A driving training intelligent monitoring and early warning system based on multi-data fusion

Abstract

The application relates to the technical field of intelligent vehicle driving training, in particular to a driving training intelligent monitoring and early warning system based on multi-data fusion, which comprises a data acquisition module used to acquire standard sitting postures of students, vehicle parameters, real-time driving postures and environmental data; a driving sitting height analysis module used to calculate the sitting height and compare the same with a standard interval to generate an abnormal result; a driving posture judgment module used to comprehensively analyze light, body inclination and face orientation to determine a posture risk when the sitting height is abnormal; a driving operation analysis module used to build a driving area and judge operation compliance in combination with obstacle positions and braking behaviors; and an abnormal state statistical module used to accumulate abnormal reminders and dynamically plan a next training duration. Through multi-dimensional data fusion analysis and time sequence behavior modeling, the application realizes full-process intelligent evaluation and individualized early warning from single sitting posture monitoring to driving operation, environmental interaction and long-term learning planning.

Description

A driver training intelligent monitoring and early warning system based on multi-data fusion Technical Field

[0001] This invention relates to the field of intelligent vehicle driver training technology, and in particular to an intelligent monitoring and early warning system for driver training based on multi-data fusion. Background Technology

[0002] With the development of intelligent driving and driver training industry, the monitoring and auxiliary judgment of driver behavior has gradually evolved from traditional manual observation to automated methods based on sensors and video monitoring. In the early days, driver training monitoring relied heavily on instructors manually recording students' operating performance, which had problems of strong subjectivity and low efficiency. To improve objectivity, some training vehicles have been equipped with basic operation recording devices that can collect data such as throttle, brake, and steering wheel angle to analyze whether the students' operation is standardized.

[0003] In recent years, with advancements in computer vision and sensor technology, methods for recognizing trainee postures based on in-vehicle cameras have emerged. These methods can detect whether trainees are looking down, yawning, or showing signs of inattention. Furthermore, some systems utilize GPS positioning and inertial navigation systems to acquire vehicle trajectory and speed information, which is used to analyze whether driving training operations are compliant. These methods, to a certain extent, improve the objectivity and intelligence of driver training.

[0004] While existing driver training monitoring technologies have evolved from manual recording to video recognition and sensor fusion, they still struggle to simultaneously address issues such as accurate posture determination, environmental interference identification, operational risk analysis, and training cycle feedback. Therefore, there is an urgent need for an intelligent monitoring and early warning system for driver training based on multi-data fusion to achieve comprehensive monitoring and scientific management of the trainee training process.

[0005] Chinese Patent Publication No. CN106651702A discloses an intelligent voice-based driver training system, including a student information storage module, an identity recognition module, an instructor matching module, a voice monitoring module, and a driving state detection module. The identity recognition module identifies the student by collecting the student's fingerprint information. After receiving the student information sent by the identity recognition module, the instructor matching module sends the instructor information associated with the idle state to the voice monitoring module. The voice monitoring module includes a human-computer interaction unit, a monitoring unit, a voice unit, and a display terminal. The human-computer interaction interface obtains the instructor information associated with the idle state sent by the instructor matching module and displays it to the student in the form of a list.

[0006] At the same time, while existing intelligent vehicle and driver training technologies have solved the problems of student identification, instructor scheduling, and remote voice monitoring, they have significant shortcomings in terms of posture detection accuracy, operation scenario fusion analysis, and long-term feedback mechanisms, making it difficult to meet the requirements of the next generation of intelligent driving training for safety and scientific management. Summary of the Invention

[0007] To address this, the present invention provides a driver training intelligent monitoring and early warning system based on multi-data fusion, which overcomes the problem of insufficient attitude detection accuracy and imperfect driving scenario analysis in the prior art, resulting in a lack of feedback mechanism in driver training and learning.

[0008] To achieve the above objectives, the present invention provides a driver training intelligent monitoring and early warning system based on multi-data fusion, comprising:

[0009] The data acquisition module is used to acquire historical driving data and collect standard sitting posture parameters, vehicle static parameters, vehicle driving parameters, target student driving posture parameters, and environmental interference parameters when the target student is in a driving state.

[0010] The driver seating height analysis module is connected to the data acquisition module and is used to calculate the first driver seating height based on the standard seating posture parameters, the driving posture parameters and the environmental interference parameters, and to calculate the standard seating height range based on historical driving data, and to obtain the first seating height anomaly result when the first driver seating height is less than the minimum value of the standard seating height range.

[0011] The driving posture judgment module is connected to the driving seat height analysis module. When a first abnormal seat height result is obtained and the facial illumination intensity is less than or equal to the facial illumination intensity threshold, the module obtains the body tilt angle analysis result obtained by comparing the body tilt angle with the safe body tilt angle value, and obtains the comparison result of the downward facing angle and the face facing angle threshold of the body tilt angle analysis result to determine whether the corresponding driving posture is abnormal.

[0012] The driving operation analysis module, connected to the driving posture judgment module, is used to construct a driving area based on vehicle coordinates and driving direction when the first face orientation analysis result is obtained after the first body tilt angle analysis result and face orientation analysis. It divides the driving area into a first driving area, a second driving area, and a third driving area, and determines whether an obstacle is located in the second driving area or the third driving area. If there is an obstacle and there is braking behavior in the time period from when the obstacle enters the area to the current time, it obtains a result of normal operation but abnormal posture.

