JP2006268824A - Specific sound identifying method of alarm sound or the like, specific sound identifying device, and specific sound identifying system in moving vehicle - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an accurate identifying method of a specific sound capable of accurately extracting and determining an alarm sound or the like even in a place such as a town having many living noises, and to provide a specific sound identifying device for certainly informing an aurally handicapped person, a visually handicapped person, an old person, or the like of approach of an urgent vehicle or the like or some imminent danger. <P>SOLUTION: Attention is focused on the fact that a target specific sound is a sound larger than a normal environmental sound (background living noise) and the background environmental sounds are various and nearly infinite frequencies, frequency components included in sound wave only in a predetermined required frequency band and the amplitude value of each frequency component are extracted. The amplitude value is compared with a reference value, and each frequency component at a predetermined level or higher is compared with the frequency of a previously stored specific sound. The background environmental sounds are filtered in the frequency band and at the amplitude value levels of the frequency components, thereby extracting sound source data only of the target specific sound. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a method for identifying a specific sound, such as an alarm sound for notifying an elderly person or other hearing-impaired person that an emergency vehicle such as a police car, an ambulance or the like is approaching or some danger is approaching, a specific sound identification device, The present invention relates to a specific sound identification system that is attached to a mobile vehicle and identifies a specific sound such as an alarm sound.

  In general, it is known that a pedestrian or a vehicle driver in the city senses the surrounding situation and notifies the pedestrian or the vehicle driver of some information depending on the degree of danger or the situation where the driver is placed. . For example, when walking in the city, devices intended for vehicle approach notifications that make pedestrians perceive that an emergency vehicle or the like has approached are ignition noise, cracking or infrared signals emitted from the vehicle, ultrasonic waves For example, Patent Document 1 or Patent Document 2 proposes a device having an optical display, mechanical vibration, or voice notification means provided on a helmet or accessory worn by a pedestrian by detecting a transmission signal such as radio waves. Has been.

  Further, Patent Document 3 or Patent Document 4 or the like is known as a means for notifying a driver who is driving a vehicle that an emergency vehicle or the like similar to Patent Documents 1 and 2 is approaching. These are equipped with detection sensors such as a plurality of microphones on mobile vehicles to detect specific sounds such as sirens emitted by emergency vehicles such as ambulances and police cars, and analyze the sound level, timbre, sound direction, etc. If the data matches the stored data of the specific sound provided in the above, the driver is notified by a display unit, alarm sound, voice or the like, vibration or the like. Furthermore, in Patent Document 3, a vehicle is identified from changes in direction, sound quality, and volume, and the approach situation is judged and notified. Similarly in Patent Document 4, a change in the sound level is captured and the approaching and separation of the specified vehicle is determined.

As a device for attaching a sound collecting microphone to a moving vehicle such as an automobile and discriminating a sound wave picked up from this microphone, it has been proposed to attach a microphone to the roof of a moving vehicle as shown in Patent Document 3 in the past. Yes. Patent Document 3 discloses a configuration in which a plurality of microphones are installed at four corners of a vehicle roof to collect sound sources from four directions.
JP 2002-13019 A JP 2002-92781 A JP-A-6-328980 JP 2000-225877 A

  In the above-mentioned patent document 1 or 2, the receiver side of the information is provided with a receiving device using radio waves, the target object to be specified has a transmitting device, and transmits a specific signal to notify necessary information. Although effective in terms of sex, there is a drawback that only a specific object having a transmitting device can be recognized. In addition, since it is used as a set of a transmitting device and a receiving device, the introduction cost increases, and the system must be limited.

  On the other hand, a specific sound such as a siren emitted by an emergency vehicle such as an ambulance or a fire engine as disclosed in Patent Documents 3 and 4 is sensed and converted into another sound source, sound, or display means for output notification. Things have been proposed. The problem here is how to detect and collect the target sound source and accurately determine that it is a specific sound in a noisy place such as a town. For this, the sound source data such as the recognition pattern of the target specific sound is stored in advance and compared with the input data of the collected sound. Remains as a problem.

