KR20140097669A - Apparatus and method for determining insurance premium based on driving pattern recognition - Google Patents

Abstract

Disclosed are an apparatus and method for determining an insurance premium based on driving pattern recognition. The method for determining the insurance premium includes the steps of: receiving driving information collected by using an acceleration sensor included in a vehicle; recognizing a driving pattern about an accident risk, among the driving habits of a vehicle driver, from the inputted driving information; and calculating the insurance premium in proportion to the accident risk rate of the driver based on the recognized driving pattern.

Description

Technical Field [0001] The present invention relates to a device and a method for determining a premium based on driving pattern recognition,

The present invention relates to a technology for determining insurance premiums, and more particularly to a technology for determining insurance premiums by collecting sensed data as a result of driving a vehicle using various sensors provided in the vehicle and recognizing a driver's operation pattern therefrom, And a recording medium on which the method is recorded.

Smart services in the automobile space are increasingly expanding from the application of telematics technology to automobiles. In recent years, automobiles are equipped with systems that can detect a vehicle accident or theft, a parking position, a remote system, or a lane departure. As the performance of the automobile becomes more advanced, interest in the stability of the driver is increased. Accordingly, studies are being conducted to recognize the behavior pattern of the driver in order to improve the stability.

Recently, smart phones, which are a kind of mobile devices using various kinds of sensors together with excellent portability, have an advantage that they can be applied to various fields, and active research using them is under way. Particularly, an acceleration sensor provided in a cellular phone is useful for recognizing an action or a state of a moving object. Techniques such as gait recognition or gesture recognition using an acceleration sensor have been proposed, and studies are being conducted to implement a management system for safe operation using an acceleration sensor or a GPS module.

In the prior art cited below, there is proposed a technique for obtaining information on the accident situation of a vehicle using an acceleration sensor and analyzing the cause of the accident using the information. However, the technology which is known so far is mostly a simple approach to acquire and utilize the driving information by utilizing the vehicle and the sensors added thereto, and it is at a level that is utilized in the field directly related to the driving of the vehicle.

Meanwhile, in connection with the calculation of the conventional automobile insurance premium, the automobile insurance premium is generally calculated based on the conventional indicators such as driver's driving experience, accident history, and vehicle value. However, there is a problem that the inherent driving habit of each driver and the risk factors inherent in such driving habit are difficult to grasp from a simple accident history. Therefore, there is a need to refine the information acquired through various sensors according to the driving of the vehicle in a highly advanced manner, and to objectify the potential risk factors inherent in the driving habits of the driver, However, there was still no technical means by which car insurers could identify the risk factors objectively indexed from the driver's actual driving habits.

Korean Published Patent No. 2011-0003449, AJIN INDUSTRIAL CO., LTD., Jan. 12, 2011 disclosed.

SUMMARY OF THE INVENTION The present invention has been made in view of the above problems, and it is an object of the present invention to overcome the limitation that the conventional automobile insurance premium calculation method can not grasp the actual driving habits of the driver alone and is based only on the accident history or driving experience, We want to solve the problem that there is no technical means to compare the driving habits of the insured and the normal driver who run the driving section to the insurance premium calculation.

According to an aspect of the present invention, there is provided a method of determining premiums, the method comprising: receiving driving information collected using an acceleration sensor provided in a vehicle; Recognizing an operation pattern related to an accident risk in the driving habit of the driver of the vehicle from the input driving information; And calculating an insurance premium proportional to an accident risk rate of the driver based on the recognized operation pattern.

The step of recognizing the operation pattern may include: preprocessing data by removing fluctuations and discontinuities caused by operation of the acceleration sensor and noise from the inputted travel information; Detecting an event interval in which a pattern occurs among the preprocessed data using end point detection; Resampling the data for each sample point at a predetermined time interval with respect to the detected event period; And detecting a driving pattern by measuring the similarity of the sequence between the feature template and the resampled data stored in advance for the driving section and storing the corresponding section and the time information together. The detecting of the event interval may include detecting only the energy value of the preprocessed data, temporarily storing a point having an energy value equal to or greater than the first threshold value, and then, when the energy value reaches or exceeds a second threshold value If the energy value reaches the first threshold value or less within a predetermined time after the energy value falls below the second threshold value, Can be determined as the end point of the event section.

