TWI820471B - Vehicle driving risk assessment system - Google Patents

Abstract

A vehicle driving risk assessment system includes a driving time analysis unit, a driving behavior analysis unit, and a risk assessment unit. The driving time analysis unit can analyze driving information to obtain all time segments of manual driving and automatic driving. The driving behavior analysis unit can analyze driving information to obtain manual driving behavior parameters and automatic driving behavior parameters. The risk assessment unit can analyze manual driving behavior parameters and automatic driving behavior parameters to assess the driver's driving accident risk. This design makes the vehicle driving risk assessment system of the present invention suitable for introducing driving behavior car insurance application services, and can be used for Provide more reasonable auto insurance premium estimates for drivers who drive vehicles with autonomous driving capabilities.

Description

Vehicle driving risk assessment system

The present invention relates to a risk assessment system, and in particular, to a vehicle driving risk assessment system for assessing traffic accident risks.

Driving behavior car insurance application service (Usage Based Insurance, referred to as UBI) is a new vehicle insurance mechanism derived from the development of vehicle monitoring and sensing technology. The evaluation design of UBI car insurance is based on the data recorded by the sensing equipment on the vehicle. Driving information is used to analyze the driver's driving behavior, and based on the various driving behaviors analyzed, the driver's accident risk is assessed to set auto insurance premiums.

However, in recent years, many vehicles have begun to introduce automatic driving technologies, such as cruise control, automatic car following, automatic navigation driving and automatic parking, etc., so that the driver may be in manual driving mode part of the time while driving. Self-driving mode is used part of the time. How to effectively assess the accident risk of drivers of vehicles with self-driving functions is an urgent issue in the insurance industry.

Therefore, the object of the present invention is to provide a method that can improve the prior art. At least one shortcoming of the vehicle driving risk assessment system.

Therefore, the vehicle driving risk assessment system of the present invention is suitable for analyzing a piece of driving information of a vehicle to evaluate the driving accident risk of the driver of the vehicle. The vehicle driving risk assessment system includes a driving time analysis unit, a driving behavior analysis unit, and a risk assessment unit.

The driving time analysis unit can be used to analyze the driving information to obtain all hand-driving time sections corresponding to manual driving, all self-driving time sections corresponding to automatic driving, and a total manual driving time obtained by summing up all hand-driving time sections. (T _H ), and a total autonomous driving time ( _TM ) obtained by summing up all self-driving time segments.

The driving behavior analysis unit can be used to analyze the driving information to obtain all manual driving behavior parameters that exist in each manual driving time segment, and to obtain all automatic driving behavior parameters that exist in each self-driving time segment.

The risk assessment unit has a risk assessment model established through machine learning, and includes an accident risk assessment module. The accident risk assessment module can analyze all manual driving behavior parameters in all driving time periods through the risk assessment model, All self-driving behavior parameters of all self-driving time segments, the total manual driving time, and the total self-driving time are used to obtain the probability of a hand-driving accident (P _H ) and the loss of a hand-driving accident corresponding to the driver and the vehicle. valuation (L _H ), a self-driving accident probability (P _M ), and a self-driving accident loss valuation (L _M ).

The effect of the present invention is: by activating the automatic driving mode for the vehicle The driving information during the automatic driving mode and the manual driving mode is analyzed to obtain the design of various automatic driving behavior parameters during the automatic driving mode and the manual driving mode, so that the vehicle driving risk assessment system of the present invention is suitable for introducing driving behavior auto insurance applications. Service (Usage Based Insurance, UBI), which can be used to provide more reasonable auto insurance premium estimates for drivers who drive vehicles with autonomous driving capabilities.

200: Vehicle driving risk assessment system

3: Driving time analysis unit

4: Driving behavior analysis unit

5:Risk Assessment Unit

501: Obtain driving information

502: Data classification

503: Manual driving mode risk assessment

504: Autonomous driving mode risk assessment

505: Comprehensive Risk Assessment

506: Insurance cost estimate

507: Analysis and suggestions for optimal driving modes

51:Accident Risk Assessment Module

52:Accident Level Assessment Module

53: Premium estimation module

6: Driving mode recommendation unit

701:Vehicle

702:Mobile device

801: Real-time driving information database

802: Historical traffic accident information database

900: Electronic devices

901:Display

Other features and effects of the present invention will be clearly presented in the embodiments with reference to the drawings, in which: Figure 1 is an architectural schematic diagram illustrating an embodiment of the vehicle driving risk assessment system of the present invention implemented on an electronic device; And when used with a real-time driving information database and a historical traffic accident information database: Figure 2 is a functional block diagram of this embodiment; and Figure 3 is a step flow chart illustrating vehicle driving risk assessment in this embodiment time steps.

