Shape estimation system, object detection system, communication system, shape estimation method, object detection method, and communication method

The shape estimation system improves vehicle control by accurately estimating object types and dimensions through wave processing and spectral comparison, addressing the limitations of existing systems in accuracy and noise resistance, and offering cost-effective solutions for vehicle control and safety systems.

WO2024236665A1PCT designated stage expired Publication Date: 2024-11-21ASTEMO LTD
View PDF 9 Cites 0 Cited by

Patent Information

Application Number
PCT/JP2023/017997
Authority / Receiving Office
WO · WO
Patent Type
Applications
Current Assignee / Owner
Filing Date
2023-05-12
Publication Date
2024-11-21

AI Technical Summary

Technical Problem

Current vehicle control systems lack the accuracy in detecting and estimating the type and dimensions of objects around a vehicle, particularly in systems that rely on theoretical calculations and do not provide reliable dimension estimation.

Method used

A shape estimation system that transmits waves to detect objects, processes the received waves to calculate distance and amplitude, resamples data for consistent intervals, converts it to the spatial frequency domain, and uses a database to estimate object shapes by comparing spectra with preset shapes, allowing for accurate type and dimension estimation.

Benefits of technology

Enables high-accuracy estimation of object types and dimensions, resistant to noise, and applicable in various scenarios, including automatic parking and adverse weather conditions, with potential for integration with existing technologies like stereo cameras and Lidar.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure JP2023017997_21112024_PF_FP_ABST
    Figure JP2023017997_21112024_PF_FP_ABST
Patent Text Reader

Abstract

According to the present invention, the distance, between a vehicle (1) and a detection object (3), acquired on the basis of the time from when a transmission wave (2A) is transmitted until when a received wave (2B) scattered / reflected from the detection object (3) is received, and an amplitude value / intensity value of the received wave (2B) corresponding to the distance are stored as detection data, the amplitude value / intensity value of the received wave (2B) corresponding to the distance between the vehicle (1) and the detection object (3) are resampled as sampling data of a desired distance, the resampled data is converted from a spatial domain to a spatial frequency domain, a set of spectra in the spatial frequency domain for a plurality of preset object shapes is saved as a database in association with the object shapes, and the shape of the detection object (3) is estimated on the basis of the spectrum of the detection object (3) and the saved set of spectra. In addition, the object shape corresponding to the spectrum closest to the spectrum of the detection object (3) among the set of spectra is estimated to be the shape of the detection object (3).
Need to check novelty before this filing date? Find Prior Art

Description

Shape estimation system, object detection system, communication system, shape estimation method, object detection method, and communication method

[0001] The present invention relates to a shape estimation system, an object detection system, a communication system, a shape estimation method, an object detection method, and a communication method for a vehicle that detects an object present around the vehicle and estimates the shape of the object.

[0002] Patent Document 1, for example, describes a technology for estimating whether an object present around a vehicle is a human body. Patent Document 1 states, "In an embodiment, a human body detection device acquires detection information from an ultrasonic sensor that receives reflected waves of ultrasonic waves transmitted to a predetermined area. The reflected wave waveform corresponding to the detection information is normalized based on the distance to the object reflecting the ultrasonic waves and attenuation characteristic information related to the ultrasonic waves according to the distance. If a peak in the normalized reflected wave waveform is equal to or greater than a predetermined threshold, it is determined that an object has been detected in the predetermined area. If an object is detected by the object detection unit, a reflected wave waveform of a predetermined range including the peak is extracted from the normalized reflected wave waveform. A predetermined frequency analysis is performed on the reflected wave waveform of the predetermined range without losing time information to generate multiple parameters. A parameter for determining whether the object is a human body is extracted from the multiple parameters. Whether the object is a human body is determined based on the extracted parameter."

[0003] Japanese Patent Application Laid-Open No. 2022-142252

[0004] In vehicle control systems, technology to detect objects around the vehicle is important in order to safely and appropriately control the vehicle to a desired position. In addition, it is desirable to not only recognize the presence or absence of objects around the vehicle, but also to be able to estimate the shape of detected objects.

[0005] For example, in vehicle control systems such as those used in automatic parking and parking assistance, it is possible to switch vehicle control methods by recognizing the shape of an object, such as whether it is a vehicle or a curb / parking stop.

[0006] Incidentally, in a shape estimation system for a vehicle control system, it is desirable to be able to estimate the type of a detected object with high accuracy, and it would be even more convenient if the system could estimate not only the type of the object but also its dimensions.

[0007] However, in the system / device / technology described in Patent Document 1 and the like, there is a possibility that the accuracy of type estimation can be improved based on theoretical calculation results and actual measurement results, but no technology for dimensional estimation is disclosed.

[0008] The present invention has been made in consideration of the above-mentioned problems, and aims to provide a shape estimation system, an object detection system, a communication system, a shape estimation method, an object detection method, and a communication method that are capable of estimating the type and dimensions of objects present around a vehicle with high accuracy.

