Vehicle transportation audit scoring method and system

Abstract

The present invention relates to a vehicle transportation audit and scoring method and system. The vehicle transportation audit and scoring system includes a cloud server for performing intersection analysis, speed analysis, and smoothness analysis to generate key features such as the intersection turning ratio, the number of intersection turning ratios without stopping, the average minimum turning speed, the duration of sudden deceleration, the duration of sudden acceleration, the duration of speeding, the speeding ratio, the average maximum speeding, the proportion of heavy throttle, the number of vehicle body forward tilts, the number of vehicle body backward tilts, the number of vehicle body side tilts, and the number of vehicle body vibrations. A supervised learning scoring process is then executed based on the key features to generate an audit score. This system provides automobile transportation operators with a way to manage drivers, avoid subjective human judgment, and audit and evaluate drivers' driving behavior in an objective and efficient manner.

Description

Technical Field

[0001] The present invention relates to a scoring method and system, and in particular to a vehicle transportation audit scoring method and system. Background Art

[0002] In recent years, many traffic accidents have occurred in the automotive industry due to poor driving behavior of vehicles such as buses, passenger buses, and tour buses. For example, when turning at an intersection, if the driver does not stop to check whether there are other pedestrians or vehicles passing, it is easy to cause a traffic accident.

[0003] While most vehicles are currently equipped with dashcams, they only record the vehicle's driving behavior and cannot record the actual driving behavior of the drivers involved in the transportation industry, making it difficult to evaluate the drivers' performance. Furthermore, while some transportation companies have incentive mechanisms, such as asking passengers to complete questionnaires to rate drivers' driving behavior, solely based on human ratings is overly subjective, and subsequent manual audits are labor-intensive and time-consuming. Consequently, there is currently a lack of effective mechanisms for evaluating drivers' driving behavior. Consequently, existing mechanisms for evaluating drivers' driving behavior require further improvement. Summary of the Invention

[0004] In view of the above problems, the present invention provides a vehicle transportation audit scoring method and system to assist automobile transportation companies or intelligent driving developers in developing an automated audit scoring system.

[0005] The vehicle transportation audit scoring method is executed by a cloud server and includes the following steps: performing intersection analysis to generate the intersection turning ratio, the number of intersection turning ratios without stopping, and the average minimum turning speed; performing speed analysis to generate the rapid deceleration duration, the rapid acceleration duration, the speeding duration, the speeding ratio, and the average maximum speeding speed; performing smoothness analysis to generate the throttle overload ratio, the number of vehicle body forward tilts, the number of vehicle body backward tilts, the number of vehicle body side tilts, and the number of vehicle body vibrations; performing a supervised learning scoring procedure based on the intersection turning ratio without stopping, the number of intersection turning ratios without stopping, the average minimum turning speed, the rapid deceleration duration, the rapid acceleration duration, the speeding duration, the speeding ratio, the average maximum speeding speed, the throttle overload ratio, the number of vehicle body forward tilts, the number of vehicle body backward tilts, the number of vehicle body side tilts, and the number of vehicle body vibrations to generate an audit score.

[0006] In addition, the vehicle transportation audit and scoring system includes a cloud server for performing intersection analysis, speed analysis, and smoothness analysis. Intersection analysis is used to generate the intersection turn-without-stop ratio, the number of intersection turn-without-stop times, and the average minimum turning speed. Speed analysis is used to generate the duration of sudden deceleration, the duration of sudden acceleration, the duration of speeding, the speeding ratio, and the average maximum speeding speed. Smoothness analysis is used to generate the throttle overload ratio, the number of vehicle body forward tilts, the number of vehicle body backward tilts, the number of vehicle body side tilts, and the number of vehicle body vibrations. The cloud server then performs a supervised learning scoring process based on the intersection turn-without-stop ratio, the number of intersection turn-without-stop times, the average minimum turning speed, the duration of sudden deceleration, the duration of sudden acceleration, the speeding ratio, the average maximum speeding speed, the throttle overload ratio, the number of vehicle body forward tilts, the number of vehicle body backward tilts, the number of vehicle body side tilts, and the number of vehicle body vibrations to generate an audit score.

[0007] The present invention uses a cloud server to obtain vehicle body information, inertial sensor information, and positioning information from onboard devices. This data is then aggregated and analyzed. This allows for an immediate and automatic evaluation of the driver's driving behavior, which can then be fed back to the driver to prompt him or her to improve his or her driving behavior. Furthermore, the results of the analysis and statistics can be stored and provided to transportation companies for driver management and to developers for scoring reference in improving intelligent driving behavior.

[0008] Since the present invention directly obtains information from the vehicle-mounted device and then audits and scores it through the cloud server, it can avoid human subjective judgment and audit and evaluate the driver's driving behavior in an objective and efficient manner. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1 1 is a flow chart of the first embodiment of the vehicle transportation audit scoring method of the present invention.

[0010] Figure 2 2 is a block diagram of the vehicle transportation audit and scoring system of the present invention.

[0011] Figure 3 It is a flow chart of the intersection analysis of the vehicle transportation audit scoring method of the present invention.

[0012] Figure 4 It is a flow chart of the speed analysis of the vehicle transportation audit scoring method of the present invention.

[0013] Figure 5 It is a flow chart of the smoothness analysis of the vehicle transportation audit scoring method of the present invention.

[0014] Figure 6 2. It is a flow chart of the first embodiment of the supervised learning scoring procedure of the vehicle transportation audit scoring method of the present invention.

[0015] Figure 7 2 is a flow chart of the second embodiment of the vehicle transportation audit scoring method of the present invention.

[0016] Figure 8 It is a flow chart of the transit analysis of the vehicle transportation audit scoring method of the present invention.

[0017] Figure 9 2 is a flow chart of the second embodiment of the supervised learning scoring procedure of the vehicle transportation audit scoring method of the present invention. DETAILED DESCRIPTION

[0018] See also Figure 1 and Figure 2 The dynamic resource adjustment method for a distributed data transmission fault-tolerant system of the present invention is executed by a cloud server 10. The distributed data transmission fault-tolerant system includes the cloud server 10 and an on-vehicle device 20. The on-vehicle device 20 is mounted on a vehicle and can sense the vehicle's status and generate various information. In this embodiment, the on-vehicle device 20 includes a positioning unit 21, a vehicle information acquisition unit 22, an inertial measurement unit 23, and a wireless transceiver unit 24.

