Heavy truck tractor and trailer collision judgment method, system and equipment and medium

Abstract

The invention relates to the technical field of Internet of Vehicles, and discloses a heavy truck tractor trailer collision judgment method, system and device and a medium. The method comprises the following steps: acquiring driving data of a tractor in the heavy truck tractor trailer; feature parameters of the driving data are extracted, and whether the driving data are abnormal driving data or not is judged according to a preset feature parameter threshold value, so that the collision possibility of the trailer in the heavy truck tractor trailer is preliminarily judged; if the possibility of the collision of the trailer is preliminarily judged to be greater than a preset threshold value, inputting the driving data into a trained collision judgment model, and extracting basic statistical characteristics, change rate characteristics, state combination characteristics, spatial change characteristics, time sequence characteristics and static characteristics of the driving data through the collision judgment model; therefore, whether the trailer collides or not is judged again, the collision judgment result of the trailer is obtained, and the accuracy of collision detection is improved on the basis that the collision detection cost of the trailer is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle networking, and in particular to a method, system, equipment and medium for distinguishing collisions between a heavy truck tractor and a trailer. Background Art

[0002] In the heavy-duty truck transport sector, the combination of a tractor and trailer is extremely common. This unique vehicle structure, with the front of the tractor and trailer connected via a connecting device, presents certain safety risks during operation. For example, under emergency braking or in complex road conditions, the trailer is prone to accidents such as tailspinning, jackknifing, or rear-end collisions.

[0003] Existing collision detection technology primarily targets non-trailer collision scenarios and relies on front-mounted sensors, such as accelerometers and cameras. However, when a trailer collides, these sensors cannot directly detect the situation. While deploying independent sensors on the trailer can address this monitoring issue, this requires additional hardware and communication modules, significantly increasing costs. Summary of the Invention

[0004] The purpose of the present invention is to provide a heavy truck tractor trailer collision determination method, system, equipment and medium, which can solve the problem of high trailer collision detection cost.

[0005] To solve the above technical problems, an embodiment of the present invention provides a method for determining a collision between a heavy truck tractor and a trailer, comprising the following steps: Obtain driving data of the tractor in a heavy truck tractor trailer; Extracting characteristic parameters of driving data and judging whether the driving data is abnormal based on preset characteristic parameter thresholds to preliminarily determine the possibility of a collision between the heavy truck tractor and trailer; If the initial judgment is that the possibility of the trailer collision is greater than the preset threshold, the driving data is input into the trained collision discrimination model, and the collision discrimination model is used to extract the basic statistical features, change rate features, state combination features, spatial change features, time series features, and static features of the driving data; Among them, basic statistical features are used to characterize the distribution and central tendency of driving data in a time period, rate of change features are used to characterize the dynamic change process of driving data, state combination features are used to characterize the interaction between driving data and the comprehensive state of the vehicle, spatial change features are used to characterize the change of the vehicle's spatial position and its relationship with the vehicle state, time series features are used to characterize the process of vehicle state evolution over time, and static features are inherent attributes of driving data; Based on basic statistical features, change rate features, state combination features, spatial change features, time series features and static features, it is re-judged whether the trailer has collided to obtain the trailer collision judgment result.

[0006] Optionally, the driving data includes: the impact force of the heavy truck tractor trailer, acceleration in all directions, vehicle speed, steering angle, sound generated during the collision, and traction at the connection between the tractor and trailer, collected in real time; The step of extracting characteristic parameters of driving data includes: Obtain the peak impact force, peak acceleration and variance, rate of change of vehicle speed, rate of change of steering angle, and location of the sound source generated during the impact; wherein, the location of the sound source is used to indicate whether the sound source generated during the impact is in the direction of the trailer of the heavy truck tractor trailer.

[0007] Optionally, the location of the sound source generated during the collision is obtained by: The sound source is located by using an array of two microphones, and the angle between the sound source and the midpoint of the line connecting the two microphones is used as the location of the sound source generated during the collision, as shown in the following formula: ; Where, the angle between the sound source and the midpoint of the line connecting the two microphones is , is the speed of sound in air, is the time difference between the sound reaching the two microphones, is the distance between the two microphones.

