JP2019144192A - Object detector - Google Patents

Abstract

To detect the state of an object with a simple configuration.SOLUTION: An object detector 100 includes: a vibrator 32; a vibration detector 33; and an object detector 116. The vibrator 32 is arranged in a first position of a mounted board 31 and provides a vibration to the mounted board 31. The vibration detector 33 is arranged in a second position, which is different from the first position of the mounted board 31, and detects a vibration of the mounted board 31. The object detector 116 detects the state of an object BJ on a mounted surface SF of the mounted board 31 according to the result of detection by the vibration detector 33.SELECTED DRAWING: Figure 3

Description

  The present invention relates to an object detection device that detects a state of an object placed on a placement surface.

  An object position recognition device described in Patent Literature 1 includes a detection panel, a plurality of first transmission / reception coils, a plurality of second transmission / reception coils, a control unit, a plurality of lateral transmission / reception units, and a plurality of vertical transmission / reception units. A transmission / reception means, a horizontal direction switching means, a vertical direction switching means, and a transmission / reception means are provided. The plurality of first transmission / reception coils are arranged in the horizontal direction of the detection panel, and the plurality of second transmission / reception coils are arranged in the vertical direction of the detection panel. The horizontal direction switching means switches a plurality of horizontal direction transmission / reception means. The vertical direction switching means switches a plurality of vertical direction transmission / reception means. A memory device that stores identification data is attached to the object. The transmission / reception means transmits a data read command and receives identification data from the memory device.

  When an object exists on the detection panel, the horizontal direction switching unit sequentially drives the horizontal direction transmission / reception unit, and the transmission / reception unit transmits a data read command. A resonance circuit provided in the memory device resonates at the carrier frequency, and the residual vibration is intermittently transmitted to transmit identification data. The vertical direction switching unit sequentially drives the vertical direction transmission / reception unit, and the transmission / reception unit transmits a data read command. A resonance circuit provided in the memory device resonates at the carrier frequency, and the residual vibration is intermittently transmitted to transmit identification data. The control means, based on the column number of the switching control signal given to the horizontal direction switching means and the row number of the switching control signal given to the vertical direction switching means when the transmission / reception means receives the identification data, The position of is detected. The control means detects the type of the object based on the identification data.

JP-A-8-220248

  However, the object position recognition device described in Patent Document 1 includes a plurality of first transmission / reception coils, a plurality of second transmission / reception coils, a plurality of horizontal transmission / reception means, a plurality of vertical transmission / reception means, and a horizontal direction. It is necessary to provide switching means, vertical direction switching means, and transmission / reception means. Therefore, the manufacturing cost of the object position recognition device increases.

  The present invention has been made in view of the above problems, and an object of the present invention is to provide an object detection device capable of detecting the state of an object with a simple configuration.

  An object detection apparatus according to the present invention includes a vibrator, a vibration detector, and an object detection unit. The said vibrator is arrange | positioned in the 1st position of a mounting board, and provides a vibration to the mounting board mentioned above. The vibration detector is disposed at a second position different from the first position of the mounting plate, and detects vibration of the mounting plate. The object detection unit detects the state of the object placed on the placement surface of the placement plate based on the detection result of the vibration detector.

  In the object detection device according to the present invention, it is preferable that the state includes at least one of a position, a type, a posture, and a content.

  In the object detection apparatus according to the present invention, it is preferable that the vibration exciter imparts a vibration whose frequency changes with time to the mounting plate.

  In the object detection apparatus according to the present invention, it is preferable that the vibrator has a first vibrating body that outputs vibration, and the first vibrating body applies the vibration to the mounting plate. The vibration detector preferably includes a second vibrating body that detects the vibration, and the second vibrating body detects the vibration of the mounting plate.

  In the object detection apparatus according to the present invention, it is preferable that the vibrator has a first piezoelectric element that outputs acoustic vibration, and the first piezoelectric element applies the acoustic vibration to the mounting plate. The vibration detector preferably includes a second piezoelectric element that detects the acoustic vibration, and the second piezoelectric element preferably detects the acoustic vibration of the mounting plate.

  In the object detection device according to the present invention, the vibration detector includes a first vibration detector disposed at the second position of the mounting plate, the first position and the second position of the mounting plate. It is preferable to have a second vibration detector disposed at a different third position.

  The object detection apparatus according to the present invention further includes a conversion unit that performs Fourier transform on the detection signal of the vibration detector and generates an amplitude spectrum, the object detection unit based on the amplitude spectrum, It is preferable to detect the state of the object.

  The object detection apparatus according to the present invention preferably further comprises a learning unit that learns the amplitude spectrum, and the object detection unit detects the state of the object based on a learning result of the learning unit.

  In the object detection device according to the present invention, it is preferable that the conversion unit generates a first amplitude spectrum and a second amplitude spectrum. The first amplitude spectrum indicates the amplitude spectrum when no object is placed on the placement surface. The second amplitude spectrum indicates the amplitude spectrum when each of a plurality of objects is placed on the placement surface. The learning unit preferably learns based on the first amplitude spectrum and the second amplitude spectrum. The object detection unit preferably detects the state of each of the plurality of objects based on a learning result of the learning unit.

  The object detection apparatus according to the present invention preferably further includes a generation unit that generates a third amplitude spectrum based on the first amplitude spectrum and the second amplitude spectrum. The third amplitude spectrum is an amplitude spectrum when the plurality of objects are placed on the placement surface. Preferably, the learning unit learns the third amplitude spectrum, and the object detection unit detects the state of each of the plurality of objects based on a learning result of the learning unit.

  In the object detection device according to the present invention, it is preferable that the conversion unit generates a fourth amplitude spectrum and a fifth amplitude spectrum. The fourth amplitude spectrum indicates the second amplitude spectrum when the first object is placed on the placement surface. The fifth amplitude spectrum indicates the second amplitude spectrum when a second object different from the first object is placed on the placement surface. The generation unit preferably generates a sum spectrum. The sum spectrum indicates a sum of amplitudes for each frequency of the fourth amplitude spectrum and the fifth amplitude spectrum. Preferably, the generation unit generates an amplitude difference for each frequency between the sum spectrum and the first amplitude spectrum as the third amplitude spectrum. It is preferable that the learning unit learns the third amplitude spectrum. Preferably, the object detection unit detects the state of the first object and the state of the second object based on a learning result of the learning unit.

  In the object detection device according to the present invention, it is preferable that the conversion unit generates a fourth amplitude spectrum, a fifth amplitude spectrum, and a sixth amplitude spectrum. The fourth amplitude spectrum indicates the second amplitude spectrum when the first object is placed on the placement surface. The fifth amplitude spectrum indicates the second amplitude spectrum when a second object different from the first object is placed on the placement surface. The sixth amplitude spectrum indicates the second amplitude spectrum when a third object different from the first object and the second object is placed on the placement surface. The generation unit preferably generates a sum spectrum and a difference spectrum. The sum spectrum indicates a sum of amplitudes for each frequency of the fourth amplitude spectrum, the fifth amplitude spectrum, and the sixth amplitude spectrum. The difference spectrum indicates an amplitude difference for each frequency between the sum spectrum and the first amplitude spectrum. It is preferable that the generation unit generates an amplitude difference for each frequency between the difference spectrum and the first amplitude spectrum as the third amplitude spectrum. It is preferable that the learning unit learns the third amplitude spectrum. Preferably, the object detection unit detects the state of the first object, the state of the second object, and the state of the third object based on a learning result of the learning unit.

  According to the object detection device of the present invention, the state of an object can be detected with a simple configuration.

It is a figure which shows an example of a structure of the object detection apparatus which concerns on embodiment of this invention. (A) is a top view which shows an example of arrangement | positioning in the mounting plate of a vibrator, a vibration detector, and a supporting member. (B) is a side view which shows an example of arrangement | positioning in the mounting plate of a vibrator, a vibration detector, and a supporting member. It is a figure which shows an example of a structure of the control part which concerns on embodiment of this invention. It is a figure which shows an example of the change of the frequency of the acoustic vibration applied to a vibrator. It is a figure which shows an example of the experimental method which detects the position of an object. It is a figure which shows an example of the relationship between the position of an object, and an amplitude spectrum. It is a figure which shows an example of the identification rate of the position of an object. It is a figure which shows an example of the experimental method which detects the kind of object. (A) is a figure which shows an example of the experiment method in case the kind of object is a mug. (B) is a figure which shows an example of the experiment method in case the kind of object is a mandarin orange. It is a figure which shows an example of the kind of object. It is a figure which shows an example of the relationship between the kind of object, and an amplitude spectrum. It is a figure which shows an example of the identification rate of the kind of object. It is a figure which shows an example of the experimental method which detects the attitude | position of an object. (A) is a figure which shows an example of the experimental method in case a building block is arrange | positioned vertically long. (B) is a figure which shows an example of the experimental method in case a building block is arrange | positioned horizontally long. It is a figure which shows an example of the relationship between the attitude | position of an object, and an amplitude spectrum. It is a figure which shows an example of the experimental method which detects the content volume of an object. (A) is a figure which shows an example of the experiment method in case water does not enter into the cup. (B) is a figure which shows an example of the experiment method in case water is contained in the cup. It is a figure which shows an example of the relationship between the content volume of an object, and an amplitude spectrum. It is a figure which shows an example of the method of detecting the position and kind of two objects. (A) is a figure which shows an example of the production | generation method of a 1st amplitude spectrum and a 2nd amplitude spectrum. (B) is a figure which shows an example of the estimation method of a 3rd amplitude spectrum. (C) is a figure which shows an example of the production | generation method of a 3rd amplitude spectrum. It is a flowchart which shows an example of a process of the control part in the case of detecting the position and kind of two objects. It is a flowchart which shows an example of the learning process of a control part. It is a flowchart which shows an example of the detection evaluation process of a control part. It is a figure which shows an example of the identification rate of the position and kind of two objects. It is a figure which shows an example of the method of detecting the position and kind of three objects. (A) is a figure which shows an example of the detection method of a 1st amplitude spectrum and a 2nd amplitude spectrum. (B) is a figure which shows an example of the estimation method of a 3rd amplitude spectrum. (C) is a figure which shows an example of the production | generation method of a 3rd amplitude spectrum. It is a figure which shows an example of the identification rate of the position and kind of three objects.

