Road surface evaluation device

Abstract

To evaluate road surface states with good accuracy.SOLUTION: A road surface evaluation device 10 comprises: an information acquisition unit 111 that acquires position information on a plurality of vehicles traveling along a road, acceleration information indicating the acceleration of the plurality of vehicles, travel image information including the captured image of a road surface, and travel sound information indicating the traveling sound of the plurality of vehicles, as travel information on the plurality of vehicles; a storage unit 120 that stores map information including roughness information indicating the roughness of a road surface; an evaluation unit that evaluates the roughness of the road surface on the basis of the travel information on the plurality of vehicles acquired by the information acquisition unit 111; and an update unit that updates roughness information corresponding to the road stored in the storage unit on the basis of the evaluation result of the evaluation unit.SELECTED DRAWING: Figure 3

Description

[Technical field]

[0001] The present invention relates to a road surface evaluation device that evaluates a road surface profile that represents the uneven shape of a road surface. [Background technology]

[0002] Conventionally, one such device is known that displays on a map a road surface roughness index calculated based on the traveling acceleration acquired from a traveling vehicle (see, for example, Patent Document 1). The device described in Patent Document 1 displays on a map the crack rate of the road surface detected based on an image captured by an on-board camera together with the road surface roughness index, so that the road surface condition can be efficiently grasped. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2019-196680 A Summary of the Invention [Problem to be solved by the invention]

[0004] However, the road surface condition detected based on the captured image may vary depending on the environment outside the vehicle, such as the weather, etc. Therefore, if the captured image is used as in the device described in Patent Document 1, there is a risk that the road surface condition cannot be evaluated with high accuracy. [Means for solving the problem]

[0005] A road surface evaluation device according to one embodiment of the present invention includes a driving information acquisition unit that acquires, as driving information of multiple vehicles, position information of multiple vehicles traveling on a road, acceleration information indicating the acceleration of the multiple vehicles, driving image information including an image of the road surface, and driving sound information indicating the driving sound of the multiple vehicles, a memory unit that stores map information including roughness information indicating the roughness of the road surface, an evaluation unit that evaluates the roughness of the road surface based on the driving information of the multiple vehicles acquired by the driving information acquisition unit, and an update unit that updates the roughness information corresponding to the road stored in the memory unit based on the evaluation result of the evaluation unit. The travel information acquisition unit further acquires travel speed information indicating the travel speeds of the multiple vehicles calculated from the time transition of the position information of the multiple vehicles, or measured values ​​of the travel speeds of the multiple vehicles transmitted from the multiple vehicles as travel information. The evaluation unit removes noise contained in the travel sound information by spectrum analysis according to the speed band based on the travel speed information, and uses the travel sound information from which the noise has been removed to evaluate the roughness of the road surface. Effect of the Invention

[0006] According to the present invention, the road surface condition can be evaluated with high accuracy. [Brief description of the drawings]

[0007] [Figure 1] 1 is a diagram showing an example of the configuration of a road surface evaluation system including a road surface evaluation device according to an embodiment of the present invention. [Diagram 2] FIG. 2 is a block diagram showing the configuration of a main part of an in-vehicle terminal. [Diagram 3] 1 is a block diagram showing a configuration of a main part of a road surface evaluation device according to an embodiment of the present invention. [Figure 4A] FIG. 2 is a diagram showing an example of a map of roads on which a vehicle travels. [Figure 4B] FIG. 4 is a diagram showing an example of travel information. [Figure 5A] FIG. 4 is a diagram for explaining a method of deriving the correlation between a road surface roughness value and a lateral acceleration. [Figure 5B] FIG. 4 is a diagram for explaining a method of deriving the correlation between a road surface roughness value and a lateral acceleration. [Figure 6] FIG. 4 is a diagram showing an example of road surface profile information. [Figure 7] FIG. 4 is a diagram for explaining updating of a road surface condition inference model. [Figure 8A] FIG. 4 is a diagram for explaining correction of vehicle vibration data. [Figure 8B] FIG. 4 is a diagram for explaining correction of driving sound data. [Figure 9A] FIG. 4 is a diagram for explaining correction of road image data. [Figure 9B] FIG. 4 is a diagram for explaining the position of a road surface imaged by a camera. [Figure 10] 8 is a diagram for explaining calculation of a road surface roughness value using the road surface condition inference model of FIG. 7. [Figure 11] 4 is a flowchart showing an example of processing executed by a calculation unit in FIG. 3. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0008] Hereinafter, an embodiment of the present invention will be described with reference to Figs. 1 to 11. A road surface evaluation device according to an embodiment of the present invention is a device for evaluating a road surface profile of a road on which a vehicle travels. Fig. 1 is a diagram showing an example of the configuration of a road surface evaluation system including a road surface evaluation device according to this embodiment. As shown in Fig. 1, the road surface evaluation system 1 includes a road surface evaluation device 10 and an in-vehicle terminal 30. The road surface evaluation device 10 is configured by, for example, a server device. The in-vehicle terminal 30 is configured to be able to communicate with the road surface evaluation device 10 via a communication network 2.