[0019] The positioning unit 21 generates vehicle location information based on the vehicle's positioning status. For example, the positioning unit 21 is a global positioning system (GPS) unit. The vehicle information acquisition unit 22 measures the vehicle's status to generate speed information, braking information, and throttle and door information. For example, the vehicle information acquisition unit 22 is an on-board diagnostics system (OBD) that senses the vehicle's current speed, throttle depth, brake level, and door status, and generates corresponding related information. The inertial measurement unit 23 measures the vehicle's inertial status to generate inertial measurement information. For example, the inertial measurement information includes angular velocity information and acceleration information. The wireless transceiver unit 24 is connected to the positioning unit 21, the vehicle information acquisition unit 22, and the inertial measurement unit 23 to receive and transmit the vehicle's location information, speed information, braking information, throttle information, door information, and inertial measurement information to the cloud server 10.

[0020] The first embodiment of the vehicle transportation audit scoring method includes the following steps:

[0021] Step S101: The cloud server 10 performs intersection analysis to generate a ratio of intersection turns without stopping, a number of intersection turns without stopping, and an average minimum cornering speed.

[0022] Please also refer to Figure 3As shown, when the cloud server 10 performs intersection analysis, it communicates with the wireless transceiver unit 24 of the vehicle-mounted device 20, as shown in step S301, and the cloud server 10 obtains vehicle location information, speed information, and braking information through the wireless transceiver unit 24. As shown in step S302, the cloud server 10 obtains traffic light location information from the driving information database 30. As shown in step S303, the cloud server 10 determines whether at least a first distance between the vehicle location and at least one traffic light location is less than the intersection threshold distance based on the vehicle location information and the traffic light location information. As shown in step S304, when the first distance is less than the intersection threshold distance, the cloud server 10 generates the number of intersection turns and generates a minimum turning speed based on the speed information. As shown in step S305, the cloud server 10 determines whether the waiting time exceeds the waiting threshold based on the braking information. When the waiting time exceeds the waiting threshold, as shown in step S306, the cloud server 10 generates the number of intersection waiting times. Then, as shown in step S307, the cloud server 10 determines whether the audit conditions are met. If the audit conditions are met, as shown in step S308, the cloud server 10 calculates the intersection turning rate and the number of intersection turning times without stopping during the audit period based on the number of intersection turning times and the number of intersection waiting times without stopping. Based on the minimum turning speed, the cloud server 10 calculates the average minimum turning speed during the audit period. If the audit conditions in step S307 are not met, the cloud server 10 re-acquires the vehicle's location, speed, and braking information via the wireless transceiver unit 24.

[0023] For example, the cloud server 10 confirms the current location of the vehicle by receiving the vehicle location information sent by the vehicle-mounted device 20, and compares it with the signal location information obtained from the driving information database 30. When the distance between the vehicle location and the signal location is less than a first distance, it means that the vehicle's current location is close enough to the signal location. Generally speaking, signal signs are set at intersections, so when the vehicle location is close enough to the signal location, it means that the vehicle is currently at the intersection. Because when a vehicle passes through an intersection, it is necessary to slow down to maintain driving safety. If the cloud server 10 determines based on the braking information that the vehicle does not slow down or stop at the intersection, it means that the driver's driving behavior is poor. Therefore, the cloud server 10 uses the number of times the vehicle does not stop at the intersection as one of the audit criteria.

[0024] Furthermore, when the vehicle is at an intersection, i.e., when the first distance is less than the intersection threshold distance, the vehicle position information can be used to determine whether the vehicle has turned. If so, the number of turns at the intersection is counted; if not, the number of turns is not counted. Furthermore, the cloud server 10 records the vehicle's lowest turning speed based on the vehicle speed information.

[0025] When the audit conditions are met, such as completing a trip from the starting point to the end point, the cloud server 10 can calculate the intersection turn percentage, the number of intersection turn percentages, and the average minimum turning speed based on the total number of turns made by the vehicle from the starting point to the end point, whether any stops or waits were made during turns, and the minimum turning speed for each turn. For example, the intersection turn percentage is the number of turns without stops or waits divided by the number of turns made at the intersection, and the average minimum turning speed is the average of the minimum turning speeds for each turn made from the starting point to the end point.

[0026] Intersection analysis uses vehicle position, speed, and braking information to determine whether a vehicle should stop before turning at an intersection to avoid traffic accidents.

[0027] Step S102: The cloud server 10 performs speed analysis to generate rapid deceleration duration, rapid acceleration duration, speeding duration, speeding ratio, and average maximum speeding speed.

[0028] Please also refer to Figure 4As shown, when the cloud server 10 performs speed analysis, it communicates with the wireless transceiver unit 24 of the onboard device 20. As shown in step S401, the cloud server 10 obtains vehicle location and speed information via the wireless transceiver unit 24. As shown in step S402, the cloud server 10 obtains route speed limit information from the driving information database 30. As shown in step S403, the cloud server 10 generates location speed limit information based on the vehicle location and route speed limit information. As shown in step S404, the cloud server 10 further determines whether the vehicle speed exceeds the location speed limit based on the speed information and location speed limit information. If the vehicle speed exceeds the location speed limit, as shown in step S405, the cloud server 10 calculates the speeding duration and generates the maximum speed exceeding the location speed based on the speed information. As shown in step S406, the cloud server 10 determines whether the audit condition is met. As shown in step S407, the cloud server 10 further determines whether the first difference between the maximum and minimum speed values of the vehicle during the acceleration time is greater than a first threshold value based on the speed information. When the first difference is greater than the first threshold value, as shown in step S408, the cloud server 10 generates a rapid acceleration duration and determines whether the audit condition is met. In addition, as shown in step S409, the cloud server 10 further determines whether the second difference between the maximum and minimum vehicle speeds during the deceleration time is greater than the second threshold value based on the speed information. When the second difference is greater than the second threshold value, as shown in step S410, the cloud server 10 generates a rapid deceleration duration and determines whether the audit condition is met. When the audit condition is met, as shown in step S411, the cloud server 10 calculates the speeding ratio within the audit duration based on the speeding duration, and calculates the average maximum speeding speed within the audit duration based on the maximum speeding speed. However, when the audit condition is not met, the cloud server 10 re-acquires the vehicle location information and speed information through the wireless transceiver unit 24.

