JP7101287B1 - Inference device, inference system, inference method, and program - Google Patents

Abstract

An inference device, an inference system, an inference method, and a program for inferring the possibility that a specific user will cause at least one of an accident and a violation are provided. Kind Code: A1 In an inference system comprising a wearable device 40 attached to a user's advance and retreat and communicating with each other via a communication network, and an inference device 100, the inference device is a vehicle that a specific user (delivery company) uses as a vehicle. an acquisition unit that acquires activity amount data related to the amount of activity other than driving a mobile object during a time period in which a specific user does not drive a mobile object on a day when the mobile object (delivery truck) is driven; Activity amount data of a specific user acquired by an acquisition unit is used as new input data based on the learning result of the correspondence relationship between the user's activity amount data and the history of at least one of accidents and violations by any user. , a reasoning unit that infers the likelihood that a particular user will cause an accident and/or violation. [Selection drawing] Fig. 2

Description

The present invention relates to inference devices, inference systems, inference methods, and programs.

Patent Document 1 states, "When an abnormal value of physical information is detected, the position information at that time is associated with map data and road data to realize safe driving support for the driver himself / herself while driving." (Paragraph 0017).
[Prior Art Document]
[Patent Document]
[Patent Document 1] Japanese Unexamined Patent Publication No. 2018-195249

In the first aspect of the present invention, an inference device is provided. The inference device acquires activity data related to the amount of activity other than driving the moving body during the time when the specific user is not driving the moving body on the day when the specific user drives the moving body. It may be provided with a part. The inference device is based on the learning result of the correspondence between the activity data of any user and the history of at least one of accidents and violations by any user, and the activity amount of a specific user acquired by the acquisition unit. With the data as new input data, it may be provided with an inference unit that infers the possibility that a specific user causes at least one of an accident and a violation.

The amount of activity other than driving a moving body is the walking distance, the number of times of going up and down stairs, the standing time, the number of times the knees are bent and stretched, the number of times the upper body is bent forward, backward, laterally or twisted, and the luggage. It may include at least one of the number of raising and lowering and the total weight of the manually carried baggage.

The acquisition unit may acquire activity data measured by a sensor mounted on the body of a specific user.

The acquisition unit may estimate the activity amount data based on the position information of at least one of a specific user and a moving body.

The moving object and the specific user may be a delivery truck and a delivery company, respectively. Based on the location information, the acquisition unit identifies at least one of the areas where there is little space where the moving object can be parked and stopped, and the area where a specific user needs to carry luggage and go up and down the stairs on multiple floors. When it is determined that the user has been active other than driving a moving object, a relatively large amount of activity data may be estimated.

The inference device further provides additional data of at least one of diagnostic data related to the diagnostic result of the driving aptitude test of a specific user and attribute data related to the attribute of the specific user, as user identification information of the specific user. It may be provided with a storage unit for storing in association with. The acquisition unit may acquire activity data together with user identification information. The inference unit is based on the activity data of any user and the learning result of the correspondence between the additional data and the history, the activity data of a specific user acquired by the acquisition unit, and the specific activity data acquired by the acquisition unit. The possibility of a specific user may be inferred by using the additional data associated with the user identification information of the user and stored in the storage unit as new input data.

The acquisition unit is further at least one of blinking, gaze, facial expression, blood pressure, heart rate, arousal, and behavior detected by the specific user during the time when the specific user is driving the moving body. Biometric data related to may be acquired. The inference unit specifies the activity data and biometric data of a specific user acquired by the acquisition unit as new input data based on the learning result of the activity data of any user and the correspondence between the biometric data and the history. You may infer the possibility of the user of.

The acquisition unit may further acquire movement route information of at least one of a specific user and a moving body. The inference unit uses the activity data and travel route information of a specific user acquired by the acquisition unit as new input data based on the learning result of the correspondence between the activity data of any user and the travel route information and the history. , You may infer the possibility of a specific user.

The inference unit has a high inference accuracy, although the inference basis is blackboxed according to at least one of the new input data types and at least one of the correspondences and possibilities. Either a learning model or a second learning model in which the basis of inference is not blackboxed and can be understood by humans but the inference accuracy is low may be used.

The first learning model may be a GBDT (Gradient Boosting Decision Tree), and the second learning model may be a random forest.

The inference device may further include an output unit that outputs the possibility inferred by the inference unit as information that can be understood by humans. The output unit may output information together with the basis of the inference when the inference unit uses the second learning model.

In the second aspect of the present invention, an inference system is provided. The inference system may include any of the above inference devices. The inference system may be equipped with a wearable device that is worn on the body of a specific user, measures the activity data of the specific user, and transmits it to the inference device.

The inference system may further include a warning device that warns a specific user if the possibility of being inferred by the inference device satisfies a predetermined condition.

In the third aspect of the present invention, an inference method is provided. The inference method is to acquire activity data related to the amount of activity other than driving the moving body during the time when the specific user is not driving the moving body on the day when the specific user drives the moving body. It may have stages. The inference method is based on the learning result of the correspondence between the activity data of any user and the history of any user causing at least one of accidents and violations, and the activity amount of a specific user acquired at the acquisition stage. The data may be used as new input data and may include an inference step that infers the possibility that a particular user will cause at least one of accidents and violations.

