Learning device, learning method, and program

Abstract

To provide a technology that makes it possible to improve the accuracy of recognizing the type of vehicle.SOLUTION: Provided is a learning device comprising: a vehicle information acquisition unit that acquires vehicle information that includes the position information of a first vehicle and type information corresponding to the type of the first vehicle; a sensor information acquisition unit that acquires sensor information that includes the distance and direction from the sensor and the representative value of received electric power in the sensor; an association processing unit that associates the type information and the representative value, on the basis of the position information and the distance and direction; and a learning unit that learns parameters for identifying the type of a second vehicle on the basis of the representative value and the type information.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a learning device, a learning method, and a program. [Background technology]

[0002] In recent years, a technology has been known that identifies the type of vehicle based on sensor data obtained by a sensor. Examples of the vehicle type include a type in which the vehicle is a large vehicle, a type in which the vehicle is not a large vehicle, etc. For example, a technology has been disclosed that identifies the type of vehicle based on the reception strength by a radar of a signal that is irradiated by the radar and reflected by an object (see, for example, Patent Document 1). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2020-3336 Summary of the Invention [Problem to be solved by the invention]

[0004] However, it is desirable to provide a technique that enables improved accuracy in identifying vehicle types.

[0005] Therefore, the present invention has been made in consideration of the above problems, and an object of the present invention is to provide a technique that makes it possible to improve the accuracy of identifying vehicle types. [Means for solving the problem]

[0006] In order to solve the above problem, according to one aspect of the present invention, there is provided a vehicle information acquisition unit that acquires vehicle information including position information of a first vehicle and type information according to the type of the first vehicle; a sensor information acquisition unit that acquires sensor information including a distance and a direction from a sensor and a representative value of received power at the sensor; and a vehicle information acquisition unit that acquires vehicle information including a distance and a direction from a sensor and a representative value of received power at the sensor based on the position information, the distance, and the direction. calculating a position of an object, and if a difference between the position of the object and the position of the first vehicle indicated by the position information is smaller than a first threshold value, A learning device is provided, which includes a linking processing unit that links the type information with the representative value, and a learning unit that learns parameters for identifying the type of a second vehicle based on the representative value and the type information.

[0007] The type of the first vehicle may include a type in which the first vehicle is a large vehicle or a type in which the first vehicle is other than a large vehicle, and the type of the second vehicle may include a type in which the second vehicle is a large vehicle or a type in which the second vehicle is other than a large vehicle.

[0008] The type information may include at least one of the model and length of the first vehicle.

[0009] The learning unit may determine a type of the first vehicle corresponding to the type information, and learn the parameters using the type of the first vehicle as teacher data and the representative value as training data.

[0010] The representative value may be a maximum value, a median value, or an average value of the received power.

[0011] The learning unit may learn the parameter based on the representative value, the distance, and the type information.

[0013] The vehicle information may include the speed of the first vehicle, and the sensor information may include the speed of the object, and the linking processing unit may link the type information with the representative value when a difference between the position of the object and the position of the first vehicle is smaller than the first threshold and when a difference between the speed of the object and the speed of the first vehicle is smaller than a second threshold.

[0014] The vehicle information may include the length of the first vehicle, and the sensor information may include the length of the object along a predetermined vehicle driving direction, and the linking processing unit may link the type information with the representative value when the difference between the position of the object and the position of the first vehicle is smaller than the first threshold value and when the difference between the length of the object and the length of the first vehicle is smaller than a third threshold value.

[0015] The vehicle information may include a first time when the position was detected, and the sensor information may include a second time when the distance and the direction were detected, and the linking processing unit may link the type information and the representative value when the difference between the position of the object and the position of the first vehicle is smaller than the first threshold value and when the difference between the first time and the second time is smaller than a fourth threshold value.

[0016] According to an aspect of the present invention, a method for detecting a vehicle speed includes acquiring vehicle information including position information of a first vehicle and type information corresponding to the type of the first vehicle, acquiring sensor information including a distance and a direction from a sensor and a representative value of received power at the sensor, and calculating a vehicle speed based on the position information, the distance, and the direction. calculating a position of an object, and if a difference between the position of the object and the position of the first vehicle indicated by the position information is smaller than a first threshold value, A learning method is provided, which includes linking the type information with the representative value, and learning a parameter for identifying the type of a second vehicle based on the representative value and the type information.

