JP4826392B2 - Object identification device - Google Patents

Description

  The present invention relates to an object identification device mounted on, for example, a vehicle.

  A human body detection device that uses an ultrasonic sensor and an infrared sensor to identify an object in a detection area as a human body is known (see, for example, Patent Document 1).

JP-A-10-186049

  However, since ultrasonic waves and infrared rays are used, a dedicated transmitter / receiver is required and a processing device for integrating information is required, which increases the size of the device and increases costs. It was.

Object identification device according to the present invention is an object identification device to be measured a person wearing fiber clothes, and transmitting means for transmitting toward ultrasound to the measurement target, a reflected wave from the measurement object Receiving means for receiving; detecting means for detecting a frequency characteristic representing a relationship between the reception intensity of the reflected wave and the frequency based on reception information of the receiving means; and linear approximation processing within a predetermined frequency band in the frequency characteristic And calculating means for outputting the gradient amount of received intensity as a feature value based on the obtained linear approximate waveform, comparing the feature value obtained by the calculation means with a predetermined reference feature value, And a determination unit that determines whether or not a difference between the reference feature amount and the reference feature amount is included in a predetermined range, and the determination unit determines that the difference is included in the predetermined range , The measurement Subject is characterized by having a determining means for determining that the person wearing the clothing.

According to the present invention, since the measurement target can be identified as an object including fibers based on the tendency of the amplitude of the reflected wave to decrease with an increase in frequency, the measurement target person can be detected as a person regardless of the number of clothes .

Hereinafter, the best mode for carrying out the present invention will be described with reference to the drawings.
-First embodiment-
FIG. 1 is a diagram showing a first embodiment of an object identification device according to the present invention, and is a block diagram showing a schematic configuration of an object identification device mounted on a vehicle. The object identification device shown in FIG. 1 includes a control device 100, an oscillator 101, a transmitter 103, and a receiver 104. A control device 100 including an input / output I / F, a memory, a CPU, and the like includes a transmission control unit 102, a signal analysis unit 105, a physical property identification unit 106, and an object classification unit 107 as functional configurations.

  The transmission control unit 102 outputs an instruction regarding the waveform of the transmission signal and the center frequency to the oscillator 101 and the signal analysis unit 105. The oscillator 101 includes an electromagnetic wave generator and a frequency controller (not shown), and outputs a transmission signal based on the waveform and the center frequency input from the transmission control unit 102 to the transmitter 103. The transmitter 103 generates an ultrasonic wave based on a transmission signal from the oscillator 101. For example, a thermally induced ultrasonic wave generating element is used.

  FIG. 2 is a diagram illustrating an example of a thermally induced ultrasonic wave generating element. In the heat-induced ultrasonic wave generating element, a heat insulating layer 103b made of porous silicon is formed on a substrate 103a made of single crystal silicon, and a heat generating portion 103c is formed on the heat insulating layer 103b. The heat generating part 103c is composed of an electrical resistor that generates Joule heat and a Peltier element that generates and absorbs heat by the Peltier effect.

  When the oscillator 101 energizes the heat generating portion 103c, the heat generating portion 103c generates heat, and when the energization is stopped, the heat generation stops. Therefore, when energization is repeatedly turned on / off, the temperature of the heat generating portion 103c repeats high / bottom according to the on / off. As the temperature of the heat generating portion 103c increases and decreases, the temperature of the medium (air) in contact with the heat generating portion 103a also repeats high and low. As a result, repetition of large and small medium densities proceeds upward in the figure, and a dense wave is generated in the medium. In this embodiment, a dense wave in the ultrasonic band is generated.

  As described above, the heat generated in the heat generating portion 103c is easily transmitted to the medium side by forming the heat generating portion 103c with porous silicon having excellent heat insulation. In addition, when the substrate 103a is made of single crystal silicon, a steady temperature increase of the substrate 103a is suppressed. The directivity of the transmitter 103 is determined by the size of the heat generating part 103a and the frequency of the traveling wave (dense wave). For example, when the size of the heat generating part 103a is a rectangle of 20 mm × 20 mm and the frequency is 160 kHz, About 10deg. (Reference; Watanabe et al., “Studies on the directivity of thermally induced nanosilicon ultrasonic sources”, pp461-464, Proceeding of “11th symposium of Microjoining and Assembly Technology in Electronics (Mate2005)”, Feb. 2005).