In the insurance premium determination method according to an embodiment, the step of calculating the insurance premium may include comparing an operation pattern according to the security level reference for each operation section collected in advance and the recognized operation pattern of the driver, Calculating a risk weight value proportional to an accident risk rate of the driver from the comparison result; And summing the risk weights to premiums calculated according to the contract conditions of the driver's insurance of the driver.

According to another aspect of the present invention, there is provided a method of determining premiums, the method comprising: receiving driving information collected using an acceleration sensor provided in a vehicle; Recognizing an operation pattern related to an accident risk in the driving habit of the driver of the vehicle from the input driving information; Matching the corresponding section information and the time information to the recognized operation pattern and displaying the same on the display means; Receiving a reflection rate for each section considering a visually recognized driving situation from the user through the display means; And calculating an insurance premium proportional to an accident risk rate of the driver on the basis of the recognized operation pattern, and adjusting the accident risk rate according to the inputted reflection rate per section.

In the method of determining the insurance premium according to another embodiment, the step of recognizing the operation pattern includes: preprocessing data by removing fluctuations and discontinuities caused by the operation of the acceleration sensor and noise from the entered travel information; Detecting an event interval in which a pattern occurs among the preprocessed data using end point detection; Resampling the data for each sample point at a predetermined time interval with respect to the detected event interval; And detecting a driving pattern by measuring the similarity of the sequence between the feature template and the resampled data stored in advance for the driving section and storing the corresponding section and the time information together.

In the insurance premium determination method according to another embodiment, the step of calculating the insurance premium may include comparing an operation pattern according to the security level reference collected in advance with the driver's recognized operation pattern, Calculating a risk weight value proportional to an accident risk rate of the driver from the comparison result; And summing the risk weights to premiums calculated according to the contract conditions of the driver's insurance of the driver.

Furthermore, a computer-readable recording medium storing a program for causing a computer to execute the methods for determining the insurance premium described above is provided.

According to an aspect of the present invention, there is provided an apparatus for determining insurance premiums, comprising: an input unit for receiving driving information collected using an acceleration sensor provided in a vehicle; And an operation unit for recognizing an operation pattern related to an accident risk in the driving habit of the driver of the vehicle from the entered travel information and calculating an insurance premium proportional to an accident risk rate of the driver based on the recognized operation pattern , The operation unit preprocesses data by removing fluctuations and discontinuities generated due to the operation of the acceleration sensor and noise from the driving information inputted and detects an event section in which the pattern is generated from the preprocessed data by using end point detection Detects a traveling pattern by resampling data for each sample point at a predetermined time interval with respect to the detected event section, measures the similarity of the sequence between the feature template previously stored with respect to the driving section and the resampled data, According to safety level standard for each section Calculating a risk weight proportional to an accident risk rate of the driver based on the comparison result and adding the risk weight to the insurance premium calculated according to a contract condition of the driver's insurance of the driver, do.

The conventional method of estimating the premium of the automobile has not been able to grasp the actual driving habits of the driver. In contrast, the embodiments of the present invention use the driving pattern extracted from the driving information collected by using the acceleration sensor provided in the vehicle, By calculating premiums proportional to the accident risk rate, it is possible to calculate more accurate insurance premiums, and it can be used as a basis for calculating individualized premiums based on driving habits for each driver.

FIG. 1 is a flowchart illustrating a method of determining insurance premiums based on operation pattern recognition according to an embodiment of the present invention.
FIG. 2 is a flowchart illustrating a process of recognizing a driving pattern in the insurance determination method of FIG. 1 according to an exemplary embodiment of the present invention.
3 is a diagram for explaining a method of detecting an event interval in recognizing an operation pattern according to an embodiment of the present invention.
4 is a diagram illustrating a detection result according to the event interval detection method of FIG. 3 according to an embodiment of the present invention.
FIG. 5 is a diagram for explaining a method of resampling data in recognizing an operation pattern according to an embodiment of the present invention. Referring to FIG.
6A to 6E are views for explaining the types of feature templates utilized in recognizing the operation pattern according to an embodiment of the present invention.
FIG. 7 is a flowchart illustrating a process of calculating a premium in proportion to an accident risk rate in the insurance premium determination method of FIG. 1 according to an embodiment of the present invention.
FIG. 8 is a flowchart illustrating a method for determining an accident risk adjusted premium in accordance with another embodiment of the present invention.
FIG. 9 is a diagram illustrating information displayed on the display means in the insurance premium determination method of FIG. 8 according to another embodiment of the present invention.
10 is a block diagram illustrating an apparatus for determining an insurance premium based on an operation pattern recognition according to an embodiment of the present invention.
11 is a diagram illustrating a coordinate system of an acceleration sensor that can be utilized in the insurance premium determination apparatus according to an embodiment of the present invention.