[0013] The abnormal status statistics module, connected to the driving operation analysis module, is used to count the total number of abnormal status reminders received by the target learner during this driving training session, and to obtain the driving training duration based on the preset mapping relationship between the total number of abnormal status reminders and the driving training duration.

[0014] Furthermore, the data acquisition module includes a parameter acquisition unit and a parameter calculation unit;

[0015] The parameter acquisition unit is used to directly acquire several parameters, including static seat height in standard sitting posture parameters, vehicle static parameters, vehicle coordinates, driving direction, driving speed, steering angle, braking signal, driving posture parameters, and environmental interference parameters in vehicle driving parameters.

[0016] The parameter calculation unit is connected to the parameter acquisition unit and is used to calculate and obtain several parameters, including the standard sitting height range in the standard sitting posture parameters, and the first driving area, the second driving area, and the third driving area in the vehicle driving parameters.

[0017] Furthermore, the driver seating height analysis module includes a seating height calculation unit and a seating height judgment unit;

[0018] The seat height calculation unit is used to calculate the corrected first driver seat height based on the seat fore-aft position and backrest angle in the vehicle's static parameters.

[0019] The seat height determination unit is connected to the seat height calculation unit and is used to compare the first driver seat height with the standard seat height range to obtain the corresponding driver seat height determination result.

[0020] Furthermore, the sitting height calculation unit includes a sitting height correction subunit and an interval mapping subunit;

[0021] The seat height correction subunit is used to perform weighted calculations on the fore-and-aft position of the seat and the backrest angle based on a preset seat height correction algorithm to obtain the corrected first driver seat height.

[0022] The interval mapping subunit, connected to the seat height correction subunit, is used to map the first driver seat height to a standard seat height interval obtained based on historical driving data.

[0023] Furthermore, the driving posture determination module includes a light intensity analysis unit and a body posture analysis unit;

[0024] The light intensity analysis unit is used to determine whether the facial light intensity is greater than the facial light intensity threshold, and analyze the light source category when it is greater.

[0025] The body posture analysis unit, connected to the light intensity analysis unit, is used to perform body tilt angle analysis when the facial light intensity is less than or equal to the facial light intensity threshold. After obtaining the body tilt angle analysis result, it performs face orientation analysis to determine whether the driving posture corresponding to the face orientation analysis result is abnormal.

[0026] Furthermore, the body posture analysis unit includes a tilt angle analysis subunit;

[0027] The tilt angle analysis subunit is used to compare the body tilt angle with the safe body tilt angle value to obtain the tilt angle comparison result.

[0028] Furthermore, the body posture analysis unit also includes a face orientation analysis subunit;

[0029] The face orientation analysis subunit is connected to the tilt angle analysis subunit. When the tilt angle analysis is obtained, the face orientation angle comparison result obtained by comparing the face orientation angle with the face orientation angle threshold is obtained, and whether the corresponding driving posture is abnormal.

[0030] Furthermore, the driving operation analysis module includes a driving area division unit and an obstacle analysis unit;

[0031] The driving area division unit is used to construct the driving area based on the vehicle's driving coordinates and direction, and to divide the driving area into the first driving area, the second driving area and the third driving area;

[0032] The obstacle analysis unit, connected to the driving area division unit, is used to determine whether an obstacle is in the second or third driving area when driving.

[0033] Furthermore, the obstacle behavior analysis unit includes a region judgment subunit and a braking analysis subunit;

[0034] The area determination subunit is used to determine whether an obstacle is located in the second or third driving area;

[0035] The braking analysis subunit, connected to the area judgment subunit, is used to analyze whether there is braking behavior in the time period from when the obstacle enters the area to the current time, and outputs the first driving operation behavior or the second driving operation behavior accordingly.

[0036] Furthermore, the abnormal state statistics module includes an abnormal counting unit and a duration mapping unit;

[0037] The anomaly counting unit is used to count the total number of abnormal status alerts received by the target learner during this driving training session;

[0038] The duration mapping unit is connected to the anomaly counting unit and is used to obtain the driving training duration based on the preset mapping relationship between the total number of anomaly status alerts and the driving training duration.

[0039] Compared with existing technologies, the beneficial effects of this invention are as follows: by integrating several parameters during driver training and combining seat height correction and interval mapping mechanisms, the accuracy of posture judgment can be effectively improved; through comprehensive analysis of the body tilt angle and the angle of facing, multiple abnormal states can be distinguished, and by combining the elimination of light intensity and noise interference, the reliability of posture recognition results can be ensured; in terms of driving scene monitoring, a partition model is constructed using the relationship between vehicle coordinates, driving direction and obstacle position, and the nature of operation is determined by braking behavior, forming more targeted early warning prompts; through the abnormal state statistics module, a quantitative mapping relationship is established between the number of abnormal reminders and training duration, providing a reference for subsequent personalized learning arrangements; while improving the safety of driver training, this system establishes a complete closed loop of monitoring, early warning and feedback, significantly enhancing the scientific and intelligent level of the training process.

