Risk Level Setting Device

Abstract

To secure reliability for a setting result in the case of setting a level of risk of a vehicle that is passing through a predetermined section on a road encountering other moving objects while during passage through the predetermined section.SOLUTION: A data processing device performs level setting processing. In the level setting processing, a risk level is set indicating a level of a risk of a vehicle that is passing through a predetermined section encountering other moving objects while passing through the predetermined section. In the level setting processing, irregular data is removed from driving history data for which the risk level is being set. The irregular data is driving history data that was collected due to a dynamic deceleration factor. In the level setting processing, variance of latent risk in the predetermined section is calculated on the basis of regular data remaining after the irregular data is removed. Then, the risk level is set when it is determined that the variance exceeds a threshold value.SELECTED DRAWING: Figure 4

Description

[Technical field]

[0001] The present invention relates to an apparatus for setting a level of risk that a vehicle passing through a predetermined section of a road will encounter another moving object while passing through the predetermined section. [Background technology]

[0002] JP 2020-140310 A discloses a device that collects deceleration behavior data of a vehicle when the vehicle passes through an intersection and classifies the level of risk present around the intersection into three levels. This conventional device extracts intersections with known risks based on map data and sets a predetermined section including the intersection as a risk section. The deceleration behavior data is collected in this risk section. The collected deceleration behavior data includes features such as the distance from the position where the accelerator pedal of the vehicle is released to the intersection.

[0003] The conventional device classifies deceleration behavior data into three types of driver types based on the feature quantities of the deceleration behavior. The three driver types include a driver type with a high level of carefulness, a driver type with a low level of carefulness, and a driver type with an intermediate level of carefulness. The conventional device normalizes the deceleration behavior data classified into the intermediate driver type. The classification into the three levels is performed by allocating each normalized value into three numerical ranges. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2020-140310 A Summary of the Invention [Problem to be solved by the invention]

[0005] The characteristic of the process performed by the above-mentioned conventional device is that data on driver types with a medium level of driving caution is extracted from the deceleration behavior data collected in the risk section and used for classification. However, there is room for improvement in the collection of deceleration behavior data. This is because, if there is a bias in the deceleration behavior data itself, the medium level data will also be biased. This makes it difficult to guarantee the reliability of the classification results using the medium level data.

[0006] One object of the present invention is to provide a technology that can ensure the reliability of the results of setting the level of risk that a vehicle passing through a specified section of a road will encounter another moving object while passing through the specified section. [Means for solving the problem]

[0007] The present invention is a risk level setting device and has the following features. The risk level setting device includes a driving history database, a potential risk database, and a data processing device. The driving history database stores driving history data of the vehicle, the driving history data including deceleration behavior data associated with each of a plurality of road sections and external condition data associated with each of the plurality of road sections. The potential risk database stores data on potential risks, and the potential risks indicate potential traffic safety risks in a specific section selected in advance from among the road sections, the potential risks being calculated based on data on static deceleration factors in the specific section. The data processing device performs a level setting process, in which a risk level indicating a level of risk that a vehicle passing through the predetermined section will encounter another moving object while passing through the predetermined section is set based on the driving history data and the potential risk data. The data processing device, in the level setting process, Extracting the driving history data for a road section corresponding to the predetermined section from the driving history database; Identifying irregular data in which the deceleration behavior data in the predetermined section is derived from a dynamic deceleration factor from the driving history data based on the external condition data included in the extracted driving history data; excluding the irregular data from the extracted driving history data and selecting regular data in which the deceleration behavior data in the predetermined section is derived from the static deceleration cause; Extracting all of the potential risk data for the road section of the regular data from the potential risk database; Calculate the variance of the extracted potential risk data; determining whether the variance exceeds a threshold; If it is determined that the variance exceeds the threshold, the risk level is set for a predetermined section corresponding to the road section of the regular data. Effect of the Invention

[0008] According to the present invention, in the level setting process, irregular data is removed from the driving history data for which a risk level is to be set. The irregular data is driving history data collected due to dynamic deceleration factors, and is therefore not suitable as a target for setting a risk level based on a potential risk. In this regard, the regular data remaining after the irregular data has been removed is driving history data collected due to static deceleration factors. Therefore, according to the present invention, it is possible to ensure the reliability of the risk level set based on the potential risk data.