  Embodiments of the present invention will be described below with reference to the drawings (FIGS. 1 to 22). In the drawings, the same or corresponding parts are denoted by the same reference numerals and description thereof is not repeated.

<Configuration of Object Detection Device 100>
First, the configuration of the object detection apparatus 100 according to the present embodiment will be described with reference to FIG. FIG. 1 is a diagram illustrating an example of the configuration of the object detection apparatus 100. As shown in FIG. 1, the object detection device 100 includes a personal computer 1, an audio interface 2, and a detection device body 3.

  The personal computer 1 includes a control unit 11 and a display 12. The control unit 11 controls the audio interface 2 and the detection device main body 3. The control unit 11 includes a processor 11A and a storage unit 11B. The processor 11A includes, for example, a CPU (Central Processing Unit). The storage unit 11B includes a memory such as a semiconductor memory, and may include an HDD (Hard Disk Drive). The storage unit 11B stores a control program. The display 12 includes, for example, an LCD (Liquid Crystal Display) and displays various images. The control unit 11 will be described in detail later with reference to FIG.

  The audio interface 2 transmits and receives signals to and from the detection device body 3 in accordance with instructions from the control unit 11. The audio interface 2 includes a signal generator 21. The signal generator 21 generates a signal (sweep signal) whose frequency F changes as time T elapses. The sweep signal will be described in detail later with reference to FIG.

  The detection device main body 3 includes a mounting plate 31, a vibrator 32, a vibration detector 33, and a support body 34. FIG. 1 shows the X axis, the Y axis, and the Z axis. The Z axis is parallel to the vertical direction. The X axis and the Y axis are parallel to the horizontal direction. The X axis, the Y axis, and the Z axis are orthogonal to each other. Specifically, the X axis indicates the longitudinal direction of the mounting plate 31. The Y axis indicates the width direction of the mounting plate 31.

  The mounting plate 31 is, for example, a rectangular acrylic plate. Specifically, the mounting plate 31 has a thickness of 5 mm, a width of 180 mm, and a length of 320 mm. The mounting plate 31 has a mounting surface SF. The placement surface SF corresponds to the upper surface of the placement plate 31 on which the object BJ is placed. In addition, a vibrator 32 and a vibration detector 33 are arranged on the mounting plate 31.

  The vibrator 32 is disposed at the first position P1 of the mounting plate 31 and applies vibration to the mounting plate 31. For example, the first position P <b> 1 indicates the position of the center in the width direction (Y-axis direction) of the mounting plate 31 and one end in the longitudinal direction of the mounting plate 31 (the negative direction end in the X-axis direction). The first position P1 is located on the lower surface of the mounting plate 31, for example. Specifically, the vibrator 32 imparts acoustic vibration corresponding to the sweep signal generated by the signal generator 21 to the mounting plate 31. The vibrator 32 includes a first piezoelectric element that outputs acoustic vibration. The first piezoelectric element is, for example, a piezo element. In other words, the vibrator 32 is a speaker. The vibrator 32 is formed in a thin circular plate shape and has a diameter of 20 mm.

  The vibration detector 33 is disposed on the mounting plate 31 and detects vibration of the mounting plate 31. The second position P2 is different from the first position P1. The second position P2 is located on the lower surface of the mounting plate 31, for example. A detection signal SG detected by the vibration detector 33 is input to the audio interface 2. The detection signal SG indicates acoustic vibration. The audio interface 2 acquires the detection signal SG detected by the vibration detector 33 at a sampling frequency of 96 kHz, for example.

  The vibration detector 33 includes a first vibration detector 331 and a second vibration detector 332. The first vibration detector 331 is disposed at the second position P <b> 2 of the placement plate 31. The second vibration detector 332 is disposed at the third position P3 of the placement plate 31. The third position P3 is different from the first position P1 and the second position P2. The third position P3 is located on the lower surface of the mounting plate 31, for example. For example, the second position P <b> 2 indicates the position of the other end in the longitudinal direction of the mounting plate 31 (the positive end in the X-axis direction) at the center in the width direction (Y-axis direction) of the mounting plate 31. The third position P3 indicates the center of the mounting plate 31 in the longitudinal direction (X-axis direction) and the position of one end in the width direction of the mounting plate 31 (the positive end in the Y-axis direction).

  Specifically, each of the first vibration detector 331 and the second vibration detector 332 includes a second piezoelectric element that detects acoustic vibration. The second piezoelectric element is, for example, a piezo element. In other words, each of the first vibration detector 331 and the second vibration detector 332 is a microphone. Each of the first vibration detector 331 and the second vibration detector 332 is formed in a thin circular plate shape and has a diameter of 20 mm.

  The support 34 supports the placement plate 31. The support body 34 includes a first support body 341, a second support body 342, a third support body 343, and a fourth support body 344. Each of the first support body 341 to the fourth support body 344 is formed in a cylindrical shape with felt. Each of the first support 341 to the fourth support 344 has a diameter of 32 mm and a thickness of 5 mm.

  The first support 341 supports the first end portion of the mounting plate 31. The first end portion is an end portion that is the negative end of the X axis of the mounting plate 31 and the negative end of the Y axis. The second support 342 supports the second end portion of the mounting plate 31. The second end portion is an end portion that is the negative end of the X axis of the mounting plate 31 and the positive end of the Y axis. The third support 343 supports the third end portion of the placement plate 31. The third end is an end that is the positive end of the X axis of the mounting plate 31 and the positive end of the Y axis. The fourth support 344 supports the fourth end portion of the mounting plate 31. The fourth end is an end that is the positive end of the X axis of the mounting plate 31 and the negative end of the Y axis.

  Next, with reference to FIG.1 and FIG.2, arrangement | positioning in the mounting plate 31 of the vibrator 32, the vibration detector 33, and the support body 34 is demonstrated concretely. FIG. 2A is a plan view showing an example of the arrangement of the vibrator 32, the vibration detector 33, and the support 34 on the mounting plate 31. FIG. 2B is a side view showing an example of the arrangement of the vibrator 32, the vibration detector 33, and the support 34 on the mounting plate 31.

  As shown in FIGS. 2A and 2B, the mounting plate 31 is formed in a rectangular flat plate shape. Each of the first support body 341 to the fourth support body 344 is disposed so that the peripheral surface of the support body 34 is in contact with the long side and the short side of the mounting plate 31. That is, each of the first support body 341 to the fourth support body 344 is disposed at the four corners of the mounting plate 31.

  The vibrator 32 is arranged such that the center of the vibrator 32 is located on a straight line connecting the center of the first support 341 and the center of the second support 342. The first vibration detector 331 is arranged so that the center of the first vibration detector 331 is positioned on a straight line connecting the center of the third support 343 and the center of the fourth support 344. The second vibration detector 332 is arranged so that the center of the second vibration detector 332 is positioned on a straight line connecting the center of the second support 342 and the center of the third support 343.

  Next, with reference to FIGS. 1-3, the structure of the control part 11 which concerns on embodiment of this invention is demonstrated. FIG. 3 is a diagram illustrating an example of the configuration of the control unit 11. As illustrated in FIG. 3, the control unit 11 includes a vibration applying unit 111, a vibration acquisition unit 112, a conversion unit 113, a generation unit 114, a learning unit 115, and an object detection unit 116. Specifically, when the processor 11A of the control unit 11 executes the control program, the control unit 11 includes the vibration applying unit 111, the vibration acquisition unit 112, the conversion unit 113, the generation unit 114, the learning unit 115, and the object detection unit. It functions as 116.

  The vibration applying unit 111 instructs the signal generator 21 so that the vibrator 32 applies acoustic vibration to the mounting plate 31.

  The vibration acquisition unit 112 acquires the detection signal SG from the vibration detector 33 via the audio interface 2.

  The conversion unit 113 performs a Fourier transform on the detection signal SG to generate an amplitude spectrum SP. Specifically, the conversion unit 113 performs FFT (Fast Fourier Transform) on the detection signal SG to generate an amplitude spectrum SP. The conversion unit 113 generates a first amplitude spectrum SP1 and a second amplitude spectrum SP2. The first amplitude spectrum SP1 indicates the amplitude spectrum SP when the object BJ is not placed on the placement surface SF. The second amplitude spectrum SP2 indicates the amplitude spectrum SP when each of the plurality of objects BJ is placed on the placement surface SF.