[0017] According to an aspect of the present invention, a computer includes a vehicle information acquisition unit that acquires vehicle information including position information of a first vehicle and type information corresponding to the type of the first vehicle, a sensor information acquisition unit that acquires sensor information including a distance and a direction from a sensor and a representative value of received power at the sensor, and a vehicle information acquisition unit that acquires vehicle information including a distance and a direction from a sensor and a representative value of received power at the sensor based on the position information, the distance, and the direction. calculating a position of an object, and if a difference between the position of the object and the position of the first vehicle indicated by the position information is smaller than a first threshold value,A program is provided that functions as a learning device that includes a linking processing unit that links the type information with the representative value, and a learning unit that learns parameters for identifying the type of a second vehicle based on the representative value and the type information. [Effects of the Invention]

[0018] As described above, according to the present invention, it is possible to improve the accuracy of identifying vehicle types. [Brief explanation of the drawings]

[0019] [Figure 1] 1 is a diagram illustrating an example of the configuration of a communication system according to an embodiment of the present invention. [Figure 2] 5 is a flowchart illustrating an example of the operation of a signal processing unit of the radar device according to the embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0020] Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. In this specification and drawings, components having substantially the same functional configurations are designated by the same reference numerals, and redundant explanations will be omitted.

[0021] In this specification and drawings, multiple components having substantially the same functional configuration are distinguished by adding different numbers after the same reference numeral. Similar components in different embodiments are distinguished by adding different letters after the same reference numeral. However, when there is no particular need to distinguish between multiple components having substantially the same functional configuration, only the same reference numeral is used.

[0022] [Summary] Next, an outline of an embodiment of the present invention will be described.

[0023] In recent years, a technology for identifying a vehicle type based on sensor data obtained by a sensor has been known. Examples of the vehicle type include a type in which the vehicle is a large vehicle, a type in which the vehicle is not a large vehicle, etc. For example, a technology for identifying a vehicle type based on the reception strength of a signal emitted by a radar and reflected by an object has been disclosed.

[0024] This specification mainly proposes a technology that enables improving the accuracy of identifying vehicle types. More specifically, the reception strength of light emitted by the radar and reflected by an object by the radar can vary depending on the directivity of the antenna, the orientation of the antenna, the environment in which the antenna is installed, etc. Therefore, it is difficult to improve the accuracy of identifying vehicle types simply based on the reception strength of the radar.

[0025] Therefore, this specification proposes a technology for automatically adjusting parameters for identifying vehicle types to appropriate values ​​depending on the directivity of the antenna, the orientation of the antenna, the environment in which the antenna is installed, etc. By automatically adjusting the parameters for identifying vehicle types to appropriate values, it becomes possible to improve the accuracy of identifying vehicle types using the adjusted parameters.

[0026] The outline of the embodiment of the present invention has been described above.

[0027] [Details of the embodiment] Next, details of the embodiment of the present invention will be described.

[0028] (Configuration explanation) In the technology according to the embodiment of the present invention, a radar device will be described. The radar device can function as a learning device, which is an example of an information processing device. The radar device is equipped with an antenna, which is an example of a sensor. In the following description, a case will be described in which an FMCW (Frequency Modulated Continuous Wave) radar is used as an example of the sensor. However, as will be described in a later modified example, the sensor equipped in the radar device is not limited to an FMCW radar.

[0029] (Example of communication system configuration) Fig. 1 is a diagram showing an example of the configuration of a communication system according to an embodiment of the present invention. As shown in Fig. 1, the communication system according to the embodiment of the present invention includes a radar device 101, a vehicle 103, and a roadside device 102. The vehicle 103 and the roadside device 102 can communicate with each other by radio signals. Furthermore, the roadside device 102 and the radar device 101 can communicate with each other via a LAN (Local Area Network) cable 113.

[0030] (Example of vehicle configuration) The vehicle 103 travels on a flat road. For example, the vehicle 103 travels on a flat road in a predetermined vehicle travel direction. The vehicle 103 is equipped with an on-board device 104. The on-board device 104 is connected to a vehicle antenna 111, and transmits information about the vehicle (first vehicle) (hereinafter also referred to as "vehicle information") via the vehicle antenna 111. The timing at which the on-board device 104 transmits the vehicle information is not particularly limited. As an example, the on-board device 104 may transmit the vehicle information periodically.