  On the other hand, the microphone means for collecting a specific sound in the above-mentioned apparatus needs to correspond to the sound source from the four sides, and at the same time, when equipped in a moving vehicle such as an automobile, the movement of the vehicle causes, for example, “go-go”, “puy-pu” The wind noise is a problem that a specific sound such as a siren sound is canceled and the target specific sound cannot be detected. In particular, when the microphone means is attached to the roof of the automobile as in the above-mentioned Patent Document 3, it is possible to collect sound from all directions, but if the automobile to which the microphone is attached moves at, for example, 30 km, wind noise and the like The noise is loud and the specific sound is canceled out. In the experiment of the present inventor, when the vehicle speed exceeded 30 km to 40 km under conditions close to no wind, we encountered a problem that the wind noise was large and it was impossible to collect specific sounds such as police cars before identifying them. .

  Therefore, the present invention detects a specific sound generated by an emergency vehicle or the like while walking or driving a car, and a specific sound identification method capable of accurately discriminating and detecting an alarm, and a specific sound using the specific sound The main issue is to provide an identification device. Furthermore, the present invention provides a specific sound identification system in a mobile vehicle that is mounted on a mobile vehicle such as an automobile and is less susceptible to wind noise such as wind noise and can accurately determine a wide range of specific sounds. That is the issue.

  In order to solve the above-described problems, the present invention employs a configuration having the following characteristics when a specific sound such as a siren sound such as an ambulance or a police car is determined by comparing its frequencies. First of all, paying attention to the fact that the target specific sound is larger than the normal environmental sound (life noise in the background), and at the same time, the environmental sound that is the background emits a myriad of infinite frequencies. The frequency component included in the sound wave and the amplitude value of each frequency component are extracted. Then, the amplitude value is compared with a reference value, and each frequency component at a predetermined level or higher is compared with the frequency of the specific sound stored and held in advance. In this way, the sound source data is extracted by filtering the environmental sound as the background by the frequency band and the amplitude value level of the frequency component so as to narrow down to the target specific sound. Therefore, the present invention employs the specific sound identification device according to claim 1. The reference value is set to an optimum value that excludes noise other than the target sound by measuring the amplitudes of the target specific sound and the environmental sound.

  Secondly, focusing on the fact that the target specific sound lasts for a certain period of time, each frequency component obtained by a plurality of sound source data extractions at a predetermined time interval when comparing the amplitude values of the respective frequency components. Is smoothed for each frequency component to obtain sound source data. For this reason, the present invention employs the specific sound identification device according to claim 2. As a result, instantaneous noise such as impact sound is substantially excluded from the sound source data. Therefore, as the time interval and the number of times, a suitable time and number of times are selected by measuring the time during which the specific sound continues and the time of the noise included in the environmental sound.

  Thirdly, the target specific sound is generated based on the natural frequency of the sound source, and at the same time as the natural frequency, an overtone frequency having an integer multiple (including 1 / integer) relationship is generated. The specific frequency set based on the data and the natural frequency of the sound source is compared. As a result, noise having a frequency that coincides with the natural frequency of the specific sound can be discriminated and excluded. That is, when the natural frequency of the specific sound is 500 Hz, even if 1000 Hz exists in the frequency component of the sound source data, it can be regarded as noise when 1500 Hz that is the harmonic frequency does not exist. For this reason, the present invention employs the specific sound identification device according to claim 5. Note that at least one harmonic frequency having an integer multiple relationship is compared with the sound source data.

  Fourth, the harmonic frequency of the target specific sound has the largest amplitude value of the same frequency as the natural frequency, and the amplitude value is relatively different, for example, the harmonic frequency, for example, the second harmonic, the third harmonic, etc. It focuses on the existence. In other words, when multiple frequencies having an integer multiple relationship are included in the sound source data, the frequency of the maximum amplitude value is compared with the natural frequency of the specific sound, and the target sound is almost present, then the other frequencies. When the amplitude value matches a predetermined rank, it can be determined that the target sound exists. For this reason, the present invention employs the specific sound identification device according to claim 6.

  Furthermore, the specific sound identification system for a mobile vehicle according to the present invention includes a sound collecting microphone means attached to a roof portion of a mobile vehicle having an appropriate structure, and whether or not a specific frequency is included in sound source data input from the microphone means. And determining means for notifying the driver based on a determination signal from the determining means. The sound collecting microphone means includes a pickup element that converts sound waves into an electrical signal, and a flat outer case that houses the pickup element. Therefore, the sound collecting microphone means constituted by the flat outer casing case is arranged so that the flat surface of the outer casing case is positioned within a range of 9 mm upward from the surface of the moving roof panel. Attach to. With such a configuration, a moving vehicle such as an automobile forms a laminar flow that is a uniform air flow above the roof during traveling.