[0009] The communication network 2 includes not only public wireless communication networks such as the Internet and mobile phone networks, but also closed communication networks established for each designated management area, such as wireless LAN, Wi-Fi (registered trademark), Bluetooth (registered trademark), etc.

[0010] The in-vehicle terminal 30 is mounted on a vehicle 20. The vehicles 20 include a plurality of vehicles 20-1, 20-2, . . . , 20-n. The vehicles 20 may be manually driven vehicles or automatically driven vehicles. The vehicles 20 may include vehicles of different models and grades.

[0011] 2 is a block diagram showing the main configuration of the in-vehicle terminal 30 according to this embodiment. The in-vehicle terminal 30 includes an electronic control unit (ECU) 31, a positioning sensor 32, an acceleration sensor 33, a steering angle sensor 34, a vehicle speed sensor 35, an illuminance sensor 36, a camera 37, a microphone (hereinafter simply referred to as a microphone) 38, and a TCU (Telematic Control Unit) 39.

[0012] The positioning sensor 32 is, for example, a GPS sensor that receives a positioning signal transmitted from a GPS satellite and detects the absolute position (latitude, longitude, etc.) of the vehicle 20. The positioning sensor 32 includes not only a GPS sensor but also a sensor that performs positioning using radio waves transmitted from satellites of various countries called GNSS satellites, including quasi-zenith orbit satellites. Hereinafter, information indicating the position of the vehicle detected by the positioning sensor 32 is referred to as position information. The position information also includes information indicating the time when the vehicle 20 traveled through that position (hereinafter, referred to as travel time information).

[0013] The acceleration sensor 33 detects the acceleration in the left-right direction of the vehicle 20, i.e., the lateral acceleration. The acceleration sensor 33 may be configured to detect the acceleration in the front-rear direction and the acceleration in the up-down direction as well as the lateral acceleration of the vehicle 20. Hereinafter, information on the acceleration detected by the acceleration sensor 33 will be referred to as acceleration information or vehicle vibration data. The steering angle sensor 34 detects the steering angle of the steering wheel (not shown) of the vehicle 20. The vehicle speed sensor 35 detects the vehicle speed of the vehicle 20.

[0014] The illuminance sensor 36 has a light receiving element and detects the brightness (illuminance) of light incident on the light receiving element. The illuminance sensor 36 is installed outside the vehicle 20 (for example, on the roof) or inside the vehicle 20 (on the dashboard) so as to be able to detect the illuminance around the vehicle 20. Note that since light passing through the windshield is attenuated to some extent by the glass, when the illuminance sensor 36 is installed inside the vehicle 20, the sensor value may be corrected taking this amount of attenuation into account.

[0015] The camera 37 has an imaging element (image sensor) such as a CCD or a CMOS. The camera 37 continuously captures images of the space ahead of the vehicle 20 to acquire image data. The camera 37 is attached to a predetermined position (front) of the vehicle 20 so that the road surface on which the vehicle 20 is traveling is included in the imaging range. Hereinafter, the captured image data acquired by the camera 37 will be referred to as road image data or simply as a road image. The camera 37 may be a monocular camera or a stereo camera.

[0016] The microphone 38 converts the input sound into an electric signal and outputs it. The microphone 38 is attached to a predetermined position of the vehicle 20 so as to be able to collect the running sound of the vehicle 20. Hereinafter, the data of the running sound output from the microphone 38 will be referred to as running sound data. The microphone 38 may be attached inside or outside the vehicle. The microphone 38 may be composed of a single microphone, or may be composed of a microphone array having multiple microphones.

[0017] 2, the ECU 31 includes a computer having a calculation unit 310 such as a CPU, a storage unit 320 such as a ROM, a RAM, and other peripheral circuits (not shown) such as an I / O interface. The calculation unit 310 functions as a sensor value acquisition unit 311 and a communication control unit 312 by executing a program stored in advance in the storage unit 320.

[0018] The sensor value acquisition unit 311 acquires the detection value (position information) of the positioning sensor 32 and the detection value (vehicle vibration data) of the acceleration sensor 33 at a predetermined sampling period. The communication control unit 312 transmits the information acquired by the sensor value acquisition unit 311 (hereinafter referred to as driving information) together with a vehicle ID capable of identifying the vehicle 20 to the road surface evaluation device 10 via the TCU 36 at a predetermined period.

