KR20230038129A - Method and Apparatus for Analying Driving Pattern - Google Patents

Abstract

Disclosed are a method for analyzing a driving pattern and a device therefor, which can improve the accuracy of driving pattern analysis. A computing device which analyzes a driving pattern of a vehicle comprises: at least one processor; and at least one memory which has a non-temporary storage medium and stores program codes executed by the processor. According to the present invention execution of the program codes, the at least one processor may perform operations which includes: a step of acquiring at least one collection information selected from a collection information group consisting of human factor information, driving record information, and vehicle information; a step of generating a trip driving feature value indicating a driving feature of a trip section unit by using the driving record information; and a step of generating a driving pattern feature value by using the human factor information or the vehicle information, and the trip driving feature value.

Description

Driving pattern analysis method and apparatus therefor {Method and Apparatus for Analyzing Driving Pattern}

The present invention relates to a method and an apparatus for analyzing a driving pattern based on a driving pattern feature value generated using at least one piece of collected information.

The contents described in this section simply provide background information on the embodiments of the present invention and do not constitute prior art.

Conventionally, there have been methods for analyzing a driving pattern of a vehicle using various sensors mounted on the vehicle. In the case of shared vehicles such as car-sharing vehicles or rental cars, various drivers may exist and the driving patterns of the vehicles may also vary. It is inevitably different from the evaluation method, but existing technologies do not suggest a special plan for this.

The present invention provides a driving pattern analysis method for generating a driving pattern feature value using at least one piece of collected information, transmitting the generated driving pattern feature value to an external device, or analyzing a driving pattern using the driving pattern feature value, and the same. Its main purpose is to provide a device for

According to one aspect of the present invention, in a computing device for analyzing a driving pattern of a vehicle to achieve the above object, the computing device includes: at least one processor; It has a non-transitory storage medium and includes at least one memory for storing program codes executed by the processor, wherein the at least one processor converts human factor information, driving record information, and vehicle information according to the execution of the program codes. obtaining at least one piece of collection information selected from the collected information group; generating a trip driving feature value representing a driving characteristic of each trip section by using the driving record information; and generating a driving pattern feature value using the human factor information or the vehicle information and the trip driving feature value.

In addition, according to another aspect of the present invention, at least one processor for achieving the above object; A driving pattern analysis method performed by a computing device having a non-transitory storage medium and including at least one memory storing program codes executed by the processor, the computing device comprising: human factor information, driving record information and obtaining at least one collection information selected from a collection information group consisting of vehicle information; generating a trip driving feature value representing a driving characteristic of each trip section by using the driving record information; and generating a driving pattern feature value using the human factor information or the vehicle information and the trip driving feature value.

As described above, the present invention evaluates the driving habits of shared vehicle users from the user's point of view, the vehicle's point of view, and the environment's point of view by considering all human factor information, driving record information, and vehicle information. Accuracy of driving pattern analysis can be improved.

In addition, the present invention can provide various additional services using the analysis result of the driving pattern, and when calculating the insurance rate, it is possible to calculate the insurance rate that better reflects the driving pattern of the shared vehicle driver.

1 is a block diagram schematically illustrating a driving pattern analysis system according to an embodiment of the present invention.
2 is a schematic block diagram of a driving pattern analysis device according to an embodiment of the present invention.
3 is a block diagram schematically showing a result data processing unit of a driving pattern analysis device according to an embodiment of the present invention.
4 is a block diagram schematically showing a hardware configuration of a driving pattern analysis device according to an embodiment of the present invention.
5 is a flowchart illustrating a driving pattern analysis method according to an embodiment of the present invention.
6 is a diagram for explaining a feature value generation operation according to an embodiment of the present invention.
7 is an exemplary view for explaining driving pattern feature values according to an embodiment of the present invention.
8 is an exemplary diagram illustrating an analysis result of driving pattern feature values according to an embodiment of the present invention.
9 is an exemplary diagram for explaining a repetitive behavior detection operation of the driving pattern analysis device according to an embodiment of the present invention.
10 is an exemplary diagram for explaining a behavior prediction processing operation of a driving pattern analysis device according to an embodiment of the present invention.
11 is an exemplary diagram for explaining a user prediction processing operation of the driving pattern analysis device according to an embodiment of the present invention.
12 to 14 are diagrams for explaining a driving pattern analysis operation based on a neural network according to an embodiment of the present invention.
15 is a reference diagram illustrating factors related to a user's service use and driving disposition evaluation according to an embodiment of the present invention.
16A, 16B, and 16C are tables showing measurement items for evaluating a user's service use and driving disposition according to an embodiment of the present invention.
17 is an example of evaluating a user's service use and driving disposition according to an embodiment of the present invention.
18 illustrates a concept related to additional information of trip data according to an embodiment of the present invention.
19 is a flowchart illustrating a method of evaluating a user driving pattern according to an embodiment of the present invention.
20 is a flowchart illustrating a method of evaluating a driving pattern of a vehicle according to an embodiment of the present invention.
21 is a flowchart illustrating a method of evaluating a driving pattern of a vehicle according to an embodiment of the present invention.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. In describing the present invention, if it is determined that a detailed description of a related known configuration or function may obscure the gist of the present invention, the detailed description will be omitted. In addition, although preferred embodiments of the present invention will be described below, the technical idea of the present invention is not limited or limited thereto and can be modified and implemented in various ways by those skilled in the art. Hereinafter, a driving pattern analysis method and an apparatus therefor proposed in the present invention will be described in detail with reference to the drawings.

1 is a block diagram schematically illustrating a driving pattern analysis system according to an embodiment of the present invention.

The driving pattern analysis system 10 according to the present embodiment includes an information collecting device 100 and a driving pattern analyzing device 200 . The driving pattern analysis system 10 of FIG. 1 is according to an embodiment, and all blocks shown in FIG. 1 are not essential components, and some blocks included in the driving pattern analysis system 10 in another embodiment are added. , may be changed or deleted.

The information collection device 100 includes at least one collection device and collects human factor information, driving record information, and vehicle information. The information collection device 100 may be included in a rental/shared vehicle or may be a terminal interworking with the vehicle.

For example, the information collection device 100 may be a sensor installed in a vehicle, a black box, or the like. In-vehicle sensors acquire information necessary for analyzing a driving pattern, such as a speed sensor, an acceleration sensor, a gyro sensor, a GPS, and a temperature sensor. According to another embodiment of the present invention, a sensor corresponding to this may already be built into the vehicle, and it is also possible to obtain a sensor value from the already built-in sensor and provide it to the driving pattern analysis device. However, there are cases in which it is difficult to share the internal information of the vehicle with the outside, and it is desirable in terms of data accuracy to obtain necessary information through a separate sensor device for cross-checking the internal information.

The black box is not an essential component of the present invention and may be a related external component, and image information obtained through the black box may be provided to the driving pattern analysis device.

In the present invention, vehicles include devices that can be ridden, such as cars and trucks. In addition, although a typical four-wheeled shared vehicle is the subject of the present invention, the present technology may also be applied to a general vehicle or a vehicle such as a taxi.

Human factor information refers to information about personal information about a driver, and may include an identification ID, gender, age, driving experience period, and the like. Information included in the human factor information may be changed according to the driver's choice of information disclosure.

The driving record information may include all information related to driving while the vehicle is driving. For example, the driving record information may include driving distance, driving time, acceleration, braking, rotation, weather environment, location information, driving time zone, and the like.

The vehicle information refers to information about a vehicle driven by a driver. For example, the vehicle information may include vehicle year, tire condition, accumulated mileage, and the like.

The driving pattern analysis device 200 refers to a device that analyzes a driving pattern of a driving vehicle.

The driving pattern analysis device 200 generates a driving pattern feature value using at least one piece of collected information, transmits the generated driving pattern feature value to an external device, or analyzes the driving pattern using the driving pattern feature value.

