information acquisition unit 41 constantly acquires vehicle information from a plurality of ECUs via the CAN communication unit 12, as indicated by an arrow A31.

[0143] Furthermore, the vehicle information acquisition unit 41 constantly acquires acceleration information indicating the acceleration detected by the acceleration sensor 35 mounted on the vehicle, as indicated by an arrow A32.

[0144] The sound acquisition unit 42 constantly acquires sonic electrical signals from the microphones 31 and 33 as indicated by arrows A33 and A34.

[0145] As indicated by arrows A35 and A36, the image acquisition unit 44 constantly acquires image data generated by the cameras 36 and 37 mounted on the vehicle.

[0146] The accident detection control unit 46 acquires vehicle information from the vehicle information acquisition unit 41 at the timing when the acceleration exceeds the threshold, as indicated by an arrow A37.

[0147] The accident detection control unit 46 acquires, from the image acquisition unit 44, image data at the timing when the acceleration exceeds the threshold value, as indicated by an arrow A38.

[0148] The accident detection control unit 46 acquires sound wave data from the sound acquisition unit 42 at the timing when the acceleration exceeds the threshold value, as indicated by an arrow A39.

[0149] After acquiring the vehicle information, the video data, and the sound wave data, the accident detection control unit 46 outputs a collision determination request to the collision determination unit 47, as indicated by an arrow A40.

[0150] When the collision determination unit 47 receives a collision determination request from the accident detection control unit 46, it outputs a sound source position estimation request to the sound analysis unit 43, as indicated by an arrow A41.

[0151] When the sound analysis unit 43 receives a sound source position estimation request from the collision determination unit 47, it estimates the position of the sound source based on the timing at which the acceleration exceeds a threshold and the timing at which the sound wave is detected by the sound acquisition unit 42. The sound analysis unit 43 notifies the collision determination unit 47 of the estimated sound source position, as indicated by arrow A42.

[0152] When the collision determination unit 47 receives a collision determination request from the accident detection control unit 46, it outputs a contact determination request to the video analysis unit 45 as indicated by an arrow A43.

[0153] When the video analysis unit 45 receives a contact determination request from the collision determination unit 47, it analyzes the video data at the time when the acceleration exceeds a threshold value and determines whether or not an object has come into contact with the vehicle. The video analysis unit 45 notifies the collision determination unit 47 of the determination result as to whether or not an object has come into contact with the vehicle, as shown by arrow A44.

[0154] When the collision determination unit 47 receives a request for collision determination from the accident detection control unit 46, it obtains vehicle body information from the vehicle body information storage unit 49, as indicated by an arrow A45.

[0155] The collision determination unit 47 determines whether an object has collided with the vehicle based on the estimated sound source position obtained from the sound analysis unit 43, the determination result obtained from the video analysis unit 45, and the vehicle body information obtained from the vehicle body information storage unit 49.

[0156] The collision determination unit 47 notifies the accident detection control unit 46 of the determination result as to whether or not an object has collided with the vehicle, as indicated by an arrow A46.

[0157] When an object collides with the vehicle, the accident detection control unit 46 generates accident information including the acquired vehicle information and video data based on the determination result acquired from the collision determination unit 47. The accident detection control unit 46 stores the generated accident information in the temporary storage unit 50, as indicated by arrow A47.

[0158] After storing the accident information in the temporary storage unit 50, the accident detection control unit 46 outputs an upload instruction to the upload unit 48, as indicated by an arrow A48.

[0159] When the upload unit 48 receives an upload instruction from the accident detection control unit 46, it retrieves the accident information from the temporary storage unit 50, as shown by arrow A49. The upload unit 48 uploads the retrieved accident information to the accident information server 3.

[0160] Next, a procedure for detecting an accident according to the second embodiment will be described. For the sake of simplicity, the present embodiment will show a form in which an accident is detected using the microphones 31 and 33.

[0161] When the accident detection process of the second embodiment is executed, the CPU 21 of the control unit 11 determines whether the detected acceleration is equal to or greater than a preset collision determination acceleration in S910, as shown in Fig. 14. If the detected acceleration is less than the collision determination acceleration, the CPU 21 ends the accident detection process.

[0162] On the other hand, if the acceleration is equal to or greater than the collision determination acceleration, the CPU 21 acquires vehicle information around the time when the acceleration becomes equal to or greater than the collision determination acceleration (hereinafter referred to as the acceleration occurrence time) at S920. The time around the acceleration occurrence time is, for example, a time range from a predetermined acquisition time before the acceleration occurrence time (e.g., one second before) to a predetermined acquisition time after the acceleration occurrence time (e.g., one second after).

[0163] In S930, the CPU 21 acquires video data around the timing at which acceleration occurs.

