Body shape estimation device and body shape estimation method, seat belt reminder system and airbag control system

Abstract

A body shape estimation device is provided, which can estimate body shape. The body shape estimation device comprises: a sensor (10) having a transmission antenna for transmitting a transmission wave and a reception antenna for receiving the transmission wave reflected by at least one object in a vehicle compartment as a reception wave; a frequency analysis unit (49) for obtaining position information of a reflection point where the transmission wave is reflected using the reception wave; and a body shape estimation unit (61) for estimating the body shape of a non-stationary object existing in the vehicle compartment using the position information.

Description

Technical Field

[0001] The present invention relates to a body shape estimation device and a body shape estimation method, and a seat belt reminder system and an air bag control system using the body shape estimation device. Background Art

[0002] Patent document 1 discloses the following occupant detection device: a Doppler sensor is provided inside a seat cushion of a vehicle, the Doppler sensor emits an electromagnetic beam toward a reflector of the seat cushion to detect a displacement speed of the seat cushion that changes due to a load, and the occupant detection device determines the presence or absence of an occupant based on the displacement speed of the seat cushion detected by the Doppler sensor.

[0003] Prior art literature

[0004] Patent Literature

[0005] Patent Document 1: Japanese Patent Application Laid-Open No. 08-127264 Summary of the invention

[0006] Problems to be solved by the invention

[0007] According to the technology of Patent Document 1, there is a problem that the presence or absence of an occupant can be detected, but the body shape of the occupant cannot be estimated.

[0008] The present invention has been made to solve the above-mentioned problems, and an object of one aspect of an embodiment is to provide a body shape estimating device capable of estimating the body shape of an occupant.

[0009] Means for solving problems

[0010] One aspect of the body shape estimation device of the embodiment comprises: a sensor having a transmitting antenna for transmitting a transmission wave and a receiving antenna for receiving the transmission wave after being reflected by at least one object in a vehicle compartment as a reception wave; a frequency analysis unit, which uses the reception wave to obtain position information of a reflection point where the transmission wave is reflected; and a body shape estimation unit, which uses the position information to estimate the shape of a non-stationary object present in the vehicle compartment.

[0011] Effects of the Invention

[0012] According to the body shape estimation device of the embodiment, it is possible to estimate the body shape of a non-stationary object. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 This is a block diagram schematically showing the structure of the body shape estimation device according to the first embodiment.

[0014] Figure 2 This is a block diagram showing the structure of the frequency analysis unit according to the first embodiment.

[0015] Figure 3 This is a block diagram showing the structure of a body shape estimation unit according to the first embodiment.

[0016] Figure 4A This is a block diagram showing a hardware configuration example of a signal processor according to Embodiment 1.

[0017] Figure 4B This is a block diagram showing another hardware configuration example of the signal processor according to the first embodiment.

[0018] Figure 5 This is a flowchart showing the operation of the body shape estimation device according to the first embodiment.

[0019] Figure 6 This is a detailed flowchart of the body shape estimation process in the first embodiment.

[0020] Fig. 7A It is a diagram showing an example of use of the body shape estimation device.

[0021] Figure 7B This is a diagram showing an example of an imaged occupied grid map.

[0022] Figure 8 This is a diagram showing an example of an imaged occupied grid map.

[0023] Fig. 9 : is a block diagram showing the structure of a body shape estimation unit according to Modification 1.

[0024] Fig.10 This is a detailed flowchart of the body shape estimation processing of Modification Example 1.

[0025] Fig.11 : is a block diagram showing the structure of a body shape estimation unit according to Modification Example 2.

[0026] Fig.12 This is a detailed flowchart of the body shape estimation processing of Modification Example 2.

[0027] Fig.13A 1 is a diagram showing simulation results of the operation of the body shape estimation device according to Modification 1. FIG.

[0028] Fig. 13B FIG. 5 is a diagram showing simulation results of the operation of the body shape estimation device according to Modification Example 2. FIG.

[0029] Fig.14 is a block diagram showing the structure of a seat belt reminder system and an air bag control system.

[0030] Fig.15 : is a block diagram showing the structure of a seat belt wearing determination device.

[0031] Fig.16is a block diagram showing the structure of an air bag control device. DETAILED DESCRIPTION

[0032] Hereinafter, various embodiments of the present invention will be described in detail with reference to the accompanying drawings. In addition, the components marked with the same reference numerals throughout the drawings have the same or similar structures or functions.

[0033] Implementation Method 1

[0034] <Structure>

[0035] Figure 1 FIG. 1 is a block diagram schematically showing the structure of a body shape estimation device 1 according to Embodiment 1 of the present invention. Figure 1 As shown, the body shape estimation device 1 is configured to include a sensor 10 and a signal processor 41 that operates in cooperation with the sensor 10. The body shape estimation device 1 is a device for estimating the body shape of a non-stationary object including a moving body such as an occupant of a vehicle, and is configured to transmit a transmission wave Tw to the occupant in the vehicle and receive a reflected wave Rw from the occupant. Here, "vehicle" refers to a means of transportation with wheels, and examples of vehicles include cars, buses, and trucks. In addition, "non-stationary object" refers to an object other than a stationary object, and examples of non-stationary objects include a moving body that can move by itself and an object that cannot move by itself but moves with the help of an external force. In the following description, for the sake of simplicity, the embodiments of the present invention are described using the terms moving body, occupant, or person as examples of a non-stationary object. In addition, "body shape" refers to the three-dimensional dimensions of a moving body such as a person.

[0036] The sensor 10 includes a transmitting circuit 21, a transmitting antenna 20, and a plurality of receiving antennas 300 to 30 Q-1 and respectively with a plurality of receiving antennas 300 to 30 Q-1 The plurality of receivers 310 to 31 are arranged correspondingly Q-1 .

[0037] The transmission circuit 21 includes a voltage generator 22, a voltage controlled oscillator 23, a distributor 24 and an amplifier 25, and generates a frequency modulated wave in a high frequency band such as a millimeter wave band (approximately 30 GHz to 300 GHz). The voltage generator 22 generates a modulation voltage according to a control signal TC supplied from a signal processor 41, and supplies the generated modulation voltage to the voltage controlled oscillator 23. The voltage controlled oscillator 23 repeatedly outputs a frequency modulated wave signal according to a predetermined frequency modulation method, and the frequency modulated wave signal has a modulation frequency that rises or falls over time according to the supplied modulation voltage. As a predetermined frequency modulation method, for example, a frequency modulated continuous wave (FMCW: Frequency Modulated Continuous Wave) method or a fast linear frequency modulation (FCM: Fast-Chirp Modulation) method can be used. According to the FMCW or FCM method, the frequency of the frequency modulated wave signal, that is, the transmission frequency, is scanned in a manner that continuously rises or falls over time within a certain frequency band. The distributor 24 distributes the frequency modulated wave signal input from the voltage controlled oscillator 23 into a transmission wave signal and a local signal. The distributor 24 supplies the transmission wave signal to the amplifier 25 and supplies the local signal to the receivers 310 to 31 Q-1 The transmission wave signal is amplified by the amplifier 25. The transmission antenna 20 transmits the transmission wave Tw (chirp) based on the output signal of the amplifier 25 to the observation space (ie, the interior of the vehicle cabin).

[0038] Receiving antenna 300~30 Q-1 It is arranged in a straight line, a plane or a curved surface to receive the reflected wave Rw generated by the reflection of the transmission wave Tw in the vehicle as a received wave. Q represents the receiving antenna 300 to 30 Q-1 An integer greater than 3 of the number of receiving channels.

[0039] Receiver 310~31 Q-1 Is with receiving antenna 300~30 Q-1 The qth receiver 31 is configured accordingly. q With low noise amplifier (LNA: Low Noise Amplifier) ​​32 q 、Mixer 33 q ,IF amplifier 34 q 、Filter 35 q and A / D converter (ADC) 36 q . Wherein, q is any integer in the range of 0 to Q-1.

[0040] Low Noise Amplifier 32 q For receiving antenna 30 q The output signal is amplified and the amplified signal is output to the mixer 33q . Mixer 33 q The amplified signal is mixed with the local signal supplied from the distributor 24 to generate a beat signal in the intermediate frequency band. q To slave mixer 33 q The input beat signal is amplified and the amplified beat signal is output to the filter 35 q Filter 35 q The unnecessary frequency components in the beat signal are suppressed and amplified, and the analog reception signal is output. ADC 36 q Convert the analog received signal into a digital received signal at a predetermined sampling rate m (k) (n, h, q), and outputs the digital reception signal to the signal processor 41. Here, k is an integer representing the frame number, n is an integer representing the sample number, and h is an integer representing the chirp number.

[0041] The signal processor 41 includes a data storage unit 46 , a signal processing unit 47 , and a control unit 45 . The control unit 45 controls the operations of the transmission circuit 21 , the data storage unit 46 , and the signal processing unit 47 .

[0042] The data storage unit 46 temporarily stores the data from the receivers 310 to 31 Q-1 As the data storage unit 46, a RAM (Random Access Memory) having a fast response performance can be used.

[0043] The control unit 45 supplies the control signal TC for generating the modulation voltage to the transmission circuit 21. The control unit 45 can also perform read control and write control of the signal with respect to the data storage unit 46.

[0044] The signal processing unit 47 performs digital signal processing on the received signal read from the data storage unit 46 to identify the target object in the observation space. Figure 1 As shown, the signal processing unit 47 includes a moving body extraction unit 48, a frequency analysis unit 49, a body shape estimation unit 61, and a map storage unit 63. In addition, the moving body extraction unit 48 is an optional component, and the signal processing unit 47 may not include the moving body extraction unit 48. The following describes a method in which the moving body extraction unit 48 is configured in the signal processing unit 47.

[0045] The moving object extraction unit 48 is a processing unit for extracting a moving object (Moving Target) by reading the sampled beat signal from the data storage unit 46. Specifically, by applying an MTI filter to the read signal, a relatively low-frequency signal component generated by a stationary object is removed, and a relatively high-frequency signal component from the moving object is extracted. The moving object extraction unit 48 supplies the extracted signal to the frequency analysis unit 49.

[0046] The frequency analysis unit 49 performs frequency analysis on the signal received from the moving body extraction unit 48, and supplies the result of the frequency analysis to the body shape estimation unit 61. In addition, in the case of an embodiment in which the moving body extraction unit 48 is not provided, the frequency analysis unit 49 reads the sampled beat signal from the data storage unit 46 and performs frequency analysis. Figure 2 , details of the processing contents of the frequency analysis unit 49 are described. Figure 2 1 is a block diagram showing the structure of the frequency analysis unit of Embodiment 1. Figure 2 As shown, the frequency analysis unit 49 includes a position information calculation unit 491 and a coordinate conversion unit 492. The position information calculation unit 491 includes a range Doppler processing unit 4911, an integration processing unit 4912, a peak extraction processing unit 4913, and a peak angle measurement processing unit 4914.