[0029] For example, the cloud server 10 confirms the vehicle's current location by receiving vehicle location information transmitted by the onboard device 20, and confirms the vehicle's current speed by using speed information. The cloud server 10 also obtains route speed limit information from the driving information database 30. Because different driving routes have different speed limits, the cloud server 10 first confirms the vehicle's current location based on the vehicle location information and then determines the local speed limit based on the vehicle's current location. For example, if the vehicle is located in an urban area, the speed limit is 50 km / h, but if the vehicle is located on a highway, the speed limit is 100 km / h. After confirming the location speed limit information, the cloud server 10 can determine whether the vehicle is speeding based on the speed information and the location speed limit information. If the vehicle's speed exceeds the local speed limit, the cloud server 10 determines that the vehicle has exceeded the speed limit. The cloud server 10 then further calculates the length of time the vehicle has exceeded the speed limit as the speeding duration. If the vehicle exceeds the speed limit, the cloud server 10 also records the highest speed at which the vehicle exceeded the speed limit.

[0030] When audit conditions are met, such as when the vehicle travels from the starting point to the end point, the cloud server 10 can calculate the average maximum speeding speed based on the total number of speeding incidents between the starting point and the end point, as well as the maximum speeding speed during each incident. Furthermore, the audit duration, for example, is the time elapsed from the starting point to the end point after the vehicle departs. The cloud server 10 calculates the speeding ratio by dividing the speeding duration by the audit duration. The average maximum speeding speed is the sum of the maximum speeding speeds during each incident divided by the number of speeding incidents.

[0031] Furthermore, after receiving the vehicle's location and speed information, the cloud server 10 further determines whether the vehicle has experienced sudden acceleration or deceleration based on the speed information. If, during the acceleration period, the first difference between the maximum and minimum vehicle speeds exceeds a first threshold, this indicates a sudden acceleration event. The cloud server 10 then generates a sudden acceleration duration based on the time difference corresponding to the first difference. If the audit conditions are met, the cloud server 10 accumulates all sudden acceleration durations within the audit period.

[0032] Similarly, when the second difference is greater than the second threshold, it indicates that the vehicle has suddenly decelerated. Therefore, the cloud server 10 also generates a sudden deceleration duration and accumulates all sudden deceleration durations within the audit duration when the audit conditions are met.

[0033] Speed analysis derives the speeding, sudden acceleration, sudden deceleration conditions and the location of the incident based on the vehicle's position and speed information.

[0034] Step S103 : The cloud server 10 performs a ride analysis to generate a throttle overload ratio, a vehicle body forward tilt frequency, a vehicle body backward tilt frequency, a vehicle body side roll frequency, and a vehicle body vibration frequency.

[0035] Please also refer to Figure 5 As shown, when the cloud server 10 performs a ride analysis, it communicates with the wireless transceiver unit 24 of the vehicle-mounted device 20. As shown in step S501, the cloud server 10 obtains inertial measurement information and vehicle throttle information via the wireless transceiver unit 24. As shown in step S502, the cloud server 10 determines whether the throttle depth is greater than the threshold depth based on the vehicle throttle information. As shown in step S503, if the throttle depth is greater than the threshold depth, the cloud server 10 generates a throttle-heavy duration. As shown in step S504, the cloud server 10 determines whether the audit conditions are met. In addition, as shown in step S505, when the cloud server 10 obtains inertial measurement information and vehicle throttle information, the cloud server 10 further determines whether the first absolute value of the first axial acceleration is greater than the first threshold acceleration based on the inertial measurement information, and as shown in step S506, the cloud server 10 determines whether the second absolute value of the second axial acceleration is greater than the second threshold acceleration based on the inertial measurement information, and as shown in step S507, the cloud server 10 also determines whether the third absolute value of the third axial acceleration is greater than a third threshold acceleration based on the inertial measurement information.

[0036] When the first absolute value is greater than the first threshold acceleration, as shown in step S508, and when the first axial acceleration is a positive number, the cloud server 10 generates the number of vehicle body tilting backward, and when the first axial acceleration is a negative number, the cloud server 10 generates the number of vehicle body tilting forward, and then as shown in step S504, the cloud server determines whether the audit conditions are met.

[0037] When the second absolute value is greater than the second threshold acceleration, as shown in step S509 , the cloud server generates the vehicle body roll count, and as shown in step S504 , the cloud server 10 determines whether the audit condition is met.

[0038] When the third absolute value is greater than the third threshold acceleration, as shown in step S510 , the cloud server generates a vehicle body vibration count, and as shown in step S504 , the cloud server 10 determines whether the audit condition is met.

[0039] As shown in step S511, if the audit conditions are met, the cloud server 10 calculates the throttle overload ratio during the audit period based on the duration of the throttle overload, and calculates the average maximum speed over the audit period based on the maximum speeding speed. However, if the audit conditions are not met, the cloud server 10 re-acquires the inertial measurement information and vehicle throttle information via the wireless transceiver unit 24.

[0040] For example, the cloud server 10 receives throttle information from the vehicle-mounted device 20 to determine the current throttle depth of the driver's throttle application. If the throttle depth is greater than the threshold depth, indicating that the driver is applying excessive throttle pressure, the cloud server 10 generates an excessive throttle application duration to record the time the driver applied excessive throttle pressure. If the audit conditions are met, the cloud server 10 records the total time the driver applied excessive throttle pressure during the audit period based on the excessive throttle application duration and calculates the excessive throttle application ratio by dividing the excessive throttle application duration by the audit period.