[0164] In S940, the CPU 21 acquires sound wave data around the timing of acceleration occurrence.

[0165] In S950, the CPU 21 executes a sound source position estimation process, which will be described later.

[0166] In S960, the CPU 21 executes the collision determination process in the same manner as in S70.

[0167] At S970, the CPU 21 determines whether or not a collision has occurred based on the determination result of S960. If a collision has not occurred, the CPU 21 ends the accident detection process.

[0168] On the other hand, if a collision has occurred, the CPU 21 generates the above-mentioned accident information and stores the generated accident information in the temporary storage unit 50 in S980.

[0169] In S990, the CPU 21 retrieves the accident information generated in S980 from the temporary storage unit 50, and uploads the retrieved accident information to the accident information server 3.

[0170] When the process of S990 ends, the CPU 21 ends the accident detection process.

[0171] Next, the procedure of the sound source position estimation process executed in S950 will be described.

[0172] When the sound source position estimation process of the second embodiment is executed, the CPU 21 determines in S1110 whether or not the above-described distance circles have been calculated for each of the microphones 31 and 33, as shown in FIG.

[0173] If the distance circle has not been calculated, the CPU 21 calculates the timing at which sound waves are detected by each of the microphones 31 and 33 (hereinafter referred to as sound wave detection timing) at S1120.

[0174] At S1130, the CPU 21 calculates the distance between the microphones 31 and 33 and the sound source for each of the microphones 31 and 33 based on the time difference between the acceleration generation timing and the sound wave detection timing.

[0175] In S1140, the CPU 21 calculates an equation representing a distance circle passing through a plurality of points that are the same distance to the sound source for each of the microphones 31 and 33 based on the distances calculated in S1130, and then proceeds to S1110.

[0176] If the distance circle is calculated in S1110, the CPU 21 calculates the phase difference between the two sound waves detected by the microphones 31 and 33 in S1150.

[0177] In S1160, the CPU 21 calculates the difference in distance from the microphones 31 and 33 to the sound source based on the phase difference calculated in S1150.

[0178] In S1170, the CPU 21 calculates an equation representing a distance difference curve that passes through a plurality of points where the difference in distance from the microphones 31 and 33 to the sound source is equal, based on the distance calculated in S1160.

[0179] In S1180, the CPU 21 estimates the coordinate position where the two distance difference curves and one distance difference curve are closest to each other as the estimated sound source position.

[0180] In S1190 , the CPU 21 acquires the vehicle body information from the vehicle body information storage unit 49 .

[0181] In S1200, the CPU 21 determines whether the first contour determination condition is met, in the same manner as in S280.

[0182] If the first contour determination condition is met, the CPU 21 determines in S1210 that the estimated sound source position estimated in S1180 is near the vehicle contour, and ends the sound source position estimation process.

[0183] On the other hand, if the first contour determination condition is not met, the CPU 21 determines in S1220 that the estimated sound source position estimated in S1180 is not near the vehicle contour, and ends the sound source position estimation process.

[0184] The vehicle accident detection device 2 configured in this manner is configured to detect the timing when the acceleration detected by the acceleration sensor 35 mounted on the vehicle becomes equal to or greater than the collision determination acceleration (i.e., the acceleration occurrence timing) as the sound occurrence timing when sound is generated from the sound source.

[0185] The vehicle accident detection device 2 is configured to estimate an estimated sound source position based on the difference between the sound generation timing and two sound detection timings at which the microphones 31, 33 detect the sound. The vehicle accident detection device 2 is configured to determine whether the estimated sound source position is on the exterior of the vehicle body. The vehicle accident detection device 2 is configured to determine that a collision accident has occurred when it is determined that the estimated sound source position is on the exterior of the vehicle body.

[0186] Such a vehicle accident detection device 2 can detect a collision accident in a vehicle using a plurality of microphones 31, 33 installed in the vehicle.

[0187] The vehicle accident detection device 2 is further configured to estimate the estimated position of the sound source based on the difference between the two sound detection timings at which the sound is detected by the two microphones 31 and 33. This enables the vehicle accident detection device 2 to improve the accuracy of estimating the estimated position of the sound source.

[0188] In the embodiment described above, steps S1110 to S1180 correspond to the processing performed by the sound source position estimation unit, steps S1190 to S1220 correspond to the processing performed by the vehicle body contour determination unit, and step S910 corresponds to the processing performed by the sound generation timing detection unit.

[0189] Third Embodiment A third embodiment of the present disclosure will be described below with reference to the drawings. In the third embodiment, differences from the first embodiment will be described. The same reference numerals will be used to designate common components.