In general, there is a method of evaluating a corresponding behavior by complexly analyzing external environment information at a given time based on information collected through a vehicle.

However, a general evaluation method is a method of judging behavior through scoring, ranking, and the like, based on a simple reference value that specifies the behavior that occurred at a corresponding vehicle and time point.

Accordingly, the general evaluation method focuses on a processing operation of synthesizing the evaluation and judgment of the occurred actions into the same result value and comparing it with the average or range value of the comparison target.

On the other hand, the driving pattern analysis device 200 according to an embodiment of the present invention can know the characteristics of the behavior, and as the behavior is repeated, 'the degree of similarity of the behavior, the degree of change in the behavior, the degree of repetition of the behavior, etc. are recorded so as to be confirmed. generated driving pattern feature values.

The driving pattern analyzing apparatus 200 may analyze or determine behaviors after a predetermined point in time for driving through driving pattern feature values.

In addition, the driving pattern analyzing apparatus 200 may predict the possibility of occurrence of a specific behavior only with the behaviors up to a specific point by generating driving pattern feature values so as to know whether the behavior has become a habit or the degree of tendency of the behavior.

In particular, the driving pattern analysis apparatus 200 generates driving pattern feature values in order to intensively determine each specific behavior occurring during driving.

Since the accumulated behavior is connected to the problem of the possibility of going to damage to others or society later, in order to solve this problem, the driving pattern analysis device 200 generates the evaluation result for prediction as a driving pattern feature value. can do.

That is, in the driving pattern analysis device 200, rather than simply calculating quick judgment results for predicting behavior, the characteristics of behaviors recorded for situations occurring during driving are relatable in a predetermined expression method (eg, code). It is important for the evaluation and judgment of behavior to perform the characterization of the base behavior and express it as a driving pattern feature value.

For example, the driving pattern analyzer 200 may perform a first action and a second action consecutively or perform a third action a predetermined number of times (eg, three times) at a first point in time in the first environment (driving environment). If it occurs, it may be recorded as a driving pattern feature value including an analysis result of the dangerous behavior of 'OOO'-'QRS' characteristics expressed based on a preset expression rule and an identification mark. In addition, a specific reason for the driving behavior may be additionally expressed in the driving pattern feature value.

The driving pattern analyzing apparatus 200 may derive an analysis result by evaluating the subsequent association or related behavior according to the identification mark so as to be intensively analyzed for a specified period range.

That is, the driving pattern analysis device 200 generates driving pattern feature values using a pre-arranged rule and coded recording method using a system algorithm, and analyzes driving behavior using the generated driving pattern feature values. can be made

In addition, the driving pattern analysis apparatus 200 may also add information on which behaviors and in which manner are accumulated to the driving pattern feature values.

A detailed description of the driving pattern analysis device 200 will be described in FIGS. 2 and 3 .

2 is a schematic block diagram of a driving pattern analysis device according to an embodiment of the present invention.

The driving pattern analysis device 200 according to the present embodiment includes a collection information acquisition unit 210 and a feature value generation unit 220 . In addition, the driving pattern analyzing device 200 further includes a pre-processing unit 222, a driving pattern analyzing unit 230, and a result data processing unit 240. The driving pattern analyzing device 200 of FIG. 2 is according to an embodiment, and all blocks shown in FIG. 2 are not essential components, and some blocks included in the driving pattern analyzing device 200 in another embodiment are added. , may be changed or deleted.

The collection information acquisition unit 210 acquires at least one piece of collection information selected from a collection information group consisting of human factor information, driving record information, and vehicle information.

The collection information acquisition unit 210 acquires collection information including human factor information about the driver, driving record information about driving of the vehicle, and vehicle information about the state of the vehicle from at least one collection device. Here, the included human factor information may be changed according to the driver's selection.

The collection information acquisition unit 210 may obtain vehicle history information using vehicle information. For example, when vehicle information for specifying a vehicle is transmitted to the processor, the collection information acquisition unit 210 may retrieve history information of the corresponding vehicle from a database, memory, or the like.

The collection information acquisition unit 210 may additionally acquire a driver profile specified from human factor information or a predictive driver profile generated to specify a driver based on the trip driving feature value. Here, the driver profile refers to information such as identification ID, gender, age, address, period of driving, details of fines and penalties, accident and insurance history, affiliation, receivable status, area of interest and involvement, payment and cost information, etc.

When human factor information is specified and input, the collection information acquisition unit 210 may retrieve a driver profile matching the input human factor information from the database. Meanwhile, when it is difficult to retrieve a driver profile because human factor information is not properly input, the collection information acquisition unit 210 may generate a predictive driver profile for predicting a driver using trip driving feature values.

The feature value generation unit 220 generates driving pattern feature values based on collected information.

The feature value generation unit 220 generates trip driving feature values representing driving characteristics in units of trip sections using the driving record information, and uses the human factor information or vehicle information and the trip driving feature values to generate driving pattern feature values. generate

The feature value generation unit 220 includes speed, travel distance, travel time, acceleration, number of accelerations, engine oil temperature, gear position, transmission oil temperature, trip data snapshot, RPM, number of constant speed driving, idling, and fuel economy in the trip section. and a warm-up time to generate a trip driving characteristic value for the driving characteristic according to one or more items selected from the group consisting of a warm-up time. Here, the trip driving characteristic value may include a trip driving characteristic value for each single item according to one selected item or a trip driving characteristic value for each plurality of items according to the selected plurality of items.

The feature value generation unit 220 may generate driving pattern feature values using vehicle history information and trip driving feature values.

Meanwhile, the feature value generation unit 220 additionally uses a driver profile specified from vehicle history information, trip driving feature values, and human factor information, or a predictive driver profile generated to specify a driver based on trip driving feature values. Thus, driving pattern feature values may be generated.

The feature value generation unit 220 may provide the generated driving pattern feature value to an external device, but is not necessarily limited thereto, and sends the driving pattern feature value generated for analysis of the driving pattern to the driving pattern analyzer 230. may be conveyed.

The driving pattern feature value generated by the feature value generation unit 220 may be generated based on a preset trip section unit. Here, the trip unit refers to a unit dividing a time interval in which the ignition of the vehicle is switched from an on state to an off state.

The feature value generation unit 220 may generate sub-trip driving feature values representing driving characteristics of a sub-trip section shorter than the trip section.

The trip driving feature value may be obtained by connecting sub-trip driving feature values obtained in the trip section.

The sub-trip driving feature value may be expressed as a sub-trip driving code in which driving characteristics in a given sub-trip section are coded. The trip driving feature value may be expressed as a trip driving code in which sub trip driving codes are connected.

Coding the driving characteristics in a given subtrip driving section may be an operation of expressing the driving characteristics with codes to correspond to the number of occurrences of the driving characteristics or the level of the driving characteristics.

Meanwhile, the sub-trip driving feature value may be a sub-trip feature vector for driving characteristics in a given sub-trip section. The trip driving feature value may be a feature vector obtained by extending the dimension of the sub-trip feature vector through a process of combining the sub-trip feature vectors.

The pre-processing unit 222 performs an operation of adjusting driving pattern feature values generated for driving pattern analysis.

The pre-processing unit 222 may summarize trip driving feature values.

The preprocessing unit 222 generates summarized trip driving feature values by excluding some sub-trip driving feature values according to a predetermined criterion among the sub-trip driving feature values included in the trip driving feature values, and generates trip driving feature values. The driving pattern feature value is adjusted using the value.

Meanwhile, the pre-processing unit 222 applies the trip driving feature values to the first neural network, which has been pre-learned to summarize the driving pattern through the first group of training data, to generate the summarized trip driving feature values. there is.

Additionally, the pre-processing unit 222 may perform pre-processing for summarizing or expanding driving pattern feature values based on preset pre-processing rules.