[0047] The range Doppler processing unit 4911 is a processing unit that generates a range Doppler map by performing range FFT (Fast Fourier Transform) and Doppler FFT on the signal received from the moving body extraction unit 48. In a sensor based on the FMCW method or the FCM method, one linear frequency modulation received by reflection from an object is received after a delay of a time proportional to the distance from the sensor to the object. In addition, different linear frequency modulations received by reflection from the object include the Doppler frequency shift between the sensor and the object. Therefore, first, in order to extract the range component, the range Doppler processing unit 4911 performs range FFT on the signal received from the moving body extraction unit 48 in units of linear frequency modulation (i.e., in terms of the number of linear frequency modulations), and performs a transformation based on equation (1) on the signal after the range FFT.

[0048]

[0049] Here, c is the speed of light, Tc is the duration of chirp, Δf is the difference between the transmission frequency and the reception frequency, and B is the bandwidth of chirp. By performing the transformation based on equation (1), data in the distance direction is generated for each chirp.

[0050] Next, in order to extract the relative velocity component, the range Doppler processing unit 4911 arranges the data in the distance direction obtained according to each linear frequency modulation in a time series. That is, the range Doppler processing unit 4911 arranges the data in the distance direction obtained according to each linear frequency modulation in the order of the linear frequency modulation number. Then, the range Doppler processing unit 4911 performs Doppler FFT on the data in the distance direction arranged in a time series. More specifically, the range Doppler processing unit 4911 performs Doppler FFT at each distance (that is, by the number of distance bins) for data of the same distance with different linear frequency modulation numbers. Thus, a data set having a correlation with the distance R and the Doppler frequency f is generated. d The generated range Doppler map is supplied to the integration processing unit 4912.

[0051] The integration processing unit 4912 performs incoherent addition processing on all the range Doppler maps received from the range Doppler processing unit 4911 to improve the SNR (Signal To Noise Ration). The range Doppler map obtained by the incoherent addition processing is supplied to the peak extraction processing unit 4913 .

[0052] The peak extraction processing unit 4913 performs two-dimensional peak extraction processing on the range Doppler map received from the integration processing unit 4912 to detect the target signal. Examples of two-dimensional peak extraction processing include two-dimensional CFAR (Constant False Alarm Rate). The two-dimensional peak extraction processing is performed on more than three antennas configured. The peak extraction processing unit 4913 supplies the extracted two-dimensional peak data to the peak angle measurement processing unit 4914.

[0053] The peak angle measurement processing unit 4914 performs angle FFT for different antennas on the two-dimensional peak data extracted for each antenna to obtain angle data. Even two-dimensional peak data at the same distance and the same Doppler measure on the range Doppler map has a phase difference ΔΦ corresponding to the distance d between the antennas. Therefore, the peak angle measurement processing unit 4914 performs angle FFT for different antennas to obtain the azimuth angle θ.

[0054]

[0055] Similarly, the peak angle measurement processing unit 4914 performs angle FFT on different antennas to obtain the elevation angle The peak angle measurement processing unit 4914 converts the polar coordinate data (R, θ, ) is supplied to the coordinate conversion unit 492.

[0056] The coordinate conversion unit 492 converts the position information calculated by the position information calculation unit 491 into the position information in the vehicle fixed coordinate system according to the following equations (3) and (4). Equation (3) is a conversion equation for converting the polar coordinate system for positioning into the rectangular coordinate system x'y'z' of the sensor. Equation (4) is a conversion equation for converting the rectangular coordinate system x'y'z' of the sensor into the fixed coordinate system xyz of the vehicle. It is assumed that the rectangular coordinate system x'y'z' of the sensor is rotated by an angle δ with the x-axis (=x' axis) as the center relative to the fixed coordinate system xyz of the vehicle. The converted position information (X, Y, Z) is supplied to the body shape estimation unit 61.

[0057]

[0058]

[0059] return Figure 1 , the structure of the signal processing unit 47 is explained. The body shape estimation unit 61 uses the position information of the reflection point obtained from the position information calculation unit 491 to generate a three-dimensional map indicating the existence position of the reflection point in the three-dimensional space as the observation object. In the present embodiment, the three-dimensional space as the observation object refers to a physical space in which it is assumed that there are people sitting in the car. The three-dimensional space as the observation object refers to a three-dimensional grid space represented as a grid having a three-dimensional shape in the observation data space. The three-dimensional grid space is represented, for example, as a three-dimensional grid map m represented in a three-dimensional rectangular coordinate system including a plurality of cells depicted in a grid at equal intervals. Each cell is represented by a number m indicating the position in the x-direction. x , m, which indicates the position in the y direction y and m, which indicates the position in the z direction z The body shape estimation unit 61 generates an occupied grid map indicating whether or not a detection target object exists in each cell of the three-dimensional grid map m.

[0060] The three-dimensional grid map m representing the vehicle interior space before observation is stored in advance in the map storage unit 63. In addition, the three-dimensional grid map m can be generated by the signal processing unit 47 when the body shape estimation device 1 is activated. When observing the vehicle interior space, the body shape estimation unit 61 reads the three-dimensional grid map m from the map storage unit 63 and plots the position coordinates of the reflection point supplied from the frequency analysis unit 49 on the three-dimensional grid map m. Based on the plotting result, an occupation grid map is generated, and the occupation grid map holds the probability of existence of the moving body in each cell constituting the three-dimensional grid map m.

[0061] Below, refer to Figure 3 , the detailed structure of the body shape estimation unit 61 is described. Figure 3As shown, the body shape estimation unit 61 includes an occupation probability calculation unit 611 , an occupation probability update unit 612 , and an occupation grid map generation unit 613 .

[0062] The occupation probability calculation unit 611 calculates the temporary occupation probability indicating the probability of the presence of a moving object in each cell in a certain measurement cycle based on the plotting result. Specifically, when the detection value vector z representing the xyz coordinates of the detection point is input in the nth measurement cycle, n When, according to the following formula (5), the calculation is expressed in relation to z n The corresponding cell m on the occupied grid map xyz The temporary mobile body occupancy probability p(m xyz |z n ). It can be considered that all the detection points represent moving objects by the moving object extraction unit 48, so the detection value vector z n The corresponding cell m xyz Assign temporary mobile body occupancy probability p(m xyz |z n )1. For other cells, the temporary mobile body occupancy probability p(m xyz |z n )0.

[0063] Similarly, when the detection value vector z representing the coordinates of the detection point is input in the n+1th measurement cycle, n When n The corresponding cell m xyz Assign temporary mobile body occupancy probability p(m xyz |z n )1, and assign temporary mobile body occupancy probability p(m xyz |z n )0. Temporary mobile body occupancy probability p(m xyz |z n ) is calculated at each input detection value z n It is carried out at the time.

[0064] The occupation probability calculation unit 611 calculates the temporary mobile object occupation probability p(m xyz |z n ) is supplied to the occupancy probability update unit 612.

[0065]

[0066] The occupation probability updating unit 612 uses the temporary mobile body occupation probability p(m xyz |z n), according to the following formula (6), the detection value z based on all measurement cycles from the 1st to the nth time is 1：n , indicating that in each cell m xyz The probability of a mobile object existing on the mobile object occupancy probability p(m xyz |z 1：n ) to update.

[0067] p(m xyz |z 1：n )=min{(1-w).p(m xyz |z 1：n-1 )+wp(m xyz |z n ),1} (6)

[0068] In addition, 1：n-1 Represents all detection values ​​from the 1st measurement cycle to the n-1th measurement cycle. The coefficient w is a coefficient representing the weight. The function min{a, b} is a min function that returns the minimum value of the independent variables a and b. As shown in formula (6), the mobile body occupancy probability is suppressed to an upper limit of 1, and the mobile body occupancy probability is adjusted to not exceed 1.

[0069] Sometimes, measurement errors occur, such as missed detection of an object that cannot be detected even though the object exists in only one measurement, or false alarm of an object that is detected even though the object does not exist. As in the present embodiment, by adding the temporary mobile body occupancy probability obtained based on the detection data of the current measurement cycle to the mobile body occupancy probability obtained based on the detection data up to the previous measurement cycle, the final mobile body occupancy probability is updated, thereby suppressing such measurement errors.

[0070] The occupied grid map generation unit 613 causes each cell to hold the obtained moving body occupation probability p(m xyz |z 1：n That is, the obtained moving object occupancy probability p(m xyz |z 1：n ) is associated with each cell. Thus, an occupation grid map is generated that indicates the probability of a mobile body existing in each cell of the space m. The mobile body occupation probability p(m) can be determined based on a predetermined threshold value ε (e.g., 0.8) or higher. xyz |z 1：n ) to estimate the shape of the moving object. Alternatively, three-dimensional imaging can be performed based on the moving object occupancy probability. For example, the moving object occupancy probability p(m) above a predetermined threshold ε (e.g., 0.8) is displayed in black. xyz |z 1：n) cells with a mobile body occupancy probability p(m) less than the threshold are displayed in white. xyz |z 1：n ) cells.

[0071] Next, refer to Figure 4A , Figure 4B , the hardware structure of the signal processor 41 is described. As an example, Figure 4A As shown, the control unit 45, the moving body extraction unit 48, the frequency analysis unit 49, and the body shape estimation unit 61 of the signal processor 41 are implemented by the processing circuit 100. The processing circuit 100 is, for example, a single circuit, a composite circuit, a programmed processor, a parallel programmed processor, an ASIC (Application Specific Integrated Circuit), an FPGA (Field-Programmable Gate Array), or a combination thereof. The functions of the control unit 45, the moving body extraction unit 48, the frequency analysis unit 49, and the body shape estimation unit 61 may be implemented by different processing circuits, or these functions may be integrated and implemented by one processing circuit.

[0072] As another example, Figure 4B As shown in FIG. 1 , the control unit 45, the moving body extraction unit 48, the frequency analysis unit 49, and the body shape estimation unit 61 of the signal processor 41 are implemented by the processor 101 and the memory 102. The control unit 45, the moving body extraction unit 48, the frequency analysis unit 49, and the body shape estimation unit 61 are implemented by the processor 101 reading and executing the program stored in the memory 102. The program is implemented by software, firmware, or a combination of software and firmware. Examples of the memory 102 include non-volatile or volatile semiconductor memories such as RAM (Random Access Memory), ROM (Read Only Memory), flash memory, EPROM (Erasable Programmable Read Only Memory), EEPROM (Electrically-EPROM), magnetic disks, floppy disks, optical disks, compact disks, mini disks, DVDs, and the like.

[0073] <Action>

[0074] Next, refer to Figure 5 and Figure 6 , describing the operation of the body shape estimation device 1. Figure 5 is a flowchart showing the operation of the body shape estimation device 1. Figure 6 yes Figure 5 A detailed flowchart of the processing of step ST14 is shown in FIG. Figure 5 The processing is started, for example, by turning on the power of the body shape estimation device 1 in conjunction with turning on the ignition of the vehicle.