[0009] The present invention includes a plurality of means for solving the above-mentioned problems, and one example thereof is a shape estimation system for a vehicle, comprising: a storage unit that stores, as detection data, a distance between the vehicle and a detected object and an amplitude value / intensity value of the wave corresponding to the distance, the distance being acquired based on the time from when a transmission unit that transmits a transmission wave to a detection area to be observed to when a reception unit that receives waves scattered / reflected from a detected object present in the detection area receives the wave; a resampling unit that, based on the detection data, resamples the amplitude value / intensity value of the wave corresponding to the distance between the vehicle and the detected object into sampling data of a desired distance; The system comprises a conversion processing unit that converts the resampled data output from the resampling unit from the spatial domain to the spatial frequency domain; a storage unit that stores groups of spectra in the spatial frequency domain for a plurality of preset object shapes as a database in association with the object shapes; and a shape estimation unit that estimates the shape of the detected object based on the spectrum of the detected object output from the conversion processing unit and the group of spectra stored in the storage unit, and the shape estimation unit estimates the object shape corresponding to the spectrum of the group of spectra that is closest to the spectrum of the detected object output from the conversion processing unit as the shape of the detected object.

[0010] According to the present invention, it is possible to estimate the type and size of an object present around a vehicle with high accuracy. Problems, configurations, and effects other than those described above will become apparent from the following description of the embodiments.

[0011] FIG. 1 is a block diagram showing an example of the overall configuration of a vehicle equipped with a shape estimation system according to a first embodiment of the present invention. FIG. 2 is a schematic diagram showing an example of the spatial arrangement of a vehicle, transmitted and received waves, and an object when an object is detected. FIG. 3 is a diagram showing example waveforms of signals detected by a vehicle and signals processed in the shape estimation system when an object is detected. FIG. 4 is an explanatory diagram showing an example of object type estimation performed by the shape estimation system. FIG. 5 is an explanatory diagram showing an example of object size estimation performed by the shape estimation system. FIG. 6 is a flowchart showing an example of a processing flow in the shape estimation system according to the first embodiment of the present invention. FIG. 7 is a diagram showing example waveforms when the technical validity of the shape estimation system according to the first embodiment of the present invention is verified by simulation. FIG. 8 is a diagram showing example waveforms when the technical validity of the shape estimation system according to the first embodiment of the present invention is verified by simulation. FIG. 9 is a diagram showing example waveforms when the technical validity of the shape estimation system according to the first embodiment of the present invention is verified by simulation. FIG. 10 is a diagram showing example estimation results when the technical validity of the shape estimation system according to the first embodiment of the present invention is verified by simulation. FIG. 11 is a diagram showing an example of the relationship between an error rate and noise amount when the technical validity of the shape estimation system according to the first embodiment of the present invention is verified by simulation. FIG. 10 is a block diagram showing an example of the overall configuration of a vehicle equipped with a shape estimation system according to a second embodiment of the present invention, and a communication system.

[0012] Below, embodiments of the shape estimation system, object detection system, and communication system, as well as the shape estimation method, object detection method, and communication method of the present invention will be described with reference to the drawings. Note that the present invention is not limited to the embodiments described below, and various modifications are possible within the scope of the technical concept. Furthermore, in the drawings used in this specification, identical or corresponding components are designated by the same or similar reference numerals, and repeated description of these components may be omitted.

[0013] First Embodiment A first embodiment of a shape estimation system, an object detection system, a communication system, a shape estimation method, an object detection method, and a communication method according to the present invention will be described with reference to FIGS. 1 to 9. FIG.

[0014] First, the overall configuration of a vehicle equipped with a shape estimation system will be described with reference to Fig. 1. Fig. 1 is a block diagram showing an example of the overall configuration of a vehicle equipped with a shape estimation system according to a first embodiment of the present invention.

[0015] 1, a vehicle 1 is configured with an object detection system 100, a vehicle control unit 130, and an HMI (Human Machine Interface) 140. Although the vehicle 1 includes many functional blocks, only functional blocks related to the present invention are shown in the figure.

[0016] The object detection system 100 is composed of a transmitter / receiver 110 and a shape estimation system 120, and is capable of detecting objects present around the vehicle and estimating the shapes of those objects.

[0017] The transmitting / receiving unit 110 has a transmitting unit 111 that transmits a transmission wave 2A (see Figure 2) to a detection area to be observed, a receiving unit 112 that receives a reception wave 2B (see Figure 2) scattered / reflected from a detection object 3 present in the detection area, and a distance calculation unit 113 that calculates the distance between the vehicle 1 and the detection object 3 based on the time from when the transmitting unit 111 transmits the transmission wave 2A to when the receiving unit 112 receives the reception wave 2B. In order to detect objects around the vehicle, the transmitting unit 111 irradiates a wave within a desired range.

[0018] Here, the transmission wave 2A sent by the transmitter 111 and the reception wave 2B received by the receiver 112 may be, for example, an ultrasonic wave, a light wave, a laser which is considered to have particularly high coherence among light waves, or an electric wave. In the following description of the present invention, ultrasonic waves will be used as an example.

[0019] If an object is present within the detection range, the receiver 112 receives the waves scattered / reflected from the object. The distance calculator 113 then inputs the output signal from the receiver 112 and calculates the distance between the vehicle 1 and the object from the time it takes for the wave to be transmitted by the transmitter to be received by the receiver. This time can be considered to be the time required for the wave to travel back and forth between the vehicle 1 and the object.

[0020] The shape estimation system 120 comprises a saving unit 121 , a memory unit 122 , a resampling unit 123 , a conversion processing unit 124 , and a comparison operation unit 125 .