[0008] Optionally, the determining whether the driving data is abnormal driving data based on a preset characteristic parameter threshold to preliminarily determine the possibility of a collision between the heavy truck tractor and trailer includes: Determine whether the peak value of the impact force, the peak value and variance of the acceleration, the rate of change of the vehicle speed, the rate of change of the steering angle, and the location of the sound source generated during the impact exceed the corresponding characteristic parameter thresholds. If so, the corresponding driving data is confirmed to be abnormal driving data; The weighted voting method is used to fuse the abnormality judgment results of all driving data to obtain the possibility of trailer collision.

[0009] Optionally, the collision determination result of the trailer is a collision level of the trailer, where the collision level includes a first-level collision and a second-level collision; After obtaining the trailer collision determination result, the method further includes: If the collision level is level one, the system stores the driving data for 30 seconds before the collision and outputs a warning message: "There may be an abnormality with the trailer, please pay attention." If the collision level is level 2, the emergency rescue channel will be connected and a warning message "The trailer may have collided, please check immediately!" will be output.

[0010] An embodiment of the present invention further provides a heavy truck tractor trailer collision identification system, comprising: A data acquisition module is used to obtain the driving data of the tractor in the heavy truck tractor trailer; A preliminary identification module is used to extract characteristic parameters of driving data and determine whether the driving data is abnormal according to a preset characteristic parameter threshold, so as to preliminarily determine the possibility of a collision between the heavy truck tractor trailer and the trailer; A model processing module is used to input the driving data into a trained collision discrimination model if the possibility of the trailer collision is initially determined to be greater than a preset threshold, and to extract the basic statistical features, change rate features, state combination features, spatial change features, time series features, and static features of the driving data through the collision discrimination model; Among them, basic statistical features are used to characterize the distribution and central tendency of driving data in a time period, rate of change features are used to characterize the dynamic change process of driving data, state combination features are used to characterize the interaction between driving data and the comprehensive state of the vehicle, spatial change features are used to characterize the change of the vehicle's spatial position and its relationship with the vehicle state, time series features are used to characterize the process of vehicle state evolution over time, and static features are inherent attributes of driving data; The final discrimination module is used to re-judge whether the trailer has collided based on basic statistical features, change rate features, state combination features, spatial change features, time series features and static features to obtain the collision discrimination result of the trailer.

[0011] An embodiment of the present invention also provides a computer device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the above-mentioned heavy-truck tractor trailer collision determination method.

[0012] An embodiment of the present invention further provides a computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the above-mentioned heavy truck tractor trailer collision determination method.

[0013] The heavy truck tractor trailer collision identification method provided by the present invention has at least the following beneficial effects: The present application does not set a sensor at the trailer position of the heavy truck tractor trailer, and only relies on the driving data of the truck head (i.e. the tractor of the heavy truck tractor trailer) to determine the collision of the trailer. First, the characteristic parameters of the driving data are analyzed to preliminarily determine the possibility of collision of the trailer in the heavy truck tractor trailer. Then, the collision determination model is combined for further determination. Without additional sensor cost, the collision of the trailer can be determined. The characteristic parameter analysis of the driving data and the collision determination model combined to determine the collision use different ways to obtain the vehicle state represented by the driving data of the truck head. Especially, the collision determination model extracts the basic statistical features, change rate features, state combination features, spatial change features, time sequence features and static features of the driving data to mine the features of the driving data of the truck head in various latitudes, analyze the vehicle change rule, and more comprehensively reflect the running state of the vehicle, so as to improve the accuracy of the collision determination of the trailer. BRIEF DESCRIPTION OF DRAWINGS

[0014] The accompanying drawings, which are included to provide a further understanding of the present application, form a part of the present application and are included to further explain the present application and, together with the specification, serve to explain the present application. In the drawings:

[0015] Figure 1 A flowchart of a heavy truck tractor trailer collision determination method provided by the present application is shown in the figure. Figure 2 A logic diagram of a heavy truck tractor trailer collision determination method provided by the present application is shown in the figure. DETAILED DESCRIPTION

[0016] To make the purpose, technical scheme and advantages of the present application clearer, the technical scheme of the present application will be described in detail below with reference to the specific embodiments of the present application and the corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0017] The technical scheme provided by each embodiment of the present application will be described in detail below with reference to the drawings.