[0009] Furthermore, in the level setting process, all potential risk data in a specified section corresponding to the road section in the regular data is extracted, and the variance of the data is calculated. If it is determined that the variance exceeds a threshold, the risk level is set. In other words, the risk level is set only if it is determined that there is no bias in the potential risk data. Therefore, according to the present invention, it is possible to increase the reliability of the risk level set based on the potential risk data. [Brief description of the drawings]

[0010] [Figure 1] FIG. 1 is a diagram illustrating an example of a configuration of a risk level setting device according to an embodiment. [Diagram 2] FIG. 2 is a diagram illustrating an example of setting road sections. [Diagram 3] 11 is a diagram for explaining an example of setting a predetermined section and an example of a method for calculating a potential risk. FIG. [Figure 4] 1 is a flowchart showing the flow of a level setting process performed by a data processing device (processor). [Diagram 5] 4 is a diagram illustrating an example of a feature amount (correlated feature amount) having a correlation with a potential risk, which is extracted from deceleration behavior data. FIG. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0011] Hereinafter, a risk level setting device according to an embodiment of the present invention will be described with reference to the drawings. In each drawing, the same or corresponding parts are denoted by the same reference numerals, and the description thereof will be simplified or omitted.

[0012] 1. Configuration Example Fig. 1 is a diagram showing an example of the configuration of a risk level setting device according to an embodiment. The risk level setting device 1 shown in Fig. 1 includes a driving history database 2, a potential risk database 3, and a data processing device 4. These components will be described below.

[0013] 1-1. Driving history database The driving history database 2 is a database for storing driving history data DHIS of multiple vehicles. The driving history data DHIS includes road section data DSEC. The road section data DSEC is data on multiple road sections SEC set on a road RD. The road sections SEC are set, for example, by dividing a road RD connected to an intersection CP at a predetermined interval. In this specification, an "intersection CP" is a portion where at least two roads RD intersect. The road section SEC of the intersection CP is set by connecting the end sections of the road RD adjacent to this intersection CP.

[0014] FIG. 2 is a diagram for explaining an example of setting a road section SEC. Roads RD1 to RD6 are depicted in FIG. 2. Roads RD1, RD2, RD4, and RD5 are connected to intersection CP1. Road RD2 is also connected to intersection CP2. Roads R3 and R6 are also connected to this intersection CP2. In the example shown in FIG. 2, a stop line SL1 and a pedestrian crossing CW1 exist in the terminal section of road RD1. A pedestrian crossing CW2 exists on one side of the terminal section of road RD2, and a stop line SL2 and a traffic light TS exist on the other side. Stop lines SL3 and SL4 exist in the terminal sections of roads RD4 and RD5, respectively. A dashed line DL exists in the terminal section of road RD6 to prompt vehicles entering road RD3 to make a safety check.

[0015] Road sections SEC1 to SEC5 are examples of road sections SEC. Road section SEC1 is set on road R1. Road section SEC1 is located to the left of stop line SL1. Road section SEC2 is set at intersection CP1. Road section SEC2 is formed by connecting end sections of roads RD1, RD2, RD4, and RD5. Road section SEC3 is set on road RD2. Road section SEC3 is located in front of the entrance / exit DW of the house HU. Road section SEC4 is set at intersection CP2. Road section SEC4 is formed by connecting end sections of roads RD2, RD3, and RD6.

[0016] Returning to FIG. 1, the description of the driving history data DHIS will be continued. The driving history data DHIS also includes deceleration behavior data DDEC. The deceleration behavior data DDEC is data indicating the history of the deceleration behavior of the vehicle. The deceleration behavior data DDEC includes, for example, data on the running speed (vehicle speed) of the vehicle on the road RD and data on the amount of operation of the accelerator pedal by the driver.