[0019] The driving information acquired by the sensor value acquisition unit 311 further includes a detection value of the vehicle speed sensor 35, i.e., a measured value of the driving speed of the vehicle 20 (hereinafter referred to as driving speed information), a detection value of the illuminance sensor 36 (hereinafter referred to as illuminance information), a detection value of the camera 37 (road image data), and a detection value of the microphone 38 (driving sound data). Note that the driving information may also include a detection value of the steering angle sensor 34, i.e., a measured value of the steering angle of the steering wheel of the vehicle 20 (hereinafter referred to as steering angle information).

[0020] The road surface evaluation device 10 detects the uneven shape of the road surface, i.e., the roughness of the road surface (hereinafter also referred to as the road surface profile), based on the detection value (vehicle vibration data) of the acceleration sensor 33 of the vehicle 20 (vehicle-mounted terminal 30). Information on this detected road surface profile is output to a terminal owned by a road management company or the like, and is used as reference data when the road management company or the like considers the need for repairs, etc. In other words, the detection value of the acceleration sensor is used to evaluate the road surface profile.

[0021] 3 is a block diagram showing the configuration of the main parts of the road surface evaluation device 10 according to this embodiment. The road surface evaluation device 10 includes a computer having a calculation unit 110 such as a CPU, a storage unit 120 such as a ROM, a RAM, and other peripheral circuits (not shown) such as an I / O interface. The storage unit 120 stores map information including a road map and various information processed by the calculation unit 110.

[0022] The calculation unit 110 executes the programs stored in the storage unit 120 to function as an information acquisition unit 111 , an evaluation unit 112 , an output unit 113 , and a communication control unit 114 .

[0023] The information acquisition unit 111 receives travel information from the in-vehicle terminal 30 of each vehicle 20 traveling on a road via the communication control unit 114. The information acquisition unit 111 can identify the vehicle 20 that is the source of the travel information by the vehicle ID associated with the travel information.

[0024] The information acquisition unit 111 stores the driving information received from multiple vehicles 20 (in-vehicle terminals 30) in the memory unit 120 in chronological order. Hereinafter, the driving information stored in the memory unit 120 in chronological order is referred to as time-series driving information. The information acquisition unit 111 also acquires map information including information on roads on which the vehicle 20 is traveling from the memory unit 120. Furthermore, the information acquisition unit 111 acquires weather information corresponding to the driving position and driving time of the vehicle 20 indicated by the driving information of the vehicle 20 from an external information distribution server (not shown) that distributes weather information. The information acquisition unit 111 stores the weather information in the memory unit 120 in association with the driving information of the vehicle 20.

[0025] The evaluation unit 112 evaluates the amount of unevenness (depth or height) of the road surface, that is, the road surface roughness, based on the travel information of the multiple vehicles 20 acquired within a predetermined period by the information acquisition unit 111. More specifically, the evaluation unit 112 calculates a road surface roughness value indicating the degree of roughness of the road surface, based on the lateral acceleration included in the vehicle vibration data of the multiple vehicles 20 acquired within a predetermined period by the information acquisition unit 111. The road surface roughness value is, for example, a value expressed by IRI (International Roughness Index), which is an international index. Hereinafter, the road surface roughness value may be simply referred to as a roughness value.

[0026] FIG. 4A is a diagram showing an example of a map of a road on which the vehicle 20 travels. FIG. 4A shows a predetermined road (latitude Y to Z of national highway X) to be evaluated for road surface roughness. In FIG. 4A, the upward direction corresponds to the north direction, and the rightward direction corresponds to the east direction. The predetermined road to be evaluated for road surface roughness (hereinafter referred to as the evaluation target road) can be specified by a user (for example, a road management company). More specifically, the evaluation target road is specified by transmitting information capable of identifying the evaluation target road, such as the name and position (latitude and longitude) of the road in the evaluation target section, from a terminal owned by the road management company or the like to the road surface evaluation device 10. When the evaluation target road has multiple lanes on one side, the user may be able to specify the lanes to be evaluated for road surface roughness. The evaluation target road may also be specified by the distance from the starting point coordinates.

[0027] The driving information acquired at a predetermined sampling period by the in-vehicle terminal 30 is transmitted to the road surface evaluation device 10 via the communication control unit 312. Fig. 4B is a diagram showing an example of time-series driving information acquired by the road surface evaluation device 10 from the in-vehicle terminal 30 of the vehicle 20-1 traveling on the evaluation target road (latitude Y to Z of national highway X) in Fig. 4A. The horizontal axis in the diagram is the position (latitude) of the vehicle 20-1 in the traveling direction along the travel lane, and the vertical axis is the lateral acceleration of the vehicle 20-1. Characteristics D1, D2,...,Dn represent the time-series driving information acquired when the vehicle 20-1 traveled on the evaluation target road for the 1st, 2nd,...,nth times, respectively.