A fourth aspect of the invention provides a program. The program gets the computer activity data related to the amount of activity other than driving the mobile body during the time when the specific user is not driving the mobile body on the day when the specific user drives the mobile body. You may execute the acquisition procedure. The program is based on the learning result of the correspondence between the activity data of any user and the history of at least one of accidents and violations by any user on the computer, and the program is based on the learning result of the specific user acquired in the acquisition procedure. The activity data may be used as new input data to perform an inference procedure that infers the possibility that a particular user will cause at least one of accidents and violations.

The outline of the above invention does not list all the features of the present invention. A subcombination of these feature groups can also be an invention.

It is a schematic diagram of the inference system 10 by one Embodiment. It is a block diagram of the inference apparatus 100 by one Embodiment. It is a flow diagram of the inference method by one Embodiment. It is a flow chart of the learning method by one Embodiment. It is a figure which shows the example of the computer 1200 which can embody a plurality of aspects of this invention in whole or part.

Hereinafter, the present invention will be described through embodiments of the invention, but the following embodiments do not limit the invention to which the claims are made. Also, not all combinations of features described in the embodiments are essential to the means of solving the invention.

FIG. 1 is a schematic diagram of an inference system 10 according to an embodiment. The inference system 10 includes a wearable device 40 and an inference device 100 that communicate with each other via the communication network 50. The inference system 10 uses machine learning by a learner, and is active in a time zone other than driving a moving body during a time period when the specific user is not driving the moving body on the day when the specific user drives the moving body. A system for inferring the likelihood that a particular user will cause at least one of an accident or a breach, based on quantity-related activity data. The communication network 50 may be wired or wireless.

As an example in the present embodiment, the specific user is the delivery company 20 that delivers the cargo, and the passenger moving body is the delivery truck 30 that the delivery company 20 uses to deliver the cargo. The moving body for passengers may be an automobile such as a motorcycle, a tricycle, or a four-wheeled vehicle, a two-wheeled bicycle, a ship, or an air vehicle, and the user drives the moving body in business. You may drive a moving body in your private life.

In the present embodiment, the amount of activity other than the operation of the delivery truck 30 during the time when the delivery company 20 is not driving the delivery truck 30 is the walking distance, the number of times of going up and down the stairs, the standing time, and the bending and stretching of the knee. It may include the number of times it has been made, and the like. In addition, the amount of activity may include the number of times the upper body is bent forward, backward, laterally or twisted, the number of times the baggage is raised and lowered, the total weight of the baggage carried by human power, and the like. Moreover, the amount of activity may be a combination of two or more of these.

The wearable device 40 is attached to the body of the delivery company 20 and has a plurality of sensors for measuring the activity data of the delivery company 20. The wearable device 40 according to the present embodiment is a communication device, and the measured activity data is transmitted to the inference device 100 via the communication network 50. For example, the wearable device 40 may transmit activity amount data to the inference device 100 in real time at regular time intervals by wireless communication. For example, when the wearable device 40 is turned off by the delivery company 20 by wireless communication, the wearable device 40 infers cumulative activity data from the time when the power is turned on to the time when the wearable device 40 is turned off. It may be transmitted to 100. For example, when the wearable device 40 is connected to the charger installed at the office of the delivery company 20, the wearable device 40 infers the cumulative activity data from the removal from the charger to the connection by wired communication. It may be transmitted to the device 100.

Further, in the present embodiment, the wearable device 40 transmits the user identification information together with the activity amount data to the inference device 100. For example, the wearable device 40 has a user interface, and user identification information may be input by the delivery company 20 before and after measuring the activity data of the delivery company 20. For example, the wearable device 40 may be input with user identification information from another device such as the inference device 100 before being rented out to the delivery company 20, assuming that the wearable device 40 is continuously worn by the delivery company 20. For example, the wearable device 40 wirelessly communicates with a communication terminal such as a smartphone owned by the delivery company 20, an in-vehicle communication device of the delivery truck 30 operated by the delivery company 20, and user identification information from these communication devices. May be received.

The inference device 100 is, for example, a server of a management company that manages the inference device 100. The inference device 100 has a learning device, and the learning device learns a correspondence relationship between activity data of an arbitrary user and a history in which the arbitrary user has caused at least one of an accident and a violation. When the inference device 100 newly acquires the activity amount data of a specific user, the specific user causes at least one of an accident and a violation by using the activity amount data as new input data to the learning device. Infer the possibility. The inference device 100 according to the present embodiment receives the activity amount data of the delivery company 20 from the wearable device 40 via the communication network 50, and based on the activity amount data, the delivery company 20 causes at least one of accidents and violations. Infer the possibility of causing a problem.

The inference device 100 may output the inferred possibility as information that can be understood by humans. As an example, the inference device 100 may output the information to the personal computer of the above management company. The possibility may be, for example, the probability itself that the user causes an accident or a violation, or may indicate whether or not the probability is equal to or higher than a predetermined threshold value. May indicate which of the plurality of predetermined ranges corresponds to. The criteria such as the threshold value and the probability range may be stored in advance in the memory of the inference device 100, for example, RAM, and may be rewritable by the above management company. Is omitted.

