JP7024352B2 - Anomaly estimation device - Google Patents

Description

The present invention relates to an abnormality estimation device that estimates an abnormality in road conditions.

The device disclosed in Patent Document 1 stores data of driving behavior patterns when a large number of drivers drive a vehicle as a normal driving model. Then, the normal driving model is compared with the driving pattern data representing the driving behavior pattern of a certain vehicle, and the degree of abnormality in the driving behavior of the vehicle is calculated.

Japanese Unexamined Patent Publication No. 2016-118906

In recent years, research and development of self-driving cars has been actively carried out, and it is expected that self-driving cars will run on public roads in the near future. The level is defined for automatic driving. In the present embodiment, when simply described as automatic driving, it means automatic driving of level 3 or higher, which is a level at which the driver does not normally need to perform a driving operation.

Further, even in an autonomous vehicle, depending on the situation, driving by the driver (hereinafter referred to as manual driving) may be performed instead of automatic driving. The state in which automatic driving is performed in the automatic driving vehicle is defined as the automatic driving mode. On the other hand, the state in which manual driving is performed in the self-driving car is set to the manual driving mode. The manual driving mode in the self-driving car is no different from the manual driving in the car not equipped with the self-driving function. Therefore, the manual driving mode is also set to the state in which the manual driving is performed in the vehicle not equipped with the automatic driving function.

In the automatic driving mode, the main body that controls the driving behavior of the vehicle is a control device, and a control device that executes the same algorithm is mounted on a large number of automatic driving vehicles. Therefore, the driving behavior of the vehicle in the automatic driving mode is similar between different vehicles. On the other hand, in the manual driving mode, individual differences are reflected in the driving behavior, so that the driving behavior has a relatively large variation as compared with the automatic driving mode.

Patent Document 1 does not distinguish between the subjects that control the driving behavior of the vehicle. Therefore, there is room for improvement in the accuracy of estimating road abnormalities.

The present disclosure has been made based on this circumstance, and an object thereof is to provide an anomaly estimation device capable of more accurately estimating an abnormality on a road.

The above object is achieved by a combination of the features described in the independent claims, and the sub-claims specify further advantageous specific examples. The reference numerals in parentheses described in the claims indicate, as one embodiment, the correspondence with the specific means described in the embodiments described later, and do not limit the disclosed technical scope.

One disclosure to achieve the above objectives is
An abnormality estimation device (200, 1100) that estimates abnormalities in road conditions.
A driving feature amount acquisition unit (201, 1102) that acquires a driving feature amount, which is a feature amount of a vehicle's driving behavior, in association with a grid that divides a region in which the vehicle travels.
With respect to the grid associated with the driving feature amount acquired by the driving feature amount acquisition unit, the driving model acquisition unit (206) that acquires the driving model representing the normal driving behavior, and the driving model acquisition unit (206).
An abnormality degree calculation unit (207) that calculates the degree of abnormality of the driving feature amount with respect to the driving model by comparing the driving feature amount acquired by the driving feature amount acquisition unit with the driving model acquired by the driving model acquisition unit.
It is equipped with an output unit (209, 1103) that outputs information that estimates road abnormalities based on the degree of abnormality.
The driving feature amount acquisition unit acquires the driving feature amount as well as whether the driving mode of the vehicle is the automatic driving mode or the manual driving mode.
The driving model acquisition unit acquires a driving model in the automatic driving mode when the driving feature amount is the driving feature amount in the automatic driving mode, and when the driving feature amount is the driving feature amount in the manual driving mode. Gets the driving model in manual driving mode ,
It is equipped with an operation feature amount storage unit (203, 204) that stores the operation feature amount in the automatic operation mode and the operation feature amount in the manual operation mode for each grid.
The driving model is determined based on the number of driving features accumulated in the driving feature storage section, which is equal to or larger than the reference number.
For the grid associated with the driving feature amount acquired by the driving feature amount acquisition unit, if the driving feature amount in the automatic operation mode stored in the driving feature amount storage unit does not reach the reference number, the driving model acquisition unit will perform the operation model acquisition unit. , About the grid associated with the operation feature amount acquired by the operation feature amount acquisition unit The operation model in the manual operation mode stored in the operation feature amount storage unit is acquired as the operation model in the automatic operation mode. It is an estimation device .
Another disclosure to achieve the above objectives is
An abnormality estimation device (200, 1100) that estimates abnormalities in road conditions.
A driving feature amount acquisition unit (201, 1102) that acquires a driving feature amount, which is a feature amount of a vehicle's driving behavior, in association with a grid that divides a region in which the vehicle travels.
With respect to the grid associated with the driving feature amount acquired by the driving feature amount acquisition unit, the driving model acquisition unit (206) that acquires the driving model representing the normal driving behavior, and the driving model acquisition unit (206).
An abnormality degree calculation unit (207) that calculates the degree of abnormality of the driving feature amount with respect to the driving model by comparing the driving feature amount acquired by the driving feature amount acquisition unit with the driving model acquired by the driving model acquisition unit.
It is equipped with an output unit (209, 1103) that outputs information that estimates road abnormalities based on the degree of abnormality.
The driving feature amount acquisition unit acquires the driving feature amount as well as whether the driving mode of the vehicle is the automatic driving mode or the manual driving mode.
The driving model acquisition unit acquires a driving model in the automatic driving mode when the driving feature amount is the driving feature amount in the automatic driving mode, and when the driving feature amount is the driving feature amount in the manual driving mode. Gets the driving model in manual driving mode,
It is equipped with an operation feature amount storage unit (203, 204) that stores the operation feature amount in the automatic operation mode and the operation feature amount in the manual operation mode for each grid.
The driving model is determined based on the number of driving features accumulated in the driving feature storage section, which is equal to or larger than the reference number.
When the grid for determining the driving model, which is determined based on the grid associated with the driving feature quantity acquired by the driving feature quantity acquisition unit, is used as the model extraction grid.
When the driving feature amount in the automatic driving mode associated with the model extraction grid, which is stored in the driving feature amount storage unit, does not reach the reference number, the driving model acquisition unit runs the vehicle in the model extraction grid. The driving feature amount in the automatic driving mode associated with the grid including other roads having a shape similar to the road supplements the driving feature amount in the automatic driving mode associated with the model extraction grid. It is an abnormality estimation device that determines.
Another disclosure to achieve the above objectives is
An abnormality estimation device (200, 1100) that estimates abnormalities in road conditions.
A driving feature amount acquisition unit (201, 1102) that acquires a driving feature amount, which is a feature amount of a vehicle's driving behavior, in association with a grid that divides a region in which the vehicle travels.
With respect to the grid associated with the driving feature amount acquired by the driving feature amount acquisition unit, the driving model acquisition unit (206) that acquires the driving model representing the normal driving behavior, and the driving model acquisition unit (206).
An abnormality degree calculation unit (207) that calculates the degree of abnormality of the driving feature amount with respect to the driving model by comparing the driving feature amount acquired by the driving feature amount acquisition unit with the driving model acquired by the driving model acquisition unit.
It is equipped with an output unit (209, 1103) that outputs information that estimates road abnormalities based on the degree of abnormality.
The driving feature amount acquisition unit acquires the driving feature amount as well as whether the driving mode of the vehicle is the automatic driving mode or the manual driving mode.
The driving model acquisition unit acquires a driving model in the automatic driving mode when the driving feature amount is the driving feature amount in the automatic driving mode, and when the driving feature amount is the driving feature amount in the manual driving mode. Gets the driving model in manual driving mode,
Pattern 1 is when the abnormality degree calculation unit calculates the degree of abnormality by comparing the operation feature amount in the manual operation mode with the operation model, and the abnormality degree is compared with the operation feature amount in the automatic operation mode and the operation model. When pattern 2 is calculated, and pattern 3 is the case where the degree of abnormality is calculated by comparing the driving feature amount and the driving model for the automatic operation mode and the manual operation mode, respectively.
The output unit is an abnormality estimation device that outputs information based on the degree of abnormality so that pattern 1, pattern 2, and pattern 3 can be distinguished.

