Method and system for monitoring, evaluating and warning of safe driving behavior of a driver

Abstract

The application provides a kind of driver safe driving behavior monitoring, evaluation and early warning method and system, it is related to online car-hailing technical field, comprising: step S1: collection driving data, and the driving data is handled;Step S2: model is established, including driving behavior score and driving result score, the driving person comprehensive driving rating is obtained after the combination of two;Step S3: according to the driving person comprehensive driving rating, different safety assistance and prompt are taken to the driver.The application can effectively identify personnel with potential driving risk, implement targeted safety prompts and education, thereby reducing traffic accidents.

Description

Technical Field

[0001] This invention relates to the field of ride-hailing technology, specifically to a method for monitoring, evaluating, and issuing early warnings based on the safe driving behavior of ride-hailing drivers, and more particularly to a method and system for monitoring, evaluating, and issuing early warnings based on the safe driving behavior of drivers. Background Technology

[0002] Ride-hailing, short for online-booked taxi service, refers to the business activities that rely on Internet technology to build a service platform, connect qualified vehicles and drivers, and provide non-cruising, booked taxi services by integrating supply and demand information.

[0003] The invention patent with publication number CN113395394A discloses a method for early warning of the safety of ride-hailing trips, including (1) calculating the straight-line distance between the end of the order location and the scheduled drop-off location, and determining whether the straight-line distance is greater than a first threshold; (2) transmitting the trip recording; (3) asking whether a safety problem has occurred. If the user chooses to end the operation normally, or if the user chooses to jump to the online customer service chat page due to an order abnormality, or if the user chooses to end the abnormality and call 110, the user enters step (4); or if the user chooses to end the abnormality and notify the emergency contact, the user enters step (5); (4) asking whether the user wants to call 110. If not, the order returns to normal and the operation ends. If the user does, the user enters the alarm interface. After the alarm is completed, the order returns to normal and the operation ends. (5) determining whether the passenger has set an emergency contact. If not, the user enters the emergency contact's mobile phone number and the abnormal information SMS is sent to that mobile phone number. Otherwise, the user enters the dialing interface and the passenger calls the emergency contact.

[0004] The core of the aforementioned patent is to determine the safety of a single trip based on the behavior of ride-hailing users and the order itself. However, since the patent only provides safety warnings based on a single trip, it cannot assess the driver's driving behavior and safety awareness over a given time period. This limits the accuracy and continuity of the evaluation, thus affecting its effectiveness. Summary of the Invention

[0005] In view of the deficiencies in the prior art, the present invention provides a method and system for monitoring, evaluating and warning of safe driving behavior of drivers.

[0006] According to the present invention, a method and system for monitoring, evaluating, and providing early warning of safe driving behavior of a driver are provided, the scheme of which is as follows:

[0007] Firstly, a method for monitoring, evaluating, and issuing early warnings of safe driving behavior is provided, the method comprising:

[0008] Step S1: Collect driving data and process the driving data;

[0009] Step S2: Establish a model, including driving behavior score and driving result score, and combine the two to obtain the driver's comprehensive driving rating;

[0010] Step S3: Based on the driver's comprehensive driving rating, provide different safety assistance and prompts to the driver.

[0011] Preferably, the driving data includes: real-time vehicle data during driving, including basic information on the OBD high-frequency tracking location and alarm information; and driving result data, including traffic violations and insurance claim data during the observation period.

[0012] Preferably, the processing of driving data includes: the OBD device performing edge computing on the driver's driving situation every second, and sending the processed alarm results back to the cloud.

[0013] Preferably, the driving behavior scoring includes: using a supervised learning algorithm to perform linear regression training on a large amount of ride-hailing vehicle trip data collected by OBD, and combining key trip features to construct a model of the probability of an accident per 100,000 kilometers for a single vehicle, thus characterizing the driving behavior and travel risk probability of each driver.

