A method and device for cumulative monitoring and identification of vehicle runout

Abstract

The application discloses a kind of vehicle deviation cumulative monitoring identification method and device, it is related to intelligent transportation and automobile electronic control technical field.The method comprises: obtaining vehicle multimodal operation data and preprocessing;When vehicle speed is greater than preset speed threshold, steering wheel angle absolute value is less than preset angle threshold, and vehicle deflection data meets deviation determination condition, deviation event is determined and primary validity verification and secondary verification are carried out, and deflection direction is determined;In preset time window, based on the duration of valid deviation event and deflection severity, the cumulative deviation representation value of each deflection direction is generated;Record the cumulative deviation representation value of continuous preset time window to form deviation trend data sequence, and carry out trend analysis, obtain change trend characteristic value;Deviation reason is identified based on vehicle state data, and deviation deterioration grade is judged according to change trend characteristic value, and output warning information.The application realizes deviation reason tracing and deterioration grade evaluation, and reduces deviation false alarm rate.

Description

Technical Field

[0001] This invention relates to the field of intelligent transportation and automotive electronic control technology, and in particular to a method and device for cumulative monitoring and identification of vehicle deviation. Background Technology

[0002] Vehicle deviation is a common abnormal phenomenon during vehicle operation. Currently, with the rapid development of advanced driver assistance systems and autonomous driving technologies, higher demands are placed on the accurate perception of vehicle posture and the early identification of abnormal states. Existing technologies already exist that utilize multi-source sensors to monitor vehicle driving status, providing a hardware foundation for the automatic detection of vehicle deviation.

[0003] Most existing vehicle drift detection technologies rely on instantaneous sensor threshold triggering, meaning that if the motion sensor signal exceeds a fixed threshold at a certain moment, the vehicle is directly determined to be drifting. This type of detection method is susceptible to interference from external conditions such as crosswinds, cornering, instantaneous noise from lane markings, and road bumps, resulting in significant false triggers and alarms. Furthermore, drift caused by chassis component performance degradation is often a gradual and slow process, making it difficult for instantaneous signals to capture early, subtle drift characteristics and detect the deterioration trend in a timely manner. In addition, most existing technologies only indicate whether drift has occurred or not, failing to differentiate the severity of the drift and unable to combine multi-source chassis status information to further trace the specific causes of the drift. They only issue alarms after the drift has become very significant, hindering early identification and proactive maintenance of chassis anomalies. Summary of the Invention

[0004] To address the problems of existing vehicle deviation detection technologies that rely on instantaneous sensor thresholds for judgment, are susceptible to external operating conditions causing false alarms, struggle to capture gradual and subtle deviations caused by chassis deterioration, cannot distinguish the degree of deviation worsening, and cannot trace the cause of deviation, this invention proposes a cumulative monitoring and identification method and device for vehicle deviation.

[0005] The present invention achieves the above objectives through the following technical solutions: A cumulative monitoring and identification method for vehicle deviation includes: Acquire and preprocess multimodal operating data during vehicle operation; the multimodal operating data includes at least steering wheel angle data, vehicle yaw data, vehicle status data, road curvature recognition data, and vehicle speed data; When the vehicle speed data is greater than the preset vehicle speed threshold, the absolute value of the steering wheel angle data is less than the preset angle threshold, and the vehicle deflection data meets the deviation judgment condition, a deviation event is determined to have occurred, and the validity of the deviation event is verified; the validity verification includes primary validity verification and secondary verification. The deviation event that passes the validity verification is regarded as a valid deviation event, and the deflection direction of the valid deviation event is determined according to the vehicle deflection data; the deflection direction includes a first direction and a second direction. Within each preset time window, based on each valid deviation event that occurs within that preset time window, a cumulative deviation characterization value corresponding to each deviation direction is generated for that preset time window; the cumulative deviation characterization value is positively correlated with the duration and severity of each valid deviation event. The cumulative deviation characterization values ​​corresponding to each deflection direction are recorded in multiple consecutive preset time windows to form deviation trend data sequences corresponding to each deflection direction. When the number of data points in any deviation trend data sequence reaches the first preset number, the trend analysis is performed on the most recent first preset number of data points in the deviation trend data sequence to obtain the change trend characteristic value of the deviation trend data sequence. The system identifies the cause of deviation based on vehicle status data, and determines the degree of deviation deterioration based on the characteristic values ​​of the changing trends corresponding to each deflection direction. It then outputs a warning message that includes the cause of deviation, the degree of deviation deterioration, and the determination of the deviation direction.

[0006] As a preferred embodiment of the present invention, the vehicle deflection data includes the vehicle's yaw rate, the vehicle's lateral acceleration, the vehicle's lateral offset relative to the lane line, and the vehicle's lateral offset velocity relative to the lane line. The vehicle status data includes at least one of the following: tire pressure monitoring data, steering wheel zero-position offset self-learning value, and suspension system status data. The preprocessing includes timing synchronization, anti-aliasing filtering, filtering and noise reduction, and outlier data removal.

[0007] As a preferred embodiment of the present invention, the deviation determination conditions include: The absolute value of the yaw rate is greater than a preset angular rate threshold, and the first duration is greater than a first preset duration threshold; or, The absolute value of the lateral offset velocity is greater than the preset offset velocity threshold, and the second duration is greater than the second preset duration threshold; Wherein, the first duration is the duration during which the absolute value of the yaw angular velocity is continuously greater than a preset angular velocity threshold; the second duration is the duration during which the absolute value of the lateral offset velocity is continuously greater than a preset offset velocity threshold.

[0008] As a preferred embodiment of the present invention, the primary validity verification includes: Within the determination window, the average yaw rate and the average lateral drift rate are calculated respectively; the determination window is the time period from the start time when any one of the deviation determination conditions is first met for the deviation event to the end time when all deviation determination conditions are no longer met. When the sign of the average yaw rate is the same as the sign of the average lateral drift rate, the misalignment event is deemed to have passed the initial validity verification; when the sign of the average yaw rate is different from the sign of the average lateral drift rate, the misalignment event is deemed to have failed the initial validity verification and is discarded. The secondary verification includes: If the high-frequency fluctuation component of the lateral acceleration is greater than the preset fluctuation threshold, it is determined to be crosswind interference, and the deviation event that has passed the primary validity verification is marked as an invalid deviation event; the high-frequency fluctuation component is the amplitude of the fluctuation component whose frequency is greater than the first preset frequency threshold after high-pass filtering of the lateral acceleration. If the road curvature recognition data indicates that the vehicle is driving on a curve, and the direction of the steering wheel angle data is consistent with the turning direction indicated by the road curvature recognition data, then it is determined to be a normal deviation caused by driving on a curve, and the deviation event that has passed the initial validity verification is marked as an invalid deviation event; the turning direction indicated by the road curvature recognition data is determined according to the road curvature.