As described above, when the driving feature amount is the driving feature amount in the automatic driving mode, the driving model acquisition unit acquires the driving model in the automatic driving mode, and the driving feature amount is the driving in the manual driving mode. If it is a feature quantity, the operation model in the manual operation mode is acquired. Therefore, in the abnormality degree calculation unit, the operation feature amount in the automatic operation mode is compared with the operation model in the automatic operation mode, or the operation feature amount in the manual operation mode is compared with the operation model in the manual operation mode. Will be done.

As described above, the abnormality estimation device separates the manual operation mode and the automatic operation mode, and compares the operation feature amount and the operation model. Therefore, it is taken into consideration that the variation of the operation model differs between the automatic operation mode and the manual operation mode. Therefore, the abnormality of the road can be estimated more accurately.

It is a figure which shows the whole structure of an abnormality estimation system 10. It is a figure which shows the structure of the vehicle side apparatus 100 and the center side apparatus 200. It is explanatory drawing which shows the correspondence with driving data, driving scene, and topic ratio 210. It is a figure which shows the distribution of the topic ratio 210 stored in the manual operation data storage part 203. It is a figure which shows the distribution of the topic ratio 210 stored in the automatic operation data storage part 204. It is a flowchart which shows the process which the operation model acquisition part 206 of FIG. 2 executes. It is a figure which illustrates the abnormal road condition map 220. It is a figure which shows another example of the abnormal road condition map 220. It is a figure which shows the information provided by the information providing part 105. It is a figure which shows the information provided by the information providing part 105. It is a figure which shows the information provided by the information providing part 105. It is a figure which shows the information provided by the information providing part 105. It is a figure which shows the information provided by the information providing part 105. It is a figure which shows the information provided by the information providing part 105. It is a figure which shows the structure of the vehicle side apparatus 1100 and the center side apparatus 1200 of 2nd Embodiment.

<First Embodiment>
[overall structure]
As shown in FIG. 1, the abnormality estimation system 10 includes a vehicle-side device 100 and a center-side device 200. In the first embodiment, the center-side device 200 has a function as an abnormality estimation device.

The vehicle-side device 100 is mounted on a vehicle 12 such as a passenger car. On the other hand, the center-side device 200 is provided in the center 13 outside the vehicle. The vehicle-side device 100 and the center-side device 200 are configured to enable data communication. For example, the vehicle-side device 100 and the center-side device 200 are connected to each other via a public communication line network or an Internet network of a mobile phone company (not shown) to perform data communication. Further, the communication may be performed by connecting to each other via the roadside unit.

Communication between the vehicle-side device 100 and the center-side device 200 is possible in both directions. That is, it is possible to transmit a signal from the vehicle-side device 100 to the center-side device 200, and to transmit a signal from the center-side device 200 to the vehicle-side device 100.

The vehicle-side device 100 and the center-side device 200 include a microcomputer including a semiconductor memory such as a CPU, RAM, ROM, and flash memory. The number of microcomputers included in the vehicle-side device 100 and the center-side device 200 may be one or a plurality. Each function realized by the vehicle-side device 100 and the center-side device 200 is realized by the CPU executing a program stored in a non-transitional substantive recording medium such as a ROM, a flash memory, or a hard disk. When this program is executed, the method corresponding to the program is executed. However, the method by which the vehicle-side device 100 and the center-side device 200 realize each function is not limited to software, and some or all of the elements thereof are combined with hardware such as a logic circuit or an analog circuit. It may be realized.

