WO2018029904A1

WO2018029904A1 - Driving tendency determination apparatus and driving tendency determination system

Info

Publication number: WO2018029904A1
Application number: PCT/JP2017/014736
Authority: WO
Inventors: 板原　弘
Original assignee: パナソニックＩｐマネジメント株式会社
Priority date: 2016-08-10
Filing date: 2017-04-11
Publication date: 2018-02-15
Also published as: JP2018024340A

Abstract

This driving tendency determination apparatus is provided with: an acquisition unit that acquires information indicating the jerk and the acceleration of a vehicle measured while the vehicle is driven by a determination subject; and an arithmetic unit that is provided with an artificial intelligence function realized through learning by using a determination image including a two-dimensional plane in which the time sequence of acceleration and jerk measured while a vehicle is driven by an arbitrarily defined driver is plotted. The arithmetic unit generates a plurality of determination images on the basis of the information indicating the acceleration and the jerk measured while the vehicle is driven by the determination subject, and determines the driving tendency of the determination subject by inputting the plurality of determination images to the artificial intelligence function.

Description

Driving tendency determination device and driving tendency determination system

The present disclosure relates to a driving tendency determination device and a driving tendency determination system that determine a driving tendency of a vehicle driver.

Patent Document 1 discloses a vehicle driving support device that determines a driving operation state of a vehicle based on an amount of change in acceleration or the like.

The vehicle driving support device of Patent Document 1 includes a change amount calculation unit that calculates a first related value related to a change amount of acceleration of a vehicle, and a jerk calculation that calculates a second related value related to the jerk of the vehicle. Means, acceleration calculating means for calculating a third related value related to the absolute value of the acceleration of the vehicle, and state determining means for determining the driving operation state of the vehicle based on the first to third related values.

When the first related value calculated by the change amount calculating means is equal to or greater than a predetermined value, the state determining means uses the first determination map and based on the first related value and the second related value, The driving operation state of the vehicle is determined according to a criterion set in advance based on the ratio of the kinetic energy of the mass point at the end of the change to the amount of change in the acceleration of the vehicle, calculated using a vibration model representing the movement of the mass point. judge. Further, when the first related value is smaller than the predetermined value, the state determining means uses a second determination map different from the first determination map, and based on the first related value and the third related value, The driving operation state is determined. With this configuration, it is possible to accurately determine the driving operation state of the vehicle regardless of the amount of change in acceleration.

JP 2012-198345 A

In a first aspect of the present disclosure, a driving tendency determination device is provided. The driving tendency determination device includes an acquisition unit that acquires information indicating acceleration and jerk of a vehicle measured when the determination target person drives the vehicle, and a time series of acceleration and jerk measured when the vehicle is driven by an arbitrary driver. And an arithmetic unit implemented with an artificial intelligence function learned using a determination image including a two-dimensional plane on which is plotted. The calculation unit generates a plurality of determination images based on the acceleration and jerk information measured during driving of the determination target person, and inputs the plurality of determination images to the artificial intelligence function to increase the driving tendency of the determination target person. judge.

In a second aspect of the present disclosure, a driving tendency determination system is provided. The driving tendency determination system includes a measuring device that measures vehicle acceleration, a portable terminal that receives information indicating acceleration measured from the measuring apparatus and transfers the information to the driving tendency determination device, and receives information indicating acceleration from the portable terminal. And the driving tendency determination device that learns the artificial intelligence function based on the received information.

According to the driving tendency detection device and the driving tendency system of the present disclosure, it is possible to determine the driving tendency of the driver of the vehicle based on information on acceleration and jerk measured during driving of the vehicle.

