Vehicle driving risk identification method and device integrated with vehicle posture monitoring

Abstract

The present invention discloses a vehicle driving risk identification method and device integrated with vehicle posture monitoring, relating to the field of data processing technology. The method comprises: deploying a sensor network for vehicle posture monitoring to collect real-time posture information of a target vehicle; constructing an intuitive assessment channel and a regression analysis channel to fuse and generate a vehicle dynamic assessment model; inputting the real-time posture information into the assessment model to obtain a dynamic assessment value; defining neighborhood control limits to extract a preceding dynamic sequence of the assessment value; and combining the real-time dynamic assessment value with the preceding dynamic assessment sequence to identify vehicle driving risks. This method achieves the technical effect of expanding the recognition model, integrating subjective behavior assessment, and improving risk identification accuracy.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a vehicle driving risk identification method and device integrated with vehicle body posture monitoring. Background Art

[0002] The dynamic behavior of a vehicle during driving has a significant impact on driving safety. Especially under extreme driving conditions, changes in vehicle posture can often be a precursor to an accident. Vehicle posture monitoring technology relies on sensor data from the vehicle, including acceleration, gyroscopes, speed, and steering angles, to assess posture changes and determine driving risks. Existing technologies often analyze data in isolation at a specific moment or over a period of time. This leads to technical issues such as a single risk identification model, a lack of subjective behavior assessment, and limited risk identification accuracy. Summary of the Invention

[0003] The present invention provides a vehicle driving risk identification method and device that integrates vehicle body posture monitoring to solve the technical problems in the existing technology of a single risk identification mode, lack of subjective behavior assessment, and impact on risk identification accuracy, and achieves the technical effect of expanding the identification mode, integrating subjective behavior assessment, and thus improving risk identification accuracy.

[0004] In a first aspect, the present invention provides a vehicle driving risk identification method integrating vehicle body posture monitoring, wherein the method comprises:

[0005] Identify the target vehicle and deploy a sensor network for vehicle posture monitoring.

[0006] The sensor network is activated to perform real-time body posture monitoring of the target vehicle, and real-time posture information of the target vehicle is calculated based on the monitoring results.

[0007] An intuitive evaluation channel based on physical property relationship and an indirect evaluation channel based on regression analysis are constructed respectively, and the intuitive evaluation channel and the indirect evaluation channel are integrated to obtain a vehicle dynamic evaluation model.

[0008] The real-time posture information is used as input information, and the vehicle dynamics evaluation model is activated to perform vehicle dynamics evaluation, and a real-time dynamics evaluation value at the current moment is obtained.

[0009] A neighborhood control limit at the current moment is defined, and a preceding dynamic evaluation sequence of the real-time dynamic evaluation value is extracted based on the neighborhood control limit.

[0010] The real-time dynamic evaluation value is combined with the previous dynamic evaluation sequence to identify the driving risk of the target vehicle.

[0011] In a second aspect, the present invention further provides a vehicle driving risk identification device integrated with vehicle body posture monitoring, wherein the device comprises:

[0012] A sensor network deployment module is used to determine the target vehicle and deploy a sensor network for vehicle body posture monitoring.

[0013] A real-time posture monitoring module is used to activate the sensor network to perform real-time body posture monitoring of the target vehicle and calculate the real-time posture information of the target vehicle based on the monitoring results.

[0014] A dynamic evaluation construction module is used to respectively construct an intuitive evaluation channel based on physical property relationships and an indirect evaluation channel based on regression analysis, and to fuse the intuitive evaluation channel and the indirect evaluation channel to obtain a vehicle dynamic evaluation model.

[0015] A dynamic evaluation module is used to use the real-time posture information as input information, activate the vehicle dynamic evaluation model to perform vehicle dynamic evaluation, and obtain the real-time dynamic evaluation value at the current moment.

[0016] An associated calling module is used to define a neighborhood control limit at a current moment and extract a preceding dynamic evaluation sequence of the real-time dynamic evaluation value according to the neighborhood control limit.

[0017] A risk identification module is used to identify the driving risk of the target vehicle by combining the real-time dynamic evaluation value with the previous dynamic evaluation sequence.

[0018] The present invention discloses a vehicle driving risk identification method and device integrated with vehicle body posture monitoring, comprising: determining a target vehicle and deploying a sensor network for vehicle body posture monitoring; activating the sensor network, monitoring the target vehicle's body posture in real time, and calculating the vehicle's real-time posture information based on the monitoring results; constructing an intuitive evaluation channel based on physical property relationships and an indirect evaluation channel based on regression analysis, and fusing the two to obtain a vehicle dynamic evaluation model; using real-time posture information as input, activating the vehicle dynamic evaluation model for dynamic evaluation to obtain a real-time dynamic evaluation value at the current moment; defining a neighborhood control limit at the current moment, and extracting a preceding dynamic evaluation sequence of the real-time dynamic evaluation value based on the control limit; combining the real-time dynamic evaluation value with the preceding dynamic evaluation sequence to identify the target vehicle's driving risk. The vehicle driving risk identification method and device integrated with vehicle body posture monitoring disclosed by the present invention solves the technical problems of a single risk identification mode, a lack of subjective behavior evaluation, and an impact on risk identification accuracy, and achieves the technical effect of expanding the identification mode, fusing subjective behavior evaluation, and thereby improving risk identification accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1Schematic diagram of the process of the vehicle driving risk identification method integrating vehicle body posture monitoring according to the present invention;

[0020] Figure 2 This is a structural diagram of a vehicle driving risk identification device integrated with vehicle body posture monitoring according to the present invention.