FIG. 2 is a block diagram of the inference device 100 according to the embodiment. The inference device 100 includes an acquisition unit 101 and an inference unit 103. The inference device 100 according to the present embodiment further includes an output unit 105 and an accident / violation information database 109.

The acquisition unit 101 is a delivery truck during a time when the delivery company 20 is not driving the delivery truck 30 on the day when a specific user, for example, the delivery company 20, is driving a passenger moving object, for example, the delivery truck 30. Acquire activity amount data related to the amount of activity other than 30 driving. In the present embodiment, the acquisition unit 101 receives the activity amount data from the wearable device 40 via the communication network 50. The acquisition unit 101 may acquire the activity amount data from another device with or without the communication network 50. The acquisition unit 101 may bidirectionally communicate with the wearable device 40 and another communication terminal, for example, the computer of the management company that manages the inference device 100, via the communication network 50.

In the present embodiment, the acquisition unit 101 can acquire activity amount data related to the amount of activity in the operation of the delivery truck 30 during the time zone in which the delivery company 20 is driving the delivery truck 30. The amount of activity in driving the delivery truck 30 during the time when the delivery company 20 is driving the delivery truck 30 includes the time when the arm was raised to operate the steering wheel, the number of times the accelerator or the brake was depressed, and the like. May include.

In this case, the acquisition unit 101 is an amount of activity other than the operation of the delivery truck 30, depending on whether or not, for example, the measured value of the magnitude of the load applied to the body of the delivery company 20 is equal to or higher than a predetermined threshold value. The activity amount data related to the activity amount data may be distinguished from the activity amount data related to the amount of activity in the operation of the delivery truck 30. As a specific example, the acquisition unit 101 is in a state where the delivery company 20 is seated in the driver's seat of the delivery truck 30 according to the measured value of the magnitude of the load applied to the waist of the delivery company 20. That is, it may be possible to determine whether or not the delivery truck 30 is in operation, whereby the above activity amount data may be distinguishable.

Further, for example, the wearable device 40 may be capable of detecting that the delivery company 20 wearing the wearable device 40 is seated and tagging the activity amount data with data indicating that fact. In this case, the acquisition unit 101 may be able to distinguish the above-mentioned activity amount data depending on whether or not the activity amount data acquired from the wearable device 40 is tagged with the data.

Further, in the present embodiment, the acquisition unit 101 acquires the activity amount data together with the user identification information. Thereby, when the inference device 100 receives the activity amount data of the plurality of delivery companies 20 from the plurality of wearable devices 40, the inference device 100 can infer the above possibility for each delivery company 20. The acquisition unit 101 outputs the acquired activity data and the like to the inference unit 103.

The reasoning unit 103 is a specific user acquired by the acquisition unit 101 based on the learning result of the correspondence between the activity data of the arbitrary user and the history of the arbitrary user causing at least one of an accident and a violation. Using the activity data of the above as new input data, it is inferred that the specific user may cause at least one of accidents and violations.

In the inference unit 103 according to the present embodiment, in addition to the activity data of the arbitrary user, the diagnostic data related to the diagnostic result of the driving aptitude test of the arbitrary user and the attribute data related to the attribute of the arbitrary user. , At least one of the additional data is used for training. That is, the inference unit 103 learns the correspondence between the activity data of any user and the additional data and the history of the arbitrary user causing at least one of an accident and a violation.

The inference unit 103 according to the present embodiment has the activity amount data of the delivery company 20 acquired by the acquisition unit 101 and the user identification information of the delivery company 20 acquired by the acquisition unit 101 based on the learning result of the correspondence relationship. Using the additional data of the delivery company 20 stored in the accident / violation information database 109 linked to the new input data, it is inferred that the delivery company 20 may cause at least one of the accident and the violation. ..

The inference unit 103 in the present embodiment may use the learning model properly according to at least one of the above-mentioned new input data type, the above-mentioned correspondence relationship, and at least one of the above-mentioned possibilities. good. As a specific example, the inference unit 103 has a first learning model in which the inference basis is blackboxed but the inference accuracy is high, and the inference basis is blackboxed according to at least one of them. Either one of the second learning models, which can be understood by humans but has low inference accuracy, may be used. The first learning model may be a GBDT (Gradient Boosting Decision Tree), and the second learning model may be a random forest.

For example, the inference unit 103 uses the second learning model when the new input data type is only activity data, and the new input data type is the activity data as well as the diagnostic result of the driving aptitude test. The first learning model may be used when the above is also included. Further, for example, the inference unit 103 uses the second learning model if the use of the above possibility is to determine the user to be educated or to alert the delivery company 20 itself, and the use of the above possibility is used. The first learning model may be used to determine insurance premiums and insurance plans.

The output unit 105 outputs the above possibility inferred by the inference unit 103 as information that can be understood by a person, for example, an employee of a management company of the inference device 100. As an example, the output unit 105 may output the information together with the above reasoning basis when the inference unit 103 uses the above-mentioned second learning model. Further, the output unit 105 may output the information and the like together with the user identification information of the user.