[0021] The storage unit 121 is a storage medium that stores, as a database, groups of spectra in the spatial frequency domain for a plurality of preset object shapes in association with the object shapes, and stores information on groups of spectra in the spatial frequency domain for scattered waves / reflected waves corresponding to a plurality of objects in association with the objects. These groups of spectra may be created, for example, based on simulation or on actual measurement data.

[0022] The memory unit 122 is a recording medium that stores, as detection data, the distance between the vehicle 1 and the detected object 3, and the amplitude value / intensity value of the received wave 2B corresponding to the distance, which are acquired based on the time from when the transmitted wave 2A is transmitted from the transmitting unit 111, which transmits the transmitted wave 2A to the detection area to be observed, to when the received wave 2B is received by the receiving unit 112, which receives the received wave 2B scattered / reflected from the detected object 3 present in the detection area.

[0023] The resampling unit 123 is a part that resamples the amplitude value / intensity value of the received wave 2B corresponding to the distance between the vehicle 1 and the detected object 3 into sampling data of a desired distance based on the detection data recorded in the memory unit 122. This is because when the driver or the system controls the vehicle, it is expected that the vehicle will repeatedly move forward and backward depending on the situation, and the data is not necessarily easy for the shape estimation system 120 to estimate the object shape, so it is necessary to process it so that it becomes the desired data.

[0024] Here, one example of the resampling process is to carry out arithmetic processing to calculate data on amplitude values / intensity values ​​when the distance between the vehicle and the object changes at regular intervals.

[0025] The conversion processing unit 124 is a part that converts the resampled data output from the resampling unit 123 from the spatial domain to the spatial frequency domain, and, for example, inputs the amplitude values / intensity values ​​resampled from the resampling unit 123 and converts, for example, the spatial domain data into spatial frequency domain data. Note that, although the conversion processing to the spatial frequency domain has been given as an example in this explanation, it is not limited to this conversion processing, and it is considered that various conversion methods can be used.

[0026] Comparison operation unit 125 is composed of type estimation unit 126 and size estimation unit 127, and is a unit that estimates the shape of detected object 3 based on the spectrum of detected object 3 output from conversion processing unit 124 and the group of spectra stored in storage unit 121. In this embodiment, comparison operation unit 125 estimates that the object shape corresponding to the spectrum closest to the spectrum of detected object 3 output from conversion processing unit 124, among the group of spectra stored in storage unit 121, is the shape of detected object 3.

[0027] The type estimation unit 126 particularly estimates the type of the detected object 3 as a shape estimation of the detected object 3. Here, in type estimation, for example, it is possible to perform machine learning of the characteristics of each type using data of a group of spectra prepared in advance as a database in the storage unit 121 and information on the corresponding object shapes as training data, and to estimate the type from the measured spectrum of the detected object 3 based on the results of the machine learning. Alternatively, it is also possible to calculate the average value of the difference between a group of spectra prepared in advance as a database in the storage unit 121 and the measured spectrum for each type of object, and estimate the type for which the average value of this difference is approximately the smallest as the type of the object being observed.

[0028] The dimension estimation unit 127 particularly estimates the dimensions of the detected object 3 as an estimation of the shape of the detected object 3. Here, for example, a conceivable method for estimating the dimensions is the so-called scatterometry method, which estimates the dimensions by comparing a database prepared in advance with actual measurement data. However, while with a normal scatterometry method, it is relatively easy to estimate the dimensions because the general shape of the object is known and the measurement method is the same each time, with a vehicle, the approach method, such as the speed at which the vehicle approaches the object and the number of times it repeats forward and backward movement, is different each time, so it is not easy to apply this method.

[0029] In addition, various cases are conceivable for the object of observation, such as a wall, a vehicle, a pole, a curb, etc., and a normal scatterometry method cannot be easily applied. Therefore, as described above, it is conceivable to perform resampling processing in the resampling unit 123 so that the data format is convenient for shape estimation, regardless of the method of approach of the vehicle.

[0030] Furthermore, in the comparison calculation unit 125, for example, the type is narrowed down in advance by the type estimation unit 126, and then the size estimation unit 127 performs size estimation, thereby making it possible to apply the scatterometry method to estimate the size of the object being observed.

[0031] One example of processing in the dimension estimation unit 127 is to extract a group of spectra prepared in advance as a database in the storage unit 121, limited to the type estimated by the type estimation unit 126, calculate the difference between the extracted group of spectra and the measured spectrum, and estimate the dimension at which this difference is approximately minimum as the dimension of the object to be observed. The above method has a high degree of freedom in operation, and furthermore, can be technically applied even to vehicles for which the object to be observed is unknown and for which application of the scatterometry method is extremely difficult, making it possible to estimate the shape of the object to be observed.

[0032] The vehicle control unit 130 receives shape information of the object to be observed from the comparison calculation unit 125, for example, and changes the vehicle control method depending on the type and dimensions of the object.

[0033] The HMI 140 receives, for example, shape information of the object to be observed from the comparison calculation unit 125 and notifies the driver of the object information.