[0021] Explanation of the accompanying drawings: sensor network deployment module 11, real-time posture monitoring module 12, dynamic assessment construction module 13, dynamic assessment module 14, association calling module 15, risk identification module 16. DETAILED DESCRIPTION

[0022] The technical solutions provided in the embodiments of the present invention are designed to address the technical issues of the prior art, such as a single risk identification model, a lack of subjective behavior assessment, and poor risk identification accuracy. The overall approach adopted is as follows:

[0023] First, a target vehicle is determined and a sensor network for vehicle body posture monitoring is deployed; then, the sensor network is activated to perform real-time vehicle body posture monitoring of the target vehicle, and the real-time posture information of the target vehicle is calculated based on the monitoring results; then, an intuitive evaluation channel based on physical property relationship and an indirect evaluation channel based on regression analysis are respectively constructed, and the intuitive evaluation channel and the indirect evaluation channel are integrated to obtain a vehicle dynamic evaluation model; then, with the real-time posture information as input information, the vehicle dynamic evaluation model is activated to perform vehicle dynamic evaluation and obtain the real-time dynamic evaluation value at the current moment; then, the neighborhood control limit at the current moment is defined, and the preceding dynamic evaluation sequence of the real-time dynamic evaluation value is extracted based on the neighborhood control limit; finally, the real-time dynamic evaluation value and the preceding dynamic evaluation sequence are combined to perform driving risk identification of the target vehicle.

[0024] The above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods of the specification to better understand the above technical solution. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments of the present invention. It should be understood that the present invention is not limited to the example embodiments used only to explain the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. In addition, it should be noted that, for the convenience of description, only the parts related to the present invention, rather than all, are shown in the drawings.

[0025] Example 1

[0026] Figure 1 The flowchart of the vehicle driving risk identification method integrated with vehicle posture monitoring according to the present invention is as follows:

[0027] Identify the target vehicle and deploy a sensor network for vehicle posture monitoring.

[0028] Specifically, first, the sensor network is adaptively deployed according to the determined target vehicle to ensure that the various monitoring data required for driving risk identification can be obtained efficiently and accurately.

[0029] In some embodiments, determining a target vehicle and deploying a sensor network for vehicle posture monitoring includes:

[0030] Based on the monitoring requirements of the target scenario, a sensor requirement form is determined; the target vehicle is interacted with to obtain the current sensor set, including sensor ID, sensor category, sensor parameters and sensor quantity; the sensor requirement form is compared with the current sensor set, and the difference items between the sensor requirement form and the current sensor set are extracted and output as a sensor deployment task form; according to the sensor deployment task form, the sensor network is deployed in combination with the current sensor set of the target vehicle.

[0031] Specifically, first, according to the actual application scenarios of vehicle posture monitoring, such as driving environment (urban roads, highways, bad weather, etc.) and vehicle characteristics (model, load, etc.), the corresponding monitoring needs are determined, and then the required monitoring indicators (such as vehicle body tilt angle, acceleration, vibration, etc.) are clarified; then, the target vehicle is interacted with and the vehicle's existing sensor data is obtained through the vehicle's on-board system, OBD interface or vehicle electronic control unit (ECU), that is, the current sensor set

[0032] Specifically, the current sensor set includes the sensor ID (a unique identifier for each sensor, used for calibration and management), sensor category, sensor parameters (such as sensitivity, measurement range, accuracy, etc.), and sensor quantity collected from the target vehicle.

[0033] Specifically, the monitoring requirement form and the existing sensor set are compared to analyze whether the existing sensors meet the requirements of vehicle body posture monitoring, and to obtain difference information (i.e., multiple difference items), and then a sensor deployment task form is obtained. The sensor deployment task form is used to define the sensors that need to be added or replaced; exemplarily, the difference items include sensor quantity difference (insufficient number of sensors), sensor category difference (missing sensor type), accuracy and parameter difference (existing sensors whose accuracy does not meet the monitoring requirements).

[0034] Optionally, the sensor deployment task form includes the type and quantity of new sensors, the deployment location of the new sensors, the deployment priority, etc., thereby providing clear guidance for the configuration of the sensor network.

[0035] Specifically, the actual sensor deployment task is performed according to the task form. First, based on the requirements of the task form, suitable sensors are selected, such as accelerometers, gyroscopes, GPS, magnetometers, etc., and then the selected sensors are installed at the designated locations of the vehicle to ensure that the sensors can accurately monitor the vehicle body posture and do not interfere with the normal operation of the vehicle; then, the newly added sensors are connected to the vehicle's on-board system to ensure that the data can be transmitted and processed in real time through the on-board bus (such as CAN bus, Ethernet, etc.) to form a sensor network.

[0036] The above steps, by comparing the task form with the difference items, can automatically identify the deficiencies of vehicle sensors and make timely supplements to ensure that monitoring needs can be met, thereby helping to obtain vehicle body dynamic information efficiently and accurately.

[0037] In some implementations, comparing the sensing requirement form with the current sensor set, extracting the differences between the sensing requirement form and the current sensor set, and outputting the differences as a sensor deployment task form includes:

[0038] Taking the sensor category as an index, extract the corresponding category sensor status information from the current sensor set; parse the category sensor status information, obtain category sensor parameters and category sensor quantity, and determine whether the category sensor parameters and category sensor quantity meet the sensor requirement form; if the category sensor parameters and category sensor quantity meet the corresponding form items in the sensor requirement form, remove the corresponding form items; if the category sensor parameters and category sensor quantity do not meet the corresponding form items in the sensor requirement form, retain the unsatisfied difference information; output the adjusted sensor requirement form as the sensor deployment task form.