[0009] As a preferred embodiment of the present invention, the valid deviation event is a deviation event that passes the initial validity verification and is not marked as invalid in the secondary verification. The step of determining the deflection direction of a valid deviation event based on vehicle deflection data includes: When a valid deviation event is determined solely based on yaw rate, if the average yaw rate is positive, the deflection direction is determined as the first direction; if the average yaw rate is negative, the deflection direction is determined as the second direction. When a valid deviation event is determined solely based on lateral offset velocity, if the average lateral offset velocity is positive, the deflection direction is determined as the first direction; if the average lateral offset velocity is negative, the deflection direction is determined as the second direction. When a valid deviation event is determined based on yaw rate and lateral offset rate, the sign of the average yaw rate or the average lateral offset rate is used to determine the deflection direction.

[0010] As a preferred embodiment of the present invention, the cumulative deviation characterization value includes a first cumulative deviation characterization value and a second cumulative deviation characterization value; the first cumulative deviation characterization value is generated based on each valid deviation event in the first direction within the preset time window; the second cumulative deviation characterization value is generated based on each valid deviation event in the second direction within the preset time window. For any deflection direction, the formula for calculating the cumulative deviation characterization value is: ; in, This represents the cumulative deviation characterization value; The total number of valid deviation events in that deflection direction within the preset time window; The sequence number of the valid deviation event; For the first The duration of each valid deviation event; For the first The normalized deflection severity index for each valid deviation event; the normalized deflection severity index is determined as follows: If the first If the determination of a valid deviation event is based solely on the yaw rate, then... ; If the first If the determination of a valid deviation event is based solely on the lateral offset speed, then... ; If the first The determination of a valid deviation event is based on the yaw rate and lateral deviation rate. The calculation formula is: ; in, For the first The absolute value of the maximum yaw rate within the judgment window for each valid deviation event; For the first The absolute value of the maximum lateral offset velocity within the judgment window for each valid deviation event; The preset yaw rate reference value; This is a preset reference value for lateral offset speed.

[0011] As a preferred embodiment of the present invention, the first direction and the second direction each maintain an independent deviation trend data sequence; the recording time of the data point is the end time of the preset time window, and each preset time window corresponds to a data point with a cumulative deviation characterization value. When there is no valid deviation event in the corresponding deviation direction within a preset time window, the cumulative deviation characterization value of the data point is recorded as zero. The trend analysis includes using the least squares method to perform linear regression fitting on the most recent first preset number of data points; the characteristic value of the trend is the slope of the straight line obtained by the linear regression fitting.

[0012] As a preferred embodiment of the present invention, the step of identifying the cause of deviation based on vehicle status data includes: If the tire pressure monitoring data in the detection window shows that the pressure of any tire is lower than the standard value by more than the preset tire pressure deviation threshold, or the pressure difference between the left and right tires on the same axle exceeds the preset pressure difference threshold, then the cause of the deviation is determined to be abnormal tire pressure; the detection window is the determination window for the most recent valid deviation event within the preset time window. If the absolute value of the self-learned value of the steering wheel zero offset is greater than the preset zero offset threshold, the cause of the deviation is determined to be abnormal steering wheel zero offset. If the suspension system status data in the detection window shows that the height difference between the left and right sides of the same axle is greater than the preset height difference threshold, or if there is a fault in the shock absorber, then the cause of the pull-off is determined to be an abnormal suspension. If the vehicle status data does not show the above-mentioned abnormalities, but in the trend analysis results of a second preset number of preset time windows, the characteristic value of the change trend in any deflection direction is continuously greater than or equal to the first characteristic threshold, then the cause of the deviation is determined to be four-wheel alignment misalignment.

[0013] As a preferred embodiment of the present invention, the step of determining the degree of deviation deterioration based on the characteristic values ​​of the changing trends corresponding to each deflection direction includes: When the characteristic value of the trend is less than the first characteristic threshold, it is determined that there is no deterioration trend; When the characteristic value of the trend is greater than or equal to the first characteristic threshold and less than the second characteristic threshold, the deviation deterioration level is determined to be Level 1. When the characteristic value of the trend is greater than or equal to the second characteristic threshold and less than the third characteristic threshold, the deviation deterioration level is determined to be level two. When the characteristic value of the trend is greater than or equal to the third characteristic threshold, the deviation deterioration level is determined to be level three; Wherein, the first feature threshold is less than the second feature threshold, the second feature threshold is less than the third feature threshold, and the first feature threshold is a positive value; The deviation direction determination result is based on the comparison results of the change trend feature values ​​corresponding to the first direction and the change trend feature values ​​corresponding to the second direction with preset consistency thresholds respectively: When the trend characteristic value corresponding to the first direction and the trend characteristic value corresponding to the second direction are both greater than or equal to the preset consistency threshold, it is determined to be a two-way deviation trend, and the higher of the deviation deterioration level determined by the first direction and the deviation deterioration level determined by the second direction is taken as the final deviation deterioration level. When the trend characteristic value of the change trend corresponding only in the first direction or the trend characteristic value of the change trend corresponding only in the second direction is greater than or equal to the preset consistency threshold, it is determined to be a one-way deviation trend, and the deviation deterioration level corresponding to that direction is taken as the final deviation deterioration level. When the trend characteristic value corresponding to the first direction and the trend characteristic value corresponding to the second direction are both less than the preset consistency threshold, it is determined that there is no obvious deviation trend and the state of no warning is maintained.

[0014] A cumulative monitoring and identification device for vehicle deviation includes: The data acquisition and processing module is used to acquire and preprocess multimodal operating data during vehicle operation; the multimodal operating data includes at least steering wheel angle data, vehicle yaw data, vehicle status data, road curvature recognition data, and vehicle speed data; The vehicle deviation event determination and verification module is used to determine that a deviation event has occurred when the vehicle speed data is greater than a preset vehicle speed threshold, the absolute value of the steering wheel angle data is less than a preset steering angle threshold, and the vehicle deviation data meets the deviation determination conditions, and to verify the validity of the deviation event; the validity verification includes primary validity verification and secondary verification. The deflection direction determination module is used to identify the deviation event that has passed the validity verification as a valid deviation event, and to determine the deflection direction of the deviation event based on the vehicle deflection data; the deflection direction includes a first direction and a second direction. The deviation accumulation module is used to generate a cumulative deviation characterization value for each deviation direction corresponding to each preset time window, based on each valid deviation event that occurs within the preset time window; the cumulative deviation characterization value is positively correlated with the duration and severity of each valid deviation event. The trend analysis module is used to record the cumulative deviation characterization values ​​corresponding to each deviation direction in multiple consecutive preset time windows, forming a deviation trend data sequence corresponding to each deviation direction. When the number of data points in any deviation trend data sequence reaches a first preset number, trend analysis is performed on the most recent first preset number of data points in the deviation trend data sequence to obtain the change trend characteristic value of the deviation trend data sequence. The early warning module is used to identify the cause of deviation based on vehicle status data, and to determine the degree of deviation deterioration based on the characteristic values ​​of the changing trends corresponding to each deflection direction. It outputs early warning information that includes the cause of deviation, the degree of deviation deterioration, and the determination result of the deviation direction.