[0034] Figure 2 shows an example of the spatial arrangement of vehicles, transmitted and received waves, and objects during object detection. A transmitted wave 2A emitted from vehicle 1 is scattered / reflected by a detected object 3 and is detected by vehicle 1 as a received wave 2B. During object detection, for example, vehicle 1 may approach detected object 3 from the positive direction toward origin O on the drawing, and it becomes possible to estimate the shape of the object that the vehicle is approaching.

[0035] FIG. 3 shows an example of the waveforms of the detection signal from the vehicle when an object is detected and the signal processed in the shape estimation system.

[0036] Figure 3(a) shows an example of the change in the distance between the vehicle and an object over time. In this example, the vehicle first approaches the object, then moves away and approaches again, which is considered to be a common case when controlling the vehicle to a target parking space in a parking lot.

[0037] Figure 3(b) shows an example of the change over time in the detection level of waves scattered / reflected from an object. The level tends to increase as the object approaches. Note that this example shows an example of a 14-bit detection range, with the detection level saturating at 16,384.

[0038] Figure 3 (c) shows an example plot of the distance between the vehicle and the object on the horizontal axis and the detection level of the waves scattered / reflected from the object on the vertical axis. It can be seen that specific noise occurs around a distance of 1080 mm. Furthermore, random noise occurs throughout the plot, and since this noise affects shape estimation, countermeasures are required. In addition, since the plotted data is not at the desired distance interval, it is desirable to perform resampling processing as described above.

[0039] Therefore, the resampling unit 123 can calculate the sampling data by calculating the average value of the amplitude / intensity values ​​of the received wave 2B included in the detection data within a predetermined distance range from the desired distance.

[0040] For example, (d) in Figure 3 shows an example of the relationship between distance and detection level after resampling processing. Resampling involves resampling the detection level values ​​so that the sampled data is at evenly spaced distances, as described above. When calculating the detection level value at a desired distance, for example, one possible method is to extract detection level values ​​that fall within a predetermined distance range from the desired distance and calculate the average value. This makes it possible to reduce the effects of random noise and specific noise, as shown in (d) in Figure 3.

[0041] 3(e) shows an example of the relationship between distance and detection level after preprocessing in the transformation process. For example, when a discrete Fourier transform is performed as the transformation process, periodic boundary conditions are assumed, so it is desirable that the detection level value at the minimum distance and the detection level value at the maximum distance are approximately the same. To meet this requirement, a window function is generally applied, but this intentionally distorts the original detection waveform, leading to a decrease in the accuracy of object shape estimation.

[0042] Therefore, the conversion processing unit 124 can combine the resampling data output from the resampling unit 123 with data in which the sign of the distance of the resampling data is inverted, and then convert the data from the spatial domain to the spatial frequency domain.

[0043] For example, one possible approach is to combine data with the sign of the distance inverted, as shown in (e) of Fig. 3, and preprocess it to generate data suitable for periodic boundary conditions. Here, the meaning of the data with the sign of the distance inverted corresponds to the data acquired in the negative distance region in Fig. 2, and can be considered to be data acquired when the same detection object 3 is detected by a transmitter / receiver provided at the rear of the vehicle 1, for example. In other words, (e) of Fig. 3 can be considered to be data assuming that data with the same detection level value as the data acquired in the positive distance region can be acquired even in the negative distance region.

[0044] FIG. 3(f) shows an example of a spectrum obtained by subjecting the detection level data shown in FIG. 3(e) to a discrete Fourier transform.

[0045] An example of object type estimation performed by the shape estimation system is shown in Figure 4. When performing type estimation, the type of the object being observed is estimated by, for example, comparing the measured spectrum with a group of spectra prepared in advance as a database.

[0046] The database stores, for example, groups of spectra obtained when dimensions such as height and thickness are changed for each type of object, such as a wall, curbstone, pole, etc. As mentioned above, for example, the spectral features of each type are machine-learned based on the information on the groups of spectra in this database, and this learning data is used to estimate the type of object being observed from the actually measured spectrum.

[0047] Alternatively, the average value of the difference between the group of spectra in the database stored in the storage unit 121 and the spectrum of the detected object 3, which is actual measurement data, can be calculated for each type, and the type for which the average value of the difference is the smallest can be estimated to be the type of the detected object 3. Note that, since an exact minimum value cannot be obtained, an approximate minimum may also be used. Here, the type for which the approximate minimum value is obtained may be estimated to be the type of the object being observed. Note that the group of spectra stored as a database may be prepared by simulation or may be prepared from actual measurement data.

[0048] FIG. 5 shows an example of object size estimation performed by the shape estimation system.

[0049] When estimating the dimensions of the detected object 3 in the dimension estimation unit 127 during shape estimation, it is considered that the type of the object being observed has already been estimated in the type estimation unit 126, and therefore the dimensions of the detected object 3 can be estimated by referring only to data from the spectrum group relating to the identified type of detected object 3.

[0050] For example, the size of the detected object 3 to be observed is estimated to be the size at which the difference between the measured spectrum and the group of spectra stored as a database is approximately minimum. Alternatively, the size may be estimated using machine learning. The group of spectra stored as a database may be prepared by simulation or may be prepared from measured data.

[0051] FIG. 6 shows an example of a processing flow in the shape estimation system according to the first embodiment of the present invention.

[0052] First, in step S601, detection data relating to an object to be observed is acquired. As described above, the detection data may include, for example, the distance between the vehicle and the object, and the amplitude and intensity values ​​of scattered and reflected waves from the object.