FIG. 1 is a diagram illustrating a configuration of a driving tendency determination system according to Embodiment 1 of the present disclosure. FIG. 2 is a block diagram showing the configuration of the telemeter unit in the first embodiment. FIG. 3 is a block diagram showing a configuration of the portable terminal in the first embodiment. FIG. 4 is a block diagram showing the configuration of the data server (driving tendency determination device) in the first embodiment. FIG. 5 is a diagram for explaining the flow of driving tendency data in the first embodiment. FIG. 6 is a diagram illustrating a change in vehicle speed when a sudden brake is applied during vehicle operation while descending a gradient. FIG. 7 is a diagram for explaining a change in acceleration when a sudden brake is applied during driving of the vehicle while descending a gradient. FIG. 8A is a view for explaining a J (jerk acceleration) -G (acceleration) plane. FIG. 8B is a diagram for explaining an area determined as dangerous driving on the J (jerk acceleration) -G (acceleration) plane. FIG. 9 is a diagram showing a convolutional neural network in the AI (artificial intelligence) function. FIG. 10 is a diagram illustrating a method of generating an image (JG plane image) used for machine learning. FIG. 11 is a diagram for explaining the learning procedure of the convolutional neural network. FIG. 12 is a flowchart showing a process of transmitting driving tendency data from the telemeter unit to the portable terminal. FIG. 13 is a flowchart showing driving tendency determination processing in the data server. FIG. 14 is a diagram illustrating a configuration of the driving tendency determination system according to the second embodiment of the present disclosure. FIG. 15 is a diagram for explaining the flow of driving tendency data in the second embodiment. FIG. 16 is a diagram illustrating a configuration of the driving tendency determination system according to the third embodiment of the present disclosure. FIG. 17 is a diagram for explaining the flow of driving tendency data in the third embodiment.

Hereinafter, embodiments will be described in detail with reference to the drawings as appropriate. However, more detailed description than necessary may be omitted. For example, detailed descriptions of already well-known matters and repeated descriptions for substantially the same configuration may be omitted. This is to avoid the following description from becoming unnecessarily redundant and to facilitate understanding by those skilled in the art.

In addition, the inventor provides the accompanying drawings and the following description in order for those skilled in the art to fully understand the present disclosure, and is not intended to limit the claimed subject matter. .

(Embodiment 1)
[1-1. Constitution]
[1-1-1. System configuration]
Embodiments of a driving tendency determination system according to the present invention will be described below with reference to the accompanying drawings. The driving tendency determination system described in the following embodiment is a system that detects the driving tendency of a vehicle driver. FIG. 1 illustrates a configuration of a driving tendency determination system according to an embodiment of the present disclosure.

As shown in FIG. 1, the driving tendency determination system 100 includes a telemeter unit 10, a mobile terminal 40, and a data server 50. The telemeter unit 10 is arranged in the vehicle 30, detects the behavior of the vehicle 30 when the vehicle 30 is driven, and data indicating the driving tendency (or driving operation state) of the driver of the vehicle based on the behavior (hereinafter, "" Driving tendency data ”) and send it to the outside. The portable terminal 40 receives the driving tendency data from the telemeter unit 10 and transmits it to the data server 50 via the network 200. The data server 50 updates the database based on the received driving tendency data.

Hereinafter, specific configurations of the elements of the telemeter unit 10, the mobile terminal 40, and the data server 50 will be described.

[1-1-2. Telemeter unit]
FIG. 2 is a diagram illustrating the configuration of the telemeter unit 10. The telemeter unit 10 includes a controller 11, an acceleration sensor 15, a communication interface 18 that enables wireless communication with other electronic devices in accordance with communication standards such as WiFi and Bluetooth (registered trademark), and a memory that stores data and the like. 17.

The controller 11 is composed of a CPU (Central Processing Unit) and an MPU (Micro Processing Unit), and implements predetermined functions to be described later by executing a program stored in the memory 17. The program executed by the controller 11 may be provided via the network 200 or may be provided by a recording medium such as a CD-ROM.

The acceleration sensor 15 is a sensor that detects the acceleration of the vehicle 30 in three orthogonal directions (X, Y, and Z directions). Here, the width direction of the vehicle 30 is defined as the X direction, the traveling direction (forward direction) of the vehicle 30 is defined as the Y direction, and the upward direction (zenith direction) of the vehicle 30 is defined as the Z direction.

The memory 17 is a recording medium for storing various data, and is composed of, for example, a semiconductor storage element such as a flash memory. The memory 17 stores programs and data executed by the controller 11. Instead of the memory 17, a recording medium such as a removable memory card or a hard disk may be used.

The communication interface 18 is a module that performs wireless communication in accordance with a communication standard such as WiFi or Bluetooth (registered trademark). The communication interface 18 may perform communication according to a communication standard such as LTE (Long Term Evolution) or 3G. Note that the communication interface 18 is not limited to wireless communication, and may be wired communication.

[1-1-3. Mobile device]
The portable terminal 40 can communicate with the telemeter unit 10. In addition, the mobile terminal 40 can transmit information to the data server 50 via the network 200. In the present embodiment, a smartphone is assumed as an example of the mobile terminal 40, but the mobile terminal 40 may be a PDA (Personal Digital Assistant), a mobile phone, or the like.