Referring to Figures 1 and 2, one embodiment of the vehicle driving risk assessment system 200 of the present invention is suitable for implementation and construction on an electronic device 900 through electronic circuits, firmware and/or program software, and can be used to obtain an autonomous driving function. Vehicle 701 transmits to a real-time driving information database via the Internet and/or mobile communication network The driving information of 801 can be analyzed to evaluate the accident risk of the driver of vehicle 701 and the auto insurance premium payable by the driver, as well as analyze the driving mode applicable to the driver, and evaluate the driver's driving information. The held mobile device 702 sends the analyzed driving mode suggestions. The electronic device 900 has a display 901 . The driving information includes data measured by various sensors installed on the vehicle 701 and data recorded by various recorders. The driving information includes, for example, but not limited to, automatic driving mode and manual driving mode. The following driving image data, distance data of the vehicle in front, distance data of the vehicle behind, accelerator pedal data, brake pedal data, and driving speed data, etc.

The vehicle driving risk assessment system 200 includes a driving time analysis unit 3, a driving behavior analysis unit 4, a risk assessment unit 5, and a driving mode recommendation unit 6.

The driving time analysis unit 3 can be used to analyze the driving information to obtain all hand-driving time segments corresponding to the manual driving mode, all self-driving time segments corresponding to the automatic driving mode, and a sum of all hand-driving time segments. The total manual driving time (T _H ), and the total automatic driving time (T _M ) obtained by summing up all self-driving time segments.

The driving behavior analysis unit 4 can be used to analyze the driving information to obtain all manual driving behavior parameters that exist in each manual driving time segment, and to obtain all automatic driving behavior parameters that exist in each self-driving time segment. In this embodiment, the driving The behavior analysis unit 4 is based on the driving behavior analysis model established in advance through machine learning, and uses image analysis technology and big data statistical analysis to obtain each driving time segment in which the driver manually drives the vehicle 701 The various manual driving behaviors performed, as well as the number and frequency of occurrences of various manual driving behaviors, are used to obtain the manual driving behavior parameters. And obtain various automatic driving behaviors performed by the vehicle 701 in each self-driving time section, as well as the number and frequency of occurrences of various automatic driving behaviors, thereby obtaining the automatic driving behavior parameters.

In this embodiment, the manual driving behavior and the automatic driving behavior among the manual driving behavior parameters and the automatic driving behavior parameters include, but are not limited to, changes in driving speed, emergency braking, heavy accelerator pedal, and distance from the vehicle in front. , distance between vehicles behind, lane changes, lane deviations, and running red lights, etc.

The risk assessment unit 5 includes an accident risk assessment module 51, an accident level assessment module 52, and a premium estimation module 53, and has a risk assessment model and an accident level assessment model established through machine learning.

In this embodiment, the risk assessment unit 5 will connect to a historical traffic accident information database 802 through the electronic device 900 signal to obtain a plurality of traffic accident data, and will analyze each traffic accident data from before the occurrence. During the period from a predetermined time to the moment when the accident occurs, such as the day before, three days before, or the week before the accident, the various manual driving behavior parameters and various automatic driving behavior parameters recorded above are recorded during the period from the predetermined time to the moment when the accident occurs, and injuries and deaths of drivers and passengers caused by various traffic accidents The degree of accident, the degree of casualties of people around the scene, the degree of damage to the accident vehicle and the degree of damage to the surrounding environment are analyzed through big data statistics, and machine learning algorithms are used for calculation training to establish the risk assessment model and the accident level assessment model. Since there are many types of machine learning algorithms used to establish such models, such as but not limited to neural network systems, support vector machines, decision trees, etc., they will not be described in detail.

The accident level assessment module 52 will use the risk assessment model to evaluate all manual driving behavior parameters in all manual driving time sections, all automatic driving behavior parameters in all self-driving time sections, the total manual driving time and the total automatic driving time. The time is calculated to obtain an accident level that represents the driving accident that the driver may have.

The accident risk assessment module 51 can use the risk assessment model to evaluate all manual driving behavior parameters in all manual driving time sections, all automatic driving behavior parameters in all self-driving time sections, the total manual driving time, the total automatic driving time Time, and the accident level are calculated to obtain the probability of a hand-driving accident (P _H ), a loss estimate of a hand-driving accident (L _H ), a probability of a self-driving accident (P _M ), and A self-driving accident loss estimate (L _M ), and the display 901 displays the hand-driving accident probability (P _H ), the hand-driving accident loss estimate (L _H ), and the self-driving accident probability (P _M ) , and the estimated loss of the self-driving accident ( _LM ).