[0053] Thereafter, in step S602, the acquired detection data is subjected to a resampling process, where the amplitude / intensity values ​​are resampled so that the data is acquired at equal intervals, for example.

[0054] Thereafter, in step S603, the resampled amplitude / intensity values ​​are subjected to a conversion process, for example, from the spatial domain to the spatial frequency domain, thereby obtaining a spectrum in the spatial frequency domain.

[0055] Thereafter, in step S604, the type of the object being observed is estimated by, for example, comparing the actually measured spectrum with a group of spectra prepared in advance as a database.

[0056] Thereafter, in S605, it is determined whether or not dimension estimation is necessary, and if dimension estimation is not necessary, the process is terminated, whereas if dimension estimation is necessary, dimension estimation is performed in S606 and then the process is terminated. Here, whether or not dimension estimation is necessary can be determined, for example, by comparing the height of the vehicle bumper with the height of the curb or bollard to determine whether or not a collision has occurred.

[0057] In the size estimation in S606, for example, a group of spectra prepared in advance as a database limited to the estimated type is extracted, and the size of the object being observed is estimated by comparing the actually measured spectrum with this group of spectra.

[0058] 7A to 7D show examples of waveforms obtained when verifying the technical validity of the shape estimation system according to Example 1 of the present invention through simulation. Calculations were performed for poles and blocks as objects, with three types of poles (40 mm, 80 mm, and 200 mm in diameter) and three types of blocks (100 mm, 200 mm, and 300 mm in height).

[0059] Figure 7A shows the relationship between the level values ​​of scattered waves / reflected waves calculated by simulation and the distance between the vehicle and the object. As shown in Figure 7A, the level values ​​of scattered waves / reflected waves generally increase as the distance between the vehicle and the object decreases, but for blocks, the level values ​​of scattered waves / reflected waves tend to decrease when the vehicle gets too close because they fall into a blind spot.

[0060] Therefore, the comparison calculation unit 125 can estimate the shape of the detected object 3 after normalizing the spectrum group and all of the spectra of the detected object 3 output from the conversion processing unit 124 by their respective maximum values.

[0061] Figure 7B shows a set of spectra after Fourier transforming the level values ​​of the scattered waves / reflected waves calculated by the simulation shown in Figure 7A into the spatial frequency domain. Note that in this figure, the maximum value is normalized to 1. The principle was verified on the assumption that this set of spectra was stored as a database.

[0062] Figure 7C shows the relationship between the simulated scattered wave / reflected wave level values ​​and the distance between the vehicle and an object. This data was created to be equivalent to actual measurements by superimposing noise on the data calculated by the simulation shown in Figure 7A. In the principle verification, noise resistance was investigated by changing the amount of added noise, and this figure shows an example.

[0063] Figure 7D shows the spectrum group after Fourier transforming the level values ​​of the simulated scattered waves / reflected waves shown in Figure 7C into the spatial frequency domain. By comparing this data with the spectrum group prepared as a database by prior simulation shown in Figure 7B, we verified whether the shape of an object could be accurately estimated.

[0064] Fig. 8 shows an example of the estimation results when verifying the technical validity of the shape estimation system according to Example 1 of the present invention through simulation. This result shows the estimation results for the data shown in Fig. 7C(d). As is clear from Fig. 8, it can be seen that the type and dimensions are all correctly estimated, demonstrating that by utilizing the present invention, it is possible to estimate the shape of an object being observed.

[0065] 9 shows an example of the relationship between the error rate and the amount of noise when the technical validity of the shape estimation system according to Example 1 of the present invention is verified by simulation. Here, the amount of noise is expressed as a percentage, assuming that the peak value of the level values ​​of the scattered wave / reflected wave at a block height of 100 mm is 100%.

[0066] It is particularly noteworthy that the type determination is accurate and the type is estimated correctly even when a large amount of noise is superimposed, and it can be said that the technology of the present invention enables type estimation with extremely high accuracy. Also, in size estimation, the size is estimated correctly and without error when the amount of noise is less than 20%, and the technology of the present invention enables accurate estimation of the size of the object being observed.

[0067] Furthermore, by taking advantage of the tendency of noise to appear on the high frequency side and performing filtering processing such as a low-pass filter to separate the signal component from the noise component, it may be possible to further improve the accuracy of dimension estimation.

[0068] Next, the effects of this embodiment will be described.

[0069] The shape estimation system 120 for the vehicle 1 according to the first embodiment of the present invention described above includes a storage unit 122 that stores, as detection data, the distance between the vehicle 1 and the detected object 3 and the amplitude / intensity value of the received wave 2B corresponding to the distance, which are acquired based on the time from when the transmitted wave 2A is transmitted from a transmitting unit 111 that transmits the transmitted wave 2A to a detection area to be observed until when the received wave 2B is received by a receiving unit 112 that receives the received wave 2B scattered / reflected from the detected object 3 present in the detection area; a resampling unit 123 that resamples the amplitude / intensity value of the received wave 2B corresponding to the distance between the vehicle 1 and the detected object 3 based on the detection data, to sampling data of a desired distance; The present invention is provided with a conversion processing unit 124 that converts the resampled data output from 123 from the spatial domain to the spatial frequency domain, a storage unit 121 that stores groups of spectra in the spatial frequency domain for a plurality of preset object shapes as a database in association with the object shapes, and a comparison calculation unit 125 that estimates the shape of the detected object 3 based on the spectrum of the detected object 3 output from conversion processing unit 124 and the group of spectra stored in storage unit 121, and comparison calculation unit 125 estimates that the object shape corresponding to the spectrum of the group of spectra stored in storage unit 121 that is closest to the spectrum of the detected object 3 output from conversion processing unit 124 is the shape of the detected object 3.