The inference device 100 does not have to include the output unit 105. For example, the inference device 100 may transmit the inferred data indicating the above possibility to the external device and cause the external device to execute each process of the output unit 105 described above. An example of the external device may be the computer of the management company described above.

The accident / violation information database 109 links the activity data of the user and the history of the user causing at least one of the accident and the violation to the user identification information of the user for each of a plurality of arbitrary users. It is attached and stored. The accident / violation information database 109 according to the present embodiment further obtains at least one additional data of diagnostic data related to the diagnostic result of the driving aptitude test of each user and attribute data related to the attribute of the user. It is stored in association with the user identification information of the user. The accident / violation information database 109 is an example of a storage unit.

The above-mentioned driving aptitude test may be, for example, an OD type safety test (https://www.dennoo.co.jp/Denn001.html) or a police agency type driving aptitude test K type. The diagnostic result of the driving aptitude test may include one or more of motor function, health / maturity, personality traits, and driving manners. The driving function may include one or more of attention, judgment, flexibility, determination, precision, stability of movement, and adaptability.

Health / maturity may include one or more of physical health, mental health, and social maturity. Personality traits may include one or more of emotional instability, impulsiveness / unintentional discharge, egocentrism, nervousness / hypersensitivity, and vanity.

The user attributes described above may also include one or more of the following plurality of items: gender, age, driving history, region, license type, past accident history, past violation history, past. Number of accidents, past working hours, past mileage, past injury / illness history, past driving record time, past driving vehicle (small / medium / large vehicle), license type, past labor company. The "past" data referred to here may be the average of the data of the user in the previous flight or the data of the past n days (n = 1, 7, 30, 365, etc.).

The inference device 100 may be provided with a memory in which a program or the like for controlling the entire inference device 100 is stored separately from the accident / violation information database 109, and the memory may include the accident / violation information database 109. good. Each configuration of the inference device 100 reads a program stored in the memory in order to execute each function described above.

The inference device 100 does not have to include the accident / violation information database 109. For example, the inference device 100 may acquire each of the above-mentioned data from the database of the external device and execute the above-mentioned inference or learning.

FIG. 3 is a flow chart of an inference method according to an embodiment. As an example, the flow shown in FIG. 3 may be started by the delivery company 20 turning on the power of the wearable device 40. In this embodiment, it is premised that the delivery company 20 wears the wearable device 40 and turns on the power of the wearable device 40 on the day when the delivery truck 30 is operated for business. The inference device 100 may acquire in advance data indicating a day when the delivery company 20 operates the delivery truck 30 and a day when the delivery truck 30 does not operate. In this case, the activity amount data acquired on the day when the delivery company 20 does not operate may be acquired. , May be discarded without being used to infer the above possibilities.

The wearable device 40 starts measuring the activity amount data of the delivery company 20 equipped with the wearable device 40 (step S101), and measures the activity amount data and the built-in memory until the power of the wearable device 40 is turned off. Continue to accumulate in (step S103: NO). The wearable device 40 may have a built-in timer, and in this case, the date and time data may be tagged with the activity amount data stored in the built-in memory.

In response to the power of the wearable device 40 being turned off (step S103: YES), the wearable device 40 uses the activity data stored in the built-in memory together with the user identification information stored in the built-in memory in the communication network. It is transmitted to the inference device 100 via 50 (step S105). The wearable device 40 may, for example, execute the process of step S105 and delete the activity amount data stored in the built-in memory. The wearable device 40 may update the activity data stored in the built-in memory when the power of the wearable device 40 is turned on and new activity data is measured next time.

The inference device 100 executes an acquisition step of acquiring the activity amount data of the delivery company 20 (step S107). More specifically, the inference device 100 acquires activity data and user identification information from the wearable device 40 via the communication network 50. As described above, the inference device 100 is based on the activity amount data acquired from the wearable device 40, and the amount of activity other than the operation of the delivery truck 30 during the time when the delivery company 20 is not operating the delivery truck 30. If the activity data related to the activity data and the activity data related to the amount of activity in the operation of the delivery truck 30 during the time when the delivery company 20 is driving the delivery truck 30 are included, it is optional. These activity data can be identified by the above method. The inference device 100 may discard the latter activity data without using it for inferring the above possibility.

The inference device 100 executes an inference step of inferring the possibility that the delivery company 20 causes at least one of an accident and a violation based on the activity data of the delivery company 20 acquired in the acquisition stage (step S109). .. More specifically, the inference device 100 acquires at the acquisition stage based on the learning result of the correspondence relationship between the activity data of any user and the history of at least one of accidents and violations by the arbitrary user. Using the activity data of the delivery company 20 as new input data, it is inferred that the delivery company 20 may cause at least one of an accident and a violation.

The inference device 100 according to the present embodiment is associated with the user identification information of the delivery company 20 from the accident / violation information database 109 by using the user identification information of the delivery company 20 acquired together with the activity amount data of the delivery company 20. The additional data of at least one of the diagnostic data and the attribute data of the delivery company 20 stored in the accident / violation information database 109 is extracted. The inference device 100 according to the present embodiment further acquires the activity data of any user and the additional data based on the learning result of the correspondence relationship between the arbitrary user and the history of causing at least one of accidents and violations. Using the activity data of the delivery company 20 acquired by the unit 101 and the additional data of the delivery company 20 extracted from the accident / violation information database 109 as new input data, the delivery company 20 selects at least one of the accident and the violation. Infer the possibility of causing it.