[0018] One embodiment of the present application relates to a heavy truck tractor trailer collision determination method. The implementation details of the heavy truck tractor trailer collision determination method of the present embodiment will be described in detail below. The following content is only provided for the convenience of understanding the implementation details, and is not necessary for implementing the present scheme.

[0019] The specific process of the heavy truck tractor trailer collision determination method of the present embodiment can be as shown in Figure 1 ​ Step 101: Acquire driving data of the tractor in the heavy truck tractor trailer.

[0020] Specifically, the collision sensor, acceleration sensor, gyroscope, traction force sensor, driving recorder, and onboard terminal installed on the front of the vehicle collect relevant data in real time during the vehicle's driving process. Among them, the collision signal sensor directly detects the impact force on the vehicle; the acceleration sensor measures the change in acceleration of the vehicle in all directions; the gyroscope is used to measure the vehicle's rotational angular velocity; the traction force sensor collects traction force data at the connection between the tractor and trailer; the driving recorder records real-time images around the vehicle; and the audio data is used to record the sound generated by the collision.

[0021] In one example, after collecting various sensor data and terminal reported data, they are preprocessed, including cleaning and filtering, time synchronization, and normalization. Cleaning and filtering: Use the moving average filtering method with a window size of 10 to remove noise and outliers in the data to improve the quality and reliability of the data. The specific formula is: , ,⋯ , ,when When the filtered data ;when hour, .

[0022] Time synchronization: Ensure that data collected by different sensors and devices are synchronized in time to facilitate subsequent fusion analysis.

[0023] Data normalization: Normalize the data to make them have the same scale range to facilitate subsequent feature extraction and comparison.

[0024] Step 102 , extracting characteristic parameters of the driving data, and determining whether the driving data is abnormal driving data based on a preset characteristic parameter threshold, so as to preliminarily determine the possibility of a collision between the heavy truck tractor trailer and the trailer.

[0025] Specifically, the key features of the collected data are extracted in combination with the time window to obtain the characteristic parameters of each of the above data: ①The peak value of the collision sensor (the peak value of the impact force on the heavy truck tractor trailer), that is ,in For the processed data.

[0026] ②Traction force change rate, calculate the change rate of the force data at the connection between the tractor and the trailer , and take the maximum absolute value .

[0027] ③ The peak value and variance of vehicle acceleration in all directions, assuming the lateral acceleration is , the longitudinal acceleration is , the vertical acceleration is At the same time, the resultant acceleration of the vehicle on the horizontal plane (XY plane) is calculated ; Calculate the peak value of acceleration in each direction and synthetic acceleration data respectively and variance get , , , , , , , .

[0028] ; in, is the mean value of acceleration data, is the number of data points.

[0029] ④Vehicle speed change rate and steering angle change rate, , .

[0030] ⑤ Sound source localization (based on arrival time difference) using two microphones and The sound source is located by an array composed of , the time difference between the sound reaching the two microphones is , the distance between the two microphones is , the angle between the sound source and the midpoint of the line connecting the two microphones is , then according to the geometric relationship: By calculating The angle can be used to roughly determine the direction of the sound source, and combined with the vehicle structure to determine whether it comes from the direction of the trailer.

[0031] After obtaining the characteristic parameters of various driving data, each driving data is judged separately to see whether it exceeds the corresponding characteristic parameter threshold, namely the peak value of the impact force, the peak value and variance of the acceleration, the rate of change of the vehicle speed, the rate of change of the steering angle, and the location of the sound source generated during the impact. If so, the corresponding driving data is confirmed to be abnormal driving data; a weighted voting method is used to fuse the abnormality judgment results of all driving data to obtain the possibility of a trailer collision.

[0032] In the specific implementation, set the peak threshold ,like , then a preliminary judgment is made that a collision is likely to occur, denoted as ; otherwise, .