[0031] As will be described later, there are no limitations on the specific types of information included in the vehicle information transmitted by the on-board device 104. Here, as an example, it is mainly assumed that the vehicle information includes vehicle location information and type information according to the type of vehicle.

[0032] The position of the vehicle can be detected by a position detection sensor. Typically, a GPS (Global Positioning System) sensor can be used as the position detection sensor. However, the position detection sensor is not limited to a GPS sensor.

[0033] Furthermore, the vehicle type may include a type indicating that the vehicle is a large vehicle or a type other than a large vehicle. Here, it is mainly assumed that the type information includes both the type of vehicle (hereinafter also referred to as "vehicle type") and the length of the vehicle in the longitudinal direction (hereinafter also referred to as "vehicle length"). However, the type information may include at least either the vehicle type or the vehicle length. The vehicle type and vehicle length may be set in advance in the onboard device 104.

[0034] (Example of roadside unit configuration) The roadside device 102 is installed at a predetermined position relative to the road. The roadside device 102 is connected to a roadside device antenna 112 and receives vehicle information from the onboard device 104 via the roadside device antenna 112. The roadside device 102 transmits the vehicle information received from the onboard device 104 to the radar device 101 via a LAN cable 113.

[0035] (Radar device configuration example) The radar device 101 includes a signal generating unit 105 (synthesizer), a transmitting antenna 106, a receiving antenna 107, a mixer 108, an ADC (Analog-Digital Converter) 109, a signal processing unit 110, a vehicle information storage unit 114, and a parameter generating unit 115.

[0036] The signal processing unit 110 and the parameter generating unit 115 include an arithmetic unit (processor) such as a CPU (Central Processing Unit), and their functions can be realized by the arithmetic unit expanding a program stored in a ROM (Read Only Memory) into a RAM and executing it. In this case, a computer-readable recording medium on which the program is recorded can also be provided. Alternatively, these blocks can be configured with dedicated hardware or a combination of multiple pieces of hardware.

[0037] The vehicle information storage unit 114 is configured by a memory, and stores the vehicle information received by the radar device 101 from the roadside device 102 via the LAN cable 113 .

[0038] The signal generating unit 105 generates a signal. More specifically, the signal generating unit 105 generates a signal whose frequency changes over time, called a chirp signal. The signal generating unit 105 is connected to the transmitting antenna 106 and outputs the generated signal to the transmitting antenna 106. The transmitting antenna 106 transmits the signal generated by the signal generating unit 105. When the chirp signal is transmitted by the transmitting antenna 106, the chirp signal is reflected by an object. The receiving antenna 107 receives the signal reflected by the object. The object may include a vehicle, but may also include objects other than vehicles.

[0039] Mixer 108 is connected to each of transmitting antenna 106 and receiving antenna 107, and receives as input a signal transmitted by transmitting antenna 106 and a signal received by receiving antenna 107. Mixer 108 generates an IF (Intermediate Frequency) signal based on the signal transmitted by transmitting antenna 106 and the signal received by receiving antenna 107.

[0040] More specifically, mixer 108 multiplies the transmission signal by the reception signal to generate an IF signal having a frequency corresponding to the difference in frequency between the transmission signal and the reception signal. Mixer 108 outputs the generated IF signal to ADC 109. ADC 109 converts the format of the IF signal output from mixer 108 from analog to digital, and outputs the digital IF signal to signal processing unit 110.

[0041] The signal processing unit 110 performs signal processing on the digital IF signal output from the ADC 109 to detect the distance from the radar device 101 to the object, the speed of the object, and the direction of the object relative to the position of the radar device 101 (hereinafter, the "direction" of the object may also be referred to as the "angle" of the object). Here, a brief description will be given of an example of detecting the distance from the radar device 101 to the object, the speed of the object, and the direction of the object relative to the position of the radar device 101.

[0042] First, the signal processing unit 110 performs a distance FFT on the IF signal to obtain a frequency spectrum. Based on the frequency spectrum obtained in this manner, the signal processing unit 110 detects the distance between the radar device 101 and the object. More specifically, the signal processing unit 110 extracts a peak from the frequency spectrum. Then, the signal processing unit 110 converts the frequency at which the peak occurs into a distance, thereby detecting the distance from the radar device 101 to the object.