P]利用駕駛時間加權的計算方式分析得到一個綜合事故風險(P)，且會根據公式

分析得到一個綜合損失金 額(L)。並根據公式[P×L=N]計算得到對應該駕駛人與該車輛701的該車險保費(N)，並於該顯示器901顯示出該車險保費(N)。 The premium estimation module 53 will be based on the formula

P] uses the driving time weighted calculation method to analyze and obtain a comprehensive accident risk (P), and will be calculated according to the formula

The analysis yields a comprehensive loss amount (L). The auto insurance premium (N) corresponding to the driver and the vehicle 701 is calculated according to the formula [P×L=N], and the auto insurance premium (N) is displayed on the display 901.

The driving mode recommendation unit 6 will calculate a hand-driving accident financial loss amount based on the hand-driving accident probability (P _H ) and the hand-driving accident loss estimate (L _H ), and will calculate the amount of financial loss based on the self-driving accident probability. (P _M ) and the self-driving accident loss estimate (L _M ) are estimated to obtain a self-driving accident financial loss amount, and the driving mode recommendation unit 6 will determine that the self-driving accident financial loss amount is greater than the self-driving accident financial loss amount. When, a recommended self-driving mode message is generated, and when it is determined that the financial loss amount of the manual driving accident is less than the financial loss amount of the self-driving accident, a recommended manual driving mode message is generated. During implementation, the driving mode recommendation unit 6 can transmit the recommended self-driving mode message or the recommended hand-driving mode message to the mobile device 702 carried by the driver through the communication function design of the electronic device 900 , or directly transmit it to the mobile device 702 carried by the driver. Vehicle 701, and a display (not shown) equipped in the vehicle 701 displays a display to remind the driver to adopt the recommended driving mode. However, in other implementation aspects of the present invention, the output method of the recommended self-driving mode message or the recommended hand-driving mode message generated by the driving mode recommendation unit 6 is not limited to the above method.

Referring to Figures 1, 2, and 3, the vehicle driving risk assessment system 200 of the present invention includes the following steps when performing vehicle driving risk assessment:

Step 501. Get driving information. The vehicle driving risk assessment system 200 is connected to the real-time driving information database 801 through the electronic device 900 to obtain the All driving information currently uploaded by vehicle 701.

Step 502. Data classification. The driving time analysis unit 3 and the driving behavior analysis unit 4 are caused to analyze the driving information and classify the data for the manual driving mode and the automatic driving mode to respectively obtain all driving time segments corresponding to the manual driving mode and the total manual driving time. Time ( _TH ) and all manual driving behavior parameters, as well as all self-driving time sections corresponding to the automatic driving mode, the total self-driving time ( _TM ) and all automatic driving behavior parameters.

Then, the manual driving mode risk assessment in step 503 and the automatic driving mode risk assessment in step 504 are performed respectively. In step 503, the risk assessment unit 5 will analyze and obtain the hand-driving accident probability ( _PH ) and the hand-driving accident loss estimate (L _H ) corresponding to the manual driving mode. In step 504, the risk assessment unit 5 will analyze and obtain the self-driving accident probability ( _PM ) and the self-driving accident loss estimate ( _LM ) corresponding to the self-driving mode.

Then, step 505 is performed. Comprehensive risk assessment. The risk assessment unit 5 will integrate the probability of occurrence of the hand-driving accident (P _H ), the estimated loss of the hand-driving accident (L _H ), the probability of occurrence of the self-driving accident (P _M ), and the estimated loss of the self-driving accident (L _M ) to perform a comprehensive risk assessment on the driver of vehicle 701 to obtain the amount of financial loss in the hand-driving accident and the amount of financial loss in the self-driving accident.

Then, the insurance cost estimation in step 506 and the analysis and recommendation of the best driving mode in step 507 are performed respectively. In step 506, the vehicle 701 will be The driver's insurance cost estimation is performed in step 507, and the analyzed recommended self-driving mode information or the recommended hand-driving mode information is sent to the mobile device 702 carried by the driver.