[0039] Specifically, first, according to the sensor category, the multiple sensors and related information in the current sensor set are divided into multiple categories of sensor current information. The multiple categories of sensor current information include the ID, parameters (such as sensitivity, range, etc.) and quantity (the actual number of sensors of this category on the vehicle) of each category of sensors; then, the parameter value of each sensor is extracted from the category current sensor information, such as the sensitivity of the accelerometer, the accuracy of the gyroscope, etc., and the number of sensors of each category is counted and compared with the requirements in the sensor requirement form to compare whether the number of existing sensors meets the required number of the category in the sensor requirement form and whether the parameters of the existing sensors meet the technical requirements (such as accuracy, range, etc.) in the sensor requirement form.

[0040] Specifically, if the parameters and quantity of existing sensors meet the requirements in the sensor requirement form, the corresponding items of this category will be removed from the sensor requirement form. If the parameters and / or quantity of existing sensors do not meet the requirements, the difference items of this category will be retained, and all unsatisfied sensor categories, parameters or quantities will be retained as difference information, indicating the type, parameters and quantity of sensors that need to be added or replaced to point out the missing sensors or parameter problems; the adjusted sensor requirement form will be output as the sensor deployment task form.

[0041] In some implementations, the sensor network includes at least an acceleration sensor and a gyroscope, which are used to obtain acceleration information, azimuth information, and pitch angle information of the target vehicle.

[0042] Specifically, the sensing network consists of at least an accelerometer and a gyroscope. The accelerometer is used to detect the acceleration of the vehicle, including the acceleration changes of the vehicle in various directions (longitudinal, lateral, and vertical directions), reflecting the dynamic response of the vehicle in different axes (such as forward acceleration, lateral acceleration, etc.); the gyroscope is used to measure the angular velocity of the vehicle, and then infer the posture changes of the vehicle, that is, the pitch, roll, yaw and other angle changes. Through the joint use of the accelerometer and the gyroscope, all-round posture monitoring of the target vehicle can be achieved.

[0043] Specifically, the vehicle's azimuth angle information is used to reflect the vehicle's rotation state on the horizontal plane; the vehicle's pitch angle, that is, the angle at which the vehicle tilts forward and backward, can be used to reflect the vehicle's posture changes when on a slope or braking. Combining the azimuth angle and the pitch angle can accurately describe the vehicle's posture changes.

[0044] The sensor network is activated to perform real-time body posture monitoring of the target vehicle, and real-time posture information of the target vehicle is calculated based on the monitoring results.

[0045] Specifically, the accelerometer and gyroscope in the sensor network are enabled to collect the vehicle's acceleration data and angular velocity data in real time, thereby providing the vehicle's original dynamic response data in all directions; then, the angular velocity data provided by the gyroscope is combined with time integration to calculate the vehicle's azimuth angle, describing the vehicle's rotational changes in the horizontal direction; the angular velocity data provided by the gyroscope is combined with the vertical acceleration information of the accelerometer to calculate the vehicle's pitch angle (the degree of front and rear tilt); by processing the lateral acceleration and lateral angular velocity data, the vehicle's roll angle is calculated to reflect the vehicle's state in terms of roll, and the output is the real-time posture information of the target vehicle.

[0046] In other words, the real-time attitude information of the target vehicle includes pitch angle (forward and backward tilt), azimuth angle (steering angle) and roll angle (side tilt).

[0047] By activating the accelerometer and gyroscope, combined with real-time data acquisition and fusion algorithms, the real-time posture information of the target vehicle can be accurately calculated, helping to monitor the dynamic behavior of the vehicle in real time and providing an accurate analysis data basis for subsequent analysis and judgment of the vehicle.

[0048] An intuitive evaluation channel based on physical property relationship and an indirect evaluation channel based on regression analysis are constructed respectively, and the intuitive evaluation channel and the indirect evaluation channel are integrated to obtain a vehicle dynamic evaluation model.

[0049] Specifically, the intuitive evaluation channel is used to evaluate based on the vehicle's physical properties and motion characteristics, and has strong physical intuitiveness. In other words, the intuitive evaluation channel relies on known vehicle dynamics models and physical laws, and conducts a preliminary evaluation of the vehicle's dynamic performance based on real-time input data, outputting a series of intuitive evaluation results, such as acceleration smoothness, stability, braking ability, etc.

[0050] Specifically, the indirect assessment channel is a regression model under a data-driven approach. By analyzing historical data, a mapping relationship is established between the subjective driving style of the target vehicle operator and the vehicle dynamics, thereby accurately predicting the corresponding risks of the vehicle under different driving styles.

[0051] By combining the intuitive evaluation channel based on physical property relationships with the indirect evaluation channel based on regression analysis, the dynamic performance of the vehicle can be evaluated more accurately and comprehensively. The integrated vehicle dynamic evaluation model can not only perform fast calculations based on physical laws, but also provide more precise dynamic behavior predictions through data-driven regression analysis.

[0052] In some embodiments, an intuitive evaluation channel based on physical property relationships and an indirect evaluation channel based on regression analysis are constructed separately, and the intuitive evaluation channel and the indirect evaluation channel are integrated to obtain a vehicle dynamic evaluation model, including:

[0053] The vehicle intrinsic parameters of the target vehicle are interactively acquired and output as a physical property parameter set; in combination with a predefined intuitive dynamic indicator set, a corresponding physical property relationship mathematical model is called and initialized based on the physical property parameter set, and relative margin operators corresponding to multiple intuitive dynamic indicators are constructed based on the initialization result; and multiple relative margin operators are weightedly fused to obtain the intuitive evaluation channel.