[0028] In general, the greater the amount of unevenness on the road surface, the greater the lateral acceleration of the vehicle 20, and there is a predetermined correlation between the road surface roughness value and the lateral acceleration. The evaluation unit 112 uses information indicating this correlation (hereinafter referred to as a road surface condition inference model or simply an inference model) to calculate a road surface roughness value corresponding to the vehicle position on the road from the lateral acceleration.

[0029] First, the evaluation unit 112 performs machine learning (machine learning LN in FIG. 7 described later) using previously measured road surface roughness values ​​and lateral acceleration as teacher data to generate a road surface condition inference model. FIGS. 5A and 5B are diagrams for explaining teacher data of road surface roughness values ​​and lateral acceleration, respectively. The vehicle V1 shown in FIG. 5A is a dedicated vehicle equipped with a measuring device MA that measures road surface roughness. The measuring device MA measures the road surface roughness value of the road RD when the vehicle V1 is traveling on a predetermined road (measurement course, etc.) RD. The characteristic P1 in FIG. 5A indicates the road surface roughness value measured at this time, i.e., the road surface roughness value used as teacher data.

[0030] Fig. 5B shows a state in which vehicle 20 in Fig. 1 travels on the same road RD as in Fig. 5A. A characteristic P2 in Fig. 5B shows the lateral acceleration detected by acceleration sensor 33 provided on vehicle 20 while vehicle 20 is traveling on a predetermined road RD at a predetermined traveling speed (hereinafter referred to as reference traveling speed) Vref, that is, the lateral acceleration used as teacher data.

[0031] The teacher data of the road surface roughness value and the lateral acceleration may be stored in the memory unit 120 of the road surface evaluation device 10, or may be stored in an external storage device. The evaluation unit 112 performs machine learning using the teacher data of the road surface roughness value and the lateral acceleration read from the memory unit 120 or the external storage device to generate a road surface condition inference model. Note that machine learning may be performed by adding the longitudinal acceleration and the steering angle as the teacher data.

[0032] The evaluation unit 112 uses the generated road surface condition inference model to calculate the road surface roughness value of the road to be evaluated based on the driving information acquired while the vehicle 20 is traveling on the road to be evaluated.

[0033] The output unit 113 executes a process (hereinafter referred to as a map information update process) of storing the roughness information of the road surface evaluated by the evaluation unit 112, i.e., the road surface roughness value calculated by the evaluation unit 112, in association with the road information included in the map information in the storage unit 120. Note that if a road surface roughness value corresponding to the road to be evaluated has already been stored, the output unit 113 updates the road surface roughness value with the road surface roughness value calculated by the evaluation unit 112.

[0034] The output unit 113 outputs the road surface roughness information evaluated by the evaluation unit 112 in association with the road information acquired by the information acquisition unit 111. The information output at this time is called road surface profile information. When the output unit 113 receives an output instruction for the road surface profile from a terminal of a road management company or the like via the communication network 2, the output unit 113 outputs the road surface profile information to the terminal that sent the output instruction or to a predetermined output destination terminal. The road surface profile information is information that can be displayed on a display device such as a display, and a user can check the road surface profile by displaying the road surface profile information on a display of the user's terminal. FIG. 6 is a diagram showing an example of road surface profile information. The characteristic P11 shown in FIG. 6 represents a road surface roughness value calculated based on the time-series driving information D1 of FIG. 4B. The characteristic P12 shown in FIG. 6 represents a road surface roughness value calculated based on the time-series driving information D2 of FIG. 4B. The characteristic P13 shown in FIG. 6 represents a road surface roughness value calculated based on the time-series driving information Dn of FIG. 4B.

[0035] Incidentally, the detection value of the acceleration sensor 33 included in the travel information of the vehicle 20 may contain noises and the like caused by the external environment of the vehicle 20, such as the weather. Therefore, even if the vehicle 20 travels on the same road, if the external environment during travel is different, different road surface roughness values ​​are calculated as shown in Fig. 6. In this case, there is a risk that the condition of the road surface cannot be evaluated with high accuracy. Therefore, in order to address such a problem, the evaluation unit 112 evaluates the road surface roughness as follows.

[0036] FIG. 7 is a diagram for explaining the update of the road surface condition inference model. The model RM in FIG. 7 is a road surface condition inference model generated by the machine learning LN using the teacher data in FIG. 5A and FIG. 5B. As shown in FIG. 7, the evaluation unit 112 inputs the vehicle vibration data acquired by the information acquisition unit 111 to the machine learning LN to update the inference model RM. At this time, the evaluation unit 112 inputs the road image data and the running sound data detected at the same timing as the vehicle vibration data to the machine learning LN together with the vehicle vibration data. In this way, by inputting the road image data and the running sound data together with the vehicle vibration data to the machine learning LN, the inference accuracy of the inference model RM can be improved.