  Returning to FIG. 1, the ultrasonic wave is reflected by the measurement object M, and the reflected wave is received by the receiver 104. For the receiver 104, for example, a microphone using a piezo element or the like is used as a sound receiving unit, and a reception signal obtained by piezoelectrically converting ultrasonic waves is amplified and detected by an amplifier. Instead of the piezoelectric conversion element, a general capacitance type element may be used for the sound receiving unit.

  Based on the signal information (transmission signal waveform and center frequency) from the transmission control unit 102 and the signal waveform of the reception signal from the receiver 104, the signal analysis unit 105 reflects the time waveform and frequency of the reflected wave from the measurement target M. Calculate the characteristics. The physical property identification unit 106 calculates a characteristic amount related to the frequency characteristic of the measurement target using a statistical method with respect to the frequency characteristic of a predetermined frequency band (for example, 60 kHz to 80 kHz band). For example, the amplitude value of the frequency characteristic is logarithmically displayed and the data is linearly approximated, and the gradient of the straight line obtained by the linear approximation is used as the feature amount. The object classifying unit 107 classifies the measurement target by comparing the feature amount calculated by the physical property identifying unit 106 with the frequency characteristic of a known perfect reflector.

<Description of operation of object identification device>
Next, the basic operation of the object identification device of the first embodiment will be described based on the flowchart shown in FIG. In the case of an object identification device mounted on a vehicle, it is necessary to determine whether or not the surrounding object is a person. In the identification processing shown in FIG. The calculation process related to identification is shown. In step S <b> 201, an ultrasonic wave is transmitted from the transmitter 103 toward the measurement target M. In response to an instruction from the transmission control unit 102, the oscillator 101 transmits a frequency-modulated electric signal to the transmitter 103, and the transmitter 103 generates an ultrasonic wave based on the input electric signal.

  The frequency modulation at this time is set so as to modulate in the 10 kHz to 50 kHz band for 100 msec, for example. The frequency modulation may be continuous or may be changed for each step frequency, and the transmission waveform may be a continuous wave or a pulse wave.

  In step S <b> 202, the reflected wave from the measurement target M is received by the receiver 104. The receiver 104 outputs an electrical signal corresponding to the reflected wave to the signal analysis unit 105. In step S <b> 203, the signal analysis unit 105 calculates the frequency characteristic of the electrical signal input from the receiver 104.

  FIG. 4 shows an example of experimental data obtained for the frequency characteristics of the received wave. The vertical axis represents the signal intensity (decibel value) output from the receiver 104, and the horizontal axis represents the frequency. As the measurement object M, three types are shown: a plywood board (L1), a case where five sheets of felt are closely stacked (L2), and a case where five sheets of felt are stacked with a gap (L3).

  When the curves L1 and L2 in FIG. 4 are linearly approximated, the approximate changes are as indicated by L10 and L20 in FIG. As shown in FIG. 4 and FIG. 5, the signal intensity decreases as the frequency of the reflected wave increases, and the characteristic curve has a downward trend. This downward slope amount is larger in the felt of the fiber structure, and when the same felt is overlapped, the slope amount is larger in the case of overlapping with a gap. When comparing these gradient amounts, as shown in FIGS. 4 and 5, the difference between the gradient amount of the fiber object (felt) and the gradient amount of the reflection object (veneer) clearly appears in the band of the frequencies f1 to f2. ing.

  The propagation characteristics of the ultrasonic wave can be roughly calculated by the “density” and “elasticity” of the medium and the “frequency” of the sound wave. In particular, in the case of fibers, the part where the ultrasonic wave propagates is the “air” part of the fiber structure. Therefore, the difference between the perfect reflector and the fiber in the reflection system is obtained by modulating the “frequency” to determine the propagation intensity. Can be classified well by “density” and “elasticity”.

  The frequency attenuation characteristic shifts to a region where the sound velocity is smaller than that of the atmosphere, and the frequency attenuation characteristic is seen on the lower frequency side than the atmospheric attenuation. For example, in the case of a fiber object such as felt, the sound wave propagation speed is 100 m / s to 400 m / s, and attenuation characteristics appear in a band of 60 kHz to 80 kHz. Considering the difference in the material and form of the clothes, it is preferable to detect the characteristics in the band range of about 50 kHz to 100 kHz. Thus, by processing the observation of the attenuation characteristic statistically, it is possible to more easily detect a person wearing a fiber.