  The generation unit 114 generates a third amplitude spectrum SP3 based on the first amplitude spectrum SP1 and the second amplitude spectrum SP2. Specifically, the generation unit 114 generates an estimated spectrum SP3G of the third amplitude spectrum SP3 based on the first amplitude spectrum SP1 and the second amplitude spectrum SP2. The third amplitude spectrum SP3 indicates the amplitude spectrum SP when a plurality of objects BJ are placed on the placement surface SF.

  The learning unit 115 learns the third amplitude spectrum SP3. Specifically, the learning unit 115 learns the estimated spectrum SP3G of the third amplitude spectrum SP3.

  The object detection unit 116 detects the state ST of the object BJ placed on the placement surface SF based on the detection result of the vibration detector 33. Specifically, the object detection unit 116 detects the state ST of the object BJ placed on the placement surface SF based on the detection signal SG acquired by the vibration acquisition unit 112. Further, the object detection unit 116 displays the detection result on the display 12. The state ST includes a position PS, a type TP, a posture PT, and an internal capacity CP. Moreover, the object detection part 116 detects each state ST of several object BJ based on 3rd amplitude spectrum SP3, for example. Moreover, the object detection part 116 detects each state ST of the some object BJ based on the learning result of the learning part 115, for example. The generation unit 114, the learning unit 115, and the object detection unit 116 will be described in detail later with reference to FIGS.

The case where the object detection unit 116 detects the position PS of one object BJ placed on the placement surface SF will be described in detail later with reference to FIGS.
The case where the object detection unit 116 detects the type TP of one object BJ placed on the placement surface SF will be described in detail later with reference to FIGS.
The case where the object detection unit 116 detects the posture PT of one object BJ placed on the placement surface SF will be described in detail later with reference to FIGS. 12 to 13.
The case where the object detection unit 116 detects the internal capacity CP of one object BJ placed on the placement surface SF will be described in detail later with reference to FIGS. 14 to 15.
The case where the object detection unit 116 detects the position PS and the type TP of each of the plurality of objects BJ will be described in detail later with reference to FIGS. 16 to 22.

  As described above with reference to FIGS. 1 to 3, in the embodiment of the present invention, the vibration exciter 32 imparts vibration to the placement plate 31, and the vibration detector 33 causes vibration of the placement plate 31. Based on the detection result of the vibration detector 33, the state ST of the object BJ placed on the placement surface SF is detected. Therefore, the state ST of the object BJ can be detected with a simple configuration. Therefore, the manufacturing cost of the object detection apparatus 100 can be reduced.

  Further, Fourier transform is performed on the detection signal SG of the vibration detector 33 to generate an amplitude spectrum SP, and the state ST of the object BJ is detected based on the amplitude spectrum SP. Therefore, the state ST of the object BJ placed on the placement surface SF can be detected more accurately with a simple configuration.

  Further, the vibration detector 33 includes a first vibration detector 331 and a second vibration detector 332 that are arranged at different positions. Therefore, the state ST of the object BJ placed on the placement surface SF is detected based on the detection signal SG of the first vibration detector 331 and the detection signal SG of the second vibration detector 332. Therefore, the state ST of the object BJ placed on the placement surface SF can be accurately detected.

  In the embodiment of the present invention, the vibration detector 33 includes two vibration detectors (the first vibration detector 331 and the second vibration detector 332), but the present invention is not limited to this. The vibration detector 33 may have one or more vibration detectors. As the number of vibration detectors increases, the state ST of the object BJ placed on the placement surface SF may be more accurately detected.

  In the embodiment of the present invention, the first vibration detector 331 is disposed at the second position P2 and the second vibration detector 332 is disposed at the third position P3. However, the present invention is not limited to this. The 1st vibration detector 331 and the 2nd vibration detector 332 should just be arrange | positioned in the position which mutually differs.

  Next, the acoustic vibration applied to the vibrator 32 will be described with reference to FIGS. FIG. 4 is a diagram illustrating an example of a change in the frequency F of the acoustic vibration applied to the vibrator 32. The horizontal axis of the graph shown in FIG. 4 indicates time T, and the vertical axis of the graph indicates frequency F. As shown by a graph G1 in FIG. 4, the frequency F of the acoustic vibration applied to the vibrator 32 changes linearly from 20 kHz to 40 kHz within a period of 1 second. In other words, the vibration exciter 32 applies acoustic vibration to the mounting plate 31 within a period of 1 second. The frequency F of the acoustic vibration is 20 kHz when the application of the acoustic vibration to the mounting plate 31 is started, and is linear so as to be 40 kHz when the application of the acoustic vibration to the mounting plate 31 is finished. To increase.

  As described above with reference to FIGS. 1 to 4, in the embodiment of the present invention, the vibration exciter 32 imparts a vibration whose frequency F changes with time T to the placement plate 31. . Therefore, vibrations having a plurality of frequencies F can be efficiently applied to the mounting plate 31. Therefore, the time required for detecting the state ST of the object BJ by the object detection apparatus 100 can be reduced.

  The first piezoelectric element applies acoustic vibration to the mounting plate 31, and the second piezoelectric element detects acoustic vibration of the mounting plate 31. Therefore, the state ST of the object BJ can be detected with a simpler configuration. Therefore, the manufacturing cost of the object detection apparatus 100 can be further reduced.

  Furthermore, since the state ST includes the position PS, the type TP, the posture PT, and the internal capacity CP, the position PS, the type TP, the posture PT, and the internal capacity CP of the object BJ can be detected. The case where the position PS of the object BJ is detected will be described in detail later with reference to FIGS. The case where the type TP of the object BJ is detected will be described in detail later with reference to FIGS. The case where the posture PT of the object BJ is detected will be described in detail later with reference to FIGS. The case of detecting the internal volume CP of the object BJ will be described in detail later with reference to FIGS.

  In the embodiment of the present invention, the vibrator 32 includes a first piezoelectric element that outputs acoustic vibration, and the first piezoelectric element imparts acoustic vibration to the mounting plate 31. It is not limited. The vibrator 32 may apply vibration to the mounting plate 31. In other words, the vibration applied to the mounting plate 31 by the vibrator 32 is not limited to acoustic vibration. The vibrator 32 may include a first vibrating body that outputs vibration, and the first vibrating body may apply vibration to the mounting plate 31. For example, when the vibrator is a crystal vibrator, the frequency F of vibration applied to the mounting plate 31 is 1 to 20 MHz.

  Further, when the first vibrating body of the vibrator 32 imparts vibration to the mounting plate 31, the vibration detector 33 has a second vibrating body that detects vibration, and the second vibrating body is It is preferable to detect the vibration of the mounting plate 31. In this case, the first vibrating body applies vibration to the mounting plate 31, and the second vibrating body detects vibration of the mounting plate 31. Therefore, the state ST of the object BJ can be detected with a simple configuration. Therefore, the manufacturing cost of the object detection apparatus 100 can be reduced.

  In the embodiment of the present invention, the vibration exciter 32 imparts vibration on the mounting plate 31 whose frequency F linearly changes with respect to time T, but the present invention is not limited to this. The vibration exciter 32 only needs to apply to the mounting plate 31 vibration whose frequency F changes as time T elapses. For example, as the frequency F is lower, vibration that changes the frequency F may be applied to the mounting plate 31 so that the change in the frequency F in a predetermined time (for example, 0, 1 second) is smaller. In this case, the FFT for the low frequency F can be performed accurately.

  In the embodiment of the present invention, the state ST includes the position PS, the type TP, the posture PT, and the internal capacity CP, but the present invention is not limited to this. The state ST may include at least one of the position PS, the type TP, the posture PT, and the internal capacity CP.

<Experiment for Detecting Position PS of One Object BJ>
Next, an experiment in which the object detection device 100 detects the position PS of one object BJ placed on the placement surface SF will be described with reference to FIGS. FIG. 5 is a diagram illustrating an example of an experimental method for detecting the position PS of the object BJ. As shown in FIG. 5, on the placement surface SF of the placement plate 31, the object BJ is placed at the center position in the width direction (Y-axis direction) of the placement surface SF, and in the longitudinal direction (X-axis direction). Moved along. As the object BJ, a cubic block having a side of 44 mm was used.

  Specifically, the amplitude spectrum SP when the object BJ is not placed on the placement surface SF is generated. Further, the object spectrum BJ is moved at an interval of 5 mm from the position of 30 mm to the position of 290 mm from the end on the negative direction side of the X axis of the placement surface SF on the center line CL to generate the amplitude spectrum SP. That is, the amplitude spectrum SP was generated in the following manner at 53 positions (= (290-30) ÷ 5 + 1) along the longitudinal direction (X-axis direction) on the center line CL. That is, the vibration imparting unit 111 imparts acoustic vibration to the mounting plate 31 with respect to the vibrator 32. The vibration acquisition unit 112 acquires the detection signal SG from the vibration detector 33. Further, the conversion unit 113 generates an amplitude spectrum SP from the detection signal SG.

  Also, 12 experiments are performed for each of the case where the object BJ is not placed and the case where the object BJ is placed at each of the 53 positions, and 648 (= (53 + 1) × 12) amplitude spectra SP are generated. did. The 648 amplitude spectra SP were divided into 6 groups, and the identification rate was obtained by cross-validation.