[0014] Preferably, the driving result score includes: a final determination based on rapid acceleration from 100 km / h, rapid deceleration from 100 km / h, sharp turns from 100 km / h, speeding from 100 km / h, the percentage of time spent driving while fatigued, and the average daily mileage, as shown in the formula:

[0015] R = a * p

[0016] p = s1 x1 *s2 x2 *s3 x3 *s4 x4 *s5 x5 *s6log x6

[0017] Where R represents the driving performance score of a certain vehicle; p represents the probability of an accident occurring every 100,000 kilometers; a represents the output coefficient of the score based on the probability of the accident; for the probability p, s1, s2, s3, s4, s5, and s6 represent the coefficients fitted based on historical vehicle operation and accident records; x1 represents the number of times the vehicle exceeded the speed limit per 100 kilometers; x2 represents the number of times the vehicle accelerated rapidly per 100 kilometers; x3 represents the number of times the vehicle decelerated rapidly per 100 kilometers; x4 represents the number of times the vehicle made a sharp turn per 100 kilometers; x5 represents the percentage of time spent driving while fatigued; and x6 represents the average daily mileage.

[0018] Preferably, the statistical period for the comprehensive driving rating of the driver is the time period required for the assessment. Combined with the labeling of the ride-hailing driver's accidents and violations, the driver's driving behavior is classified by Bayesian network through semi-supervised learning to obtain high-risk, medium-risk, low-risk, and normal vehicles.

[0019] Secondly, a system for monitoring, evaluating, and issuing early warnings for safe driving behavior of drivers is provided, the system comprising:

[0020] Module M1: Collects driving data and processes the driving data;

[0021] Module M2: Establishes a model, including driving behavior scores and driving outcome scores, which are combined to obtain the driver's comprehensive driving rating;

[0022] Module M3: Based on the driver's comprehensive driving rating, different safety assistance and prompts are provided to the driver.

[0023] Preferably, the driving data includes: real-time vehicle data during driving, including basic information on the OBD high-frequency tracking location and alarm information; and driving result data, including traffic violations and insurance claim data during the observation period.

[0024] Preferably, the processing of driving data includes: the OBD device performing edge computing on the driver's driving situation every second, and sending the processed alarm results back to the cloud.

[0025] Preferably, the driving behavior scoring includes: using a supervised learning algorithm to perform linear regression training on a large amount of ride-hailing vehicle trip data collected by OBD, and combining key trip features to construct a model of the probability of an accident per 100,000 kilometers for a single vehicle, thus characterizing the driving behavior and travel risk probability of each driver;

[0026] The driving result score includes: a final determination based on rapid acceleration from 100 km / h, rapid deceleration from 100 km / h, sharp turns from 100 km / h, speeding from 100 km / h, the percentage of time spent driving while fatigued, and the average daily mileage. The formula is as follows:

[0027] R = a * p

[0028] p = s1 x1 *s2 x2 *s3 x3 *s4 x4 *s5 x5 *s6log x6

[0029] Where R represents the driving performance score of a certain vehicle; p represents the probability of an accident occurring every 100,000 kilometers; a represents the output coefficient of the score based on the probability of the accident; for the probability p, s1, s2, s3, s4, s5, and s6 represent the coefficients fitted based on historical vehicle operation and accident records; x1 represents the number of times the vehicle exceeded the speed limit per 100 kilometers; x2 represents the number of times the vehicle accelerated rapidly per 100 kilometers; x3 represents the number of times the vehicle decelerated rapidly per 100 kilometers; x4 represents the number of times the vehicle made a sharp turn per 100 kilometers; x5 represents the percentage of time spent driving while fatigued; and x6 represents the average daily mileage.

[0030] The statistical period for the comprehensive driving rating of drivers is the time period required for the assessment. Combined with the labeling of accidents and violations by ride-hailing drivers, the driving behavior of drivers is classified by Bayesian network through semi-supervised learning to obtain high-risk, medium-risk, low-risk, and normal vehicles.

[0031] Compared with the prior art, the present invention has the following beneficial effects:

[0032] Based on user data collection, this invention analyzes and evaluates the safe driving level of various drivers, effectively identifies individuals with potential driving risks, and implements targeted safety tips and education to reduce the occurrence of traffic accidents.

[0033] Other beneficial effects of the present invention will be explained in detail through the introduction of specific technical features and technical solutions in specific embodiments. Those skilled in the art should be able to understand the beneficial technical effects brought about by these technical features and technical solutions through the introduction of these technical features and technical solutions. Attached Figure Description

[0034] Other features, objects, and advantages of the present invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0035] Figure 1 This is a flowchart of the present invention. Detailed Implementation

[0036] The present invention will now be described in detail with reference to specific embodiments. These embodiments will help those skilled in the art to further understand the present invention, but do not limit the invention in any way. It should be noted that those skilled in the art can make several changes and improvements without departing from the concept of the present invention. These all fall within the scope of protection of the present invention.