As described above, by analyzing the driving information using the vehicle and the sensor, it is possible to evaluate the driving pattern and the inclination of the driver by analyzing the dangerous driving which can cause a traffic accident from the driver's ordinary driving information. The dangerous driving can be suppressed. Therefore, the embodiments of the present invention are derived from the above-mentioned problem consciousness. The embodiments of the present invention recognize the traveling pattern of the vehicle by using the data obtained from the acceleration sensor provided in the vehicle, By identifying the risk factors, we propose a technological means to calculate the differential premium for each driver.

Hereinafter, embodiments of the present invention for solving the above-mentioned technical problems will be described in detail with reference to the drawings. In the following description and the accompanying drawings, detailed description of well-known functions or constructions that may obscure the subject matter of the present invention will be omitted. It is to be noted that the same components are denoted by the same names and reference numerals as possible throughout the drawings.

FIG. 1 is a flowchart illustrating a method of determining insurance premiums based on operation pattern recognition according to an embodiment of the present invention, and includes the following series of steps. These processes can be performed through an electronic data processing means, and the operating entity can be a person who intends to calculate the insurance premium. For example, an automobile insurance company may be an operating entity of such a data processing means, and may be physically implemented as a premium calculation server having at least one processor. Although the present invention has been described with reference to specific embodiments thereof, it will be apparent to those skilled in the art that various modifications and variations can be made in the present invention without departing from the spirit or scope of the invention. .

In step S110, the premium calculation server of the automobile insurance company receives the driving information collected using the acceleration sensor provided in the vehicle. To this end, an acceleration sensor should be provided in the vehicle, and various types of products such as an acceleration sensor incorporated in a portable terminal, as well as an acceleration sensor implemented in a stand-alone manner, may be utilized. The driving information recorded during the operation of the vehicle using these acceleration sensors is input to the insurance premium calculation server through step S110. From the viewpoint of the implementation, the driving record stored in the portable terminal may be uploaded to the server or data of the electronic behavior may be received from a separate recording device provided in the vehicle.

In step S120, the insurance premium calculation server recognizes an operation pattern related to an accident risk during driving habits of the driver of the vehicle from the travel information input through step S110. In step S120, the premium calculation server extracts only the information on the accident risk among the driving records recorded in a time-series manner, rather than utilizing the entire driving information inputted, and recognizes the driving pattern capable of discriminating the driver's driving habit therefrom do. For example, a situation associated with an accident risk can be a specific situation, such as a U-turn or an emergency stop, and the embodiments of the present invention classify and store the types of situations associated with such an accident risk in advance, And the driving pattern is analyzed by concentrating on the corresponding records. The operation pattern recognition will be described later in more detail with reference to FIG.

In step S130, the insurance premium calculation server calculates the insurance premium proportional to the accident risk rate of the driver based on the operation pattern recognized through step S120. The conventional insurance premium calculation method is devoid of the essence of the present invention, and a detailed explanation thereof will be omitted here. However, in step S130, the driver's personal insurance premium is calculated according to the normal insurance premium calculation method. In step S120, the driver's accident risk rate is derived from the recognized driving pattern, and the premium differentiated according to the driver's individual is taken into account Respectively. That is, the auto insurance premium is adjusted to reflect the accident risk rate of each driver. The calculation of the insurance premium considering the accident risk rate will be described later in more detail with reference to FIG.

FIG. 2 is a flowchart illustrating a process of recognizing an operation pattern in the insurance premium determination method of FIG. 1 according to an embodiment of the present invention. In order to avoid duplication of description, Assume the situation and describe only step S120 (recognition of the operation pattern). Step S120 is again composed of the following four steps.

[1] Preprocessing process

In step S121, the data is pre-processed by removing fluctuations and discontinuities caused by the operation of the acceleration sensor and noise from the driving information input through step S110.