  FIG. 6 is a diagram illustrating an example of the relationship between the position PS of the object BJ and the amplitude spectrum SP. The horizontal axis of the graph shown in FIG. 6 indicates the frequency F, and the vertical axis indicates the amplitude B. A graph G11 shown in FIG. 6 shows an amplitude spectrum SP when the object BJ is placed at a position of 95 mm from the end of the placement surface SF on the negative direction side of the X axis. A graph G12 shown in FIG. 6 shows an amplitude spectrum SP when the object BJ is placed at a position of 160 mm from the end of the placement surface SF on the negative direction side of the X axis. A graph G13 illustrated in FIG. 6 illustrates an amplitude spectrum SP when the object BJ is placed at a position 225 mm from the end of the placement surface SF on the negative side of the X axis.

  As shown in FIG. 6, the amplitude B at the frequency F of 37.5 kHz decreased as the distance from the end on the negative direction side of the X axis of the mounting surface SF increased. Further, the amplitude B near the frequency F of 40 kHz increased as the distance from the end of the placement surface SF on the negative direction side of the X axis increased. Thus, the amplitude spectrum SP changed according to the distance from the end portion on the negative direction side of the X axis of the mounting surface SF. Therefore, as described in detail below with reference to FIG. 7, the object detection apparatus 100 can detect the position PS of one object BJ placed on the placement surface SF based on the amplitude spectrum SP. found.

  Next, with reference to FIG. 7, an experimental result of an experiment in which the object detection device 100 detects the position PS of one object BJ placed on the placement surface SF will be described. FIG. 7 is a diagram illustrating an example of the identification rate of the position PS of the object BJ. As shown in FIG. 7, when the amplitude spectrum SP is generated using the detection signal SG of the first vibration detector 331, the identification rate is 85.5%, the average absolute error is 2.35 mm, The standard deviation was 4.73 mm. The “average absolute error” indicates an average value of absolute values of errors in the detection result of the position PS of the object BJ. When the amplitude spectrum SP is generated using the detection signal SG of the second vibration detector 332, the identification rate is 85.5%, the average absolute error is 0.621 mm, and the standard deviation is 0.516 mm. there were. When the amplitude spectrum SP is generated using the detection signal SG of the first vibration detector 331 and the detection signal SG of the second vibration detector 332, the identification rate is 89.2% and the average absolute error is 0. .377 mm with a standard deviation of 0.411 mm.

  As described above with reference to FIGS. 1 to 7, it has been found that the position PS of one object BJ placed on the placement surface SF can be accurately detected in the embodiment of the present invention.

  Further, the case where the amplitude spectrum SP is generated using the detection signal SG of the second vibration detector 332 is compared with the case where the amplitude spectrum SP is generated using the detection signal SG of the first vibration detector 331. It has been found that the average absolute error and the standard deviation are small, and the position PS of the object BJ can be detected more accurately. This is because the first vibration detector 331 is disposed at a position symmetrical to the vibrator 32 with respect to a center line extending in the width direction through the center position in the longitudinal direction of the mounting plate 31. Presumed.

  Further, when the amplitude spectrum SP is generated using the detection signal SG of the first vibration detector 331 and the detection signal SG of the second vibration detector 332, the detection signal SG of the first vibration detector 331 is used. Compared to the case where the amplitude spectrum SP is generated and the case where the amplitude spectrum SP is generated using the detection signal SG of the second vibration detector 332, the discrimination rate is high, and the position PS of the object BJ is more accurately determined. It turned out to be detectable.

<Experiment for Detecting Position PS of One Object BJ>
Next, an experiment in which the object detection device 100 detects the type TP of one object BJ placed on the placement surface SF will be described with reference to FIGS. FIG. 8 is a diagram illustrating an example of an experimental method for detecting the type TP of the object BJ. FIG. 8A is a diagram illustrating an example of an experiment method when the type TP of the object BJ is “mug cup”. In FIG. 8A, the “mug cup” is described as an object BJA. FIG. 8B is a diagram illustrating an example of an experimental method when the type TP of the object BJ is “mandarin orange”. In FIG. 8B, “mandarin orange” is described as an object BJB. As shown in FIGS. 8A and 8B, at the center position in the width direction (Y-axis direction) of the placement surface SF and at the center position in the longitudinal direction (X-axis direction) of the placement surface SF. The object BJ was placed and the type TP of the object BJ was changed.

  Specifically, the amplitude spectrum SP when the object BJ is not placed on the placement surface SF is generated. Further, each of the 26 types of objects BJ shown in FIG. 9 is placed at the center position in the width direction (Y-axis direction) of the placement surface SF and the center position in the longitudinal direction (X-axis direction) of the placement surface SF. To generate an amplitude spectrum SP. More specifically, each of the 26 types of objects BJ was placed at the center position of the placement surface SF, and the amplitude spectrum SP was generated as follows. That is, the vibration imparting unit 111 imparts acoustic vibration to the mounting plate 31 with respect to the vibrator 32. The vibration acquisition unit 112 acquires the detection signal SG from the vibration detector 33. Further, the conversion unit 113 generates an amplitude spectrum SP from the detection signal SG.

  Further, 12 experiments were performed for each of the case where the object BJ was not placed on the placement surface SF and the case where each of the 26 types of objects BJ was placed, and 324 pieces (= (26 + 1) × 12). Amplitude spectrum SP was generated. An experiment for generating 324 amplitude spectra SP (referred to as “one-session experiment”) was performed twice a day for 3 days, and 1944 (= 324 × 6) amplitude spectra SP were generated. In the following description, an experiment that generates 324 amplitude spectra SP is referred to as “one session experiment”. 1944 amplitude spectra SP were divided into 6 groups, and the identification rate was obtained by cross-validation.

  As a method of dividing into six groups, an identification rate was obtained by cross-validation for “a method of randomly dividing” and “a method for each session”. The “method of randomly dividing” refers to a method of randomly dividing 1944 amplitude spectra SP into six groups. The “method of dividing for each session” indicates a method of dividing into 6 groups by dividing into 6 sessions performed twice a day for 3 days.

  FIG. 9 shows the type TP of the object BJ used in the experiment in which the object detection apparatus 100 detects the type TP of the object BJ. In the embodiment of the present invention, the object BJ placed on the desk and the object BJ stored in the refrigerator are mainly selected as the type TP of the object BJ. For example, “scissors”, “pens”, “watches” and “notebooks” are objects BJ placed on a desk, and “potatoes”, “mandarin oranges”, “apples” and “onions” are stored in a refrigerator. Is the object BJ.

  FIG. 10 is a diagram illustrating an example of the relationship between the type TP of the object BJ and the amplitude spectrum SP. The horizontal axis of the graph shown in FIG. 10 indicates the frequency F, and the vertical axis indicates the amplitude B. A graph G21 illustrated in FIG. 10 illustrates an amplitude spectrum SP when a “mug” is placed on the placement surface SF as the object BJ. A graph G22 illustrated in FIG. 10 illustrates an amplitude spectrum SP when a “note” is placed on the placement surface SF as the object BJ.

  As shown in FIG. 10, when the frequency F is 2.1 kHz and the frequency F is 2.4 kHz, the amplitude B of the amplitude spectrum SP when the “mug” is placed on the placement surface SF is “note”. It is larger than the amplitude B of the amplitude spectrum SP when placed on the placement surface SF. When the frequency F is 2.25 kHz, the amplitude B of the amplitude spectrum SP when the “mug” is placed on the placement surface SF is the same as the amplitude spectrum SP when the “note” is placed on the placement surface SF. Is smaller than the amplitude B. Therefore, as described in detail below with reference to FIG. 11, the object detection apparatus 100 can detect the type TP of one object BJ placed on the placement surface SF based on the amplitude spectrum SP. found.

  Next, with reference to FIG. 11, an experimental result of an experiment in which the object detection device 100 detects the type TP of the object BJ will be described. FIG. 11 is a diagram illustrating an example of the identification rate of the type TP of the object BJ. As shown in FIG. 11, when the amplitude spectrum SP is generated using the detection signal SG of the first vibration detector 331, the identification rate when “randomly dividing” is 98.2%, In the case of “divided into two”, the identification rate was 87.9%. In the case where the amplitude spectrum SP is generated using the detection signal SG of the second vibration detector 332, the identification rate when “randomly dividing” is 95.5%, and the identification when “dividing for each session” is performed. The rate was 74.7%. When the amplitude spectrum SP is generated using the detection signal SG of the first vibration detector 331 and the detection signal SG of the second vibration detector 332, the identification rate when “randomly dividing” is 98.9%. Yes, the identification rate when “divided for each session” was 87.2%.

  As described above with reference to FIGS. 1 to 4 and FIGS. 8 to 11, in the embodiment of the present invention, the type TP of one object BJ placed on the placement surface SF can be accurately detected. There was found.

  The identification rate when the amplitude spectrum SP is generated using the detection signal SG of the second vibration detector 332 is the identification rate when the amplitude spectrum SP is generated using the detection signal SG of the first vibration detector 331. It was higher. That is, in the case where the amplitude spectrum SP is generated using the detection signal SG of the second vibration detector 332, the object is generated in comparison with the case where the amplitude spectrum SP is generated using the detection signal SG of the first vibration detector 331. It has been found that the position PS of BJ can be detected more accurately.