[0015] The beneficial effects of this invention are as follows: By acquiring multimodal operational data, this invention comprehensively depicts the vehicle's dynamic attitude and driving environment. Simultaneously, in the initial judgment stage, it introduces dual pre-screening conditions: vehicle speed exceeding a preset speed threshold and the absolute value of the steering wheel angle being less than a preset angle threshold. This excludes low-speed invalid driving conditions and scenarios involving driver-initiated lane changes and turns, improving the steady-state nature of the conditions used for deviation monitoring and ensuring the effectiveness and accuracy of the judgment basis from the outset. Based on the initial judgment of a deviation event, a cascaded dual verification mechanism of primary validity verification and secondary verification is further introduced. This mechanism verifies the determined deviation events step by step, ensuring that all deviation events entering subsequent cumulative calculations are valid deviation events that have passed validity verification. This fundamentally reduces the false alarm rate and improves the reliability of the identification results. Furthermore, by combining the duration of effective deviation events with the severity of the deviation, a cumulative deviation characterization value is generated, comprehensively accumulating all effective deviation events occurring within a preset time window. Based on this, deviation trend data sequences are independently formed for the first and second directions respectively. When the number of data points reaches a first preset number, trend analysis is performed to obtain characteristic values ​​representing the changing trend of the deviation severity. This technique enables the method to keenly capture and quantify the tracking of subtle, gradual vehicle deviation and its evolutionary trend, providing early warning before severe deviation occurs and improving active driving safety. Finally, the cause of deviation is identified based on vehicle status data, and the degree of deviation deterioration is objectively judged according to the characteristic values ​​of the changing trend in each deflection direction. The final output is a warning message containing the cause of deviation, the degree of deterioration, and the determination of the deviation direction. This warning message is rich in dimensions and clearly hierarchical, providing not only intuitive risk warnings for drivers but also operational guidance for maintenance personnel, achieving a technological leap from passive alarm to proactive identification and maintenance guidance. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein: Figure 1 A flowchart of a cumulative monitoring and identification method for vehicle deviation provided by the present invention; Figure 2 This is a flowchart of the deviation event determination and validity verification in an embodiment of the present invention; Figure 3 This is a schematic diagram of the modular structure of a vehicle deviation cumulative monitoring and identification device provided by the present invention. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the described embodiments of the present invention are within the scope of protection of the present invention.

[0018] like Figure 1 As shown, this is an embodiment of the present invention, which provides a cumulative monitoring and identification method for vehicle deviation. This method is executed by a vehicle electronic control unit (ECU) or an on-board domain controller, and includes: S1: Acquire multimodal operating data during vehicle operation and perform preprocessing.

[0019] Multimodal operating data includes at least steering wheel angle data, vehicle yaw data, vehicle status data, road curvature recognition data, and vehicle speed data.

[0020] Steering wheel angle data is collected in real time by the vehicle's steering wheel angle sensor, with a sampling frequency of no less than 100Hz. This data is used to identify the driver's active steering operations, exclude vehicle directional deviations caused by manual lane changes or turns, and retain only the judgment scenario of the vehicle veering off course on its own.

[0021] Vehicle yaw data includes the vehicle's yaw rate, lateral acceleration, lateral offset relative to the lane lines, and lateral offset velocity relative to the lane lines. The yaw rate is acquired using an onboard yaw rate sensor, and the lateral acceleration is acquired using an onboard lateral acceleration sensor; both are sampled at a raw frequency of 100Hz. Lateral offset and lateral offset velocity are acquired by a lane-keeping camera mounted on the upper part of the vehicle's windshield. This lane-keeping camera captures images of the road ahead, uses the Canny edge detection algorithm combined with Hough transform to identify lane boundaries, and calculates the perpendicular distance between the vehicle's centerline and the lane centerline as the lateral offset. The instantaneous lateral offset velocity is obtained by performing a first-order difference operation on the lateral offsets of adjacent frames and using a length of... The frame's moving average window smooths the instantaneous lateral offset velocity to suppress speed noise caused by lane line recognition jumps; the formula for calculating the smoothed lateral offset velocity is: ; in, For the first The lateral offset speed of the frame; For the first The horizontal offset of the frame; For the first The horizontal offset of the frame; The frame rate for the lane-keeping camera is 30fps in this embodiment. Approximately 0.033s; The moving average window frame number is 5 in this embodiment.

[0022] After the lateral offset speed is calculated at the original frame rate of the lane keeping camera, it is uniformly resampled to 50Hz during the preprocessing stage.

[0023] Vehicle status data includes at least one of the following: tire pressure monitoring data, steering wheel zero-position offset self-learning value, and suspension system status data. Tire pressure monitoring data is collected by the vehicle's direct tire pressure monitoring system and includes the real-time pressure and temperature of each tire; the steering wheel zero-position offset self-learning value is stored and output by the electronic power steering system controller; and the suspension system status data is collected by the suspension height sensor or continuous damping control system to help identify the causes of vehicle drift due to chassis abnormalities.

[0024] Road curvature recognition data is obtained by identifying the geometry of lane lines through lane keeping cameras, or by fusing road topology information from ADAS navigation maps. The output parameter is road curvature. This road curvature recognition data is used to describe the degree of curvature and turning direction of the road the vehicle is currently traveling on.

[0025] Vehicle speed data is acquired through the vehicle's electronic stability system or wheel speed sensors, with a sampling frequency of no less than 100Hz.

[0026] Since the multimodal operational data comes from different sensors and has inconsistent sampling frequencies, preprocessing is required. Preprocessing includes timing synchronization, anti-aliasing filtering, noise reduction, and outlier removal. Details are as follows: For signals with an original sampling frequency higher than 50Hz, an anti-aliasing low-pass filter with a cutoff frequency of 20Hz is first applied. Then, using the original timestamps of each sensor as a reference, a first-order linear interpolation algorithm is employed to uniformly resample all multimodal operating data to 50Hz for strict timing alignment. Next, different types of data are filtered separately. Specifically, for steering wheel angle data and yaw rate, a second-order Butterworth low-pass filter with a cutoff frequency of 10Hz is used to filter out high-frequency excitation noise caused by road bumps and vehicle vibration. For lane lateral offset, a sliding window mid-range filtering algorithm with a window size of 5 frames is used to suppress jump noise generated by lane line recognition. For lateral acceleration, in addition to the 20Hz anti-aliasing low-pass filter, the 10Hz low-pass filter is not applied to retain high-frequency fluctuation components in the range of 2Hz to 20Hz for subsequent crosswind interference recognition. In addition, physical boundary thresholds are set for each sensor data, and data exceeding the normal driving physical range of the vehicle is marked as invalid and discarded. At the same time, a sliding window with a window length of 1 second is used to remove transient spikes from the multimodal operating data according to the 3σ criterion, retaining only the steady-state valid driving data for subsequent deviation determination. In a specific embodiment, the normal driving physical range of the vehicle includes: steering wheel angle [-900°, +900°], yaw rate [-100° / s, +100° / s], lateral acceleration [-15 m / s², +15 m / s²], lateral offset [-10m, +10m], lateral offset velocity [-2m / s, +2m / s], and vehicle speed [0, 300 km / h].