라고 할 때, 스무딩은 다음의 수학식 1과 같이 정의된다.In the course of acquiring sensor data, each sensor is sensitive to the road surface condition and the shaking of a small vehicle. Therefore, it is necessary to preprocess the collected data before recognizing the driving pattern. In this preprocessing process, it is desirable to process the data through smoothing and normalization. The smoothing operation is applied to remove fine fluctuations and discontinuities that may occur due to sensor operation and noise in the course of acquiring sensor data. The acceleration sensor stream smoothens the average of k intervals (where k is a positive integer) for each axis. The sensor stream

, The smoothing is defined by the following equation (1).

The normalization process is applied to the same processing because the reference values of the x, y, and z axes of the three-axis acceleration sensor data are different. Through the normalization, the reference value of each axis can be moved to the origin. The normalization can be defined as Equation (2) below.

[2] Event interval detection process

In step S122, an event interval in which a pattern is generated in the data preprocessed in step S121 is detected using the end point detection. Refer to FIG. 3 illustrating a method of detecting an event interval.

In order to recognize the driving pattern, it is necessary to detect the section where the pattern occurs. The event interval detection method can borrow the end point detection method used in speech recognition. However, in the speech recognition field, the energy value and the zero crossing rate method are applied. Since the voice data and the acceleration sensor data have different characteristics, the energy value alone is applied except for the zero crossing rate method.

In addition, embodiments of the present invention use dual threshold values (Threshold_1, Threshold_2) to detect a start point (SP) and an end point (EP) of a more accurate event section. The detection method is as follows.

First, the points p1 that satisfy the first threshold value Threshold_1 are temporarily stored. Then, if the point (p2) which is equal to or more than Threshold_2 for the first time is found, the most recently stored point (p1) is determined as the start point of the event section. Then, the detection of the end point determines the point (p4) which is less than the threshold value Threshold_1 as the end point of the event section, while the threshold value Threshold_2 is not satisfied during the predetermined time period (t).

FIG. 4 is a diagram illustrating a detection result according to the event interval detection method of FIG. 3 according to an embodiment of the present invention, and actually includes three patterns in the collected data. In FIG. 4, E1 denotes a rapid start, E2 denotes a right turn, E3 denotes deceleration, and SP and EP respectively denote a start point and an end point of the E2 section. Three pattern intervals are classified by applying the event interval detection method, and the bottom part is an example of extending the sensor stream of the right turn pattern (E2) section among the classified results.

In summary, the process of detecting an event interval in the insurance premium calculation method according to the embodiments of the present invention is performed by using only the energy value of the preprocessed data, temporarily storing a point having an energy value equal to or greater than the first threshold value The energy value is determined as a start point of an event section when the energy value reaches a second threshold value or more, and when the energy value falls below the second threshold value, The point at this time can be determined as the end point of the event section when the first threshold value is reached.

Returning to Fig. 2, step S123 will be described.

[3] Data Resampling  process

In step S123, data is resampled for each sample point of a predetermined time interval with respect to the event interval detected through step S122.

In order to generate the feature template in the proposed method, all input data lengths are treated the same. To do this, the length of the data is resampled to 60 sample points (for example, can be 3 seconds). The resampling uses the data of the interval in which the pattern is generated through the event interval detection process and can apply linear interpolation. FIG. 5 is a diagram illustrating the result of applying such resampling.

The graph in the lower part of FIG. 5 shows the result after resampling and the result that the length is reduced from 182 sample points to 60 sample points as compared with the data before the upper processing, and the result of filtering the data stream by the interpolation method is also confirmed .

Returning to Fig. 2, step S124 will be described.

[4] Dynamic time Warping ( Dynamic Time Warping , DTW ) process

In step S124, the travel pattern is detected by measuring the similarity of the sequence between the template stored in advance for the driving section and the data resampled through the step S123, and the corresponding section and the time information are stored together. In this case, it is preferable that the interval pattern is an operation pattern represented by at least one selected from a rotation section, an acceleration section, a deceleration section, and an overspeed inhibition section. Such a section pattern is selected in consideration of a specific situation in which a dangerous element is to be detected in the driver's driving habit, and there may be a situation where two or more sections compete in parallel as well as the single section presented above. It is also necessary to consider the pattern by

The feature template is generated from the average of the training sets collected for each interval pattern in the driving interval. Hereinafter, a method for constructing the feature template is described in the order of priority (DTW) First, let me explain.