[0017] The deceleration behavior data DDEC is associated with the road section data DSEC. That is, the deceleration behavior data DDEC is associated with the road section SEC. The deceleration behavior data DDEC is associated with the road section data DSEC, for example, based on the vehicle's traveling speed data and the vehicle's position and attitude data. Note that the traveling speed data is data acquired by an on-board sensor (e.g., a wheel speed sensor), and the position and attitude data is also data acquired by an on-board sensor (e.g., a GPS sensor).

[0018] The generation process and the association process of the deceleration behavior data DDEC may be performed in the data processing device 4, or may be performed in an external data processing device different from the data processing device 4. In the association process of the deceleration behavior data DDEC, association in combination with a vehicle or a driver of this vehicle may be performed. In this case, the deceleration behavior data DDEC is associated with the road section SEC and the vehicle (or the driver of this vehicle).

[0019] The driving history data DHIS further includes external condition data DEXT. The external condition data DEXT is data indicating the history of the external conditions of the vehicle. The external condition data DEXT is acquired, for example, by an on-board sensor (for example, a camera and a radar). The external condition data DEXT includes data of moving objects and data of stationary objects. Examples of moving objects include pedestrians, two-wheeled vehicles, and four-wheeled vehicles. Examples of stationary objects include road markings (for example, regulatory markings and advance warning markings) and structures on the road. Examples of regulatory markings include speed limits, dividing lines (white lines), and pedestrian crossings. Examples of advance warning markings include diamond marks displayed in front of pedestrian crossings. Examples of structures on the road include traffic facilities such as guardrails and traffic lights, and general buildings around the road.

[0020] The external condition data DEXT is associated with the road section data DSEC. That is, the external condition data DEXT is associated with the road section SEC. The association between the external condition data DEXT and the road section data DSEC is performed, for example, based on the position and attitude data of the vehicle when the external condition data DEXT is acquired.

[0021] The association process of the external condition data DEXT may be performed in the data processing device 4, or may be performed in an external data processing device different from the data processing device 4. In addition, in the association process of the external condition data DEXT, association in combination with a vehicle or a driver of this vehicle may be performed. In this case, the road section SEC and the vehicle (or the driver of this vehicle) are associated with the external condition data DEXT.

[0022] 1-2. Potential risk database The potential risk database 3 stores potential risk data DRSK. The potential risk data DRSK is data on the traffic safety risk potentially present in a specific section PSEC on the road RD (hereinafter also referred to as "potential risk RSK"). The specific section PSEC is set on the road RD, like the road section SEC. In the embodiment, the specific section PSEC is selected from among the road sections SEC. The potential risk RSK is calculated based on data on "static deceleration factors" in the specific section PSEC. The data on static deceleration factors will be described later.

[0023] 3 is a diagram for explaining an example of setting a predetermined section PSEC and an example of a method of calculating a potential risk RSK. The roads RD1 to RD6 and the road sections CS1 to CS5 shown in FIG. 3 are the same as those shown in FIG.

[0024] 3 illustrates a vehicle VH traveling on road RD1 (road section SEC1). The vehicle VH travels on roads RD1, RD2, and RD3 in that order. The predetermined section PSEC is set to a road section SEC where there is a potential risk of encountering a moving object other than the vehicle VH while the vehicle VH is traveling. In other words, the predetermined section PSEC is not set to a road section SEC where there is absolutely no risk of encountering another moving object.

[0025] In the example shown in FIG. 3, road sections SEC2, SEC3, and SEC4 are set as the predetermined section PSEC. This is because at intersections CP1 and CP2, there is a risk of encountering other moving objects (such as pedestrians) that jump out at these intersections while passing through road sections SEC2 and SEC4. Similarly, in front of the entrance / exit EN, there is a risk of encountering other moving objects (such as residents of the house HU) that jump out from the entrance / exit EN while passing through road section SEC3. On the other hand, road section SEC1 is not set as the predetermined section PSEC. This is because the obstructions SH present around road section SEC1 essentially eliminate the risk of encountering other moving objects that jump out while passing through road section SEC1.