[0043] To detect the speed of an obstacle, a collection of time-continuous chirp signals called a chirp frame is used. The chirp frames are transmitted by the transmitting antenna 106, and an IF signal for each chirp is obtained by the mixer 108. The signal processing unit 110 performs a range FFT on each IF signal to obtain a frequency spectrum for each IF signal. Peaks indicating the presence of an object appear in the frequency spectrum.

[0044] The signal processing unit 110 then performs an FFT (velocity FFT) on all of the data for peaks that exist at similar positions within the chirp frame. As a result, the signal processing unit 110 obtains an angular frequency spectrum. The signal processing unit 110 detects the velocity of the object based on the angular frequency spectrum. More specifically, the signal processing unit 110 extracts peaks from the angular frequency spectrum. The signal processing unit 110 then converts the angular frequency at which the peak occurs into a velocity, thereby detecting the velocity of the object.

[0045] To detect the direction of an object relative to the position of the radar device 101, multiple receiving antennas 107 are used. The signal processing unit 110 further applies an FFT (angle FFT) to the results of the velocity FFT corresponding to each of the multiple receiving antennas 107. In this way, the signal processing unit 110 obtains a spectrum.

[0046] The signal processing unit 110 extracts a peak from the spectrum thus obtained. This peak corresponds to the phase difference between the multiple receiving antennas 107. The signal processing unit 110 then converts this phase difference into an angle to detect the direction of the object relative to the position of the radar device 101.

[0047] The above has been a brief description of an example of detecting the distance from the radar device 101 to an object, the object's speed, and the object's direction based on the position of the radar device 101. In addition, the radar device 101 measures the received power of a signal reflected by the object and received by the receiving antenna 107 for each same object. The signal processing unit 110 then calculates a representative value of the received power for each object. For example, objects whose position change during the measurement interval by the radar device 101 is smaller than a predetermined movement amount may be considered to be the same object. The representative value of the received power may be the maximum value, median value, or average value of the received power.

[0048] The signal processing unit 110 functions as an example of a sensor information acquisition unit that acquires, for each object, the distance from the radar device 101 to the object, the direction of the object relative to the position of the radar device 101, and a representative value of the received power of a signal reflected by the object as sensor information. Note that a case in which the sensor information also includes the speed of the object will be described in detail later in a modified example. Furthermore, the signal processing unit 110 functions as an example of a vehicle information acquisition unit that acquires vehicle information from the vehicle information storage unit 114.

[0049] Furthermore, the signal processing unit 110 functions as an example of a linking processing unit that links vehicle information with sensor information. Here, it is mainly assumed that the signal processing unit 110 links type information included in the vehicle information with a representative value of received power included in the sensor information, based on vehicle position information included in the vehicle information and the distance and direction included in the sensor information.

[0050] More specifically, the position information of the radar device 101 may be measured in advance and set in the radar device 101. In such a case, the signal processing unit 110 can detect the position of the object relative to the position of the radar device 101 based on the previously set position information of the radar device 101 and the distance and direction included in the sensor information.

[0051] Then, the signal processing unit 110 calculates the difference between the object's position and the vehicle's position indicated by the vehicle's position information. If the difference between the object's position and the vehicle's position is smaller than a preset first threshold, the signal processing unit 110 determines that the object and the vehicle are the same subject, and associates the vehicle information with the sensor information. That is, the signal processing unit 110 associates the vehicle type information with the representative value of the received power.

[0052] The parameter generating unit 115 functions as an example of a learning unit that learns parameters for identifying the type of a vehicle (second vehicle) based on the representative value of the received power and the vehicle type information linked by the signal processing unit 110. Similarly to the type of the vehicle (first vehicle) related to the learning stage, the type of the vehicle (second vehicle) related to the identification stage may include a type in which the vehicle is a large vehicle or a type in which the vehicle is not a large vehicle.

[0053] More specifically, the parameter generating unit 115 determines the type of vehicle (first vehicle) corresponding to the type information included in the vehicle information. Type information and types may be associated in advance. For example, assume that "type = vehicle type A" and "vehicle length ≧ L1" are associated with "type = large vehicle." In this case, if the type indicated by the type information included in the vehicle information is "vehicle type A" and the vehicle length is L1 or more, the parameter generating unit 115 may determine the type of the vehicle as "large vehicle."