In order to perform step S124, it is necessary to configure the feature template to be compared before applying the DTW. The feature template is constructed by applying the average calculation of the training set which is the same length data through resampling. The number of data used in the training set to generate the feature template through the present embodiment is 30 each.

FIGS. 6A to 6E are diagrams for explaining a type of a feature template used in recognizing an operation pattern according to an exemplary embodiment of the present invention, and illustrate a feature template created by averaging a training set for each pattern.

6A is a diagram illustrating a feature template in a left turn and a U-turn situation in order. In the U-turn graph at the bottom, the portion indicated by the graph is not a result displayed by the brake operation, The result is similar to the deceleration effect by the force (leaning forward).

FIG. 6B is a diagram illustrating feature templates in right turn and rapid-deceleration situations, respectively, and FIG. 6C is a diagram illustrating feature templates in rapid-start and rapid-acceleration situations, respectively.

FIGS. 6D and 6E are characteristic templates of the overspeed inhibition tile, and are a result of examining the flow of the data stream to obtain a more accurate recognition result and structuring it into four categories. FIG. 6D is a diagram illustrating a feature template in the fast-speed-braking-during-deceleration state and the low-speed-running state, respectively, in the order, and FIG. 6E is a diagram illustrating a feature template in the overspeed-

In step S124, the similarity of the sequence is measured (DTW) by utilizing the feature template previously configured as described above, and the traveling pattern is detected. This DTW is one of the pattern recognition methods applicable to time series patterns. The DTW algorithm is suitable for a matching method for measuring the similarity of two sequences whose lengths do not match. In the embodiments of the present invention, the DTW algorithm is applied to evaluate the similarity, and the data of the test set is also subjected to linear interpolation Lt; RTI ID = 0.0 > resampling < / RTI > The reason for this is that the number of patterns to be compared is considerably smaller than that of pattern recognition such as speech recognition. Unlike speech data, the change in acceleration of each pattern is visually easily distinguishable. It is considered to be irrelevant. In addition, through the resampling, the computation process of the DTW algorithm is shortened and the processing speed is improved.

와

같이 주어졌을 때, DTW 알고리즘은 다음의 수학식 3에 의해 결정된다.Two sequences

Wow

When given together, the DTW algorithm is determined by the following equation (3).

는 각 시퀀스의 요소들 간의 거리(cost)이며, 이러한

는 다음의 수학식 4에 의해 구해진다.In Equation 3,

Is the distance between the elements of each sequence,

Is obtained by the following equation (4).

Then, the similarity of the two sequences X and Y is determined by the following equation (5).

Therefore, the smaller the value of the result of Equation (5), the higher the degree of similarity. That is, the inputted data and the feature template are compared through the DTW algorithm to determine a pattern having a high degree of similarity. At this time, it is preferable that the travel pattern result detected by the DTW algorithm is stored together with the time on the sensor stream and the detection interval.

FIG. 7 is a flowchart specifically illustrating a process of calculating a premium in proportion to an accident risk rate in the insurance premium determination method of FIG. 1 according to an embodiment of the present invention. Here, in order to avoid duplication of explanation, ), And the description will be focused on only step S130 (calculation of insurance premium).

In step S131, the operation pattern according to the safety level criterion collected beforehand and the recognized operation pattern of the driver are compared with each other. In this case, the preliminarily collected safety standard for each operating section means a statistical value of an operation pattern collected by operating a normal driver at a safe level for each section of an actual road, and the driving behavior of the insured person As a reference for comparison with. For this purpose, it is desirable for the insurer to use the test vehicle to comply with the prescribed road traffic laws for each section of the actual road, and to extract the operation pattern from the collected operation records and perform statistical processing.

In step S132, a risk weight value proportional to the accident risk rate of the driver is calculated from the comparison result through step S131. Since the operation pattern has been recognized through step S120, if the operation patterns according to the safety standard in the section corresponding to the operation pattern are compared with each other in operation S131, it is possible to know how far the operation record of the actual driver exceeds the safety standard . Therefore, in step S132, the risk weight value proportional to the quantitative difference value obtained through the comparison result of step S131 can be calculated.