[0033] Set the threshold value of each angle acceleration peak , , , and the threshold value of variance , , , . If and , and , and , and , then a preliminary judgment is made that a collision is likely to occur, denoted as ; otherwise, .

[0034] Set the threshold value of force change rate . If , then a preliminary judgment is made that a collision is likely to occur, denoted as ; otherwise, .

[0035] Set the threshold value of vehicle speed change rate and the threshold value of steering angle change rate . If and , then a preliminary judgment is made that a collision is likely to occur, denoted as ; otherwise, .

[0036] Set the threshold value of sound source positioning . If , then a preliminary judgment is made that a collision is likely to occur, denoted as ; otherwise, .

[0037] If a certain feature parameter exceeds its corresponding threshold value, it is considered to be abnormal.

[0038] When at least one feature parameter is abnormal, a weighted voting method is used to fuse the preliminary judgment results. A weight is assigned to each preliminary judgment result, satisfying . The comprehensive score is calculated as

[0039] ; When ≥ At this time, it is considered that the trailer has a greater possibility of collision.

[0040] Step 103, if the preliminary determination of the trailer collision possibility is greater than the preset threshold, the driving data is input into the trained collision determination model, and the basic statistical features, rate of change features, state combination features, spatial change features, time sequence features and static features of the driving data are extracted through the collision determination model; Among them, the basic statistical features are used to represent the distribution and concentration trend of the driving data in a time period, the rate of change features are used to represent the dynamic change process of the driving data, the state combination features are used to represent the interaction between the driving data and the comprehensive state of the vehicle, the spatial change features are used to represent the change of the vehicle in the spatial position and its relationship with the vehicle state, the time sequence features are used to represent the evolution process of the vehicle state with time, and the static features are the inherent attributes of the driving data.

[0041] Step 104, according to the basic statistical features, the rate of change features, the state combination features, the spatial change features, the time sequence features and the static features, the collision of the trailer is determined again to obtain the collision determination result of the trailer.

[0042] Specifically, if the above preliminary determination considers that the trailer has a greater possibility of collision, the machine learning model (collision determination model) is entered into the determination step, which is realized by the following steps: (a) Deep feature mining based on existing data: Deep feature mining is performed based on existing data. From the six latitudes of basic statistical features, rate of change features, state combination features, spatial feature changes, time sequence features, static features, the vehicle change law is analyzed. The running state of the vehicle is comprehensively reflected.

[0043] (b) Model selection: The deep learning model (SVM) and the ensemble learning model (XGBOOST) are combined. The multi-dimensional data collected after optimization is input into the model for training, and the model automatically mines the potential collision features.

[0044] (c) Model fusion and cross-validation: Model fusion: The prediction results of the deep learning model and the ensemble learning model are fused. Through weighted average, according to the performance of different models on the validation set, different weights are assigned to them, and their prediction results are combined into the final collision recognition result.

[0045] Cross-validation: Use k-fold cross-validation to evaluate and optimize the model. This involves splitting the data into training and validation sets multiple times, training and validating the model on different subsets of the data. Based on the validation results, the model is adjusted and optimized to avoid overfitting and improve generalization.

[0046] (d) Model discrimination: The trained model is used to perform a final judgment on the feature parameters. The extracted feature parameters are combined with the window calculation and input into the trained model, which then outputs the judgment result. If the judgment result is 1, the trailer is considered to have collided; if the judgment result is 0, the trailer is considered to have not collided.

[0047] In one example, the trailer collision determination result is the collision level of the trailer collision, which includes level one collision and level two collision. After obtaining the trailer collision determination result, different responses are triggered based on the severity, i.e., the collision level: Level 1 collision (matching degree > 70%): Stores data from the 30 seconds before the accident, outputs "There may be an abnormality with the trailer, please pay attention.", and sends the vehicle ID, collision time, and location.

[0048] Level 2 collision (matching degree > 90% + video verification): Automatically connect to the emergency rescue channel, output "The trailer may have collided, please check immediately!", and send the vehicle identification, collision time and location.