  In the case of many road objects such as vehicles and walls among objects around the vehicle, the ultrasonic waves are totally reflected on the object boundary surface due to the difference in the specific acoustic impedance. On the other hand, in the case of an object with large depletion such as clothing, the sound wave also has a property of transmitting through the boundary surface of the object, and arrives as a reflected wave after repeating scattering and absorption inside the measurement target. Therefore, the reflected wave includes the acoustic characteristic of the measuring object M as information, and this appears as a change in the frequency characteristic as shown in FIGS. In the case of a fiber object, it is considered that a sound wave propagates through the gap space of the fiber. Therefore, the propagation characteristics are low, and the reflected wave from the inside of the fiber object becomes small.

  In step S204, the frequency characteristic calculated in step S203 is processed using a statistical method in the physical property identification unit 106, and a feature amount related to the measurement target M is calculated. For example, the received wave frequency characteristics in the frequency bands f1 to f2 are linearly approximated, and the gradient amount is calculated as the feature amount.

  In step S205, the object classification unit 107 compares the feature quantity of the measurement target M calculated in step S204 with the feature quantity (reference feature quantity) of the complete reflector, and “| reference feature quantity−feature quantity | ≧ predetermined value”. Is determined. Here, the predetermined amount is a reference amount for determining whether or not the feature amount of the measurement target M corresponds to the feature amount of the reflected wave of the clothing, and “| reference feature amount−feature amount |” In the case of “≧ predetermined value”, it is determined that it corresponds to the feature amount of the reflected wave of the clothes. For example, when the above-described gradient amount is used as the feature amount, determination is made in step S205 as “| gradient amount of complete reflector−gradient amount | ≧ predetermined value”. Here, the predetermined value is determined by, for example, the sensitivity frequency characteristic of the transmission / reception system, the distance of the complete reflector, and the like.

  If it is determined in step S205 that “| reference feature amount−feature amount | ≧ predetermined value” and the process proceeds to step S206, it is determined in step S206 that the measurement target M is an object including clothing. On the other hand, if it is determined in step S205 that “| reference feature amount−feature amount | <predetermined value” and the process proceeds to step S207, the measurement target M is an object having the same properties as a complete reflector in step S207. It is determined that

  As described above, in the first embodiment, since the frequency characteristic of the reflected wave received by the receiver 104 changes depending on the structure of the measurement target, a complicated structure such as a flat object and a fiber is determined from the change. The object can be easily identified. Therefore, it is possible to classify the measurement object simply by providing an ultrasonic device. Further, since the measurement target can be identified as an object including fibers based on the tendency of the amplitude of the reflected wave to decrease with an increase in frequency, the measurement target person can be detected as a person regardless of the number of clothes. In particular, by setting the frequency band to 50 kHz to 100 kHz, it is difficult to be affected by atmospheric attenuation, and it is possible to detect far, and the characteristic attenuation due to fibers can be observed, so that fibers such as clothes can be detected efficiently. Can do.

-Second Embodiment-
FIG. 6 is a diagram showing a second embodiment of the object identification device, and is a block diagram similar to FIG. The basic configuration and operation are the same as those in the block diagram shown in FIG. 1 except that the frequency domain setting unit 108 is provided in the control device 100. The frequency domain setting unit 108 detects the amplitude of the reflected wave, that is, the amplitude of the reception signal output from the receiver 103, and the frequency at which the physical property identification unit 106 calculates the feature amount according to the detected amplitude. Set the bandwidth. The physical property identification unit 106 calculates a feature amount based on the signal in the frequency band set by the frequency domain setting unit 108.

  FIG. 7 is a flowchart showing the basic operation of the object identification device according to the second embodiment, in which the processing in step S301 related to the frequency domain setting unit 108 is added to the flowchart shown in FIG. In step S301, based on the reception result of the reflected wave in step S202, a frequency band for calculating the feature amount by the physical property identification unit 106 is set.