[0040] Furthermore, through the collaborative action of the parameter acquisition unit and the parameter calculation unit, the fusion processing of multi-source data in the driving training scenario is realized; the acquisition unit can acquire the driver's posture, vehicle operating status and environmental parameters in real time to ensure data integrity; the calculation unit combines the vehicle's static dimensions and steering parameters to dynamically generate the first driving area and accurately divide the second and third driving areas, thereby realizing the rapid identification and judgment of obstacle risks and improving the accuracy and real-time performance of anomaly detection in driving training.

[0041] Furthermore, by setting up a seat height calculation unit and a seat height judgment unit, the first driver's seat height can be corrected based on the collected seat front and back position and backrest angle, and compared with the standard seat height range to output an intuitive seat height status result. This ensures the objectivity and consistency of the sitting posture analysis and provides more accurate basic data support for subsequent abnormal posture identification, thereby significantly improving the accuracy and practicality of the system in the posture monitoring stage. Attached Figure Description

[0042] Figure 1 is a schematic diagram of the intelligent monitoring and early warning system for driver training based on multi-data fusion according to an embodiment of the present invention;

[0043] Figure 2 is a logic diagram of the driver's seating height determination unit in an embodiment of the present invention;

[0044] Figure 3 is a schematic diagram of the sitting height calculation unit according to an embodiment of the present invention;

[0045] Figure 4 is a logic decision diagram of the light intensity analysis unit in an embodiment of the present invention. Detailed Implementation

[0046] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0047] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0048] It should be noted that in the description of this invention, the terms "upper", "lower", "left", "right", "inner", "outer", etc., which indicate directions or positional relationships, are based on the directions or positional relationships shown in the accompanying drawings. This is only for the convenience of description and is not intended to indicate or imply that the device or element must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation of this invention.

[0049] Furthermore, it should be noted that, in the description of this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0050] Please refer to Figure 1, which is a structural schematic diagram of the intelligent monitoring and early warning system for driver training based on multi-data fusion according to an embodiment of the present invention. The present invention provides an intelligent monitoring and early warning system for driver training based on multi-data fusion, comprising:

[0051] The data acquisition module is used to acquire historical driving data and collect standard sitting posture parameters, vehicle static parameters, vehicle driving parameters, target student driving posture parameters, and environmental interference parameters when the target student is in a driving state.

[0052] The driver seating height analysis module is connected to the data acquisition module and is used to calculate the first driver seating height based on the standard seating posture parameters, the driving posture parameters and the environmental interference parameters, and to calculate the standard seating height range based on historical driving data, and to obtain the first seating height anomaly result when the first driver seating height is less than the minimum value of the standard seating height range.

[0053] The driving posture judgment module is connected to the driving seat height analysis module. When a first abnormal seat height result is obtained and the facial illumination intensity is less than or equal to the facial illumination intensity threshold, the module obtains the body tilt angle analysis result obtained by comparing the body tilt angle with the safe body tilt angle value, and obtains the comparison result of the downward facing angle and the face facing angle threshold of the body tilt angle analysis result to determine whether the corresponding driving posture is abnormal.

[0054] The driving operation analysis module, connected to the driving posture judgment module, is used to construct a driving area based on vehicle coordinates and driving direction when the first face orientation analysis result is obtained after the first body tilt angle analysis result and face orientation analysis. It divides the driving area into a first driving area, a second driving area, and a third driving area, and determines whether an obstacle is located in the second driving area or the third driving area. If there is an obstacle and there is braking behavior in the time period from when the obstacle enters the area to the current time, it obtains a result of normal operation but abnormal posture.

[0055] The abnormal status statistics module, connected to the driving operation analysis module, is used to count the total number of abnormal status reminders received by the target learner during this driving training session, and to obtain the driving training duration based on the preset mapping relationship between the total number of abnormal status reminders and the driving training duration.

[0056] In this embodiment, the driving state is the state of the vehicle being driven or temporarily stopped when the vehicle is ignited and not parked.

[0057] The standard sitting posture parameters include static sitting height and standard sitting height range; vehicle static parameters include seat fore-and-aft position, backrest angle, vehicle length, and vehicle geometric center; vehicle driving parameters include vehicle driving direction and maximum vehicle steering angle; driving posture parameters include driving sitting height, face orientation angle, body tilt angle, and head and neck angle; environmental interference parameters are facial light intensity.

[0058] The vehicle-mounted camera in the data acquisition module obtains the seat motor position signal based on computer vision algorithms, and the standard sitting posture parameters are obtained together with the OBD interface.

[0059] After obtaining the vehicle's VIN code through the OBD interface, the vehicle's static parameters are retrieved from the database; driving posture parameters are collected through the vehicle's onboard camera and IMU in the data acquisition module; environmental interference parameters are collected through the ambient light sensor and vehicle's onboard camera in the data acquisition module; and vehicle driving parameters are directly provided by the GPS, IMU, OBD, and steering wheel sensors in the data acquisition module.

[0060] The high-quality data processed by the data fusion and processing module, such as normalized seat height values ​​and filtered posture angles, are inputs to the driver seat height analysis module, driver posture judgment module, driver operation analysis module, and behavior recognition and risk assessment module.

[0061] A lightweight machine learning model, SVM, is used to compare real-time seating height with a standard range to analyze driver seating height.

[0062] The system primarily uses a computer vision CNN model to analyze camera video streams and identify lighting, body tilt, and facial orientation to determine driving posture.