[0037] On the other hand, the vehicle vibration data, road image data, and running sound data change depending on the running speed of the vehicle 20 even if the vehicle 20 runs on the same road under the same environment. Therefore, if the vehicle vibration data, road image data, and running sound data are input directly to the machine learning LN, there is a possibility that a desired road surface condition inference model RM cannot be obtained. In consideration of this point, the evaluation unit 112 performs corrections CR1, CR2, and CR3 based on the running speed of the vehicle 20 on the road image data, the vehicle vibration data, and the running sound data, as shown in FIG. 7.

[0038] FIG. 8A is a diagram for explaining the correction CR2 of the vehicle vibration data. FIG. 8A shows how the time axis of the vehicle vibration data before correction, that is, the vehicle vibration data (left diagram) included in the traveling information acquired by the information acquisition unit 111, is multiplied by a correction coefficient to obtain the vehicle vibration data after correction (right diagram). The correction coefficient is a value obtained by dividing the reference traveling speed Vref by the traveling speed indicated by the traveling speed information acquired by the information acquisition unit 111, that is, the traveling speed V of the vehicle 20 at the time when the vehicle vibration data is detected. FIG. 8B is a diagram for explaining the correction CR3 of the traveling sound data. The correction of the traveling sound data is performed in the same manner as the correction of the vehicle vibration data. Specifically, the time axis of the traveling sound data before correction, that is, the traveling sound data (left diagram) included in the traveling information acquired by the information acquisition unit 111, is multiplied by a correction coefficient (Vref / V) to obtain the traveling sound data after correction (right diagram).

[0039] FIG. 9A is a diagram for explaining the correction CR1 of the road image data. As shown in FIG. 9A, the road image data SD is video image data consisting of a plurality of frames. In the diagram, the numbers in parentheses indicate the frame numbers. Frame N is the current frame, and frames N-1, . . . , N-5 are past frames consecutive to frame N. FIG. 9B is a diagram for explaining the position of the road surface captured by the camera 37 of the in-vehicle terminal 30. As shown in FIG. 9B, the position P1 of the road surface captured by the camera 37 is located a distance L ahead of the vehicle 20. The distance L is calculated based on the position and attitude of the camera 37 relative to the road surface. The road surface (more specifically, the road surface directly under the tires of the vehicle 20) at the time when the current frame N is captured by the camera 37 is included in the frame (frame N-3 in the example of FIG. 9A) captured a predetermined time Δt (=L / V) before the current frame N. Therefore, the evaluation unit 112 captures road image data with the time axis shifted by Δt back as the corrected road image data.

[0040] 7, the evaluation unit 112 assigns weights W1, W2, and W3 to the road image data, the vehicle vibration data, and the running sound data corrected based on the running speed of the vehicle 20. The calculation of the weights W1, W2, and W3 will be described.

[0041] The evaluation unit 112 calculates the weight W1 to be applied to the road image data based on the illuminance information included in the driving information acquired by the information acquisition unit 111. Specifically, the evaluation unit 112 determines the brightness (illuminance) of the surroundings of the vehicle 20 based on the illuminance information, and when the illuminance is equal to or less than a predetermined value, the weight W1 is made smaller than when the illuminance is greater than the predetermined value. The evaluation unit 112 may make the weight W1 smaller as the illuminance decreases. In addition, since the image recognition accuracy (the recognition accuracy of the road surface condition based on the camera image) decreases when the driving time zone is in bad weather compared to when the weather is fine, the evaluation unit 112 may recognize the driving position of the vehicle 20 and the weather during the driving time zone based on the weather information acquired by the information acquisition unit 111, and may make the weight W1 smaller when the weather is in bad weather than when the weather is fine. Bad weather is weather that reduces the image recognition accuracy, such as rainfall, snowfall, snow accumulation, or fog.

[0042] Furthermore, the evaluation unit 112 calculates a weight W2 to be applied to the vehicle vibration data based on the traveling speed of the vehicle 20. The vibration of the vehicle 20 caused by the unevenness of the road surface changes depending on the traveling speed of the vehicle 20. Specifically, the higher the traveling speed, the greater the vehicle vibration caused by the unevenness of the road surface. Taking this into consideration, the evaluation unit 112 reduces the weight W2 as the traveling speed of the vehicle 20 increases.

[0043] Furthermore, the evaluation unit 112 calculates the weight W3 of the running sound data based on the weather during the driving time period of the vehicle 20. For example, if it is raining during the driving time period, noise due to rain sounds or splashes of water may be included in the running sound data. Taking this into consideration, the evaluation unit 112 determines whether or not noise due to the weather is included in the running sound data based on the weather during the driving time period of the vehicle 20. Then, the evaluation unit 112 reduces the weight W3 when noise is included compared to when noise is not included.