[0049] In one embodiment, the heavy truck tractor trailer collision identification method of the present invention can be as follows: Figure 2 The logical implementation shown in the figure innovatively combines multi-dimensional features such as front-end sensor data, data reported by the on-board terminal, and vehicle static data to build a real-time analysis method to accurately and promptly determine whether a trailer has collided, thereby improving the safety of heavy-duty truck transportation.

[0050] It specifically includes the following parts: 0. Data collection: The collision sensor, acceleration sensor, gyroscope, traction sensor and driving recorder data installed on the front of the heavy-duty truck tractor and the on-board terminal data (such as vehicle speed, engine speed, torque, brakes, handbrake, double flash, etc.) are used to collect relevant data during the vehicle's driving process in real time and transmit the data to the data processing unit.

[0051] 1. Data preprocessing: After receiving the collected data, the data processing unit cleans and filters it. Taking vibration data as an example, a moving average filter method with a window size of 10 is used to remove noise and outliers in the data, making the data smoother and more reliable.

[0052] Data fusion and data normalization are performed on the cleaned and filtered data.

[0053] 2. Key feature extraction: Key features are extracted from the pre-processed data for threshold judgment. For example, for vibration data, find its peak value; for traction data, calculate its rate of change. These features are combined into a feature vector for subsequent discriminant analysis.

[0054] 3. Threshold judgment: Set a reasonable threshold for the extracted feature parameters. When a certain feature parameter exceeds the threshold, it is considered abnormal.

[0055] 4. Multi-parameter correlation analysis: Integrate multiple feature parameters for correlation analysis, and statistically calculate the number of abnormal feature parameters. If the weighted result exceeds the threshold, it is considered that the trailer is more likely to have a collision.

[0056] 5. Machine learning model discrimination: Based on existing data, from the six latitudes of basic statistical features, rate of change features, state combination features, spatial feature changes, time series features, and static features, the extracted feature vectors are input into the pre-trained model. The model calculates the parameters obtained by training and outputs the discrimination result. If the discrimination result is 1, it is determined that the trailer has a collision; if the discrimination result is 0, it is determined that the trailer has no collision.

[0057] 6. Collision classification and disposal: trigger different responses according to the severity.

[0058] The heavy truck tractor trailer collision discrimination method of the present application has the following beneficial effects: (1) Improve discrimination accuracy: comprehensively utilize the data of multiple sensors on the vehicle head, and discriminate through the combination of multi-parameter correlation analysis and machine learning model, effectively improve the accuracy and reliability of trailer collision discrimination, and reduce the situation of misjudgment and omission.

[0059] (2) Real-time monitoring: can real-time collect and analyze vehicle head data, timely find the signs of trailer collision, provide timely warning information for the driver, and help to avoid further deterioration of the accident.

[0060] (3) Low cost: without installing a large number of monitoring equipment on the trailer, only using the existing sensor data of the vehicle head for analysis, reducing the cost and installation complexity of the system.

[0061] The step division of the above various methods is only for the purpose of clear description, and can be combined into one step or split into multiple steps in implementation, as long as the same logical relationship is included, and all are within the protection scope of the present application; adding irrelevant modifications or introducing irrelevant designs in the algorithm or flow, but not changing the core design of the algorithm and flow, are within the protection scope of the present application.

[0062] Another embodiment of the present application relates to a heavy truck tractor trailer collision discrimination system, and the implementation details of the heavy truck tractor trailer collision discrimination system of the present embodiment are specifically described below. The following content is only provided for the implementation details for the convenience of understanding, and is not essential for implementing the present solution. The heavy truck tractor trailer collision discrimination system of the present embodiment comprises: A data acquisition module is configured to acquire driving data of a tractor in a heavy truck tractor trailer. A preliminary discrimination module is configured to extract feature parameters of the driving data, and determine whether the driving data is abnormal driving data according to a preset feature parameter threshold, so as to preliminarily discriminate the possibility of a collision of a trailer in the heavy truck tractor trailer. A model processing module is configured to input the driving data into a trained collision discrimination model if the possibility of the collision of the trailer is greater than a preset threshold, and extract basic statistical features, rate of change features, state combination features, spatial change features, time sequence features and static features of the driving data through the collision discrimination model. The basic statistical features are used to represent the distribution and central tendency of the driving data in a time period, the rate of change features are used to represent the dynamic change process of the driving data, the state combination features are used to represent the interaction between the driving data and the comprehensive state of the vehicle, the spatial change features are used to represent the change of the vehicle in the spatial position and the relationship between the vehicle state and the vehicle state, the time sequence features are used to represent the process of the evolution of the vehicle state with time, and the static features are inherent properties of the driving data. A final discrimination module is configured to re-discriminate whether the trailer has a collision according to the basic statistical features, the rate of change features, the state combination features, the spatial change features, the time sequence features and the static features, so as to obtain a collision discrimination result of the trailer.