[0026] The potential risk RSK is calculated using data on static deceleration factors. Static deceleration factors are structural factors in the specified section PSEC that cause the driver of the vehicle VH to assume the risk of running out onto the road RD on which the vehicle VH is traveling. Static deceleration factors are identified based on data that does not change in the short term, such as position data of stationary objects (e.g., latitude and longitude data) and structural data of stationary objects (size, shape).

[0027] Examples of static deceleration factors include the following: The width of the road RD on which the vehicle VH is traveling (i.e., the width of roads RD1 to RD3) - The speed limit on the road RD on which the vehicle VH is traveling The lateral distance (distance in the vehicle width direction) from the vehicle VH to structures (e.g., obstructions SH) around the road RD on which the vehicle VH is traveling - The width of the shoulder of the road RD on which the vehicle VH is traveling The shape of the pedestrian lane on the road RD on which the vehicle VH is traveling (including the presence or absence of dividing lines and curbs) - Presence or absence of road markings (pedestrian crossing CW, diamond mark, broken line DL) around the specified section PSEC - Shape of a given section PSEC (crossroads, T-junctions, entrance / exit DW, etc.) Height of structures (e.g., shielding SH) around a given section PSEC -Whether or not there is a traffic light TS -Whether or not there are traffic signs -Whether or not there is a convex mirror

[0028] In calculating the potential risk RSK, the static deceleration factor in the specified section PSEC is normalized, and the potential risk RSK is calculated, for example, using the following formula (1). RSK = ΣFACk × wk (1) In equation (1), FACk is the value of the normalized kth (1≦k≦n) static deceleration factor, and wk is the weighting coefficient of the kth static deceleration factor.

[0029] The potential risk RSK is calculated for each predetermined section PSEC. The calculation process of the potential risk RSK may be performed in the data processing device 4, or may be performed in an external data processing device different from the data processing device 4.

[0030] 1-3.Data processing device The data processing device 4 performs a "level setting process" for setting a risk level LRSK based on the driving history data DHIS and the potential risk data DRSK. In this specification, the risk level RLV indicates the level of risk that a vehicle passing through a predetermined section PSEC will encounter another moving object while passing through the predetermined section PSEC.

[0031] The data processing device 4 may perform a process of providing the risk level RLV set by the level setting process to the vehicle. Examples of vehicles to which the risk level RLV is provided include new cars and vehicles to which driving history data DHIS has been provided. The data processing device 4 can provide the risk level RLV each time the level setting process is performed.

[0032] As a configuration for performing the level setting process and associated processes, the data processing device 4 includes at least one processor 5 and at least one memory 6. The processor 5 includes a CPU (Central Processing Unit). The memory 6 is a volatile memory such as a DDR memory, and expands programs used by the processor 5 and temporarily stores various data. The various data include driving history data DHIS and potential risk data DRSK. The content of the level setting process will be described later.

[0033] 2. Level setting process Fig. 4 is a flowchart showing the flow of the level setting process performed by the data processing device 4 (processor 5). The routine shown in Fig. 4 is executed when a predetermined timing arrives. An example of the predetermined timing is the timing when the setting (or updating) of the potential risk data DRSK stored in the potential risk database 3 is completed. Another example of the predetermined timing is the timing when the amount of new data (or the amount of updated data) of the driving history data DHIS stored in the driving history database 2 exceeds a certain amount.

[0034] In the routine shown in Fig. 4, first, the deceleration behavior data DDEC and the external condition data DEXT in the predetermined section PSEC are extracted (step S11). As already explained, the deceleration behavior data DDEC and the external condition data DEXT are associated with the road section data DSEC. Furthermore, the predetermined section PSEC is selected from among the road sections SEC. In the processing of step S11, the driving history data DHIS in the predetermined section PSEC is extracted from the driving history database 2 using the data of the road section SEC selected as the predetermined section PSEC (i.e., the road section data DSEC) as a search key. The driving history data DHIS is extracted for all road sections SEC selected as the predetermined section PSEC.