  In the present invention, a laminar flow means a velocity distribution with a velocity difference of 99% or less, and in the experiment of the present inventors, a laminar boundary layer appears 10 mm to 20 mm above the roof plate when the traveling speed is 40 km. It has been found. The sound collecting microphone means configured in the above-described flat shape can be mounted on the roof plate of the mobile vehicle so that the flat surface is located below the laminar boundary layer, so that wind noise during traveling can be reduced. In the sound collecting microphone means, the side wall surface forming a flat shape is formed of an irregular uneven surface such as a texture. As a result, when the vehicle is traveling at high speed, the flow of wind along the roof of the vehicle causes a peeling phenomenon that peels off the surface of the roof plate, but this can be prevented by the uneven surface. In addition, the laminar flow formed in the moving vehicle when traveling is generated from the vicinity of the edge of the windshield surface, develops in the downstream direction, increases in thickness, and becomes a boundary at the rear edge of the roof (rear edge in the traveling direction). The layer ends.

  Accordingly, the sound collecting microphone means is arranged in the center of the roof width in the range of 1/3 to 1/2 of the roof length from the end position of the windshield surface so as to be disposed inside the laminar boundary layer. Attach to. Thus, the influence of the wind noise can be further reduced by disposing the sound collecting microphone means in a position avoiding the vicinity of the end portion of the windshield and the vicinity of the rear end portion of the roof roof.

  The present invention extracts a specific frequency set based on the natural frequency of a specific sound, a plurality of frequency components constituting the sound wave signal from the sound source input means and the amplitude value of each frequency component, and specifies the extracted sound source data. When determining whether or not the frequency is included, it is determined whether or not the amplitude value of each frequency component is greater than or equal to a preset reference value, the natural frequency of the specific sound in the sound source data, and the natural frequency Compared to the conventional method of simply comparing frequencies, the amplitude value of each frequency component is compared with the frequency compared to the conventional method of comparing the frequencies. Even if there is a noise frequency close to a specific frequency, it is specified by whether or not there is a harmonic-related frequency, and further specified by whether or not the amplitude value of the harmonic-related frequency is in a certain order. Do Meyori precise determination becomes possible. Therefore, even in places where there is a lot of noise such as in the city, or even in mobile vehicles such as automobiles, it is possible to identify specific sounds emitted by police cars, ambulances, etc. relatively accurately without being affected by engine noise etc. It is possible to inform.

  Furthermore, the specific sound identification system for a mobile vehicle according to the present invention is such that the microphone means having a flat outer casing case is mounted on the roof of a mobile vehicle such as an automobile so that the flat surface is located within 9 mm above the roof plate. Since the sound source data is collected from this microphone means, it is possible to collect and discriminate specific sounds without being affected by wind noise or other noise even when the moving vehicle moves at high speed. As a vehicle specific sound discriminating apparatus and system, it has a remarkable effect.

  Hereinafter, embodiments of a specific sound identification method such as an alarm sound and a specific sound identification device 21 according to the present invention will be described in detail with reference to the accompanying drawings. Here, FIG. 1 is a block diagram showing a configuration of a specific sound identification device 21 such as an alarm sound according to the present invention, FIGS. 2 and 3 are processing flows of the specific sound identification device 21, and FIGS. It is the graph which analyzed the input sound source data.

  The specific sound identification device 21 of the present invention is mainly intended to prevent dangers caused by the approach of emergency vehicles such as ambulances and fire engines, or other general vehicles while visually impaired persons and elderly people are walking around the city. The alarm sound generated by the vehicle is discriminated and judged as ambient noise, and a detection signal 46 is output to prompt danger notification. As shown in FIG. 1, the specific sound identification device 21 stores a specific frequency set based on a specific frequency of a specific sound such as a siren sound or an alarm sound of an ambulance or a fire engine as specific sound source data in a predetermined order. Specific frequency storage means 22, sound source input means 23 that senses and selectively captures an audible sound in the perceived frequency band in a normal living environment, and sound waves captured by the sound source input means 23 A plurality of frequency components constituting the sound wave signal by A / D conversion of the signal and an amplitude value of each frequency component, and a specific frequency of the plurality of sound sources stored in advance in the specific frequency storage unit 22 The data and the sound source data 35 extracted by the sound source data extraction means 24 are compared and collated, and it is identified whether or not the sound wave signal taken in from the degree of coincidence of the data is the target alarm sound. It is composed of a discriminating means 25 for disconnection.