[0041] In addition, the cloud server 10 receives inertial measurement information transmitted by the vehicle-mounted device 20 and determines the smoothness of the vehicle's driving based on the acceleration of the inertial measurement information. For example, the first axis represents the vehicle's forward direction, the second axis represents the vehicle's lateral direction, and the third axis represents the vehicle's vertical direction. When the first absolute value of the first axial acceleration is greater than the first threshold acceleration, it indicates that the vehicle is moving forward or backward rapidly, causing the vehicle body to tilt forward or backward. The cloud server 10 further determines whether the vehicle body is tilting forward or backward based on the positive or negative value of the first axial acceleration. For example, when the first axial acceleration is positive, it indicates that the vehicle is moving forward rapidly, causing the vehicle body to tilt backward; conversely, it indicates that the vehicle body will tilt forward. When the second absolute value of the second axial acceleration is greater than the second threshold acceleration, it indicates that the vehicle is turning left or right rapidly, causing the vehicle body to tilt. When the third absolute value of the third axial acceleration is greater than the third threshold acceleration, it indicates that the vehicle is vibrating violently up and down.

[0042] Therefore, whenever the first absolute value exceeds the first acceleration threshold, the cloud server 10 determines whether the vehicle is tilting forward or backward based on the positive or negative value of the first-axis acceleration, and accumulates the number of forward and backward tilts. Similarly, whenever the second absolute value exceeds the second acceleration threshold, the cloud server 10 accumulates the number of vehicle rolls, and whenever the third absolute value exceeds the third acceleration threshold, the cloud server 10 accumulates the number of vehicle vibrations.

[0043] When the audit conditions are met, the cloud server 10 accumulates the number of vehicle body forward tilts, vehicle body backward tilts, vehicle body side tilts, and vehicle body vibrations within the audit period.

[0044] In this embodiment, the first threshold acceleration is the square of the first axial acceleration multiplied by a dynamic adjustment parameter, the second threshold acceleration is the square of the second axial acceleration multiplied by the dynamic adjustment parameter, and the third threshold acceleration is the square of the third axial acceleration multiplied by the dynamic adjustment parameter. The dynamic adjustment parameter depends on the type of vehicle; for example, the dynamic adjustment parameter for a large bus is different from that for a small bus.

[0045] Furthermore, in other embodiments, the audit duration can be a fixed time length, such as 24 hours, and the cloud server 10 further determines whether any of the first analysis duration for performing the intersection analysis, the second analysis duration for performing the speed analysis, or the third analysis duration for performing the smoothness analysis exceeds the audit duration. When any of the first analysis duration, the second analysis duration, or the third analysis duration exceeds the audit duration, the cloud server 10 determines that the audit condition is satisfied. For example, from 12:00 noon on the first day to 1:00 p.m. on the second day, the cloud server 10 has performed any of the intersection analysis, speed analysis, or smoothness analysis for 25 hours, exceeding the audit duration of 24 hours, and the cloud server 10 determines that the audit condition is satisfied.

[0046] Smoothness analysis uses vehicle throttle information to determine the driver's habit of pressing the accelerator, whether he often presses the accelerator hard, and uses inertial measurement information to distinguish the vehicle's backward tilt, forward tilt, side tilt, and vibration to analyze whether the ride experience is good.

[0047] Step S104: The cloud server 10 executes a supervised learning scoring procedure to generate an audit score based on the ratio of non-stop turning at the intersection, the number of non-stop turning at the intersection, the average minimum turning speed, the duration of sudden deceleration, the duration of sudden acceleration, the duration of speeding, the speeding ratio, the average maximum speeding speed, the ratio of excessive throttle, the number of vehicle body forward tilts, the number of vehicle body backward tilts, the number of vehicle body side tilts, and the number of vehicle body vibrations.

[0048] Please also refer to Figure 6 As shown in step S601, the first embodiment of the supervised learning scoring program is that the cloud server 10 generates an intersection analysis score using an intersection machine learning model based on the intersection turning ratio, the intersection turning number and the average minimum turning speed. And as shown in step S602, the cloud server 10 generates a speed analysis score using a speed machine learning model based on the sudden deceleration duration, the sudden acceleration duration, the speeding duration, the speeding ratio and the average maximum speeding speed. And as shown in step S603, the cloud server 10 generates a smoothness analysis score using a smoothness machine learning model based on the throttle overload ratio, the number of vehicle body forward tilts, the number of vehicle body backward tilts, the number of vehicle body rolls and the number of vehicle body vibrations. As shown in step S604, the cloud server 10 performs weighted summation based on the intersection analysis score, the speed analysis score and the smoothness analysis score to generate an audit score.

[0049] For example, the intersection analysis machine learning model, the speeding analysis machine learning model, and the ride smoothness machine learning model are neural network models using supervised learning or multivariate regression. The present invention uses the intersection turn-stop ratio, the number of intersection turn-stops, and the average minimum cornering speed as key features for the intersection analysis machine learning model; the duration of sudden deceleration, the duration of sudden acceleration, the duration of speeding, the speeding ratio, and the average maximum speeding as key features for the speeding analysis machine learning model; and the throttle overload ratio, the number of vehicle forward tilts, the number of vehicle backward tilts, the number of vehicle rolls, and the number of vehicle vibrations as key features for the ride smoothness machine learning model. These models are trained to generate intersection analysis scores, speed analysis scores, and ride smoothness analysis scores.

[0050] The audit score is calculated as follows:

[0051] Audit score = aw a +bw b +cw c ;

[0052] Among them, a, b, and c represent the intersection analysis score, speed analysis score, and smoothness analysis score respectively, and w a 、w b 、w c These represent the weights for intersection analysis, speed analysis, and ride safety analysis, respectively. Weights can be calculated based on the number of traffic incidents and complaints, or the weights of key audit items. For example, the intersection analysis weight can be determined based on the number of accidents caused by failure to yield or improper turns. The speed analysis weight can be determined based on the number of accidents caused by speeding or failure to maintain safe distance. The ride safety analysis weight can be determined based on the number of accidents caused by sudden braking and the number of complaints caused by improper driving.

[0053] For further information, see Figure 1 and Figure 7 As shown, steps S701, S702, and S703 of the second embodiment of the vehicle transportation audit scoring method are the same as steps S101, S102, and S103 of the first embodiment, and are not repeated here. The difference between the second embodiment of the vehicle transportation audit scoring method and the first embodiment is that the second embodiment further includes the following steps:

[0054] Step S704: The cloud server 10 performs a stop-over analysis to generate a stop-over ratio, a stop-over number, and an average minimum stop-over speed.