[0037] This invention provides a method for monitoring, evaluating, and issuing early warnings about safe driving behavior of drivers, referring to... Figure 1 As shown, the present invention specifically includes:

[0038] Step S1: Collect driving data and process the driving data.

[0039] Data range:

[0040] 1) Real-time vehicle data during driving mainly consists of basic information on the location of high-frequency OBD markers and alarm information. In this embodiment, OBD refers to the on-board diagnostic system, used for collecting driving behavior data in the Internet of Vehicles.

[0041] 2) The driving results data consists of traffic violations and insurance claims data during the observation period.

[0042] Data collection method:

[0043] 1) Basic information of the location: device number, longitude, latitude, device reporting time, resolution time, processing time, GPS speed (km / h), number of satellites, signal strength, altitude (m), direction (0~359, due north is 0, clockwise), total distance (km), power-on flag, whether to retransmit.

[0044] 2) Alarm information: Alarm flags (including equipment power failure alarm, collision alarm, overspeed alarm, low voltage alarm, flameout alarm, ignition alarm, insertion alarm, removal alarm, rapid acceleration alarm, rapid deceleration alarm, and sharp turn alarm).

[0045] 3) Violation details: violation time, violation type, violation content, violation location, violation amount, demerit points, and annual inspection time.

[0046] 4) Insurance claim data: policy number, type of insurance, accident number, report number, report time, time of incident, case type, case status, reporting channel, estimated loss amount, compensation amount, vehicle price, vehicle model, type of liability, whether there was personal injury, location of incident, and details of the accident.

[0047] Data processing: The OBD device performs edge computing on the driver's driving situation every second. The result of the edge computing is the vehicle driving behavior alarm result. The result and other device data are transmitted to the database. Combined with the existing vehicle and accident and violation data in the database, driving behavior is predicted and evaluated.

[0048] Step S2: Model Building

[0049] The model is divided into two parts: the first part is the driving behavior process score, and the second part is the driving result score. Finally, the first part and the second part are combined to obtain the driver's comprehensive driving rating.

[0050] (1) Driving behavior score:

[0051] This scoring section utilizes supervised learning algorithms to train linear regression on a large amount of unlabeled and anonymized trip data from tens of thousands of existing ride-hailing vehicles collected by OBD devices. Combined with key trip features—rapid acceleration, rapid deceleration, emergency braking, sharp turns, speeding, fatigue driving, average daily driving time, and average daily mileage—it constructs a model of the probability of an accident occurring per vehicle every 100,000 kilometers, accurately characterizing each driver's driving behavior and the probability of travel risks.

[0052] (2) Driving result scoring:

[0053] This section of the score is ultimately determined by factors including rapid acceleration from 0 to 100 km / h, rapid deceleration from 0 to 100 km / h, sharp turns from 0 to 100 km / h, speeding at 0 to 100 km / h, the percentage of time spent driving while fatigued, and average daily mileage. The formula is as follows:

[0054] R = a * p

[0055] p = s1 x1 *s2 x2 *s3 x3 *s4 x4 *s5 x5 *s6log x6

[0056] Where R represents the driving performance score of a vehicle; p represents the probability of an accident occurring every 100,000 kilometers; and a represents the scoring coefficient mapped based on this probability of an accident. For the probability p, s1, s2, s3, s4, s5, and s6 represent coefficients fitted based on historical vehicle operation and accident records; x1 represents the number of speeding violations per 100 kilometers (total speeding warnings / total mileage); x2 represents the number of rapid accelerations per 100 kilometers (total rapid acceleration warnings / total mileage); x3 represents the number of rapid decelerations per 100 kilometers (total rapid deceleration warnings / total mileage); x4 represents the number of sharp turns per 100 kilometers (total sharp turn warnings / total mileage); x5 represents the percentage of time spent driving while fatigued (total fatigue driving time / total driving time); and x6 represents the average daily mileage.