  The specific frequency storage means 22 has a plurality of specific sound sources having a natural frequency of a specific sound source such as a siren sound, a traveling sound or a brake sound of an emergency vehicle such as an ambulance or a fire truck and an overtone frequency that is an integer multiple of the natural frequency. It consists of a database stored for each. The database is composed of a removable medium such as a semiconductor memory or a small-sized memory card that is detachable using a magnetic or light storage medium (not shown). The harmonic frequencies having an integer multiple relationship with the natural frequency of the specific sound source are arranged in a predetermined order and stored in a database.

  The sound source input means 23 amplifies the microphone 31 having a wide directivity (the microphone means 61 in FIG. 7) and a sound wave signal captured from the microphone 31 to a certain sound pressure level that can be signal-processed according to frequency characteristics. And an A / D converter 34 that converts the analog sound wave signal amplified by the amplifier 33 into a digital sound wave signal. The audible sound picked up by the microphone 31 may contain a noise component in a high frequency band. In order to reduce such misrecognition due to noise, the frequency band that passes through the filter 32 to remove high-frequency noise and is actually captured before the amplification process in the amplifier 33 is narrowed down to a range of 300 to 8500 Hz.

  The sound source data extracting unit 24 applies a window function to the digital sound wave signal captured and A / D converted by the sound source input unit 23 to cut unnecessary noise components included in the high frequency, and then performs high-speed Fourier transform (FFT) processing. The plurality of frequency components constituting the sound wave signal and the amplitude value of each frequency component are extracted as a spectral distribution. The sound source data extraction is performed a predetermined number of times at predetermined time intervals, averaged, and instantaneous noise such as an impact sound is substantially excluded from the sound source data.

  The discriminating means 25 obtains an average spectrum extracted by the sound source data extracting means 24, and calculates the main frequency components and amplitude values constituting the sound source data and the natural frequency and natural frequency of the specific sound source stored in the specific frequency storage means 22. Analytical work to identify the type of input sound source by comparing and collating overtone frequencies that have an integer multiple relationship. This comparison and collation includes the structure of the frequency components of the sound source data, the arrangement of the frequency components arranged in the order of their amplitude values, the specific frequency and the specific frequency of the specific sound source previously stored in the specific frequency storage means 22 in a predetermined order. By comparing and collating the arrangement of overtone frequencies having an integer multiple relationship with the frequency, the degree of coincidence is determined.

  The sound source data extracting means 24 is composed of a CPU (central processing unit) that performs input / output control and various calculations, and a DSP (digital signal processing unit) that specializes in analyzing sound source data. The CPU performs input / output control of the specific sound source data 42 stored in the sound source data and the specific frequency storage means 22 and controls the interface unit with the notification means 51 when the input sound source data matches the specific sound, and the DSP. Then, the analysis of the signal by fast Fourier transform (FFT) for specifying the input sound source is mainly performed. This FFT processing is a technique widely used in sound and vibration analysis. The input sound source data that has been captured and digitized has its own fundamental wave, fundamental wave × second order, fundamental wave × third order, etc. The fundamental wave × n-order frequency (sine wave) is decomposed into frequency components, and the peak amplitude value of each frequency component is calculated.

  FIG. 5 shows the frequency components constituting the siren sound by performing FFT processing on the ambulance siren sound. The magnitude of the amplitude value of each frequency component is expressed by the thickness of the line of the graph. As shown in the graph, the horizontal axis is the time axis, and the vertical axis is the frequency, and the siren sound repeats high and low sounds. The frequency of the high sound range is remarkable in the vicinity of 1000 Hz, 2000 Hz, and 3000 Hz. Looking at the thickness of the line, it can be seen that the one with a large amplitude value is 1000 Hz. Similarly, the frequencies in the low sound range are prominent in the vicinity of 800 Hz, 1600 Hz, and 2400 Hz, and 800 Hz has the largest amplitude value from the thickness of the line. Furthermore, it can be seen from the graph that the high-frequency range and the low-frequency range have a harmonic relationship that is almost an integral multiple of each.