For example, the pre-processing unit 222 records behavior numerical values measured in the terminal by pre-classification settings such as 'general range' and 'caution/danger behavior range', and records data corresponding to an unnecessary range according to the utilization method. Except, driving pattern feature values may be generated.

In addition, the pre-processing unit 222 may assign action rules according to the state of the vehicle, such as 'measurement immediately/when switching to parking mode/when trip ends/booking ends' in processing behavior record data coded with measured information. there is.

In addition, the pre-processing unit 222 performs a vehicle such as a 'forced switching mode to end driving' when the same pattern behavior is measured in a driving state of a user whose 'weighting' based on 'repetition status' and 'risk level reading' of a specific behavior is recorded. Driving pattern feature values can be created by adding code for state transition control. For example, the preprocessing unit 222 may generate a code for utilizing an immobilizer function for preventing theft through an existing additionally installed device such as cutting off fuel supply. The pre-processing unit 222 may generate a code for Edge AI to determine and control as soon as the activation of the condition 'behavior pattern of the driver + degree of fixation + degree of similarity is measured.

In addition, the pre-processing unit 222 performs reading based on the driver's previous driving record (previous recent trip evaluation code and cumulative driving evaluation code: download) according to the reservation information collected from the terminal of the shared/rental vehicle. Thus, driving pattern feature values may be generated. Since the pre-processing unit 222 directly links/connects the necessary numerical values with the coded data, it can be sufficiently processed even with the logic calculation capability of the low-end terminal even during later evaluation and recording.

The driving pattern analyzer 230 analyzes the driving pattern using the driving pattern characteristic value for each trip unit.

The driving pattern analyzer 230 may generate an analysis result for a driving pattern in the trip section using the summarized trip driving characteristic value included in the driving pattern characteristic value.

The driving pattern analyzer 230 may determine a single attribute class according to each item selected from the trip driving characteristic values or a complex attribute class defined by considering a plurality of items selected from the trip driving characteristic values. For example, the single attribute class may be a rapid acceleration attribute class generated based on the feature value of a single item for 'sudden acceleration', and the composite attribute class may be a plurality of items such as rapid acceleration, rapid deceleration, and average speed below average. It may be a driving inexperienced attribute class generated based on feature values.

Meanwhile, the driving pattern analyzer 230 uses a driver profile specified from the summarized trip driving feature values, vehicle history information, and human factor information, or a predictive driver profile generated to specify a driver based on the trip driving feature values. In order to determine a driving pattern through training data of a second group that is at least partially different from the first group, it may be applied to a second neural network in which learning has been performed in advance to generate an analysis result for the driving pattern.

The resulting data processing unit 240 may provide driving pattern detection result data including all or part of the driving pattern analysis result to an external device. Here, the result data processing unit 240 may provide driving pattern detection result data including all or part of driving pattern feature values.

The resulting data processing unit 240 may generate driving pattern detection result data by filtering the driving pattern analysis result to correspond to a request from an external device or a preset data format.

In addition, the result data processing unit 240 may perform a repeat behavior detection step of extracting data on the driver's repetitive behavior from the driving pattern analysis result, thereby generating driving pattern detection result data including the repetitive behavior identification result.

In addition, the result data processing unit 240 performs a behavior prediction processing step of extracting data on the driver's behavior prediction from the driving pattern analysis result, and generates driving pattern detection result data including the derived behavior prediction result for a specific driver. can do.

In addition, the resultant data processing unit 240 may perform a user prediction processing step of extracting data for identifying or estimating a specific behavioral user from the driving pattern analysis result to generate driving pattern detection result data including the user identification result. .

A detailed description of the result data processing unit 240 will be described in FIG. 3 .

3 is a block diagram schematically showing a result data processing unit of a driving pattern analysis device according to an embodiment of the present invention.

The result data processing unit 240 of the driving pattern analysis device 200 according to the present embodiment includes the driving pattern detection unit 242, the repetitive behavior detection unit 244, the behavior prediction processing unit 246, and the user prediction processing unit 248. include

The driving pattern detection unit 242 generates driving pattern detection result data by filtering the driving pattern analysis result to correspond to a request from an external device or a preset data format.

The repetitive behavior detecting unit 244 extracts data on the driver's repetitive behavior from the driving pattern analysis result and generates driving pattern detection result data including a repetitive behavior identification result.

The behavior prediction processing unit 246 extracts the driver's behavior prediction data from the driving pattern analysis result and generates driving pattern detection result data including a behavior prediction result for a specific driver.

The user prediction processing unit 248 extracts data for identifying or estimating a specific behavioral user from the driving pattern analysis result and generates driving pattern detection result data including a user identification result.

4 is a block diagram schematically showing a hardware configuration of a driving pattern analysis device according to an embodiment of the present invention.

The driving pattern analysis device 200 according to the present embodiment includes an input unit 410, an output unit 420, a processor 430, a neural network processor 432, a memory 440, and a database 450. The driving pattern analyzing device 200 of FIG. 4 is according to an embodiment, and all blocks shown in FIG. 4 are not essential components, and some blocks included in the driving pattern analyzing device 200 in another embodiment are added. , may be changed or deleted. Meanwhile, the driving pattern analyzing device 200 may be implemented as a computing device, and each component included in the driving pattern analyzing device 200 is implemented as a separate software device or as a separate hardware device combined with software. can be implemented

The input unit 410 means a means for inputting or acquiring signals or data for performing operations such as driving pattern characteristic value generation and driving pattern analysis of the driving pattern analysis device 200 . The input unit 410 may interwork with the processor 430 to input various types of signals or data, or interwork with an external device to directly obtain data and transmit it to the processor 430 .

The input unit 410 may be implemented as a sensing unit including a driver sensing unit, a driving information acquisition unit, a vehicle information acquisition unit, and an environment information acquisition unit. The sensor may directly or indirectly obtain driver information, vehicle information, and driving information for analyzing a driving pattern.

The driver detector acquires driver identification information, and also obtains driver's touch sensing information obtained from a separately provided touch sensor or driver's image information acquired from a camera.

The driving information acquisition unit acquires driving information related to driving of the vehicle. The driving information may include speed, acceleration, angular velocity, and direction vector values according to steering wheel manipulation. The driving information may be obtained through a speed sensor, a gyro sensor, and a speed sensor installed in the driving pattern analyzer 200, obtained through a separately provided in-vehicle sensor, or originally installed in the vehicle from the ECU of the vehicle. Driving information may be obtained from the detected sensor values.

The vehicle information acquisition unit acquires information about the vehicle in motion. For example, the vehicle information acquisition unit receives information related to the vehicle, such as the vehicle identification number, vehicle model, year, cumulative distance, consumables information, fuel economy information, etc., from a stored memory or from an external communication network or from a processor installed in the vehicle. can

The environmental information acquisition unit acquires weather, road condition information, time information, location information, emergency situation information, etc. from various sensors built into the driving pattern analysis device, acquired through a separately provided in-vehicle sensor, or directly from the vehicle. You may receive information.

The output unit 420 may output or display various information such as driving pattern characteristic values, driving pattern analysis results, and final result data in conjunction with the processor 430 . The output unit 420 is implemented to include a touch screen capable of displaying various information through a display (not shown) provided in the driving pattern analysis device 200 and capable of touch sensing, and a display panel module providing images. It may be, but is not necessarily limited thereto.

The input unit 410 and the output unit 420 may be connected to an input/output interface unit (not shown). The input/output interface unit transmits information obtained from the input unit 410 to the processor 430 or receives a control signal from the processor 430 to substantially control the input unit 410 and the output unit 420. It may be arranged to convert into a signal.

The processor 430 performs a function of executing at least one instruction or program included in the memory 440 .