[0057] (3) Driver's overall rating:

[0058] The statistical period for this evaluation section is the time frame required for the assessment (generally one year). It combines the labeling of ride-hailing drivers' accidents and violations, calculates a prior probability matrix based on historical driver behavior, constructs a Bayesian network DAG (Directed Acyclic Graph) based on the matrix, and calculates the weight of each variable x until convergence is obtained, yielding the probabilities of each vehicle as high-risk (prone to accidents / violations), medium-risk (prone to accidents), low-risk (prone to violations), and normal. The highest probability value is output as the classification rating.

[0059] Step S3: Based on the driver's overall driving rating, different safety assistance and prompts are provided to the driver, as shown in the table below:

[0060]

[0061]

[0062] Scoring results:

[0063] This model uses quantitative methods to comprehensively evaluate the driving behavior of ride-hailing drivers. Considering the challenges of managing fatigued driving and the need to manage daily mileage in actual ride-hailing operations, it objectively demonstrates the driving risks of drivers. Through regular training by fleet managers, in-vehicle safety electronic audible prompts, speed exceeding 80km / h warnings, and remote power-off functions, it provides a comprehensive safety protection system integrating prompting, supervision, education, and enforcement, thereby reducing the probability of accidents.

[0064] This invention also provides a system for monitoring, evaluating, and issuing early warnings for safe driving behavior. This system can be implemented by executing the steps of the method for monitoring, evaluating, and issuing early warnings for safe driving behavior. That is, those skilled in the art can understand the method for monitoring, evaluating, and issuing early warnings for safe driving behavior as a preferred embodiment of the system. The system specifically includes the following:

[0065] Module M1: Collects driving data and processes the driving data.

[0066] Data range:

[0067] 1) Real-time vehicle data during driving mainly consists of basic information on the location of OBD high-frequency markers and alarm information.

[0068] 2) The driving results data consists of traffic violations and insurance claims data during the observation period.

[0069] Data collection method:

[0070] 1) Basic information of the location: device number, longitude, latitude, device reporting time, resolution time, processing time, GPS speed (km / h), number of satellites, signal strength, altitude (m), direction (0~359, due north is 0, clockwise), total distance (km), power-on flag, whether to retransmit.

[0071] 2) Alarm information: Alarm flags (including equipment power failure alarm, collision alarm, overspeed alarm, low voltage alarm, flameout alarm, ignition alarm, insertion alarm, removal alarm, rapid acceleration alarm, rapid deceleration alarm, and sharp turn alarm).

[0072] 3) Violation details: violation time, violation type, violation content, violation location, violation amount, demerit points, and annual inspection time.

[0073] 4) Insurance claim data: policy number, type of insurance, accident number, report number, report time, time of incident, case type, case status, reporting channel, estimated loss amount, compensation amount, vehicle price, vehicle model, type of liability, whether there was personal injury, location of incident, and details of the accident.

[0074] Data processing: The OBD device performs edge computing on the driver's driving situation every second and sends the processed alarm results back to the cloud.

[0075] Module M2: Model Building:

[0076] The model is divided into two parts: the first part is the driving behavior process score, and the second part is the driving result score. Finally, the first part and the second part are combined to obtain the driver's comprehensive driving rating.

[0077] (1) Driving behavior score:

[0078] This scoring section utilizes a supervised learning algorithm to train linear regression on a large amount of unlabeled and anonymized trip data from tens of thousands of existing ride-hailing vehicles collected via OBD. Combined with key trip features—rapid acceleration, rapid deceleration, emergency braking, sharp turns, speeding, fatigue driving, average daily driving time, and average daily mileage—it constructs a model of the probability of an accident occurring per vehicle every 100,000 kilometers, accurately characterizing each driver's driving behavior and the probability of travel risks.