[0190] 16 , the accident information collection system 1 of the third embodiment differs from the first embodiment in that the microphone 32 is omitted and the microphone 31 is installed in the center of the dashboard. The camera 36 is installed, for example, above the center of the dashboard to capture images of the rear of the vehicle. The camera 37 is installed, for example, above the rear of the luggage compartment of the vehicle to capture images of the front of the vehicle.

[0191] The vehicle accident detection device 2 of the third embodiment determines whether or not an object has come into contact with the vehicle by analyzing video data captured by the cameras 36, 37. If an object has come into contact with the vehicle, the vehicle accident detection device 2 estimates the position where the object came into contact with the vehicle as the estimated sound source position, and estimates the timing when the object came into contact with the vehicle as the sound generation timing. If the first boundary determination condition is satisfied for the estimated sound source position, the vehicle accident detection device 2 determines that the estimated sound source position is near the vehicle boundary.

[0192] As shown in Figure 17, the vehicle accident detection device 2 includes functional blocks that are realized by the CPU 21 executing a program stored in the ROM 22, such as a vehicle information acquisition unit 41, a sound acquisition unit 42, a sound analysis unit 43, a video acquisition unit 44, a video analysis unit 45, an accident detection control unit 46, a collision determination unit 47, and an upload unit 48.

[0193] The vehicle information acquisition unit 41 constantly acquires vehicle information from a plurality of ECUs via the CAN communication unit 12, as indicated by an arrow A51.

[0194] Furthermore, the vehicle information acquisition unit 41 constantly acquires acceleration information indicating the acceleration detected by the acceleration sensor 35 mounted on the vehicle, as indicated by an arrow A52.

[0195] The sound acquisition unit 42 constantly acquires sonic electrical signals from the microphones 31 and 33 as indicated by arrows A53 and A54.

[0196] As indicated by arrows A55 and A56, the image acquisition unit 44 constantly acquires image data generated by the cameras 36 and 37 mounted on the vehicle.

[0197] The video acquisition unit 44 outputs the acquired video data to the video analysis unit 45, as indicated by an arrow A57.

[0198] The video analysis unit 45 analyzes the video data acquired by the video acquisition unit 44 and determines whether or not an object has come into contact with the vehicle. If the video analysis unit 45 determines that an object has come into contact with the vehicle, the video analysis unit 45 notifies the video acquisition unit 44 of the contact timing of the object with the vehicle, as shown by arrow A58.

[0199] The image acquisition unit 44 notifies the accident detection control unit 46 of the contact timing, as indicated by an arrow A59.

[0200] When the accident detection control unit 46 is notified of the contact timing by the image acquisition unit 44, it acquires the vehicle information, acceleration information, and sound wave data at the contact timing, as shown by arrows A60 and A61.

[0201] After acquiring the vehicle information, acceleration information, and sound wave data, the accident detection control unit 46 outputs a collision determination request to the collision determination unit 47, as indicated by an arrow A62.

[0202] When the collision determination unit 47 receives a request for collision determination from the accident detection control unit 46, it obtains vehicle body information from the vehicle body information storage unit 49, as indicated by an arrow A63.

[0203] The collision determination unit 47 outputs a contact position provision request to the video analysis unit 45, as indicated by an arrow A64, requesting provision of the contact position where the object came into contact with the vehicle.

[0204] When the video analysis unit 45 receives a contact position request from the collision determination unit 47, it notifies the collision determination unit 47 of the contact position as shown by arrow A65.

[0205] The collision determination unit 47 outputs a sound wave analysis request to the sound analysis unit 43 to request analysis of the characteristics of the waveform of the sound wave at the contact timing, as indicated by an arrow A66.

[0206] When the sound analysis unit 43 receives the sound wave analysis request from the collision determination unit 47, it outputs a sound wave data provision request to the sound acquisition unit 42, as shown by arrow A67, requesting the provision of sound wave data at the contact timing.

[0207] When the sound acquisition unit 42 acquires the request to provide sound wave data from the sound analysis unit 43, it outputs the sound wave data at the contact timing to the sound analysis unit 43, as shown by arrow A68.

[0208] When the sound analysis unit 43 acquires the sound wave data from the sound acquisition unit 42, it analyzes the characteristics of the waveform of the sound wave and notifies the collision determination unit 47 of the analysis result of the sound wave data, as shown by arrow A69.

[0209] The collision determination unit 47 determines whether an object has collided with the vehicle based on the vehicle information, acceleration information, contact position, vehicle body information, and waveform characteristics of the sound wave. The collision determination unit 47 notifies the accident detection control unit 46 of the determination result as to whether an object has collided with the vehicle, as indicated by arrow A70.