[0054] Specifically, first, the physical characteristics of the target vehicle are obtained, such as vehicle geometric parameters, mass distribution, center of gravity position, suspension system stiffness and damping coefficient, tire size and parameters and other vehicle inherent parameters, and a data set containing all the basic physical and geometric characteristics of the vehicle is formed, which is called the physical property parameter set.

[0055] Specifically, based on the inherent physical parameters and geometric characteristics of the vehicle, a predefined mathematical model of physical property relationships is used to describe and calculate multiple intuitive dynamic indicators in the intuitive dynamic indicator set. Among them, the intuitive dynamic indicators are key dynamic indicators that can directly reflect the driving stability, controllability and safety of the vehicle, and exemplary include steering response time, vehicle stability, acceleration performance, braking performance, etc.

[0056] Specifically, based on the output results of the physical property relationship model, a "margin" evaluation is performed on each dynamic indicator, where the margin refers to the distance between the vehicle's dynamic performance and the safety boundary under specific driving behavior. Specifically, for each predefined dynamic indicator (such as steering response, acceleration performance, braking performance, etc.), the relative margin of each dynamic indicator is calculated through the output of the physical property relationship model. Exemplary examples include: steering response margin, which reflects whether the vehicle can respond quickly and stably under steering operations. Too small a margin may lead to steering out of control; acceleration margin, which evaluates the response and stability of the vehicle during acceleration. Too small a margin may lead to unstable acceleration or loss of control; braking margin, which reflects whether the vehicle can brake stably during braking. Too small a margin may lead to brake failure or loss of control.

[0057] Optionally, the relative margin may be a difference between a maximum stable performance value and a current performance value, or a difference between a performance threshold value and a current performance value.

[0058] Furthermore, by comprehensively considering the importance of various indicators to the vehicle dynamic evaluation, a weighted fusion relative margin operator is performed. For example, the acceleration margin may be more important at high speeds, so it can be given a higher weight, while the braking margin may be more critical in a mountainous environment, so it can be given a higher weight. Through weighted fusion, a comprehensive evaluation mathematical model of vehicle dynamic performance can be obtained as an intuitive evaluation channel, which can intuitively reflect the dynamic performance, handling and safety of the target vehicle under different driving conditions, and provide a basis for subsequent decision-making.

[0059] In some implementations, constructing an intuitive evaluation channel based on physical property relationships and an indirect evaluation channel based on regression analysis, respectively, and fusing the intuitive evaluation channel and the indirect evaluation channel to obtain a vehicle dynamic evaluation model further includes:

[0060] Based on the vehicle category information of the target vehicle, corresponding historical driving data is collected, the historical driving data is marked as unstable, and historical posture information is extracted; through regression analysis, the historical posture information is combined with the instability mark to establish a probability mapping relationship between posture information and dynamic instability, and the relationship is stored as the indirect evaluation channel; the intuitive evaluation channel and the indirect evaluation channel are integrated to obtain the vehicle dynamic evaluation model, wherein the input of the indirect evaluation channel is the adjustment coefficient of the output of the intuitive evaluation channel.

[0061] Specifically, the category information of the target vehicle is obtained. Vehicles of different categories have significant differences in driving characteristics and dynamic behaviors, and these differences need to be considered separately in the evaluation model. Then, based on the target vehicle category, historical data related to vehicle driving is collected, including historical vehicle status (i.e., historical posture information), driving risk conditions (whether it is unstable), driver input, etc.

[0062] Specifically, historical driving data is analyzed to identify unstable events such as skidding / slipping, rollover, and control failure, and the historical driving data at the time of these unstable events is marked. At the same time, relevant posture information is extracted from the historical driving data.

[0063] Specifically, a mapping relationship between historical posture information and dynamic instability is established through regression analysis methods such as linear regression, polynomial regression, and support vector regression. The extracted historical posture information and the corresponding driver input are used as input variables for the regression analysis, and the instability mark is used as the output variable. The output variable is a combination of a binary variable (instability / no instability) and a categorical variable (such as mild instability, moderate instability, severe instability, etc.).

[0064] By combining the driver's driving style with the vehicle's dynamic behavior regression analysis, we can obtain the probabilistic relationship between historical posture information and dynamic instability events, and evaluate the risk coefficient faced by the vehicle during real-time driving. In other words, we predict the probability of vehicle instability under a specific posture. For example, for the same set of acceleration / angle data, if the driver of the target vehicle is an aggressive driver, the corresponding risk coefficient of the vehicle dynamics is actually higher, which means that it is more likely to generate risks.

[0065] The real-time posture information is used as input information, and the vehicle dynamics evaluation model is activated to perform vehicle dynamics evaluation, and a real-time dynamics evaluation value at the current moment is obtained.

[0066] Specifically, the real-time posture information is input into the vehicle dynamic evaluation model and passed to the intuitive evaluation channel and the indirect evaluation channel for parallel processing respectively to obtain the intuitive evaluation results and the indirect evaluation results. Then, the outputs from the two evaluation channels are weightedly fused to calculate the final dynamic evaluation value.

[0067] A neighborhood control limit at the current moment is defined, and a preceding dynamic evaluation sequence of the real-time dynamic evaluation value is extracted based on the neighborhood control limit.

[0068] Specifically, the neighborhood control limit refers to the traceable range of the vehicle dynamic evaluation value at the current moment, that is, the length of the time window for extracting historical dynamic evaluation values; after determining the neighborhood control limit at the current moment, the previous dynamic evaluation sequence is extracted from the historical data according to the neighborhood control limit, where the previous dynamic evaluation sequence refers to several historical dynamic evaluation values ​​obtained according to a certain time window before the current moment, which are used to analyze the changing trend of the current dynamic evaluation value.