[0027] Furthermore, in the preprocessing stage, coordinate transformation is performed on the raw sensor data; after coordinate transformation, positive values ​​of yaw rate correspond to the vehicle turning right and negative values ​​correspond to the vehicle turning left; after coordinate transformation, positive values ​​of lateral offset rate correspond to the vehicle shifting to the right side of the lane and negative values ​​correspond to the vehicle shifting to the left side of the lane; after coordinate transformation, positive values ​​of steering wheel angle data correspond to turning the steering wheel to the right and negative values ​​correspond to turning the steering wheel to the left.

[0028] Different vehicle dynamics standards and sensor manufacturers may have different definitions of the positive and negative directions for yaw rate and lateral offset rate. For example, the ISO 8855 vehicle coordinate system specifies that the z-axis points downwards, and under the right-hand rule, a clockwise yaw rate when viewed from above is a positive value, corresponding to a rightward yaw. Some manufacturers use a coordinate system with the z-axis pointing upwards, where a positive yaw rate corresponds to a counterclockwise rotation, corresponding to a leftward yaw. Lateral offset and lateral offset rate are usually based on the lane centerline, and their positive and negative directions are determined by the direction of the y-axis of the coordinate system. Those skilled in the art can also make adaptive adjustments according to the actual coordinate system used, and the definition of the positive and negative directions does not constitute a limitation of the present invention.

[0029] S2, when the vehicle speed data is greater than the preset vehicle speed threshold, the absolute value of the steering wheel angle data is less than the preset angle threshold, and the vehicle deflection data meets the deviation judgment condition, a deviation event is determined to have occurred, and the validity of the deviation event is verified.

[0030] The preset vehicle speed threshold is calibrated according to the vehicle's driving conditions, with a range of 20km / h to 40km / h. In this embodiment, 30km / h is used to ensure that the vehicle is in an effective driving state and that the lane keeping camera can stably identify lane lines.

[0031] The preset steering angle threshold is calibrated based on the vehicle type and steering system transmission ratio, and its value ranges from 3° to 10°. In this embodiment, the preset steering angle threshold is 5°, which is applicable to conventional passenger vehicles. When the absolute value of the steering wheel angle data is greater than or equal to the preset steering angle threshold, it is determined that the driver has actively intervened in the steering, the current determination cycle is terminated, and all timers are reset to zero. The determination cycle refers to the continuous monitoring period from the time the last deviation event ended or the time the determination started to the current time; all timers include a timer for accumulating a first duration and a timer for accumulating a second duration.

[0032] like Figure 2 As shown, the process for determining and validating deviation events is as follows: A vehicle deviation event is determined to have occurred when the vehicle speed data exceeds a preset speed threshold, the absolute value of the steering wheel angle data is less than a preset angle threshold, and the vehicle deflection data meets the deviation determination criteria. The deviation determination criteria include: Condition A: The absolute value of the yaw rate is greater than a preset angular rate threshold, and the first duration is greater than a first preset duration threshold; or, Condition B: The absolute value of the lateral offset speed is greater than the preset offset speed threshold, and the second duration is greater than the second preset duration threshold.

[0033] The first duration is the duration during which the absolute value of the yaw angular velocity is continuously greater than a preset angular velocity threshold; the second duration is the duration during which the absolute value of the lateral offset velocity is continuously greater than a preset offset velocity threshold. Both the first and second durations are accumulated using the sampling period after 50Hz resampling as the timing unit. When the absolute value of the corresponding physical quantity falls below the threshold, the corresponding timer is reset to zero.

[0034] In a preferred embodiment, when the vehicle speed is below 80 km / h, the preset angular velocity threshold is set to 2° / s; when the vehicle speed is above 80 km / h, the preset angular velocity threshold adaptively decreases to 1.5° / s; the first preset duration threshold is set to 1.0s. The preset offset speed threshold is related to the lane width and camera performance, and its value ranges from 0.1m / s to 0.3m / s, preferably 0.15m / s; the second preset duration threshold is set to 1.2s, which can avoid false triggering caused by instantaneous lane line recognition noise and minor adjustments by the driver within the lane.

[0035] Validation includes primary validity validation and secondary validity validation.

[0036] Primary validity verification includes: Within the decision window, the average yaw rate and the average lateral drift rate are calculated separately. The decision window is the time period from the start time when any of the deviation criteria is first met in the deviation event to the end time when all deviation criteria are no longer met. If the length of this time period is less than 2 seconds, the decision window is calculated by looking back 2 seconds from the end time. The extended decision window may include data before the start time of the event; this data is used to provide a more stable orientation estimation benchmark. The backtracking window does not cross the end time of the previous deviation event. Those skilled in the art can adjust the backtracking duration as needed.

[0037] When the sign of the average yaw rate is the same as the sign of the average lateral drift rate, the misalignment event is deemed to have passed the initial validity verification; when the sign of the average yaw rate is different from the sign of the average lateral drift rate, the misalignment event is deemed to have failed the initial validity verification and is discarded.

[0038] Secondary verification includes: If the high-frequency fluctuation component of the lateral acceleration exceeds a preset fluctuation threshold, it is determined to be crosswind interference, and the deviation event that passed the initial validity verification is marked as an invalid deviation event. The high-frequency fluctuation component is the amplitude of the fluctuation component with a frequency greater than the first preset frequency threshold obtained after high-pass filtering of the lateral acceleration. The lateral acceleration signal within the determination window is filtered by a fourth-order Butterworth high-pass filter with a cutoff frequency of 2Hz to extract the high-frequency fluctuation component with a frequency greater than 2Hz. The amplitude of the high-frequency fluctuation component is taken as the 90th percentile of the absolute value of the high-pass filtered signal within the determination window. In a preferred embodiment, the first preset frequency threshold is 2Hz, and the preset fluctuation threshold is 0.5m / s².

[0039] If the road curvature recognition data indicates that the vehicle is traveling on a curve, and the direction of the steering wheel angle data is consistent with the turning direction indicated by the road curvature recognition data, then it is determined to be a normal deviation caused by curve driving. This deviation event, which passes the initial validity verification, is marked as an invalid deviation event. Specifically, the maximum absolute value of the road curvature within the determination window is extracted from the road curvature recognition data. If the maximum absolute value of the road curvature is greater than a preset curvature threshold, then the vehicle is determined to be traveling on a curve. This preset curvature threshold is set to 0.002. This corresponds to curves with a radius of curvature less than 500m. If the steering wheel angle direction matches the turning direction indicated by the road curvature recognition result (i.e., the steering wheel turns right and the road curves to the right, or the steering wheel turns left and the road curves to the left), then the current deviation is determined to be due to normal curve driving and is not considered vehicle deviation. The direction of the steering wheel angle data is determined by the positive or negative sign after coordinate transformation via S1 preprocessing; a positive value corresponds to turning the steering wheel to the right, and a negative value corresponds to turning the steering wheel to the left. The turning direction indicated by the road curvature recognition data is determined based on the road curvature.