[0063] It can be found that the present embodiment is a system embodiment corresponding to the above method embodiment, and the present embodiment can be implemented in cooperation with the above method embodiment. The related technical details and technical effects mentioned in the above embodiments are still effective in the present embodiment. In order to reduce repetition, they will not be described here. Correspondingly, the related technical details mentioned in the present embodiment can also be applied in the above embodiments.

[0064] It is worth mentioning that each module involved in the embodiment is a logical module, and in actual application, one logical unit can be one physical unit, or a part of one physical unit, or realized in combination of multiple physical units. In addition, in order to highlight the innovative part of the present application, units not closely related to solving the technical problems proposed by the present application are not introduced in the embodiment, but this does not mean that there are no other units in the embodiment.

[0065] Another embodiment of the present application relates to a computer device, comprising: at least one processor; and a memory connected with the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the heavy truck tractor trailer collision discrimination method in each of the above embodiments.

[0066] The memory and the processor are connected in a bus mode, the bus can include any number of interconnected buses and bridges, and the bus connects various circuits of the one or more processors and the memory together. The bus can also connect various other circuits such as peripheral devices, voltage regulators and power management circuits together, which are well known in the art, and therefore, they will not be further described herein. The bus interface provides an interface between the bus and the transceiver. The transceiver can be one element or multiple elements such as multiple receivers and transmitters, which provide a unit for communicating with various other devices on the transmission medium. The data processed by the processor is transmitted on the wireless medium through the antenna, and further, the antenna also receives data and transmits the data to the processor.

[0067] The processor is responsible for managing the bus and general processing, and can also provide various functions including timing, peripheral interface, voltage regulation, power management and other control functions. And the memory can be used to store data used by the processor in performing operations.

[0068] Another embodiment of the present application relates to a computer readable storage medium, which stores a computer program. The computer program is executed by the processor to realize the method embodiments.

[0069] That is, a person skilled in the art can understand that all or part of the steps in the above-mentioned embodiment methods can be completed by a program instructing relevant hardware, the program being stored in a storage medium and including a plurality of instructions for causing a device (which can be a single-chip microcomputer, a chip, etc.) or a processor to execute all or part of the steps of the methods described in various embodiments of the present application. The aforementioned storage medium includes a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various media that can store program codes.

[0070] A person of ordinary skill in the art can understand that the above-mentioned embodiments are specific embodiments for implementing the present application, and in actual applications, various changes can be made in form and details without departing from the spirit and scope of the present application.

Claims

1. A heavy truck tractor trailer collision identification method, characterized in that: The method comprises: Obtain driving data of the tractor in a heavy truck tractor trailer; Extracting characteristic parameters of driving data and judging whether the driving data is abnormal based on preset characteristic parameter thresholds to preliminarily determine the possibility of a collision between the heavy truck tractor and trailer; If the initial judgment is that the possibility of the trailer collision is greater than the preset threshold, the driving data is input into the trained collision discrimination model, and the collision discrimination model is used to extract the basic statistical features, change rate features, state combination features, spatial change features, time series features, and static features of the driving data; Among them, basic statistical features are used to characterize the distribution and central tendency of driving data in a time period, rate of change features are used to characterize the dynamic change process of driving data, state combination features are used to characterize the interaction between driving data and the comprehensive state of the vehicle, spatial change features are used to characterize the change of the vehicle's spatial position and its relationship with the vehicle state, time series features are used to characterize the process of vehicle state evolution over time, and static features are inherent attributes of driving data; Based on basic statistical features, change rate features, state combination features, spatial change features, time series features and static features, it is re-judged whether the trailer has collided to obtain the trailer collision judgment result.