[0063] Rule-based logical judgment and LSTM time series analysis model are used to comprehensively judge the relationship between obstacle position, vehicle trajectory and braking operation.

[0064] By integrating standard sitting posture parameters, vehicle static parameters, driving posture parameters, vehicle driving parameters, and environmental interference parameters collected during driver training, and combining these with a sitting height correction and interval mapping mechanism, the accuracy of sitting posture determination can be effectively improved. Through comprehensive analysis of the body tilt angle and facing angle, multiple abnormal states can be distinguished, and by eliminating light intensity and noise interference, the reliability of posture recognition results is ensured. In terms of driving scene monitoring, a zoning model is constructed using the relationship between vehicle coordinates, driving direction, and obstacle positions. Combined with braking behavior to determine the nature of the operation, more targeted early warning prompts are generated. An abnormal state statistics module establishes a quantitative mapping relationship between the number of abnormal alerts and training duration, providing a reference for subsequent personalized learning arrangements. This system improves driver training safety while establishing a complete closed loop of monitoring, early warning, and feedback, significantly enhancing the scientific and intelligent level of the training process.

[0065] Specifically, the data acquisition module includes a parameter acquisition unit and a parameter calculation unit;

[0066] The parameter acquisition unit is used to directly acquire several parameters, including static seat height in standard sitting posture parameters, vehicle static parameters, vehicle coordinates, driving direction, driving speed, steering angle, braking signal, driving posture parameters, and environmental interference parameters in vehicle driving parameters.

[0067] The parameter calculation unit is connected to the parameter acquisition unit and is used to calculate and obtain several parameters, including the standard sitting height range in the standard sitting posture parameters, and the first driving area, the second driving area, and the third driving area in the vehicle driving parameters.

[0068] In this embodiment, the parameter acquisition unit acquires the driver's line of sight parameters through the vehicle camera, the vehicle coordinates and speed parameters through the GPS module, the engine speed and throttle opening parameters through the OBD interface, the distance parameters of surrounding obstacles through the radar sensor, and the steering wheel angle parameters through the steering wheel sensor.

[0069] The parameter calculation unit obtains the above parameters and calls the preset vehicle static parameter database to obtain the current vehicle body length and body width data; with the real-time vehicle coordinates as the center, it calculates the radius of the outer circle based on the body length and width values ​​to generate the first driving area circular boundary that completely surrounds the vehicle outline.

[0070] By combining the steering wheel angle parameters and the vehicle's minimum turning radius data, the instantaneous turning radius of the vehicle is calculated. Starting from the vehicle coordinates, along the current driving direction axis, the motion trajectory of the vehicle when traveling at the maximum steering angle is simulated forward and backward respectively, generating two symmetrical arc-shaped trajectory lines. These two trajectory lines divide the space around the vehicle into four fan-shaped regions, of which the left and right fan-shaped regions in the forward direction are marked as the second driving region, and the left and right fan-shaped regions in the backward direction are marked as the third driving region.

[0071] The system acquires obstacle distance and orientation data collected by radar sensors in real time and converts it into planar coordinate data. It then compares these coordinates with the boundary coordinates of the second and third driving areas provided by the calculation unit. When the system detects that the obstacle coordinates are located within any risk area, it immediately queries historical data from the time the obstacle entered the area to the current time.

[0072] Through the collaborative action of the parameter acquisition unit and the parameter calculation unit, the fusion processing of multi-source data in the driving training scenario is realized. The acquisition unit can acquire the driver's posture, vehicle operating status and environmental parameters in real time to ensure data integrity. The calculation unit combines the vehicle's static dimensions and steering parameters to dynamically generate the first driving area and accurately divide the second and third driving areas, thereby realizing the rapid identification and judgment of obstacle risks and improving the accuracy and real-time performance of anomaly detection in driving training.

[0073] Referring to Figure 2, which is a logic determination diagram of the sitting height determination unit in an embodiment of the present invention;

[0074] Specifically, the driver seating height analysis module includes a seating height calculation unit and a seating height judgment unit;

[0075] The seat height calculation unit is used to calculate the corrected first driver seat height based on the seat fore-aft position and backrest angle in the vehicle's static parameters.

[0076] The seat height determination unit is connected to the seat height calculation unit and is used to compare the first driver seat height with the standard seat height range to obtain the corresponding driver seat height determination result.

[0077] In this embodiment, the first driver's seat height is compared with the standard seat height range to determine whether the first driver's seat height obtained after correction during driving is within the standard seat height range.

[0078] If the first driver's seat height is less than or equal to the minimum value of the standard seat height range, the driving posture is abnormal. The first driver's seat height status is obtained, and the corresponding first seat height abnormality result is further analyzed.

[0079] If the first driver's seating height is within the standard seating height range and the driving posture is normal, the second driver's seating height status is obtained, and no abnormal status alert is sent.

[0080] If the first driver's seat height is greater than or equal to the maximum value of the standard seat height range, the driving posture is abnormal. The third driver's seat height status is obtained, corresponding to the second abnormal seat height result. The driving posture is abnormal, and the first abnormal status reminder is sent, that is, the driver's seat height is too high. Adjust the driving posture and adjust the seat back if necessary.