[0040] S3, the deviation event that passes the validity verification is taken as a valid deviation event, and the deviation direction of the valid deviation event is determined based on the vehicle deviation data.

[0041] A valid deviation event is one that passes the initial validity verification and is not marked as invalid in the secondary verification. Each valid deviation event includes an event number, deviation direction, absolute value of maximum yaw rate, absolute value of maximum lateral drift rate, duration, trigger mode identifier, and event termination time. The trigger mode identifier indicates whether the event was triggered by condition A, condition B, or both conditions A and B. The absolute values ​​of maximum yaw rate, maximum lateral drift rate, and duration are all calculated within the decision window.

[0042] The direction of deflection for valid deviation events is determined based on vehicle deflection data, including: When a valid deviation event is determined solely based on yaw rate, if the average yaw rate is positive, the deflection direction is determined as the first direction; if the average yaw rate is negative, the deflection direction is determined as the second direction.

[0043] When a valid deviation event is determined solely based on lateral offset speed, if the average lateral offset speed is positive, the deflection direction is determined as the first direction; if the average lateral offset speed is negative, the deflection direction is determined as the second direction.

[0044] When a valid deviation event is determined based on yaw rate and lateral drift rate, the deflection direction is determined by the sign of either the average yaw rate or the average lateral drift rate. Since the valid deviation event has passed the initial validity verification, the signs of the average yaw rate and the average lateral drift rate are consistent, and either the sign of the average yaw rate or the average lateral drift rate can be used to determine the deflection direction.

[0045] In this embodiment, the first direction is the direction in which the vehicle veers / shifts to the right, and the second direction is the direction in which the vehicle veers / shifts to the left.

[0046] S4. Within each preset time window, based on each valid deviation event that occurs within the preset time window, generate the cumulative deviation characterization value corresponding to each deviation direction for that preset time window.

[0047] In this embodiment, a continuous sliding statistical window is used as the preset time window, with a window length of 60 seconds and a sliding step size of 1 second. The preset time window length is adjusted within the range of 30 seconds to 120 seconds, and the specific value is calibrated according to the vehicle driving scenario and diagnostic response speed requirements. When the vehicle's real-time speed is continuously lower than the preset minimum speed of 20 km / h for 5 seconds, the preset time window is automatically extended to 120 seconds, and the sliding step size is adjusted accordingly to 2 seconds to ensure that there are still enough effective deviation events included in the statistics under low-speed conditions; when the vehicle speed recovers to above 20 km / h, the preset time window automatically returns to 60 seconds, and the sliding step size returns to 1 second to avoid sudden changes in diagnostic data caused by abrupt changes in operating conditions.

[0048] Each valid deviation event is included in the preset time window where the termination time is located for cumulative calculation based on the termination time of the valid deviation event; if the duration of a valid deviation event exceeds the boundary of the preset time window, it is included in the preset time window according to the complete duration of the valid deviation event, without being truncated by the preset time window.

[0049] The cumulative deviation characterization value is positively correlated with the duration and severity of each valid deviation event, including a first cumulative deviation characterization value and a second cumulative deviation characterization value; the first cumulative deviation characterization value is generated based on each valid deviation event in the first direction within the preset time window; the second cumulative deviation characterization value is generated based on each valid deviation event in the second direction within the preset time window.

[0050] For any deflection direction, the formula for calculating the cumulative deviation characterization value is: ; in, The cumulative deviation characterization value represents the overall intensity of the deviation event in the corresponding deviation direction. The larger the value, the stronger the overall deviation effect in that deviation direction. The total number of valid deviation events in that deflection direction within the preset time window; The sequence number of the valid deviation event; For the first Normalized deflection severity index for each effective deviation event; For the first The duration of a valid deviation event is defined as the time interval from the start time when the deviation event first meets the deviation judgment condition to the end time when the deviation event no longer meets all deviation judgment conditions.

[0051] The normalized deflection severity index is determined as follows: If the first If the determination of a valid deviation event is based solely on the yaw rate, then... .

[0052] If the first If the determination of a valid deviation event is based solely on the lateral offset speed, then... .

[0053] If the first The determination of a valid deviation event is based on the yaw rate and lateral deviation rate. The calculation formula is: ; in, For the first The absolute value of the maximum yaw rate within the judgment window for each valid deviation event; For the first The absolute value of the maximum lateral offset velocity within the judgment window for each valid deviation event; The preset yaw rate reference value; This is a preset reference value for lateral offset speed.

[0054] In a preferred embodiment ,in, A preset angular velocity threshold is set for real-time operating conditions. This preset angular velocity threshold is adaptively selected based on the current vehicle speed. When the vehicle speed is greater than the preset speed threshold but less than 80 km / h, the preset angular velocity threshold is 2° / s. When the vehicle speed is greater than or equal to 80 km / h, the preset angular velocity threshold is 1.5° / s. ,in, The preset offset velocity threshold is preferably 0.15 m / s. When the normalized deflection severity index is calculated... When the value is greater than 1, force to Set to 1.

[0055] Furthermore, if the first cumulative deviation characterization value Or the second cumulative deviation characterization value In the middle, any value is greater than or equal to the preset cumulative threshold. If this is the case, it is determined that the vehicle has a tendency to veer off course in the corresponding direction within the current preset time window. The preset cumulative threshold is... The value is obtained through real vehicle road tests and vehicle dynamics model calibration; in this embodiment, 15 seconds is used.

[0056] Optionally, the first cumulative deviation characterization value can also be calculated. Compared with the second cumulative deviation characterization value The difference .when Further investigation confirmed that the vehicle was continuously veering off course on the side with the larger cumulative deviation value. The preset difference threshold is set to 10s in this embodiment.

[0057] S5, record the cumulative deviation characterization values ​​corresponding to each deflection direction for multiple consecutive preset time windows, forming a deviation trend data sequence corresponding to each deflection direction. When the number of data points in any deviation trend data sequence reaches the first preset number, perform trend analysis on the most recent first preset number of data points in the deviation trend data sequence to obtain the change trend characteristic value of the deviation trend data sequence.

[0058] The first and second directions each maintain an independent deviation trend data sequence. The recording time of the data point is the end time of the preset time window, and each preset time window corresponds to a data point with a cumulative deviation characterization value. When there is no valid deviation event in the corresponding deviation direction within a preset time window, the cumulative deviation characterization value of the data point is recorded as zero. In this embodiment, the sliding step size of the sliding statistical window is 1 second, therefore, the time interval between adjacent data points is 1 second. When the sliding step size is adjusted to 2 seconds under low-speed conditions, the time interval between adjacent data points is correspondingly 2 seconds. The trend analysis uses the actual recording time of the data point as the time coordinate and is not affected by changes in the data point interval. The first preset quantity is 30 in this embodiment.