[Structure of vehicle-side device 100]
As shown in FIG. 2, the vehicle-side device 100 includes a driving data acquisition unit 101, a symbolization processing unit 102, a vehicle signal transmission unit 103, an abnormal road condition map reception unit 104, and an information provision unit 105.

The driving data acquisition unit 101 repeatedly acquires driving operation data related to the driving operation by the driver and vehicle behavior data related to the behavior of the vehicle 12 appearing as a result of the driving operation via various sensors mounted on the vehicle 12. Further, the driving data acquisition unit 101 generates differential data obtained by differentiating the driving operation data and the vehicle behavior data, and outputs multidimensional data including the driving operation data, the vehicle behavior data, and the differential data as driving data. ..

It is desirable that the driving data acquisition unit 101 normalizes and outputs the driving data with the normalization parameters prepared in advance for each vehicle type or vehicle type. By normalizing the driving data, it is possible to absorb the difference in vehicle type and vehicle type.

As the driving operation data, for example, the accelerator opening according to the amount of depression of the accelerator pedal, the brake pressure according to the amount of depression of the brake pedal, the steering angle of the steering wheel, and the like can be used. As vehicle behavior data, for example, vehicle speed, yaw rate, acceleration, and the like can be used.

Further, the driving data acquisition unit 101 acquires the position of the vehicle 12 from a position detection device such as a GPS device. Further, the traveling route of the vehicle to the destination may be acquired from a navigation device or the like, or image data obtained by capturing the surroundings of the vehicle 12 may be acquired from an in-vehicle camera. Travel routes and image data are also examples of driving data.

The symbolization processing unit 102 finds a time-series pattern that repeatedly occurs from the driving data output by the driving data acquisition unit 101, and sets a series of vehicle driving behaviors as a driving scene. FIG. 3 illustrates driving data and driving scenes that can be extracted from the driving data. In FIG. 3, five driving scenes are extracted from the driving data. Since the technique of collecting the driving data that can be acquired in time series into the driving scene is described in Patent Document 1 and the like, further detailed description thereof will be omitted.

The symbolization processing unit 102 further extracts a driving feature amount for each driving scene. The driving feature amount is a value representing the feature of driving behavior. For example, one or more of the accelerator opening, the brake pressure, the steering angle, the vehicle speed, the yaw rate, and their differential data are used as driving data, and the topic ratio 210 is used as the driving feature amount of each driving scene. FIG. 3 also shows the topic ratio 210 determined for each driving scene.

When the topic ratio 210 is used as the driving feature amount, the division unit of the range of the driving data is set as a bin as the feature amount distribution of each driving data in each segmented driving scene, and the appearance frequency of the driving data in each bin is set. A representational histogram is used.

The symbolization processing unit 102 prepares in advance a plurality of basal feature distributions representing characteristic patterns of driving data used for expressing one or more types of feature quantity distributions for one or more types of driving data as driving topics. I will do it. Then, a plurality of basis feature distributions are used to express one or more types of feature amount distributions, and a topic ratio 210, which is a low-order vector represented by a mixture ratio of the basis feature distributions, is calculated. Since this topic ratio 210 is also described in Patent Document 1 and the like, further detailed description thereof will be omitted.

The driving feature amount is not limited to the above, and a histogram showing the appearance frequency for each of the above-mentioned driving data may be used as the feature amount distribution, and the feature amount distribution of one or more types of driving data may be arranged and used as the driving feature amount.

The vehicle signal transmission unit 103 outputs a vehicle signal including the topic ratio 210 obtained by the symbolization processing unit 102, the vehicle position from which the topic ratio 210 is obtained, and the driver main information of the vehicle 12 to the center side device 200. Send to. The vehicle position is expressed in coordinates. The driving subject state is information indicating whether the driving subject of the vehicle 12 is a control device or a person. Specifically, the vehicle 12 is an automatic driving vehicle or is equipped with a manual driving vehicle, that is, an automatic driving function. The information indicating the vehicle type indicating the non-driving vehicle and the driving mode is the driving subject information. The operation mode is either an automatic operation mode or a manual operation mode.

The abnormal road condition map receiving unit 104 receives the abnormal road condition map 220 (see FIGS. 7 and 8), which is a map that visualizes the abnormal portion of the road generated by the center side device 200, from the center side device 200. Then, an image signal or an audio signal based on the received abnormal road condition map 220 is generated and sent to the information providing unit 105.

The information providing unit 105 is configured to include a display and a speaker, and provides an image or sound based on an image signal or an audio signal received from the abnormal road condition map receiving unit 104 to the occupants of the vehicle 12.

[Configuration of center side device 200]
The center-side device 200 includes a vehicle signal receiving unit 201, a vehicle determination unit 202, a manual driving data storage unit 203, an automatic driving data storage unit 204, a driving model acquisition unit 206, an abnormality determination unit 207, and an abnormal road condition map creation unit 208. It is equipped with an abnormal road condition map transmission unit 209 and a road map storage unit 230.

The road map storage unit 230 stores the road map data and the grid which is the division when the area on the road map is divided. For example, a grid is an area on a map divided by 10 m square.

The vehicle signal receiving unit 201 receives a vehicle signal from the vehicle side device 100. The vehicle signal receiving unit 201 identifies the grid including the vehicle position included in the received vehicle signal with reference to the grid stored in the road map storage unit 230. By this process, the topic ratio 210, which is a driving feature amount, is associated with the grid. The vehicle signal receiving unit 201 corresponds to a driving feature amount acquisition unit. The vehicle signal receiving unit 201 sends the received vehicle signal and the specified grid to the vehicle determination unit 202 and the abnormality determination unit 207.