[0070] According to this embodiment, by utilizing not only actual measurement data but also a database prepared in advance using simulations, etc., it is possible to estimate the type / dimensions of the object being observed with higher accuracy than the estimation techniques disclosed in the past. Furthermore, it has been shown through principle verification that error-free estimation is possible even when the amount of noise is large, which is an advantage of enabling high noise resistance.

[0071] Although it is possible to estimate the type and size using a stereo camera / multi-camera, the technology of the present invention has the advantage of being usable at night or in bad weather, which is one of the challenges of the above camera technology.Furthermore, by fusing the detection information from the stereo camera / multi-camera with the detection information from the technology of the present invention, it is expected that the accuracy of the system will be improved.

[0072] Furthermore, although sensing using Lidar has been gaining popularity in recent years, the technology of the present invention can also utilize relatively inexpensive ultrasonic sonar, which has the advantage of enabling systems to be constructed at low cost.

[0073] Furthermore, while multiple transceivers are often used for sensing dimension estimation, the technology of the present invention can be realized with a single transceiver, and can be realized by improving the signal processing unit while maintaining the conventionally proposed transceiver arrangement configuration, offering the advantages of ease of implementation and excellent cost performance.

[0074] Furthermore, there is an advantage that additional information such as the type and dimensions of an object can be notified to the driver not only in vehicle control systems but also in clearance sonars and the like that have been widely used in the past.

[0075] Furthermore, while conventional scatterometry methods acquire spectral data by changing the wave frequency itself, the technology of the present invention utilizes waves of a single frequency and achieves this using spectral data of spatial frequencies, which has the advantage of being simple in configuration.

[0076] In addition, the resampling unit 123 calculates the sampling data by calculating the average value of the amplitude / intensity values ​​of the received wave 2B included in the detection data that fall within a predetermined distance range from the desired distance, thereby reducing the effects of random noise and specific noise.

[0077] Furthermore, the conversion processing unit 124 combines the resampling data output from the resampling unit 123 with data in which the sign of the distance of the resampling data is inverted, and then performs a conversion process from the spatial domain to the spatial frequency domain, thereby preprocessing the data into data suitable for periodic boundary conditions.Since there is no longer a need to intentionally distort the original detection waveform when performing a discrete Fourier transform as a conversion process, improved accuracy can be achieved.

[0078] Furthermore, the comparison calculation unit 125 normalizes the spectrum group and all of the spectra of the detected object 3 output from the conversion processing unit 124 by their respective maximum values, and then estimates the shape of the detected object 3, thereby being able to handle both increases and decreases in the level values ​​of the scattered waves / reflected waves and realizing spectrum processing of detected objects 3 of various heights.

[0079] Furthermore, by estimating the type of detected object 3 as a shape estimate of the detected object 3, the comparison calculation unit 125 can narrow down the types of objects whose dimensions are to be estimated in the subsequent dimension estimation process, thereby speeding up processing and further reducing the probability of erroneous judgment.

[0080] Furthermore, by estimating that the type of the detected object 3 is at least a curb / car stop, the comparison calculation unit 125 can estimate the shape of the curb / car stop with high accuracy, which was previously difficult.

[0081] Furthermore, the comparison calculation unit 125 performs machine learning to learn the characteristics of each type of object using the spectrum group as training data, and estimates the type from the spectrum of the detected object 3 based on the results of the machine learning, thereby efficiently improving the accuracy of type estimation.

[0082] Furthermore, the comparison calculation unit 125 calculates the average value of the difference between the spectrum group and the spectrum of the detected object 3 for each type, and estimates that the type for which the average value of the difference is approximately the smallest is the type of the detected object 3, thereby improving the accuracy of type estimation.

[0083] Furthermore, the comparison operation unit 125 estimates the dimensions of the detected object 3 as a shape estimate of the detected object 3, thereby making it possible to more accurately estimate what kind of object the detected object 3 is.

[0084] Furthermore, when estimating the shape of the detected object 3, the comparison calculation unit 125 estimates the type of the detected object 3, and then estimates the dimensions of the detected object 3 by referring only to the data in the spectrum group that relates to the identified type of detected object 3, thereby enabling the estimation process to be performed more quickly and with higher accuracy.

[0085] Furthermore, the comparison calculation unit 125 can further improve the accuracy of size estimation by estimating that the size at which the difference between the group of spectra and the spectrum of the detected object 3 is smallest is the size of the detected object 3.

[0086] Second Embodiment A shape estimation system, an object detection system, a communication system, a shape estimation method, an object detection method, and a communication method according to a second embodiment of the present invention will be described with reference to Fig. 10. Fig. 10 is a block diagram showing an example of the overall configuration of a vehicle equipped with a shape estimation system according to the second embodiment of the present invention, and a communication system.