[0079] (2) Driving result scoring:

[0080] This section of the score is ultimately determined by factors including rapid acceleration from 0 to 100 km / h, rapid deceleration from 0 to 100 km / h, sharp turns from 0 to 100 km / h, speeding at 0 to 100 km / h, the percentage of time spent driving while fatigued, and average daily mileage. The formula is as follows:

[0081] R = a * p

[0082] p = s1 x1 *s2 x2 *s3 x3 *s4 x4 *s5 x5 *s6log x6

[0083] Where R represents the driving behavior score of a vehicle; p represents the probability of an accident occurring every 100,000 kilometers; and a represents the scoring coefficient mapped based on this probability of an accident. For the probability p, s1, s2, s3, s4, s5, and s6 represent coefficients fitted based on historical vehicle operation and accident records; x1 represents the number of speeding violations per 100 kilometers (total speeding warnings / total mileage); x2 represents the number of rapid accelerations per 100 kilometers (total rapid acceleration warnings / total mileage); x3 represents the number of rapid decelerations per 100 kilometers (total rapid deceleration warnings / total mileage); x4 represents the number of sharp turns per 100 kilometers (total sharp turn warnings / total mileage); x5 represents the percentage of time spent driving while fatigued (total fatigue driving time / total driving time); and x6 represents the average daily mileage.

[0084] (3) Driver's overall rating:

[0085] The statistical period for this evaluation is the time period required for assessment (generally one year). Combining the labeling of ride-hailing drivers' accidents and violations, semi-supervised learning is used to classify drivers' driving behavior using a Bayesian network, resulting in high-risk (prone to accidents / violations), medium-risk (prone to accidents), low-risk (prone to violations), and normal vehicles.

[0086] Module M3: Based on the driver's overall driving rating, different safety assistance and prompts are provided to the driver, as shown in the table below:

[0087]

[0088] This invention provides a method and system for monitoring, evaluating, and issuing early warnings about safe driving behavior. Due to the diverse types of car rental users in mobile travel, their differences lie in driving skills, personality traits, and physiological characteristics, ultimately affecting their driving behavior. Based on user data collection, this invention analyzes and evaluates the safe driving skills of various types of drivers, effectively identifying individuals with potential driving risks, and implementing targeted safety tips and education to reduce the occurrence of traffic accidents.

[0089] This invention utilizes high-frequency data acquisition from OBD devices and edge computing to obtain ride-hailing driving behavior data. It integrates insurance claims and traffic violation data, along with overall data from multiple trips within a specific time period, to establish a ride-hailing driving behavior evaluation model based on supervised / semi-supervised accident probability and scoring tags. By combining the driving process and driving results, the model evaluates the driver's behavior and establishes a driving behavior profile for that driver.

[0090] Those skilled in the art will understand that, besides implementing the system and its various devices, modules, and units provided by this invention in the form of purely computer-readable program code, the same functions can be achieved entirely through logical programming of the method steps, making the system and its various devices, modules, and units of this invention function in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers. Therefore, the system and its various devices, modules, and units provided by this invention can be considered as a hardware component, and the devices, modules, and units included therein for implementing various functions can also be considered as structures within the hardware component; alternatively, the devices, modules, and units for implementing various functions can be considered as both software modules implementing the method and structures within the hardware component.

[0091] Specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art can make various changes or modifications within the scope of the claims, which do not affect the essence of the present invention. Unless otherwise specified, the embodiments and features described in this application can be arbitrarily combined with each other.

Claims

1. A method for monitoring, evaluating, and providing early warning of safe driving behavior of drivers, characterized in that, include: Step S1: Collect driving data and process the driving data; Step S2: Establish a model, including driving behavior score and driving result score, and combine the two to obtain the driver's comprehensive driving rating; Step S3: Based on the driver's comprehensive driving rating, provide different safety assistance and prompts to the driver; The driving behavior score includes: using a supervised learning algorithm, linear regression training is performed on a large amount of ride-hailing vehicle trip data collected by OBD, and combined with key trip features, a model is constructed to depict the probability of an accident per 100,000 kilometers for a single vehicle, thus characterizing the driving behavior and travel risk probability of each driver.

2. The method for monitoring, evaluating, and issuing early warnings of driver safe driving behavior according to claim 1, characterized in that, The driving data includes: real-time vehicle data during driving, including basic information on the location of OBD high-frequency markers and alarm information; driving result data includes traffic violations and insurance claim data during the observation period.

3. The method for monitoring, evaluating, and providing early warning of driver safety behavior according to claim 1, characterized in that, The processing of driving data includes: the OBD device performing edge computing on the driver's driving situation every second, and sending the processed alarm results back to the cloud.