FIG. 3 is a diagram illustrating the configuration of the mobile terminal 40. The portable terminal 40 includes an imaging unit 42 that captures an image, a display unit 43 that displays information such as an image, and an operation unit 45. Furthermore, the portable terminal 40 includes a first communication interface 48 that communicates to connect to a network, and a second communication interface 49 that communicates with other electronic devices. Furthermore, the mobile terminal 40 includes a RAM (Random Access Memory) 46 and a data storage unit 47 that store data and the like, and a controller 41 that controls the overall operation of the mobile terminal 40. The portable terminal 40 may include an acceleration sensor as will be described later.

The image pickup unit 42 includes an image sensor such as a CCD (Charge Coupled Device) or a CMOS (Complementary Metal Oxide Semiconductor) image sensor, and picks up a subject to generate image data.

The display unit 43 is composed of, for example, a liquid crystal display or an organic EL display. The operation unit 45 includes a touch panel that is arranged on the upper surface of the display unit 43 and receives a touch operation by a user. The operation unit 45 further includes operation buttons. The operation buttons include buttons physically provided on the main body of the mobile terminal 40 and virtual buttons realized by the display unit 43 and the touch panel function.

The first communication interface 48 is a communication module for connecting to a network, and performs communication according to a communication standard such as LTE or 3G. The second communication interface 49 is a communication module for wirelessly communicating with other electronic devices at a relatively short distance, and performs communication according to a communication standard such as WiFi or Bluetooth (registered trademark). The second communication interface 49 may communicate with other electronic devices via a cable. For example, the second communication interface 49 may perform data communication in accordance with a standard such as USB (Universal Serial Bus).

The RAM 46 is a storage element that temporarily stores programs and data, and functions as a work area for the controller 41. The data storage unit 47 is a recording medium that stores programs and various data, and may be configured by a recording medium such as a hard disk, a semiconductor memory, or a removable memory card. The data storage unit 47 stores programs executed by the controller 41 (OS: Operating System, application program 47a) and data.

The controller 41 includes a CPU and an MPU, and implements a predetermined function to be described later by executing an application program 47a stored in the data storage unit 47. The application program 47a may be provided via the network 200 or may be provided by a recording medium such as a CD-ROM.

[1-1-4. Data server]
FIG. 4 is a diagram illustrating the configuration of the data server 50. The data server 50 is constituted by an information processing apparatus such as a personal computer. The data server 50 includes a controller 51 that controls the overall operation thereof, a display unit 53 that performs screen display, an operation unit 55 that is operated by a user, a RAM 56 that stores data and programs, and a data storage unit 57. The display unit 53 is configured by, for example, a liquid crystal display or an organic EL display. The operation unit 55 includes a keyboard, a mouse, a touch panel, and the like.

The data server 50 further includes a device interface 58 for connecting to an external device such as a printer, and a network interface 59 for connecting to a network. The device interface 58 is a communication module that performs data communication in accordance with USB, HDMI (registered trademark) (High-Definition Multimedia Interface), IEEE 1394, or the like. The network interface 59 is a communication module that performs data communication in accordance with standards such as IEEE 802.11 and WiFi.

The controller 51 is composed of a CPU and an MPU, and realizes a predetermined function by executing a predetermined control program 57a stored in the data storage unit 57. The control program executed by the controller 51 may be provided via the network 200 or may be provided on a recording medium such as a CD-ROM.

The RAM 56 is a storage element that temporarily stores programs and data, and functions as a work area for the controller 51. The data storage unit 57 is a recording medium that stores parameters, data, and programs necessary for realizing the functions, and stores a control program executed by the controller 51 and various data. The data storage unit 57 includes, for example, a hard disk (HDD: Hard Disk Drive) or a semiconductor storage device (SSD: Solid State Drive). The data storage unit 57 is installed with a control program 57a for realizing functions described later. The controller 51 implements various functions to be described later by executing the control program 57a. The data storage unit 57 also stores a database 57b regarding the driving tendency of the driver. The data storage unit 57 also functions as a work area for the controller 51.

The database 57b manages data indicating driving tendency for each driver (hereinafter referred to as “driving tendency data”). The driving tendency data includes a driver ID for identifying the driver, a vehicle ID for identifying the vehicle, a data sampling date, information indicating acceleration, and information indicating jerk.

[1-2. Operation]
The operation of the driving tendency determination system 100 configured as described above will be described below.