[0044] The evaluation unit 112 grows the road surface condition inference model RM by periodically or intermittently repeating machine learning as shown in Fig. 7. For example, the evaluation unit 112 executes machine learning as shown in Fig. 7 at predetermined intervals or every time a predetermined amount of new travel information corresponding to the road to be evaluated is accumulated in the storage unit 120. Alternatively, the evaluation unit 112 executes the machine learning as shown in Fig. 7 based on a user instruction input via the communication control unit 114.

[0045] Fig. 10 is a diagram for explaining the calculation of the road surface roughness value using the road surface condition inference model RM of Fig. 7. The evaluation unit 112 inputs the vehicle vibration data, road image data, and running sound data acquired while the vehicle 20 is traveling on the road to be evaluated to the road surface condition inference model RM, and calculates the road surface roughness value of the road to be evaluated. At this time, as shown in Fig. 10, the evaluation unit 112 applies corrections CR1, CR2, and CR3 to the vehicle vibration data, road image data, and running sound data, and assigns weights W1, W2, and W3 to the data, and then inputs the data to the road surface condition inference model RM. This makes it possible to accurately calculate the road surface roughness value without being affected by the traveling speed of the vehicle 20 or the external environment.

[0046] FIG. 11 is a flowchart showing an example of processing (map information update processing) executed by the calculation unit 110 (CPU) of the road surface evaluation device 10 in accordance with a predetermined program. The processing shown in this flowchart is repeated at a predetermined cycle while the road surface evaluation device 10 is running. First, in step S11, it is determined whether or not driving information has been received from the in-vehicle terminal 30 of the vehicle 20. If the result in step S11 is negative, the processing is terminated. If the result in step S11 is positive, in step S12, the vehicle information received in step S11 is stored in the storage unit 120. In step S13, it is determined whether or not the road to be evaluated has been specified by the user.

[0047] If the result in step S13 is negative, the process is terminated. If the result in step S13 is positive, then in step S14, it is determined whether or not to update the road surface roughness value included in the map information in the memory unit 120. More specifically, it is determined whether or not a predetermined period of time has elapsed since the last update (the last time step S18 was executed). It may also be determined whether or not a predetermined amount of new driving information corresponding to the road to be evaluated has been accumulated in the memory unit 120. It may also be determined whether or not an update instruction has been received from a terminal of a road management company or the like via the communication network 2.

[0048] If the result in step S14 is negative, the process is terminated. If the result in step S14 is positive, in step S15, the travel information of the vehicle 20 corresponding to the road to be evaluated is acquired from the storage unit 120. In step S16, the vehicle vibration data, road image data, and travel sound data included in the travel information acquired in step S15 are subjected to corrections CR1, CR2, and CR3, and weights W1, W2, and W3 are assigned. In step S17, the road surface roughness is evaluated. Specifically, the vehicle vibration data, road image data, and travel sound data corrected and weighted in step S16 are input to the road surface condition inference model RM to calculate a road surface roughness value. In step S18, the road surface roughness value calculated in step S17 is stored in association with the road information included in the map information of the storage unit 120. As a result, the road surface roughness value included in the map information and corresponding to the road to be evaluated is updated.

[0049] According to the embodiment of the present invention, the following advantageous effects can be obtained. (1) The road surface evaluation device 10 includes an information acquisition unit 111 that acquires position information of a plurality of vehicles 20 traveling on a road, acceleration information indicating the acceleration of the plurality of vehicles 20, traveling image information including an image of the road surface, and traveling sound information indicating the traveling sound of the plurality of vehicles 20 as traveling information of the plurality of vehicles 20, a storage unit 120 that stores map information including roughness information indicating the roughness of the road surface, an evaluation unit 112 that evaluates the roughness of the road surface based on the traveling information of the plurality of vehicles 20 acquired by the information acquisition unit 111, and an output unit 113 that updates the roughness information corresponding to the road stored in the storage unit 120 based on the evaluation result of the evaluation unit 112. The information acquisition unit 111 acquires external environment information indicating the external environment of the plurality of vehicles 20 traveling on the road. The evaluation unit 112 weights the acceleration information, traveling image information, and traveling sound information used to evaluate the roughness of the road surface. The evaluation unit 112 changes the weights assigned to the acceleration information, the running image information, and the running sound information based on the external environment information acquired by the information acquisition unit 111. This makes it possible to accurately evaluate the state of the road surface without being influenced by the external environment.

[0050] (2) The external environment information includes illuminance information indicating the illuminance outside the multiple vehicles 20 traveling on the road. When the illuminance indicated by the illuminance information indicates a predetermined value or less, the evaluation unit 112 assigns a smaller weight to the traveling image information than when the illuminance is greater than the predetermined value. This allows the road surface condition to be evaluated with high accuracy regardless of the brightness outside the vehicle 20.