As an example of the frequency band setting, the following formulas (1) to (3) are used so that there is no influence of distance attenuation on a known perfect reflector and statistical calculation for classifying the measurement target M can be performed efficiently. Set as follows. S11 is the amplitude ratio of the reflected wave to the amplitude of the perfect reflector. The start frequency fa is a lower frequency in the frequency band. As the measurement object M has a larger amplitude S11 of the reflected wave, the band moves to the higher frequency side, and the band width also increases.
Incremental frequency fa = S11 × 100kHz (1)
Start frequency fs = S11 × 100kHz + 40kHz (2)
Incremental band α = S11 × 50kHz (3)

  Further, as described above, the band settings are not different for different amplitudes S11, but may be classified into a plurality of types of band settings. In the example shown in FIG. 8, the band setting is set to one of two types of a standard mode and a broadband high-frequency mode that are determined in advance depending on the magnitude of the reflectance (the amplitude of the reflected wave) of the clothes. With this setting, the bandwidth setting can be simplified and the calculation load on the CPU can be reduced.

  As shown in FIG. 8B, in the case of clothes A such as a thin T-shirt, the reflectance is large. When the reflectivity is large in this way, the difference in inclination is small from that in the case of the reference reflection. Therefore, the band setting is set to the broadband high frequency mode so that the gradient amount can be calculated with higher accuracy. On the other hand, in the case of clothes B with a low reflectance such as a trainer, the band setting is set to the standard mode. FIG. 9 shows the frequency characteristics of clothes A and B with respect to the reference reflection and the respective band settings. In the broadband high-frequency mode, the start frequency is increased by fa, and the width of the frequency band is increased by α. The settings of fa and α may be set by the above-described equations (1) and (3).

  The reason why the band is set according to the measurement target is as follows. When an object identification device on the road or indoors is assumed, as shown in FIG. 8A, a human body exists on the back surface of the fiber (clothing). Therefore, in order to easily detect the absorption characteristics and efficiently analyze the frequency of a plurality of fibers, a band is set so that the difference in attenuation characteristics is likely to occur based on the reflectance.

  In the second embodiment, in addition to the effects of the first embodiment described above, the frequency band is set based on the amplitude, so that the frequency band of the transmission wave can be selected efficiently and quickly. Objects can be classified.

-Third embodiment-
FIG. 10 is a diagram showing a third embodiment of the object identification device, and is a block diagram similar to FIG. The basic configuration and operation are the same as those in the block diagram shown in FIG. 1 except that a distance measuring unit 109 and an atmospheric attenuation correction unit 110 are further provided. The distance measurement unit 109 calculates the distance to the measurement target M based on the delay time between the signal information from the transmission control unit 102 and the reception signal of the receiver 104. The atmospheric attenuation correction unit 110 corrects the frequency characteristics of the reflected wave based on the distance to the measurement target M detected by the distance measuring unit 109 and the propagation attenuation of the reflected wave due to the atmosphere.

FIG. 11 is a flowchart showing the basic operation of the object identification device according to the third embodiment, and steps S401 and S402 related to the distance measuring unit 109 and the atmospheric attenuation correction unit 110 are added to the flowchart shown in FIG. Is. In step S401, the distance measuring unit 109 calculates the distance to the measurement object M described above. In step S402, the attenuation correction unit 110 corrects atmospheric attenuation using the following equation (4). In Equation (4), η is the atmospheric viscosity coefficient, and ρ is the atmospheric density.

  The attenuation amount β in the equation (4) represents the attenuation amount per 1 m, and differs depending on the frequency of the reflected wave. Β = 0 at the frequency f = 0, but β <0 when f> 0, and the frequency characteristic calculated in step S202 is a value that is smaller by the attenuation β due to the influence of the atmosphere. Therefore, in the correction in step S402, a corrected frequency characteristic that eliminates the influence of atmospheric attenuation is obtained by subtracting the attenuation amount β from the calculated frequency characteristic.

  As the frequency of the reflected wave increases and as the distance increases, the effect of atmospheric attenuation becomes greater. Therefore, even if it is a perfect reflector, if the distance changes, the frequency characteristics will change, causing false detection. However, by using the attenuation amount β of the above-described equation (4), the influence of atmospheric attenuation can be corrected based on the distance. As a result, it is possible to classify and detect an object with higher accuracy by holding only one frequency characteristic of a known perfect reflector.