[0075] In step ST10, a transmission wave is transmitted to the vehicle interior space via the transmission antenna 20. Specifically, a frequency-modulated chirp is transmitted to the vehicle interior space via the transmission antenna 20.

[0076] In step ST11, the receiving antennas 300 to 30 Q-1 Receives chirps that are reflected by objects inside the vehicle.

[0077] In step ST12, the mixers 330 to 33 Q-1 The linear frequency modulation is sent and received by the receiving antenna 300~30 Q-1 The received linear frequency modulation is mixed to generate a signal that is transmitted to the receiving antenna 300 to 30 Q-1 The number of corresponding beat signals.

[0078] In step ST13, the three-dimensional position information of the reflection point is calculated by the frequency analysis unit 49. In addition, according to the embodiment including the moving object extraction unit 48, the reflection point is the reflection point of the moving object.

[0079] In step ST14, the body shape estimation unit 61 estimates the body shape of the moving object. Figure 6 The details of the processing in step ST14 are described below. In step ST141, the occupation probability calculation unit 611 calculates the temporary mobile body occupation probability p(m xyz |z n ).

[0080] In step ST142, the occupation probability updating unit 612 calculates the moving body occupation probability p(m xyz |z 1：n-1 ) and the temporary mobile body occupancy probability p(m xyz |z n ) as the weighted average of the moving object occupancy probability p(m xyz |z 1：n ).

[0081] In step ST143, the occupation grid map generation unit 613 stores the mobile body occupation probability p(m) up to the current measurement period in each cell. xyz |z 1：n ), and generate an occupied grid map. xyz |z 1：n )’s spatial width of the cell and estimate the shape of the moving object. Figure 5 The processing is ended, for example, by turning off the power of the body shape estimation device 1 in conjunction with turning off the ignition of the vehicle.

[0082] Here, refer to Fig. 7A , Figure 7B , which illustrates an example of generating an occupied grid map. Fig. 7A This is a diagram of the space INT in the vehicle as viewed from the front of the vehicle. The sensor 10 of the body estimation device 1 is arranged on the line where the plane between the front seat and the rear seat of the vehicle and the plane between the rear seat RS1 and the rear seat RS2 intersect and on the position of the roof of the vehicle in a manner that emits radio waves into the vehicle compartment. In addition, such a configuration is only an example, and the body estimation device 1 can also be arranged at other positions, such as a position near the rearview mirror or on the position of the ceiling of the rear seat. In the space INT, a person P1 sits on the rear seat RS1, and a person P2 sits on the rear seat RS2. Luggage L is placed between the sensor 10 and the person P2. The person P2 is, for example, a child or an infant of short stature. When observed from the sensor 10, the person P2 is in a state of being hidden in the shadow of the luggage L.

[0083] Under such circumstances, according to the above body shape estimation device 1 having the moving body extraction unit 48, the person P1 is detected as a moving body, and the occupancy probability of the cell corresponding to the space where the person P1 exists is 1. The luggage L is a stationary object, so it is not detected as a moving body, and the person P2 is hidden in the shadow of the luggage L, so it is also not detected as a moving body. Therefore, the moving body occupancy probability of the cell of the space where the luggage exists and the space of the shadow of the luggage when viewed from the sensor 10 is calculated to be 0. Therefore, the generated Figure 7B The occupation grid map shown. Figure 7B In FIG. 1 , an example of imaging based on the mobile body occupancy probability maintained in each cell is shown. Figure 7B In the figure, black cells indicate the presence of a moving object, and white cells indicate the absence of a moving object.

[0084] <Modification 1>

[0085] According to the above-described embodiment in which the signal processing unit 47 includes the moving object extraction unit 48, the generated occupation grid map is Figure 7B , person P2 is not detected. However, the reason why person P2 is not detected is not that person P2 does not exist, but that person P2 is located in a blind spot due to the presence of luggage L. In such a case where a blind spot is generated, it may be better to determine that it is unclear whether a moving object exists rather than to determine that there is no moving object at the position that becomes the blind spot. That is, Figure 8 That kind of grid map is better. Figure 8In FIG. 1 , a black cell indicates that a moving object exists, a white cell indicates that a moving object does not exist, and a gray cell indicates that it is unclear whether a moving object exists. Therefore, a modification example in which it is possible to determine that it is unclear whether a moving object exists will be described.

[0086] <Structure>

[0087] The body shape estimation device 1A of the modification 1 is different from the embodiment 1 in that the signal processing unit 47 does not include the moving body extraction unit 48. Therefore, the position information supplied from the frequency analysis unit 49 to the body shape estimation unit includes the position information of the reflection point of the stationary object in addition to the position information of the reflection point of the moving object. In the modification 1, the frequency analysis unit 49 also includes the polar coordinate data (R, θ, ) and Doppler frequency f d The modified example 1 is different from the embodiment 1 in that the data is supplied to the body shape estimation unit. The modified example 1 is the same as the embodiment 1 in other configurations. For the same configurations, repeated descriptions are omitted.

[0088] like Fig. 9 As shown, the body shape estimation unit 61A of the body shape estimation device 1A of the modification example 1 includes an object determination unit 614, an occupation probability calculation unit 611A, an occupation probability update unit 612A, and an occupation grid map generation unit 613A. The body shape estimation unit 61A obtains the position information of the reflection point from the frequency analysis unit 49. The signal processing unit 47 of the body shape estimation device 1A does not include the moving body extraction unit 48. Therefore, in addition to the position information of the reflection point of the moving body, the position information also includes the position information of the reflection point of the stationary object. In addition, the body shape estimation unit 61A obtains the polar coordinate data (R, θ, θ) in addition to the rectangular coordinate data (X, Y, Z). ). In addition, the body shape estimation unit 61A also obtains the Doppler frequency f of the reflection point. d .

[0089] The object determination unit 614 determines the Doppler frequency f at the reflection point. d , and determines whether the object present at the reflection point is a moving object or a stationary object. The object determination unit 614 supplies the determination result to the occupation probability calculation unit 611A.

[0090] The occupation probability calculation unit 611A uses equation (7) to calculate the temporary mobile body occupation probability p(m xyz |Z n ), the occupation probability update unit 612A uses equations (8) and (9) to obtain the final mobile body occupation probability p(m xyz |Z 1：n ).

[0091] When K detection value sets Zn={z 1 n , z 2 n , ..., z K n} after the specific cell m xyz The mobile body occupancy probability is set to p(m xyz |Z 1：n ), p(m xyz |Z 1：n ) is derived indirectly from the following equations (7) and (8). As shown in equation (7), as a general formula, the temporary mobile body occupancy probability p(m xyz |Z n ) is calculated as the product of K three-dimensional mixed Gaussian distributions for each detection point. When it is determined that the object at a certain detection point is a moving object, the three-dimensional mixed Gaussian distribution for the detection point is expressed as a three-dimensional mixed Gaussian distribution with peaks at the position of the moving object and the position of the sensor. When it is determined that the object at a certain detection point is a stationary object, the three-dimensional mixed Gaussian distribution for the detection point is expressed as a three-dimensional Gaussian distribution with peaks at the position of the sensor. That is, when a stationary object is detected, the three-dimensional mixed Gaussian distribution in the general formula is a normal three-dimensional Gaussian distribution. Formula (7) is an update formula based on the Bayesian estimation of occupancy probability. The newly obtained data of this detection cycle is added to the probability information up to the previous detection cycle to obtain new probability information up to this detection cycle. ln(m) in formula (8) xyz ) is defined as the moving body occupancy probability p(m xyz |Z 1：n ) divided by the probability of non-occupancy of the mobile body as its negation 1-p(m xyz |Z 1：n ) and the logarithm of the ratio obtained. In addition, the initial value of the mobile body occupancy probability of all cells is set to, for example, 0.5 (ie, unclear state).

[0092]

[0093] Here,

[0094]

[0095]

[0096]

[0097] Among these formulas, is the polar coordinate display of the detection value vector, R k is the distance from the sensor to the detection value k, θk is the azimuth, is the elevation angle. In addition, N(μ,Σ) represents a three-dimensional polar coordinate Gaussian distribution with mean μ and error covariance Σ. Σocc is the error covariance of the occupied probability in polar coordinates, and Σemp is the error covariance of the non-occupied probability in polar coordinates.

[0098] As shown in formula (8), the logarithmic odds l of the probability of mobile body occupation based on the detection values ​​up to the current measurement cycle n is n (m xyz ) is calculated as the logarithmic odds l of the probability of mobile body occupation based on the detection values ​​up to the previous measurement cycle n-1 n-1 (m xyz ) and the logarithmic odds of the temporary mobile body occupation probability log{[p(m xyz |Z n )] / [1-p(m xyz |Z n Based on the calculation result, the moving body occupancy probability p(m) based on the detection values ​​up to the current measurement cycle n is calculated using equation (9) which is the definition of logarithmic odds. xyz |Z 1：n ). That is, the definition of equation (9) is transformed, and the moving body occupancy probability p(m xyz |Z 1：n ) can be obtained by formula (10).

[0099]

[0100] Thus, when a moving body is detected, the temporary moving body occupancy probability of each cell is set as follows. The temporary moving body occupancy probability of one or more cells corresponding to the position of the moving body is set to a value close to 1. It can be considered that only air exists, so the temporary moving body occupancy probability of the cells representing the position between the position of the moving body and the sensor is set to a value that asymptotically becomes 0 from the position of the moving body toward the sensor. It can be considered that the existence of the moving body is unclear because the radio wave has not arrived, so the temporary moving body occupancy probability of the cells representing the position farther than the position of the moving body when observed from the sensor is set to 0.5. Here, "corresponding to the position of the moving body" refers to the cell of the position of the moving body and the cells within a range with a certain width from the position of the moving body. Therefore, "cells representing the position farther than the position of the moving body when observed from the sensor" do not include "cells corresponding to the position of the moving body".

[0101] Furthermore, when a stationary object is detected, the temporary moving body occupancy probability of one or more cells corresponding to the position of the stationary object and cells indicating a position farther than the position of the stationary object when observed from the sensor is set to 0.5, and the temporary moving body occupancy probability of cells between the position of the stationary object and the sensor is set to a value that asymptotically approaches 0 from the position of the stationary object toward the sensor. "Corresponding to the position of the stationary object" refers to the cell at the position of the stationary object and cells within a range of a certain width from the position of the stationary object, and "cells indicating a position farther than the position of the stationary object when observed from the sensor" do not include "cells corresponding to the position of the stationary object".

[0102] When multiple moving bodies are detected, the product of one or more temporary moving body occupancy probabilities that have been set when a moving body is detected is calculated for each cell and updated as a new temporary moving body occupancy probability of the cell. On the other hand, when at least one moving body and at least one stationary object are detected, the product of one or more temporary moving body occupancy probabilities that have been set when a moving body is detected and one or more temporary moving body occupancy probabilities that have been set when a stationary object is detected is calculated for each cell and calculated as a new temporary moving body occupancy probability of the cell.