  Also, from this graph, the ambulance siren sound is a sound composed of two basic natural frequencies, and the basic natural frequencies can be regarded as 1000 Hz and 800 Hz having the largest amplitude values, respectively. This is used as a specific frequency of the ambulance siren sound, and a database including the order of the amplitude value levels of the overtone frequencies can be used as comparison data for identifying the specific sound.

  FIG. 6 shows the frequency components constituting the police car siren sound. This is an example in which a police car sounds a siren repeatedly and repeatedly ON / OFF at intervals of about 8 seconds. In this graph, a frequency of about 900 Hz, which is the thickest line, is seen as a basic natural frequency, and a frequency seen as its harmonic frequency can be observed and extracted in a superimposed manner. Other specific alarm sounds other than the above examples can be identified by previously observing and creating similar data and creating a database.

  The identification of the sound source in the actual scene is performed by sampling a frequency at which a specific peak of the sound source appears as a spectrum distribution as shown in FIG. 4, a frequency equal to or higher than a predetermined level that is a preset reference value, This is done by comparing with the data stored in the specific frequency storage means. Although the frequency component of the sound other than the specific sound source is also observed on the spectrum distribution, it can be discarded as noise by extraction with a preset reference value.

  Note that the specific sound identification device 21 of the present embodiment is an application example for detecting sounds emitted from vehicles traveling in the city and perceiving visually impaired persons such as visually impaired persons and elderly people, or elderly people, depending on the approaching state. However, it can also be applied for other purposes. For example, in addition to facilitating the handling of various electrical appliances even in deaf elderly people by making it possible to sense the notification sound emitted by washing machines and microwave ovens in the home, the sound of breaking window glass, By making it possible to detect the sound of opening the door, it can be applied as a security system. In addition, by installing in a vehicle, the driver can sense the presence of an emergency vehicle such as an ambulance, an alarm sound of surrounding vehicles, etc. to help prevent danger. In the case of application to such notification of electrical products, security, or traffic accident prevention, specific sound source data to be compared by a removable storage medium such as a memory card provided in the specific frequency storage means 22 in the specific sound identification device 21 It can be easily realized by replacing.

  Next, the specific sound identification method employed in the present invention will be described in detail. The sound source discrimination process of the specific sound discrimination device 21 will be described based on the process flow of FIGS. The specific sound identification device 21 starts by turning on the power, first performs an initial process 41, and then shifts to an actual process. In the initial processing 41, initialization of a data area inside a CPU or DSP (not shown) and setting of an identification number unique to the apparatus are performed automatically.

  FIG. 2 shows an operation flow as the sound source data extraction step 11. As the sound source data extraction step 11, an audible sound is actually captured from the surroundings with a microphone 31 or the like, passed through a filter 32 for the purpose of removing unnecessary high-frequency noise, amplified to a level at which signal processing is possible with an amplifier 33, and A / D converted. To convert it into a digital sound wave signal. Next, the A / D converted sound wave signal multiplied by the optimum window function for the target specific sound source is subjected to a fast Fourier transform (FFT) process, and the frequency component and each frequency component included in the acquired sound wave signal are processed. Are extracted as a spectrum distribution as shown in FIG. At this time, the audible sound is captured a predetermined number of times at predetermined time intervals, the obtained spectrum distribution data is averaged, and the average spectrum distribution << Fn, Vn >> (<< >> indicates the average spectrum distribution, and so on. And a plurality of frequency components of the observed sound source and the amplitude value of each frequency component are extracted. By averaging as the average spectrum distribution as described above, the influence of disturbance noise such as instantaneous noise can be removed.

  Further, here, as the sound source data processing step 12, a frequency component having a spectrum intensity (indicating the magnitude of the amplitude value, the same shall apply hereinafter) having a predetermined reference level or higher, which is a preset reference value, with respect to the average spectrum distribution. , And an array [Fn] of frequency components ([] indicates an array of frequency components arranged in a predetermined level order, the same shall apply hereinafter) to create sound source data 35.

  Next, the comparison [Fn] of the frequency component extracted and created as the determination step 13 is compared with the specific frequency set based on the natural frequency of the specific sound such as an alarm sound stored in the specific frequency storage means 22 To do. In the specific frequency storage means 22, the array element {Fcn} ({}, which is composed of harmonic frequencies including the natural frequencies of a plurality of specific sound sources, indicates each element of the set of frequency components constituting the array, and so on. Are stored as an array [Fcn] in a predetermined level order.