[0055] Please also refer to Figure 8As shown, when the cloud server 10 performs a stop-pass analysis, it communicates with the wireless transceiver unit 24 of the onboard device 20, as shown in step S801. The cloud server 10 obtains vehicle location information, speed information, brake information, and door information via the wireless transceiver unit 24. As shown in step S802, the cloud server 10 obtains stop location information from the driving information database 30. As shown in step S803, the cloud server 10 determines whether at least a second distance between the vehicle location and at least one stop location is less than a stop threshold distance based on the vehicle location information and the stop location information. As shown in step S804, if the second distance is less than the stop threshold distance, the cloud server 10 generates a stop-pass count and generates a minimum stop-pass speed based on the speed information. As shown in step S805, the cloud server 10 determines whether to stop and open the door based on the brake information and door information. If the vehicle does not stop and the door is open, as shown in step S806, the cloud server 10 generates the number of stops without door openings. In step S807, the cloud server 10 determines whether the audit conditions are met. If the audit conditions are met, the cloud server 10 calculates the percentage of stops without stopping and the number of stops without door openings during the audit period based on the number of stops passed and the number of stops without door openings. Based on the minimum stop speed, the cloud server 10 calculates the average minimum stop speed during the audit period. If the audit conditions are not met, the cloud server 10 re-acquires the vehicle's location information, speed information, brake information, and door information via the wireless transceiver unit.

[0056] For example, the cloud server 10 confirms the current location of the vehicle by receiving the vehicle location information sent by the vehicle-mounted device 20, and compares it with the station location information obtained from the driving information database 30. When the distance between the vehicle location and the station location is less than the second distance, it means that the current location of the vehicle is close enough to the location of the station. Generally speaking, when the vehicle location is close enough to the station location, it means that the vehicle is about to enter the station. Because when a vehicle passes a station, it must slow down and stop at the station to allow passengers to get on and off. If the cloud server 10 determines that the vehicle does not stop at the station location and opens the door based on the brake information and the door information, it means that the driver's driving behavior is poor. Therefore, the cloud server 10 also uses the number of times the vehicle passes the station without stopping as one of the audit criteria.

[0057] When the audit conditions are met, the cloud server 10 can calculate the non-stop stop ratio, the number of non-stop stop ratios, and the average minimum stop speed based on the total number of stops the vehicle passed during the audit period, and whether the vehicle stopped and waited at each stop and opened its doors. For example, the non-stop stop ratio is the number of non-stop stop ratios divided by the total number of stops passed during the audit period, and the average minimum stop speed is the average of the minimum stop speeds for each stop during the audit period.

[0058] The stop-over analysis uses vehicle location information and station location information to analyze whether the passenger vehicle passes through the station without stopping when it arrives at the station.

[0059] Step S705: When the cloud server 10 executes the supervised learning scoring program, the cloud server 10 generates an audit score based on the proportion of intersection turns without stopping, the number of intersection turns without stopping, the average minimum cornering speed, the duration of sudden deceleration, the duration of sudden acceleration, the duration of speeding, the speeding proportion, the average maximum speeding speed, the proportion of excessive throttle, the number of vehicle forward tilts, the number of vehicle backward tilts, the number of vehicle side tilts, the number of vehicle vibrations, the proportion of stop-stops, the number of stop-stops and the average minimum stop-stop speed.

[0060] Please also refer to Figure 6 and Figure 9 Steps S901, S902, and S903 of the second embodiment of the supervised learning scoring process are the same as steps S601, S602, and S603 of the first embodiment, and are not described here in detail. The difference between the second embodiment of the supervised learning scoring process and the first embodiment lies in the following steps:

[0061] Step S904: The cloud server 10 generates a stop-stop analysis score using a stop-stop machine learning model based on the stop-stop ratio, the number of stop-stop times, and the average minimum stop-stop speed.

[0062] Step S905: The cloud server performs weighted summation based on the intersection analysis score, speed analysis score, smoothness analysis score, and stop-over analysis score to generate the audit score.

[0063] In this embodiment, a stopover analysis is further added, and the machine learning for this analysis utilizes a supervised neural network model or a supervised multivariate regression model. Similarly, the present invention utilizes the stopover percentage, the number of stopovers without stopping, and the average minimum stopover speed as key technologies for the machine learning model to generate a stopover score.

[0064] The audit score is calculated as follows:

[0065] Audit score = aw a +bw b +cw c +dw d ;

[0066] Among them, d represents the site analysis score, w d The weight of the stop-by analysis is determined based on the number of complaints: the reason for the stop-by analysis is the number of stops not stopped or the number of stops not made according to regulations.

[0067] In summary, the present invention provides an automated evaluation method for driving behavior. In addition to transmitting the results to the passenger transport operator for audit, allowing the passenger transport operator to audit the vehicle driver, the results can also be displayed on the head-up display installed in the vehicle for the current score and parity for the vehicle driver's reference, which is used to remind the vehicle driver and allow the driver to confirm whether his driving behavior is good or not.

[0068] Furthermore, the present invention establishes four types of analysis, primarily focusing on the dangers of on-road passenger transport behavior and providing appropriate analysis of passenger perception. The results of the four analyses are then classified using a separate method to determine the score for each analysis. After weighted summation, a final score and evaluation are given, allowing passenger transport operators to objectively and efficiently audit and evaluate drivers' driving behavior.

[0069] The above description is merely a preferred embodiment of the present invention and does not constitute any form of limitation to the present invention. Although the present invention has been disclosed as a preferred embodiment as above, it is not intended to limit the present invention. Any person skilled in the art can make some changes or modifications to equivalent embodiments using the technical contents disclosed above without departing from the scope of the technical solution of the present invention. However, any simple modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention are still within the scope of the technical solution of the present invention.