[0059] Trend analysis includes using the least squares method to perform linear regression fitting on the most recent first preset number of data points; the characteristic value of the trend is the slope of the straight line obtained by the linear regression fitting.

[0060] In one specific embodiment, the change trend characteristic value The average change in the cumulative deviation value in this direction per second is calculated using the following formula: ; in, The index of the data point; For the first The recording time corresponding to each data point; For the first The cumulative deviation characteristic value corresponding to each data point; For the most recent The average value at each data point is recorded at that time. For the most recent The average of the cumulative deviation characteristics of each data point; The number of data points used for fitting is equal to the first preset number.

[0061] when When the value is positive, it indicates that the cumulative deviation in that direction increases over time, and the degree of deviation shows a worsening trend; when... When the value is negative, it indicates that the cumulative deviation characteristic value decreases over time, and the deviation degree shows an improving trend; when When the value is close to 0, it indicates that the deviation remains stable.

[0062] S6 identifies the cause of deviation based on vehicle status data, and determines the degree of deviation deterioration based on the characteristic values ​​of the changing trends corresponding to each deflection direction, outputting warning information including the cause of deviation, the degree of deviation deterioration, and the determination result of the deviation direction.

[0063] Identifying the causes of vehicle deviation based on vehicle status data includes: If the tire pressure monitoring data in the detection window shows that the pressure of any tire is lower than the standard value by more than the preset tire pressure deviation threshold, or the pressure difference between the left and right tires on the same axle exceeds the preset pressure difference threshold, then the cause of the vehicle pulling to one side is determined to be abnormal tire pressure. In this embodiment, the preset tire pressure deviation threshold is set to 20 kPa, and the preset pressure difference threshold is set to 15 kPa.

[0064] If the absolute value of the self-learned value of the steering wheel zero-position offset is greater than the preset zero-position offset threshold, the cause of the deviation is determined to be an abnormal steering wheel zero-position offset. In this embodiment, the preset zero-position offset threshold is set to 3°.

[0065] If the suspension system status data in the detection window shows that the height difference between the left and right sides of the vehicle body on the same axle is greater than a preset height difference threshold, or if there is a malfunction in the shock absorber, then the cause of the pull-off is determined to be a suspension malfunction. In this embodiment, the preset height difference threshold is set to 15mm.

[0066] If the vehicle status data shows no abnormalities, but the trend analysis results for a second preset number of preset time windows show that the characteristic value of the change trend in any deflection direction is consistently greater than or equal to the first characteristic threshold, then the cause of the deviation is determined to be a misalignment of the four-wheel alignment. The second preset number is an independent calibration parameter, independent of the first preset number participating in trend fitting in S5. In this embodiment, both are the same value, 30.

[0067] The detection window is the judgment window for the most recent valid deviation event within the preset time window; when there are multiple valid deviation events within the preset time window, the vehicle status data corresponding to the union of all valid deviation event judgment windows is selected for judgment.

[0068] Obtain the trend characteristic values ​​for the first direction and the second direction respectively. For any trend characteristic value, determine the degree of deviation worsening in that direction according to the following rules: When the characteristic value of the trend is less than the first characteristic threshold, it is determined that there is no deterioration trend.

[0069] When the characteristic value of the trend is greater than or equal to the first characteristic threshold and less than the second characteristic threshold, the deviation deterioration level is determined to be Level 1.

[0070] When the trend characteristic value is greater than or equal to the second characteristic threshold and less than the third characteristic threshold, the deviation deterioration level is determined to be level two.

[0071] When the characteristic value of the trend is greater than or equal to the third characteristic threshold, the deviation deterioration level is determined to be level three.

[0072] Among them, the first feature threshold is less than the second feature threshold, the second feature threshold is less than the third feature threshold, and the first feature threshold is a positive value.

[0073] In this embodiment, the first feature threshold is set to 0.01, the second feature threshold is set to 0.03, and the third feature threshold is set to 0.06. These feature thresholds are calibrated based on real-vehicle road tests and vehicle dynamics models.

[0074] The deviation direction determination result is based on the comparison results of the change trend feature values ​​corresponding to the first direction and the change trend feature values ​​corresponding to the second direction with preset consistency thresholds respectively: When the trend characteristic value corresponding to the first direction and the trend characteristic value corresponding to the second direction are both greater than or equal to the preset consistency threshold, it is determined to be a two-way deviation trend, and the higher of the deviation deterioration level determined by the first direction and the deviation deterioration level determined by the second direction is taken as the final deviation deterioration level.

[0075] When the trend characteristic value corresponding only to the first direction or the trend characteristic value corresponding only to the second direction is greater than or equal to the preset consistency threshold, it is determined to be a unidirectional deviation trend, and the deviation deterioration level corresponding to that direction is taken as the final deviation deterioration level.

[0076] When the trend characteristic value corresponding to the first direction and the trend characteristic value corresponding to the second direction are both less than the preset consistency threshold, it is determined that there is no obvious deviation trend and the state of no warning is maintained.

[0077] In this embodiment, the preset consistency threshold is set to 0.01, which is the same as the first feature threshold. That is, when the characteristic value of the change trend in a certain direction reaches the first feature threshold, it is considered that there is an identifiable deviation trend in that direction. The preset consistency threshold needs to be adjusted according to the actual vehicle calibration.

[0078] The final output warning information can be sent to the vehicle's instrument panel for human-machine interaction prompts, or stored in the vehicle's fault memory for external maintenance and diagnostic equipment to read. For example, the cause of the pull-off is: misalignment of the four wheels; the level of pull-off deterioration is: level two; the direction of pull-off is determined as: one-way - pulling to the right.

[0079] like Figure 3 As shown, another embodiment of the present invention provides a vehicle drift cumulative monitoring and identification device for performing a vehicle drift cumulative monitoring and identification method, including: The data acquisition and processing module is used to acquire and preprocess multimodal operating data during vehicle operation. The multimodal operating data includes at least steering wheel angle data, vehicle yaw data, vehicle status data, road curvature recognition data, and vehicle speed data.

[0080] The vehicle deviation event detection and verification module is used to determine that a deviation event has occurred when the vehicle speed data is greater than a preset vehicle speed threshold, the absolute value of the steering wheel angle data is less than a preset steering angle threshold, and the vehicle deviation data meets the deviation detection conditions. The module then verifies the validity of the deviation event. Validity verification includes primary validity verification and secondary verification.

[0081] The deflection direction determination module is used to identify valid deviation events as effective deviation events and determine the deflection direction of the deviation event based on vehicle deflection data. The deflection direction includes a first direction and a second direction.