[0054] On the other hand, if the vehicle type indicated by the type information included in the vehicle information is "other than vehicle type A," or if the vehicle length is less than L1, the parameter generation unit 115 may determine the vehicle type as "other than a large vehicle." Then, the parameter generation unit 115 learns parameters using the determined vehicle type as teacher data and the representative value of the received power as training data.

[0055] This allows the parameters for identifying the type of vehicle to be automatically adjusted, taking into consideration the directivity of the antenna provided in the radar device 101, the orientation of the antenna, the environment in which the antenna is installed, etc. Therefore, it becomes possible to improve the accuracy of identifying the type of vehicle using the adjusted parameters.

[0056] The parameter generating unit 115 may learn parameters based on the distance from the radar device 101 to the object, the representative value of the received power, and the vehicle type information. In this case, the parameter generating unit 115 learns the parameters using the determined vehicle type as teacher data and the distance from the radar device 101 to the object and the representative value of the received power as training data. This allows learning to be performed taking into account the distance from the radar device 101 to the object.

[0057] The parameters may be parameters for determining boundaries that divide the training data into teacher data. The type of machine learning algorithm is not particularly limited. For example, a support vector machine (SVM) may be used as the machine learning algorithm.

[0058] The configuration example of the communication system according to the embodiment of the present invention has been described above.

[0059] (Explanation of operation) Next, an example of the operation of the signal processing unit 110 of the radar device 101 according to the embodiment of the present invention will be described.

[0060] 2 is a flowchart showing an example of the operation of the signal processing unit 110 of the radar device 101 according to the embodiment of the present invention. As shown in Fig. 2, the signal processing unit 110 acquires one or more pieces of vehicle information from the vehicle information storage unit 114 (S201). Then, the signal processing unit 110 repeatedly executes processes (S202 to S206) for each vehicle corresponding to the acquired one or more pieces of vehicle information.

[0061] First, the signal processing unit 110 extracts sensor information related to the vehicle information (S203). More specifically, the signal processing unit 110 detects the position of an object relative to the position of the radar device 101 based on preset position information of the radar device 101 and the distance and direction included in the sensor information. Then, the signal processing unit 110 calculates the difference between the position of the object and the position of the vehicle indicated by the vehicle position information.

[0062] If the difference between the object position and the vehicle position is smaller than a first threshold value set in advance, the signal processing unit 110 determines that the object and the vehicle are the same object, and extracts sensor information related to the vehicle information. The signal processing unit 110 extracts a representative value of the received power from the extracted sensor information (S204).

[0063] Then, the signal processing unit 110 generates a set of the distance from the radar device 101 to the object, which is included in the sensor information, the vehicle length and vehicle type, which are included in the vehicle information, and the representative value of the received power extracted from the sensor information, and outputs the generated set to the parameter generating unit 115 (S205). When the repeated processing (S202 to S206) for each vehicle is completed, the parameter generating unit 115 learns parameters based on each set output from the signal processing unit 110.

[0064] An example of the operation of the signal processing unit 110 of the radar device 101 according to the embodiment of the present invention has been described above.

[0065] The details of the embodiment of the present invention have been described above.

[0066] [Effect description] As described above, according to an embodiment of the present invention, a learning device is provided, which includes: a vehicle information acquisition unit that acquires vehicle information including position information of a first vehicle and type information corresponding to the type of the first vehicle; a sensor information acquisition unit that acquires sensor information including a distance and direction from a sensor and a representative value of the received power at the sensor; a linking processing unit that links the type information and the representative value based on the position information, the distance, and the direction; and a learning unit that learns parameters for identifying the type of a second vehicle based on the representative value and the type information.

[0067] Furthermore, according to the embodiment of the present invention, the parameters for identifying the type of vehicle can be automatically adjusted while taking into consideration the directivity of the antenna provided in the radar device, the orientation of the antenna, the environment in which the antenna is installed, etc. Therefore, it is possible to improve the accuracy of identifying the type of vehicle using the adjusted parameters.

[0068] [Description of Modifications] Although the preferred embodiments of the present invention have been described in detail above with reference to the accompanying drawings, the present invention is not limited to these examples. It is clear that a person skilled in the art to which the present invention pertains can conceive of various modifications and alterations within the scope of the technical ideas set forth in the claims, and it is understood that these also naturally fall within the technical scope of the present invention.