[0069] Optionally, by analyzing the previous dynamic evaluation sequence, the changing trend of vehicle stability and risk can be identified. For example, if the previous sequence and the current evaluation value show a clear trend of gradual instability, the target vehicle is more likely to be at risk.

[0070] The real-time dynamic evaluation value is combined with the previous dynamic evaluation sequence to identify the driving risk of the target vehicle.

[0071] In some embodiments, the real-time dynamic evaluation value and the previous dynamic evaluation sequence are combined to identify the driving risk of the target vehicle, including:

[0072] Based on the real-time dynamic evaluation value and the previous dynamic evaluation sequence, the dynamic trend evaluation value of the target vehicle is calculated; a risk discrimination threshold is defined, wherein the risk discrimination threshold includes a multi-level dynamic threshold and a multi-level dynamic trend threshold corresponding to multiple risk levels; if the real-time dynamic evaluation value is greater than any one of the multi-level dynamic thresholds and / or the dynamic trend evaluation value is greater than any one of the multi-level dynamic trend thresholds, it is considered that the target vehicle has a corresponding level of driving risk, and a driving risk identification result including a risk level mark is output; if the real-time dynamic evaluation value and the dynamic trend evaluation value are both less than or equal to the multi-level dynamic threshold and the multi-level dynamic trend threshold, it is considered that the target vehicle has no driving risk.

[0073] Specifically, first, by fitting the changing trend of the preceding sequence, the slope is calculated as the dynamic trend evaluation value of the target vehicle.

[0074] Specifically, the multi-level dynamic threshold is the threshold of the real-time dynamic evaluation value defined for different risk levels, and the multi-level dynamic trend threshold is the threshold of the dynamic trend evaluation value defined for different risk levels. The driving risk of the target vehicle can be judged through the numerical relationship between the real-time dynamic evaluation value and the dynamic trend evaluation value and the risk judgment threshold.

[0075] Specifically, if the real-time dynamic evaluation value exceeds any multi-level dynamic threshold, and / or the dynamic trend evaluation value exceeds any multi-level dynamic trend threshold, the target vehicle is considered to have a driving risk. For example, if the real-time dynamic evaluation value exceeds the first-level dynamic threshold but does not exceed the second-level dynamic threshold, it is judged as low risk; if the dynamic trend evaluation value exceeds the second-level dynamic trend threshold, it is judged as medium risk.

[0076] Specifically, if the real-time dynamic evaluation value is less than or equal to all multi-level dynamic thresholds, and the dynamic trend evaluation value is less than or equal to all multi-level dynamic trend thresholds, it is considered that the target vehicle does not have a driving risk.

[0077] Furthermore, according to the risk discrimination logic, the driving risk identification result of the target vehicle is output, including: a risk level mark used to identify the driving risk level of the target vehicle (such as low risk, medium risk, high risk); a real-time dynamic evaluation value and a dynamic trend evaluation value.

[0078] The above method steps dynamically calculate the trend evaluation value of the target vehicle by combining the real-time dynamic evaluation value with the previous dynamic evaluation sequence, and perform risk judgment based on multi-level dynamic thresholds and multi-level dynamic trend thresholds, so as to identify the driving risk of the target vehicle in real time and accurately.

[0079] In some embodiments, the real-time dynamic evaluation value is combined with the previous dynamic evaluation sequence to identify the driving risk of the target vehicle. Thereafter, the method further includes:

[0080] Based on the driving risk identification results, the risk plan library is accessed to call the corresponding risk reminder plan, and multi-dimensional risk reminders are performed according to the risk reminder plan, wherein the multi-dimensional risk reminders include at least audio warnings, visual warnings, and vibration warnings.

[0081] Specifically, after completing the driving risk identification of the target vehicle, the pre-defined risk plan library is accessed according to the identified driving risk level (low, medium, high, etc.) and related risk type, and the matching risk reminder plan is called. The risk plan library stores corresponding processing solutions (i.e., risk reminder plans) for different risk levels and risk types, involving specific reminder methods and multi-dimensional reminder content, such as audio, vision, vibration, etc.

[0082] Furthermore, multi-dimensional risk reminders are given to the driver or passengers in the car based on the risk reminder plan called. The multi-dimensional risk reminders include at least the following three methods: audio warnings, that is, playing voice prompts or warning sounds through speakers; visual warnings, including displaying warning information on the vehicle's display screen, head-up display (HUD) or instrument panel; vibration warnings, that is, providing tactile feedback through the steering wheel, seat or other vibration devices.

[0083] In some embodiments, the real-time dynamic evaluation value is combined with the previous dynamic evaluation sequence to identify the driving risk of the target vehicle. Thereafter, the method further includes:

[0084] Configure active intervention constraints, where the active intervention constraints include risk severity constraints and intervention delay constraints; after performing multi-dimensional risk reminders, continuously monitor the posture information of the target vehicle and update the driving risk identification results accordingly; if the updated driving risk identification results meet the risk severity constraints and the duration of the multi-dimensional risk reminders meets the intervention delay constraints, activate the active safety configuration of the target vehicle and perform active intervention.

[0085] Specifically, after completing the driving risk identification and multi-dimensional risk reminders of the target vehicle, an active intervention constraint mechanism is further introduced to trigger active intervention when the active intervention constraints are met.