[0051] (3) The external environment information includes weather information indicating the weather when the multiple vehicles 20 are traveling on the road. When the weather information indicates bad weather, the evaluation unit 112 assigns a smaller weight to the traveling image information than when the weather is not bad. Bad weather includes rainfall, snowfall, snow accumulation, and fog. This allows the road surface condition to be evaluated with high accuracy regardless of the weather when the vehicles 20 are traveling.

[0052] (4) The information acquisition unit 111 further acquires, as the traveling information, traveling speed information indicating the traveling speeds of the multiple vehicles calculated from the time transition of the position information of the multiple vehicles 20, or measured values ​​of the multiple vehicle speeds transmitted from the multiple vehicles 20. The evaluation unit 112 removes noise contained in the traveling sound information according to the speed band based on the traveling speed information by spectrum analysis, and uses the traveling sound information from which the noise has been removed for evaluating the roughness of the road surface. This makes it possible to appropriately remove noise contained in the traveling sound used for evaluating the road surface roughness according to the traveling speed of the vehicle 20. As a result, it is possible to accurately evaluate the condition of the road surface.

[0053] The above embodiment can be modified in various ways. Modifications will be described below.

[0054] In the above embodiment, the evaluation unit 112 inputs the vehicle vibration data acquired by the information acquisition unit 111 to the machine learning LN to update the inference model RM. Normally, even if a plurality of vehicles 20 travel on the same road, if the models and grades of the vehicles 20 are different, the road surface roughness values ​​calculated by the evaluation unit 112 may differ. This is because the parts that affect the vehicle motion, such as the suspensions and tires mounted on each vehicle 20, differ for each model and grade. In consideration of this point, the evaluation unit may correct the vehicle vibration data included in the travel information of each vehicle 20 according to the model and grade of each vehicle 20, and then input it to the machine learning LN. In general, the lower the shock absorption performance (vertical shock absorption performance) of the suspensions and tires, the more easily shocks and vibrations caused by unevenness in the road surface are transmitted to the vehicle, and the larger the lateral acceleration detected by the acceleration sensor 33 of the vehicle 20. In addition, normally, the shock absorption performance of the suspensions and tires is higher as the grade is higher among the same model, and is higher among different models for models that emphasize ride comfort. As a result, even when each vehicle 20 travels on the same road, the lateral acceleration detected in each vehicle 20 varies. Therefore, the information acquisition unit 111 may identify the model and grade of the vehicle 20 based on the vehicle ID (e.g., chassis number) of the vehicle 20 associated with the travel information, recognize the shock absorbing performance of the suspension and tires based on the identified model and grade, and assign a weight based on the corresponding shock absorbing performance to the vehicle vibration data as the shock absorbing performance increases. Specifically, the lower the shock absorbing performance, the smaller the weight assigned to the vehicle vibration data may be.

[0055] In the above embodiment, the evaluation unit 112 evaluates the road surface roughness. However, the evaluation unit may evaluate a road surface condition other than the road surface roughness. That is, data representing other road surface conditions may be input to the road surface condition inference model RM. The other road surface conditions include, for example, the crack rate of the road surface, rutting, flatness, potholes, steps, dry road surface, wet road surface, snow-covered road surface, and frozen road surface. In addition, the road surface evaluation device 10 may use MCI (Maintenance Control Index) instead of IRI as an index representing road surface roughness.

[0056] In the above embodiment, the information acquisition unit 111 as the traveling information acquisition unit acquires the detection value of the vehicle speed sensor 35 as the traveling speed information. However, the traveling information acquisition unit may calculate the traveling speed of the vehicle 20 based on the time transition of the position information of the vehicle 20, and acquire the calculation result as the traveling speed information.

[0057] In the above embodiment, the evaluation unit 112 performs the correction CR2 and the correction CR3 based on the traveling speed on the vehicle vibration data and the running sound data as shown in Fig. 7. However, the evaluation unit may perform machine learning for each speed band instead of performing the correction based on the traveling speed on the vehicle vibration data and the running sound data. Specifically, the vehicle vibration data and the running sound data may be divided into speed bands based on the traveling speed of the vehicle 20 at the time when the vehicle vibration data and the running sound data are detected, and the vehicle vibration data and the running sound data corresponding to each speed band may be input to the machine learning LN prepared for each speed band.

[0058] In the above embodiment, the information acquisition unit 111 as the environmental information acquisition unit acquires weather information indicating the weather when the vehicle 20 is traveling on the road, and the evaluation unit 112 recognizes the weather during the traveling time of the vehicle 20 based on the weather information, and weights the traveling sound data using the weight W3 calculated based on the weather. However, instead of weighting the traveling sound data, the evaluation unit may remove noise included in the traveling sound information by spectrum analysis based on the weather information. More specifically, the evaluation unit may use spectrum analysis to detect noise (e.g., splashing sound) corresponding to the weather during the traveling time included in the traveling sound data, and remove the detected noise from the traveling sound data. In addition, when machine learning is performed for each speed band as described above, the evaluation unit may remove noise included in the traveling sound data by spectrum analysis according to the speed band corresponding to the traveling speed of the vehicle 20 based on the traveling speed information.