In step S133, the final driver premium is calculated by adding the risk weights calculated through step S132 to the premium calculated according to the contract conditions of the driver's insurance of the driver. Here, the premium calculated based on the terms and conditions of the driver's insurance means the premium according to the normal method of calculating premiums. That is, without considering the actual driving data of the driver, the insurance premium calculated based on the normal insurance premium calculation standards (for example, the vehicle price, the driving experience, the accident history, etc.) The final driver's premium is determined by reflecting the risk weight. If the driver is a safe driver, the final driver's premium may be equal to or lower than the normal premium, as reflected by a relatively low risk weighting. On the other hand, if the driver has a risky driving habit, the driver's premium will ultimately be higher than the normal premium, reflecting a relatively high risk weighting.

FIG. 8 is a flowchart illustrating a method for determining an accident rate-adjusted premium according to another embodiment of the present invention. The steps S110 and S120 of the insurance premium determination method of FIG. 1 described above are performed in the same manner. Therefore, in order to avoid duplication of explanation, it is assumed here that step S120 is completed, and only the subsequent steps will be described. The embodiment of FIG. 8 differs from the insurance determination method of FIG. 1 in that an insurance company employee visually confirms additional data about a driver's driving record, thereby taking into account an exceptional situation that may appear during the mechanical premium calculation process, It suggests a technical means to adjust.

In step S830, the insurance premium server of the insurance company matches the corresponding section information and the time information together with the recognized operation pattern through step S120, and displays the matching information on the display unit. That is, the time information of the detected section is used to match the video image corresponding to each running pattern and provided to the driver, so that it can be utilized for analyzing the traveling pattern. To this end, the vehicle should be provided with a camera capable of capturing an image in addition to the acceleration sensor. It is also preferable that a GPS module is utilized to obtain position information on the travel section.

FIG. 9 is a view illustrating information displayed through a display unit in the insurance premium determination method of FIG. 8 according to another embodiment of the present invention. The image information, acceleration sensor information, and GPS information acquired using the smart phone are collected It is an image that is matched to one screen and displayed on the screen. In other words, it is confirmed that the image of the section running through the camera of the smartphone is recorded, the sensor stream of the pattern collected through the acceleration sensor of the smartphone is recorded, and the position information collected through the GPS module is displayed on the map .

Returning to Fig. 8, the flow advances to step S840.

In step S840, the insurance premium calculation server of the insurance company inputs a reflection rate for each section considering the visually recognized driving situation from the user through the display means. That is, the insurance company employee can check the image displayed on the display means through step S830, determine whether there is an unexpected situation in the driving situation, and input the risk reflection rate in the corresponding interval in consideration of the unexpected situation. For example, if the driving information input through step S110 is due to an unexpected situation such as a situation in which a dog is leaping on a road as a result of checking an actual image, It would be appropriate to exclude the data from reflecting the risk for calculating premiums. To this end, the embodiment of the present invention provides the image information about the actual recording situation through step S830, and receives the risk reflection rate from the user through step S840.

In step S850, the insurance premium calculation server of the insurance company calculates the insurance premium proportional to the accident risk rate of the driver based on the recognized operation pattern through step S120, and adjusts the accident risk rate according to the reflection rate To determine the final driver 's premium. Of course, the process of calculating the insurance premium prior to adjusting the accident risk rate may be performed by comparing the operation pattern according to the safety level criterion according to the preliminarily collected driving range with the recognized driving pattern of the driver, And the risk weight is added to the insurance premium calculated according to the contract condition of the driver's insurance of the driver.

FIG. 10 is a block diagram illustrating an apparatus for determining insurance premiums based on operation pattern recognition according to an embodiment of the present invention, and includes a structure for performing operations corresponding to the respective steps of FIGS. 1 and 8 Therefore, in order to avoid duplication of explanations, we focus on the device characteristics and the flow of data, and outline the outline only.

In FIG. 10, the device 20 for determining insurance premiums may be an insurance premium calculation server of an insurance company, and receives driving information sensed from the vehicle 10 equipped with various sensors 11, 13, and 15.

The input unit 21 of the insurance premium calculation server receives the travel information collected by using the acceleration sensor provided in the vehicle 10. The input unit 21 may be implemented as various input terminals or communication means for facilitating input of electronic data. For example, a wired / wireless network interface or a USB to serial terminal.

The operation unit 23 of the premium calculation server recognizes an operation pattern related to an accident risk in the driving habit of the driver of the vehicle from the driving information inputted through the input unit 21, Calculate premiums that are proportional to the accident rate.