[0082] The deviation accumulation module is used to generate cumulative deviation characterization values ​​for each deviation direction within each preset time window, based on each valid deviation event that occurs within that preset time window. The cumulative deviation characterization value is positively correlated with the duration and severity of each valid deviation event.

[0083] The trend analysis module is used to record the cumulative deviation characterization values ​​corresponding to each deviation direction in multiple consecutive preset time windows, forming a deviation trend data sequence corresponding to each deviation direction. When the number of data points in any deviation trend data sequence reaches a first preset number, trend analysis is performed on the most recent first preset number of data points in the deviation trend data sequence to obtain the change trend characteristic value of the deviation trend data sequence.

[0084] The early warning module is used to identify the cause of deviation based on vehicle status data, and to determine the degree of deviation deterioration based on the characteristic values ​​of the changing trends corresponding to each deflection direction. It outputs early warning information that includes the cause of deviation, the degree of deviation deterioration, and the determination result of the deviation direction.

[0085] In summary, this invention first acquires multimodal operational data and performs rigorous preprocessing, including time synchronization, anti-aliasing filtering, and noise reduction, to suppress data interference caused by inconsistent sampling frequencies of multi-source heterogeneous sensors and vehicle vibration noise, thus building a high-quality data foundation for accurate monitoring and identification of vehicle deviation. Secondly, it proposes a dual validity verification mechanism for deviation events. Primary verification is performed through the sign consistency of yaw angular velocity and lateral offset velocity, followed by secondary verification by integrating high-frequency components of lateral acceleration and road curvature. This eliminates non-genuine deviation events such as crosswind interference and normal driving on curves, thereby reducing the false alarm rate at its source. Furthermore, it abandons the traditional single-instantaneous judgment mode and constructs a cumulative deviation characterization value that couples the duration of valid deviation events with the normalized deflection severity. Long-term trend fitting analysis is then performed using a sliding time window and least squares method, enabling the keen capture and quantification of subtle, gradual deviations and their worsening trends at the vehicle window scale. Finally, it deeply integrates vehicle state data and trend feature values ​​to achieve the identification of the physical root causes of deviation and the objective classification of deviation deterioration levels. The entire technical solution not only provides drivers with multi-dimensional early warning information, but also provides scientific and quantitative decision-making basis for the proactive maintenance and condition assessment of the vehicle chassis system.

[0086] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various variations or substitutions within the technical scope disclosed in this application, and these should all be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A cumulative monitoring and identification method for vehicle deviation, characterized in that, include: Acquire and preprocess multimodal operational data during vehicle operation; The multimodal operation data includes at least steering wheel angle data, vehicle yaw data, vehicle status data, road curvature recognition data, and vehicle speed data; When the vehicle speed data is greater than the preset vehicle speed threshold, the absolute value of the steering wheel angle data is less than the preset angle threshold, and the vehicle deflection data meets the deviation judgment condition, a deviation event is determined to have occurred, and the validity of the deviation event is verified; the validity verification includes primary validity verification and secondary verification. The deviation event that passes the validity verification is regarded as a valid deviation event, and the deflection direction of the valid deviation event is determined according to the vehicle deflection data; the deflection direction includes a first direction and a second direction. Within each preset time window, based on each valid deviation event that occurs within that preset time window, a cumulative deviation characterization value corresponding to each deviation direction is generated for that preset time window; the cumulative deviation characterization value is positively correlated with the duration and severity of each valid deviation event. The cumulative deviation characterization values ​​corresponding to each deflection direction are recorded in multiple consecutive preset time windows to form deviation trend data sequences corresponding to each deflection direction. When the number of data points in any deviation trend data sequence reaches the first preset number, the trend analysis is performed on the most recent first preset number of data points in the deviation trend data sequence to obtain the change trend characteristic value of the deviation trend data sequence. The system identifies the cause of deviation based on vehicle status data, and determines the degree of deviation deterioration based on the characteristic values ​​of the changing trends corresponding to each deflection direction. It then outputs a warning message that includes the cause of deviation, the degree of deviation deterioration, and the determination of the deviation direction.

2. The cumulative monitoring and identification method for vehicle deviation according to claim 1, characterized in that, The vehicle yaw data includes the vehicle's yaw rate, lateral acceleration, lateral offset relative to the lane lines, and lateral offset velocity relative to the lane lines. The vehicle status data includes at least one of the following: tire pressure monitoring data, steering wheel zero-position offset self-learning value, and suspension system status data. The preprocessing includes timing synchronization, anti-aliasing filtering, filtering and noise reduction, and outlier data removal.

3. The cumulative monitoring and identification method for vehicle deviation according to claim 2, characterized in that, The conditions for determining deviation include: The absolute value of the yaw rate is greater than a preset angular rate threshold, and the first duration is greater than a first preset duration threshold; or, The absolute value of the lateral offset velocity is greater than the preset offset velocity threshold, and the second duration is greater than the second preset duration threshold; Wherein, the first duration is the duration during which the absolute value of the yaw angular velocity is continuously greater than a preset angular velocity threshold; the second duration is the duration during which the absolute value of the lateral offset velocity is continuously greater than a preset offset velocity threshold.

4. The cumulative monitoring and identification method for vehicle deviation according to claim 3, characterized in that, The primary validity verification includes: Within the determination window, the average yaw rate and the average lateral drift rate are calculated respectively; the determination window is the time period from the start time when any one of the deviation determination conditions is first met for the deviation event to the end time when all deviation determination conditions are no longer met. When the sign of the average yaw rate is the same as the sign of the average lateral drift rate, the misalignment event is deemed to have passed the initial validity verification; when the sign of the average yaw rate is different from the sign of the average lateral drift rate, the misalignment event is deemed to have failed the initial validity verification and is discarded. The secondary verification includes: If the high-frequency fluctuation component of the lateral acceleration is greater than the preset fluctuation threshold, it is determined to be crosswind interference, and the deviation event that has passed the primary validity verification is marked as an invalid deviation event; the high-frequency fluctuation component is the amplitude of the fluctuation component whose frequency is greater than the first preset frequency threshold after high-pass filtering of the lateral acceleration. If the road curvature recognition data indicates that the vehicle is driving on a curve, and the direction of the steering wheel angle data is consistent with the turning direction indicated by the road curvature recognition data, then it is determined to be a normal deviation caused by driving on a curve, and the deviation event that has passed the initial validity verification is marked as an invalid deviation event; the turning direction indicated by the road curvature recognition data is determined according to the road curvature.

5. The cumulative monitoring and identification method for vehicle deviation according to claim 4, characterized in that, The valid deviation event is a deviation event that passes the initial validity verification and is not marked as invalid in the secondary verification. The step of determining the deflection direction of a valid deviation event based on vehicle deflection data includes: When a valid deviation event is determined solely based on yaw rate, if the average yaw rate is positive, the deflection direction is determined as the first direction; if the average yaw rate is negative, the deflection direction is determined as the second direction. When a valid deviation event is determined solely based on lateral offset velocity, if the average lateral offset velocity is positive, the deflection direction is determined as the first direction; if the average lateral offset velocity is negative, the deflection direction is determined as the second direction. When a valid deviation event is determined based on yaw rate and lateral offset rate, the sign of the average yaw rate or the average lateral offset rate is used to determine the deflection direction.