[0087] As shown in FIG. 10, this embodiment differs from the first embodiment in that a communication system 4 capable of communicating with the storage unit 121 in the vehicle 1 is added.

[0088] The communication system 4 is a system that communicates with a vehicle 1 equipped with an object detection system 100 via a network, and is composed of, for example, a database update unit 410 that updates the database in the database storage unit 420, a database storage unit 420 that stores the database used by the vehicle 1 when estimating the shape of a detected object 3, and a database transmission unit 430 that transmits the database stored in the database storage unit 420 to the storage unit 121 of the vehicle 1.

[0089] The database update unit 410 updates, as necessary, the groups of spectra in the spatial frequency domain relating to various objects that should be prepared in advance as a database for the vehicle. For example, when preparing this database through simulation, it is possible to update the database when the simulation accuracy can be improved compared to the initial state or when errors in the initial database become a problem.

[0090] The database storage unit 420 receives the updated database from the database update unit 410 and stores it.

[0091] The database transmission unit 430 receives the updated database from the database storage unit 420 and updates the database stored in the storage unit 121 of the vehicle 1. When the database update unit 410 updates the database, it is desirable that this database transmission unit 430 transmits the updated database to the vehicle 1. Here, a configuration that does not involve the database storage unit 420 can also be realized, but it is desirable that the updated database be managed by the database storage unit 420 in order to make it possible to track the updated database.

[0092] For this reason, the database stored in the storage unit 121 can be updated via the communication system 4 or a physical medium.

[0093] Other configurations and operations are substantially the same as those of the shape estimation system, object detection system, and communication system, as well as the shape estimation method, object detection method, and communication method of the first embodiment, and details thereof will be omitted.

[0094] The shape estimation system, object detection system, and communication system, as well as the shape estimation method, object detection method, and communication method of Example 2 of the present invention, can also achieve effects substantially similar to those of the shape estimation system, object detection system, and communication system, as well as the shape estimation method, object detection method, and communication method of Example 1 described above.

[0095] In addition, since the spectrum group of scattered waves / reflected waves prepared in advance as a database can be updated, there is the advantage that it can be used when the accuracy of estimating the object shape becomes an issue or when the estimation accuracy can be improved.

[0096] Although an example in which the database is updated via a communication system has been shown, it is of course also possible to update the database via a physical medium.

[0097] <Others> The above describes embodiments of the present invention. Note that the present invention is not limited to the above-described embodiments, and includes various modifications and equivalent configurations within the spirit of the appended claims. For example, the above-described embodiments have been described in detail to clearly explain the present invention, and the present invention is not necessarily limited to configurations including all of the described configurations. Furthermore, part of the configuration of one embodiment may be replaced with the configuration of another embodiment. Furthermore, the configuration of another embodiment may be added to the configuration of one embodiment. Furthermore, part of the configuration of each embodiment may be added, deleted, or replaced with other configurations.

[0098] Furthermore, the aforementioned configurations, functions, processing units, processing means, etc. may be realized in part or in whole in hardware, for example by designing them as integrated circuits, or may be realized in software by having a processor interpret and execute a program that realizes each function.

[0099] Information such as programs, tables, and files that realize each function can be stored in a storage device such as a memory, hard disk, or SSD (Solid State Drive), or in a recording medium such as an IC card, SD card, or DVD.

[0100] In addition, the control lines and information lines shown are those that are considered necessary for explanation, and do not necessarily represent all the control lines and information lines that are necessary for implementation. In reality, it can be assumed that almost all components are interconnected.

[0101] 1... Vehicle 2A... Transmitted wave 2B... Received wave 3... Detected object 4... Communication system 100... Object detection system 110... Transmitting / receiving unit 111... Transmitting unit 112... Receiving unit 113... Distance calculation unit 120... Shape estimation system 121... Saving unit 122... Memory unit 123... Resampling unit 124... Conversion processing unit 125... Comparison calculation unit (shape estimation unit) 126... Type estimation unit 127... Dimension estimation unit 130... Vehicle control unit 140... HMI 410... Database update unit 420... Database saving unit 430... Database transmission unit

Claims

1. A shape estimation system for a vehicle, comprising: a memory unit that stores as detection data a distance between the vehicle and the detected object and an amplitude / intensity value of the wave corresponding to the distance, the distance being acquired based on the time from when a transmitting unit that transmits a transmission wave to a detection area to be observed transmits the transmission wave to when a receiving unit that receives waves scattered / reflected from a detected object present in the detection area receives the wave; a resampling unit that resamples the amplitude / intensity value of the wave corresponding to the distance between the vehicle and the detected object based on the detection data to sampling data of a desired distance; a conversion processing unit that converts the resampling data output from the resampling unit from a spatial domain to a spatial frequency domain; a storage unit that associates a group of spectra in the spatial frequency domain for a plurality of predetermined object shapes with the object shapes and stores them as a database; and a shape estimation unit that estimates a shape of the detected object based on the spectrum of the detected object output from the conversion processing unit and the group of spectra stored in the storage unit, and the shape estimation unit estimates the object shape corresponding to the spectrum of the group of spectra that is closest to the spectrum of the detected object output from the conversion processing unit to be the shape of the detected object. Shape estimation system.