FIG. 5 is a diagram for explaining the flow of driving tendency data in the driving tendency determination system 100. In the driving tendency determination system, the telemeter unit 10 calculates driving tendency data indicating the driving tendency of the driver based on the acceleration data of the vehicle 30 (S1), and transmits the driving tendency data to the portable terminal 40 via the communication interface 18 (S2). ). The portable terminal 40 receives the driving tendency data from the telemeter unit 10 via the second communication interface 49 (S2), and transmits the received driving tendency data to the data server 50 via the first communication interface 48 (S3). . The data server 50 receives the driving tendency data from the portable terminal 40 (S3), and updates the database 57b based on the received driving tendency data (S4). As described above, the driving tendency data detected by the telemeter unit 10 is transmitted to the data server 50, and the database 57b is updated with the driving tendency data. By referring to the database 57b, the driving tendency of the driver of the vehicle can be recognized.

[1-2-1. Driving tendency data]
Driving tendency data used for determination of driving tendency will be described.

Generally, it is known that there is a correlation between acceleration during driving and driving tendency. For example, when driving in an ordinary city, the maximum acceleration in a car is about 0.15 G, and if it exceeds 0.3 G, it can be determined that the driving is rough.

FIG. 6 is a diagram for explaining changes in vehicle speed measured when sudden braking is applied during vehicle operation while going down a gentle slope. In this example, an accelerometer is used to drive a minivan (two passengers) down a gentle slope, and then the vehicle is decelerated from a speed of approximately 50 km / h and suddenly braked when the speed reaches 20 km / h. The change in the acceleration of the vehicle when the vehicle was decelerated to 0 km / h was measured. The measurement was performed for about 11 seconds at a sampling interval of 20 ms.

In FIG. 6, “safe driving” is performed until the speed of 20 km / h is reached and sudden braking is applied, and “dangerous driving” is performed until the braking is stopped after sudden braking is started.

FIG. 7 is a diagram for explaining changes in acceleration measured by the acceleration sensor with respect to the behavior of the vehicle as shown in FIG. Plot X shows acceleration in the X direction (vehicle width direction), plot Y shows acceleration in the Y direction (vehicle traveling (front-rear) direction), and plot Z shows acceleration in the Z direction (upward direction of the vehicle). ing. As shown in FIG. 7, when sudden braking is started at time t1, the acceleration in the Y direction (traveling direction) changes greatly. At this time, since the vehicle is gently descending the slope, an offset of 0.3 always occurs in the acceleration in the Y direction. The magnitude of this offset varies depending on the slope angle of the slope and the weight of the vehicle. As described above, since an offset occurs in the acceleration, it is difficult to accurately determine whether the state is the “safe driving state” or the “dangerous driving state” only by the acceleration.

Therefore, as a result of earnest research, the present inventor has determined the driving tendency in consideration of the jerk (also referred to as “jump”) in addition to the acceleration, and accurately determines the driving tendency (that is, the driving state). It was found that it could be detected. The jerk is an amount indicating the rate of change of acceleration, and is obtained by differentiation of acceleration.

FIG. 8A shows a change in acceleration and jerk of a sampling point measured in a Y-axis direction (traveling direction) for a predetermined period (for example, 11 seconds) with respect to the behavior of the vehicle shown in FIG. (Hereinafter referred to as “JG plane”). In the GG plane shown in FIG. 8A, the horizontal axis represents jerk (J), and the vertical axis represents acceleration (G).

The present inventor has found that the plot of sampling points draws a unique trajectory due to a sudden change in the behavior of the vehicle on such a GG plane. Specifically, in a situation where safe driving is performed as shown in FIGS. 6 and 7, jerk (J) and acceleration (G) are relatively small on the GG plane as shown in FIG. 8B. The plots are concentrated in the range P (the region indicated by the alternate long and short dash line). Thereafter, when sudden braking is performed and a dangerous driving situation is reached, the locus Q of the plot is drawn counterclockwise far beyond the range P on the GG plane.

The image in which the sampling points are plotted on the GG plane in this way includes information corresponding to the driving tendency of the driver, such as the locus Q that appears corresponding to the sudden braking. The inventor pays attention to this point, and performs machine learning on an image on the JG plane (hereinafter referred to as “JG plane image”) using a convolutional neural network. We thought that driving tendency could be classified. Therefore, the controller 51 of the data server 50 has an AI (Artificial Intelligence) function. Hereinafter, the AI function of the controller 51 will be described.