[0210] When an object collides with the vehicle, the accident detection control unit 46 generates accident information including the acquired vehicle information and video data based on the determination result acquired from the collision determination unit 47. The accident detection control unit 46 stores the generated accident information in the temporary storage unit 50, as indicated by arrow A71.

[0211] After storing the accident information in the temporary storage unit 50, the accident detection control unit 46 outputs an upload instruction to the upload unit 48, as indicated by an arrow A72.

[0212] When the upload unit 48 receives an upload instruction from the accident detection control unit 46, it retrieves the accident information from the temporary storage unit 50, as shown by arrow A73. The upload unit 48 uploads the retrieved accident information to the accident information server 3.

[0213] Next, the procedure of the accident detection process of the third embodiment will be described.

[0214] 18, when the accident detection process of the third embodiment is executed, the CPU 21 of the control unit 11 determines whether or not an object has come into contact with the vehicle by analyzing the video data generated by the cameras 36 and 37. If no object has come into contact with the vehicle, the CPU 21 ends the accident detection process.

[0215] On the other hand, if an object has come into contact with the vehicle, the CPU 21 determines the timing of contact of the object with the vehicle and the contact position of the object with the vehicle at S1320 based on video data (hereinafter referred to as contact video data) capturing the situation in which the object has come into contact with the vehicle.

[0216] At S1330, the CPU 21 acquires vehicle information around the contact timing. The term "around the contact timing" refers to, for example, a time range from a predetermined acquisition time before the contact timing (e.g., one second before) to a predetermined acquisition time after the contact timing (e.g., one second after).

[0217] In S1340, the CPU 21 acquires video data around the contact timing.

[0218] In S1350, the CPU 21 acquires sound wave data around the contact timing.

[0219] In S1360, the CPU 21 estimates the contact position identified in S1320 as the position of the collision sound source.

[0220] At S1370, the CPU 21 executes a collision determination process, which will be described later.

[0221] At S1380, the CPU 21 determines whether or not a collision has occurred based on the determination result of S1370. If a collision has not occurred, the CPU 21 ends the accident detection process.

[0222] On the other hand, if a collision has occurred, the CPU 21 generates the above-mentioned accident information in S1390 and stores the generated accident information in the temporary storage unit 50.

[0223] In S1400, the CPU 21 retrieves the accident information generated in S1390 from the temporary storage unit 50, and uploads the retrieved accident information to the accident information server 3.

[0224] When the process of S1400 ends, the CPU 21 ends the accident detection process.

[0225] Next, the procedure of the collision determination process executed in S1370 will be described.

[0226] When the collision determination process of the third embodiment is executed, the CPU 21 determines at S1510 whether the collision sound source position is near the vehicle body exterior, as shown in Fig. 19. Specifically, the CPU 21 determines whether the first exterior boundary determination condition is met, in the same manner as S1200 in the second embodiment. If the first exterior boundary determination condition is met, the CPU 21 determines that the collision sound source position is near the vehicle body exterior, and if the first exterior boundary determination condition is not met, the CPU 21 determines that the collision sound source position is not near the vehicle body exterior.

[0227] If the collision sound source position is not near the exterior of the vehicle body, the CPU 21 proceeds to S1600.

[0228] On the other hand, if the collision sound source position is near the exterior of the vehicle body, the CPU 21 acquires acceleration information around the time of contact at S1520.

[0229] At S1530, the CPU 21 determines whether the acceleration indicated by the acceleration information acquired at S1520 is equal to or greater than a preset collision determination acceleration. If the acceleration is less than the collision determination acceleration, the CPU 21 proceeds to S1600.

[0230] On the other hand, if the acceleration is equal to or greater than the collision determination acceleration, the CPU 21 executes the specific operation determination process at S1540 in the same manner as at S470.

[0231] At S1550, the CPU 21 determines whether or not a specific operation is being performed based on the determination result at S1540. If a specific operation is being performed, the CPU 21 proceeds to S1600.

[0232] On the other hand, if the specific operation is not being performed, the CPU 21 executes the filter setting process in S1560 in the same manner as in S500.

[0233] At S1570, the CPU 21 analyzes the waveform of the sound wave for which the filter setting process at S1560 has been executed, thereby determining whether or not the sound has characteristics of a collision sound or a non-collision sound.

[0234] At S1580, the CPU 21 determines whether or not the sound waves around the contact timing have the characteristics of a non-collision sound based on the determination result of S1570. If the sound waves have the characteristics of a non-collision sound, the CPU 21 proceeds to S1600.

[0235] On the other hand, if the sound wave does not have the characteristics of a non-collision sound, the CPU 21 determines at S1590 whether or not the sound wave near the contact timing has the characteristics of a collision sound based on the determination result of S1570. If the sound wave does not have the characteristics of a collision sound, the CPU 21 proceeds to S1600. On the other hand, if the sound wave has the characteristics of a collision sound, the CPU 21 proceeds to S1610.