[0069] (Modified sensor) In the above description, an FMCW radar has been taken as an example of an object detection sensor mounted on the radar device 101. However, the object detection sensor mounted on the radar device 101 is not limited to this example. For example, the object detection sensor mounted on the radar device 101 may be a radar using another method, a LiDAR (Light Detection and Ranging), or another sensor.

[0070] (Modification of type information) The above mainly describes the case where the type information includes both the vehicle model and the vehicle length. However, the type information may include the vehicle model without including the vehicle length. For example, consider a case where "type = vehicle model A" is associated with "type = large vehicle." In this case, if the vehicle model indicated by the type information included in the vehicle information is "vehicle model A," the parameter generation unit 115 may determine the vehicle model as "large vehicle." On the other hand, if the vehicle model indicated by the type information included in the vehicle information is "other than vehicle model A," the parameter generation unit 115 may determine the vehicle model as "other than large vehicle."

[0071] Alternatively, the type information may include the vehicle length without including the vehicle model. For example, consider a case where "type = large vehicle" is associated with "vehicle length ≧ L1." In this case, the parameter generation unit 115 may determine the vehicle type as "large vehicle" if the vehicle length indicated by the type information included in the vehicle information is L1 or more. On the other hand, the parameter generation unit 115 may determine the vehicle type as "other than large vehicle" if the vehicle length indicated by the type information included in the vehicle information is less than L1.

[0072] (Matching Variation 1) In the above, the case where the vehicle position is used to match the vehicle information with the sensor information has been mainly described. However, in addition to the vehicle position, other additional information may be used to match the vehicle information with the sensor information. For example, the additional information may include the vehicle speed.

[0073] More specifically, the vehicle information transmitted from the on-board device 104 may include the speed of the vehicle (first vehicle). As described above, the signal processing unit 110 may detect the speed of the object. That is, the sensor information obtained by the radar device 101 may include the speed of the object. In such a case, the signal processing unit 110 may calculate the difference between the speed of the object included in the sensor information and the speed of the vehicle (first vehicle) included in the vehicle information.

[0074] Then, when the difference between the position of the object included in the sensor information and the position of the vehicle (first vehicle) included in the vehicle information is smaller than a preset first threshold, and when the difference between the speed of the object included in the sensor information and the speed of the vehicle (first vehicle) included in the vehicle information is smaller than a preset second threshold, the signal processing unit 110 determines that the object and the vehicle are the same object, and associates the vehicle information with the sensor information. That is, the signal processing unit 110 associates the vehicle type information with the representative value of the received power.

[0075] (Matching variation 2) Alternatively, the additional information may include the vehicle length.

[0076] More specifically, the vehicle information transmitted from the on-board device 104 may include the length of the vehicle (first vehicle). Also, the signal processing unit 110 may detect the length of an object along a vehicle traveling direction that is set in advance. That is, the sensor information obtained by the radar device 101 may include the length of an object along the vehicle traveling direction. In such a case, the signal processing unit 110 may calculate the difference between the length of the object included in the sensor information and the length of the vehicle (first vehicle) included in the vehicle information.

[0077] Then, when the difference between the position of the object included in the sensor information and the position of the vehicle (first vehicle) included in the vehicle information is smaller than a preset first threshold, and when the difference between the length of the object included in the sensor information and the length of the vehicle (first vehicle) included in the vehicle information is smaller than a preset third threshold, the signal processing unit 110 determines that the object and the vehicle are the same object, and associates the vehicle information with the sensor information. That is, the signal processing unit 110 associates the vehicle type information with the representative value of the received power.

[0078] (Matching Variation 3) Alternatively, the additional information may include the time of day.

[0079] More specifically, the vehicle information transmitted from the on-board device 104 may include the time (first time) at which the position of the vehicle (first vehicle) is detected. Also, the signal processing unit 110 may acquire the time (second time) at which the distance and direction are detected. That is, the sensor information acquired by the radar device 101 may include the time at which the distance and direction are detected. In such a case, the signal processing unit 110 may calculate the difference between the time included in the sensor information and the time included in the vehicle information.