Based on the driving subject information included in the vehicle signal, the vehicle determination unit 202 determines whether the topic ratio 210 included in the vehicle signal is when the vehicle 12 is traveling in the automatic driving mode. , It is determined whether the vehicle 12 is traveling in the manual driving mode. If the topic ratio 210 is when the vehicle 12 is traveling in the automatic driving mode, the topic ratio 210 is stored in the automatic driving data storage unit 204. On the other hand, the topic ratio 210 when the vehicle 12 is traveling in the manual driving mode is stored in the manual driving data storage unit 203. In any case, the topic ratio 210 is saved together with the vehicle position or is saved in association with the grid determined from the vehicle position.

The manual operation data storage unit 203 stores the topic ratio 210 when traveling in the manual operation mode. In the automatic driving data storage unit 204, the topic ratio 210 when the vehicle is running in the automatic driving mode is stored. Since the topic ratio 210 corresponds to the operation feature amount, the manual operation data storage unit 203 and the automatic operation data storage unit 204 that store the topic ratio 210 correspond to the operation feature amount storage unit.

FIG. 4 conceptually shows the distribution of the topic ratio 210 stored in the manual operation data storage unit 203, and FIG. 5 conceptually shows the distribution of the topic ratio 210 stored in the automatic operation data storage unit 204. Although FIGS. 4 and 5 are two-dimensional coordinate systems, the topic ratio 210 has the same number of dimensions as the number of parameters included in the topic ratio 210.

The driving behavior cannot be expressed only by the topic ratio 210, and the driving behavior of the vehicle 12 can be expressed by the topic ratio 210 and each driving topic constituting the topic ratio 210. However, each driving topic itself is prepared in advance. Therefore, the topic ratio 210 may be stored in the manual operation data storage unit 203 and the automatic operation data storage unit 204.

The topic ratio 210 stored in the manual operation data storage unit 203 or the automatic operation data storage unit 204 may be only the topic ratio 210 that is not determined to be abnormal by the abnormality determination unit 207 described later. Further, the topic ratio 210 after a certain period of time may be deleted from the manual operation data storage unit 203 or the automatic operation data storage unit 204.

The driving model acquisition unit 206 identifies the grid from the vehicle position included in the vehicle signal received by the vehicle signal receiving unit 201. Then, the operation model for the grid is determined and provided to the abnormality determination unit 207.

The processing of the operation model acquisition unit 206 will be described with reference to FIG. The process shown in FIG. 6 starts when the vehicle position is supplied from the vehicle signal receiving unit 201 and the vehicle type and the driving mode are acquired from the vehicle determination unit 202.

Step (hereinafter, step is omitted) In S1, the vehicle type is acquired. In S2, it is determined whether or not the vehicle type acquired in S1 is an autonomous driving vehicle. If the vehicle type is an autonomous driving vehicle, the process proceeds to S3, and if the vehicle type is a manually driven vehicle, the process proceeds to S4.

In S3, it is determined whether or not the operation mode is manual operation. If it is a manual operation mode, the process proceeds to S4 to generate a manual operation model. Further, even in the case of a manually driven vehicle, since S4 is executed, a manually driven model is generated. The manual operation model is an operation model in the manual operation mode, and the automatic operation model is an operation model in the automatic operation mode. After executing S4, the process proceeds to S8.

In this embodiment, the operation model is represented by a variance 211 and an average vector 212 as shown in FIG. In order to calculate the variance 211 and the average vector 212, the driving model acquisition unit 206 first determines the vehicle position grid in S4. The vehicle position grid is a grid containing vehicle positions included in the vehicle signal. Then, a total of 9 grids of this vehicle position grid and 8 grids surrounding the vehicle position grid are used as model extraction grids. When the topic ratio 210 is equal to or larger than the reference number in this model extraction grid, the variance 211 and the average vector 212 are calculated. The variance 211 is represented by a variance-covariance matrix of a set of topic ratios 210. The variance 211 and the average vector 212 are probabilistic models of normal driving behavior.

If the number of topic ratios 210 is less than the reference number, the operation model acquisition unit 206 does not calculate the operation model. This is because the reliability of the operation model is not high when the number of data is small. The reference number is determined based on the reliability required for the driving model. The reference number may be set as appropriate, and is, for example, 10.

As a result of the determination of S3, if it is the automatic operation mode, the process proceeds to S5 and an automatic operation model is generated. In S5, as in S4, the vehicle position grid and the model extraction grid are first determined. Then, when the determined model extraction grid has a topic ratio 210 equal to or larger than the reference number, the variance 211 and the average vector 212 are calculated. FIG. 5 shows a variance 211 and an average vector 212, which are automatic driving models.

As can be seen from the comparison of FIGS. 4 and 5, the degree of variation in the topic ratio 210 differs between the manual operation mode and the automatic operation mode even in the normal operation mode. Therefore, in the present embodiment, the driving model is calculated by distinguishing between the topic ratio 210 obtained when driving in the manual driving mode and the topic ratio 210 obtained when driving in the automatic driving mode.

However, as described above, if the number of topic ratios 210 does not reach the reference number, the operation model is not calculated. Therefore, in S6, it is determined whether or not the amount of data is insufficient to generate the automatic operation model.

If the judgment of S6 is NO, the process proceeds to S8. On the other hand, if the judgment of S6 is YES, the process proceeds to S7. In S7, a process of substituting the operation model is executed. In the present embodiment, the first substitute process and the second substitute process described below are possible. When both the first substitute process and the second substitute process are possible, which one should be prioritized may be set in advance.

The first substitute process uses the topic ratio 210 stored in the manual operation data storage unit 203 for the model extraction grid. When self-driving cars are not widespread, it is assumed that there is a large amount of data with a topic ratio of 210 for the manual driving mode for the same grid. Therefore, this first substitute process is executed.

The second substitute process searches for a grid in the model extraction grid that includes other roads that are similar in shape to the road on which the vehicle 12 travels. Whether or not the shapes are similar is judged from the curvature of the road, the number of lanes, and the like. The grid found as a result of the search is used as a substitute grid.