The processor 430 according to the present embodiment performs operations for generating driving pattern feature values and analyzing driving patterns based on collected information obtained from the input unit 410 or the database 450 .

The processor 430 may perform operations corresponding to all or part of the blocks of FIGS. 2 and 3 and may process operations using some neural networks in conjunction with the neural network processor 432 .

Meanwhile, although the processor 430 and the neural network processor 432 are described as different modules, they are not necessarily limited thereto and may be combined into one module.

The neural network processor 432 receives human factor information, driving record information, and vehicle information transmitted from the processor 430 and information such as a driving pattern score or a median score calculated by the processor. The neural network of the neural network processor 432 has an input node, an intermediate node, and an output node, and has a structure specified by a decision weight pre-learned through training data as a connection weight connecting each node. The output value of the neural network may be the final score for the user's driving pattern and vehicle driving pattern evaluated according to the existing driving and the current driving, or may be implemented in the form of a feature value matrix for the final score. .

There are many types of data acquired through sensors, and a lot of data can be obtained as time-series information. If all of these data are used as inputs to the neural network, too much noise may interfere with accurately calculating the driving pattern score. In this regard, it is preferable to perform filtering to selectively adjust the input of the neural network by filtering a lot of negative information in the case of a driver with high reliability and filtering relatively little in the case of a driver with low reliability through a separate filter unit. do.

The memory 440 includes at least one instruction or program executable by the processor 430 . The memory 440 may include commands or programs for generating driving pattern feature values and analyzing driving patterns. In addition, the memory 140 may include instructions or programs for operations for calculating neural network learning results and final results.

The database 450 refers to a general data structure implemented in the storage space (hard disk or memory) of a computer system using a database management program (DBMS), such as data search (extraction), deletion, editing, addition, etc. It means a data storage format that can freely perform, Oracle, Informix, Sybase, Relational Database Management System (RDBMS) such as DB2, Gemston, Orion ( Orion), O2, etc., and XML Native Databases, such as Excelon, Tamino, Sekaiju, etc. It can be implemented according to the purpose and has appropriate fields or elements to achieve its function.

The database 450 according to the present embodiment may store data related to generation and analysis of feature values of driving patterns, and may provide data related to driving patterns. Although the database 450 is described as being implemented within the driving pattern analysis device 200, it is not necessarily limited thereto and may be implemented as a separate data storage device.

5 is a flowchart illustrating a driving pattern analysis method according to an embodiment of the present invention.

The driving pattern analysis device 200 acquires collection information (S510). The driving pattern analysis apparatus 200 obtains at least one piece of collected information selected from a collected information group consisting of human factor information, driving record information, and vehicle information.

The driving pattern analysis device 200 generates driving pattern feature values based on the collected information (S520). The driving pattern analysis apparatus 200 generates a trip driving feature value representing a driving characteristic of each trip section using the driving record information, and uses the human factor information or vehicle information and the trip driving feature value to generate a driving pattern feature value. generate

The driving pattern analysis device 200 may provide the generated driving pattern feature values to an external device, but is not necessarily limited thereto.

The driving pattern analysis device 200 additionally performs the following operation when performing driving pattern analysis based on driving pattern feature values.

The driving pattern analysis device 200 determines whether pre-processing is performed based on preset conditions (S542). Here, the preset conditions may include a data amount of driving pattern feature values, a pre-classification setting for driving pattern analysis, and the like.

When pre-processing is performed in step S542, the driving pattern analysis apparatus 200 excludes some sub-trip driving characteristic values from among the sub-trip driving characteristic values included in the trip driving characteristic values according to a predetermined criterion and summarizes the trip driving. Characteristic values are generated, and driving pattern characteristic values adjusted using the generated trip driving characteristic values are output.

If preprocessing is not performed in step S542, the driving pattern analyzing device 200 outputs the driving pattern characteristic value generated in step S520.

The driving pattern analysis device 200 analyzes the driving pattern based on the driving pattern feature values (S540).

The driving pattern analyzing apparatus 200 may generate an analysis result for a driving pattern in a trip section by using the summarized trip driving characteristic value included in the driving pattern characteristic value.

The driving pattern analysis apparatus 200 may determine a single attribute class according to each item selected from the trip driving characteristic values, or a complex attribute class defined by considering a plurality of items selected from the trip driving characteristic values.

The driving pattern analysis device 200 provides driving pattern detection result data including all or part of the driving pattern analysis result to an external device (S550).

In FIG. 6, it is described that each step is sequentially executed, but is not necessarily limited thereto. In other words, since it will be applicable to changing and executing the steps described in FIG. 6 or executing one or more steps in parallel, FIG. 6 is not limited to a time-series sequence.

The driving pattern analysis method according to the present embodiment described in FIG. 6 may be implemented as an application (or program) and recorded on a recording medium readable by a terminal device (or computer). All types of recording devices in which applications (or programs) for implementing the driving pattern analysis method according to the present embodiment are recorded and data readable by a computing system are stored in a recording medium readable by a terminal device (or computer). or medium.

6 is a diagram for explaining a feature value generation operation according to an embodiment of the present invention.

The method of analyzing driving patterns by generating driving pattern feature values in the driving pattern analysis device 200 is to perform automatic analysis through comparison between detailed numerical values (behavioral data) for each driving habit configuration item, and to apply a code + numerical expression evaluation method. can

The driving pattern analyzer 200 may use a method of 'evaluation of the degree of behavior' through direct comparison between the measured change values (acceleration, inclination, etc.) of the main event.

In addition, the cumulative value of each item of 'evaluation by trip (driving record from engine on to off)' can be coded as 'driving behavior evaluation' by comparing and evaluating the cumulative value of each item with situations such as occurrence intervals.

In addition, the driving pattern analysis apparatus 200 can also 'compare and express behavior levels' compared with average driving data of feature groups within 'human factors/service use behavior/driving behavior' of the evaluation system. For example, the driving pattern analysis device 200 groups users who have a record of using a two-wheeled vehicle among males aged 32 to 34 in the Seocho A and Gangnam A areas, have a risk behavior frequency of 60% or more, and a degree of fixation of 3/5 or more in driving behavior evaluation. And it is possible to create a sequence list based on grouped items.

In addition, the driving pattern analyzing apparatus 200 may generate a driving pattern feature value by adding a 'fixed' value for each item to the driving behavior evaluation for driving more than N times (where N is a natural number equal to or greater than 2).

7 is an exemplary view for explaining driving pattern feature values according to an embodiment of the present invention.

Referring to FIG. 7 , driving pattern feature values represent feature values for a trip section.

The driving pattern feature values are composed of trip driving feature values representing driving characteristics within the trip section.

The trip driving feature values are configured in the form of connecting trip driving feature values representing driving characteristics of each of a plurality of sub-trip sections.

Hereinafter, codes for generating driving pattern feature values will be described.

The driving pattern feature value (User.code) may be generated using a human factor code (Code _{human factor} ), a driving trip code (Code _Trip ), and a vehicle factor code (Code _{vehicle factor} ). Driving pattern feature values may be defined as in [Equation 1]. Here, the weight can be applied if necessary, and can be reflected as weight or addition or subtraction according to the evaluation criteria.

Hereinafter, the human factor code (Code _{Human Factors} ) will be explained. The human factor code may be defined as in [Equation 2].

(SE = gender code (because gender does not change, it is excluded from item comparison evaluation after one measurement), AG = age code, OP = driving period code)

In the human factor code, information included in feature value creation may be changed according to whether the user agrees to the information. On the other hand, if the user does not agree to the disclosure of all information, human factor information may not exist. In this case, the human factor code is excluded from the driving pattern feature value, or information obtained through user sensing in the vehicle is used. A human factor code may be generated using the user prediction identification information predicted.

Hereinafter, the driving trip code (Code _Trip ) will be described. The driving trip code may be defined as in [Equation 3].