[0086] Specifically, active intervention constraints are used to define the circumstances under which the vehicle's active safety configuration is triggered to ensure the accuracy and rationality of the intervention. The constraints include risk severity constraints for the risk levels in the driving risk identification results and intervention delay constraints that define the duration requirements for multi-dimensional risk reminders. This ensures that active intervention is only allowed to be activated when the updated risk identification results meet the preset risk severity and the driver has been fully reminded before active intervention.

[0087] Specifically, after the multi-dimensional risk reminder, the posture information of the target vehicle is continuously monitored, and the driving risk identification results are dynamically updated. If the updated driving risk identification results meet or exceed the risk severity constraint standard, and the duration of the multi-dimensional risk reminder reaches or exceeds the time threshold of the intervention delay constraint, if the driver has been reminded for more than 3 seconds but has not taken effective measures, the active safety configuration of the target vehicle will be activated.

[0088] Optional active safety configuration of the target vehicle includes emergency braking function, lane keeping function, power output limitation, traction control, etc.

[0089] Through the above steps, it is possible to ensure that the active safety configuration or system of the target vehicle automatically intervenes when necessary on the basis of fully reminding the driver, thereby ensuring driving safety to the greatest extent.

[0090] In summary, the vehicle driving risk identification method integrated with vehicle body posture monitoring provided by the present invention has the following technical effects:

[0091] The method identifies the target vehicle and deploys a sensor network for body posture monitoring. The sensor network is activated to monitor the target vehicle's body posture in real time, and the vehicle's real-time posture information is calculated based on the monitoring results. An intuitive evaluation channel based on physical property relationships and an indirect evaluation channel based on regression analysis are constructed and integrated to form a vehicle dynamic evaluation model. The vehicle dynamic evaluation model is activated using real-time posture information as input to perform dynamic evaluation and obtain the current real-time dynamic evaluation value. Neighborhood control limits are defined for the current moment, and based on these control limits, a preceding dynamic evaluation sequence of the real-time dynamic evaluation value is extracted. The real-time dynamic evaluation value and the preceding dynamic evaluation sequence are combined to identify the target vehicle's driving risk. This achieves the technical effect of expanding the recognition mode, integrating subjective behavior evaluation, and thus improving the accuracy of risk identification.

[0092] Example 2

[0093] Figure 2 This is a schematic diagram of the structure of the vehicle driving risk identification device integrated with vehicle body posture monitoring of the present invention. For example, Figure 1 The flow chart of the vehicle driving risk identification method integrating vehicle body posture monitoring in the present invention can be shown as follows: Figure 2 The structure shown is implemented.

[0094] Based on the same concept as the vehicle driving risk identification method integrating vehicle body posture monitoring in the above embodiment, the present invention also provides a vehicle driving risk identification device integrating vehicle body posture monitoring, comprising:

[0095] The sensor network deployment module 11 is used to determine the target vehicle and deploy a sensor network for vehicle body posture monitoring.

[0096] The real-time posture monitoring module 12 is used to activate the sensor network to perform real-time body posture monitoring of the target vehicle and calculate the real-time posture information of the target vehicle based on the monitoring results.

[0097] The dynamic evaluation construction module 13 is used to respectively construct an intuitive evaluation channel based on physical property relationships and an indirect evaluation channel based on regression analysis, and to fuse the intuitive evaluation channel and the indirect evaluation channel to obtain a vehicle dynamic evaluation model.

[0098] The dynamic evaluation module 14 is configured to use the real-time posture information as input information, activate the vehicle dynamic evaluation model to perform vehicle dynamic evaluation, and obtain a real-time dynamic evaluation value at the current moment.

[0099] The associated calling module 15 is used to define the neighborhood control limit at the current moment and extract the preceding dynamic evaluation sequence of the real-time dynamic evaluation value according to the neighborhood control limit.

[0100] The risk identification module 16 is configured to identify the driving risk of the target vehicle by combining the real-time dynamic evaluation value with the previous dynamic evaluation sequence.

[0101] The execution steps of the sensor network deployment module 11 include:

[0102] Determine the sensing requirements form based on the monitoring requirements of the target scenario.

[0103] Interact with the target vehicle to obtain the current sensor set, including sensor ID, sensor category, sensor parameters and sensor quantity.

[0104] The sensing requirement form is compared with the current sensor set, and the difference items between the sensing requirement form and the current sensor set are extracted and output as a sensor deployment task form.

[0105] The sensor network is deployed according to the sensor deployment task form and in combination with the current sensor set of the target vehicle.

[0106] In some implementations, the steps executed by the sensor network deployment module 11 further include:

[0107] Using the sensor category as an index, the corresponding category sensor status information is extracted from the status sensor set.

[0108] The status information of the category sensors is parsed to obtain category sensor parameters and the number of category sensors, and it is determined whether the category sensor parameters and the number of category sensors meet the sensing requirement form.

[0109] If the category sensor parameters and the category sensor quantity meet the corresponding form items in the sensing requirement form, the corresponding form items are removed; if the category sensor parameters and the category sensor quantity do not meet the corresponding form items in the sensing requirement form, the unsatisfied difference information is retained.

[0110] The adjusted sensing requirement form is output as the sensor placement task form.

[0111] Furthermore, the sensor network includes at least an acceleration sensor and a gyroscope, which are used to obtain acceleration information, azimuth angle information and pitch angle information of the target vehicle.

[0112] In some embodiments, the execution steps of the dynamic evaluation construction module 13 include:

[0113] The intrinsic parameters of the target vehicle are interactively acquired and output as a set of physical property parameters.

[0114] Combined with a predefined intuitive dynamic indicator set, the corresponding physical property relationship mathematical model is called and initialized based on the physical property parameter set. According to the initialization result, relative margin operators corresponding to multiple intuitive dynamic indicators are constructed.