2. A shape estimation system according to claim 1, wherein the resampling unit calculates the sampling data by calculating the average value of the amplitude / intensity values ​​of the waves included in the detection data that are within a specified distance range from the desired distance.

3. A shape estimation system according to claim 1, wherein the conversion processing unit combines the resampling data output from the resampling unit with data obtained by inverting the sign of the distance of the resampling data, and then converts the combined data from the spatial domain to the spatial frequency domain.

4. A shape estimation system according to claim 1, wherein the shape estimation unit normalizes the group of spectra and all of the spectra of the detected object output from the conversion processing unit by their respective maximum values, and then estimates the shape of the detected object.

5. A shape estimation system according to claim 1, wherein the shape estimation unit estimates a type of the detected object as a shape estimate of the detected object.

6. A shape estimation system according to claim 5, wherein the shape estimation unit estimates at least a curb / car stop as the type of the detected object.

7. A shape estimation system according to claim 5, wherein the shape estimation unit performs machine learning to learn characteristics of each type of object using the group of spectra as training data, and estimates the type of the object from the spectrum of the detected object based on the results of the machine learning.

8. A shape estimation system according to claim 5, wherein the shape estimation unit calculates an average value of the difference between the group of spectra and the spectrum of the detected object for each type, and estimates that the type for which the average value of the difference is the smallest is the type of the detected object.

9. A shape estimation system according to claim 1, wherein the shape estimation unit estimates dimensions of the detected object as an estimate of the shape of the detected object.

10. A shape estimation system as described in claim 1, wherein the shape estimation unit, when estimating the shape of the detected object, estimates the type of the detected object, and then estimates the dimensions of the detected object by referring only to data from the group of spectra relating to the identified type of the detected object.

11. A shape estimation system according to claim 9, wherein the shape estimation unit estimates the dimensions of the detected object as the dimensions that minimize the difference between the group of spectra and the spectrum of the detected object.

12. A shape estimation system according to claim 1, wherein the group of spectra is created based on a simulation.

13. A shape estimation system according to claim 1, wherein the group of spectra is created based on actual measurement data.

14. A shape estimation system according to claim 1, wherein the database stored in the storage unit is updateable via a communication system or a physical medium.

15. A shape estimation system according to claim 1, wherein the transmission wave sent by the transmitting unit and the wave received by the receiving unit are ultrasonic waves.

16. An object detection system for a vehicle, comprising: a transmitting unit that transmits a transmission wave to a detection area to be observed; a receiving unit that receives waves scattered / reflected from a detection object present in the detection area; a distance calculation unit that calculates the distance between the vehicle and the detection object based on the time from when the transmission wave is transmitted from the transmitting unit to when the wave is received by the receiving unit; and a shape estimation system described in any one of claims 1 to 15.

17. A communication system which communicates via a network with a vehicle equipped with the object detection system described in claim 16, comprising: a database storage unit which stores a database used by the vehicle when estimating the shape of a detected object; a database update unit which updates the database in the database storage unit; and a database transmission unit which transmits the database stored in the database storage unit to the storage unit of the vehicle, wherein the database transmission unit transmits the updated database to the vehicle when the database update unit updates the database.

18. A shape estimation method for a vehicle, comprising: a storage step of storing, as detection data, the distance between the vehicle and the detected object, and the amplitude / intensity value of the wave corresponding to the distance, which are acquired based on the time from when a transmission unit that transmits a transmission wave to a detection area to be observed transmits the transmission wave to when a receiving unit that receives waves scattered / reflected from a detected object present in the detection area receives the wave; a resampling step of resampling the amplitude / intensity value of the wave corresponding to the distance between the vehicle and the detected object to sampling data of a desired distance based on the detection data; a conversion step of converting the resampling data output in the resampling step from a spatial domain to a spatial frequency domain; a storage step of storing, as a database, a group of spectra in the spatial frequency domain for a plurality of preset object shapes in association with the object shapes; and a shape estimation step of estimating the shape of the detected object based on the spectrum of the detected object output in the conversion step and the group of spectra stored in the storage step, In the shape estimation step, the object shape corresponding to a spectrum among the group of spectra stored in the storage step that is closest to the spectrum of the detected object output in the conversion processing step is estimated to be the shape of the detected object.

19. An object detection method comprising: a transmitting step of transmitting waves to a detection area to be observed; a receiving step of receiving waves scattered / reflected from a detection object present in the detection area; a distance calculation step of determining the distance between the vehicle and the detection object based on the time from when the waves are transmitted in the transmitting step to when the waves are received in the receiving step; and each step of the shape estimation method described in claim 18.

20. A communication system that communicates via a network with a vehicle having a processing device that executes the shape estimation method described in claim 18, comprising: a database storage step for storing a database used by the vehicle when estimating the shape of a detected object; a database update step for updating the database; and a database transmission step for transmitting the database to a storage unit of the processing device, wherein in the database transmission step, when the database is updated in the database update step, the updated database is transmitted to the storage unit.

Citation Information

Patent Citations

  • Pavement type identifying method and device based on laser radar

    CN104463217A

  • Method and apparatus for detection of target signal

    JP2000171545A

  • Image recognition device and its method

    JP2003178305A

  • Object detection device and system, and object detection method

    JP2005106478A

  • Radar image processor

    JP2010048651A