(DR = driving record (evaluation of factors affecting risky driving such as mileage/time by road type), DB x Level = risky driving behavior (sudden acceleration, sudden braking, etc.) x risk assessment, EF x Level = environmental factor (weather information) x risk assessment)

For example, DR can be expressed as EW.105.70.8_DT.11.30.4 - LT, 'highway code/mileage/time traveled (minutes)/average speed range level 8/country road (city) code/driving Distance / Driving time (minutes) / Average speed Range level 4 / Driving time - Lunch'.

For example, DB x Level can be expressed as HA3 ^S .6.8 ² -- 2.5, 'Rapid acceleration code/level 3 within 30 minutes immediately after driving, level 6 within 30 minutes, level 8 twice in a row/ Level 2/30 minutes after 1 hour means level 5'.

In addition, DB x Level can be expressed as HD.5 -- 2.4 ² , meaning 'emergency braking code/level 5 within 30 minutes after driving/level 2 after 1 hour/level 4 within 30 minutes twice in a row' .

For example, EF x Level can be expressed as RF8, which means 'rainfall code/rainfall level level 8'.

Here, Trip ₁ means a level evaluation according to a pre-determined standard, and Trip ₂ means a level evaluation according to a pre-determined criterion and an item-by-item comparative evaluation with Trip ₁ .

Hereinafter, the vehicle factor code (Code _{vehicle factor} ) will be described. The vehicle factor code may be defined as in [Equation 4].

(VMY x Level = Vehicle year code x Year level, TA x Level = Tire wear code x Wear condition level, AM x Level = Accumulated mileage x Accumulated mileage level)

8 is an exemplary diagram illustrating an analysis result of driving pattern feature values according to an embodiment of the present invention.

The driving pattern feature value is generated through a predetermined rule that directly expresses numerical changes and degree values for each evaluation action item, rather than an arithmetic total score or average formula.

It is desirable to refer to standard values such as 'Road Traffic Safety Authority' for the criteria for risky behavior in driving pattern feature values, but the range can be adjusted and set by the system operator according to data accumulation. In addition, the degree of the action can be evaluated and expressed through the list of level values according to the detailed measurement values in the risky action included in the driving pattern feature value. For example, the risk action item for the trip-by-trip evaluation of driving pattern feature values can be expressed as 'HA3 ^S .6.8 ² - 2', and 'Record of rapid acceleration: Level 3 immediately after driving starts, 1 action / 30 Level 6 within minutes, 1 action/Level 8 within 30 minutes, 2 consecutive actions/Level 2 after 1 hour, 1 action'.

The evaluation of driving pattern feature values for each trip may include independent evaluation items, comparison evaluation items, connection evaluation items, and the like, as shown in [Table 1].

In the evaluation for each trip, previous values and 'automatic comparison evaluation items' (sudden change event_sudden acceleration/deceleration, sharp turn, etc.) and only the corresponding result 'independent evaluation items' (Trip mileage/time_18km/1h = low-speed driving: city center) 80% pass the congestion section).

Driving behavior evaluation may update driving pattern feature values after secondary evaluation by applying items requiring connection evaluation, vehicle factors, and environmental factors among evaluation items for each trip.

In addition, for the intuitive understanding of the user of the evaluation system, as shown in FIG. 8, a substitution-expression method may be applied, and a tag substitution + numerical value or degree value exposure method may be applied.

9 is an exemplary diagram for explaining a repetitive behavior detection operation of the driving pattern analysis device according to an embodiment of the present invention.

The driving pattern analysis device 200 performs a repetitive behavior detection step of extracting data on the driver's repetitive behavior from the driving pattern analysis result.

The driving pattern analysis device 200 may read and identify a 'repetitive' behavior in order to determine to what extent and how often a person performs the corresponding behavior.

For example, the driving pattern analysis device 200 sets A (coded behavior record_fixed numerical value and occurrence pattern) + B (location, time information _GPS log) + C (environmental information such as weather) within Trip, action items Detect the repetitive tendency of star ratings and record them separately.

Thereafter, the driving pattern analysis device 200 extracts repetitive situations from detailed driving records of a paid transport driver and a general driver with multiple usage records. The driving pattern analysis device 200 may compare the driving pattern information with the driver profile information and read whether the driving pattern is repetitive or habitual behavior due to inexperienced driving.

The driving pattern analysis device 200 can classify and manage risk factors, such as guiding 'restrictions on rental' according to the range of risky behaviors that occur thereafter, or post-measurement of compliance after 'restrictions on driving methods in a specific section' do.

For example, 1) When a repetitive feature is read with only A: A specific single action or a numerical value and occurrence cycle between specific actions are repeated and patterned.

[Level 6 sudden start when turning left, continuous behavior of sudden turn at level 5, 87% frequency per trip, 93% frequency when left turn]

On the other hand, 2) When the repetition feature is read as A + B: The numerical value and occurrence pattern of a specific action repeatedly occur at a specific place and at a specific time.

[Gangnam Station Bangbang Intersection, between 9:20 and 10:05 am, left turn (8 out of 10 times) + Rapid departure 6. Rapid turn 5], [Seocho Station Intersection, between 9:54 and 10:16 am, left turn ( 9 out of 10) + quick start 6. quick turn 5]

Meanwhile, when repeating features are read as A + B + C: Among the repeated features read as 1) and 2), refinement of feature reading is performed by evaluating correlations with external variables such as weather.

[Repeat a specific driving behavior at a specific location only on rainy days (regardless of the rainfall range)]

10 is an exemplary diagram for explaining a behavior prediction processing operation of a driving pattern analysis device according to an embodiment of the present invention.

The driving pattern analysis device 200 performs a behavior prediction processing step of extracting data on driver's behavior prediction from the driving pattern analysis result.

The driving pattern analysis device 200 predicts whether the driving pattern is a general behavior of a 21-year-old man, whether user A's rapid acceleration-level 7 or higher behavior will increase/repeat/decrease in the future after 3 months.

The driving pattern analysis device 200 codes and records the behavioral values and behavioral tendencies of a specific range for each behavioral item, so during machine learning for prediction, it is easy to compare and learn relationships with data outside the range, improving the efficiency of the prediction model It is possible.

The driving pattern analysis device 200 can predict the 'probability of occurrence of a specific behavior-risk value' of a 'specific user group' by adding additional information such as 'driver profile and environment information' to the recorded behavior.

The driving pattern analysis device 200 reflects the 'pattern and correlation weight of similar behavior users' to predict the 'probability of occurrence and expected numerical value' in order to predict the future risk of 'specific user' exhibiting risky behavior.

The driving pattern analysis device 200 analyzes the association between behaviors and codes the repetitive relationship pattern derivation method to read the trend of behavior standards even when 'merging heterogeneous data' other than driving information, so it is easy to infer the possibility of occurrence of the same behavior in the future. one result can be drawn.

11 is an exemplary diagram for explaining a user prediction processing operation of the driving pattern analysis device according to an embodiment of the present invention.

The driving pattern analysis device 200 performs a user prediction processing step of extracting data for identifying or estimating a specific behavioral user from the driving pattern analysis result.

The driving pattern analysis device 200 may identify and predict a 'specific behavioral user' based on unique behavioral characteristics.

For example, the driving pattern analysis device 200 selects a user of a specific action, such as 'August 3 at 9:00 am, passing through Seocho Station intersection with rapid acceleration, and a driver who caused an accident between other vehicles' at that time, among a plurality of users. It is possible to extract users with the highest frequency of 'sudden acceleration' record and range rapid acceleration pass behavior.

In addition, the driving pattern analysis device 200 determines whether a user who has performed a 'specific dangerous behavior (repeated 10 or more times. dangerous)' in service A is in 'recent user usage records' in service B, even based on 'non-identified user information'. + 'Behavioral similarity analysis' is possible.