Then, the topic ratio 210 stored in the automatic operation data storage unit 204 and the topic ratio 210 stored in the substitute grid supplements the topic ratio 210 stored in the model extraction grid. Calculate the driving model.

In S8, the operation model determined as described above is provided to the abnormality determination unit 207. The explanation is returned to FIG.

The abnormality determination unit 207 compares the driving model provided by the driving model acquisition unit 206 with the topic ratio 210 included in the vehicle signal provided by the vehicle signal receiving unit 201. From the result of this comparison, how much the topic ratio 210 deviates from the driving model is calculated as the degree of abnormality. Since the abnormality degree is calculated, the abnormality determination unit 207 is also an abnormality degree calculation unit.

The degree of anomaly is, for example, a value obtained by normalizing the distance between the topic ratio 210 and the average vector 212 this time by a variance-covariance matrix. The abnormality determination unit 207 further compares the calculated abnormality degree with the threshold value, and if the abnormality degree exceeds the threshold value, determines that the topic ratio 210 this time is abnormal, and if the abnormality degree does not exceed the threshold value. It is determined that the topic ratio 210 this time is normal. The determination of the degree of abnormality can be used depending on whether or not the topic ratio 210 is stored in the manual operation data storage unit 203 or the automatic operation data storage unit 204.

The abnormal road condition map creation unit 208 acquires the degree of abnormality and the grid on which the degree of abnormality is calculated from the abnormality determination unit 207. Then, the cumulative abnormality degree is calculated for each grid. For the cumulative abnormality degree, for example, as described in Patent Document 1, the latest predetermined number of abnormality degrees in the nine grids centered on the grid on which the abnormality degree is calculated are selected, and the selected abnormality degree is selected. Is calculated by increasing the weight as the degree of abnormality increases and adding. Then, when the cumulative abnormality degree exceeds the cumulative determination value, it is determined that the road condition on the road on which the vehicle 12 included in the grid has traveled is abnormal.

However, in the present embodiment, the cumulative abnormality degree is calculated by distinguishing whether the operation mode is the automatic operation mode or the manual operation mode. Then, it is also determined whether or not the cumulative determination value is exceeded by distinguishing between the automatic operation mode and the manual operation mode.

The abnormal road condition map creation unit 208 creates an abnormal road condition map 220 showing a grid in which the cumulative abnormality degree exceeds the cumulative determination value. The abnormal road condition map 220 is information that estimates the abnormalities of the road.

The abnormal road condition map 220 has both a pattern 1 in which the cumulative abnormality degree exceeds the cumulative determination value is only the manual driving mode, a pattern 2 in which only the automatic driving mode is used, and both the manual driving mode and the automatic driving mode. Pattern 3 is shown separately.

FIG. 7 shows an example of the mode of distinction. In FIG. 7, patterns 1, 2, and 3 are distinguished by shades of color. It should be noted that the color may be distinguished by the difference in color instead of the shade of color. When distinguishing patterns 1, 2 and 3 by the difference in color, for example, pattern 1 is green or blue, pattern 2 is yellow, and pattern 3 is red.

FIG. 8 is another example of the mode of distinguishing patterns 1, 2, and 3. In FIG. 8, pattern 1 is indicated by “?”, Pattern 2 is indicated by “!”, And pattern 3 is indicated by a color that calls attention (for example, red) although it has no symbol.

In the manual driving mode, even if the driver travels on the same road, the driving behavior varies greatly depending on the driver. Therefore, even if the cumulative abnormality degree exceeds the cumulative determination value, the reason may be due to the variation in the driving behavior of the driver. On the other hand, in the automatic driving mode, there is little variation in driving behavior among vehicles when traveling on the same road. Therefore, when the cumulative abnormality degree exceeds the cumulative determination value in the automatic driving mode, it is more likely that there is an abnormality in the road condition than when the cumulative abnormality degree exceeds the cumulative determination value in the manual driving mode. Therefore, the pattern 2 is shown in a mode in which the degree of attention is higher than that of the pattern 1. In pattern 3, in both the manual driving mode and the automatic driving mode, the cumulative abnormality degree exceeds the cumulative judgment value, so that the road condition is most likely to be abnormal in patterns 1, 2, and 3. .. Therefore, the pattern 3 is shown in a mode in which the degree of attention is higher than that of the pattern 2.

The abnormal road condition map transmission unit 209 corresponds to an output unit, and transmits the abnormal road condition map 220 created by the abnormal road condition map creation unit 208 to the vehicle side device 100.

In the vehicle-side device 100, the abnormal road condition map receiving unit 104 receives the abnormal road condition map 220. Then, an image signal or an audio signal based on the received abnormal road condition map 220 is generated. The image signal to be generated is, for example, a signal indicating an abnormal road condition map 220 including a point where the vehicle 12 is traveling. In addition, a message indicating that there is an abnormality in the road in the direction of travel may be generated. Even when the message is generated, the patterns 1, 2, and 3 are generated in a distinguishable manner. 9, 10, and 11 are examples in which the background colors of the messages are different in patterns 1, 2, and 3. 9 is an example of the case of pattern 1, FIG. 10 is an example of the case of pattern 2, and FIG. 11 is an example of the case of pattern 3. As the specific background color, as in FIG. 7, the color in which the degree of attention is increased in the order of patterns 1, 2, and 3.

12, 13, and 14 are examples in which the expression of the message is expressed in the order of patterns 1, 2, and 3, and the degree of attention is increased. The image signal and audio signal generated in this way are output to the information providing unit 105. Therefore, the occupants of the vehicle 12 are also notified of the abnormal state of the road by distinguishing the patterns 1, 2, and 3.