[0081] The system is equipped with a seat height calculation unit and a seat height judgment unit. Based on the collected seat fore-and-aft position and backrest angle, the system can correct the first driver's seat height and compare it with the standard seat height range to output an intuitive seat height status result. This ensures the objectivity and consistency of the sitting posture analysis and provides more accurate basic data support for subsequent abnormal posture identification, thereby significantly improving the accuracy and practicality of the system in the posture monitoring stage.

[0082] Referring to Figure 3, which is a schematic diagram of the sitting height calculation unit in an embodiment of the present invention;

[0083] Specifically, the sitting height calculation unit includes a sitting height correction subunit and an interval mapping subunit;

[0084] The seat height correction subunit is used to perform weighted calculations on the fore-and-aft position of the seat and the backrest angle based on a preset seat height correction algorithm to obtain the corrected first driver seat height.

[0085] The interval mapping subunit, connected to the seat height correction subunit, is used to map the first driver seat height to a standard seat height interval obtained based on historical driving data.

[0086] In this embodiment, the vehicle-mounted camera that acquires standard sitting posture parameters is located at the geometric center of the top of the vehicle, and uses computer vision algorithms to detect the position of the seat motor and the static sitting height of the target student in real time.

[0087] The weighted correction algorithm was used to calculate the first driver's seating height.

[0088]

[0089] in, This refers to the static sitting height. For seat displacement; Backrest angle; To ensure the accuracy of the correction coefficients, which are obtained by machine learning fitting based on historical training samples, at least 500 sets of historical student data are collected and determined based on least squares fitting. In this embodiment, α∈[0.85,1.05], β∈[0.90,1.10], preferably, α=0.9, β=0.95.

[0090] The fore-aft position of the seat refers to the position of the driver's seat relative to a fixed reference point on the vehicle after displacement. This represents the midpoint of the seat rail; a negative value indicates that the seat is behind the reference point, i.e., away from the steering wheel; a positive value indicates that the seat is in front of the reference point, i.e., close to the steering wheel.

[0091] The formula for calculating seat displacement is:

[0092]

[0093] Among them, the vehicle fixed reference point is , The fore-and-aft position of the seat;

[0094] The seat rails are linear tracks. It directly represents the coordinate values ​​of the seat, that is, one-dimensional coordinates, rather than spatial coordinates;

[0095] The interval mapping subunit receives the first driver's seat height output by the seat height correction subunit and maps it to the system's preset standard seat height interval [Hmin, Hmax]. This standard seat height interval is determined based on historical driving data through statistical learning methods and is used to characterize the reasonable seat height range for trainees in normal driving postures. A large number of trainees' static seat heights, corresponding seat displacements, and backrest angles in standard postures are collected, and the corresponding historical first driver's seat heights are calculated. All historical corrected seat heights are sorted from shortest to tallest and their distribution is analyzed. If there are one hundred historical first driver's seat height data, the 5th percentile is taken as the lower limit of the interval, and the 95th percentile is taken as the upper limit of the interval to form the standard seat height interval [Hmin, Hmax], covering 95% of normal seat height situations. If the 5th percentile of the corrected seat height in the historical data is 85cm and the 95th percentile is 95cm, then the standard seat height interval is [85, 95]. In practical applications, this interval can be adaptively adjusted according to different vehicle models, seat designs, and user characteristics.

[0096] By using statistical learning methods based on historical big data to determine the standard sitting height range, the interference of extreme individual differences is objectively eliminated, significantly improving the scientific nature and universality of sitting posture judgment. This range can be adaptively adjusted according to different vehicle models and population characteristics, greatly enhancing the accuracy and reliability of abnormal driver sitting height monitoring and the system's adaptability in different application scenarios.

[0097] Referring to Figure 4, which is a logic decision diagram of the light intensity analysis unit in an embodiment of the present invention;

[0098] Specifically, the driving posture determination module includes a light intensity analysis unit and a body posture analysis unit;

[0099] The light intensity analysis unit is used to determine whether the facial light intensity is greater than the facial light intensity threshold, and analyze the light source category when it is greater.

[0100] The body posture analysis unit, connected to the light intensity analysis unit, is used to perform body tilt angle analysis when the facial light intensity is less than or equal to the facial light intensity threshold. After obtaining the body tilt angle analysis result, it performs face orientation analysis to determine whether the driving posture corresponding to the face orientation analysis result is abnormal.

[0101] In this embodiment, the facial illumination intensity threshold is set to 400 Lux. This threshold is set based on research data of driving safety human factors engineering and represents the illumination level of the driver's face for comfortable reading of the dashboard under normal sunlight or in-vehicle lighting. Facial illumination intensity recognition is completed in collaboration between the ambient light sensor in the data acquisition module and the vehicle camera.

[0102] Compare the facial light intensity with the facial light intensity threshold;

[0103] If the facial illumination intensity is greater than the facial illumination intensity threshold, the first facial illumination intensity comparison result is obtained, and the light source category is further determined, including the first light source category and the second light source category;

[0104] The first category of light sources is natural ambient light that originates from outside the vehicle or cannot be directly controlled by the driver.

[0105] The second category of light sources is artificial light sources that originate from inside the vehicle and can be controlled or introduced by the driver;

[0106] If the first light source category is determined and the driving posture is abnormal, a second abnormal status alert is sent, that is, the driving posture is adjusted and the sun visor is adjusted if necessary;

[0107] If the second light source category is determined and the driving posture is abnormal, a third abnormal status alert is sent, which means adjusting the driving posture and adjusting the in-vehicle light source when necessary.