Also, the driving pattern analysis device 200 may specify a user group or a user exhibiting a specific behavior pattern.

Example 1) Presumed threat driving user group = (within 50m distance between vehicles) advanced data (after rapid acceleration, sudden stop) A group of users measured with a frequency of 80% or more for each trip of 3 or more consecutive trips => 'User group estimated to threaten driving Can be classified according to the condition to be specified, such as listing the occurrence frequency of 'by road type' such as 'while driving on a highway' in 'Driving on the highway' = Case of performing the relevant act under a specific condition or case of performing the relevant act regardless of the condition + condition of human factors (34 years old) , Seoul. Living in Gangnam, male, etc.) + Intersection of information with point of view, location information, etc. Separate extraction of specific conditions is possible

Example 2) Identification of a user presumed to be driving in Seocho A area in August 2022, among 2W vehicles, carrying out transportation for a fee => Daily first driving time and last driving end time, no movement within 3 to 15 minutes Subjects can be selected and read based on the location of the viewpoint, the position of each viewpoint when switched to parking mode, the time required for each movement, and whether or not the movement route overlaps.

12 to 14 are diagrams for explaining a driving pattern analysis operation based on a neural network according to an embodiment of the present invention.

The driving pattern analysis device 200 may perform driving pattern feature value preprocessing and driving pattern analysis using the neural networks 1120 and 1130 . Here, it is preferable to use a convolutional neural network (CNN) as the neural network, but it is not necessarily limited thereto.

The driving pattern analysis device 200 applies the trip driving feature values to the first neural network 1120, which has been pre-learned to summarize the driving pattern through the first group of training data, to obtain the summarized trip driving feature values. can create

The driving pattern analysis device 200 provides a driver profile specified from the summarized trip driving feature values, vehicle history information and human factor information, or a predictive driver profile generated to specify a driver based on the trip driving feature values. In order to determine a driving pattern through training data of a second group that is at least partially different from the first group, it may be applied to the second neural network 1130 in which learning has been performed in advance to generate an analysis result for the driving pattern.

15 is a reference diagram illustrating factors related to a user's service use and driving disposition evaluation according to an embodiment of the present invention.

User behavior evaluation of services that require safe use must be accompanied by subdivided measurement and expression of the meaning-degree of items in order to predict and manage risk factors. In this embodiment, the goal is to evaluate according to the 'accident record analysis' method of the traffic accident analysis system based on the analysis of human factors/vehicle factors/environmental factors + driving record information. Aggregate evaluation based on continuous/accumulative use is necessary, but simple cumulative evaluation of each occurrence may result in dilution of habit and temperament points that require actual attention. For example, if a speeding event occurs 1 time out of 3 driving records in the same section, 2 non-speeding times pass through the congestion time zone, and 1 speeding time is driving at a high risk value at 03:00 in the middle of the night. The 'speeding temperament degree' according to can be highly evaluated. Depending on the user usage cycle of the domain to be measured, the detailed rules of the cumulative evaluation formula and temperament evaluation + dilution/attenuation formula of measurement items may be changed.

Referring to FIG. 15 , human factors constitute user information about a user who is a driver. The score for the temperament of the user may be calculated by adding up a past cumulative record and a current record obtained during current driving according to a weight. Human factors vary depending on the driver, so an accurate basic identification of the driver is necessary.

16A, 16B, and 16C are tables showing measurement items for evaluating a user's service use and driving disposition according to an embodiment of the present invention. For example, the user profile (evaluation of the quality and degree of risk factor disposition items) can be calculated by using Trip Record evaluation (applying absolute evaluation method for each item) + Usage history evaluation (Grade + Subtractive Score). The current driver profile value may be calculated by summing the trip records and usage history evaluation information according to weights or using long-term and short-term moving averages for each.

Risk factors include information on patterns that may be related to traffic accidents, such as sudden acceleration patterns, sudden deceleration patterns, speeding patterns, lane-changing cycles, excessive cut-in patterns, and distance from the vehicle in front. Through these items, insurance premium calculation, ranking for operator user management, and standard scoring are not performed, and propensity and temperament indicators for each user can be provided for distribution and management and response by category.

17 is an example of evaluating a user's service use and driving disposition according to an embodiment of the present invention. Referring to FIG. 17 , the following evaluation method may be performed in relation to user A evaluation.

1. Trip-Trip record-based driving evaluation by usage unit such as booking, absolute evaluation method: 'Traveling distance, driving time', 'actual driving distance and distance of one trip, driving date and time zone', 'driving record using acceleration information Assessing the degree of '

2. (Service)Usage history evaluation is based on the relative evaluation method with the service user record: 'tendency comparison with previous records', 'comparative evaluation before/after use', and 'service rule-based evaluation'.

Here, the user temperament evaluation is 1 speeding, attention driving, sudden change needle... , 2 VOCs of the next user related to repeated cancellation of imminent time, repeated delayed returns, and returns… Temperament evaluation can be performed using index values of detailed items such as

Regarding Driving Behavior, user A - Trip score sum of Booking ID + all connected evaluation items can be reflected.

Specifically, Trip 1 Score may reflect 00.00 points <-> Trip 2 Score: connection evaluation items among Trip 1 items.

- Total mileage/time: date/time based travel distance/average speed exposure. scoring

- Idle time. Section : Idle time. Position exposure. scoring

- Highest/lowest speed: Exposure of corresponding speed information. Scoring by time zone/location

- Driving path: Route, road, enforcement camera pass, speed limit section exposure

- Sudden acceleration, rapid deceleration: scoring based on event time, location, acceleration change value

- Rapid right/left turn/U-turn: scoring based on event time, location, acceleration change value

- Speeding: Scoring based on maximum speed, duration, and time

- Amount of fuel: Amount of fuel at start/end of engine, fuel consumption standard score

- Mobile phone manipulation: Scoring based on the number of times allowed (measurement plan in progress)

- Accident/shock detection: EDR inactive memory information or impact degree/direction detection

- Proximity detection: Used when measurement is possible. Scoring each number of measurements based on the near degree value

18 illustrates a concept related to additional information of trip data according to an embodiment of the present invention. Trip data Additional measurement desired information is as follows.

- Transmitted to the server whenever the vehicle engine is turned ON ~ OFF

- Trip data Snapshot time point record, multiple/simultaneous items are recorded at the same time

- Engine oil temperature / RPM / vehicle speed / Fuel Cut status

- Gear position (auto vehicles only)

- Transmission oil temperature (auto vehicles only)

- Total distance traveled by vehicle: Total distance traveled after device power is applied

- Top speed: based on trip

- Highest RPM: Based on Trip

- Average speed: based on trip

- Travel distance: based on trip

- Operation time: based on trip

- Number of rapid accelerations: Number of accelerations of 8 km/h or more per second (Activated at 5 km/h or more? - Start measuring acceleration values?)

- Number of rapid decelerations: Number of decelerations of 14km/h or more per second (same as above)

- 60~70 / 70~80 / 80~90 / 90~100 / 100~110 / Accumulated time driving at speeds above 110: Accumulated time when maintaining speed within each section

- Idle rotation time: the time when the speed is 0 state after detecting the speed over 0

- Warm-up time: Time after ignition is turned on until the speed reaches 0 or higher

-Fuel economy index: distance traveled by consumed oil (comparison of wheel revolutions, etc., simple calculation rather than measurement?)

- Fuel level / EV battery level

- Reserved: Reserved? Y/N?

- Total fuel consumption: Total fuel consumption after device power is applied

- Fuel Consumption: Based on Trip

- Reservation mileage: mileage based on the current borrower (final)

- Reserved fuel consumption: Same as above

- Whether or not the EV charger is detachable: whether it is fast/slow and connected

- Vehicle fuel type: 0-gasoline, 1-LPG, 2-diesel, 3-hybrid, 4-electric vehicle

- Number of gears: gear position, record of change time?