[Summary of embodiments]
In the first embodiment described above, the driving model acquisition unit 206 uses the driving model in the automatic driving mode when the topic ratio 210 acquired this time from the vehicle 12 is the topic ratio 210 in the automatic driving mode. If the topic ratio is 210 in the manual operation mode, it is decided to use the operation model in the manual operation mode. Therefore, the abnormality determination unit 207 compares the topic ratio 210 in the automatic operation mode with the operation model in the automatic operation mode, or compares the topic ratio 210 in the manual operation mode with the operation model in the manual operation mode. Will be done.

That is, the center-side device 200 separates the manual operation mode and the automatic operation mode, and compares the topic ratio 210 with the operation model. Therefore, it is taken into consideration that the variation of the operation model differs between the automatic operation mode and the manual operation mode. Therefore, the abnormality of the road can be estimated more accurately.

Further, in the present embodiment, when S6 in FIG. 6 is YES, that is, when it is determined to use the automatic driving model but the automatic driving model cannot be used due to insufficient data amount, a process of substituting the driving model is executed. By executing the process of substituting the operation model, the abnormality determination unit 207 can perform the abnormality determination even when the automatic operation model cannot be used.

<Second Embodiment>
Next, the second embodiment will be described. In the following description of the second embodiment, the element having the same number as the code used so far is the same as the element having the same code in the previous embodiments, unless otherwise specified. Further, when only a part of the configuration is described, the embodiment described above can be applied to the other parts of the configuration.

FIG. 15 shows a configuration diagram of the abnormality estimation system 1000 of the second embodiment. The abnormality estimation system 1000 includes a vehicle-side device 1100 and a center-side device 1200. The vehicle-side device 1100 is mounted on each of the plurality of vehicles 12 as in the vehicle-side device 100 of the first embodiment. In the second embodiment, the vehicle-side device 1100 has a function as an abnormality estimation device.

The vehicle-side device 1100 includes the same driving data acquisition unit 101, abnormal road condition map receiving unit 104, and information providing unit 105 as in the first embodiment. Further, it includes a symbolization processing unit 1102 whose processing is similar to that of the symbolization processing unit 102 of the first embodiment.

Similar to the symbolization processing unit 102 of the first embodiment, the symbolization processing unit 1102 calculates the topic ratio 210, which is an operating feature amount. The topic ratio 210 and the grid from which the topic ratio 210 is obtained are associated with each other and provided to the abnormality determination unit 207. Further, the symbolization processing unit 1102 also acquires whether the driving mode of the vehicle 12 is the automatic driving mode or the manual driving mode from the control device mounted on the vehicle 12 and controlling the behavior of the vehicle 12, and the acquired driving. The mode is also provided to the abnormality determination unit 207. Further, the acquired operation mode is also provided to the operation model acquisition unit 206. In the second embodiment, the symbolization processing unit 1102 corresponds to the operation feature amount acquisition unit.

In addition, the vehicle-side device 1100 includes a manual driving data storage unit 203, an automatic driving data storage unit 204, a driving model acquisition unit 206, and an abnormality determination unit 207 provided in the center-side device 200 in the first embodiment. .. The contents stored in the manual operation data storage unit 203 and the automatic operation data storage unit 204 are transmitted from the center side device 1200 at appropriate timings such as periodic or the timing requested by the vehicle side device 1100. Further, it may be stored in the vehicle-side device 1100 in advance.

Further, the vehicle-side device 1100 includes a vehicle signal transmission unit 1103. When the abnormality determination unit 207 determines that the road condition of the road on which the vehicle 12 has traveled is abnormal, the vehicle signal transmission unit 1103 centered on a signal indicating the degree of abnormality and a grid on which the degree of abnormality is calculated. It is transmitted to the side device 1200. In the second embodiment, the vehicle signal transmission unit 1103 corresponds to the output unit. Further, the signal indicating the degree of abnormality and the grid on which the degree of abnormality is calculated is the information for estimating the abnormality of the road.

The center-side device 1200 includes the same abnormal road condition map creating unit 208, abnormal road condition map transmitting unit 209, manual driving data storage unit 203, and automatic driving data storage unit 204 as in the first embodiment.

In addition, the center-side device 1200 includes a vehicle signal receiving unit 1201 and a driving model transmitting unit 1210. The vehicle signal receiving unit 1201 receives a signal indicating an abnormality degree transmitted from the vehicle side device 1100 and a grid on which the abnormality degree is calculated. Then, the received abnormality degree and the grid on which the abnormality degree is calculated are output to the abnormal road condition map creating unit 208. As a result, the abnormal road condition map creating unit 208 can create the abnormal road condition map 220 as in the first embodiment.

The driving model transmission unit 1210 transmits the contents stored in the manual driving data storage unit 203 and the automatic driving data storage unit 204 to the vehicle-side device 1100 at the appropriate timing described above. In the second embodiment, the stored contents of the manual operation data storage unit 203 and the automatic operation data storage unit 204 are pre-stored by the method described in the first embodiment or the like.

In this second embodiment, the vehicle-side device 1100 determines the road abnormality, and when the vehicle-side device 1100 determines the road abnormality, a signal is transmitted from the vehicle-side device 1100 to the center-side device 1200. Therefore, the amount of communication can be suppressed as compared with the abnormality estimation system 10 of the first embodiment.

Although the embodiments have been described above, the disclosed technology is not limited to the above-described embodiment, and the following modifications are also included in the disclosed scope, and further, within the scope not deviating from the gist other than the following. It can be changed in various ways.

<Modification 1>
In the embodiment, the topic ratio 210 when the driving data is represented by the driving topic is used as the driving feature amount, but the driving data may be used as the driving feature amount.

<Modification 2>
In the first embodiment, the coordinates of the vehicle 12 are used as the vehicle position to be included in the vehicle signal, but the grid on which the vehicle 12 is located may be the vehicle position and the grid may be included in the vehicle signal.