[1-2-2. Convolutional neural network]
As shown in FIG. 9, the controller 51 of the data server 50 according to the present embodiment uses a convolutional neural network (CNN) model in the AI function. The convolutional neural network is made to learn an image (JG plane image) generated from long-term driving tendency data of the driver. Data of an image (JG plane image) used for machine learning for a convolutional neural network is obtained as follows.

FIG. 10 is a diagram illustrating a method for generating an image (JG plane image) used for machine learning. It is assumed that long-term (for example, 30 hours) acceleration and jerk data (driving tendency data) 300 for a driver is measured by the telemeter unit 10. From this long-term driving tendency data 300, the acceleration and jerk data measured during a predetermined period (for example, 11 seconds) are plotted on the JG plane to obtain a JG plane image 310. create. As shown in FIG. 8B, the predetermined period is set so that information on the behavior of the vehicle from sudden braking to vehicle stop is included in one JG plane (for example, 11 seconds). At this time, a plurality of JG plane images 310 are generated from the driving tendency data 300 measured over a long period (for example, 30 hours) (see FIG. 10). That is, for a single driver, a plurality of JG plane images (JG plane image group) are generated from the driving tendency data measured over a long period of time. Then, the JG plane image group for a large number of drivers is machine-learned by a convolutional neural network.

FIG. 11 is a diagram for explaining the learning procedure of the convolutional neural network. Convolutional neural network learning is performed by executing three steps: (1) unsupervised deep learning (classification), (2) supervised deep learning (weighting clustered groups), and (3) inference. . Each step will be described.

(1) Unsupervised deep learning (classification)
A JG plane image group for a large number of drivers is input to a convolutional neural network. The convolutional neural network performs clustering and automatically constructs a feature group (clustering) from the JG plane image group. For example, as shown in FIG. 9, clusters (feature groups) C1, C2, and C3 are constructed.

(2) Supervised deep learning (weighting clustered groups)
The data of the driver who actually caused the accident is input to the convolutional neural network as a teacher signal. The convolutional neural network classifies highly responsive clusters as “accident reserve groups”, which are groups of drivers with a high probability of accident occurrence. That is, by giving history information (accident occurrence date and time) that an arbitrary driver has caused an accident as a teacher signal, the cluster group having the driving tendency is weighted with the accident preliminary group. For example, the cluster C1 is classified as an “accident reserve group” that is a group of drivers who have a high probability of accident occurrence. The cluster C2 is classified as a “slightly dangerous driving group” that is a group of drivers having a slightly higher probability of accident occurrence. Cluster C3 is classified as a “safe driving group” which is a group of drivers who perform safe driving.

(3) Inference After learning in (1) and (2) above, an arbitrary JG plane image group is input to the convolutional neural network, and a determination result based on the JG plane image group is inferred. The correctness of the inference result is fed back as a teacher signal. In this way, the inference accuracy of the convolutional neural network is increased by repeating the inference and correction by the teacher signal while inferring the convolutional neural network by the JG plane image group. For example, for the cluster C1 classified as the accident reserve group, when the inference result is incorrect, the weight of the sub-cluster C1-1 in the cluster C1 is reduced, and when the inference result is correct, the sub-cluster in the cluster This is done by increasing the weight of the cluster (for example, C1-2). For example, if the subject was classified into an “accident reserve group” as an inference result for a subject, but the subject actually did not have an accident, the teacher informed that the reasoning was incorrect. Feedback with a signal. As a result, the clusters C1-1 and C1-2 are updated.

In this way, image data is repeatedly input and deep learning is performed while teaching correct inferences and errors as appropriate while performing inference, and cluster groups are subdivided by Similarity Matching to extract the preliminary accident group with higher accuracy. Converge as you can. When the correct answer rate of the inference result is equal to or higher than a predetermined probability, the inference step is ended and the learning of the convolutional neural network is ended.

As described above, a convolutional neural network is learned using a sufficient amount of JG plane images.

[1-2-2. Data transmission processing from telemeter unit]
FIG. 12 is a flowchart showing a driving tendency data transmission process from the telemeter unit 10 to the portable terminal 40. Hereinafter, with reference to FIG. 12, the driving tendency data transmission processing in the telemeter unit 10 will be described. The process shown in FIG. 12 is mainly executed by the controller 11 of the telemeter unit 10.