-Fourth Embodiment-
FIG. 12 is a diagram showing a fourth embodiment of the object identification device, and is a block diagram similar to FIG. The basic configuration and operation are the same as those in the block diagram shown in FIG. 1 except that an azimuth detecting unit 111 and an azimuth direction correcting unit 112 are provided, and two detectors 104a and 104b are provided.

  Here, two detectors 104 a and 104 b are provided to detect the azimuth of the measurement object M, and each detection signal is input to the azimuth detection unit 111. The signal analyzer 105 receives a signal from one detector 104a. Between the reflected waves detected by the detectors 104a and 104b, a reaching path difference corresponding to the distance d between the receivers 104a and 104b occurs, and a phase difference occurs between the received signals of the detectors 104a and 104b. The azimuth detector 111 detects this phase difference and measures the arrival direction angle θd of the reflected wave based on the distance between the receivers. The azimuth direction correction unit 112 calculates the influence of directivity when the frequency is changed using the calculated arrival direction angle θd in order to eliminate the influence of the directivity of the transmitter / receiver finger on the frequency characteristics. Correct the characteristics.

FIG. 13 is a flowchart showing the basic operation of the object identification device according to the fourth embodiment. In the flowchart shown in FIG. 3, the processes of steps S501 and S502 related to the direction detection unit 111 and the direction correction unit 112 are added. Is. In step S501, the phase difference φ is calculated based on the received signals from the two receivers 104a and 104b, and the arrival direction angle θd of the reflected wave is calculated based on the equation (5).

  Here, the arrival direction angle θd is calculated from the phase difference φ, but the method of measuring the arrival direction angle θd is not limited to this method. For example, other methods for measuring the direction of arrival angle, such as a monopulse method based on the difference in intensity between the received signals of the two receivers 104a and 104b, a measurement of the arrival time difference by the pulse method, and an arrival direction estimation method using the MUSIC algorithm. It may be used.

In step S502, in order to eliminate the influence due to the frequency change of the transmitter directivity, the influence due to the directivity when the frequency is changed is calculated using the arrival direction angle θd. Note that the relationship between the frequency and directivity in the transmitter 103 is expressed by equation (6) (Reference; Watanabe et al., “Study on directivity of thermally induced nanosilicon ultrasonic source”, pp461-464, Proceeding of “11th symposium of Microjoining and Assembly Technology in Electronics (Mate2005)”, Feb. 2005). In equation (6), D (θd) is the relative intensity at angle θd (1 when θ = 90 degrees (front)), f is the frequency, and a is 1/2 times one side of the heat generating portion 103c (ultrasound generating portion). , C represents the speed of sound.

  In step S502, correction is performed from this relational expression and the frequency characteristic calculated by the signal analysis unit 105 so as to eliminate the influence of the directivity of the transmitter 103 on the frequency characteristic. As described above, in the fourth embodiment, since the frequency characteristic is corrected based on the correlation between the direction and the directivity of the transmitter 103, the influence of the directivity is removed, and the object is accurately detected in all directions. can do.

-Fifth embodiment-
FIG. 14 is a diagram showing a fourth embodiment of the object identification device, and is a block diagram similar to FIG. The basic configuration and operation are the same as those in the block diagram shown in FIG. 1 except that a speed detection unit 113 and a clothing person classification unit 114 are provided. The speed detection unit 113 calculates the Doppler shift based on the transmission signal frequency information of the transmission control unit 102 and the frequency information of the signal analysis unit 105, and calculates the relative speed of the measurement target M. The clothing person classification unit 114 is based on the clothing detection information from the object classification unit 107, the relative speed by the speed detection unit 113, and the own vehicle speed information from a vehicle speed sensor (not shown) mounted on the vehicle. Detection is performed.

  FIG. 15 is a flowchart showing the basic operation of the object identification device according to the fifth embodiment, and steps S601 to S603 related to the speed detection unit 113 and the clothing person classification unit 114 are added to the flowchart shown in FIG. Is. If it is determined in step S206 that the measurement target M is a clothing object, the process proceeds to step S601, and the speed detection unit 113 calculates the relative speed between the host vehicle and the measurement target M.

  In subsequent step S602, based on the difference between the speed of the host vehicle and the relative speed calculated in step S601, the clothing person classification unit 114 determines whether or not the measurement object M is moving. If “YES” is determined in the step S602, the process proceeds to a step S603 to determine the measurement object M as a clothing person. On the other hand, if NO is determined in step S602, step S603 is skipped, and the series of processes is ended.