Claims

1. A vehicle transportation audit scoring method, characterized in that: Executed by a cloud server and includes: Performing an intersection analysis to generate a non-stop cornering ratio, a non-stop cornering frequency, and an average minimum cornering speed; Performing a speed analysis to generate a sudden deceleration duration, a sudden acceleration duration, an overspeeding duration, an overspeeding ratio, and an average maximum overspeeding speed; Performing a ride analysis to generate a throttle overweight ratio, a vehicle body forward tilt number, a vehicle body backward tilt number, a vehicle body side roll number, and a vehicle body vibration number; executing a supervised learning scoring process to generate an audit score based on the uninterrupted cornering ratio, the number of uninterrupted cornering times at the intersection, the average minimum cornering speed, the rapid deceleration duration, the rapid acceleration duration, the speeding duration, the speeding ratio, the average maximum speeding speed, the throttle overload ratio, the number of vehicle body forward tilts, the number of vehicle body rearward tilts, the number of vehicle body side tilts, and the number of vehicle body vibrations; The supervised learning scoring procedure includes: Generating an intersection analysis score using an intersection machine learning model based on the intersection turning rate without stopping, the number of turning without stopping, and the average minimum turning speed; generating a speed analysis score using a speed machine learning model according to the rapid deceleration duration, the rapid acceleration duration, the speeding duration, the speeding ratio, and the average maximum speeding speed; generating a ride analysis score using a ride machine learning model based on the throttle overweight ratio, the number of vehicle body forward tilts, the number of vehicle body rearward tilts, the number of vehicle body side tilts, and the number of vehicle body vibrations; The audit score is generated by performing weighted summation based on the intersection analysis score, the speed analysis score, and the smoothness analysis score.

2. The vehicle transportation audit scoring method according to claim 1 is characterized in that: The intersection analysis includes: Obtaining vehicle position information, speed information, and braking information; Get the No. 1 location information; Determining, based on the vehicle position information and the signal position information, whether at least a first distance between a vehicle position and at least one signal position is less than an intersection threshold distance; When the at least one first distance is less than the intersection threshold distance, generating an intersection turning number, generating a minimum turning speed based on the speed information, and determining whether the waiting time exceeds a waiting threshold based on the braking information; When the waiting time exceeds the waiting threshold, the number of times the vehicle does not stop at an intersection is generated, and it is determined whether an audit condition is met; When the audit condition is met, the non-stop ratio of intersection turns and the number of intersection turns within an audit period are calculated based on the number of intersection turns and the number of times the driver did not stop, and the average minimum cornering speed within the audit period is calculated based on the minimum cornering speed; When the audit condition is not satisfied, the vehicle position information, the speed information, and the braking information are re-acquired.

3. The vehicle transportation audit scoring method according to claim 1, characterized in that: The speed analysis includes: Obtaining vehicle position information and speed information; Get the speed limit information of a route; generating location speed limit information according to the vehicle location information and the route speed limit information; Determining whether a vehicle speed is greater than a speed limit at a location based on the speed information and the location speed limit information; When the vehicle speed exceeds the speed limit at the location, the speeding duration is calculated, a maximum speeding speed is generated based on the speed information, and whether an audit condition is met is determined; determining, based on the speed information, whether a first difference between a maximum value and a minimum value of the vehicle speed within an acceleration time is greater than a first threshold; When the first difference is greater than the first threshold, the rapid acceleration duration is generated, and it is determined whether the audit condition is met; determining, based on the speed information, whether a second difference between a maximum value and a minimum value of the vehicle speed within a deceleration time is greater than a second threshold; When the second difference is greater than the second threshold, the rapid deceleration duration is generated, and it is determined whether the audit condition is met; When the audit condition is met, the speeding ratio within an audit period is calculated based on the speeding duration, and the average maximum speeding speed within the audit period is calculated based on the maximum speeding speed; When the audit condition is not satisfied, the vehicle position information and the speed information are reacquired.

4. The vehicle transportation audit scoring method according to claim 1, characterized in that: The smooth analysis includes: Obtaining inertial measurement information and vehicle throttle information; determining, based on the vehicle throttle information, whether a throttle depth is greater than a threshold depth; When the throttle depth is greater than the threshold depth, a throttle-excessive-duration is generated, and it is determined whether an audit condition is met; Determining, based on the inertial measurement information, whether a first absolute value of a first axial acceleration is greater than a first threshold acceleration, determining whether a second absolute value of a second axial acceleration is greater than a second threshold acceleration, and determining whether a third absolute value of a third axial acceleration is greater than a third threshold acceleration; When the first absolute value of the first axial acceleration is greater than the first threshold acceleration and the first axial acceleration is a positive number, the vehicle body backward tilt count is generated, and whether the audit condition is satisfied is determined; When the first absolute value of the first axial acceleration is greater than the first threshold acceleration and when the first axial acceleration is a negative number, the vehicle body forward tilt count is generated, and it is determined whether the audit condition is met; When the second absolute value of the second axial acceleration is greater than the second threshold acceleration, the vehicle body roll count is generated, and it is determined whether the audit condition is met; When the third absolute value of the third axial acceleration is greater than the third threshold acceleration, the vehicle body vibration frequency is generated, and it is determined whether the audit condition is met; When the audit condition is met, the throttle overload ratio within an audit duration is calculated based on the throttle overload duration; When the audit condition is not met, the inertial measurement information and the vehicle throttle information are reacquired.

5. The vehicle transportation audit scoring method according to claim 4, characterized in that: The first threshold acceleration is the square of the first axial acceleration multiplied by a dynamic adjustment parameter: The second threshold acceleration is the square of the second axial acceleration multiplied by the dynamic adjustment parameter: The third threshold acceleration is the square of the third axial acceleration multiplied by the dynamic adjustment parameter.

6. The vehicle transportation audit scoring method according to claim 2, characterized in that: The audit condition is to determine whether a first analysis time for executing the intersection analysis, a second analysis time for executing the speed analysis, or a third analysis time for executing the smoothness analysis exceeds the audit time; When the first analysis time, the second analysis time or the third analysis time exceeds the audit time, it is determined that the audit condition is met.

7. The vehicle transportation audit scoring method according to claim 1, characterized in that: Further included are: Performing a stop-over analysis to generate a stop-over ratio, a stop-over number, and an average minimum stop-over speed; When the supervised learning scoring procedure is executed, the audit score is further generated based on the non-stop stop ratio, the number of non-stop stop times and the average minimum stop speed.