<Modification 3>
In the embodiment, as the degree of anomaly, a value obtained by normalizing the distance between the topic ratio 210 and the average vector 212 this time by a variance-covariance matrix is shown. However, the deviance degree disclosed in Patent Document 1 may be used as the degree of abnormality.

<Modification example 4>
In the embodiment, an example in which the model extraction grid is nine grids including the vehicle position grid has been described. However, the number and size of the model extraction grid can be various, such that the model extraction grid may be only the vehicle position grid.

<Modification 5>
In the embodiment, the abnormality determination unit 207 compares the cumulative abnormality degree with the cumulative determination value. However, the cumulative anomaly may be output without performing this comparison.

<Modification 6>
In the embodiment, the operation model is calculated every time the degree of abnormality is determined. However, the present invention is not limited to this, and the operation model may be calculated in advance before determining the degree of abnormality. For example, the driving model may be calculated each time the vehicle signal is received once or a predetermined time greater than one. The calculated driving model is stored in the manual driving data storage unit 203 or the automatic driving data storage unit 204 in association with the vehicle position grid. In this case, the driving model acquisition unit 206 can identify the vehicle position from the vehicle position included in the vehicle signal received by the vehicle signal receiving unit 201 from the manual driving data storage unit 203 or the automatic driving data storage unit 204. Get a driving model for the grid.

10: Abnormality estimation system 12: Vehicle 13: Center 100: Vehicle side device 101: Driving data acquisition unit 102: Symbolization processing unit 103: Vehicle signal transmission unit 104: Abnormal road condition map reception unit 105: Information provision unit 200: Center Side device (abnormality estimation device) 201: Vehicle signal receiving unit (driving feature amount acquisition unit) 202: Vehicle determination unit 203: Manual driving data storage unit (driving feature amount storage unit) 204: Automatic driving data storage unit (driving feature amount) Accumulation unit) 206: Driving model acquisition unit 207: Abnormality judgment unit 208: Abnormal road condition map creation unit 209: Abnormal road condition map transmission unit (output unit) 210: Topic ratio 211: Dispersion 212: Average vector 220: Abnormal road condition Map 230: Road map storage unit 1000: Abnormality estimation system 1100: Vehicle side device (abnormality estimation device) 1102: Symbolization processing unit (driving feature amount acquisition unit) 1103: Vehicle signal transmission unit (output unit) 1200: Center side device 1201: Vehicle signal receiver 1210: Driving model transmitter

Claims (9)