[0108] By introducing facial illumination intensity threshold judgment and light source category recognition, the abnormal posture caused by external strong light glare and in-vehicle artificial light sources can be effectively distinguished. It can not only accurately trigger differentiated warnings, but also eliminate the influence of ambient light interference on visual recognition algorithms, which greatly improves the accuracy of driving posture monitoring and the effectiveness of reminders, and enhances the intelligence and practicality of the system.

[0109] Specifically, the body posture analysis unit includes a tilt angle analysis subunit;

[0110] The tilt angle analysis subunit is used to compare the body tilt angle with the safe body tilt angle value to obtain the tilt angle comparison result.

[0111] In this embodiment, the threshold for the facing angle is 30°, which represents the extreme angle at which the driver's line of sight deviates from the center line directly in front of the vehicle.

[0112] When the facial illumination intensity is less than or equal to the facial illumination intensity threshold, the body tilt angle is obtained and compared with the safe body tilt angle value.

[0113] The vehicle's camera captures images of the driver's face, and computer vision algorithms are used to identify facial feature points such as the eyes, nose tip, and mouth corners. The angle of deflection of the plane formed by these feature points relative to the camera's imaging plane, i.e. the direction of the vehicle's movement, is calculated to obtain the angle of the face.

[0114] If the body tilt angle is greater than the safe value for body tilt angle, the first body tilt angle comparison result is obtained. Based on the first body tilt angle analysis result, the facing angle is further obtained, and the facing angle is compared with the facing angle threshold.

[0115] If the body tilt angle is less than or equal to the body tilt angle threshold, a second body tilt angle comparison result is obtained. Based on the second body tilt angle analysis result, the facing angle is further obtained, and the facing angle is compared with the facing angle threshold.

[0116] By setting dual quantitative thresholds for body tilt and facial orientation, a multi-dimensional and high-precision abnormal driving posture judgment mechanism has been established. While overcoming the limitations of single-indicator judgment, it effectively distinguishes scenarios with different risk levels, provides more scientific and reliable safety status monitoring, and greatly reduces false alarm and missed alarm rates.

[0117] Specifically, the body posture analysis unit also includes a face orientation analysis subunit;

[0118] The face orientation analysis subunit is connected to the tilt angle analysis subunit. When the tilt angle analysis is obtained, the face orientation angle comparison result obtained by comparing the face orientation angle with the face orientation angle threshold is obtained, and whether the corresponding driving posture is abnormal.

[0119] In this embodiment, the safe value for the body tilt angle is 15°, and the threshold value for the facing angle is 30°.

[0120] The body tilt angle safety value measures trunk stability; as long as the trunk tilt exceeds the body tilt angle safety value, regardless of the direction the head is facing, it will affect the control of the vehicle; the face orientation angle threshold measures the range of visual attention; as long as the face orientation deviates from the front by more than the face orientation angle threshold, the driver's core field of vision has already left the main road ahead, and he will be unable to effectively deal with emergencies.

[0121] The face orientation angle is compared with the face orientation angle threshold.

[0122] If the face angle is greater than the face angle threshold, the first face angle comparison result is obtained. According to the first face analysis result, the driving posture is abnormal, and further analysis of driving operation is required.

[0123] If the facing angle is less than or equal to the facing angle threshold, the second facing angle comparison result is obtained. According to the second facing analysis result, the driving posture is abnormal, and the fourth abnormal status reminder is sent, that is, the body tilt angle is too large, and the driving posture is adjusted.

[0124] If the body tilt angle is less than or equal to the body tilt angle threshold, a second body tilt angle comparison result is obtained, and the facing angle is further obtained and compared with the facing angle threshold.

[0125] If the facing angle is greater than the facing angle threshold, the third facing angle comparison result is obtained, the driving posture is abnormal, and the fifth abnormal status reminder is sent, that is, not looking ahead, and the driving posture is adjusted.

[0126] If the facing angle is less than or equal to the facing angle threshold, the fourth facing angle comparison result is obtained. The driving posture is normal, and no abnormal status reminder is sent.

[0127] By establishing fixed dual thresholds for body tilt and gaze deflection, an objective and quantitative standard for judging abnormal driving posture was established. Its core advantage lies in not judging posture in isolation, but providing accurate input signals for subsequent scenario-based comprehensive analysis.

[0128] Specifically, the driving operation analysis module includes a driving area division unit and an obstacle analysis unit;

[0129] The driving area division unit is used to construct the driving area based on the vehicle's driving coordinates and direction, and to divide the driving area into the first driving area, the second driving area and the third driving area;

[0130] The obstacle analysis unit, connected to the driving area division unit, is used to determine whether an obstacle is in the second or third driving area when driving.

[0131] In this embodiment, the vehicle coordinates are obtained from the GPS module and are the projection points of the vehicle's geometric center on the plane.

[0132] Vehicle direction of travel: from GPS or IMU module, indicating the azimuth angle of the vehicle's direction of travel; Maximum steering wheel rotation angle: from steering wheel sensor, a fixed attribute of the vehicle.

[0133] Obtain the vehicle coordinates and the vehicle's direction of travel, using the coordinates of the point projected onto the plane from the vehicle's geometric center as the vehicle coordinates;

[0134] With the vehicle coordinates as the center, and the vehicle length and width as the simplified shape of the vehicle, draw the circumcircle of the simplified shape of the vehicle. The area inside the circumcircle is the first driving area.