[0236] When the process proceeds to S1600, the CPU 21 determines that a collision has not occurred and ends the collision determination process.

[0237] When the process proceeds to S1610, the CPU 21 determines that a collision has occurred and ends the collision determination process.

[0238] Although one embodiment of the present disclosure has been described above, the present disclosure is not limited to the above embodiment and can be implemented in various modifications.

[0239] [Variant 1] In the above second embodiment, the point where the distance circle CL1 and the distance difference curve L11 intersect, or the point where the distance circle CL2 and the distance difference curve L11 intersect, or the area surrounded by the distance circle CL1, the distance circle CL2, and the distance difference curve L11 is estimated as the two-dimensional position of the sound source SS1.

[0240] However, as shown in Figure 20, microphones 31 and 33 detect the sound waves generated by the sound source with a delay from the timing at which the sound waves are generated by the sound source. This time difference can be regarded as the distance from the sound source. Therefore, for microphones 31 and 33, distance circles CL1 and CL3 that estimate the position of the sound source can be calculated based on the distance from the sound source.

[0241] When the sound source position is estimated using only the distance circles CL1 and CL3, the two intersection points IP1 and IP2 between the distance circles CL1 and CL3 become candidate sound source positions. In this way, when there are two candidate sound source positions, the sound source may be determined to be located on the vehicle body exterior if the following first collision determination condition or second collision determination condition is met. The first collision determination condition is that both candidates are located on the vehicle body exterior. The second collision determination condition is that one of the candidates is located on the vehicle body exterior and the waveform of the sound wave detected by the microphones 31 and 33 has the characteristics of a typical collision sound.

[0242] In the above embodiment, the sound analysis unit 43, the video analysis unit 45, the accident detection control unit 46, and the collision determination unit 47 are mounted on a vehicle. However, the accident information server 3 may be provided with the sound analysis unit 43, the video analysis unit 45, the accident detection control unit 46, and the collision determination unit 47.

[0243] In the second embodiment, the sound generation timing is detected based on the acceleration detected by the acceleration sensor 35 mounted on the vehicle. However, the sound generation timing may be detected based on video data generated by cameras 36 and 37 mounted on the vehicle and configured to capture at least the exterior of the vehicle.

[0244] [Variation 4] In the above embodiment, the estimated sound source position is determined to be on the vehicle body exterior when the first and second preset exterior boundary determination conditions, which indicate that the estimated sound source position is near the vehicle body exterior boundary, are satisfied. However, instead of the estimated sound source position, the collision sound source position estimated by the video analysis unit 45 may be used to determine whether the first and second exterior boundary determination conditions are satisfied. In this case, the video analysis unit 45 estimates the collision sound source position when an object collides with the vehicle based on the video data generated by the cameras 36 and 37.

[0245] In the above embodiment, the sound analysis unit 43 determines whether the sound wave around the sound generation timing has the characteristics of a non-collision sound. However, the sound analysis unit 43 may determine whether the sound detected by the multiple microphones 31 to 34 is a speaking voice when the estimated sound source position is near the head of the driver of the vehicle or near a speaker mounted on the vehicle.

[0246] The control unit 11 and the method described herein may be implemented by a special-purpose computer configured by configuring a processor and memory programmed to execute one or more functions embodied in a computer program. Alternatively, the control unit 11 and the method described herein may be implemented by a special-purpose computer configured by configuring a processor with one or more dedicated hardware logic circuits. Alternatively, the control unit 11 and the method described herein may be implemented by one or more special-purpose computers configured by combining a processor and memory programmed to execute one or more functions with a processor configured with one or more hardware logic circuits. Furthermore, the computer program may be stored as instructions executed by a computer on a computer-readable non-transitory tangible recording medium. The method for implementing the functions of each unit included in the control unit 11 does not necessarily need to include software; all of the functions may be implemented using one or more hardware components.

[0247] In the above embodiments, multiple functions of one component may be realized by multiple components, or one function of one component may be realized by multiple components. Furthermore, multiple functions of multiple components may be realized by one component, or one function realized by multiple components may be realized by one component. Furthermore, part of the configuration of the above embodiments may be omitted. Furthermore, at least part of the configuration of the above embodiments may be added to or substituted for the configuration of another of the above embodiments.