An abnormality estimation device (200, 1100) that estimates abnormalities in road conditions.
A driving feature amount acquisition unit (201, 1102) that acquires a driving feature amount, which is a feature amount of a vehicle's driving behavior, in association with a grid that divides a region in which the vehicle travels.
With respect to the grid associated with the driving feature amount acquired by the driving feature amount acquisition unit, a driving model acquisition unit (206) that acquires a driving model representing the driving behavior in a normal state, and a driving model acquisition unit (206).
An abnormality degree calculation unit (anomaly degree calculation unit) that compares the operation feature amount acquired by the operation feature amount acquisition unit with the operation model acquired by the operation model acquisition unit and calculates the degree of abnormality of the operation feature amount with respect to the operation model. 207) and
It is provided with an output unit (209, 1103) that outputs information for estimating road abnormalities based on the degree of abnormality.
The driving feature amount acquisition unit acquires, together with the driving feature amount, whether the driving mode of the vehicle is the automatic driving mode or the manual driving mode.
When the driving feature amount is the driving feature amount in the automatic driving mode, the driving model acquisition unit acquires the driving model in the automatic driving mode, and the driving feature amount is the manual driving mode. If it is the operation feature amount of the above, the operation model in the manual operation mode is acquired , and the operation model is acquired.
The operation feature amount accumulating unit (203, 204) which stores the operation feature amount in the automatic operation mode and the operation feature amount in the manual operation mode for each grid is provided.
The operation model is determined based on the operation feature amount of the reference number or more accumulated in the operation feature amount storage unit.
With respect to the grid associated with the driving feature amount acquired by the driving feature amount acquisition unit, the driving feature amount in the automatic operation mode stored in the driving feature amount storage unit does not meet the reference number. In the case, the operation model acquisition unit is the operation model in the manual operation mode stored in the operation feature amount storage unit for the grid associated with the operation feature amount acquired by the operation feature amount acquisition unit. Is acquired as the operation model in the automatic operation mode . An abnormality estimation device (200, 1100) that estimates abnormalities in road conditions.
A driving feature amount acquisition unit (201, 1102) that acquires a driving feature amount, which is a feature amount of a vehicle's driving behavior, in association with a grid that divides a region in which the vehicle travels.
With respect to the grid associated with the driving feature amount acquired by the driving feature amount acquisition unit, a driving model acquisition unit (206) that acquires a driving model representing the driving behavior in a normal state, and a driving model acquisition unit (206).
An abnormality degree calculation unit (anomaly degree calculation unit) that compares the operation feature amount acquired by the operation feature amount acquisition unit with the operation model acquired by the operation model acquisition unit and calculates the degree of abnormality of the operation feature amount with respect to the operation model. 207) and
It is provided with an output unit (209, 1103) that outputs information for estimating road abnormalities based on the degree of abnormality.
The driving feature amount acquisition unit acquires, together with the driving feature amount, whether the driving mode of the vehicle is the automatic driving mode or the manual driving mode.
When the driving feature amount is the driving feature amount in the automatic driving mode, the driving model acquisition unit acquires the driving model in the automatic driving mode, and the driving feature amount is the manual driving mode. If it is the operation feature amount of the above, the operation model in the manual operation mode is acquired , and the operation model is acquired.
The operation feature amount accumulating unit (203, 204) which stores the operation feature amount in the automatic operation mode and the operation feature amount in the manual operation mode for each grid is provided.
The operation model is determined based on the operation feature amount of the reference number or more accumulated in the operation feature amount storage unit.
When the grid for determining the operation model, which is determined based on the grid associated with the operation feature amount acquired by the operation feature amount acquisition unit, is used as a model extraction grid.
When the driving feature amount in the automatic driving mode associated with the model extraction grid stored in the driving feature amount storage unit does not meet the reference number, the driving model acquisition unit extracts the model. The driving feature amount in the automatic driving mode associated with the grid, including other roads similar in shape to the road on which the vehicle traveled in the grid, associated with the model extraction grid. An abnormality estimation device that determines the operation model by supplementing the operation feature amount in the automatic operation mode. An abnormality estimation device (200, 1100) that estimates abnormalities in road conditions.
A driving feature amount acquisition unit (201, 1102) that acquires a driving feature amount, which is a feature amount of a vehicle's driving behavior, in association with a grid that divides a region in which the vehicle travels.
With respect to the grid associated with the driving feature amount acquired by the driving feature amount acquisition unit, a driving model acquisition unit (206) that acquires a driving model representing the driving behavior in a normal state, and a driving model acquisition unit (206).
An abnormality degree calculation unit (anomaly degree calculation unit) that compares the operation feature amount acquired by the operation feature amount acquisition unit with the operation model acquired by the operation model acquisition unit and calculates the degree of abnormality of the operation feature amount with respect to the operation model. 207) and
It is provided with an output unit (209, 1103) that outputs information for estimating road abnormalities based on the degree of abnormality.
The driving feature amount acquisition unit acquires, together with the driving feature amount, whether the driving mode of the vehicle is the automatic driving mode or the manual driving mode.
When the driving feature amount is the driving feature amount in the automatic driving mode, the driving model acquisition unit acquires the driving model in the automatic driving mode, and the driving feature amount is the manual driving mode. If it is the operation feature amount of the above, the operation model in the manual operation mode is acquired , and the operation model is acquired.
Pattern 1 is when the abnormality degree calculation unit calculates the abnormality degree by comparing the operation feature amount in the manual operation mode with the operation model, and the operation feature amount and the operation in the automatic operation mode. Pattern 2 is the case where the abnormality degree is calculated by comparing with the model, and the abnormality degree is calculated by comparing the operation feature amount and the operation model with respect to the automatic operation mode and the manual operation mode, respectively. When pattern 3 is used,
The output unit is an abnormality estimation device that outputs information based on the degree of abnormality so that the pattern 1, the pattern 2, and the pattern 3 can be distinguished. The operation feature amount accumulating unit (203, 204) which stores the operation feature amount in the automatic operation mode and the operation feature amount in the manual operation mode for each grid is provided.
The operation model acquisition unit acquires the operation feature amount of the grid corresponding to the operation feature amount acquired by the operation feature amount acquisition unit from the operation feature amount storage unit, and generates the operation model. When the operation feature amount is the operation feature amount in the automatic operation mode, the operation feature amount in the automatic operation mode is acquired from the operation feature amount storage unit, and the operation feature amount is the manual operation. The abnormality estimation device according to claim 3 , wherein when the operation feature amount is in the mode, the operation feature amount in the manual operation mode is acquired from the operation feature amount storage unit. The operation feature amount accumulating unit (203, 204) which stores the operation feature amount in the automatic operation mode and the operation feature amount in the manual operation mode for each grid is provided.
The operation model is determined based on the operation feature amount of the reference number or more accumulated in the operation feature amount storage unit.
With respect to the grid associated with the driving feature amount acquired by the driving feature amount acquisition unit, the driving feature amount in the automatic driving mode stored in the driving feature amount storage unit does not meet the reference number. In the case, the operation model acquisition unit is the operation model in the manual operation mode stored in the operation feature amount storage unit for the grid associated with the operation feature amount acquired by the operation feature amount acquisition unit. The abnormality estimation device according to claim 3 or 4 , wherein the above is acquired as the operation model in the automatic operation mode. The operation feature amount accumulating unit (203, 204) which stores the operation feature amount in the automatic operation mode and the operation feature amount in the manual operation mode for each grid is provided.
The operation model is determined based on the operation feature amount of the reference number or more accumulated in the operation feature amount storage unit.
When the grid for determining the operation model, which is determined based on the grid associated with the operation feature amount acquired by the operation feature amount acquisition unit, is used as a model extraction grid.
When the driving feature amount in the automatic driving mode associated with the model extraction grid stored in the driving feature amount storage unit does not meet the reference number, the driving model acquisition unit extracts the model. The driving feature amount in the automatic driving mode associated with the grid, including other roads similar in shape to the road on which the vehicle traveled in the grid, associated with the model extraction grid. The abnormality estimation device according to any one of claims 3 to 5 , wherein the operation model is determined by supplementing the operation feature amount in the automatic operation mode. The abnormality estimation device according to any one of claims 3 to 6 , wherein the information based on the degree of abnormality is information for notifying the occupant of the vehicle of an abnormality occurring on the road. The abnormality estimation device is located outside the vehicle.
The driving feature amount, the vehicle position for associating the driving feature amount with the grid, and the driving mode of the vehicle are transmitted from the vehicle.
The abnormality estimation device according to any one of claims 1 to 7, wherein the driving feature amount acquisition unit (201) acquires the driving feature amount, the vehicle position, and the driving mode transmitted from the vehicle. The abnormality estimation device is mounted on the vehicle and is mounted on the vehicle.
A driving data acquisition unit (101) for acquiring driving data indicating the driving behavior of the vehicle is provided.
The abnormality estimation device according to any one of claims 1 to 7, wherein the operation feature amount acquisition unit (1102) generates the operation feature amount based on the operation data acquired by the operation data acquisition unit.

Images