[0135] Based on the vehicle's current direction of travel and the maximum steering wheel angle, the space outside the circular area is divided. When the vehicle is traveling at its maximum steering angle, its forward and backward trajectories are both arc-shaped paths, which divide the space around the vehicle into four fan-shaped areas.

[0136] When a vehicle is moving forward, the fan-shaped areas on the left and right that can be reached using the maximum steering angle are called the second driving area; when a vehicle is moving backward, the fan-shaped areas on the left and right that can be reached using the maximum steering angle are called the third driving area.

[0137] Determine whether the obstacle is in the second or third driving area;

[0138] If the obstacle risk result is obtained, further information should be obtained on whether there was braking behavior during the time from when the obstacle entered the second or third driving area as the vehicle was driving until the data was obtained.

[0139] If the obstacle is neither in the second driving area nor in the third driving area, the result of no obstacle risk is obtained. The driving operation is normal but the driving posture is abnormal. The eighth abnormal status reminder is sent, that is, the driver is exhibiting probing behavior. Please maintain the standard driving posture.

[0140] The existence of braking behavior is determined by a time-series query logic without a fixed threshold. The system will backtrack to query whether the OBD system or brake pedal sensor has recorded braking signals during the period from the historical time point when the obstacle enters the risk area to the current time. As long as there is braking behavior within this period, it is determined to exist.

[0141] By leveraging the inherent physical parameters and real-time data of each vehicle, there is no need to preset a fixed safety distance threshold. Instead, a personalized safety warning zone that conforms to kinematic principles is generated based on the specific characteristics and real-time status of each vehicle, making the risk assessment of obstacles more accurate.

[0142] Specifically, the obstacle behavior analysis unit includes a region judgment subunit and a braking analysis subunit;

[0143] The area determination subunit is used to determine whether an obstacle is located in the second or third driving area;

[0144] The braking analysis subunit, connected to the area judgment subunit, is used to analyze whether there is braking behavior in the time period from when the obstacle enters the area to the current time, and outputs the first driving operation behavior or the second driving operation behavior accordingly.

[0145] In this embodiment, it is determined whether the obstacle is in the second driving area or the third driving area;

[0146] If so, further obtain whether there was braking behavior during the time from when the vehicle entered the second or third driving area as the obstacle entered the second or third driving area until the data was obtained;

[0147] If braking occurs, the first driving operation behavior is obtained under the result of obstacle risk. If the driving operation is normal but the driving posture is abnormal, the sixth abnormal status reminder is sent, that is, there is probe behavior during driving, and the driving posture is adjusted.

[0148] If there is no braking action, and there is an obstacle risk, the second driving operation action is obtained. If the driving operation is abnormal and the driving posture is abnormal, the seventh abnormal status reminder is sent, that is, there is a collision risk and there is a probe action. Brake in time and adjust the driving posture.

[0149] By accurately determining the location of obstacles and analyzing braking behavior, dual monitoring of driving operation and posture can be achieved. This can effectively distinguish between abnormal postures during normal operation and completely abnormal states, provide early warnings of probe behavior and collision risks, significantly improve driving safety, and guide drivers to adjust their posture and braking in a timely manner to avoid accidents and cultivate standardized driving habits.

[0150] Specifically, the obstacle behavior analysis unit includes a region judgment subunit and a braking analysis subunit;

[0151] The area determination subunit is used to determine whether an obstacle is located in the second or third driving area;

[0152] The braking analysis subunit, connected to the area judgment subunit, is used to analyze whether there is braking behavior in the time period from when the obstacle enters the area to the current time, and outputs the first driving operation behavior or the second driving operation behavior accordingly.

[0153] In this embodiment, the duration of the next driver training session = base duration + total number of anomalies × unit anomaly increase duration;

[0154] The base duration is set at 120 minutes; the time increment for unit anomalies is 5 minutes, meaning that for each anomaly alert received, the next training session will be extended by 5 minutes.

[0155] Meanwhile, the preset driver training duration is determined based on the total number of abnormal status alerts received during the previous driver training session.

[0156] By automatically counting the number of driving errors and linearly increasing the duration of the next training session, a data-driven feedback adjustment mechanism is established. This effectively motivates trainees to operate in a standardized manner to reduce mistakes, and quantifies the learning effect in training time, thereby achieving personalized and adaptive adjustment of training intensity and significantly improving training efficiency and safety.