[0080] Then, when the difference between the position of the object included in the sensor information and the position of the vehicle (first vehicle) included in the vehicle information is smaller than a preset first threshold, and when the difference between the time (first time) included in the sensor information and the time (second time) included in the vehicle information is smaller than a preset fourth threshold, the signal processing unit 110 determines that the object and the vehicle are the same object, and associates the vehicle information with the sensor information. That is, the signal processing unit 110 associates the vehicle type information with the representative value of the received power. [Explanation of symbols]

[0081] 101 Radar equipment 102 Roadside machine 103 vehicles 104 Onboard equipment 105 Signal generation unit 106 Transmitting Antenna 107 Receiving Antenna 108 Mixer 109 ADC 110 Signal processing section 111 Vehicle Antenna 112 Roadside unit antenna 113 LAN cable 114 Vehicle information storage unit 115 Parameter Generation Unit

Claims

1. a vehicle information acquisition unit that acquires vehicle information including location information of a first vehicle and type information according to the type of the first vehicle; a sensor information acquisition unit that acquires sensor information including a distance and a direction from a sensor and a representative value of received power at the sensor; a linking processing unit that calculates a position of an object based on the position information, the distance, and the direction, and links the type information with the representative value when a difference between the position of the object and the position of the first vehicle indicated by the position information is smaller than a first threshold; a learning unit that learns a parameter for identifying a second vehicle type based on the representative value and the type information; A learning device comprising:

2. The type of the first vehicle includes a type in which the first vehicle is a large vehicle or a type in which the first vehicle is other than a large vehicle, The type of the second vehicle includes a type in which the second vehicle is a large vehicle or a type in which the second vehicle is other than a large vehicle. The learning device according to claim 1 .

3. The type information includes at least one of a vehicle model and a vehicle length of the first vehicle. The learning device according to claim 1 or 2.

4. the learning unit determines a type of the first vehicle corresponding to the type information, and learns the parameters using the type of the first vehicle as teacher data and the representative value as training data. The learning device according to any one of claims 1 to 3.

5. The representative value is a maximum value, a median value, or an average value of the received power. The learning device according to any one of claims 1 to 4.

6. the learning unit learns the parameters based on the representative value, the distance, and the type information. The learning device according to any one of claims 1 to 5.

7. the vehicle information includes a speed of the first vehicle; the sensor information includes a velocity of an object; the linking processing unit links the type information with the representative value when a difference between the position of the object and the position of the first vehicle is smaller than the first threshold value and when a difference between a speed of the object and a speed of the first vehicle is smaller than a second threshold value; The learning device according to any one of claims 1 to 6.

8. the vehicle information includes a vehicle length of the first vehicle; the sensor information includes a length of the object along a predetermined vehicle travel direction; the linking processing unit links the type information with the representative value when a difference between the position of the object and the position of the first vehicle is smaller than the first threshold value and when a difference between the length of the object and the length of the first vehicle is smaller than a third threshold value. The learning device according to any one of claims 1 to 6.

9. the vehicle information includes a first time at which the location was detected; the sensor information includes a second time at which the distance and the direction are detected; the linking processing unit links the type information with the representative value when a difference between the position of the object and the position of the first vehicle is smaller than the first threshold value and when a difference between the first time and the second time is smaller than a fourth threshold value. The learning device according to any one of claims 1 to 6.

10. acquiring vehicle information including location information of a first vehicle and type information according to the type of the first vehicle; acquiring sensor information including a distance and a direction from a sensor and a representative value of received power at the sensor; calculating a position of the object based on the position information, the distance, and the direction, and associating the type information with the representative value when a difference between the position of the object and the position of the first vehicle indicated by the position information is smaller than a first threshold value; learning a parameter for identifying a second vehicle type based on the representative value and the type information; A learning method that provides:

11. Computer, a vehicle information acquisition unit that acquires vehicle information including location information of a first vehicle and type information according to the type of the first vehicle; a sensor information acquisition unit that acquires sensor information including a distance and a direction from a sensor and a representative value of received power at the sensor; a linking processing unit that calculates a position of an object based on the position information, the distance, and the direction, and links the type information with the representative value when a difference between the position of the object and the position of the first vehicle indicated by the position information is smaller than a first threshold; a learning unit that learns a parameter for identifying a second vehicle type based on the representative value and the type information; A program that functions as a learning device equipped with the above.

Images