  The discrimination rate when the amplitude spectrum SP is generated using the detection signal SG of the first vibration detector 331 and the detection signal SG of the second vibration detector 332 is the detection rate SG of the first vibration detector 331. It was higher than the discrimination rate when the amplitude spectrum SP was generated by using the detection signal SG of the second vibration detector 332 and the discrimination rate when the amplitude spectrum SP was generated. That is, the detection signal SG of the first vibration detector 331 and the second vibration detector are compared with the case where the amplitude spectrum SP is generated using the detection signals SG of the first vibration detector 331 and the second vibration detector 332. It was found that the position PS of the object BJ can be detected more accurately when the amplitude spectrum SP is generated using the detection signal SG of 332.

  In addition, it was found that the identification rate was higher in the case of “random division” than in the case of “division for each session”. This is presumably because the amplitude spectrum SP is affected by the environmental temperature at which the experiment was conducted. In other words, the reason why the classification rate decreases when “divide by session” is that the amplitude spectrum SP measured in the same environment as the training case set in cross-validation is not included in the test case set in cross-validation. The training case set indicates data that is learned in advance when performing identification in cross-validation, and the test case set indicates data that is actually identified at that time.

<Experiment for detecting the posture PT of one object BJ>
Next, an experiment in which the object detection apparatus 100 detects the posture PT of one object BJ placed on the placement surface SF will be described with reference to FIGS. FIG. 12 is a diagram illustrating an example of an experimental method for detecting the posture PT of the object BJ. A block BJC was used as the object BJ. Fig.12 (a) is a figure which shows an example of the experiment method in case the building blocks BJC are arrange | positioned vertically long. FIG. 12B is a diagram illustrating an example of an experiment method when the building blocks BJC are arranged horizontally. As shown in FIG. 12A and FIG. 12B, at the center position in the width direction (Y-axis direction) of the placement surface SF and at the center position in the longitudinal direction (X-axis direction) of the placement surface SF. The object BJ was placed, and the posture PT of the building block BJC was changed.

  Specifically, as illustrated in FIG. 12A, the block BJC is arranged vertically in the center of the placement surface SF, and the amplitude spectrum SP is generated using the detection signal SG of the first vibration detector 331. . The building blocks BJC were formed in a regular quadrangular prism shape, had a width of 44 mm, a height of 90 mm, and a length of 44 mm. In FIG. 12A, the building block BJC was placed so that one side of the bottom surface of the building block BJC was 44 mm and the other side was 44 mm. Further, the building blocks BJC were placed so that the side of 44 mm of the bottom face was parallel to the Y axis. Next, as illustrated in FIG. 12B, the block BJC is disposed horizontally in the center of the placement surface SF, and the amplitude spectrum SP is generated using the detection signal SG of the first vibration detector 331. In FIG. 12B, the building block BJC is placed so that one side of the bottom surface of the building block BJC is 44 mm and the other side is 90 mm. Moreover, the building blocks BJC were placed so that one side of the bottom surface was parallel to the Y axis.

  FIG. 13 is a diagram illustrating an example of the relationship between the posture PT of the building blocks BJC and the amplitude spectrum SP. The horizontal axis shown in FIG. 13 indicates the frequency F, and the vertical axis indicates the amplitude B. A graph G31 shown in FIG. 13 shows the amplitude spectrum SP when the building blocks BJC are arranged vertically, and a graph G32 shows the amplitude spectrum SP when the building blocks BJC are arranged horizontally.

  As shown in FIG. 13, when the frequency F is 33 kHz, the amplitude B of the amplitude spectrum SP when the blocks BJC are arranged vertically is larger than the amplitude B of the amplitude spectrum SP when the blocks BJC are arranged horizontally. Therefore, it has been found that the object detection apparatus 100 can detect the posture PT of one object BJ placed on the placement surface SF based on the amplitude spectrum SP.

<Experiment for detecting the content CP of one object BJ>
Next, an experiment in which the object detection device 100 detects the internal capacity CP of the object BJ placed on the placement surface SF will be described with reference to FIGS. 1 to 4 and FIGS. 14 to 15. FIG. 14 is a diagram illustrating an example of an experimental method for detecting the internal volume CP of the object BJ. In FIG. 14, the object BJ is a cup BJD. Fig.14 (a) is a figure which shows an example of the experiment method in case water does not enter in the cup BJD. FIG.14 (b) is a figure which shows an example of the experiment method in case water is contained in the cup BJD. The cup BJD includes a cup BJD1 and a cup BJD2. A cup BJD1 indicates an empty cup. Cup BJD2 indicates a cup containing 100 mL of water. As shown in FIGS. 14 (a) and 14 (b), at the center position in the width direction (Y-axis direction) of the placement surface SF and at the center position in the longitudinal direction (X-axis direction) of the placement surface SF. Each of the cup BJD1 and the cup BJD2 was placed.

  Specifically, as shown in FIG. 14A, the cup BJD1 was placed at the center of the placement surface SF, and the amplitude spectrum SP was generated using the detection signal SG of the first vibration detector 331. Next, the cup BJD2 was placed in the center of the placement surface SF, and the amplitude spectrum SP was generated using the detection signal SG of the first vibration detector 331.

  FIG. 15 is a diagram illustrating an example of the relationship between the internal capacity CP of the cup BJD and the amplitude spectrum SP. The horizontal axis of the graph shown in FIG. 15 indicates the frequency F, and the vertical axis indicates the amplitude B. A graph G41 illustrated in FIG. 15 illustrates the amplitude spectrum SP when the cup BJD1 is disposed, and a graph G42 illustrates the amplitude spectrum SP when the cup BJD2 is disposed.

  As shown in FIG. 15, in the frequency F of 28 kHz to 29 kHz, the amplitude B of the amplitude spectrum SP when the cup BJD2 is arranged is larger than the amplitude B of the amplitude spectrum SP when the cup BJD1 is arranged. Therefore, it has been found that the object detection apparatus 100 can detect the internal capacity CP of one object BJ (for example, a cup BJD) placed on the placement surface SF based on the amplitude spectrum SP.

<Detection of Position PS and Type TP of Two Objects BJ>
Next, referring to FIGS. 1 to 4 and FIGS. 16 to 20, when two objects BJ (object BJ1 and object BJ2) are placed on the placement surface SF, each of the objects BJ1 and BJ2 A process for detecting the position PS and the type TP will be described.

First, a 0th aspect, a 1st aspect, a 2nd aspect, and a 3rd aspect are prescribed | regulated as follows.
0th aspect: The aspect in which the object BJ is not mounted in mounting surface SF.
First aspect: An aspect in which only the first object BJ1 is placed on the placement surface SF.
Second aspect: An aspect in which only the second object BJ2 is placed on the placement surface SF.
Third aspect: an aspect in which the first object BJ1 and the second object BJ2 are placed on the placement surface SF.

When the object BJ is vibrated by an external force, the state is expressed by an equation of motion represented by the following equation (1).
M × U ″ + C × U ′ + K × U = P (1)
Here, the matrix M represents a mass matrix, the matrix C represents a damping matrix, and the matrix K represents a stiffness matrix. A vector U represents a displacement vector, and a vector P represents an external force vector.

In each of the 0th aspect to the 3rd aspect, since the mass matrix M, the damping matrix C, and the stiffness matrix K are constant, the equation of motion corresponding to each of the 0th aspect to the 3rd aspect is expressed by the following equation (2). -It represents with Formula (5).
M × U0 ″ + C × U0 ′ + K × U0 = I (2)
M × U1 ″ + C × U1 ′ + K × U1 = I + W1 (3)
M × U2 ″ + C × U2 ′ + K × U2 = I + W2 (4)
M × U3 ″ + C × U3 ′ + K × U3 = I + W1 + W2 (5)
Here, the displacement vector U0 indicates the displacement vector in the zeroth aspect, the displacement vector U1 indicates the displacement vector in the first aspect, the displacement vector U2 indicates the displacement vector in the second aspect, and the displacement vector U3 is The displacement vector in a 3rd aspect is shown. The external force I indicates the external force that the vibrator 32 applies to the placement surface SF. The external force W1 indicates the external force that the first object BJ1 applies to the placement surface SF. The external force W2 indicates the external force that the second object BJ2 applies to the placement surface SF. """Indicates a second derivative, and"'"indicates a first derivative.

Considering the linearity of the differentiation, the following formula (6) (= formula (4) + formula (3) −formula (2)) is derived from formulas (2) to (4).
M × (U2 + U1-U0) ”+ C × (U2 + U1-U0) ′
+ K × (U2 + U1-U0) = I + W1 + W2 (6)
By comparing the formula (5) and the formula (6), the following formula (7) is obtained.
U3 = U2 + U1-U0 (7)
From Expression (7), the displacement of the placement plate 31 when the first object BJ1 and the second object BJ2 are placed on the placement surface SF is obtained as follows. That is, the displacement of the placement plate 31 when the first object BJ1 is placed on the placement surface SF and the displacement of the placement plate 31 when the second object BJ2 is placed on the placement surface SF. Is obtained by subtracting the displacement of the placement plate 31 when the object BJ is not placed on the placement surface SF.

  Further, since it is presumed that the relationship similar to the equation (7) is also established for the amplitude spectrum SP, the third amplitude when the first object BJ1 and the second object BJ2 are placed on the placement surface SF. The spectrum SP3 is estimated as follows with reference to FIGS. 16 (a) and 16 (b).