[0248] In addition to the vehicle accident detection device 2 described above, the present disclosure can also be realized in various forms, such as a system including the vehicle accident detection device 2 as a component, a program for causing a computer to function as the vehicle accident detection device 2, a non-transient physical recording medium such as a semiconductor memory on which this program is recorded, and a vehicle accident detection method. [Technical Ideas Disclosed in the Present Specification] [Item 1] A vehicle accident detection device (2) comprising: a sound source position estimation unit (S210 to S250, S1110 to S1180) configured to estimate an estimated sound source position, which is the position of a sound source, based on detection results of a plurality of sound detection units (31 to 34) installed in a vehicle and configured to detect sound; a vehicle body shell determination unit (S260 to S330, S1190 to S1220) configured to determine whether the estimated sound source position is on the vehicle body shell of the vehicle; and a collision determination unit (S410, S450, S460, S480, S520 to S550) configured to determine that a collision accident has occurred when the vehicle body shell determination unit determines that the estimated sound source position is on the vehicle body shell.

[0249] [Item 2] The vehicle accident detection device according to Item 1, wherein the sound source position estimation unit (S210 to S250) is configured to estimate the estimated sound source position based on a difference in sound detection timing at which each of the three or more sound detection units detects a sound.

[0250] [Item 3] The vehicle accident detection device according to Item 2, wherein the sound source position estimation unit is configured to estimate a two-dimensional position of the sound source as the estimated sound source position using three or more of the sound detection units.

[0251] [Item 4] The vehicle accident detection device according to Item 2, wherein the sound source position estimation unit is configured to estimate a three-dimensional position of the sound source as the estimated sound source position using four or more of the sound detection units.

[0252] [Item 5] The vehicle accident detection device according to Item 1, further comprising a sound generation timing detection unit (S910) configured to detect a sound generation timing at which a sound is generated from the sound source, and the sound source position estimation unit (S1110 to S1180) is configured to estimate the estimated sound source position based on a difference between the sound generation timing detected by the sound generation timing detection unit and two or more sound detection timings at which two or more sound detection units respectively detect sound.

[0253] [Item 6] The vehicle accident detection device according to Item 5, wherein the sound source position estimation unit is configured to estimate a two-dimensional position of the sound source as the estimated sound source position using two or more of the sound detection units.

[0254] [Item 7] The vehicle accident detection device according to Item 5, wherein the sound source position estimation unit is configured to estimate a three-dimensional position of the sound source as the estimated sound source position using three or more of the sound detection units.

[0255] [Item 8] The vehicle accident detection device according to any one of items 5 to 7, wherein the sound generation timing detection unit is configured to detect the sound generation timing based on acceleration detected by an acceleration sensor mounted on the vehicle.

[0256] [Item 9] The vehicle accident detection device according to any one of Items 5 to 7, wherein the sound generation timing detection unit is configured to detect the sound generation timing based on video data generated by an imaging unit (36, 37) that is mounted on the vehicle and configured to capture at least an image of the outside of the vehicle.

[0257] [Item 10] The vehicle accident detection device according to any one of items 5 to 9, wherein the sound source position estimation unit is further configured to estimate the estimated sound source position based on a difference between two or more sound detection timings at which the two or more sound detection units respectively detected sounds.

[0258] [Item 11] The vehicle accident detection device according to any one of items 1 to 10, wherein the vehicle body shell determination unit is configured to determine that the estimated sound source position is on the vehicle body shell when a preset shell determination condition indicating that the estimated sound source position is near the vehicle body shell is satisfied.

[0259] [Item 12] The vehicle accident detection device according to Item 11, wherein the contour determination condition is set based on a plurality of distances from each of the plurality of sound detection units to the sound source estimated position.

[0260] [Item 13] The vehicle accident detection device according to Item 11, further comprising a vehicle body information storage unit (49) configured to store vehicle body information indicating a plurality of coordinate positions on the vehicle body exterior, wherein the exterior determination condition is set based on the vehicle body information to indicate that the estimated sound source position is near the vehicle body exterior.

[0261] [Item 14] The vehicle accident detection device according to Item 11, further comprising: a video analysis unit (45) configured to estimate a collision sound source position, which is the position where an object collided with the vehicle, by analyzing video data generated by an imaging unit (36, 37) mounted on the vehicle and configured to capture at least the exterior of the vehicle; and a vehicle body information storage unit (49) configured to store vehicle body information indicating a plurality of coordinate positions on the vehicle body exterior, wherein the exterior shell determination condition is set to indicate that the collision sound source position is near the vehicle body exterior, based on the vehicle body information.

[0262] [Item 15] The vehicle accident detection device according to any one of items 1 to 14, further comprising a sound analysis unit (43) configured to analyze waveforms of sounds detected by the plurality of sound detection units, wherein the collision determination unit is configured to determine that the collision accident has occurred based on the analysis result by the sound analysis unit when the vehicle body shell determination unit determines that the estimated sound source position is on the vehicle body shell and the sounds detected by the plurality of sound detection units have collision sound characteristics that are preset as characteristics of a collision sound.