For example, the inference device 100 may learn that, as a result of learning the correspondence, there is a correlation between the combination of activity data and additional data and the frequency of at least one of the accidents and violations. good. In this case, the inference device 100 indicates, for example, that the diagnostic result of the driving aptitude test shown in the diagnostic data of the specific user has low mental health based on the correlation, and the activity of the specific user. If the quantity data indicates that fatigue is accumulating, it may be inferred that the particular user is relatively prone to accidents.

The inference device 100 executes an output step of outputting the inferred possibility as information that can be understood by humans (step S111), and ends the flow. As an example, the flow may be repeated every time the power of the wearable device 40 is turned on.

FIG. 4 is a flow chart of a learning method according to an embodiment. The inference device 100 executes the learning method shown in FIG. 4 before executing the inference method shown in FIG. In addition to this, the inference device 100 may execute the learning method shown in FIG. 4 while the inference method shown in FIG. 3 is not being executed.

The inference device 100 acquires the activity amount data and additional data of an arbitrary user as teacher data from the accident / violation information database 109, and the arbitrary user causes at least one of the accident and the violation as the correct answer data. Acquire the history (step S201). The inference device 100 causes the learner to perform machine learning (step) by inputting activity data and additional data of an arbitrary user and outputting a history of at least one of accidents and violations by the arbitrary user as an output. S203), steps S201 to S203 are repeated until sufficient learning is performed (step S205), and then the flow is terminated.

As described above, according to the inference device 100 according to the plurality of embodiments described above, the operation of the moving body is performed during the time period when the specific user is not driving the moving body on the day when the specific user drives the moving body. Get activity data related to the amount of activity other than. According to the inference device 100, the specific user acquired above is further based on the learning result of the correspondence between the activity data of any user and the history of at least one of accidents and violations by the arbitrary user. Using the activity data of the above as new input data, it is inferred that the specific user may cause at least one of accidents and violations.

According to the inference device 100 having the configuration, data indicating the inferred possibility can be provided to the management company of the inference device 100, whereby, for example, an employee of the management company may have an accident and a violation. Focus on training the delivery company 20 that will cause at least one of the above, or instructing the amount of activity during the time when the delivery company 20 is not driving the delivery truck 30. Can be done. Further, for example, an employee of the management company can use the inference device 100 to perform the above-mentioned with respect to another delivery company 20 having an activity similar in tendency to the activity shown in the activity amount data of the delivery company 20. Similar instruction can be given by omitting sexual reasoning. Further, according to the inference device 100, as input data for inferring the above possibility of a specific user, by using additional data such as diagnostic data and attribute data of the specific user in addition to the activity amount data. , The inference accuracy can be further improved.

In the above-mentioned plurality of embodiments, the acquisition unit 101 has been described as a configuration for acquiring activity amount data measured by a sensor mounted on the body of a specific user. In other words, the acquisition unit 101 is configured to acquire the activity amount data directly measured from a specific user. As an example, the activity data may be measured directly from a specific user by the wearable device 40, or may be measured directly from a specific user by, for example, a pedometer (registered trademark) or a load sensor.

Instead of this, the acquisition unit 101 may acquire the activity amount data based on the data obtained without being directly measured by a specific user. For example, the acquisition unit 101 acquires data related to the amount of activity other than the operation of the moving body during the time period when the specific user is not driving the moving body, and the activity amount data is based on the data. May be estimated.

More specifically, the acquisition unit 101 may estimate the activity data of the specific user based on the position information of at least one of the specific user and the moving body. For example, when the specific user is the delivery company 20 and the moving object is the delivery truck 30, the acquisition unit 101 can park and stop the delivery truck 30 based on the position information. It is determined that the delivery company 20 has been active other than driving the delivery truck 30 in at least one of the areas where there are few vehicles and the area where the delivery company 20 needs to carry luggage and go up and down the stairs on multiple floors. If so, a relatively large amount of activity data may be estimated. In this case, a GPS may be provided on the wearable device 40 worn by the delivery company 20 or the in-vehicle device of the delivery truck 30, and the acquisition unit 101 may acquire the above-mentioned position information from the GPS. .. The area where the delivery company 20 needs to carry luggage and go up and down the plurality of floors by stairs may include an area where many office buildings, apartments and condominiums are constructed, and an area where the installation rate of elevators is low. The area where the delivery company 20 needs to carry the cargo and go up and down the plurality of floors on the stairs may be an area designated in advance by the delivery company 20 or the place of work of the delivery company 20, and the inference device 100 maps. It may be an area judged based on information.

Further, in the above-mentioned plurality of embodiments, the acquisition unit 101 may further acquire movement route information of at least one of a specific user and a passenger moving body. The movement route information may be a history of location information acquired as described above, and is preset as a delivery route information of a package by a specific user or a specific user's office, that is, a specific user. It may be information on the road that is driven by a moving object.