  FIG. 16A is a diagram illustrating an example of a method for generating the first amplitude spectrum SP1, the fourth amplitude spectrum SP4, and the fifth amplitude spectrum SP5. As shown in FIG. 16A, the first amplitude spectrum SP1 indicates the amplitude spectrum SP when the object BJ is not placed on the placement surface SF. The fourth amplitude spectrum SP4 indicates the amplitude spectrum SP when only the first object BJ1 is placed on the placement surface SF. The fifth amplitude spectrum SP5 indicates the amplitude spectrum SP when only the second object BJ2 is placed on the placement surface SF.

  FIG. 16B is a diagram illustrating an example of a method for estimating the third amplitude spectrum SP3. As illustrated in FIG. 16B, the generation unit 114 illustrated in FIG. 4 generates the third amplitude spectrum SP3 based on the first amplitude spectrum SP1 and the second amplitude spectrum SP2. Specifically, the generation unit 114 generates a sum spectrum SPS indicating the sum of the amplitudes B for each frequency F of the fourth amplitude spectrum SP4 and the fifth amplitude spectrum SP5. And the production | generation part 114 produces | generates the difference of the amplitude B for every frequency F of the sum spectrum SPS and 1st amplitude spectrum SP1 as estimated spectrum SP3G of 3rd amplitude spectrum SP3. “Estimated spectrum SP3G” indicates a spectrum obtained by estimating the third amplitude spectrum SP3 based on the first amplitude spectrum SP1 and the second amplitude spectrum SP2.

  FIG. 16C is a diagram illustrating an example of a method for generating the third amplitude spectrum SP3. As shown in FIG. 16C, when the first object BJ1 and the second object BJ2 are placed on the placement surface SF, the vibration acquisition unit 112 shown in FIG. 4 detects the detection signal SG, The conversion unit 113 performs Fourier transform on the detection signal SG, and generates an actually measured spectrum SP3A of the third amplitude spectrum SP3. “Measured spectrum SP3A” indicates a third amplitude spectrum SP3 generated by performing Fourier transform on the detection signal SG. The third amplitude spectrum SP3 includes “estimated spectrum SP3G” and “measured spectrum SP3A”.

  The learning unit 115 learns the third amplitude spectrum SP3. Specifically, the learning unit 115 learns the estimated spectrum SP3G. For example, the learning unit 115 learns the estimated spectrum SP3G by SVM (Support Vector Machine).

  Based on the learning result of the learning unit 115, the object detection unit 116 detects the position PS1 and type TP1 of the object BJ1, and the position PS2 and type TP2 of the object BJ2.

Next, processing of the control unit 11 will be described with reference to FIGS. FIG. 17 is a flowchart illustrating an example of processing of the control unit 11 when detecting the position PS and the type TP of two objects BJ.
As shown in FIG. 17, in step S <b> 101, the vibration acquisition unit 112 acquires a detection signal SG from the vibration detector 33 when the object BJ is not placed on the placement surface SF.
Next, in step S103, the generation unit 114 generates a first amplitude spectrum SP1.
Next, in step S105, the vibration acquisition unit 112 acquires the detection signal SG when the first object BJ1 is placed on the placement surface SF from the vibration detector 33.
Next, in step S107, the generation unit 114 generates a fourth amplitude spectrum SP4.

Next, in step S109, the vibration acquisition unit 112 acquires the detection signal SG when the second object BJ2 is placed on the placement surface SF from the vibration detector 33.
Next, in step S111, the generation unit 114 generates a fifth amplitude spectrum SP5.
Next, in step S113, the generation unit 114 generates an estimated spectrum SP3G based on the first amplitude spectrum SP1, the fourth amplitude spectrum SP4, and the fifth amplitude spectrum SP5.
Next, in step S115, the control unit 11 executes a “learning process”. “Learning process” indicates a process of learning the estimated spectrum SP3G. The “learning process” will be described in detail later with reference to FIG.
Next, in step S117, the control unit 11 executes “detection evaluation process”, and the process ends. “Detection evaluation processing” indicates processing for detecting the position PS1 and type TP1 of the object BJ1 and the position PS2 and type TP2 of the object BJ2, and evaluating the detection result. The “detection evaluation process” will be described in detail later with reference to FIG.

FIG. 18 is a flowchart illustrating an example of the “learning process” of the control unit 11.
As shown in FIG. 18, first, in step S <b> 201, the vibration acquisition unit 112 receives a detection signal SG when the first object BJ <b> 1 and the second object BJ <b> 2 are placed on the placement surface SF as a vibration detector. 33.
Next, in step S203, the conversion unit 113 performs FFT on the detection signal SG to generate an actually measured spectrum SP3A.
Next, in step S205, the learning unit 115 learns the estimated spectrum SP3G, and the process returns to step S117 in FIG.

FIG. 19 is a flowchart illustrating an example of the “detection evaluation process” of the control unit 11.
As shown in FIG. 19, first, in Step S301, the first object BJ1 and the second object BJ2 are placed on the placement surface SF, and the vibration acquisition unit 112 acquires the detection signal SG from the vibration detector 33. To do.
Next, in step S303, the conversion unit 113 performs FFT on the detection signal SG to generate an actually measured spectrum SP3A.
Next, in step S305, the object detection unit 116, based on the measured spectrum SP3A and the learning result of the estimated spectrum SP3G, the position PS1 and type TP1 of the first object BJ1, the position PS2 and type TP2 of the object BJ2, and Is detected.
Next, in step S307, the detection result of the object detection unit 116 is evaluated, and the process ends.

<Experiment for detecting position PS and type TP of two objects BJ>
First, the experimental method will be described. The position where the object BJ is placed is any one of the three positions. The three positions are a position 30 mm from the end in the X-axis direction on the center line CL of the mounting surface SF, a position 160 mm from the end in the X-axis direction, and a position 290 mm from the end in the X-axis direction. It was. As the type TP of the object BJ, the object BJ placed on the desk and the object BJ stored in the refrigerator were selected. Specifically, "Mug", "Note" and "Pen" are selected as the object BJ placed on the desk, and "Apple" and "Onion" are selected as the object BJ stored in the refrigerator. And “Mikan” were selected. In the following, the case where three types of objects BJ of “mug cup”, “notebook” and “pen” are used will be referred to as “desk pattern”, and three types of objects BJ of “apple”, “onion” and “mandarin orange” will be described. The case of using is described as “refrigerator pattern”.

  For each of the “desk pattern” and the “refrigerator pattern”, the experiment was performed 12 times for each of the nine cases in which one type of the object BJ out of the three types of objects BJ was arranged at one position, and 108 pieces ( = 9 × 12) amplitude spectrum SP was generated. One position indicates any one of the three positions. In addition, each of the two types of objects BJ out of the three types of objects BJ is subjected to 12 experiments for 18 cases where each of the two types of objects BJ is arranged at one position, and 216 (= 18 × 12) amplitude spectra. SP was generated.

  Then, the estimated spectrum SP3G of the third amplitude spectrum SP3 was learned for each of 18 cases in which two of the three types of objects BJ were arranged at two positions. Then, based on the measured spectrum SP3A of the third amplitude spectrum SP3 and the learning result of the estimated spectrum SP3G of the third amplitude spectrum SP3, the position PS1, the position PS2, the type TP1, and the type TP2 are detected, and the detection result is evaluated. . Specifically, the identification rates of the position PS1, the position PS2, the type TP1, and the type TP2 were obtained by cross-validation.

  Next, the result of an experiment for detecting the position PS and the type TP of the two objects BJ will be described with reference to FIG. As shown in FIG. 20, the identification rate in the “desk pattern” was as follows. The identification rate when the amplitude spectrum SP was generated using the detection signal SG of the first vibration detector 331 was 86.1%. The identification rate when the amplitude spectrum SP was generated using the detection signal SG of the second vibration detector 332 was 95.4%. The discrimination rate when the amplitude spectrum SP was generated using the detection signal SG of the first vibration detector 331 and the detection signal SG of the second vibration detector 332 was 96.8%.

  The identification rate in the “refrigerator pattern” was as follows. The discrimination rate when the amplitude spectrum SP was generated using the detection signal SG of the first vibration detector 331 was 95.8%. The discrimination rate when the amplitude spectrum SP was generated using the detection signal SG of the second vibration detector 332 was 97.2%. The discrimination rate when the amplitude spectrum SP was generated using the detection signal SG of the first vibration detector 331 and the detection signal SG of the second vibration detector 332 was 98.6%.

  Thus, compared with the case where only the detection signal SG of the first vibration detector 331 is used and the case where only the detection signal SG of the second vibration detector 332 is used, the detection signal SG of the first vibration detector 331 and By using the detection signal SG of the second vibration detector 332, a high identification rate was obtained.

  As described above, as described with reference to FIGS. 1 to 4 and FIGS. 16 to 20, in the embodiment of the present invention, a plurality of pieces placed on the placement surface SF based on the learning result of the amplitude spectrum SP. The state ST (here, position PS and type TP) of the object BJ (two here) is detected. Therefore, the state ST of the plurality of objects BJ placed on the placement surface SF can be detected more accurately.