[0263] [Item 16] The vehicle accident detection device according to any one of items 1 to 15, further comprising a sound analysis unit (43) configured to analyze waveforms of sounds detected by the plurality of sound detection units, wherein the collision determination unit is configured to determine that the collision accident has occurred based on the analysis result by the sound analysis unit when the vehicle body shell determination unit determines that the estimated sound source position is on the vehicle body shell and the sounds detected by the plurality of sound detection units do not have non-collision sound characteristics that are preset as characteristics of non-collision sounds.

[0264] [Item 17] The vehicle accident detection device according to Item 16, wherein the non-collision sound includes a vehicle operation sound generated when a driver of the vehicle operates the vehicle.

[0265] [Item 18] The vehicle accident detection device according to item 16 or 17, wherein the non-collision sound includes a voice.

[0266] [Item 19] The vehicle accident detection device according to Item 18, wherein the sound analysis unit is configured to determine whether or not the sound detected by the plurality of sound detection units is the speaking voice when the estimated sound source position is near the head of a driver of the vehicle or near a speaker mounted on the vehicle.

[0267] [Item 20] The vehicle accident detection device according to any one of items 15 to 19, further comprising an exclusion unit (S710 to S820) configured to apply a filter process to remove non-collision sounds from the sounds detected by the plurality of sound detection units.

[0268] [Item 21] The vehicle accident detection device according to any one of items 1 to 20, further comprising a specific operation determination unit (S610 to S660) configured to determine whether or not a driver of the vehicle has performed a predetermined specific operation, wherein the collision determination unit is configured to determine that the collision accident has not occurred if the vehicle body exterior determination unit determines that the estimated sound source position is on the vehicle body exterior and, based on the determination result by the specific operation determination unit, if the driver has performed the specific operation.

[0269] [Item 22] The vehicle accident detection device according to any one of items 1 to 21, wherein the collision determination unit is configured to determine that the collision accident has not occurred if the vehicle body shell determination unit determines that the estimated sound source position is on the vehicle body shell, and if the acceleration at the sound generation timing when the sound is generated from the sound source is less than a preset collision determination acceleration.

[0270] [Item 23] The vehicle accident detection device according to any one of Items 1 to 22, further comprising a video analysis unit (45) configured to estimate a collision sound source position, which is a position where an object collided with the vehicle, by analyzing video data generated by an imaging unit (36, 37) mounted on the vehicle and configured to capture at least an image of the outside of the vehicle, wherein the collision determination unit is further configured to determine that the collision accident has not occurred if the vehicle body exterior determination unit determines that the estimated sound source position is on the vehicle body exterior and, based on the analysis result of the video analysis unit, if an object did not collide with the vehicle at the sound generation timing when sound was generated from the sound source.

[0271] [Item 24] The vehicle accident detection device according to any one of items 1 to 23, wherein the plurality of sound detection units are installed inside the vehicle.

Claims

1. a sound source position estimation unit (S210 to S250, S1110 to S1180) configured to estimate a sound source estimated position, which is the position of a sound source, based on detection results of a plurality of sound detection units (31 to 34) installed in a vehicle and configured to detect sound; a vehicle body exterior determination unit (S260 to S330, S1190 to S1220) configured to determine whether the estimated sound source position is on the vehicle body exterior; a collision determination unit (S410, S450, S460, S480, S520 to S550) configured to determine that a collision accident has occurred when the vehicle body exterior determination unit determines that the sound source estimated position is on the vehicle body exterior; Equipped with Further provided is a sound analysis unit (43) configured to analyze the waveforms of the sounds detected by the plurality of sound detection units, the collision determination unit is configured to determine that the collision accident has occurred when the vehicle body exterior determination unit determines that the sound source estimated position is on the vehicle body exterior, and further based on the analysis result by the sound analysis unit, when the sounds detected by the plurality of sound detection units do not have non-collision sound characteristics that are set in advance as characteristics of non-collision sounds, the non-collision sounds include speech; The sound analysis unit is configured to determine whether the sound detected by the plurality of sound detection units is the speaking voice when the estimated sound source position is near the head of the driver of the vehicle or near a speaker mounted on the vehicle.

2. 2. The vehicle accident detection device according to claim 1, The sound source position estimation unit (S210 to S250) is configured to estimate the sound source position based on the difference in sound detection timing at which each of the three or more sound detection units detects a sound.

3. 3. The vehicle accident detection device according to claim 2, The vehicle accident detection device is configured such that the sound source position estimation unit uses three or more of the sound detection units to estimate a two-dimensional position of the sound source as the estimated sound source position.