6. The cumulative monitoring and identification method for vehicle deviation according to claim 5, characterized in that, The cumulative deviation characterization value includes a first cumulative deviation characterization value and a second cumulative deviation characterization value; the first cumulative deviation characterization value is generated based on each valid deviation event in the first direction within the preset time window; The second cumulative deviation characterization value is generated based on each valid deviation event in the second direction within the preset time window; For any deflection direction, the formula for calculating the cumulative deviation characterization value is: ； in, This represents the cumulative deviation characterization value; The total number of valid deviation events in that deflection direction within the preset time window; The sequence number of the valid deviation event; For the first The duration of each valid deviation event; For the first The normalized deflection severity index for each valid deviation event; the normalized deflection severity index is determined as follows: If the first If the determination of a valid deviation event is based solely on the yaw rate, then... ; If the first If the determination of a valid deviation event is based solely on the lateral offset speed, then... ; If the first The determination of a valid deviation event is based on the yaw rate and lateral deviation rate. The calculation formula is: ； in, For the first The absolute value of the maximum yaw rate within the judgment window for each valid deviation event; For the first The absolute value of the maximum lateral offset velocity within the judgment window for each valid deviation event; The preset yaw rate reference value; This is a preset reference value for lateral offset speed.

7. The cumulative monitoring and identification method for vehicle deviation according to claim 6, characterized in that, The first direction and the second direction each maintain an independent deviation trend data sequence; the recording time of the data point is the end time of the preset time window, and each preset time window corresponds to a data point with a cumulative deviation characterization value. When there is no valid deviation event in the corresponding deviation direction within a preset time window, the cumulative deviation characterization value of the data point is recorded as zero. The trend analysis includes using the least squares method to perform linear regression fitting on the most recent first preset number of data points; the characteristic value of the trend is the slope of the straight line obtained by the linear regression fitting.

8. The cumulative monitoring and identification method for vehicle deviation according to claim 7, characterized in that, The method of identifying the cause of deviation based on vehicle status data includes: If the tire pressure monitoring data in the detection window shows that the pressure of any tire is lower than the standard value by more than the preset tire pressure deviation threshold, or the pressure difference between the left and right tires on the same axle exceeds the preset pressure difference threshold, then the cause of the deviation is determined to be abnormal tire pressure; the detection window is the determination window for the most recent valid deviation event within the preset time window. If the absolute value of the self-learned value of the steering wheel zero offset is greater than the preset zero offset threshold, the cause of the deviation is determined to be abnormal steering wheel zero offset. If the suspension system status data in the detection window shows that the height difference between the left and right sides of the same axle is greater than the preset height difference threshold, or if there is a fault in the shock absorber, then the cause of the pull-off is determined to be an abnormal suspension. If the vehicle status data does not show the above-mentioned abnormalities, but in the trend analysis results of a second preset number of preset time windows, the characteristic value of the change trend in any deflection direction is continuously greater than or equal to the first characteristic threshold, then the cause of the deviation is determined to be four-wheel alignment misalignment.

9. The cumulative monitoring and identification method for vehicle deviation according to claim 8, characterized in that, The method of determining the degree of deviation deterioration based on the characteristic values ​​of the changing trends corresponding to each deflection direction includes: When the characteristic value of the trend is less than the first characteristic threshold, it is determined that there is no deterioration trend; When the characteristic value of the trend is greater than or equal to the first characteristic threshold and less than the second characteristic threshold, the deviation deterioration level is determined to be Level 1. When the characteristic value of the trend is greater than or equal to the second characteristic threshold and less than the third characteristic threshold, the deviation deterioration level is determined to be level two. When the characteristic value of the trend is greater than or equal to the third characteristic threshold, the deviation deterioration level is determined to be level three; Wherein, the first feature threshold is less than the second feature threshold, the second feature threshold is less than the third feature threshold, and the first feature threshold is a positive value; The deviation direction determination result is based on the comparison results of the change trend feature values ​​corresponding to the first direction and the change trend feature values ​​corresponding to the second direction with preset consistency thresholds respectively: When the trend characteristic value corresponding to the first direction and the trend characteristic value corresponding to the second direction are both greater than or equal to the preset consistency threshold, it is determined to be a two-way deviation trend, and the higher of the deviation deterioration level determined by the first direction and the deviation deterioration level determined by the second direction is taken as the final deviation deterioration level. When the trend characteristic value of the change trend corresponding only in the first direction or the trend characteristic value of the change trend corresponding only in the second direction is greater than or equal to the preset consistency threshold, it is determined to be a one-way deviation trend, and the deviation deterioration level corresponding to that direction is taken as the final deviation deterioration level. When the trend characteristic value corresponding to the first direction and the trend characteristic value corresponding to the second direction are both less than the preset consistency threshold, it is determined that there is no obvious deviation trend and the state of no warning is maintained.

10. A cumulative monitoring and identification device for vehicle deviation, used to execute the cumulative monitoring and identification method for vehicle deviation as described in any one of claims 1 to 9, characterized in that, include: The data acquisition and processing module is used to acquire and preprocess multimodal operating data during vehicle operation. The multimodal operation data includes at least steering wheel angle data, vehicle yaw data, vehicle status data, road curvature recognition data, and vehicle speed data; The vehicle deviation event determination and verification module is used to determine that a deviation event has occurred when the vehicle speed data is greater than a preset vehicle speed threshold, the absolute value of the steering wheel angle data is less than a preset steering angle threshold, and the vehicle deviation data meets the deviation determination conditions, and to verify the validity of the deviation event; the validity verification includes primary validity verification and secondary verification. The deflection direction determination module is used to identify the deviation event that has passed the validity verification as a valid deviation event, and to determine the deflection direction of the deviation event based on the vehicle deflection data; the deflection direction includes a first direction and a second direction. The deviation accumulation module is used to generate a cumulative deviation characterization value for each deviation direction corresponding to each preset time window, based on each valid deviation event that occurs within the preset time window; the cumulative deviation characterization value is positively correlated with the duration and severity of each valid deviation event. The trend analysis module is used to record the cumulative deviation characterization values ​​corresponding to each deviation direction in multiple consecutive preset time windows, forming a deviation trend data sequence corresponding to each deviation direction. When the number of data points in any deviation trend data sequence reaches a first preset number, trend analysis is performed on the most recent first preset number of data points in the deviation trend data sequence to obtain the change trend characteristic value of the deviation trend data sequence. The early warning module is used to identify the cause of deviation based on vehicle status data, and to determine the degree of deviation deterioration based on the characteristic values ​​of the changing trends corresponding to each deflection direction. It outputs early warning information that includes the cause of deviation, the degree of deviation deterioration, and the determination result of the deviation direction.

Images