In summary, by analyzing the driving information of the vehicle 701 during the activation of the automatic driving mode and the manual driving mode, various automatic driving behavior parameters during the automatic driving mode are obtained, and the various automatic driving behavior parameters during the manual driving mode are obtained. The design of various manual driving behavior parameters, and the analysis of the risk assessment model and the accident level assessment model to obtain the probability of occurrence of the manual driving accident (P _H ), the loss estimate of the manual driving accident (L _H ), and the self-driving accident The probability of occurrence (P _M ) and the self-driving accident loss estimate (L _M ) are further analyzed to obtain the design corresponding to the driver's auto insurance premium, making the vehicle driving risk assessment system 200 of the present invention suitable for introducing driving behavior auto insurance application services ( Usage Based Insurance (UBI for short) can be used to make more reasonable auto insurance premium estimates for drivers who drive vehicles 701 with autonomous driving functions.

In addition, by further analyzing the difference between the amount of financial losses in the hand-driving accident and the amount of financial losses in the self-driving accident, the recommended self-driving mode message or the design of the recommended hand-driving mode message can be provided to the driver according to the driver driving the vehicle. The manual driving behavior of the vehicle 701 and the automatic driving behavior of the vehicle 701 during automatic driving recommend a relatively safe driving mode suitable for the vehicle 701 to the driver, thereby helping to reduce the risk of traffic accidents. Therefore, the vehicle driving risk assessment system 200 of the present invention is indeed A quite innovative creation that can indeed achieve the purpose of this invention.

However, the above are only examples of the present invention and should not be used to limit the scope of the present invention. All simple equivalent changes and modifications made based on the patent scope of the present invention and the content of the patent specification are still within the scope of the present invention. within the scope covered by the patent of this invention.

200: Vehicle driving risk assessment system

3: Driving time analysis unit

4: Driving behavior analysis unit

5:Risk Assessment Unit

51:Accident Risk Assessment Module

52:Accident Level Assessment Module

53: Premium estimation module

6: Driving mode recommendation unit

900: Electronic devices

901:Display

Claims (4)

A vehicle driving risk assessment system is suitable for analyzing a piece of driving information of a vehicle to assess the driving accident risk of the driver of the vehicle, and includes: a driving time analysis unit that can be used to analyze the driving information to obtain a corresponding manual driving All manual driving time segments, all self-driving time segments corresponding to automatic driving, a total manual driving time (T H ) obtained by summing all manual driving time segments, and a total manual driving time (T _H ) obtained by summing all self-driving time segments. Total self-driving time ( _TM ); a driving behavior analysis unit can be used to analyze the driving information to obtain all manual driving behavior parameters that exist in each self-driving time section, and to obtain all manual driving behavior parameters that exist in each self-driving time section. All autonomous driving behavior parameters; a risk assessment unit with a risk assessment model established through machine learning and including an accident risk assessment module that can analyze all driving time periods through the risk assessment model All manual driving behavior parameters, all self-driving behavior parameters of all self-driving time segments, the total manual driving time, and the total self-driving time are used to obtain a hand-driving accident probability (P _H ), a hand-driving accident loss estimate (L _H ), a self-driving accident probability (P _M ), and a self-driving accident loss estimate (L _M ); and a driving mode recommendation unit that will use the P _H and the L _H calculates the amount of financial losses in a hand-driving accident, and calculates an amount of financial losses in a self-driving accident based on the P _M and the L _M , and determines that the amount of financial losses in a hand-driving accident is greater than the amount of financial losses in a self-driving accident , a recommended self-driving mode message is generated, and when it is determined that the financial loss amount of the manual driving accident is less than the financial loss amount of the self-driving accident, a recommended manual driving mode message is generated. The vehicle driving risk assessment system as described in claim 1, wherein the risk assessment unit also has an accident level assessment model established through machine learning, and also includes an accident level assessment module, which uses The accident level assessment model analyzes all manual driving behavior parameters in all manual driving time segments, all self-driving behavior parameters in all self-driving time segments, the total manual driving time, and the total self-driving time to obtain a representative value of the driving time. The accident risk assessment module will also analyze the accident level through the risk assessment model to obtain the loss estimate of the hand-driving accident (L _H ) and the self-driving accident loss estimate ( L _M ). The vehicle driving risk assessment system as described in request item 1 or 2, wherein the risk assessment unit also includes a premium estimation module, and the premium estimation module will calculate according to the formula

The analysis results in a comprehensive accident risk (P), which is calculated according to the formula

L _M =L] analysis to obtain a comprehensive loss amount (L), and based on the comprehensive accident risk (P) and the comprehensive loss amount (L), a car insurance premium corresponding to the driver is calculated. The vehicle driving risk assessment system as described in claim 3, wherein the premium estimation module calculates the auto insurance premium (N) according to the formula [P×L=N].

Images