[0059] In the above embodiment, the evaluation unit 112 as a weighting unit calculates the weight W1 to be applied to the road image data based on the detection value of the illuminance sensor 36. However, the weighting unit may calculate the weight W1 to be applied to the road image data based on the driving time information included in the driving information acquired by the information acquisition unit 111. More specifically, since the image recognition accuracy is lower at night than during the day, when the driving time period of the vehicle 20 is included in the nighttime, the evaluation unit 112 may set the weight W1 smaller than when the driving time period is included in the daytime.

[0060] Furthermore, in the above embodiment, the output unit 113 stores the road surface roughness value calculated by the evaluation unit 112 in association with the road information included in the map information in the storage unit 120. However, the output unit may store the road surface roughness value in association with the road information so that information (text information or color information) indicating the road surface roughness value is superimposed on the position of the road to be evaluated on the map. Also, the output unit 113 as an update unit updates the roughness value corresponding to the road stored in the storage unit 120 based on the evaluation result of the evaluation unit 112, but when updating the roughness value, the update unit may store information indicating that the roughness value has been updated in association with the roughness value in the storage unit 120.

[0061] The above description is merely an example, and the present invention is not limited to the above-mentioned embodiment and modifications, as long as the features of the present invention are not impaired. One or more of the above-mentioned embodiment and modifications can be arbitrarily combined, and modifications can also be combined with each other. [Explanation of symbols]

[0062] 10 Road surface evaluation device, 20, 20-1 to 20-n Vehicles, 30 Vehicle-mounted terminal, 110 Calculation unit, 111 Information acquisition unit, 112 Evaluation unit, 113 Output unit, 114 Communication control unit, 120 Storage unit

Claims

1. a driving information acquisition unit that acquires, as driving information of the plurality of vehicles, position information of a plurality of vehicles traveling on a road, acceleration information indicating the acceleration of the plurality of vehicles, driving image information including an image of a road surface of the road, and driving sound information indicating the driving sound of the plurality of vehicles; A storage unit that stores map information including roughness information indicating the roughness of the road surface; an evaluation unit that evaluates roughness of a road surface based on the travel information of the plurality of vehicles acquired by the travel information acquisition unit; an update unit that updates the roughness information corresponding to the road stored in the storage unit based on an evaluation result of the evaluation unit, The travel information acquisition unit further acquires, as the travel information, travel speed information indicating the travel speeds of the plurality of vehicles calculated from a time transition of the position information of the plurality of vehicles, or measurement values ​​of the travel speeds of the plurality of vehicles transmitted from the plurality of vehicles; The road surface evaluation device is characterized in that the evaluation unit removes noise contained in the running sound information by spectrum analysis according to a speed band based on the running speed information, and uses the running sound information from which the noise has been removed to evaluate the roughness of the road surface.

2. The road surface evaluation device according to claim 1, an environmental information acquisition unit that acquires external environmental information indicating an external environment of the plurality of vehicles traveling on the road; a weighting unit that weights the acceleration information, the running image information, and the running sound information used in the evaluation unit to evaluate the roughness of the road surface, A road surface evaluation device characterized in that the weighting unit changes the weights assigned to the acceleration information, the driving image information, and the driving sound information based on the external environmental information acquired by the environmental information acquisition unit.

3. 3. The road surface evaluation device according to claim 2, the external environment information includes illuminance information indicating an external illuminance of the plurality of vehicles traveling on the road; A road surface evaluation device characterized in that the weighting unit reduces the weight assigned to the driving image information when the illuminance indicated by the illuminance information is below a predetermined value compared to when the illuminance is greater than the predetermined value.

4. 3. The road surface evaluation device according to claim 2, the external environment information includes weather information indicating the weather when the plurality of vehicles are traveling on the road; the weighting unit, when the weather indicated by the weather information is bad weather, reduces the weight assigned to the driving image information compared to when the weather is not bad weather; The bad weather includes rainfall, snowfall, snow accumulation, or fog.

5. The road surface evaluation device according to claim 4, The road surface evaluation device is characterized in that the evaluation unit removes noise contained in the running sound information by spectrum analysis based on the weather information, and uses the running sound information from which the noise has been removed to evaluate the roughness of the road surface.

6. The road surface evaluation device according to claim 5, The road surface evaluation device, wherein the evaluation unit removes the noise caused by the weather indicated by the weather information from the running sound information by spectrum analysis.

Images