[0103] In this way, for each cell, the logarithmic odds of the mobile body occupancy probability up to the current measurement period are calculated as the sum of the logarithmic odds of the mobile body occupancy probability up to the previous measurement period and the logarithmic odds of the temporary mobile body occupancy probability in the current measurement period, and based on the calculated logarithmic odds of the mobile body occupancy probability up to the current measurement period, the mobile body occupancy probability up to the current measurement period is found.

[0104] Instead of the occupation probability updating unit 612A, an occupation probability updating unit 612Aa may be provided, and the occupation probability updating unit 612Aa updates the occupation probability according to the equation (6) by calculating the weighted average using the following (11). In this case, the equation (8) is not used.

[0105] p(m xyz |z 1:n )=min{(1-w)·p(m xyz |z 1：n-1 )+w·p(m xyz |z n ), 1} (11)

[0106] In this way, for the mobile body occupancy probability of each cell, a weighted average of the mobile body occupancy probability until the previous measurement cycle and the temporary mobile body occupancy probability in the current measurement cycle may be calculated as the mobile body occupancy probability until the current measurement cycle.

[0107] The mobile body occupancy probability up to the current measurement period is allocated to each cell as the mobile body occupancy probability, thereby generating an occupation grid map.

[0108] <Action>

[0109] Next, refer to Fig.10 The operation of body shape estimation apparatus 1A according to Modification 1 will be described. According to body shape estimation apparatus 1A, the detailed operation of step ST14 related to body shape estimation is different from that of Embodiment 1. Therefore, the operation of the different part will be described.

[0110] In step ST140, the object determination unit 614 determines the object based on the Doppler frequency f at the reflection point. d , determine whether the object at the reflection point is a moving object or a stationary object.

[0111] In step ST141A, the occupation probability calculation unit 611A calculates the temporary mobile body occupation probability p(m xyz |Z n ).

[0112] In step ST142A, the occupancy probability update unit 612A calculates the logarithmic odds of the mobile body occupancy probability up to the previous measurement period. n-1 (m xyz ) and the logarithmic odds of the temporary mobile body occupation probability in the current measurement period log{[p(m xyz |Z n )] / [1-p(m xyz |Z n )]}, and the logarithmic odds l of the probability of mobile body occupation up to the current measurement period is obtained. n (m xyz ). Based on the logarithmic odds of the mobile body occupation probability up to the current measurement period, n (m xyz ), the mobile body occupancy probability p(m xyz |Z 1：n The occupancy probability can be updated by the occupancy probability updating unit 612Aa calculating the weighted average according to equation (11).

[0113] In step ST143A, the occupation grid map generation unit 613A causes each cell to hold the moving body occupation probability p(m xyz |Z 1：n ), and generate an occupied grid map. xyz |Z1：n )’s spatial width of the cell and estimate the shape of the moving object.

[0114] According to the above body shape estimation device 1A, in Fig. 7A Under the same conditions, the occupation grid map ( Figure 8 ).according to Figure 8 In the occupied grid map, the gray cells indicate that it is unclear whether there is a moving object for the luggage L as a stationary object and the space S hidden in the shadow of the luggage L. That is, the space in which it is unclear whether there is a moving object in the car is determined. Therefore, it is possible to avoid the situation where the moving object is displayed as not existing even though there is a moving object.

[0115] <Modification 2>

[0116] When using laser to detect an object of the size of a person, the laser has a high straightness, so it is difficult for the laser to go around the back of the object and detect other objects. In contrast, when using electromagnetic waves such as radio waves with a frequency lower than that of laser to detect an object of the size of a person, the radio waves sometimes go around the back of the object. Therefore, sometimes a mobile body B that exists behind a certain mobile body A is detected. In such a case, according to the body shape estimation device 1A of the above-mentioned modification example 1, the temporary mobile body occupancy probability is set to a value that asymptotically becomes 0 from the position of the mobile body B toward the sensor through equation (6), so the final mobile body occupancy probability of the cell corresponding to the detected mobile body A will decrease. However, since mobile body A is detected, it may be better not to reduce the mobile body occupancy probability of the cell corresponding to mobile body A. Therefore, the following modification example 2 is described: when mobile body A and mobile body B that exists behind it are detected, the mobile body occupancy probability of mobile body A will not decrease. In modification example 2, Dempster-Shafer's evidence theory is used.

[0117] The body shape estimation device 1B of the second modification is different from the first embodiment in that the signal processing unit 47 does not include the moving body extraction unit 48. Therefore, the position information supplied from the frequency analysis unit 49 to the body shape estimation unit includes the position information of the reflection point of the stationary object in addition to the position information of the reflection point of the moving object. In addition, in the second modification, the frequency analysis unit 49 also includes the polar coordinate data (R, θ, ) and Doppler frequency f d The modified example 2 is different from the embodiment 1 in that the data is supplied to the body shape estimation unit. The modified example 2 is the same as the embodiment 1 in other configurations. The repeated description of the same configurations will be omitted.

[0118] like Fig.11As shown, the body shape estimation unit 61B includes an object determination unit 614, a basic probability distribution calculation unit 615, a basic probability distribution update unit 616, a probability conversion unit 617, and an occupied grid map generation unit 613B.

[0119] The object determination unit 614 determines the Doppler frequency f at the reflection point. d , and determines whether the object present at the reflection point is a moving object or a stationary object. The object determination unit 614 supplies the determination result to the occupation probability calculation unit 611A.

[0120] The basic probability distribution calculation unit 615 uses Dempster-Shafer's evidence theory to calculate the temporary mobile body occupied basic probability distribution, the temporary mobile body non-occupied basic probability distribution and the temporary unclear basic probability distribution for each cell. The basic probability distribution update unit 616 uses Dempster-Shafer's evidence theory to combine the three basic probability distributions (mobile body occupied basic probability distribution, mobile body non-occupied basic probability distribution and unclear basic probability distribution) of each cell in the previous measurement cycle with the temporary mobile body occupied basic probability distribution, temporary mobile body non-occupied basic probability distribution and temporary unclear basic probability distribution in this measurement cycle according to each cell, and finally calculates the three basic probability distributions of each cell in this measurement cycle.

[0121] Here, Dempster-Shafer's theory of evidence is explained. Consider all sets consisting of 2 elements Θ = {E, O}. E refers to the unoccupied state, and O refers to the occupied state. The power set of all sets Θ is X = {Φ, E, O, Θ}. Φ refers to the empty set, and Θ refers to the state where it is unclear whether it is E or O. According to Dempster-Shafer's theory of evidence, for the power set X, a function equivalent to the probability distribution in probability theory, namely the basic probability distribution, is defined as the following formula (12).

[0122]

[0123] Next, according to Dempster-Shafer's evidence theory, the combination operation of multiple basic probability distributions is defined as the following formula (13).

[0124]

[0125] Finally, according to formula (14), the basic probability distribution m(X) is converted into the probability value p(X) in probability theory.

[0126]

[0127] Assume that based on the Dempster-Shafer evidence theory as described above, in the nth measurement cycle, K detection value sets Zn={z 1 n , z 2 n , …, z K n}. The first M of the K is a moving object, and the rest are stationary objects. xyz The mobile body occupancy probability is set to p(m xyz |Z 1：n ), p(m xyz |Z 1：n ) is directly derived by the following equations (15) and (16). Equation (15) is an equation using a combination of basic probability distributions based on Dempster-Shafer. Equation (16) is a conversion equation that converts basic probability distributions m(X) into probability values ​​p(X).

[0128]

[0129] p(m xyz |z 1:n )=m n (O|m xyz )+0.5·m n (Θ|m xyz ) (16)

[0130] Among these formulas, is the polar coordinate display of the detection value vector, R k is the distance from the sensor to the detection value k, θ k is the azimuth, is the elevation angle. In addition, N(μ,Σ) represents a three-dimensional polar coordinate Gaussian distribution with a mean value μ and an error covariance Σ. Σocc is the error covariance of the occupancy probability in polar coordinates, and Σemp is the error covariance of the non-occupancy probability in polar coordinates. Different from the case of variant 1 determined by a single mixed Gaussian distribution, the moving body occupancy probability P occ (k) (m xyz ) and non-occupancy probability P emp (k) (m xyz ) are updated by separate operations.

[0131] The basic probability distribution calculation unit 615 calculates the temporary mobile body occupied basic probability distribution, the temporary mobile body non-occupied basic probability distribution, and the temporary unclear basic probability distribution for each cell according to formula (15). That is, when a mobile body is detected, the basic probability distribution calculation unit 615 sets the temporary mobile body occupied basic probability distribution of one or more cells corresponding to the position of the mobile body to be close to 1, and the temporary mobile body occupied basic probability distribution of cells other than the one or more cells corresponding to the position of the mobile body asymptotically becomes 0 as the distance from the position of the mobile body increases (m occ k (X|m xyz )=P occ (k) (m xyz )), for all cells, the temporary unclear basic probability distribution is set to the value obtained by subtracting the temporary mobile body occupancy basic probability distribution of the cell from 1 (m occ k (X|m xyz )=1-P occ (k) (m xyz )). In addition, when a stationary object is detected, the basic probability distribution calculation unit 615 sets the temporary mobile body non-occupancy basic probability distribution of one or more cells corresponding to the position of the sensor to be close to 1, and the temporary mobile body non-occupancy basic probability distribution of cells other than the one or more cells corresponding to the position of the sensor asymptotically becomes 0 from the position of the sensor toward the position of the stationary object (m emp k (X|m xyz )=P emp (k) (m xyz )), for all cells, the temporary unclear basic probability distribution is set to the value obtained by subtracting the temporary mobile body non-occupancy basic probability distribution of the cell from 1 (m emp k (X|m xyz )=1-P emp (k) (m xyz )). Moreover, in the case where the multiple objects include multiple mobile bodies, or at least one mobile body and at least one stationary object, the basic probability distribution calculation unit 615 combines the temporary mobile body occupied basic probability distribution or the temporary mobile body non-occupied basic probability distribution related to each object with the temporary unclear basic probability distribution for each cell, and calculates a new temporary mobile body occupied basic probability distribution, a new temporary mobile body non-occupied basic probability distribution, and a new temporary unclear basic probability distribution for each cell.

[0132] The basic probability distribution updating unit 616 updates the basic probability distribution m of each cell up to the previous measurement cycle for each cell according to equation (15). n-1 (X|m xyz ) and the temporary basic probability distribution in this measurement cycle (mobile body occupied basic probability distribution, temporary mobile body non-occupied basic probability distribution and temporary unclear basic probability distribution), and finally calculate the three basic probability distributions of each cell in this measurement cycle.

[0133] The probability conversion unit 617 converts the final three basic probability distributions of each cell in the current measurement cycle into the mobile body occupancy probability p(m xyz |Z 1：n ).

[0134] The occupied grid map generation unit 613B causes each cell to hold the obtained moving body occupation probability p(m xyz |Z 1：n ). Thus, an occupation grid map is generated that indicates the probability that a mobile object exists in each cell of the space m.