The operation unit 23 preprocesses data by eliminating fluctuations and discontinuities caused by the operation of the acceleration sensor and the noise from the input driving information, and detects an event in which the pattern is generated from the preprocessed data Detects the travel pattern by resampling the data for each sample point at a predetermined time interval with respect to the detected event segment, and measuring the similarity of the sequence between the feature template previously stored for the driving segment and the resampled data. Preferably, the feature template is stored in the database 25 in advance.

The operation unit 23 compares the operation pattern according to the safety level criterion collected in advance with the operation pattern recognized by the driver in advance and calculates a risk weight value proportional to the accident risk rate of the driver from the comparison result, The risk weights are added to the insurance premium calculated according to the contract conditions of the driver's insurance of the driver. For this purpose, the operation pattern according to the safety level criterion preliminarily collected in advance may be stored in the database 25.

Meanwhile, the insurance premium determination apparatus according to another embodiment of the present invention may further include a display means 30. In addition to the operation pattern identified through the display means 30, the actual photographed image and the positional information are matched and displayed so that the responsible person of the insurance company can be guided to receive the risk reflectance rate during the driving situation.

Hereinafter, the performance evaluation of the embodiments of the present invention and the results thereof will be introduced by producing software according to a series of calculation methods defined in the insurance premium determination method and the sensors provided in the vehicle.

The data acquisition environment of this experiment is as follows. The LG-LU3000 Android smartphone was installed on the front glass mounted on the front glass of the vehicle, and the driving information was obtained through actual driving at Kunsan National University campus. Seven pattern data were collected from left to right, U turn, right turn, rapid acceleration, rapid acceleration, speed braking (decelerating speed braking, speeding braking during low speed driving) All patterns were recorded at a speed of less than 60 km. Details of the acquired data are shown in Table 1 below.

In this experiment, only 3 - axis acceleration sensor was used and video data and time information were additionally acquired to match video images. The video data obtained 16 frames (frames per second) images (320 * 240) in the form of mp4 file.

FIG. 11 is a view illustrating a coordinate system of an acceleration sensor that can be utilized in the insurance premium determination apparatus according to an embodiment of the present invention. In FIG. 11, a coordinate system of a smart phone equipped with a three-axis acceleration sensor actually used in the present experiment is shown. The smartphone is installed horizontally on the mount, and the X axis corresponds to the gravity direction, the Y axis corresponds to the left and right direction of the vehicle, and the Z axis corresponds to the traveling direction of the vehicle.

The experiment determined the recognition rate based on all the data including the data of the training set. The discrimination of the recognition rate is performed by comparing the data of the recognized segment with the video stream by applying the event segment detection algorithm in the entire data stream, classifying each pattern by hand, and then calculating the accuracy through the recognizer. When we look at the recognition results, we showed a good recognition rate of about 97% in this experiment.

According to the embodiments of the present invention described above, the driving pattern extracted from the driving information collected using the acceleration sensor provided in the vehicle can be used , It is possible to estimate the premium more accurately by calculating the insurance premium that is proportional to the accident risk rate of the actual driver and can be utilized as a basis for calculating the individual premium based on the operating humidity per driver. Particularly, the advantages of the embodiments of the present invention proposed above are that a method of detecting and recognizing a driving pattern so as to judge the driving tendency for each driving pattern is proposed, and the recognition rate is improved by acquiring and training data for each pattern This method can induce objective premium calculation.

Meanwhile, the embodiments of the present invention can be embodied as computer readable codes on a computer readable recording medium. A computer-readable recording medium includes all kinds of recording apparatuses in which data that can be read by a computer system is stored.

Examples of the computer-readable recording medium include a ROM, a RAM, a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device and the like, and also a carrier wave (for example, transmission via the Internet) . In addition, the computer-readable recording medium may be distributed over network-connected computer systems so that computer readable codes can be stored and executed in a distributed manner. In addition, functional programs, codes, and code segments for implementing the present invention can be easily deduced by programmers skilled in the art to which the present invention belongs.

The present invention has been described above with reference to various embodiments. It will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined by the appended claims. Therefore, the disclosed embodiments should be considered in an illustrative rather than a restrictive sense. The scope of the present invention is defined by the appended claims rather than by the foregoing description, and all differences within the scope of equivalents thereof should be construed as being included in the present invention.