4. The method for monitoring, evaluating, and issuing early warnings of driver safe driving behavior according to claim 1, characterized in that, The driving result score includes: a final determination based on rapid acceleration from 100 km / h, rapid deceleration from 100 km / h, sharp turns from 100 km / h, speeding from 100 km / h, the percentage of time spent driving while fatigued, and the average daily mileage. The formula is as follows: R = a * p p=s1 x1 *s2 x2 *s3 x3 *s4 x4 *s5 x5 *s6log x6 Where R represents the driving performance score of a certain vehicle; p represents the probability of an accident occurring every 100,000 kilometers; a represents the output coefficient of the score based on the probability of the accident; for the probability p, s1, s2, s3, s4, s5, and s6 represent the coefficients fitted based on historical vehicle operation and accident records; x1 represents the number of times the vehicle exceeded the speed limit per 100 kilometers; x2 represents the number of times the vehicle accelerated rapidly per 100 kilometers; x3 represents the number of times the vehicle decelerated rapidly per 100 kilometers; x4 represents the number of times the vehicle made a sharp turn per 100 kilometers; x5 represents the percentage of time spent driving while fatigued; and x6 represents the average daily mileage.

5. The method for monitoring, evaluating, and issuing early warnings of driver safe driving behavior according to claim 1, characterized in that, The statistical period for the comprehensive driving rating of drivers is the time period required for the assessment. Combined with the labeling of accidents and violations by ride-hailing drivers, the driving behavior of drivers is classified by Bayesian network through semi-supervised learning to obtain high-risk, medium-risk, low-risk, and normal vehicles.

6. A system for monitoring, evaluating, and providing early warning of safe driving behavior of drivers, characterized in that, include: Module M1: Collects driving data and processes the driving data; Module M2: Establishes a model, including driving behavior scores and driving outcome scores, which are combined to obtain the driver's comprehensive driving rating; Module M3: Based on the driver's comprehensive driving rating, different safety assistance and prompts are provided to the driver; The driving behavior score includes: using a supervised learning algorithm, linear regression training is performed on a large amount of ride-hailing vehicle trip data collected by OBD, and combined with key trip features, a model is constructed to depict the probability of an accident per 100,000 kilometers for a single vehicle, thus characterizing the driving behavior and travel risk probability of each driver.

7. The monitoring, evaluation, and early warning system for driver safe driving behavior according to claim 6, characterized in that, The driving data includes: real-time vehicle data during driving, including basic information on the OBD high-frequency tracking location and alarm information; driving result data includes traffic violations and insurance claim data during the observation period.

8. The monitoring, evaluation, and early warning system for driver safe driving behavior according to claim 6, characterized in that, The processing of driving data includes: the OBD device performing edge computing on the driver's driving situation every second, and sending the processed alarm results back to the cloud.

9. The monitoring, evaluation, and early warning system for driver safe driving behavior according to claim 6, characterized in that, The driving result score includes: a final determination based on rapid acceleration from 100 km / h, rapid deceleration from 100 km / h, sharp turns from 100 km / h, speeding from 100 km / h, the percentage of time spent driving while fatigued, and the average daily mileage. The formula is as follows: R = a * p p=s1 x1 *s2 x2 *s3 x3 *s4 x4 *s5 x5 *s6log x6 Where R represents the driving performance score of a vehicle; p represents the probability of an accident occurring every 100,000 kilometers; a represents the output coefficient of the score mapped based on the probability of the accident; for the probability p, s1, s2, s3, s4, s5, and s6 represent the coefficients fitted based on historical vehicle operation and accident records; x1 represents the number of times the vehicle exceeded the speed limit per 100 kilometers; x2 represents the number of times the vehicle accelerated rapidly per 100 kilometers; x3 represents the number of times the vehicle decelerated rapidly per 100 kilometers; x4 represents the number of times the vehicle made sharp turns per 100 kilometers; x5 represents the percentage of time spent driving while fatigued; and x6 represents the average daily mileage. The statistical period for the comprehensive driving rating of drivers is the time period required for the assessment. Combined with the labeling of accidents and violations by ride-hailing drivers, the driving behavior of drivers is classified by Bayesian network through semi-supervised learning to obtain high-risk, medium-risk, low-risk, and normal vehicles.

Images