  Further, the learning unit 115 calculates the amplitude spectrum SP when the first object BJ1 and the second object BJ2 are placed on the placement surface SF based on the first amplitude spectrum SP1 and the second amplitude spectrum SP2. To learn. Then, based on the learning result of the learning unit 115, the object detection unit 116 detects the state ST (here, position PS and type TP) of the first object BJ1 and the second object BJ2. Therefore, each state ST of 1st object BJ1 and 2nd object BJ2 is detectable.

  Further, based on the first amplitude spectrum SP1 and the second amplitude spectrum SP2, a third amplitude spectrum SP3 (specifically, an estimated spectrum SP3G) is generated. The third amplitude spectrum SP3 indicates the amplitude spectrum SP when the first object BJ1 and the second object BJ2 are placed on the placement surface SF. The learning unit 115 learns the third amplitude spectrum SP3, and the object detection unit 116, based on the learning result, each state ST (here, position PS and type TP) of the first object BJ1 and the second object BJ2. Is detected. Therefore, each state ST of the first object BJ1 and the second object BJ2 can be accurately detected with a simple configuration.

  Further, a fourth amplitude spectrum SP4 indicating the amplitude spectrum SP when the first object BJ1 is placed on the placement surface SF, and a case where the second object BJ2 is placed on the placement surface SF. A fifth amplitude spectrum SP5 indicating the amplitude spectrum SP is generated. And the sum spectrum SPS which shows the sum of the amplitude B for every frequency F of 4th amplitude spectrum SP4 and 5th amplitude spectrum SP5 is produced | generated, and the amplitude B for every frequency F of sum spectrum SPS and 1st amplitude spectrum SP1 is produced | generated. The difference is generated as a third amplitude spectrum SP3 (specifically, an estimated spectrum SP3G). Further, the third amplitude spectrum SP3 is learned, and the position PS1 and type TP1 of the first object BJ1 and the position PS2 and type TP2 of the second object BJ2 are detected based on the learning result. Therefore, the position PS1 and type TP1 of the first object BJ1 and the position PS2 and type TP2 of the second object BJ2 can be accurately detected with a simple configuration.

  In the embodiment of the present invention, the position PS1 and type TP1 of the first object BJ1 and the position PS2 and type TP2 of the second object BJ2 are detected, but the present invention is not limited to this. The state ST of the first object BJ1 and the state ST of the second object BJ2 may be detected.

<Detection of Position PS and Type TP of Three Objects BJ>
Next, referring to FIG. 1 to FIG. 4 and FIG. 21 to FIG. 22, when three objects BJ (object BJ1, object BJ2, and object BJ3) are placed on placement surface SF, object BJ1 and object Processing for detecting the position PS and the type TP of each of the BJ2 and the object BJ3 will be described.

With reference to FIGS. 1 to 4 and FIGS. 16 to 20, Expression (7) is derived for the case where two objects BJ (object BJ1 and object BJ2) are placed on the placement surface SF. Thus, the following (8) is obtained.
U4 = U3 + U2 + U1-2 × U0 (8)
However, the displacement vector U4 indicates a displacement vector in a state where the first object BJ1, the second object BJ2, and the third object BJ3 are placed on the placement surface SF, and the displacement vector U3 is the first on the placement surface SF. This is different from Expression (7) in that it shows a displacement vector in a state where the two objects BJ2 are placed.

  From Expression (8), the displacement when the first object BJ1, the second object BJ2, and the third object BJ3 are placed on the placement surface SF is obtained as follows. That is, the displacement when the first object BJ1 is placed on the placement surface SF, the displacement when the second object BJ2 is placed on the placement surface SF, and the third object on the placement surface SF. The sum with the displacement when BJ3 is placed is obtained. It is obtained by subtracting twice the displacement when the object BJ is not placed on the placement surface SF from the sum of the three displacements.

  Moreover, since it is presumed that the same relationship as in the equation (8) is established for the amplitude spectrum SP, the first object BJ1, the second object BJ2, and the third object BJ3 are placed on the placement surface SF. The third amplitude spectrum SP3 in this case is estimated as follows with reference to FIG.

  FIG. 21A is a diagram illustrating an example of a method for generating the first amplitude spectrum SP1, the fourth amplitude spectrum SP4, the fifth amplitude spectrum SP5, and the sixth amplitude spectrum SP6. As shown in FIG. 21A, the first amplitude spectrum SP1 indicates the amplitude spectrum SP when the object BJ is not placed on the placement surface SF. The fourth amplitude spectrum SP4 indicates the amplitude spectrum SP when only the first object BJ1 is placed on the placement surface SF. The fifth amplitude spectrum SP5 indicates the amplitude spectrum SP when only the second object BJ2 is placed on the placement surface SF. The sixth amplitude spectrum SP6 indicates the amplitude spectrum SP when only the third object BJ3 is placed on the placement surface SF.

  FIG. 21B is a diagram illustrating an example of a method for estimating the third amplitude spectrum SP3. As illustrated in FIG. 21B, the generation unit 114 illustrated in FIG. 4 generates the third amplitude spectrum SP3 based on the first amplitude spectrum SP1 and the second amplitude spectrum SP2. Specifically, the generation unit 114 generates a sum spectrum SPS indicating the sum of the amplitudes B for each frequency F of the fourth amplitude spectrum SP4, the fifth amplitude spectrum SP5, and the sixth amplitude spectrum SP6. Then, the generation unit 114 generates, as the estimated spectrum SP3G of the third amplitude spectrum SP3, the difference in the amplitude B for each frequency F between the sum spectrum SPS and the amplitude spectrum SP having the amplitude B twice the first amplitude spectrum SP1.

  FIG. 21C is a diagram illustrating an example of a method for generating the third amplitude spectrum SP3. As shown in FIG. 21C, when the first object BJ1, the second object BJ2, and the third object BJ3 are placed on the placement surface SF, the vibration acquisition unit shown in FIG. 112 detects, and the conversion part 113 performs Fourier transformation with respect to the detection signal SG, and produces | generates measured spectrum SP3A of 3rd amplitude spectrum SP3.

  The learning unit 115 learns the third amplitude spectrum SP3. Specifically, the learning unit 115 learns the estimated spectrum SP3G. For example, the learning unit 115 learns the estimated spectrum SP3G by SVM.

  The object detection unit 116 detects the position PS1 and type TP1 of the object BJ1, the position PS2 and type TP2 of the object BJ2, and the position PS3 and type TP3 of the object BJ3 based on the learning result of the learning unit 115.

<Experiment for detecting position PS and type TP of three objects BJ>
First, the experimental method will be described. The position where the object BJ is placed is any one of the three positions. The three positions are 30 mm from the end in the X-axis direction on the center line CL of the mounting surface SF, 160 mm from the end in the X-axis direction, and 290 mm from the end in the X-axis direction. Was in position. As the type TP of the object BJ, the object BJ placed on the desk and the object BJ stored in the refrigerator were selected. Specifically, "Mug", "Note" and "Pen" are selected as the object BJ placed on the desk, and "Apple" and "Onion" are selected as the object BJ stored in the refrigerator. And “Mikan” were selected. In the following, the case where three types of objects BJ of “mug cup”, “notebook” and “pen” are used will be referred to as “desk pattern”, and three types of objects BJ of “apple”, “onion” and “mandarin orange” will be described. The case of using is described as “refrigerator pattern”.

  For each of the “desk pattern” and the “refrigerator pattern”, the experiment was performed 12 times for each of the nine cases in which one type of the object BJ out of the three types of objects BJ was arranged at one position, and 108 pieces ( = 9 × 12) amplitude spectrum SP was generated. One position indicates any one of the three positions. In addition, each of the three types of objects BJ was arranged in one position, and the experiment was performed 12 times for each of the six cases to generate 72 (= 6 × 12) amplitude spectra SP.

  Then, the estimated spectrum SP3G of the third amplitude spectrum SP3 was learned for each of the six cases in which each of the three types of objects BJ is arranged at one position. Then, the position PS1, the position PS2, the position PS3, the type TP1, the type TP2, and the type TP3 are detected based on the actual measurement spectrum SP3A of the third amplitude spectrum SP3 and the learning result of the estimated spectrum SP3G of the third amplitude spectrum SP3. The detection results were evaluated. Specifically, the identification rate was obtained by cross-validation.

  Next, the result of an experiment for detecting the position PS and the type TP of the two objects BJ will be described with reference to FIG. As shown in FIG. 20, the identification rate in the “desk pattern” was as follows. The discrimination rate when the amplitude spectrum SP was generated using the detection signal SG of the first vibration detector 331 was 97.2%. The identification rate when the amplitude spectrum SP was generated using the detection signal SG of the second vibration detector 332 was 98.6%. The discrimination rate when the amplitude spectrum SP was generated using the detection signal SG of the first vibration detector 331 and the detection signal SG of the second vibration detector 332 was 100%.

  The identification rate in the “refrigerator pattern” was as follows. The identification rate when the amplitude spectrum SP was generated using the detection signal SG of the first vibration detector 331 was 91.7%. The identification rate when the amplitude spectrum SP was generated using the detection signal SG of the second vibration detector 332 was 98.6%. The discrimination rate when the amplitude spectrum SP was generated using the detection signal SG of the first vibration detector 331 and the detection signal SG of the second vibration detector 332 was 98.6%.