During driving of the vehicle, in the telemeter unit 10, the acceleration sensor 15 detects (measures) the acceleration (G) of the vehicle 30 at a predetermined sampling interval (for example, 20 msec), and information indicating the detected (measured) acceleration is stored in the memory 17. Accumulated.

When the process shown in FIG. 12 is started, the controller 11 of the telemeter unit 10 reads and acquires information on the acceleration (G) measured within a predetermined period (for example, one month) from the memory 17 (S11).

Controller 11 calculates jerk (J) at the sampling point from each acceleration (G) in the read Y direction (S12). Driving tendency data is generated from the acceleration and jerk information of each sampling point.

Thereafter, the controller 11 transmits the driving tendency data to the portable terminal 40 via the communication interface 18 (S13). At this time, the driving tendency data obtained based on the acceleration data stored in the memory 17 of the telemeter unit 10, that is, the driving tendency data measured in a certain period is transmitted to the portable terminal 40.

The portable terminal 40 receives the driving tendency data from the telemeter unit 10 via the second communication interface 49, and transmits the received driving tendency data to the data server 50 connected to the network 200 via the first communication interface 48. To do. The data server 50 receives the driving tendency data via the network interface 59 and accumulates the received driving tendency data in the database 57b (see FIGS. 4 and 5).

[1-2-3. Determination of driving tendency]
The driving tendency determination process in the data server 50 will be described with reference to the flowchart of FIG.

The user inputs information (ID number, name, etc.) specifying the target person who wants to determine the driving tendency through the operation unit 55 of the data server 50. When the controller 51 of the data server 50 receives the information specifying the target person, the controller 51 accesses the database 57b to acquire long-term driving tendency data (for example, for 30 hours) for the target person, and the acquired driving tendency data Based on the above, a JG plane image group is created (S21).

The controller 51 inputs the JG plane image group generated for the subject to the convolutional neural network, and determines the driving tendency for the subject (S22). The convolutional neural network outputs the result of classification based on the JG plane image of the subject as a result of determining the driving tendency of the subject for each JG plane image. The controller 51 outputs information indicating the driving tendency determination result obtained from the convolutional neural network to the display unit 53 (S23).

For example, when a plurality of JG images are input to the convolutional neural network to determine driving tendency, each JG image is classified into one of the clusters (for example, C1 to C3) and classified. The information indicating the cluster is output. In such a case, the ratio of each output cluster to the total number of input JG images may be obtained and displayed as a determination result on the display unit. Specifically, the results output for each image of the JG image group input to the convolutional neural network are counted for each cluster. Then, the ratio (%) of the number of results classified into cluster C3 to the total number of JG images input for determination is obtained as “safe driving rate” (a rate indicating that the probability of causing an accident is extremely low). . Further, the ratio (%) of the number of results classified into the cluster C2 with respect to the total number of JG image groups is defined as “slightly dangerous driving rate” (a rate indicating that the probability of causing an accident is somewhat high). Further, the ratio (%) of the number of results classified into the cluster C1 with respect to the total number of JG image groups is defined as “dangerous driving rate” (a rate indicating that the probability of causing an accident is very high). Then, the ratio to each item may be displayed on the display unit 53 as the determination result of the driving tendency. For example, “safe driving rate: 80%, somewhat dangerous driving rate: 15%, dangerous driving rate: 5%” may be displayed.

Note that the determination result of the driving tendency may be stored in the data storage unit 57 or may be transmitted to another device via the network interface 59 or the device interface 58.

As described above, the data server 50 of the present embodiment can determine the driving tendency of the subject from the subject's JG plane image group.

[1-3. Effect]
As described above, the data server 50 (an example of the driving tendency determination device) according to the present embodiment acquires an information (network interface) that acquires information indicating the acceleration and jerk of the vehicle measured when the determination target person drives the vehicle. 59, a data storage unit 57, etc.) and an artificial intelligence function learned using a determination image including a two-dimensional plane in which a time series of acceleration and jerk measured when a vehicle is driven by an arbitrary driver is plotted A controller 51 (calculation unit). The controller 51 generates a plurality of JG plane images (an example of a determination image) based on acceleration and jerk information measured when the determination target person drives the vehicle, and uses the plurality of JG plane images as an artificial intelligence function. To determine the driving tendency of the person to be determined.

The driving tendency determination system 100 includes a telemeter unit 10 (an example of a measurement device) that measures vehicle acceleration, a mobile terminal 40 that receives and transfers information indicating acceleration measured from the telemeter unit 10, and a mobile terminal 40. And a data server 50 that receives information indicating acceleration from the computer and learns the artificial intelligence function based on the received information.