  In the comparison and collation, the extracted sound source data 35 and the specific sound source data 42 to be compared may be searched one by one. However, as the number of the specific sound source data 42 to be compared increases, the processing becomes much slower. For this reason, if the specific sound source data 42 to be compared is sorted in advance in ascending or descending order of data values and the binary search method is used, the collation can be completed with a small number of searches. In the comparison and collation in the determination step 13, the frequency component value multiplied by the Doppler coefficient (center) is taken into consideration that the captured sound source itself is moving, instead of using the frequency component value itself of the extracted sound source data. A comparison search is made from the specific sound source data 42 of the specific frequency storage means 22 within a range expanded to (frequency ± 3%).

  As the first judgment 43, the array element {Fn} of the natural frequency and the harmonic frequency of any array [Fc] of a plurality of specific sound sources is extracted and created the array element {Fn of the array [Fn] of frequency components } Is determined. At this time, if neither of the specific sound source frequency component array elements {Fcn} is included, it is determined that the sound is not the target specific sound, and the input of the sound source input means 23 is awaited. If all the array elements {Fcn} of frequency components of any specific sound source are included, there is a possibility that the sound source is that specific sound source, and the matching degree is verified by adding a condition.

  As the second determination 44, when there are a frequency component corresponding to the natural frequency Fc of the specific sound source and a frequency component corresponding to the harmonic frequency (Fc, 2Fc, 3Fc, etc.), the amplitude value of the frequency component of the sound source data 35 is further included. Is substantially equal to the natural frequency Fc for the maximum frequency. If they do not match, there is a possibility that the sound source may be a specific sound source, but it cannot be determined and the input of the sound source input means 23 is awaited. If they match, it is regarded as a specific sound source that is almost the object, and the degree of match is checked in more detail.

  As the third determination 45, the amplitude values of the corresponding frequency components of the matched specific sound source frequency component array [Fcn] and the extracted frequency component array [Fn] are compared with each other. Then, it is determined whether or not it matches the frequency component array [Fcn] of the specific sound source arranged in a predetermined level order set in advance. If they do not match, it cannot be determined that the sound source is a specific sound source, but it can be determined that the sound is a specific sound of the specific sound source and the detection signal 46 can be output to the notification means 51. If they match, it is considered to be a complete match and is determined to be a specific sound of a specific sound source, and a detection signal 46 is output to the notification means 51. After outputting the detection signal 46, the input of the sound source input means 23 is again waited for.

  When the above-described specific sound discrimination method and identification device are mounted on a mobile vehicle such as an automobile, a specific (alarm) sound from the outside is picked up by the microphone means 61 (the microphone 31 in the apparatus of FIG. 1). In some cases, a specific sound cannot be collected due to the effects of noise such as wind noise. This is because if there is a sound source such as noise around the microphone, the alarm sound in the distance is canceled out. In the case of a microphone in a traveling vehicle, wind noise has the greatest influence.

  Therefore, in the present invention, the microphone means 61 is first formed in a flat shape. FIG. 9 shows the structure, FIG. 9A is a cross-sectional view thereof, and FIG. 9B is a plan view of the rear surface thereof. As shown in the figure, the microphone means 61 has a flat shape formed by resin or the like and a part of a sphere is cut off, and a sound collecting member is formed at the center of the upper surface of the outer casing 62 formed with a recess for accommodating components on the back surface. There is an opening 69, and an entrance portion of the sound collection opening 69 has a conical sound collection opening slope 70. A condenser microphone (pickup element) 65 is disposed so as to sandwich the waterproof sheet 63 on the back side of the outer casing 62 so as to face the sound collection opening 69, and a buffer filler 64 is packed around the condenser microphone 65. Fixed.

  The condenser microphone 65 is supplied with power by a microphone operating battery 66 in the outer case 62 and is connected to the circuit portion of the sound source input means 23 via an output cord 67. Further, a thin donut-shaped attracting magnet 68 for attaching the microphone means 61 to the vehicle 81 is bonded and fixed. The upper surface of the microphone means 61 is composed of an outer casing case upper plane portion 71 concentric with the sound collection opening 69 and an outer casing spherical surface portion 72 with a surface / projection 73 formed on the surface.