8. The vehicle transportation audit scoring method according to claim 7, characterized in that: The supervised learning scoring procedure includes: Generating an intersection analysis score using an intersection machine learning model based on the intersection turning rate without stopping, the number of turning without stopping, and the average minimum turning speed; generating a speed analysis score using a speed machine learning model according to the rapid deceleration duration, the rapid acceleration duration, the speeding duration, the speeding ratio, and the average maximum speeding speed; generating a ride analysis score using a ride machine learning model based on the throttle overweight ratio, the number of vehicle body forward tilts, the number of vehicle body rearward tilts, the number of vehicle body side tilts, and the number of vehicle body vibrations; generating a stop-stop analysis score using a stop-stop machine learning model based on the stop-stop ratio, the number of stop-stop times, and the average minimum stop-stop speed; The audit score is generated by performing weighted summation based on the intersection analysis score, the speed analysis score, the smoothness analysis score, and the stop-over analysis score.

9. The vehicle transportation audit scoring method according to claim 7, characterized in that: The transit analysis includes: Obtaining vehicle position information, speed information, brake information, and door information; Obtaining a site location information; Determining, based on the vehicle location information and the site location information, whether at least a second distance between a vehicle location and at least one site location is less than a site threshold distance; When the at least one second distance is less than the station threshold distance, a passing number is generated, a minimum passing speed is generated based on the speed information, and whether to stop and open the door is determined based on the brake information and the door information; When the vehicle does not stop and the door is open, a number of times the vehicle passes through the station without opening the door is generated, and it is determined whether an audit condition is met; When the audit conditions are met, the non-stop transit ratio and the number of non-stop transits within an audit period are calculated based on the number of transits and the number of transits without door openings, and the average minimum transit speed within the audit period is calculated based on the minimum transit speed. When the audit condition is not satisfied, the vehicle position information, the speed information, the brake information, and the door information are re-acquired.

10. A vehicle transportation audit scoring system, characterized in that: Includes: A cloud server performs an intersection analysis to generate an intersection turn-through ratio, a number of intersection turn-through times without stopping, and an average minimum cornering speed; performs a speed analysis to generate a sudden deceleration duration, a sudden acceleration duration, a speeding duration, a speeding ratio, and an average maximum speeding speed; and performs a ride analysis to generate an over-throttle ratio, a number of vehicle body forward tilts, a number of vehicle body backward tilts, a number of vehicle body rolls, and a number of vehicle body vibrations; wherein the cloud server executes a supervised learning scoring process based on the uninterrupted cornering ratio, the number of uninterrupted cornering times at the intersection, the average minimum cornering speed, the rapid deceleration duration, the rapid acceleration duration, the speeding duration, the speeding ratio, the average maximum speeding speed, the throttle overload ratio, the number of vehicle body forward tilts, the number of vehicle body rearward tilts, the number of vehicle body side tilts, and the number of vehicle body vibrations to generate an audit score; The supervised learning scoring procedure is performed by the cloud server using an intersection machine learning model to generate an intersection analysis score based on the intersection turning rate, the intersection turning number and the average minimum turning speed; and the cloud server using a speed machine learning model to generate a speed analysis score based on the sudden deceleration duration, the sudden acceleration duration, the speeding duration, the speeding rate and the average maximum speeding speed; and the cloud server using a smoothness machine learning model to generate a smoothness analysis score based on the throttle overload rate, the vehicle body forward tilt times, the vehicle body backward tilt times, the vehicle body side tilt times and the vehicle body vibration times; and then the cloud server performs weighted summation based on the intersection analysis score, the speed analysis score and the smoothness analysis score to generate the audit score.

11. The vehicle transportation audit scoring system according to claim 10, characterized in that: Further included are: A vehicle-mounted device, mounted on a vehicle, having: a positioning unit, generating vehicle position information according to a positioning state of the vehicle; a vehicle information acquisition unit for measuring a vehicle state of the vehicle to generate speed information and braking information; a wireless transceiver unit connected to the positioning unit and the vehicle information acquisition unit to receive and send the vehicle position information, the speed information, and the braking information to the cloud server; wherein the cloud server is communicatively connected to the wireless transceiver unit of the vehicle-mounted device, and when the cloud server performs the intersection analysis, the cloud server obtains the vehicle location information, the speed information, and the braking information through the wireless transceiver unit, and obtains signal location information from a driving information database; wherein the cloud server determines whether at least a first distance between a vehicle position and at least one signal position is less than an intersection threshold distance based on the vehicle position information and the signal position information; When the at least one first distance is less than the intersection threshold distance, the cloud server generates a number of intersection turns, generates a minimum turning speed based on the speed information, and determines whether the waiting time exceeds a waiting threshold based on the braking information; When the waiting time exceeds the waiting threshold, the cloud server generates a number of times the vehicle has not stopped at the intersection, and determines whether an audit condition is met; When the audit condition is met, the cloud server calculates the intersection turning ratio and the number of intersection turnings without stopping within an audit period based on the number of intersection turnings and the number of intersection turnings without stopping, and calculates the average minimum turning speed within the audit period based on the minimum turning speed; When the audit condition is not met, the cloud server re-acquires the vehicle position information, the speed information, and the braking information through the wireless transceiver unit.

12. The vehicle transportation audit scoring system according to claim 10, characterized in that: Further included are: A vehicle-mounted device, mounted on a vehicle, having: a positioning unit for generating vehicle position information according to a positioning state of the vehicle; a vehicle information acquisition unit for measuring a vehicle state of the vehicle to generate speed information; a wireless transceiver unit connected to the positioning unit and the vehicle information acquisition unit to receive and send the vehicle position information and the speed information to the cloud server; wherein the cloud server is communicatively connected to the wireless transceiver unit of the vehicle-mounted device, and when the cloud server performs the speed analysis, the cloud server obtains the vehicle position information and the speed information through the wireless transceiver unit, and obtains route speed limit information from a driving information database; wherein the cloud server generates location speed limit information based on the vehicle location information and the route speed limit information, and determines whether a vehicle speed is greater than a location speed limit based on the speed information and the location speed limit information; When the vehicle speed exceeds the speed limit at the location, the cloud server calculates the speeding duration, generates a maximum speeding speed based on the speed information, and determines whether an audit condition is met; wherein the cloud server determines, based on the speed information, whether a first difference between a maximum value and a minimum value of the vehicle speed within an acceleration time is greater than a first threshold value; When the first difference is greater than the first threshold, the cloud server generates the rapid acceleration duration and determines whether the audit condition is met; wherein the cloud server determines, based on the speed information, whether a second difference between a maximum value and a minimum value of the vehicle speed within a deceleration time is greater than a second threshold value; When the second difference is greater than the second threshold, the cloud server generates the rapid deceleration duration and determines whether the audit condition is met; When the audit condition is met, the cloud server calculates the speeding ratio within an audit period based on the speeding duration, and calculates the average maximum speeding speed within the audit period based on the maximum speeding speed; When the audit condition is not met, the cloud server re-obtains the vehicle position information and the speed information through the wireless transceiver unit.