[0157] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

[0158] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A driver training intelligent monitoring and early warning system based on multi-data fusion, characterized in that, include: The data acquisition module is used to acquire historical driving data and, when the target learner is in a driving state, collect standard sitting posture parameters, vehicle static parameters, vehicle driving parameters, the target learner's driving posture parameters, and environmental interference parameters. The driving seat height analysis module, connected to the data acquisition module, is used to calculate a first driving seat height based on the standard sitting posture parameters, the driving posture parameters, and the environmental interference parameters, and to calculate a standard seat height range based on historical driving data, obtaining a first seat height anomaly result when the first driving seat height is less than the minimum value of the standard seat height range. The driving posture judgment module, connected to the driving seat height analysis module, is used to, when a first seat height anomaly result is obtained and the facial illumination intensity is less than or equal to a facial illumination intensity threshold, obtain a judgment based on body tilt. The body tilt angle analysis result is obtained by comparing the oblique angle with the body tilt angle safety value. The comparison result of the downward facing angle and the face facing angle threshold of the body tilt angle analysis result is obtained to determine whether the corresponding driving posture is abnormal. The driving operation analysis module is connected to the driving posture judgment module. When the first face facing analysis result is obtained under the first body tilt angle analysis result, the driving area is constructed based on the vehicle coordinates and driving direction, and the first driving area, the second driving area and the third driving area are divided. It is determined whether the obstacle is located in the second driving area or the third driving area. When there is an obstacle and there is braking behavior in the time period from the time the obstacle enters the area to the current time, the result of normal operation but abnormal posture is obtained. The abnormal status statistics module is connected to the driving operation analysis module. It is used to count the total number of abnormal status reminders received by the target learner in this driving training session, and to obtain the driving training duration based on the preset mapping relationship between the total number of abnormal status reminders and the driving training duration. The driver seating height analysis module includes a seating height calculation unit and a seating height judgment unit; the seating height calculation unit is used to calculate the corrected first driver seating height based on the seat fore-aft position and backrest angle in the vehicle static parameters. The seat height determination unit is connected to the seat height calculation unit and is used to compare the first driver seat height with the standard seat height range to obtain the corresponding driver seat height determination result.

2. The intelligent monitoring and early warning system for driver training based on multi-data fusion as described in claim 1, characterized in that, The data acquisition module includes a parameter acquisition unit and a parameter calculation unit. The parameter acquisition unit is used to directly acquire several parameters, including static seat height, vehicle static parameters, vehicle coordinates, driving direction, driving speed, steering angle, braking signal, driving posture parameters, and environmental interference parameters in the standard sitting posture parameters. The parameter calculation unit is connected to the parameter acquisition unit and is used to calculate and obtain several parameters, including the standard seat height range in the standard sitting posture parameters and the first driving area, second driving area, and third driving area in the vehicle driving parameters.

3. The intelligent monitoring and early warning system for driver training based on multi-data fusion according to claim 2, characterized in that, The seat height calculation unit includes a seat height correction subunit and an interval mapping subunit; the seat height correction subunit is used to perform weighted calculations on the fore-and-aft position of the seat and the backrest angle based on a preset seat height correction algorithm to obtain the corrected first driver seat height. The interval mapping subunit, connected to the seat height correction subunit, is used to map the first driver seat height to a standard seat height interval obtained based on historical driving data.

4. The intelligent monitoring and early warning system for driver training based on multi-data fusion as described in claim 1, characterized in that, The driving posture judgment module includes a light intensity analysis unit and a body posture analysis unit. The light intensity analysis unit is used to determine whether the facial light intensity is greater than the facial light intensity threshold, and analyzes the light source category when it is greater. The body posture analysis unit is connected to the light intensity analysis unit and is used to perform body tilt angle analysis when the facial light intensity is less than or equal to the facial light intensity threshold. After obtaining the body tilt angle analysis result, it performs face orientation analysis to determine whether the driving posture corresponding to the face orientation analysis result is abnormal.

5. The intelligent monitoring and early warning system for driver training based on multi-data fusion according to claim 4, characterized in that, The body posture analysis unit includes a tilt angle analysis subunit; the tilt angle analysis subunit is used to compare the body tilt angle with the body tilt angle safety value to obtain the tilt angle comparison result.

6. The intelligent monitoring and early warning system for driver training based on multi-data fusion according to claim 5, characterized in that, The body posture analysis unit also includes a face orientation analysis subunit; the face orientation analysis subunit is connected to the tilt angle analysis subunit, and is used to obtain the face orientation angle comparison result obtained by comparing the face orientation angle with the face orientation angle threshold when the tilt angle analysis is obtained, and to determine whether the corresponding driving posture is abnormal.

7. The intelligent monitoring and early warning system for driver training based on multi-data fusion according to claim 1, characterized in that, The driving operation analysis module includes a driving area division unit and an obstacle analysis unit; The driving area division unit is used to construct the driving area based on the vehicle's driving coordinates and direction, and to divide the driving area into the first driving area, the second driving area and the third driving area; The obstacle analysis unit, connected to the driving area division unit, is used to determine whether an obstacle is in the second or third driving area when driving.

8. The intelligent monitoring and early warning system for driver training based on multi-data fusion according to claim 7, characterized in that, The obstacle behavior analysis unit includes a region judgment subunit and a braking analysis subunit; The area determination subunit is used to determine whether an obstacle is located in the second or third driving area; The braking analysis subunit, connected to the area judgment subunit, is used to analyze whether there is braking behavior in the time period from when the obstacle enters the area to the current time, and outputs the first driving operation behavior or the second driving operation behavior accordingly.

9. The intelligent monitoring and early warning system for driver training based on multi-data fusion according to claim 1, characterized in that, The abnormal status statistics module includes an abnormal counting unit and a duration mapping unit; the abnormal counting unit is used to count the total number of abnormal status reminders received by the target learner during this driving training session. The duration mapping unit is connected to the anomaly counting unit and is used to obtain the driving training duration based on the preset mapping relationship between the total number of anomaly status alerts and the driving training duration.

Images