[0035] Following the processing of step S11, irregular data DIRE is identified (step S12). The irregular data DIRE is driving history data DHIS in which the deceleration behavior data DDEC in the specified section PSEC is derived from a "dynamic deceleration factor". The dynamic deceleration factor is a traffic factor in the specified section PSEC that makes the driver of the vehicle VH assume the risk of jumping out onto the road RD on which the vehicle VH is traveling. The dynamic deceleration factor is identified based on data that changes in the short term, such as recognition data of moving objects present in the specified section PSEC and recognition data of the lighting state of the light-emitting parts of traffic lights (e.g., the lighting state of the yellow or red light-emitting parts).

[0036] In the process of step S12, if the external condition data DEXT in the predetermined section PSEC includes data of a moving object (for example, a pedestrian, a vehicle parked on the roadside), the driving history data DHIS in the predetermined section PSEC is determined to correspond to irregular data DIRE.

[0037] Following the process of step S12, the regular data DREG is selected (step S13). The regular data DREG is the driving history data DHIS in which the deceleration behavior data DDEC in the road section SEC selected as the predetermined section PSEC is derived from a static deceleration factor. In the process of step S13, the regular data DREG is selected by excluding the irregular data DIRE identified in step S12 from the driving history data DHIS extracted in step S11.

[0038] Following the processing of step S13, data of the potential risk RSK in the predetermined section PSEC (i.e., potential risk data DRSK) is extracted (step S14). In the processing of step S14, the road section SEC of the regular data DREG selected in the processing of step S13 is used as a search key to extract the potential risk data DRSK in the road section SEC from the potential risk database 3. The extracted potential risk data DRSK is all data of the potential risk RSK in the road section SEC of the regular data DREG.

[0039] Following the processing of step S14, the variance σ of the potential risk RSK 2 is calculated (step S15). 2 is calculated, for example, by using the following formula (2) based on the potential risk data DRSK extracted in step S14. σ 2 =1 / m×Σ(DRSK(i)-μ) 2 (2) In formula (2), DRSK(i) is the potential risk data DRSK of the i-th (1≦i≦m) predetermined section PSEC, and μ is the average value of these data DRSK.

[0040] Following step S15, the variance σ 2 It is determined whether or not the variance σ exceeds a threshold TH (step S16). The threshold TH is set in advance as a value that ensures that the potential risk data DRSK is not biased. 2 If is equal to or smaller than the threshold value TH, it is determined that there is a bias in the potential risk data DRSK. Therefore, in this case, the level setting process is terminated.

[0041] If the determination result of step S16 is positive, a risk level RLV is set (step S17). The risk level RLV is set by extracting a feature quantity (hereinafter also referred to as a "correlated feature quantity") having a correlation with the potential risk RSK from the deceleration behavior data DDEC included in the driving history data DHIS selected as the regular data DREG.

[0042] Fig. 5 is a diagram for explaining an example of a correlation feature extracted from the deceleration behavior data DDEC. Fig. 5 shows a history of accelerator pedal operations performed in a predetermined section PSEC and a history of traveling speed. In the example shown in Fig. 5, an operation of switching the accelerator pedal from on to off is performed while passing through the predetermined section PSEC, and an operation of returning the accelerator pedal to on is performed after passing through a predetermined position PP. An example of the predetermined position PP is the position of an end of an intersection IS or an entrance / exit DW.

[0043] An example of a correlation feature is the distance DIS from the position Poff where the accelerator pedal is turned off to the predetermined position PP. The higher the potential risk RSK, the longer the distance DS tends to be. Therefore, it can be said that the distance DIS corresponds to a correlation feature. Another example is the average speed AVE immediately before the accelerator pedal is turned from on to off. The higher the potential risk RSK, the lower the average speed AVE tends to be. Therefore, it can be said that the average speed AVE also corresponds to a correlation feature. Another example is the minimum speed MIN. The higher the potential risk RSK, the lower the minimum speed MIN tends to be. Therefore, it can be said that the minimum speed MIN also corresponds to a correlation feature.