4. 3. The vehicle accident detection device according to claim 2, The vehicle accident detection device is configured such that the sound source position estimation unit uses four or more of the sound detection units to estimate a three-dimensional position of the sound source as the estimated sound source position.

5. 2. The vehicle accident detection device according to claim 1, The apparatus further includes a sound generation timing detection unit (S910) configured to detect a sound generation timing when a sound is generated by the sound source, The sound source position estimation unit (S1110 to S1180) is configured to estimate the sound source position based on the difference between the sound occurrence timing detected by the sound occurrence timing detection unit and two or more sound detection timings at which the two or more sound detection units respectively detected sounds. A vehicle accident detection device.

6. 6. The vehicle accident detection device according to claim 5, The vehicle accident detection device is configured such that the sound source position estimation unit uses two or more of the sound detection units to estimate a two-dimensional position of the sound source as the estimated sound source position.

7. 6. The vehicle accident detection device according to claim 5, The vehicle accident detection device is configured such that the sound source position estimation unit uses three or more of the sound detection units to estimate a three-dimensional position of the sound source as the estimated sound source position.

8. The vehicle accident detection device according to any one of claims 5 to 7, The vehicle accident detection device is configured so that the sound occurrence timing detection unit detects the sound occurrence timing based on acceleration detected by an acceleration sensor mounted on the vehicle.

9. The vehicle accident detection device according to any one of claims 5 to 7, The sound generation timing detection unit is configured to detect the sound generation timing based on video data generated by an imaging unit (36, 37) mounted on the vehicle and configured to capture at least the exterior of the vehicle.

10. The vehicle accident detection device according to any one of claims 5 to 7, The sound source position estimation unit is further configured to estimate the estimated sound source position based on a difference between two or more sound detection timings at which the two or more sound detection units respectively detected sounds.

11. The vehicle accident detection device according to any one of claims 1 to 7, The vehicle accident detection device is configured so that the vehicle body exterior determination unit determines that the estimated sound source location is on the vehicle body exterior when a predetermined exterior determination condition indicating that the estimated sound source location is near the vehicle body exterior is satisfied.

12. The vehicle accident detection device according to claim 11, The vehicle accident detection device, wherein the contour determination condition is set based on a plurality of distances from each of the plurality of sound detection units to the sound source estimated position.

13. The vehicle accident detection device according to claim 11, Further provided is a vehicle body information storage unit (49) configured to store vehicle body information indicating a plurality of coordinate positions on the vehicle body exterior, The vehicle accident detection device, wherein the outer shell determination condition is set based on the vehicle body information to indicate that the estimated sound source position is near the vehicle body outer shell.

14. The vehicle accident detection device according to claim 11, further comprising: a video analysis unit (45) configured to estimate a collision sound source position, which is a position where an object has collided with the vehicle, by analyzing video data generated by a photographing unit (36, 37) mounted on the vehicle and configured to photograph at least the outside of the vehicle; a vehicle body information storage unit (49) configured to store vehicle body information indicating a plurality of coordinate positions on the vehicle body exterior, The vehicle accident detection device, wherein the outer shell determination condition is set based on the vehicle body information to indicate that the collision sound source position is near the vehicle body outer shell.

15. The vehicle accident detection device according to any one of claims 1 to 7, The vehicle further includes a specific operation determination unit (S610 to S660) configured to determine whether or not a driver of the vehicle has performed a predetermined specific operation; The collision determination unit is configured to determine that the collision accident has not occurred if the vehicle body exterior determination unit determines that the estimated sound source position is on the vehicle body exterior and, based on the determination result by the specific operation determination unit, the driver performs the specific operation.

16. The vehicle accident detection device according to any one of claims 1 to 7, The collision determination unit is configured to determine that the collision accident has not occurred if the vehicle body exterior determination unit determines that the estimated sound source position is on the vehicle body exterior, and further if the acceleration at the sound generation timing when the sound is generated by the sound source is less than a predetermined collision determination acceleration.

17. The vehicle accident detection device according to any one of claims 1 to 7, The vehicle further includes a video analysis unit (45) configured to estimate a collision sound source position, which is a position where an object has collided with the vehicle, by analyzing video data generated by a photographing unit (36, 37) mounted on the vehicle and configured to photograph at least the outside of the vehicle, The collision determination unit is configured to determine that the collision accident has not occurred if the vehicle body exterior determination unit determines that the estimated sound source position is on the vehicle body exterior, and further, based on the analysis results of the video analysis unit, if an object has not collided with the vehicle at the time when the sound was generated by the sound source.

18. The vehicle accident detection device according to any one of claims 1 to 7, The vehicle accident detection device includes a plurality of sound detection units installed inside the vehicle.