- Key status / brake operation status / electric car driving distance / GPS-based reservation user driving distance / remaining fuel / minimum voltage

- Sudden start/sudden stop/sudden left/right turn/sudden U-turn/sudden course change/sudden overtaking frequency: Is it possible to obtain variable acceleration values and displacement values based on true north rotation?

- Diagnostic code: Can it be converted to manufacturer code?

- Time / location information (latitude, longitude)

- Number of GPS receiving satellites / GPS Hdop (horizontal position error correction) / Vehicle speed based on GPS

19 is a flowchart illustrating a method of evaluating a user driving pattern according to an embodiment of the present invention. The user driving pattern evaluation method shown in FIG. 19 includes the following steps performed time-sequentially by the driving pattern analyzing apparatus 200 . This method is a method of evaluating driving patterns based on driver information.

In step 110, the driver detecting unit detects the driver, and the processor reads a profile according to the detected driver from the memory.

In step 120, the driving information acquisition unit acquires driving information related to the current driving, and the processor checks detailed driving information about the current driving through the acquired driving information. The driving information may include, for example, current vehicle location, destination/originating location information, driving route, driving speed, acceleration, lane change, distance from a vehicle in front, steering wheel operation information, and the like.

In step 130, the sensor acquires user-related information, vehicle information, and environment information obtained during driving of the vehicle, and transmits the acquired information to the processor.

In step 140, the processor reads the existing history information and updates the driver profile and usage history information using the information obtained in step 130 in real time.

In step 150, the processor evaluates the current driving pattern of the user by using the updated driver profile and usage history information.

20 is a flowchart illustrating a method of evaluating a driving pattern of a vehicle according to an embodiment of the present invention. The driving pattern evaluation method of a vehicle shown in FIG. 20 includes the following steps performed time-sequentially by a processor. This method is a method of evaluating the driving pattern from the viewpoint of the vehicle rather than the driver.

In step 210, the processor reads driver profile and vehicle profile information from memory. More specifically, the driver detection unit detects the driver, the vehicle information acquisition unit specifies vehicle information, and the processor reads the detected driver profile and vehicle profile from the memory.

In step 220, the driving information acquisition unit acquires driving information related to the current driving, and the processor checks detailed driving information about the current driving through the acquired driving information. The driving information may include, for example, current vehicle location, destination/originating location information, driving route, driving speed, acceleration, lane change, distance from a vehicle in front, steering wheel operation information, and the like.

In step 230, the processor updates the driving record by receiving the driving information acquired through the driving information acquisition unit.

In step 240, the processor updates the vehicle profile information using the driving record.

In step 250, the processor evaluates the driving pattern of the vehicle according to the current vehicle usage using the updated vehicle profile.

21 is a flowchart illustrating a method of evaluating a driving pattern of a vehicle according to an embodiment of the present invention. The driving pattern evaluation method of a vehicle shown in FIG. 21 includes the following steps performed time-sequentially by a processor. This method is a method for evaluating the driving pattern from the viewpoints of the driver and the vehicle.

In step 310, the processor reads driver profile and vehicle profile information from memory. More specifically, the driver detection unit detects the driver, the vehicle information acquisition unit specifies vehicle information, and the processor reads the detected driver profile and vehicle profile from the memory.

In step 320, the sensor acquires action information based on the user's action.

In step 330, the sensing unit obtains driving information based on actual driving information of the vehicle.

In step 340, the processor updates the driver profile, usage history, driving record and vehicle profile, respectively.

In step 350, the processor comprehensively evaluates the driving pattern of the user using the updated stomach value.

According to another embodiment of the present invention, the driving pattern analysis device of the present invention may perform the following operations.

[Driving pattern analysis method based on driver profile and usage history (User profile, Usage history)]

When analyzing a specific item, it is possible to perform analysis on a specific item by considering values of items highly related to the specific item, rather than simply analyzing the specific item using only the value of the specific item.

For example, in the case of analyzing the “rapid acceleration” item, the “rapid acceleration” item is not analyzed simply by considering the number of “rapid acceleration” items, but rather, the item highly related to the “rapid acceleration” item, “location of rapid acceleration” ” item, “Weather corresponding to the location where rapid acceleration occurred” item, “Speed for a certain period (last 6 months) at the location where sudden acceleration occurred” item, “Congestion corresponding to the location where rapid acceleration occurred It is possible to perform an analysis on the “Rapid Acceleration” item by considering the value of the “Situation” item.

In other words, the driver who frequently “rapidly accelerates” in a specific location does not actively accelerate by his or her own will, but is forced to accelerate by external factors such as congestion on the road or heavy rain. Since acceleration may not occur, the analysis of the “rapid acceleration” item may be performed by considering the value of the item highly related to the “rapid acceleration” item.

[Driving pattern analysis method based on driving records and vehicle profiles (Trip records, VIN profile)]

When analyzing a specific item, the criterion for a specific item may be varied by considering the value of an item highly related to the specific item, rather than analyzing the specific item using a fixed criterion.

For example, when analyzing the “speeding” item, the “speeding” item is not analyzed using a fixed reference speed, but the value of the “weather” item, which is an item highly related to the “speeding” item. The reference speed for can be varied.

That is, in the case of "sunny weather", the standard speed for the "speeding" item uses the "maximum speed" of the corresponding road, and in the case of "rainy weather", the standard speed for the "speeding" item uses the "maximum speed" of the corresponding road. / 2" is available. Of course, even in the case of "rainy weather", the reference speed for the "overspeed" item may be changed according to the amount of precipitation.

According to another embodiment of the present invention, it is also possible to determine a driving pattern according to the following concept.

Driver profile: includes user basic information including user identification information, economic credit information, driving ability information, driving risk information, driving reliability information, driving concentration information, and compliance information. The driver profile includes evaluation information on the driver's temperament, which can be updated based on existing driver profile information, user history information, and driving information according to current driving. In calculating insurance premiums, driving reliability information, driving concentration information, driving risk information, driving ability information, and norm compliance information can be very useful.

User history: includes basic user history information including usage period and distance, reservation fulfillment information, time observance information, and unauthorized delay/cancellation information. This is objective information rather than a driver profile, and is difficult to view as individual evaluation information, and is more like event record information.

Driving reliability information is not only meaningful as reliability itself, but can be used at least in part to calculate accident risk information and norm compliance information. The driving reliability information is for determining how reliable the driver is, and is basically calculated as a low reliability index in the case of a driver with a history of sudden braking, rapid acceleration, frequent lane changes, and frequent accidents. In addition, the driving reliability information can be divided into long-term driving reliability information and short-term driving reliability information. Long-term driving reliability information can be used to calculate insurance premiums, and short-term driving reliability information can be used important in post-analysis of a specific event. For example, when a traffic accident occurs, even if a person has high driving reliability in the long term, there is a case where the driving reliability is low due to a lack of attention on the day. In this case, if the decision is made in consideration of only the long-term driving reliability information, an error may occur in which the driver's behavior is judged to be highly reliable even on the day of the accident. The driving reliability of can be more accurately calculated. Driving reliability is, for example, whether you drive well in the center of the lane, whether you turn on and change the blinker in advance when changing lanes, whether you drive while maintaining a safe distance between the cars in front, and whether you make left/right turns without overdoing it. It is an index that reflects factors necessary for safe driving, such as

Even if a situation arises in which the accident risk index needs to be increased as a driver makes a sharp acceleration or a sharp turn one day, the driver's long-term driving reliability information and short-term driving reliability information are high, and the rate of increase in the accident risk index can be handled relatively low.