In this case, the inference unit 103 uses the acquisition unit 101 based on the learning result of the correspondence between the activity data and the movement route information of any user and the history of causing at least one of accidents and violations by the arbitrary user. It may be inferred that the specific user may cause at least one of an accident and a violation by using the activity amount data and the movement route information of the specific user acquired by the above as new input data. For example, the inference unit 103 learns that as a result of learning the correspondence relationship described above, there is a correlation between the combination of activity data and a specific road and the frequency of at least one of accidents and violations. May be good. In this case, for example, the inference unit 103 indicates that the road shown in the travel route information of the specific user is a road in which at least one of an accident and a violation occurs frequently based on the correlation, and the specific user. If the activity data of the above indicates that fatigue is accumulated, it may be inferred that the specific user is in a state of being relatively prone to an accident.

Further, in the above-mentioned plurality of embodiments, the acquisition unit 101 further includes blinking, staring, facial expression, blood pressure, and heartbeat detected by the specific user during the time period when the specific user is driving the passenger moving body. , Arousal, and behavior, biometric data related to at least one of the following may be acquired. The biometric data may be detected by the wearable device 40, or may be detected by a driver status monitor mounted on the delivery truck 30 or the like. As a specific example, biometric data includes the number of blinks of a specific user, the state of blood pressure, heartbeat, etc., sadness, joy, anger, facial expression, and impatience while the specific user is driving a moving body. It may include facial expressions such as, and the number of times the line of sight points to a specific place. The specific portion may be, for example, a meter, a rear-view mirror, a side mirror, or the like.

In this case, the inference unit 103 uses the acquisition unit 101 based on the learning result of the correspondence between the activity data and the biometric data of any user and the history of at least one of the accidents and violations by the arbitrary user. Using the acquired activity data and biometric data of a specific user as new input data, it may be inferred that the specific user may cause at least one of an accident and a violation.

Further, in the above-mentioned plurality of embodiments, the reasoning system 10 satisfies the condition inferred by the reasoning device 100, which satisfies a predetermined condition that a specific user may cause at least one of an accident and a violation. Further, a warning device may be provided to warn a specific user. The warning device may be a communication terminal such as a smartphone or tablet owned by the specific user, or may be a device mounted on a mobile body driven by the specific user, and the specific user may use the warning device. It may be attached to the wearable device 40 to be attached. The warning device may store the condition in a built-in memory, for example, RAM in advance, and the condition may be rewritable by the management company, the specific user, or the like.

Further, in the above plurality of embodiments, the inference device 100 has been described as an example as a server of a management company that manages the inference device 100. Instead, the inference device 100 may be a device mounted on a passenger moving body driven by a specific user, or may be attached to a wearable device 40.

Various embodiments of the present invention may be described with reference to flowcharts and block diagrams, wherein the block is (1) a stage of the process in which the operation is performed or (2) a device having a role of performing the operation. May represent a section of. Specific stages and sections are implemented by dedicated circuits, programmable circuits supplied with computer-readable instructions stored on computer-readable media, and / or processors supplied with computer-readable instructions stored on computer-readable media. It's okay. Dedicated circuits may include digital and / or analog hardware circuits, and may include integrated circuits (ICs) and / or discrete circuits. A programmable circuit is a memory element such as logical AND, logical OR, logical XOR, logical NAND, logical NOR, and other logical operations, flip-flops, registers, field programmable gate arrays (FPGAs), programmable logic arrays (PLA), etc. May include reconfigurable hardware circuits, including, etc.

The computer readable medium may include any tangible device capable of storing instructions executed by the appropriate device, so that the computer readable medium having the instructions stored therein is specified in a flow chart or block diagram. It will be equipped with a product that contains instructions that can be executed to create means for performing the operation. Examples of computer-readable media may include electronic storage media, magnetic storage media, optical storage media, electromagnetic storage media, semiconductor storage media, and the like. More specific examples of computer-readable media include floppy (registered trademark) disks, diskettes, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), Electrically erasable programmable read-only memory (EEPROM), static random access memory (SRAM), compact disc read-only memory (CD-ROM), digital versatile disc (DVD), Blu-ray (RTM) disc, memory stick, integrated A circuit card or the like may be included.

Computer-readable instructions include assembler instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcodes, firmware instructions, state-setting data, or Smalltalk®, JAVA®, C ++, etc. Object-oriented programming languages, and either source code or object code written in any combination of one or more programming languages, including traditional procedural programming languages such as the "C" programming language or similar programming languages. May include.

Computer-readable instructions are used locally or to a local area network (LAN), wide area network (WAN) such as the Internet, to a general purpose computer, a special purpose computer, or the processor or programmable circuit of another programmable data processing device. ) May execute computer-readable instructions to create means for performing the operations specified in the flowchart or block diagram. Examples of processors include computer processors, processing units, microprocessors, digital signal processors, controllers, microcontrollers, and the like.

FIG. 5 shows an example of a computer 1200 in which a plurality of aspects of the present invention can be embodied in whole or in part. The program installed on the computer 1200 causes the computer 1200 to function as an operation associated with the device according to an embodiment of the present invention or as one or more "parts" of the device, or the operation or the one or more "parts". A unit can be run and / or a computer 1200 can be run a process according to an embodiment of the invention or a step in the process. Such a program may be executed by the CPU 1212 to cause the computer 1200 to perform a specific operation associated with some or all of the blocks of the flowcharts and block diagrams described herein.