[0115] The plurality of relative margin operators are weightedly fused to obtain the intuitive evaluation channel.

[0116] In some embodiments, the execution steps of the dynamic evaluation construction module 13 further include:

[0117] Based on the vehicle category information of the target vehicle, corresponding historical driving data is collected, the historical driving data is marked as unstable, and historical posture information is extracted.

[0118] By combining the historical posture information and the instability mark through regression analysis, a probability mapping relationship between posture information and dynamic instability is established and stored as the indirect evaluation channel.

[0119] The intuitive evaluation channel and the indirect evaluation channel are integrated to obtain the vehicle dynamic evaluation model, wherein the input of the indirect evaluation channel is the adjustment coefficient of the output of the intuitive evaluation channel.

[0120] In some embodiments, the steps performed by the risk identification module 16 include:

[0121] A dynamic trend evaluation value of the target vehicle is calculated based on the real-time dynamic evaluation value and the previous dynamic evaluation sequence.

[0122] A risk discrimination threshold is defined, wherein the risk discrimination threshold includes a multi-level dynamic threshold and a multi-level dynamic trend threshold corresponding to multiple risk levels.

[0123] If the real-time dynamic evaluation value is greater than any one of the multi-level dynamic thresholds and / or the dynamic trend evaluation value is greater than any one of the multi-level dynamic trend thresholds, it is considered that the target vehicle has a corresponding level of driving risk, and a driving risk identification result including a risk level mark is output.

[0124] If the real-time dynamic evaluation value and the dynamic trend evaluation value are both less than or equal to the multi-level dynamic threshold and the multi-level dynamic trend threshold, it is considered that the target vehicle does not have a driving risk.

[0125] Furthermore, the device also includes a risk reminder module, which is used to:

[0126] Based on the driving risk identification results, the risk plan library is accessed to call the corresponding risk reminder plan, and multi-dimensional risk reminders are performed according to the risk reminder plan, wherein the multi-dimensional risk reminders include at least audio warnings, visual warnings, and vibration warnings.

[0127] Furthermore, the device further includes an active intervention module, which is used to:

[0128] Active intervention constraints are configured, wherein the active intervention constraints include risk severity constraints and intervention delay constraints.

[0129] After the multi-dimensional risk reminder is performed, the posture information of the target vehicle is continuously monitored, and the driving risk identification result is updated accordingly.

[0130] If the updated driving risk identification result meets the risk severity constraint and the duration of the multi-dimensional risk reminder meets the intervention delay constraint, the active safety configuration of the target vehicle is activated and active intervention is performed.

[0131] It should be understood that the embodiments mentioned in this specification focus on their differences from other embodiments. The specific embodiments in the aforementioned embodiment one are also applicable to the vehicle driving risk identification device that integrates vehicle body posture monitoring as described in embodiment two. For the sake of brevity of the specification, no further elaboration will be given here.

[0132] It should be understood that the embodiments disclosed in the present invention and the above description can enable those skilled in the art to use the present invention to implement the present invention. At the same time, the present invention is not limited to the embodiments mentioned above. It should be understood that those skilled in the art can still modify the technical solutions described in the above embodiments or replace some of the technical features therein with equivalents; and such modifications or replacements do not deviate from the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention and are all included in the scope of protection of the present invention.

Claims

1. A vehicle driving risk identification method integrating vehicle posture monitoring is characterized by: The method comprises: Identify the target vehicle and deploy a sensor network for vehicle posture monitoring; activating the sensor network to perform real-time body posture monitoring of the target vehicle, and calculating real-time posture information of the target vehicle based on the monitoring results; Separately constructing an intuitive evaluation channel based on physical property relationships and an indirect evaluation channel based on regression analysis, and fusing the intuitive evaluation channel and the indirect evaluation channel to obtain a vehicle dynamic evaluation model; Using the real-time posture information as input information, activating the vehicle dynamics assessment model to perform vehicle dynamics assessment, and obtaining a real-time dynamics assessment value at the current moment; Defining a neighborhood control limit at the current moment, and extracting a preceding dynamic evaluation sequence of the real-time dynamic evaluation value according to the neighborhood control limit; The real-time dynamic evaluation value is combined with the previous dynamic evaluation sequence to identify the driving risk of the target vehicle.

2. The vehicle driving risk identification method integrating vehicle body posture monitoring according to claim 1 is characterized in that: Identify the target vehicle and deploy a sensor network for vehicle posture monitoring, including: Determine the sensing requirements form based on the monitoring requirements of the target scenario; Interact with the target vehicle to obtain the current sensor set, including sensor ID, sensor category, sensor parameters and sensor quantity; Comparing the sensor requirement form with the current sensor set, extracting the difference items between the sensor requirement form and the current sensor set, and outputting them as a sensor deployment task form; The sensor network is deployed according to the sensor deployment task form and in combination with the current sensor set of the target vehicle.

3. The vehicle driving risk identification method integrating vehicle body posture monitoring according to claim 2 is characterized in that: Comparing the sensor requirement form with the current sensor set, extracting the difference items between the sensor requirement form and the current sensor set, and outputting them as a sensor deployment task form, including: Taking the sensor category as an index, extracting the corresponding category sensor status information from the current sensor set; parsing the status information of the category sensors, obtaining category sensor parameters and the number of category sensors, and determining whether the category sensor parameters and the number of category sensors meet the sensing requirement form; If the category sensor parameters and the category sensor quantity satisfy the corresponding form items in the sensing requirement form, the corresponding form items are removed; if the category sensor parameters and the category sensor quantity do not satisfy the corresponding form items in the sensing requirement form, the unsatisfied difference information is retained; The adjusted sensing requirement form is output as the sensor placement task form.