[0135] <Action>

[0136] Next, refer to Fig.12 Next, the operation of body shape estimation apparatus 1B according to Modification 2 will be described. According to body shape estimation apparatus 1B, the detailed operation of step ST14 related to body shape estimation is different from that of Embodiment 1. Therefore, the operation of the different part will be described.

[0137] In step ST140, similarly to the case of Modification 1, the object determination unit 614 determines the object based on the Doppler frequency f at the reflection point. d , determine whether the object at the reflection point is a moving object or a stationary object.

[0138] In step ST145 , the basic probability distribution calculation unit 615 calculates the temporary basic probability distribution (the moving body occupied basic probability distribution, the temporary moving body non-occupied basic probability distribution, and the temporary unclear basic probability distribution) in the current measurement period.

[0139] In step ST146, the basic probability distribution updating unit 616 updates the basic probability distribution m of each cell up to the previous measurement cycle for each cell. n-1 (X|m xyz ) and the temporary basic probability allocation in this measurement cycle (mobile body occupied basic probability allocation, temporary mobile body non-occupied basic probability allocation and temporary unclear basic probability allocation) are combined to update the three basic probability allocations of each cell in this measurement cycle.

[0140] In step ST147, the probability conversion unit 617 converts the final three basic probability distributions of each cell in the current measurement cycle into the mobile body occupancy probability p(m xyz |Z 1：n ).

[0141] In step ST143B, the occupation grid map generation unit 613B causes each cell to retain the moving body occupation probability p(m xyz |Z 1：n ), and generate an occupied grid map. xyz |Z 1：n )’s spatial width of the cell and estimate the shape of the moving object.

[0142] Below, refer to Fig.13A and Fig. 13B , explaining the advantages of variant example 2. Fig.13A and Fig. 13B Each of the diagrams shows the results of simulations of the following cases: when moving object A is detected at a point 10 m from the sensor and moving object B is detected at a point 30 m from the sensor, what kind of distribution of moving object occupancy probability is formed. For simplicity, the simulation is performed in a one-dimensional space. Fig.13A The simulation results for Modification 1 are shown. Fig. 13B The simulation results in the case of Modification Example 2 are shown.

[0143] exist Fig.13A In the figure, Occupancy represents the distribution of the probability of mobile occupancy when mobile B is detected, Empty represents the distribution of the probability of mobile non-occupancy when mobile B is detected, Temp. represents the overlap of the distribution of these mobile occupancy probabilities and the distribution of mobile non-occupancy probabilities, and All represents the sum of the distribution of the mobile occupancy probability when mobile A is detected and the distribution of the mobile occupancy probability when mobile B is detected.

[0144] On the other hand, Fig. 13B In the above, Occupancy represents the distribution of mobile body occupation basic probability distribution, Empty represents the distribution of mobile body non-occupying basic probability distribution, Unknown represents the distribution of unclear basic probability distribution, and All represents the distribution of mobile body occupation probability calculated based on these basic probability distributions.

[0145] contrast Fig.13A and Fig. 13B It is known that in Fig.13AIn , the probability of mobile body occupation at the point 10m away is lower than the probability of mobile body occupation at the point 30m away. This is considered to be because the probability of mobile body existence between the sensor and mobile body B is reduced by detecting mobile body B. In contrast, in Fig. 13B In the example, the probability of mobile body occupation at a location 10 m away is the same as the probability of mobile body occupation at a location 30 m away. This is because, unlike the occupied state or the unoccupied state, m(θ) is clearly prepared as another variable for the state where it is unclear whether a mobile body exists, and calculation is performed.

[0146] As described above, according to body shape estimation apparatus 1B of Modification 2, when a plurality of moving objects are detected by radio wave diffraction, the moving object occupancy probability of one moving object is not affected by the detection of another moving object. Therefore, body shape estimation apparatus 1B can estimate the shapes of moving objects more appropriately.

[0147] According to the above-described body shape estimation devices 1 , 1A, and 1B, there is no need to process the vehicle seat for detecting the body shape as in the prior art such as Patent Document 1, and therefore there is an advantage that the introduction cost of the device can be suppressed.

[0148] Implementation Method 2

[0149] Next, refer to Fig.14 , Fig.15 and Fig.16 , explaining the seat belt reminder system and air bag control system. Fig.14 FIG. 2 is a block diagram showing the configuration of a seat belt reminder system and an air bag control system according to Embodiment 2. Fig.14 As shown, the body shape estimation device 1 (1A, 1B) outputs the body shape estimation result DD to one or both of the seat belt wearing determination device 200 and the air bag control device 300 via the output control unit (not shown) of the body shape estimation device 1 (1A, 1B). By outputting the body shape estimation result DD to the seat belt wearing determination device 200, a seat belt reminder system 400 including the body shape estimation device 1 (1A, 1B) and the seat belt wearing determination device 200 is formed. By outputting the body shape estimation result DD to the air bag control device 300, an air bag control system 500 including the body shape estimation device 1 (1A, 1B) and the air bag control device 300 is formed.

[0150] like Fig.15As shown, the seat belt wearing determination device 200 includes a determination unit 201 and a notification unit 202. The determination unit 201 determines whether the seat belt corresponding to the position of the moving body whose body shape is estimated is worn based on the body shape estimation result DD, and outputs the determination result to the notification unit 202. The notification unit 202 notifies an alarm indicating that the seat belt is not worn by sound or display in a prescribed case based on the determination result. As an example, when a body shape of a prescribed size or more is estimated, if it is determined that the seat belt corresponding to the position of the moving body whose body shape is estimated is not worn, the notification that the seat belt is not worn is given.

[0151] As another example, in the case where a body size smaller than a specified size is estimated, a notification indicating that the seat belt is not worn is not made. A seat for infants or children (hereinafter, seats for infants and seats for children are simply collectively referred to as "child seats") is placed on the seat of the vehicle, and sometimes infants or children sit in the child seats. In such a situation, it is inappropriate to notify that the vehicle's seat belt is not worn even when a moving object is detected. This is because infants or children are scheduled to wear the seat belt of the child seat but are not scheduled to wear the vehicle's seat belt. In addition, regardless of the setting of the child seat, when the height of the occupant is less than a fixed length, wearing a seat belt is sometimes not a legal obligation. Even in such a case, when a body size smaller than the size corresponding to such a height is estimated, by not making a notification indicating that the seat belt is not worn, legal restrictions can be appropriately met.

[0152] The predetermined size may be determined based on the number of cells having a moving body existence probability greater than or equal to a predetermined threshold, or may be determined based on the shape of the edges of cells having a moving body existence probability greater than or equal to the predetermined threshold and cells having a moving body existence probability less than the threshold.

[0153] like Fig.16 As shown, the airbag control device 300 has a body shape determination unit 301 and an airbag control unit 302. The body shape determination unit 301 determines the size of the body shape based on the body shape estimation result DD, and outputs the determination result to the airbag control unit 302. The airbag control unit 302 controls the movement of the airbag corresponding to the position of the moving body whose body shape is estimated based on the determination result. For example, the airbag control unit 302 changes the pressure (deployment expansion force) of the airbag when it is working according to the body shape of the occupant. As an example, when a body shape larger than a specified size is estimated, a relatively high pressure is set, and when a body shape smaller than a specified size is estimated, a relatively low pressure is set. By performing such control, the possibility of infants or children whose bodies are not fully grown being injured by the deployment of the airbag can be reduced.

[0154] The seat belt wearing determination device 200 and the air bag control device 300 can both be connected to the air bag Figure 4A or Figure 4B The same hardware structure is used to implement it.

[0155] <Notes>

[0156] Several aspects of the various embodiments described above are summarized below.

[0157] <Additional Note 1>

[0158] The body shape estimation device (1, 1A, 1B) of Supplementary Note 1 comprises:

[0159] A sensor (10) having a transmitting antenna for transmitting a transmission wave and a receiving antenna for receiving the transmission wave after being reflected by at least one object in the vehicle compartment as a receiving wave;

[0160] a frequency analysis unit (49) which uses the received wave to obtain position information of a reflection point where the transmitted wave is reflected; and

[0161] A body shape estimation unit (61) estimates the body shape of a non-stationary object existing in the vehicle compartment using the position information.

[0162] <Note 2>

[0163] The body shape estimation device of Supplementary Note 2 is the body shape estimation device described in Supplementary Note 1, wherein:

[0164] The body shape estimation unit determines that a space in the vehicle compartment where a non-stationary object exists is unclear.

[0165] <Addendum 3>

[0166] The body shape estimation device of Supplement 3 is the body shape estimation device described in Supplement 1 or Supplement 2, wherein:

[0167] The transmission wave is a plurality of linear frequency modulations with increasing or decreasing frequency.

[0168] The sensor further includes a mixer that mixes the transmission wave and the reception wave to generate a beat signal.

[0169] The position information is obtained by analyzing the beat signal,

[0170] The body shape estimation unit uses the position information to calculate the moving body occupancy probability representing the existence probability of a non-stationary object for each cell in the three-dimensional grid space having multiple cells representing the interior of the vehicle, and estimates the spatial width of the cell where the moving body occupancy probability is above a threshold value as the body shape of the non-stationary object existing in the vehicle.

[0171] <Addendum 4>

[0172] The body shape estimation device according to Supplement 4 is the body shape estimation device according to any one of Supplements 1 to 3, wherein:

[0173] The frequency analysis unit analyzes the beat signal by performing range FFT, Doppler FFT, and angle FFT on the beat signal.

[0174] <Addendum 5>

[0175] The body shape estimation device (1) of Supplement 5 is the body shape estimation device described in Supplement 3 or Supplement 4, wherein:

[0176] The body shape estimation device further comprises a moving body extraction unit (48) which performs moving body extraction processing on the beat signal to extract non-stationary objects existing in the vehicle compartment.

[0177] The body shape estimation unit comprises:

[0178] an occupation probability calculation unit (611) for each of the plurality of cells, which sets the temporary mobile body occupation probability in the current measurement period to 1 when a non-stationary object is detected, and sets the temporary mobile body occupation probability in the current measurement period to 0 when no non-stationary object is detected;

[0179] an occupancy probability updating unit (612) for calculating, for each cell, a weighted average of the mobile body occupancy probability up to the previous measurement cycle and the temporary mobile body occupancy probability in the current measurement cycle as the mobile body occupancy probability up to the current measurement cycle; and

[0180] An occupation grid map generating unit (613) generates an occupation grid map by assigning the mobile body occupation probability up to the current measurement period to each cell as the mobile body occupation probability.

[0181] <Addendum 6>

[0182] The body shape estimation device (1A) of Supplementary Note 6 is the body shape estimation device described in Supplementary Note 3 or Supplementary Note 4, wherein:

[0183] The at least one object is a plurality of objects,

[0184] The body shape estimation unit comprises:

[0185] an object determination unit (614) for determining whether the detected object is a non-stationary object or a stationary object;

[0186] The occupancy probability calculation unit (611A) detects a plurality of non-stationary objects, or at least one non-stationary object and at least one stationary object as the plurality of objects.