  As described above, in the fifth embodiment, since the clothing person is detected on the basis of the relative speed with respect to the measurement target M, the accuracy is improved even in a situation where there are large uncertainties in the measurement environment such as when applied to a vehicle. A pedestrian (clothed person) can often be detected.

In the above-described embodiment, the vehicle-mounted object identification device has been described as an example, but the present invention is not limited to being mounted on a vehicle.

It is a block diagram which shows 1st Embodiment of the object identification apparatus by this invention. It is a figure which shows an example of a heat induction type ultrasonic wave generation element, (a) is sectional drawing, (b) is a top view. It is a flowchart explaining the identification operation of 1st Embodiment. It is a figure which shows the frequency characteristic of a received wave. It is a figure which shows the case where the characteristic of FIG. 4 is linearly approximated. It is a block diagram which shows 2nd Embodiment of an object identification apparatus. It is a flowchart explaining the identification operation of 2nd Embodiment. It is a figure explaining the case where a band setting is set to two types, (a) shows a human body and clothes, (b) shows the band setting with respect to clothes A and B. It is a figure which shows the frequency characteristic of clothes A and B with respect to a reference | standard reflection, and each setting band. It is a block diagram which shows 3rd Embodiment of an object identification apparatus. It is a flowchart explaining the identification operation of 3rd Embodiment. It is a block diagram which shows 4th Embodiment of an object identification device. It is a flowchart explaining the identification operation | movement of 4th Embodiment. It is a block diagram which shows 5th Embodiment of an object identification device. It is a flowchart explaining the identification operation of 5th Embodiment.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 100: Control apparatus, 101: Oscillator, 103: Transmitter, 104, 104a, 104b: Receiver, 102: Transmission control part, 105: Signal analysis part, 106: Physical property identification part 106, 107: Object classification part, 108: Frequency domain setting unit, 109: ranging unit, 110: atmospheric attenuation correction unit, 111: azimuth detection unit, 112: azimuth direction correction unit, 113: speed detection unit, 114: clothing person classification unit, M: measurement target

Claims (6)

An object identification device for measuring a person wearing fiber clothes,
And transmitting means for transmitting toward the ultrasound to the measurement object,
Receiving means for receiving a reflected wave from the measurement object;
Detection means for detecting a frequency characteristic representing a relationship between the reception intensity of the reflected wave and the frequency based on reception information of the reception means;
An arithmetic unit that performs linear approximation processing within a predetermined frequency band in the frequency characteristics, and outputs a received intensity gradient amount as a feature amount based on the obtained linear approximation waveform;
A determination unit that compares the feature amount obtained by the calculation unit with a predetermined reference feature amount and determines whether or not a difference between the feature amount and the reference feature amount is included in a predetermined range ;
An object comprising: a determining unit that determines that the measurement target is a person wearing the clothes when the determining unit determines that the difference is included in a predetermined range. Identification device. The object identification device according to claim 1,
An object identification apparatus using a gradient amount of a perfect reflector as the predetermined reference feature amount . The object identification device according to claim 2, further comprising:
Comprising a speed detection means for detecting a relative speed of the measurement object;
Said determining means, said determining means determines the result and based on the detection result of the speed detecting means, object identification apparatus characterized by determining the measurement target and clothes person. In the object identification device according to any one of claims 1 to 3 ,
The said determination means determines based on the said gradient amount in the frequency band of 50 kHz to 100 kHz among the said frequency characteristics, The object identification device characterized by the above-mentioned . In the object identification device according to any one of claims 1 to 4 , in addition,
A distance measuring means for detecting a distance between the measurement object and the receiving means;
An object identification apparatus comprising: atmospheric attenuation correction means for correcting the frequency characteristic based on a correlation between a distance detected by the distance measuring means and atmospheric attenuation of the reflected wave. In the object identification device according to any one of claims 1 to 5 , in addition,
Azimuth detecting means for detecting the azimuth of the measurement object;
An object identification apparatus comprising: an azimuth correction unit that corrects the frequency characteristic based on a correlation between a detection result of the azimuth detection unit and directivity of the transmission unit.

Images