13. The vehicle transportation audit scoring system according to claim 10, characterized in that: Further included are: A vehicle-mounted device, mounted on a vehicle, having: a vehicle information acquisition unit for measuring a vehicle state of the vehicle to generate vehicle throttle information; an inertial measurement unit, and measuring an inertial state of the vehicle to generate inertial measurement information; a wireless transceiver unit connected to the vehicle information acquisition unit and the inertial measurement unit to receive and send the vehicle throttle information and the inertial measurement information to the cloud server; wherein the cloud server is communicatively connected to the wireless transceiver unit of the vehicle-mounted device, and when the cloud server performs the ride analysis, the cloud server obtains the inertial measurement information and the vehicle throttle information through the wireless transceiver unit, and determines whether a throttle depth is greater than a threshold depth based on the vehicle throttle information; When the throttle depth is greater than the threshold depth, the cloud server generates an excessive throttle duration and determines whether an audit condition is met; wherein the cloud server determines, based on the inertial measurement information, whether a first absolute value of a first axial acceleration is greater than a first threshold acceleration, determines whether a second absolute value of a second axial acceleration is greater than a second threshold acceleration, and determines whether a third absolute value of a third axial acceleration is greater than a third threshold acceleration; When the first absolute value of the first axial acceleration is greater than the first threshold acceleration and the first axial acceleration is a positive number, the cloud server generates the number of vehicle body backward tilts and determines whether the audit condition is met; When the first absolute value of the first axial acceleration is greater than the first threshold acceleration and when the first axial acceleration is a negative number, the cloud server generates the number of vehicle body forward tilts and determines whether the audit condition is met; When the second absolute value of the second axial acceleration is greater than the second threshold acceleration, the cloud server generates the vehicle body roll count and determines whether the audit condition is met; When the third absolute value of the third axial acceleration is greater than the third threshold acceleration, the cloud server generates the number of vehicle body vibrations and determines whether the audit condition is met; When the audit condition is met, the cloud server calculates the throttle overload ratio within an audit period based on the throttle overload duration; When the audit condition is not met, the cloud server re-obtains the inertial measurement information and the vehicle throttle information through the wireless transceiver unit.

14. The vehicle transportation audit and scoring system according to claim 13, characterized in that: The first threshold acceleration is the square of the first axial acceleration multiplied by a dynamic adjustment parameter: The second threshold acceleration is the square of the second axial acceleration multiplied by the dynamic adjustment parameter: The third threshold acceleration is the square of the third axial acceleration multiplied by the dynamic adjustment parameter.

15. The vehicle transportation audit scoring system according to claim 11, characterized in that: The audit condition is determined by the cloud server as to whether a first analysis time for executing the intersection analysis, a second analysis time for executing the speed analysis, or a third analysis time for executing the smoothness analysis exceeds the audit time; When the first analysis time, the second analysis time or the third analysis time exceeds the audit time, the cloud server determines that the audit condition is met.

16. The vehicle transportation audit scoring system according to claim 10, characterized in that: The cloud server further performs a stop-stop analysis to generate a stop-stop ratio, a stop-stop number, and an average minimum stop-stop speed; When the cloud server executes the supervised learning scoring program, the cloud server further generates the audit score based on the non-stop stop ratio, the number of non-stop stop times and the average minimum stop speed.

17. The vehicle transportation audit scoring system according to claim 16, characterized in that: The supervised learning scoring procedure is performed by the cloud server using an intersection machine learning model to generate an intersection analysis score based on the intersection turning rate, the number of intersection turning rates, and the average minimum turning speed; and the cloud server using a speed machine learning model to generate a speed analysis score based on the sudden deceleration duration, the sudden acceleration duration, the speeding duration, the speeding rate, and the average maximum speeding speed; and the cloud server using a smoothness machine learning model to generate a smoothness analysis score based on the throttle overload rate, the number of vehicle body forward tilts, the number of vehicle body rearward tilts, the number of vehicle body side tilts, and the number of vehicle body vibrations; and the cloud server using a stop-stop analysis score based on the stop-stop rate, the number of stop-stop times, and the average minimum stop-stop speed; and then the cloud server performs weighted summation based on the intersection analysis score, the speed analysis score, the smoothness analysis score, and the stop-stop analysis score to generate the audit score.

18. The vehicle transportation audit scoring system according to claim 16, characterized in that: Further included are: A vehicle-mounted device, mounted on a vehicle, having: a positioning unit for generating vehicle position information according to a positioning state of the vehicle; a vehicle information acquisition unit for measuring a vehicle state of the vehicle to generate speed information, brake information, and door information; a wireless transceiver unit connected to the positioning unit and the vehicle information acquisition unit to receive and send the vehicle position information, the speed information, the brake information, and the door information to the cloud server; wherein the cloud server is communicatively connected to the wireless transceiver unit of the vehicle-mounted device, and when the cloud server performs the stop-over analysis, the cloud server obtains the vehicle position information, the speed information, the braking information, and the door information through the wireless transceiver unit, and obtains stop location information from a driving information database, and determines whether at least a second distance between a vehicle position and at least one stop location is less than a stop threshold distance based on the vehicle position information and the stop location information; When the at least one second distance is less than the station threshold distance, the cloud server generates a station passing count, generates a minimum station passing speed based on the speed information, and determines whether to stop and open the door based on the brake information and the door information; When the vehicle does not stop and the door is open, the cloud server generates a number of times the vehicle passes through the station without opening the door, and determines whether an audit condition is met; When the audit condition is met, the cloud server calculates the non-stop transit ratio and the number of non-stop transits within an audit period based on the number of transits and the number of transits without door openings, and calculates the average minimum transit speed within the audit period based on the minimum transit speed. When the audit condition is not met, the cloud server re-acquires the vehicle position information, the speed information, the brake information, and the door information through the wireless transceiver unit.

Images