  The microphone means 61 is arranged so that the special magnet alarm sound from all directions around the vehicle can be dealt with by the single microphone means 61 on the surface of the roof plate 82 of the vehicle 81 shown in FIG. Attach and fix with. The position to be arranged is attached to substantially the center of the roof width within the range of 1/3 to 1/2 of the roof length L from the position of the top side end of the windshield surface 83 in the traveling direction.

  As shown in FIG. 7, there is a main flow portion of a laminar flow 84, which is an air flow having a substantially uniform velocity distribution along the streamlined vehicle body, on the roof plate 82 of the traveling vehicle 81, as shown in FIG. As described above, the vicinity of the upper surface of the roof plate 82 to which the microphone means 61 is attached is a boundary layer 85 that is a region in which the velocity distribution rapidly decreases as the distance from the main flow portion approaches the roof plate surface. The boundary layer 85 on the upper surface of the roof plate is generated from the vicinity of the end portion of the windshield surface 83 on the top side, develops in the downstream direction, increases in thickness, and ends at the rear edge of the roof plate 82. The microphone means 61 is disposed in the region of the boundary layer 85.

The block diagram which shows the structure of the specific sound identification apparatus concerning this invention. The sound source data extraction flow of the apparatus of FIG. The sound source discrimination flow of the apparatus of FIG. Graph spectrum distribution of input sound source data (ambulance siren) FFT processing graph of ambulance siren sound. FFT processing graph of police car siren sound. It is structure explanatory drawing which attaches a sound collection microphone means to a mobile vehicle, (a) is a roof-plate top view, (b) is a vehicle side view. It is explanatory drawing which shows the flow of the wind in the microphone means attachment part of FIG. 7, (a) And (b) is principal part expansion explanatory drawing. The structure of a sound collection microphone means is shown, (a) is the longitudinal cross-sectional view, (b) is a top view. FIG. 10 shows an outer case of the sound collecting microphone means of FIG. 9, (a) is a plan view, and (b) is a side view. It is explanatory drawing of the effect | action of the wrinkle (uneven | corrugated surface) in the outer casing case of FIG. 10, (a) is the description which shows the flow of the wind when there is no wrinkle, (b) shows the flow of the wind when there is a wrinkle. Figure.

Explanation of symbols

21 specific sound identification device 22 specific frequency storage means 23 sound source input means 24 sound source data extraction means 25 discrimination means 31 (61) microphone (microphone means)
32 Filter 33 Amplifier 34 A / D converter 35 Sound source data 42 Specific sound source data 46 Detection signal 51 Notification means 62 Outer casing 63 Waterproof sheet 64 Buffer filling material 65 Condenser microphone 66 Microphone operating battery 67 Output code 68 Adsorption magnet 69 Sound collecting opening 70 Sound collecting opening slope 71 Outer casing upper flat surface portion 72 Outer casing spherical surface portion 73 Wrinkles / protrusions 81 Vehicle 82 Vehicle top portion 83 Windshield surface 84 Laminar flow 85 Boundary layer

Claims (15)