2. The heavy truck tractor trailer collision determination method according to claim 1, characterized in that: The driving data includes: the real-time collection of the impact force on the heavy truck tractor trailer, acceleration in all directions, vehicle speed, steering angle, sound generated during the collision, and traction at the connection between the tractor and trailer; The step of extracting characteristic parameters of driving data includes: Obtain the peak impact force, peak acceleration and variance, rate of change of vehicle speed, rate of change of steering angle, and location of the sound source generated during the impact; wherein, the location of the sound source is used to indicate whether the sound source generated during the impact is in the direction of the trailer of the heavy truck tractor trailer.

3. The heavy truck tractor trailer collision determination method according to claim 2, characterized in that: The location of the sound source generated during the collision is obtained by: The sound source is located by using an array of two microphones, and the angle between the sound source and the midpoint of the line connecting the two microphones is used as the location of the sound source generated during the collision, as shown in the following formula: ； Where, the angle between the sound source and the midpoint of the line connecting the two microphones is , is the speed of sound in air, is the time difference between the sound reaching the two microphones, is the distance between the two microphones.

4. The heavy truck tractor trailer collision determination method according to claim 3, characterized in that: The determining whether the driving data is abnormal driving data based on a preset characteristic parameter threshold to preliminarily determine the possibility of a collision between the heavy truck tractor and trailer includes: Determine whether the peak value of the impact force, the peak value and variance of the acceleration, the rate of change of the vehicle speed, the rate of change of the steering angle, and the location of the sound source generated during the impact exceed the corresponding characteristic parameter thresholds. If so, the corresponding driving data is confirmed to be abnormal driving data; The weighted voting method is used to fuse the abnormality judgment results of all driving data to obtain the possibility of trailer collision.

5. The heavy truck tractor trailer collision determination method according to claim 1, characterized in that: The trailer collision determination result is a collision level of the trailer, which includes a first-level collision and a second-level collision. After obtaining the trailer collision determination result, the method further includes: If the collision level is level one, the system stores the driving data for 30 seconds before the collision and outputs a warning message: "There may be an abnormality with the trailer, please pay attention." If the collision level is level 2, the emergency rescue channel will be connected and a warning message "The trailer may have collided, please check immediately!" will be output.

6. A heavy truck tractor trailer collision identification system, characterized in that: The system comprises: A data acquisition module is used to obtain the driving data of the tractor in the heavy truck tractor trailer; A preliminary identification module is used to extract characteristic parameters of driving data and determine whether the driving data is abnormal according to a preset characteristic parameter threshold, so as to preliminarily determine the possibility of a collision between the heavy truck tractor trailer and the trailer; A model processing module is used to input the driving data into a trained collision discrimination model if the possibility of the trailer collision is initially determined to be greater than a preset threshold, and to extract the basic statistical features, change rate features, state combination features, spatial change features, time series features, and static features of the driving data through the collision discrimination model; Among them, basic statistical features are used to characterize the distribution and central tendency of driving data in a time period, rate of change features are used to characterize the dynamic change process of driving data, state combination features are used to characterize the interaction between driving data and the comprehensive state of the vehicle, spatial change features are used to characterize the change of the vehicle's spatial position and its relationship with the vehicle state, time series features are used to characterize the process of vehicle state evolution over time, and static features are inherent attributes of driving data; The final discrimination module is used to re-judge whether the trailer has collided based on basic statistical features, change rate features, state combination features, spatial change features, time series features and static features to obtain the collision discrimination result of the trailer.

7. A computer device, characterized in that: include: at least one processor; And, a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the heavy-duty truck tractor trailer collision determination method as described in any one of claims 1 to 5.

8. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the heavy truck tractor trailer collision determination method according to any one of claims 1 to 5 is implemented.

Images