[0187] When a non-stationary object is detected, the occupation probability calculation unit (611A) sets the temporary mobile body occupation probability of one or more cells corresponding to the position of the detected non-stationary object to a value close to 1, sets the temporary mobile body occupation probability of cells between the position of the detected non-stationary object and the sensor to a value that asymptotically becomes 0 from the position of the detected non-stationary object toward the sensor, and sets the temporary mobile body occupation probability of cells farther than the position of the detected non-stationary object when viewed from the sensor to 0.5.

[0188] When a stationary object is detected, the temporary moving body occupancy probability of one or more cells corresponding to the position of the detected stationary object and cells farther away than the position of the detected stationary object when observed from the sensor is set to 0.5, and the temporary moving body occupancy probability of cells between the position of the detected stationary object and the sensor is set to a value that asymptotically becomes 0 from the position of the detected stationary object toward the sensor,

[0189] When the plurality of non-stationary objects are detected, the product of one or more temporary mobile body occupancy probabilities set when a non-stationary object is detected is calculated for each cell, and the product is updated as a new temporary mobile body occupancy probability of the cell.

[0190] When the at least one non-stationary object and the at least one stationary object are detected, for each cell, the product of one or more temporary mobile body occupancy probabilities that have been set when the non-stationary object is detected and one or more temporary mobile body occupancy probabilities that have been set when the stationary object is detected is calculated as a new temporary mobile body occupancy probability of the cell;

[0191] an occupancy probability updating unit (612Aa) for calculating, for each cell, a weighted average of the occupancy probability of the mobile body up to the previous measurement cycle and the temporary occupancy probability of the mobile body in the current measurement cycle as the occupancy probability of the mobile body up to the current measurement cycle; and

[0192] An occupation grid map generation unit (613A) generates an occupation grid map by assigning the mobile body occupation probability up to the current measurement period to each cell as the mobile body occupation probability.

[0193] <Addendum 7>

[0194] The body shape estimation device (1A) of Supplement 7 is the body shape estimation device described in Supplement 3 or Supplement 4, wherein:

[0195] The at least one object is a plurality of objects,

[0196] The body shape estimation unit comprises:

[0197] an object determination unit (614) for determining whether the detected object is a non-stationary object or a stationary object;

[0198] The occupancy probability calculation unit (611A) detects a plurality of non-stationary objects, or at least one non-stationary object and at least one stationary object as the plurality of objects.

[0199] When a non-stationary object is detected, the occupation probability calculation unit (611A) sets the temporary mobile body occupation probability of one or more cells corresponding to the position of the detected non-stationary object to a value close to 1, sets the temporary mobile body occupation probability of cells between the position of the detected non-stationary object and the sensor to a value that asymptotically becomes 0 from the position of the detected non-stationary object toward the sensor, and sets the temporary mobile body occupation probability of cells farther than the position of the detected non-stationary object when viewed from the sensor to 0.5.

[0200] When a stationary object is detected, the temporary moving body occupancy probability of one or more cells corresponding to the position of the detected stationary object and cells farther away than the position of the detected stationary object when observed from the sensor is set to 0.5, and the temporary moving body occupancy probability of cells between the position of the detected stationary object and the sensor is set to a value that asymptotically becomes 0 from the position of the detected stationary object toward the sensor,

[0201] When the plurality of non-stationary objects are detected, the product of one or more temporary mobile body occupancy probabilities set when a non-stationary object is detected is calculated for each cell, and the product is updated as a new temporary mobile body occupancy probability of the cell.

[0202] When the at least one non-stationary object and the at least one stationary object are detected, for each cell, the product of one or more temporary mobile body occupancy probabilities that have been set when the non-stationary object is detected and one or more temporary mobile body occupancy probabilities that have been set when the stationary object is detected is calculated as a new temporary mobile body occupancy probability of the cell;

[0203] an occupation probability updating unit (612A) that calculates, for each cell, a logarithmic odds of the mobile body occupation probability up to the current measurement period as the sum of the logarithmic odds of the mobile body occupation probability up to the previous measurement period and the logarithmic odds of the temporary mobile body occupation probability in the current measurement period, and obtains the mobile body occupation probability up to the current measurement period based on the calculated logarithmic odds of the mobile body occupation probability up to the current measurement period; and

[0204] An occupation grid map generation unit (613A) generates an occupation grid map by assigning the mobile body occupation probability up to the current measurement period to each cell as the mobile body occupation probability.

[0205] <Addendum 8>

[0206] The body shape estimation device (1B) of Supplement 8 is the body shape estimation device described in Supplement 3 or Supplement 4, wherein:

[0207] The at least one object is a plurality of objects,

[0208] The body shape estimation unit comprises:

[0209] an object determination unit (614) for determining whether the detected object is a non-stationary object or a stationary object;

[0210] A basic probability distribution calculation unit (615) calculates three basic probability distributions including a mobile body occupation basic probability distribution indicating the presence of a non-stationary object, a mobile body non-occupancy basic probability distribution indicating the absence of a non-stationary object, and an unclear basic probability distribution indicating the presence or absence of a non-stationary object, for each measurement period.

[0211] When a non-stationary object is detected, the basic probability distribution calculation unit (615) sets the temporary moving body occupation basic probability distribution of one or more cells corresponding to the position of the detected non-stationary object to be close to 1, and the temporary moving body occupation basic probability distribution of cells other than the one or more cells corresponding to the position of the detected non-stationary object asymptotically becomes 0 as the distance from the position of the detected non-stationary object increases, and for all cells, the temporary unclear basic probability distribution is set to a value obtained by subtracting the temporary moving body occupation basic probability distribution of the cell from 1,

[0212] When a stationary object is detected, the basic probability distribution calculation unit (615) sets the temporary mobile body non-occupancy basic probability distribution of one or more cells corresponding to the position of the sensor to be close to 1, and the temporary mobile body non-occupancy basic probability distribution of cells other than the one or more cells corresponding to the position of the sensor asymptotically becomes 0 from the position of the sensor toward the position of the detected stationary object, and sets the temporary unclear basic probability distribution for all cells to a value obtained by subtracting the temporary mobile body non-occupancy basic probability distribution of the cell from 1,

[0213] In the case where the plurality of objects include a plurality of non-stationary objects, or at least one non-stationary object and at least one stationary object, the basic probability distribution calculation unit (615) combines the temporary mobile body occupation basic probability distribution or the temporary mobile body non-occupancy basic probability distribution and the temporary unclear basic probability distribution related to each object for each cell, and calculates a new temporary mobile body occupation basic probability distribution, a new temporary mobile body non-occupancy basic probability distribution and a new temporary unclear basic probability distribution for each cell;

[0214] a basic probability allocation updating unit (616) which combines the three basic probability allocations of each cell in the previous measurement cycle with the new temporary mobile body occupied basic probability allocation, the new temporary mobile body non-occupied basic probability allocation and the new temporary unclear basic probability allocation in the current measurement cycle for each cell, and finally calculates the three basic probability allocations of each cell in the current measurement cycle;

[0215] a probability conversion unit (617) that converts the final three basic probability distributions of each cell in the current measurement period into a mobile body occupancy probability of each cell up to the current measurement period; and

[0216] An occupation grid map generating unit (613B) generates an occupation grid map by assigning the mobile body occupation probability up to the current measurement period to each cell as the mobile body occupation probability.

[0217] <Addendum 9>

[0218] The seat belt reminder system (400) of Supplement 9 comprises: the body shape estimation device (1, 1A, 1B) as described in any one of Supplements 1 to 8; and a seat belt wearing determination device (200) for determining whether the seat belt is worn, wherein:

[0219] The body shape estimation device outputs an estimation result,

[0220] The seat belt wearing determination device comprises:

[0221] a determination unit (201) for determining whether a seat belt corresponding to a position of a non-stationary object existing in the vehicle compartment is worn based on the estimation result; and

[0222] A notification unit (202) notifies an alarm by sound or display when the determination unit determines that the person is not wearing a seat belt.

[0223] <Addendum 10>

[0224] The airbag control system (500) of Supplement 10 comprises: the body shape estimation device (1, 1A, 1B) as described in any one of Supplements 1 to 8; and an airbag control device (300) for controlling the movement of the airbag, wherein:

[0225] The body shape estimation device outputs an estimation result,

[0226] The airbag control device comprises:

[0227] a body shape determination unit (301) which determines the size of the body shape of the non-stationary object existing in the vehicle compartment based on the estimation result and outputs the determination result; and

[0228] An airbag control unit (302) controls the operation of the airbag corresponding to the position of the non-stationary object existing in the vehicle compartment based on the determination result.

[0229] <Addendum 11>

[0230] The body shape estimation method of Appendix 11 has the following steps:

[0231] Send a send wave;

[0232] receiving the transmitted wave reflected by at least one object in the vehicle compartment as a received wave;

[0233] Using the received wave, obtaining position information of a reflection point where the transmitted wave is reflected; and

[0234] Using the position information, the size of the non-stationary object in the vehicle compartment is estimated.

[0235] Furthermore, the embodiment modes may be combined or each embodiment may be appropriately modified or omitted.

[0236] Industrial Applicability

[0237] The body shape estimating device of the present invention can estimate the body shape of a moving object, and therefore can be used as a device for estimating the body shape of an occupant in a vehicle cabin.

[0238] Description of symbols

[0239] 1 (1A, 1B): body shape estimation device; 10: sensor; 20: transmitting antenna; 21: transmitting circuit; 22: voltage generator; 23: voltage controlled oscillator; 24: distributor; 25: amplifier; 300-30 Q-1 : Receiving antenna; 310~31 Q-1 : Receiver; 320~32 Q-1 : Low noise amplifier; 330~33 Q-1 :Mixer; 340~34 Q-1 :IF amplifier; 350~35 Q-1 : Filter; 360~36 Q-1 : ADC; 41: signal processor; 45: control unit; 46: data storage unit; 47: signal processing unit; 48: moving body extraction unit; 49: frequency analysis unit; 61 (61A, 61B): body shape estimation unit; 63: map storage unit; 100: processing circuit; 101: processor; 102: memory; 200: seat belt wearing determination device; 300: air bag control device; 400: seat belt reminder system; 491: position information calculation unit; 492: coordinate conversion unit; 500: Airbag control system; 611 (611A): occupancy probability calculation unit; 612 (612A, 612Aa): occupancy probability update unit; 613 (613A, 613B): occupancy grid map generation unit; 614: object determination unit; 615: basic probability allocation calculation unit; 616: basic probability allocation update unit; 617: probability conversion unit; 4911: range Doppler processing unit; 4912: integration processing unit; 4913: peak extraction processing unit; 4914: peak angle measurement processing unit.