As described above, the data server 50 according to the present embodiment is a convolutional neural that has previously learned a JG plane image in which the acceleration and jerk of the vehicle measured for a determination target person are plotted on a two-dimensional plane. By inputting to the network, the determination result of the driving tendency of the determination target person can be obtained. In particular, in this embodiment, a plurality of JG plane images, that is, driving tendency data measured over a long period of time is used as information to be input when learning the artificial intelligence function. For this reason, since the tendency of a driver | operator's driving | operation is included more reliably in those data, a highly accurate determination is attained.

Various methods and knowledge related to deep learning of image big data, API (Application Program Interface), etc. can be applied in driving-oriented deep learning and accident probability inference, etc., without requiring special algorithm development, The driver's driving tendency that is likely to cause an accident can be automatically classified into clusters, and the association between the cluster and the accident can be automated. In addition, it can be implemented on a common platform, including the final goal of inference (accident prediction).

(Embodiment 2)
FIG. 14 is a diagram illustrating a configuration of the driving tendency determination system 100b according to the second embodiment of the present disclosure. In the first embodiment, the driving tendency determination system 100 is configured by the telemeter unit 10, the mobile terminal 40, and the data server 50. On the other hand, the driving tendency determination system 100b of the present embodiment includes a telemeter unit 10 mounted on a vehicle and a data server 50. That is, the telemeter unit 10 transmits data directly to the data server 50 without using the mobile terminal 40. The telemeter unit 10 can be connected to the network 200 via the communication interface 18 in order to communicate with the data server 50.

FIG. 15 is a diagram for explaining the flow of driving tendency data in the second embodiment. As shown in the figure, the telemeter unit 10 calculates driving tendency data (S41), and transmits the calculated driving tendency data to the data server 50 (S42). The data server 50 updates the database 57b based on the received data (S43).

The telemeter unit 10 may transmit driving tendency data to the data server 50 every predetermined period (for example, one month). Alternatively, the telemeter unit 10 may transmit data when there is an instruction from the user or a predetermined operation of the vehicle (engine start or the like).

(Embodiment 3)
In the present embodiment, a configuration in which the mobile terminal 40 includes the above functions of the telemeter unit 10 will be described. That is, in Embodiment 1, the telemeter unit 10 detects the acceleration of the vehicle 30 and calculates the driving tendency data, but the mobile terminal 40 may execute these functions of the vehicle 30.

FIG. 16 is a diagram illustrating a configuration of the driving tendency determination system 100c according to the third embodiment of the present disclosure. The driving tendency determination system 100 c according to the present embodiment includes a mobile terminal 40 and a data server 50. The mobile terminal 40 of this embodiment includes an acceleration sensor that can detect acceleration in three orthogonal directions (XYZ directions). The portable terminal 40 is disposed in the vehicle 30 such that the traveling direction (forward direction) of the vehicle 30 is the Y direction and the upward direction of the vehicle is the Z direction. The portable terminal 40 periodically measures and records the acceleration of the vehicle 30 during the driving of the vehicle 30 in the state of being arranged in the vehicle 30 as described above.

FIG. 17 is a diagram for explaining the flow of driving tendency data in the third embodiment. As shown in the figure, the mobile terminal 40 calculates driving tendency data from the measured acceleration data according to the above-described method (S51), and transmits the calculated driving tendency data to the data server 50 (S52). The data server 50 updates the database 57b based on the received driving tendency data (S53).

The portable terminal 40 may transmit driving tendency data to the data server 50 every predetermined period (for example, one month), or transmit data when a predetermined operation (such as a transmission instruction) is performed by the user. You may do it.

(Other embodiments)
As described above, Embodiments 1 to 3 have been described as examples of the technology disclosed in the present application. However, the technology in the present disclosure is not limited to this, and can also be applied to an embodiment in which changes, replacements, additions, omissions, and the like are appropriately performed. Also, it is possible to combine the components described in the first to third embodiments to form a new embodiment. Therefore, other embodiments will be exemplified below.

In Embodiments 1 to 3, the telemeter unit 10 has been described as an example of a driving tendency detection device. The driving tendency detection device is not limited to the telemeter unit 10. The portable terminal 40 and the data server 50 can also be configured as a driving tendency detection device.