The computer 1200 according to this embodiment includes a CPU 1212, a RAM 1214, a graphic controller 1216, and a display device 1218, which are interconnected by a host controller 1210. The computer 1200 also includes input / output units such as a communication interface 1222, a hard disk drive 1224, a DVD-ROM drive 1226, and an IC card drive, which are connected to the host controller 1210 via the input / output controller 1220. The computer also includes legacy input / output units such as the ROM 1230 and keyboard 1242, which are connected to the input / output controller 1220 via the input / output chip 1240.

The CPU 1212 operates according to the programs stored in the ROM 1230 and the RAM 1214, thereby controlling each unit. The graphic controller 1216 acquires image data generated by the CPU 1212 in a frame buffer or the like provided in the RAM 1214 or the graphic controller 1216 itself, and displays the image data on the display device 1218.

The communication interface 1222 communicates with other electronic devices via the network. The hard disk drive 1224 stores programs and data used by the CPU 1212 in the computer 1200. The DVD-ROM drive 1226 reads the program or data from the DVD-ROM 1201 and provides the program or data to the hard disk drive 1224 via the RAM 1214. The IC card drive reads the program and data from the IC card and / or writes the program and data to the IC card.

The ROM 1230 internally stores a boot program or the like executed by the computer 1200 at the time of activation, and / or a program depending on the hardware of the computer 1200. The input / output chip 1240 may also connect various input / output units to the input / output controller 1220 via a parallel port, a serial port, a keyboard port, a mouse port, and the like.

The program is provided by a computer-readable storage medium such as a DVD-ROM 1201 or an IC card. The program is read from a computer-readable storage medium, installed in a hard disk drive 1224, RAM1214, or ROM1230, which is also an example of a computer-readable storage medium, and executed by the CPU 1212. The information processing described in these programs is read by the computer 1200 and provides a link between the program and the various types of hardware resources described above. The device or method may be configured to implement the operation or processing of information in accordance with the use of the computer 1200.

For example, when communication is executed between the computer 1200 and an external device, the CPU 1212 executes a communication program loaded in the RAM 1214, and performs communication processing on the communication interface 1222 based on the processing described in the communication program. You may order. Under the control of the CPU 1212, the communication interface 1222 reads and reads transmission data stored in a transmission buffer area provided in a recording medium such as a RAM 1214, a hard disk drive 1224, a DVD-ROM 1201, or an IC card. The data is transmitted to the network, or the received data received from the network is written to the reception buffer area or the like provided on the recording medium.

Further, the CPU 1212 makes the RAM 1214 read all or necessary parts of the file or the database stored in the external recording medium such as the hard disk drive 1224, the DVD-ROM drive 1226 (DVD-ROM1201), and the IC card. Various types of processing may be performed on the data on the RAM 1214. The CPU 1212 may then write back the processed data to an external recording medium.

Various types of information, such as various types of programs, data, tables, and databases, may be stored on recording media for information processing. The CPU 1212 describes various types of operations, information processing, conditional judgment, conditional branching, unconditional branching, and information retrieval described in various parts of the present disclosure with respect to the data read from the RAM 1214. Various types of processing may be performed, including / replacement, etc., and the results are written back to the RAM 1214. Further, the CPU 1212 may search for information in a file, a database, or the like in the recording medium. For example, when a plurality of entries each having an attribute value of the first attribute associated with the attribute value of the second attribute are stored in the recording medium, the CPU 1212 is the first of the plurality of entries. The attribute value of the attribute of is searched for the entry that matches the specified condition, the attribute value of the second attribute stored in the entry is read, and the attribute value of the second attribute is changed to the first attribute that satisfies the predetermined condition. You may get the attribute value of the associated second attribute.

The program or software module according to the above description may be stored on a computer 1200 or in a computer-readable storage medium near the computer 1200. Further, a recording medium such as a hard disk or RAM provided in a dedicated communication network or a server system connected to the Internet can be used as a computer-readable storage medium, whereby the program can be sent to the computer 1200 via the network. offer.

Although the present invention has been described above using the embodiments, the technical scope of the present invention is not limited to the scope described in the above embodiments. It will be apparent to those skilled in the art that various changes or improvements can be made to the above embodiments. It is clear from the description of the claims that the form with such changes or improvements may be included in the technical scope of the present invention.

The order of execution of each process such as operation, procedure, step, and step in the apparatus, system, program, and method shown in the claims, specification, and drawings is particularly "before" and "prior to". It should be noted that it can be realized in any order unless the output of the previous process is used in the subsequent process. Even if the scope of claims, the specification, and the operation flow in the drawings are explained using "first", "next", etc. for convenience, it means that it is essential to carry out in this order. It's not a thing.