In addition, a driver made a lot of sudden acceleration and sharp turns one day, and in the case of the driving section, this information, generated by driving in an area where rapid acceleration and sharp turns frequently occur as a bottleneck section, increases the driver's driving reliability. It will be used to lower and increase the accident risk index, but at least the reflection ratio can be kept low considering the specificity of the driving section.

Driving information: information related to actual driving of the vehicle acquired through a sensor in the vehicle or a sensor inside the present module. It includes basic driving information, emergency operation (sudden acceleration, sudden braking, sharp turn) information, lane change information, driving route information, and driving time information.

Environmental information: Information related to the external environment during driving of the vehicle, including route information, weather information, temperature information, altitude information, road condition information, and road surface contact information.

The above description is only illustrative of the technical idea of the embodiment of the present invention, and those skilled in the art to which the embodiment of the present invention pertains may make various modifications and modifications within the scope not departing from the essential characteristics of the embodiment of the present invention. transformation will be possible. Therefore, the embodiments of the present invention are not intended to limit the technical idea of the embodiment of the present invention, but to explain, and the scope of the technical idea of the embodiment of the present invention is not limited by these examples. The protection scope of the embodiments of the present invention should be construed according to the claims below, and all technical ideas within the equivalent range should be construed as being included in the scope of the embodiments of the present invention.

10: driving pattern analysis system
100: information collection device 200: driving pattern analysis device
210: collection information acquisition unit 220: feature value generation unit
222: pre-processing unit 230: driving pattern analysis unit
240: result data processing unit

Claims (21)

A computing device for analyzing a driving pattern of a vehicle, the computing device comprising: at least one processor; It has a non-transitory storage medium and includes at least one memory for storing program codes executed by the processor, wherein the at least one processor according to the execution of the program codes,
obtaining at least one piece of collected information selected from a collected information group consisting of human factor information, driving record information, and vehicle information;
generating a trip driving feature value representing a driving characteristic of each trip section by using the driving record information; and
Generating a driving pattern feature value using the human factor information or the vehicle information and the trip driving feature value
Characterized in that performing an operation comprising a, computing device. According to claim 1,
Further comprising obtaining vehicle history information using the vehicle information,
The step of generating the trip driving feature value,
Select from the group consisting of speed, travel distance, travel time, acceleration, number of accelerations, engine oil temperature, gear position, transmission oil temperature, trip data snapshot, RPM, number of constant speed driving, idling, fuel economy, and warm-up time in the trip section. Generating the trip driving feature value for the driving characteristic according to one or more items that are
The generating of driving pattern feature values may include generating driving pattern feature values using the vehicle history information and the trip driving feature values. According to claim 2,
The trip driving feature value,
The computing device characterized in that it includes a trip driving characteristic value for each single item according to the selected one item or a trip driving characteristic value for each plurality of items according to the selected plurality of items. According to claim 2,
The step of generating the driving pattern feature value,
the vehicle history information;
the trip driving characteristic value; and
The driving pattern feature value is generated by further using a driver profile specified from the human factor information or a predictive driver profile generated to specify a driver based on the trip driving feature value. According to claim 2,
summarizing the trip driving characteristic values; and
generating an analysis result for a driving pattern in the trip section using the summarized trip driving feature values;
Further comprising a, computing device. According to claim 5,
The step of summarizing the trip driving feature values,
generating the summarized trip driving feature values by excluding some sub-trip driving feature values from among the sub-trip driving feature values included in the trip driving feature values according to a predetermined criterion;
The generating of the analysis result may include determining a single attribute class according to each of the selected items or determining a composite attribute class defined by considering a plurality of the selected items. According to claim 1,
Further comprising generating sub-trip driving feature values representing driving characteristics of a sub-trip section shorter than the trip section;
The computing device, characterized in that the trip driving feature value is obtained by connecting the generated sub-trip driving feature values obtained in the trip section. According to claim 7,
The sub-trip driving feature value is expressed as a sub-trip driving code in which driving characteristics in a given sub-trip section are coded, and the trip driving feature value is expressed as a trip driving code obtained by connecting the sub-trip driving codes;
Characterized in that, coding the driving characteristics in the given sub-trip driving section is an operation of expressing the driving characteristics with codes corresponding to the number of occurrences of the driving characteristics or the level of the driving characteristics. According to claim 7,
The sub-trip driving feature value is a sub-trip feature vector for driving characteristics in a given sub-trip section;
The computing device, characterized in that the trip driving feature value is a feature vector obtained by extending a dimension of the sub-trip feature vector through a process of combining the sub-trip feature vectors. According to claim 9,
Applying the trip driving feature values to a first neural network pre-learned to summarize driving patterns through a first group of training data to generate summarized trip driving feature values;
The computing device further comprising generating an analysis result for a driving pattern in the trip section using the summarized trip driving feature values. According to claim 9,
the trip driving characteristic values summarized above;
the vehicle history information; and
A driver profile specified from the human factor information or a predictive driver profile generated to specify a driver based on the trip driving feature value is a driving pattern through training data of a second group that is at least partially different from the first group. The computing device further comprising generating an analysis result for a driving pattern by applying the second neural network to a second neural network on which learning has been performed in advance to determine . According to claim 7,
The computing device may further include a result data processing step of providing driving pattern detection result data including all or part of the analysis result to an external device. According to claim 12,
The result data processing step,
The computing device characterized in that the driving pattern detection result data is generated by filtering the analysis result to correspond to a request of the external device or a predetermined data form. According to claim 12,
The result data processing step,
and performing a repetitive behavior detection step of extracting data on the driver's repetitive behavior from the analysis result to generate the driving pattern detection result data including a repetitive behavior identification result. According to claim 12,
The result data processing step,
The computing device characterized in that performing a behavior prediction processing step of extracting data on the driver's behavior prediction from the analysis result to generate the driving pattern detection result data including the derived behavior prediction result for the specific driver. According to claim 12,
The result data processing step,
and generating the driving pattern detection result data including a user identification result by performing a user prediction processing step of extracting data for identifying or estimating a specific behavioral user from the analysis result. at least one processor; A driving pattern analysis method performed by a computing device having a non-transitory storage medium and including at least one memory storing program codes executed by the processor, the method comprising:
The computing device,
obtaining at least one piece of collected information selected from a collected information group consisting of human factor information, driving record information, and vehicle information;
generating a trip driving feature value representing a driving characteristic of each trip section by using the driving record information; and
Generating a driving pattern feature value using the human factor information or the vehicle information and the trip driving feature value
Characterized in that performing, driving pattern analysis method. According to claim 17,
Further comprising obtaining vehicle history information using the vehicle information,
The step of generating the trip driving feature value,
Select from the group consisting of speed, travel distance, travel time, acceleration, number of accelerations, engine oil temperature, gear position, transmission oil temperature, trip data snapshot, RPM, number of constant speed driving, idling, fuel economy, and warm-up time in the trip section. Generating the trip driving feature value for the driving characteristic according to one or more items that are
The generating of the driving pattern feature values may include generating driving pattern feature values using the vehicle history information and the trip driving feature values. According to claim 18,
The step of generating the driving pattern feature value,
the vehicle history information;
the trip driving characteristic value; and
Characterized in that the driving pattern feature value is generated by additionally using a driver profile specified from the human factor information or a predictive driver profile generated to specify a driver based on the trip driving feature value, method. According to claim 18,
summarizing the trip driving characteristic values; and
generating an analysis result for a driving pattern in the trip section using the summarized trip driving feature values;
Further comprising a, driving pattern analysis method. According to claim 20,
The step of summarizing the trip driving feature values,
generating the summarized trip driving feature values by excluding some sub-trip driving feature values from among the sub-trip driving feature values included in the trip driving feature values according to a predetermined criterion;
The generating of the analysis result may include determining a single attribute class according to each of the selected items or determining a composite attribute class defined by considering a plurality of the selected items.

Images