Claims

1. A body shape estimation device, wherein: The body shape estimation device comprises: A sensor having a transmitting antenna, a receiving antenna, and a mixer, wherein the transmitting antenna transmits a transmission wave as a plurality of linear frequency modulations with increasing or decreasing frequencies, the receiving antenna receives the transmission wave reflected by a plurality of objects in a vehicle compartment as a reception wave, and the mixer mixes the transmission wave and the reception wave to generate a beat signal; a frequency analysis unit that uses the received wave to obtain position information of a reflection point where the transmitted wave is reflected by analyzing the beat signal; as well as a body shape estimation unit that estimates the body shape of the non-stationary object existing in the vehicle compartment using the position information, The body shape estimation unit calculates the moving body occupancy probability indicating the existence probability of the non-stationary object for each cell of the three-dimensional grid space having a plurality of cells representing the interior of the vehicle using the position information, and estimates the spatial width of the cell where the moving body occupancy probability is greater than a threshold value as the body shape of the non-stationary object existing in the vehicle. The body shape estimation unit comprises: an object determination unit that determines whether the detected object is a non-stationary object or a stationary object; The occupancy probability calculation unit detects a plurality of non-stationary objects, or at least one non-stationary object and at least one stationary object as the plurality of objects. When a non-stationary object is detected, the occupation probability calculation unit sets the temporary mobile body occupation probability of one or more cells corresponding to the position of the detected non-stationary object to a value close to 1, sets the temporary mobile body occupation probability of cells between the position of the detected non-stationary object and the sensor to a value that asymptotically approaches 0 from the position of the detected non-stationary object toward the sensor, and sets the temporary mobile body occupation probability of cells farther away than the position of the detected non-stationary object when viewed from the sensor to 0.

5. When a stationary object is detected, the occupation probability calculation unit sets the temporary mobile body occupation probability of one or more cells corresponding to the position of the detected stationary object and cells farther away than the position of the detected stationary object when observed from the sensor to 0.5, and sets the temporary mobile body occupation probability of cells between the position of the detected stationary object and the sensor to a value that asymptotically becomes 0 from the position of the detected stationary object toward the sensor, When the plurality of non-stationary objects are detected, the occupation probability calculation unit calculates, for each cell, the product of one or more temporary mobile body occupation probabilities that have been set when the non-stationary objects are detected, and updates the product as a new temporary mobile body occupation probability of the cell. When the at least one non-stationary object and the at least one stationary object are detected, the occupancy probability calculation unit calculates, for each cell, a product of one or more temporary mobile body occupancy probabilities that have been set when the non-stationary object is detected and one or more temporary mobile body occupancy probabilities that have been set when the stationary object is detected, and calculates the product as a new temporary mobile body occupancy probability of the cell; an occupation probability updating unit for calculating, for each cell, a weighted average of the mobile body occupation probability up to the previous measurement cycle and the temporary mobile body occupation probability in the current measurement cycle as the mobile body occupation probability up to the current measurement cycle; as well as The occupation grid map generating unit allocates the mobile body occupation probability up to the current measurement period to each cell as the mobile body occupation probability, and generates an occupation grid map.

2. A body shape estimation device, wherein: The body shape estimation device comprises: A sensor having a transmitting antenna, a receiving antenna, and a mixer, wherein the transmitting antenna transmits a transmission wave as a plurality of linear frequency modulations with increasing or decreasing frequencies, the receiving antenna receives the transmission wave reflected by a plurality of objects in a vehicle compartment as a reception wave, and the mixer mixes the transmission wave and the reception wave to generate a beat signal; a frequency analysis unit that uses the received wave to obtain position information of a reflection point where the transmitted wave is reflected by analyzing the beat signal; as well as a body shape estimation unit that estimates the body shape of the non-stationary object existing in the vehicle compartment using the position information, The body shape estimation unit calculates the moving body occupancy probability indicating the existence probability of the non-stationary object for each cell of the three-dimensional grid space having a plurality of cells representing the interior of the vehicle using the position information, and estimates the spatial width of the cell where the moving body occupancy probability is greater than a threshold value as the body shape of the non-stationary object existing in the vehicle. The body shape estimation unit comprises: an object determination unit that determines whether the detected object is a non-stationary object or a stationary object; The occupancy probability calculation unit detects a plurality of non-stationary objects, or at least one non-stationary object and at least one stationary object as the plurality of objects. When a non-stationary object is detected, the occupation probability calculation unit sets the temporary mobile body occupation probability of one or more cells corresponding to the position of the detected non-stationary object to a value close to 1, sets the temporary mobile body occupation probability of cells between the position of the detected non-stationary object and the sensor to a value that asymptotically approaches 0 from the position of the detected non-stationary object toward the sensor, and sets the temporary mobile body occupation probability of cells farther away than the position of the detected non-stationary object when viewed from the sensor to 0.

5. When a stationary object is detected, the occupation probability calculation unit sets the temporary mobile body occupation probability of one or more cells corresponding to the position of the detected stationary object and cells farther away than the position of the detected stationary object when observed from the sensor to 0.5, and sets the temporary mobile body occupation probability of cells between the position of the detected stationary object and the sensor to a value that asymptotically becomes 0 from the position of the detected stationary object toward the sensor, When the plurality of non-stationary objects are detected, the occupation probability calculation unit calculates, for each cell, the product of one or more temporary mobile body occupation probabilities that have been set when the non-stationary objects are detected, and updates the product as a new temporary mobile body occupation probability of the cell. When the at least one non-stationary object and the at least one stationary object are detected, the occupancy probability calculation unit calculates, for each cell, a product of one or more temporary mobile body occupancy probabilities that have been set when the non-stationary object is detected and one or more temporary mobile body occupancy probabilities that have been set when the stationary object is detected, and calculates the product as a new temporary mobile body occupancy probability of the cell; an occupation probability updating unit, which calculates the logarithmic odds of the mobile body occupation probability up to the current measurement cycle by summing the logarithmic odds of the mobile body occupation probability up to the previous measurement cycle and the logarithmic odds of the temporary mobile body occupation probability in the current measurement cycle for each cell, and obtains the mobile body occupation probability up to the current measurement cycle based on the calculated logarithmic odds of the mobile body occupation probability up to the current measurement cycle; as well as The occupation grid map generating unit allocates the mobile body occupation probability up to the current measurement period to each cell as the mobile body occupation probability, and generates an occupation grid map.

3. A body shape estimation device, wherein: The body shape estimation device comprises: A sensor having a transmitting antenna, a receiving antenna, and a mixer, wherein the transmitting antenna transmits a transmission wave as a plurality of linear frequency modulations with increasing or decreasing frequencies, the receiving antenna receives the transmission wave reflected by a plurality of objects in a vehicle compartment as a reception wave, and the mixer mixes the transmission wave and the reception wave to generate a beat signal; a frequency analysis unit that uses the received wave to obtain position information of a reflection point where the transmitted wave is reflected by analyzing the beat signal; as well as a body shape estimation unit that estimates the body shape of the non-stationary object existing in the vehicle compartment using the position information, The body shape estimation unit calculates the moving body occupancy probability indicating the existence probability of the non-stationary object for each cell of the three-dimensional grid space having a plurality of cells representing the interior of the vehicle using the position information, and estimates the spatial width of the cell where the moving body occupancy probability is greater than a threshold value as the body shape of the non-stationary object existing in the vehicle. The body shape estimation unit comprises: an object determination unit that determines whether the detected object is a non-stationary object or a stationary object; a basic probability distribution calculation unit that calculates, for each measurement period, three basic probability distributions including a mobile body occupation basic probability distribution indicating the presence of a non-stationary object, a mobile body non-occupancy basic probability distribution indicating the absence of a non-stationary object, and an unclear basic probability distribution indicating the presence or absence of an unclear non-stationary object; In the case where a non-stationary object is detected, the basic probability distribution calculation unit is set so that the temporary moving body occupation basic probability distribution of one or more cells corresponding to the position of the detected non-stationary object is close to 1, and the temporary moving body occupation basic probability distribution of cells other than the one or more cells corresponding to the position of the detected non-stationary object asymptotically becomes 0 as the distance from the position of the detected non-stationary object increases, and for all cells, the temporary unclear basic probability distribution is set to a value obtained by subtracting the temporary moving body occupation basic probability distribution of the cell from 1, When a stationary object is detected, the basic probability distribution calculation unit sets the temporary mobile body non-occupancy basic probability distribution of one or more cells corresponding to the position of the sensor to be close to 1, and the temporary mobile body non-occupancy basic probability distribution of cells other than the one or more cells corresponding to the position of the sensor asymptotically becomes 0 from the position of the sensor toward the position of the detected stationary object, and sets the temporary unclear basic probability distribution to a value obtained by subtracting the temporary mobile body non-occupancy basic probability distribution of the cell from 1 for all cells. In a case where the plurality of objects include a plurality of non-stationary objects, or at least one non-stationary object and at least one stationary object, the basic probability distribution calculation unit combines the temporary mobile body occupation basic probability distribution or the temporary mobile body non-occupancy basic probability distribution and the temporary unclear basic probability distribution related to each object for each cell, and calculates a new temporary mobile body occupation basic probability distribution, a new temporary mobile body non-occupancy basic probability distribution, and a new temporary unclear basic probability distribution for each cell; a basic probability allocation updating unit, which combines the three basic probability allocations of each cell in the previous measurement cycle with the new temporary mobile body occupied basic probability allocation, the new temporary mobile body non-occupied basic probability allocation and the new temporary unclear basic probability allocation in the current measurement cycle for each cell, and finally calculates the three basic probability allocations of each cell in the current measurement cycle; a probability conversion unit that converts the final three basic probability distributions of each cell in the current measurement period into a mobile body occupancy probability of each cell up to the current measurement period; as well as The occupation grid map generating unit allocates the mobile body occupation probability up to the current measurement period to each cell as the mobile body occupation probability, and generates an occupation grid map.

4. The body shape estimation device according to any one of claims 1 to 3, wherein: The frequency analysis unit analyzes the beat signal by performing range FFT, Doppler FFT, and angle FFT on the beat signal.

5. A seat belt reminder system, comprising: a body shape estimation device according to any one of claims 1 to 4; and a seat belt wearing determination device for determining whether a seat belt is worn, wherein: The body shape estimation device outputs an estimation result, The seat belt wearing determination device comprises: a determination unit that determines, based on the estimation result, whether a seat belt corresponding to a position of a non-stationary object present in the vehicle compartment is worn; as well as A notification unit notifies an alarm by sound or display when the determination unit determines that the seat belt is not being worn.

6. An airbag control system, comprising: the body shape estimation device according to any one of claims 1 to 4; and an airbag control device for controlling the movement of the airbag, wherein: The body shape estimation device outputs an estimation result, The airbag control device comprises: a body shape determination unit that determines the size of the body shape of the non-stationary object existing in the vehicle compartment based on the estimation result and outputs the determination result; as well as The airbag control unit controls the operation of the airbag corresponding to the position of the non-stationary object existing in the vehicle cabin based on the determination result.