4. The vehicle driving risk identification method integrating vehicle body posture monitoring as claimed in claim 3 is characterized in that: The sensor network includes at least an acceleration sensor and a gyroscope, which are used to obtain acceleration information, azimuth angle information and pitch angle information of the target vehicle.

5. The vehicle driving risk identification method integrating vehicle body posture monitoring as claimed in claim 4 is characterized in that: An intuitive evaluation channel based on physical property relationships and an indirect evaluation channel based on regression analysis are respectively constructed, and the intuitive evaluation channel and the indirect evaluation channel are integrated to obtain a vehicle dynamic evaluation model, including: Interactively obtain the intrinsic parameters of the target vehicle and output them as a set of physical property parameters; In combination with a predefined intuitive dynamic indicator set, a corresponding physical property relationship mathematical model is called and initialized based on the physical property parameter set, and relative margin operators corresponding to multiple intuitive dynamic indicators are constructed according to the initialization results; The plurality of relative margin operators are weightedly fused to obtain the intuitive evaluation channel.

6. The vehicle driving risk identification method integrating vehicle body posture monitoring as claimed in claim 5 is characterized in that: Separately constructing an intuitive evaluation channel based on physical property relationships and an indirect evaluation channel based on regression analysis, and fusing the intuitive evaluation channel and the indirect evaluation channel to obtain a vehicle dynamic evaluation model, further comprising: Based on the vehicle category information of the target vehicle, corresponding historical driving data is collected, the historical driving data is marked as unstable, and historical posture information is extracted; By combining the historical posture information and the instability mark through regression analysis, a probability mapping relationship between posture information and dynamic instability is established and stored as the indirect evaluation channel; The intuitive evaluation channel and the indirect evaluation channel are integrated to obtain the vehicle dynamic evaluation model, wherein the input of the indirect evaluation channel is the adjustment coefficient of the output of the intuitive evaluation channel.

7. The vehicle driving risk identification method integrated with vehicle body posture monitoring according to claim 6 is characterized in that: Combining the real-time dynamic evaluation value with the previous dynamic evaluation sequence to identify the driving risk of the target vehicle includes: Calculating a dynamic trend evaluation value of the target vehicle based on the real-time dynamic evaluation value and the previous dynamic evaluation sequence; Defining risk discrimination thresholds, wherein the risk discrimination thresholds include multi-level dynamic thresholds and multi-level dynamic trend thresholds corresponding to multiple risk levels; If the real-time dynamic assessment value is greater than any one of the multi-level dynamic thresholds and / or the dynamic trend assessment value is greater than any one of the multi-level dynamic trend thresholds, it is considered that the target vehicle has a corresponding level of driving risk, and a driving risk identification result including a risk level mark is output; If the real-time dynamic evaluation value and the dynamic trend evaluation value are both less than or equal to the multi-level dynamic threshold and the multi-level dynamic trend threshold, it is considered that the target vehicle does not have a driving risk.

8. The vehicle driving risk identification method integrating vehicle body posture monitoring according to claim 1 is characterized in that: Combining the real-time dynamic evaluation value with the preceding dynamic evaluation sequence to identify the driving risk of the target vehicle, the method further includes: Based on the driving risk identification results, the risk plan library is accessed to call the corresponding risk reminder plan, and multi-dimensional risk reminders are performed according to the risk reminder plan, wherein the multi-dimensional risk reminders include at least audio warnings, visual warnings, and vibration warnings.

9. The vehicle driving risk identification method integrating vehicle body posture monitoring as claimed in claim 8, characterized in that: Combining the real-time dynamic evaluation value with the preceding dynamic evaluation sequence to identify the driving risk of the target vehicle, the method further includes: Configuring active intervention constraints, wherein the active intervention constraints include risk severity constraints and intervention delay constraints; After the multi-dimensional risk reminder is issued, the posture information of the target vehicle is continuously monitored and the driving risk identification result is updated accordingly; If the updated driving risk identification result meets the risk severity constraint and the duration of the multi-dimensional risk reminder meets the intervention delay constraint, the active safety configuration of the target vehicle is activated and active intervention is performed.

10. A vehicle driving risk identification device integrating vehicle body posture monitoring is characterized in that: The device is used to execute the vehicle driving risk identification method integrated with vehicle body posture monitoring according to any one of claims 1 to 9, and the device includes: A sensor network deployment module, the sensor network deployment module is used to determine the target vehicle and deploy a sensor network for vehicle body posture monitoring; A real-time posture monitoring module, wherein the real-time posture monitoring module is used to activate the sensor network to perform real-time body posture monitoring of the target vehicle and calculate the real-time posture information of the target vehicle based on the monitoring results; A dynamic evaluation construction module, the dynamic evaluation construction module is used to respectively construct an intuitive evaluation channel based on physical property relationships and an indirect evaluation channel based on regression analysis, and to fuse the intuitive evaluation channel and the indirect evaluation channel to obtain a vehicle dynamic evaluation model; A dynamic evaluation module, configured to use the real-time posture information as input information, activate the vehicle dynamic evaluation model to perform vehicle dynamic evaluation, and obtain a real-time dynamic evaluation value at a current moment; An associated calling module, the associated calling module is used to define a neighborhood control limit at a current moment and extract a preceding dynamic evaluation sequence of the real-time dynamic evaluation value according to the neighborhood control limit; A risk identification module is used to identify the driving risk of the target vehicle by combining the real-time dynamic evaluation value with the previous dynamic evaluation sequence.

Images