Specific frequency storage means for storing a specific frequency set based on a natural frequency of a specific sound such as an alarm sound;
Sound source input means for converting sound waves from the sound source into electrical signals;
Sound source data extraction means for extracting a plurality of frequency components constituting the sound wave signal from the sound source input means and the amplitude value of each frequency component;
A specific sound identification device comprising: a determination unit that determines whether the specific frequency is included in the sound source data from the sound source data extraction unit;
The determination means includes a first determination for determining whether the amplitude value of each frequency component is equal to or greater than a preset reference value for the sound source data, and the frequency of each frequency component substantially matches the specific frequency. A second determination for determining whether or not to do,
In the comparison between the frequency of the sound source data and the specific frequency in the second determination, does the sound source data include the natural frequency of the specific sound and a harmonic frequency that is an integer multiple of the natural frequency? A specific sound identification device characterized by determining whether or not. 2. The sound source data extracting unit extracts amplitude values of a plurality of frequency components at predetermined time intervals, and smoothes the plurality of amplitude values to extract sound source data. Specific sound identification device. Whether the amplitude value for each frequency component in the first determination is greater than or equal to a preset reference value is determined by amplifying the sound wave signal from the sound source input means according to the attenuation rate of the frequency component, The specific sound identification device according to claim 1, wherein an amplitude value for each frequency component is compared with a reference value. The determination means performs the second determination on a frequency component having an amplitude value equal to or greater than a preset reference value after executing the first determination on the sound source data, and is substantially the same as the specific frequency. It is discriminate | determined whether it is. The specific sound identification device in any one of Claims 1 thru | or 3 characterized by the above-mentioned. The second determination includes determining whether or not the sound source data includes a natural frequency of the specific sound and a harmonic frequency that is an integer multiple of the natural frequency, and a frequency component of the sound source data. 2. The specific sound identifying apparatus according to claim 1, wherein it is determined whether or not an amplitude value substantially coincides with the natural frequency for a maximum frequency. In the second determination, when the sound source data includes a frequency component corresponding to the natural frequency of the specific sound and a frequency component corresponding to the harmonic frequency, the amplitude values of the corresponding frequency components are compared with each other. The specific sound identification device according to claim 1, wherein it is determined whether or not a predetermined order is set in advance. 7. The specific sound identification apparatus according to claim 6, wherein the predetermined order is set to maximize an amplitude value of a frequency component corresponding to a natural frequency of the specific sound. The determination means performs the second determination for a frequency component having an amplitude value equal to or greater than a preset reference value after executing the first determination for the sound source data, and in the second determination, the sound source When the data includes a natural frequency of the specific sound and a harmonic frequency that is an integral multiple of the natural frequency, it is determined whether or not the sound source data includes a frequency other than the harmonic frequency. The specific sound identification device according to claim 1. When the sound source data includes a natural frequency of the specific sound and a harmonic frequency that is an integer multiple of the natural frequency based on a determination result from the determination unit, vibration, visual information, and other notification means are provided. The specific sound identification device according to any one of claims 1 to 8, wherein the specific sound identification device operates. A sound source data extraction step for converting sound waves from the sound source into an electrical signal and extracting the frequency component and the amplitude value of each frequency component;
Sound source data processing for extracting the amplitude value of each frequency component a plurality of times at predetermined time intervals and smoothing, and then comparing the smoothed amplitude value with a preset reference value to extract an amplitude value greater than or equal to a predetermined value Steps,
A determination step of comparing the frequency component extracted in the sound source data processing step with a specific frequency set based on a natural frequency of a specific sound such as an alarm sound,
The comparison between the sound source data and the specific frequency in the determination step determines whether or not the sound source data includes a natural frequency of the specific sound and a harmonic frequency that is an integer multiple of the natural frequency. A method for identifying a specific sound. In the determination step, when the frequency component corresponding to the natural frequency of the specific sound and the frequency component corresponding to the harmonic frequency exist in the sound source data, the amplitude values of the corresponding frequency components are mutually compared and set in advance. The specific sound identification method according to claim 10, wherein it is determined whether or not the predetermined level order is satisfied. A specific sound identification system for a mobile vehicle that forms a laminar flow of air above a roof plate,
Sound collecting microphone means attached to the roof of the mobile vehicle;
Determining means for determining whether or not the specific frequency is included in the sound source data input from the sound collecting microphone means;
A notification means for notifying the driver based on a determination signal from the determination means,
The sound collecting microphone means is composed of a pickup element that converts sound waves into an electrical signal, and a flat outer case that incorporates the pickup element,
The sound collecting microphone means is attached to the roof plate of the mobile vehicle so that the flat surface of the outer casing case is located within a range of 9 mm upward from the surface of the roof plate forming the roof portion of the mobile vehicle. A specific sound identification system. 13. The specific sound in the mobile vehicle according to claim 12, wherein the flat outer casing case has an irregular uneven surface such as a texture to prevent air vortex flow at least on a side wall surface thereof. Identification system. The sound collecting microphone means has a substantially central portion of the roof width at a position 1/3 to 1/2 of the roof length from the end position of the windshield surface forming a laminar flow above the roof plate of the mobile vehicle. The specific sound identification system for a mobile vehicle according to claim 12, wherein the specific sound identification system is disposed in the vehicle. The determination means determines whether or not the amplitude value of each frequency component is equal to or greater than a preset reference value for a plurality of frequency components constituting the sound wave signal input from the sound collecting microphone means; A second determination for determining whether or not the frequency of each frequency component substantially matches a preset specific frequency,
In the comparison between the frequency of the sound wave signal and the specific frequency in the second determination, does the sound wave signal include a natural frequency of the specific sound and a harmonic frequency that is an integer multiple of the natural frequency? 15. The specific sound identification system for a mobile vehicle according to any one of claims 12 to 14, characterized by determining whether or not.

Images