10 Inference system 20 Delivery company 30 Delivery truck 40 Wearable device 50 Communication network 100 Inference device 101 Acquisition unit 103 Inference unit 105 Output unit 109 Accident / violation information database 1200 Computer 1201 DVD-ROM
1210 Host controller 1212 CPU
1214 RAM
1216 Graphic controller 1218 Display device 1220 I / O controller 1222 Communication interface 1224 Hard disk drive 1226 DVD-ROM drive 1230 ROM
1240 I / O chip 1242 keyboard

Claims (15)

Acquisition unit that acquires activity amount data related to the amount of activity other than driving the moving body during the time period when the specific user is not driving the moving body on the day when the specific user drives the moving body. When,
The activity of the specific user acquired by the acquisition unit based on the learning result of the correspondence between the activity amount data of the arbitrary user and the history of the arbitrary user causing at least one of an accident and a violation. An inference device including an inference unit that infers the possibility that the specific user causes at least one of the accident and the violation by using the quantity data as new input data. The amount of activity other than driving the moving body is the walking distance, the number of times of going up and down stairs, the standing time, the number of times the knees are bent and stretched, the number of times the upper body is bent forward, backward, laterally or twisted, and luggage. Including at least one of the number of times the item was raised or lowered and the total weight of the manually carried baggage.
The inference device according to claim 1. The acquisition unit acquires the activity amount data measured by a sensor mounted on the body of the specific user.
The inference device according to claim 1 or 2. The acquisition unit estimates the activity amount data based on the position information of at least one of the specific user and the moving body.
The inference device according to claim 1 or 2. The moving object and the specific user are delivery trucks and delivery companies, respectively.
Based on the location information, the acquisition unit has at least an area where the space where the moving body can be parked and stopped is small, and an area where the specific user needs to carry luggage and go up and down the stairs on a plurality of floors. When it is determined that the specific user has been active other than the operation of the moving body, a relatively large amount of the activity data is estimated.
The inference device according to claim 4. At least one additional data of the diagnostic data related to the diagnostic result of the driving aptitude test of the specific user and the attribute data related to the attribute of the specific user is linked with the user identification information of the specific user. It also has a storage unit that can be attached and stored.
The acquisition unit acquires the activity amount data together with the user identification information of the specific user.
The inference unit receives the activity data of the specific user and the activity data of the specific user acquired by the acquisition unit based on the learning result of the correspondence between the activity data of the arbitrary user and the additional data and the history. Inferring the possibility of the specific user by using the additional data stored in the storage unit associated with the user identification information of the specific user acquired by the acquisition unit as the new input data. do,
The inference device according to any one of claims 1 to 5. The acquisition unit further includes blinking, staring, facial expression, blood pressure, heart rate, arousal, and behavior detected by the specific user during the time when the specific user is driving the moving body. Obtain biometric data related to at least one of
The inference unit has the activity data of the specific user and the living body acquired by the acquisition unit based on the learning result of the activity data of the arbitrary user and the correspondence between the biometric data and the history. Using the data as the new input data, infer the possibility of the specific user.
The inference device according to any one of claims 1 to 6. The acquisition unit further acquires movement route information of at least one of the specific user and the moving body.
The inference unit has the activity data of the specific user and the activity data of the specific user acquired by the acquisition unit based on the learning result of the correspondence between the activity data of the arbitrary user and the movement route information and the history. Using the travel route information as the new input data, the possibility of the specific user is inferred.
The inference device according to any one of claims 1 to 7. The inference unit has blackboxed grounds for inference according to at least one of the new input data type, the correspondence and at least one of the possibilities, and the inference accuracy. Use either the first learning model with high inference or the second learning model with low inference accuracy, although the grounds for inference are not blackboxed and can be understood by humans.
The inference device according to any one of claims 1 to 8. The first learning model is a GBDT (Gradient Boosting Decision Tree), and the second learning model is a random forest.
The inference device according to claim 9. Further provided with an output unit that outputs the possibility inferred by the inference unit as information that can be understood by humans.
The output unit outputs the information together with the basis of the inference when the inference unit uses the second learning model.
The inference device according to claim 9 or 10. The inference device according to any one of claims 1 to 11.
An inference system including a wearable device that is attached to the body of the specific user, measures the activity amount data of the specific user, and transmits the activity data to the inference device. The inference system according to claim 12, further comprising a warning device that warns the specific user when the possibility inferred by the inference device satisfies a predetermined condition. Acquisition stage to acquire activity amount data related to the amount of activity other than driving the moving body during the time period when the specific user is not driving the moving body on the day when the specific user drives the moving body. When,
The activity of the specific user acquired at the acquisition stage based on the learning result of the correspondence between the activity amount data of the arbitrary user and the history of causing at least one of the accidents and violations of the arbitrary user. An inference method comprising an inference step of inferring the possibility that the particular user will cause at least one of the accident and the violation, using the quantity data as new input data. On the computer
Acquisition procedure for acquiring activity amount data related to the amount of activity other than driving the moving body during the time period when the specific user is not driving the moving body on the day when the specific user drives the moving body. When,
The activity of the specific user acquired in the acquisition procedure based on the learning result of the correspondence between the activity data of the arbitrary user and the history of causing at least one of the accidents and violations of the arbitrary user. A program that uses quantitative data as new input data to execute an inference procedure that infers the possibility that the specific user causes at least one of the accident and the violation.

Images