10: vehicle 11, 13, 15: sensors
20: Premium determination device 21: Input unit
23: operation unit 25:
30: Display means

Claims (10)

In a method for determining premiums,
Receiving driving information collected using an acceleration sensor provided in the vehicle;
Recognizing an operation pattern related to an accident risk in the driving habit of the driver of the vehicle from the input driving information; And
And calculating an insurance premium proportional to an accident risk rate of the driver based on the recognized operation pattern. The method according to claim 1,
The step of recognizing the operation pattern includes:
Preprocessing data by removing fluctuations and discontinuities caused by operation and noise of the acceleration sensor from the input driving information;
Detecting an event interval in which a pattern occurs among the preprocessed data using end point detection;
Resampling the data for each sample point at a predetermined time interval with respect to the detected event period; And
Detecting a driving pattern by measuring a similarity of a sequence between a feature template and a resampled data stored in advance for the driving section and storing the corresponding section and the time information together. 3. The method of claim 2,
The step of detecting the event period includes:
Detecting only the energy value of the preprocessed data,
Temporarily storing a point having an energy value equal to or greater than a first threshold value and then determining the temporarily stored point as a start point of an event interval when the energy value reaches a second threshold value or more,
Wherein when the energy value falls below the second threshold value and the energy value reaches the first threshold value or less within a predetermined time period, the point at that time is determined as the end point of the event period . 3. The method of claim 2,
Wherein the feature template is generated from an average of training sets collected for each interval pattern of the driving interval,
Wherein the interval pattern is an operation pattern represented by at least one selected from a rotation section, an acceleration section, a deceleration section, and a overspeed inhibition section. The method according to claim 1,
The step of calculating the premium includes:
A step of comparing an operation pattern according to a safety degree criterion collected in advance for each driving section with a recognized operation pattern of the driver;
Calculating a risk weight value proportional to an accident risk rate of the driver from the comparison result; And
And adding the risk weight to the insurance premium calculated according to a contract condition of the driver's insurance of the driver. In a method for determining premiums,
Receiving driving information collected using an acceleration sensor provided in the vehicle;
Recognizing an operation pattern related to an accident risk in the driving habit of the driver of the vehicle from the input driving information;
Matching the corresponding section information and the time information to the recognized operation pattern and displaying the same on the display means;
Receiving a reflection rate for each section considering a visually recognized driving situation from the user through the display means; And
Calculating an insurance premium proportional to an accident risk rate of the driver based on the recognized operation pattern, and adjusting the accident risk rate according to the inputted reflection rate per section. The method according to claim 6,
The step of recognizing the operation pattern includes:
Preprocessing data by removing fluctuations and discontinuities caused by operation and noise of the acceleration sensor from the input driving information;
Detecting an event interval in which a pattern occurs among the preprocessed data using end point detection;
Resampling the data for each sample point at a predetermined time interval with respect to the detected event interval; And
Detecting a traveling pattern by measuring the similarity of the sequence between the feature template previously stored for the driving section and the resampled data, and storing the corresponding section and the time information together. The method according to claim 6,
The step of calculating the premium includes:
A step of comparing an operation pattern according to a safety degree criterion collected in advance for each driving section with a recognized operation pattern of the driver;
Calculating a risk weight value proportional to an accident risk rate of the driver from the comparison result; And
And adding the risk weight to the insurance premium calculated according to a contract condition of the driver's insurance of the driver. A computer-readable recording medium storing a program for causing a computer to execute the method according to any one of claims 1 to 8. An apparatus for determining insurance premium,
An input unit for receiving driving information collected using an acceleration sensor provided in the vehicle; And
And an operation unit for recognizing an operation pattern related to an accident risk in the driving habit of the driver of the vehicle from the input driving information and calculating an insurance premium proportional to an accident risk rate of the driver based on the recognized operation pattern,
The operation unit,
Processing the data by removing fluctuations and discontinuities caused by operation and noise of the acceleration sensor from the input driving information, detecting an event section where a pattern occurs in the preprocessed data by using end point detection, The traveling pattern is detected by resampling the data for each sample point at predetermined time intervals with respect to the event section, measuring the similarity of the sequence between the feature template previously stored with respect to the driving section and the resampled data,
The method includes the steps of: comparing an operation pattern according to a safety level criterion according to a driving range preliminarily collected and a recognized driving pattern of the driver; calculating a risk weight proportional to an accident risk rate of the driver from the comparison result; And adds the risk weights to the insurance premium calculated according to Equation (1).

Images