For example, in the first embodiment, as shown in FIG. 5, the driving tendency data is generated in the telemeter unit 10, but may be generated in a device other than the telemeter unit 10, that is, the portable terminal 40 or the data server 50.

More specifically, when the driving tendency data is calculated in the portable terminal 40, the telemeter unit 10 transmits a time series of detected acceleration data to the portable terminal 40. The portable terminal 40 receives the time series of acceleration data via the second communication interface 49 (an example of an acquisition unit) and stores it in the data storage unit 47 for work. The portable terminal 40 calculates jerk (J) based on the received acceleration data, and transmits data indicating the acceleration (G) and jerk (J) to the data server 50 as driving tendency data (FIG. 12). (See Steps S12 and S13). With this configuration, the mobile terminal 40 can operate as a driving tendency detection device.

Further, when the driving tendency data is calculated in the data server 50, the telemeter unit 10 transmits the acceleration data to the portable terminal 40. The portable terminal 40 transfers the received acceleration data to the data server 50. The data server 50 receives a time series of acceleration data via the network interface 59 (an example of an acquisition unit) and stores it in the data storage unit 57 for work. The data server 50 calculates jerk (J) based on the received acceleration data, generates driving tendency data from the data indicating the acceleration and jerk, and updates the database 57b with the generated driving tendency data. With such a configuration, the data server 50 can operate as a driving tendency detection device.

In the above embodiment, the

controller

11, 41, 51 includes a CPU or MPU, and an example in which a predetermined function (described later) is realized by executing a predetermined program (software) has been described. As described above, the functions of the

controllers

11, 41, and 51 may be realized by cooperation of hardware and software, or may be realized only by a hardware circuit. For example, the

controllers

11, 41, and 51 are a CPU, MPU, DSP (Digital Signal Processor), FPGA (Field-Programmable Gate Array), ASIC (Application Specific Integrated Circuit), ASP Can do.

As described above, the embodiments have been described as examples of the technology in the present disclosure. For this purpose, the accompanying drawings and detailed description are provided.

Accordingly, among the components described in the accompanying drawings and the detailed description, not only the components essential for solving the problem, but also the components not essential for solving the problem in order to illustrate the above technique. May also be included. Therefore, it should not be immediately recognized that these non-essential components are essential as those non-essential components are described in the accompanying drawings and detailed description.

In addition, since the above-described embodiment is for illustrating the technique in the present disclosure, various modifications, replacements, additions, omissions, and the like can be performed within the scope of the claims or an equivalent scope thereof.

The present disclosure can be applied to an apparatus that determines the driving tendency of a vehicle driver.

DESCRIPTION OF SYMBOLS 10 Telemeter unit 11 Telemeter unit controller 15 Acceleration sensor 17 Memory 18 Communication interface 30 Vehicle 40 Portable terminal 41 Portable terminal controller 47a Application program 48 First communication interface 49 Second communication interface 50 Data server 51 Data server controller 57a Control program 59 Network interface 100 Driving tendency determination system 200 Network

Claims

An acquisition unit that acquires information indicating the acceleration and jerk of the vehicle measured during driving of the determination target person;
An arithmetic unit equipped with an artificial intelligence function learned using a determination image including a two-dimensional plane in which a time series of acceleration and jerk measured when a vehicle is driven by an arbitrary driver is provided,
The computing unit is
A plurality of determination images are generated based on the information on the acceleration and jerk measured during the vehicle driving of the determination target person,
Determining the driving tendency of the person to be determined by inputting the plurality of determination images to the artificial intelligence function;
Driving tendency determination device.
The calculation unit uses the information indicating the acceleration of the vehicle measured within a first period and the information indicating the jerk corresponding to the information to be a second shorter than the first period. Generating a plurality of determination images by plotting a time series of a set of the acceleration and jerk of the vehicle measured in the second period on a two-dimensional plane for each period of
The driving tendency determination device according to claim 1.
The acceleration is measured every predetermined sampling period,
The driving tendency determination apparatus according to claim 2, wherein the second period is set to a period including a plurality of sets of plots of the acceleration and the jerk.
The driving tendency determination device according to claim 1, wherein the artificial intelligence function includes a convolutional neural network model.
A measuring device for measuring the acceleration of the vehicle;
A portable terminal that receives and transfers the information indicating the acceleration measured from the measuring device;
The driving tendency determination device according to any one of claims 1 to 4, wherein the information indicating the acceleration is received from the portable terminal, and the artificial intelligence function is learned based on the received information.
Driving tendency determination system.