[0044] In setting the risk level RLV, the extracted correlation feature is normalized. Then, the predetermined section PSEC is allocated to a plurality of risk level RLVs based on the normalized correlation feature. When there are a plurality of types of correlation feature, the average value of the correlation feature is calculated for each predetermined section PSEC, and allocation is performed based on this average value. Consider a case where three risk levels RLV are set. In this case, the predetermined section PSEC is allocated to a high level, a low level, and a middle level. The boundary value between the middle level and the high level or the low level can be appropriately set in consideration of the total number of regular data DREG, the normalized correlation feature (or the average value thereof), and the like.

[0045] In setting the risk level RLV, the normalized correlation feature may be further classified based on the type of the driver's driving prudence. In this case, the risk level RLV is allocated based on the correlation feature after being classified into the type of driving prudence. Note that the allocation of the risk level RLV based on the type of driving prudence is the same as that disclosed in JP2020-140310A, and therefore a detailed description thereof will be omitted here.

[0046] 3.Effects According to the embodiment described above, in the level setting process, the irregular data DIRE is removed from the driving history data DHIS (deceleration behavior data DDEC) that is the target for setting the risk level RLV. The irregular data DIRE is driving history data DHIS collected due to dynamic deceleration factors, and is therefore not suitable as a target for setting the risk level RLV based on the potential risk RSK. In this regard, the regular data DREG that remains after the irregular data DIRE is removed is driving history data DHIS collected due to static deceleration factors. Therefore, according to the embodiment, it is possible to ensure the reliability of the risk level RLV that is set based on the potential risk data DRSK.

[0047] In addition, according to the level setting process, all potential risk data DRSK in a predetermined section PSEC corresponding to a road section SEC of the regular data DREG is extracted, and the variance σ 2 is calculated. And the variance σ 2 If it is determined that the potential risk data DRSK exceeds the threshold value TH, the risk level RLV is set. In other words, the risk level RLV is set only if it is determined that the potential risk data DRSK is not biased. Therefore, according to the embodiment, it is also possible to improve the reliability of the risk level RLV set based on the potential risk data DRSK. [Explanation of symbols]

[0048] 1. Risk level setting device 2. Driving history database 3 Potential risk database 4 Data Processing Device 5 Processors 6. Memory RD,RD1~RD6 Road RLV Risk Level RSK Potential Risk SEC,SEC1~SEC5 Road section PSEC designated section DDEC deceleration behavior data DEXT External situation data DHIS Driving History Data DIRE Irregular Data DSEC Road Section Data DREG Regular Data DRSK Potential Risk Data

Claims

[Claim 1] a driving history database in which driving history data of a vehicle is stored, the driving history data including deceleration behavior data associated with each of a plurality of road sections and external condition data associated with each of the plurality of road sections; a potential risk database storing potential risk data indicating potential traffic safety risks in a specific section preselected from among the road sections, the potential risk data being calculated based on static deceleration factor data in the specific section; a data processing device that performs a level setting process to set a risk level indicating a level of risk that a vehicle passing through the predetermined section will encounter another moving object while passing through the predetermined section, based on the driving history data and the potential risk data; Equipped with The data processing device, in the level setting process, Extracting the driving history data for a road section corresponding to the predetermined section from the driving history database; Identifying irregular data in which the deceleration behavior data in the predetermined section is derived from a dynamic deceleration factor from the driving history data based on the external condition data included in the extracted driving history data; excluding the irregular data from the extracted driving history data and selecting regular data in which the deceleration behavior data in the predetermined section is derived from the static deceleration cause; extracting all of the potential risk data in a predetermined section corresponding to the road section of the regular data from the potential risk database; Calculate the variance of the extracted potential risk data; determining whether the variance exceeds a threshold; When it is determined that the variance exceeds the threshold, the risk level is set for a predetermined section corresponding to the road section of the regular data. A risk level setting device comprising:

Images