  As described above with reference to FIGS. 1 to 4 and FIGS. 21 to 22, in the embodiment of the present invention, the fourth amplitude spectrum SP4, the fifth amplitude spectrum SP5, and the sixth amplitude spectrum SP6 are obtained. Generate. The fourth amplitude spectrum SP4 indicates the amplitude spectrum SP when the first object BJ1 is placed on the placement surface SF. The fifth amplitude spectrum SP5 indicates the amplitude spectrum SP when the second object BJ2 is placed on the placement surface SF. The sixth amplitude spectrum SP6 indicates the amplitude spectrum SP when the third object BJ3 is placed on the placement surface SF. Then, a sum spectrum SPS is generated. The sum spectrum SPS indicates the sum of the amplitudes B for each frequency F of the fourth amplitude spectrum SP4, the fifth amplitude spectrum SP5, and the sixth amplitude spectrum SP6. Next, a difference spectrum SPD is generated. The difference spectrum SPD indicates the difference in the amplitude B for each frequency F between the sum spectrum SPS and the first amplitude spectrum SP1. Further, a difference in amplitude B for each frequency F between the difference spectrum SPD and the first amplitude spectrum SP1 is generated as a third amplitude spectrum SP3 (specifically, an estimated spectrum SP3G). Further, the third amplitude spectrum SP3 is learned, and based on the learning result, the state ST1 of the first object BJ1 (here, position PS1 and type TP1) and the state ST2 of the second object BJ2 (here, position PS2 and The type TP2) and the state ST3 of the third object BJ3 (here, the position PS3 and the type TP3) are detected. Therefore, the state ST1 of the first object BJ1, the state ST2 of the second object BJ2, and the state ST3 of the third object BJ3 can be accurately detected with a simple configuration.

  In the embodiment of the present invention, the position PS1 and type TP1 of the first object BJ1, the position PS2 and type TP2 of the second object BJ2, and the position PS3 and type TP3 of the object BJ3 are detected. It is not limited to this. The state ST of the first object BJ1, the state ST of the second object BJ2, and the state ST3 of the third object BJ3 may be detected.

  The embodiments of the present invention have been described above with reference to the drawings. However, the present invention is not limited to the above-described embodiment, and can be implemented in various modes without departing from the gist thereof (for example, (1) to (5) shown below). In order to facilitate understanding, the drawings schematically show each component as a main component, and the thickness, length, number, and the like of each component shown in the drawings are different from the actual for convenience of drawing. . Moreover, the shape, dimension, etc. of each component shown by said embodiment are an example, Comprising: It does not specifically limit, A various change is possible in the range which does not deviate substantially from the structure of this invention.

  (1) In the present embodiment, as described with reference to FIG. 1, the mounting surface SF is the upper surface of the mounting plate 31 and the mounting plate 31 is a rectangular acrylic plate. It is not limited to this. The placement plate 31 may have a placement surface SF. For example, the mounting plate 31 may be circular or the mounting plate 31 may be made of metal.

  (2) In the present embodiment, as described with reference to FIG. 1, the vibrator 32 is disposed on the lower surface of the mounting plate 31, but the present invention is not limited to this. The vibrator 32 may be disposed on the mounting plate 31. For example, the vibrator 32 may be disposed on the placement surface SF of the placement plate 31.

  (3) In the present embodiment, as described with reference to FIG. 1, the vibration detector 33 is disposed on the lower surface of the mounting plate 31, but the present invention is not limited to this. The vibration detector 33 may be disposed on the mounting plate 31. For example, the vibration detector 33 may be disposed on the placement surface SF of the placement plate 31.

  (4) In the present embodiment, as described with reference to FIG. 1, the control unit 11 is arranged in the personal computer 1, but the object detection device 100 may include the control unit 11. The control part 11 may be arrange | positioned at a tablet terminal device, and the control part 11 may be arrange | positioned at a server apparatus.

  (5) In this embodiment, as described with reference to FIG. 9, the object BJ is the object BJ placed on the desk or the object BJ stored in the refrigerator, but the present invention is not limited to this. The object BJ may be placed on the placement surface SF. For example, the object BJ may be an object BJ displayed on a shelf.

  The present invention can be used for an object detection device that detects the state of an object placed on a placement surface.

DESCRIPTION OF SYMBOLS 100 Object detection apparatus 1 Personal computer 11 Control part 11A Processor 11B Storage part 111 Vibration provision part 112 Vibration detection part 113 Conversion part 114 Generation part 115 Learning part 116 Object detection part 12 Display 2 Audio interface 21 Signal generator 3 Detection apparatus main body 31 Mounting plate 32 Exciter 33 Vibration detector 331 First vibration detector 332 Second vibration detector 34, 341, 342, 343, 344 Support BJ object BJ1 First object BJ2 Second object BJ3 Third object SF Mount Placement surface ST state PS position TP type PT posture CT content SP amplitude spectrum SP1 first amplitude spectrum SP2 second amplitude spectrum SP3 third amplitude spectrum SP4 fourth amplitude spectrum SP5 fifth amplitude spectrum SP6 sixth amplitude spectrum SPS sum spectrum SPD difference spectrum SG acoustic signal

Claims (12)

A vibration exciter that is arranged at a first position of the mounting plate and applies vibration to the mounting plate;
A vibration detector disposed at a second position different from the first position of the mounting plate, and detecting vibration of the mounting plate;
An object detection apparatus comprising: an object detection unit configured to detect a state of an object placed on the placement surface of the placement plate based on a detection result of the vibration detector.   The object detection apparatus according to claim 1, wherein the state includes at least one of position, type, posture, and content.   The object detection device according to claim 1, wherein the vibration exciter imparts vibration, the frequency of which varies with time, to the mounting plate. The vibrator has a first vibrating body that outputs the vibration,
The first vibrating body applies the vibration to the mounting plate,
The vibration detector includes a second vibrating body that detects the vibration,
4. The object detection device according to claim 1, wherein the second vibrating body detects the vibration of the mounting plate. 5. The vibrator has a first piezoelectric element that outputs acoustic vibrations,
The first piezoelectric element imparts the acoustic vibration to the mounting plate,
The vibration detector includes a second piezoelectric element that detects the acoustic vibration,
4. The object detection device according to claim 1, wherein the second piezoelectric element detects the acoustic vibration of the mounting plate. 5. The vibration detector is
A first vibration detector disposed at the second position of the mounting plate;
The object detection according to any one of claims 1 to 5, further comprising a second vibration detector disposed at a third position different from the first position and the second position of the mounting plate. apparatus. A Fourier transform is performed on the detection signal of the vibration detector, and further includes a conversion unit that generates an amplitude spectrum,
The object detection device according to claim 1, wherein the object detection unit detects the state of the object based on the amplitude spectrum. A learning unit for learning the amplitude spectrum;
The object detection device according to claim 7, wherein the object detection unit detects the state of the object based on a learning result of the learning unit. The conversion unit includes a first amplitude spectrum indicating the amplitude spectrum when the object is not placed on the placement surface, and each of the plurality of objects placed on the placement surface. A second amplitude spectrum indicating the amplitude spectrum of the case,
The learning unit learns based on the first amplitude spectrum and the second amplitude spectrum,
The object detection device according to claim 8, wherein the object detection unit detects the state of each of the plurality of objects based on a learning result of the learning unit. A generating unit configured to generate a third amplitude spectrum indicating the amplitude spectrum when the plurality of objects are placed on the placement surface based on the first amplitude spectrum and the second amplitude spectrum; Prepared,
The learning unit learns the third amplitude spectrum;
The object detection device according to claim 9, wherein the object detection unit detects the state of each of the plurality of objects based on a learning result of the learning unit. The conversion unit includes a fourth amplitude spectrum that indicates the second amplitude spectrum when the first object is placed on the placement surface, and a second amplitude that is different from the first object on the placement surface. Generating a fifth amplitude spectrum indicating the second amplitude spectrum when an object is placed;
The generation unit generates a sum spectrum indicating a sum of amplitudes for each frequency of the fourth amplitude spectrum and the fifth amplitude spectrum,
The generation unit generates a difference in amplitude for each frequency between the sum spectrum and the first amplitude spectrum as the third amplitude spectrum,
The learning unit learns the third amplitude spectrum;
The object detection device according to claim 10, wherein the object detection unit detects the state of the first object and the state of the second object based on a learning result of the learning unit. The conversion unit includes a fourth amplitude spectrum that indicates the second amplitude spectrum when the first object is placed on the placement surface, and a second amplitude that is different from the first object on the placement surface. A fifth amplitude spectrum indicating the second amplitude spectrum when an object is placed, and a third object different from the first object and the second object is placed on the placement surface. And a sixth amplitude spectrum indicating the second amplitude spectrum of
The generation unit generates a sum spectrum indicating a sum of amplitudes for each frequency of the fourth amplitude spectrum, the fifth amplitude spectrum, and the sixth amplitude spectrum,
The generation unit generates a difference spectrum indicating a difference in amplitude for each frequency between the sum spectrum and the first amplitude spectrum,
The generation unit generates a difference in amplitude for each frequency between the difference spectrum and the first amplitude spectrum as the third amplitude spectrum,
The learning unit learns the third amplitude spectrum;
The said object detection part detects the said state of the said 1st object, the said state of the said 2nd object, and the said state of the said 3rd object based on the learning result of the said learning part. Object detection device.

Images