7. A body shape estimation method performed by a body shape estimation device, the body shape estimation device comprising: a sensor having a transmitting antenna, a receiving antenna and a mixer; Frequency Analysis Unit; and A body shape estimation unit includes an object determination unit, an occupation probability calculation unit, an occupation probability update unit, and an occupation grid map generation unit, wherein: The body shape estimation method comprises the following steps: The transmitting antenna transmits a transmission wave as a plurality of linear frequency modulations with increasing or decreasing frequencies; The receiving antenna receives the transmission wave reflected by a plurality of objects in the vehicle compartment as a reception wave; The mixer mixes the transmission wave and the reception wave to generate a beat signal; The frequency analysis unit uses the received wave to obtain position information of a reflection point where the transmitted wave is reflected by analyzing the beat signal; The body shape estimation unit estimates the body shape of the non-stationary object existing in the vehicle compartment using the position information; The body shape estimation unit calculates the moving body occupancy probability indicating the existence probability of the non-stationary object for each cell of the three-dimensional grid space having a plurality of cells representing the interior of the vehicle using the position information, and estimates the spatial width of the cell where the moving body occupancy probability is greater than a threshold value as the body shape of the non-stationary object existing in the vehicle; The object determination unit of the body shape estimation unit determines whether the detected object is a non-stationary object or a stationary object; When a plurality of non-stationary objects or at least one non-stationary object and at least one stationary object are detected as the plurality of objects, When a non-stationary object is detected, the occupation probability calculation unit of the body shape estimation unit sets the temporary mobile body occupation probability of one or more cells corresponding to the position of the detected non-stationary object to a value close to 1, sets the temporary mobile body occupation probability of cells between the position of the detected non-stationary object and the sensor to a value that asymptotically becomes 0 from the position of the detected non-stationary object toward the sensor, and sets the temporary mobile body occupation probability of cells farther than the position of the detected non-stationary object when observed from the sensor to 0.

5. When a stationary object is detected, the occupation probability calculation unit of the body shape estimation unit sets the temporary mobile body occupation probability of one or more cells corresponding to the position of the detected stationary object and cells farther away than the position of the detected stationary object when observed from the sensor to 0.5, and sets the temporary mobile body occupation probability of cells between the position of the detected stationary object and the sensor to a value that asymptotically becomes 0 from the position of the detected stationary object toward the sensor, When the plurality of non-stationary objects are detected, the occupation probability calculation unit of the body shape estimation unit calculates, for each cell, the product of one or more temporary mobile body occupation probabilities that have been set when the non-stationary objects are detected, and updates the product as a new temporary mobile body occupation probability of the cell. When the at least one non-stationary object and the at least one stationary object are detected, the occupancy probability calculation unit of the body shape estimation unit calculates, for each cell, the product of one or more temporary mobile body occupancy probabilities that have been set when the non-stationary object is detected and one or more temporary mobile body occupancy probabilities that have been set when the stationary object is detected, and calculates the product as a new temporary mobile body occupancy probability of the cell; The occupancy probability updating unit of the body shape estimation unit calculates, for the occupancy probability of the mobile body in each cell, a weighted average of the occupancy probability of the mobile body up to the previous measurement cycle and the temporary occupancy probability of the mobile body in the current measurement cycle as the occupancy probability of the mobile body up to the current measurement cycle; and The occupied grid map generating unit of the body shape estimation unit allocates the mobile body occupation probability up to the current measurement period to each cell as the mobile body occupation probability, and generates an occupied grid map.

8. A body shape estimation method performed by a body shape estimation device, the body shape estimation device comprising: a sensor having a transmitting antenna, a receiving antenna and a mixer; Frequency Analysis Unit; and A body shape estimation unit includes an object determination unit, an occupation probability calculation unit, an occupation probability update unit, and an occupation grid map generation unit, wherein: The body shape estimation method comprises the following steps: The transmitting antenna transmits a transmission wave as a plurality of linear frequency modulations with increasing or decreasing frequencies; The receiving antenna receives the transmission wave reflected by a plurality of objects in the vehicle compartment as a reception wave; The mixer mixes the transmission wave and the reception wave to generate a beat signal; The frequency analysis unit uses the received wave to obtain position information of a reflection point where the transmitted wave is reflected by analyzing the beat signal; The body shape estimation unit estimates the body shape of the non-stationary object existing in the vehicle compartment using the position information; The body shape estimation unit calculates the moving body occupancy probability indicating the existence probability of the non-stationary object for each cell of the three-dimensional grid space having a plurality of cells representing the interior of the vehicle using the position information, and estimates the spatial width of the cell where the moving body occupancy probability is greater than a threshold value as the body shape of the non-stationary object existing in the vehicle; The object determination unit of the body shape estimation unit determines whether the detected object is a non-stationary object or a stationary object; When a plurality of non-stationary objects or at least one non-stationary object and at least one stationary object are detected as the plurality of objects, When a non-stationary object is detected, the occupation probability calculation unit of the body shape estimation unit sets the temporary mobile body occupation probability of one or more cells corresponding to the position of the detected non-stationary object to a value close to 1, sets the temporary mobile body occupation probability of cells between the position of the detected non-stationary object and the sensor to a value that asymptotically becomes 0 from the position of the detected non-stationary object toward the sensor, and sets the temporary mobile body occupation probability of cells farther than the position of the detected non-stationary object when observed from the sensor to 0.

5. When a stationary object is detected, the occupation probability calculation unit of the body shape estimation unit sets the temporary mobile body occupation probability of one or more cells corresponding to the position of the detected stationary object and cells farther away than the position of the detected stationary object when observed from the sensor to 0.5, and sets the temporary mobile body occupation probability of cells between the position of the detected stationary object and the sensor to a value that asymptotically becomes 0 from the position of the detected stationary object toward the sensor, When the plurality of non-stationary objects are detected, the occupation probability calculation unit of the body shape estimation unit calculates, for each cell, the product of one or more temporary mobile body occupation probabilities that have been set when the non-stationary objects are detected, and updates the product as a new temporary mobile body occupation probability of the cell. When the at least one non-stationary object and the at least one stationary object are detected, the occupancy probability calculation unit of the body shape estimation unit calculates, for each cell, the product of one or more temporary mobile body occupancy probabilities that have been set when the non-stationary object is detected and one or more temporary mobile body occupancy probabilities that have been set when the stationary object is detected, and calculates the product as a new temporary mobile body occupancy probability of the cell; The occupation probability updating unit of the body shape estimation unit calculates the logarithmic odds of the mobile body occupation probability up to the current measurement cycle by summing the logarithmic odds of the mobile body occupation probability up to the previous measurement cycle and the logarithmic odds of the temporary mobile body occupation probability in the current measurement cycle for each cell, and obtains the mobile body occupation probability up to the current measurement cycle based on the calculated logarithmic odds of the mobile body occupation probability up to the current measurement cycle; and The occupied grid map generating unit of the body shape estimation unit allocates the mobile body occupation probability up to the current measurement period to each cell as the mobile body occupation probability, and generates an occupied grid map.

9. A body shape estimation method performed by a body shape estimation device, the body shape estimation device comprising: a sensor having a transmitting antenna, a receiving antenna and a mixer; Frequency Analysis Unit; and A body shape estimation unit includes an object determination unit, a basic probability distribution calculation unit, a basic probability distribution update unit, a frequency analysis unit, and an occupation grid map generation unit, wherein: The body shape estimation method comprises the following steps: The transmitting antenna transmits a transmission wave as a plurality of linear frequency modulations with increasing or decreasing frequencies; The receiving antenna receives the transmission wave reflected by a plurality of objects in the vehicle compartment as a reception wave; The mixer mixes the transmission wave and the reception wave to generate a beat signal; The frequency analysis unit uses the received wave to obtain position information of a reflection point where the transmitted wave is reflected by analyzing the beat signal; The body shape estimation unit estimates the body shape of the non-stationary object existing in the vehicle compartment using the position information; The body shape estimation unit calculates the moving body occupancy probability indicating the existence probability of the non-stationary object for each cell of the three-dimensional grid space having a plurality of cells representing the interior of the vehicle using the position information, and estimates the spatial width of the cell where the moving body occupancy probability is greater than a threshold value as the body shape of the non-stationary object existing in the vehicle; The object determination unit of the body shape estimation unit determines whether the detected object is a non-stationary object or a stationary object; The basic probability distribution calculation unit of the body shape estimation unit calculates three basic probability distributions including a mobile body occupation basic probability distribution indicating the presence of a non-stationary object, a mobile body non-occupancy basic probability distribution indicating the absence of a non-stationary object, and an unclear basic probability distribution indicating the presence of an unclear non-stationary object, for each measurement period. In the case where a non-stationary object is detected, the basic probability distribution calculation unit of the body shape estimation unit is set so that the temporary moving body occupation basic probability distribution of one or more cells corresponding to the position of the detected non-stationary object is close to 1, and the temporary moving body occupation basic probability distribution of cells other than the one or more cells corresponding to the position of the detected non-stationary object asymptotically becomes 0 as the distance from the position of the detected non-stationary object increases, and for all cells, the temporary unclear basic probability distribution is set to a value obtained by subtracting the temporary moving body occupation basic probability distribution of the cell from 1, When a stationary object is detected, the basic probability distribution calculation unit of the body shape estimation unit is set so that the temporary mobile body non-occupancy basic probability distribution of one or more cells corresponding to the position of the sensor is close to 1, and the temporary mobile body non-occupancy basic probability distribution of cells other than the one or more cells corresponding to the position of the sensor asymptotically becomes 0 from the position of the sensor toward the position of the detected stationary object, and for all cells, the temporary unclear basic probability distribution is set to a value obtained by subtracting the temporary mobile body non-occupancy basic probability distribution of the cell from 1, In a case where the plurality of objects include a plurality of non-stationary objects, or at least one non-stationary object and at least one stationary object, the basic probability distribution calculation unit of the body shape estimation unit combines the temporary mobile body occupation basic probability distribution or the temporary mobile body non-occupancy basic probability distribution and the temporary unclear basic probability distribution related to each object for each cell, and calculates a new temporary mobile body occupation basic probability distribution, a new temporary mobile body non-occupancy basic probability distribution, and a new temporary unclear basic probability distribution for each cell; The basic probability distribution updating unit of the body shape estimation unit combines the three basic probability distributions of each cell in the previous measurement cycle with the new temporary mobile body occupied basic probability distribution, the new temporary mobile body non-occupied basic probability distribution and the new temporary unclear basic probability distribution in the current measurement cycle for each cell, and finally calculates the three basic probability distributions of each cell in the current measurement cycle; The probability conversion unit of the body shape estimation unit converts the final three basic probability distributions of each cell in the current measurement cycle into the mobile body occupancy probability of each cell up to the current measurement cycle; and The occupied grid map generating unit of the body shape estimating unit allocates the mobile body occupation probability up to the current measurement period to each cell as the mobile body occupation probability, and generates an occupied grid map.

Images