A vehicle collision intelligent identification method, system, terminal device and storage medium

Abstract

The present invention relates to the technical field of vehicle collision recognition, and specifically includes a vehicle collision intelligent recognition method, system, terminal equipment and storage medium, including: marking M speed limit points in a collision recognition area, and identifying the lane and motion state when a target vehicle enters; using a bidirectional linked list and a sliding time window technology to respectively configure a speed monitoring mechanism for U maximum speed limit points and V minimum speed limit points; combining the vehicle state, building a collision risk recognition model, and sending the collision risk result to the target vehicle in real time, solving the technical problem that the collision risk cannot be accurately assessed in the case of high-speed driving or dense traffic and large speed differences, realizing the establishment of a vehicle motion state simulation model, combining the acceleration and angular velocity data obtained by an IMU sensor, accurately extracting the dynamic characteristics of the target vehicle, and being able to more comprehensively and accurately analyze the vehicle's motion state, thereby more effectively assessing the technical effect of the collision risk.

Description

Technical Field

[0001] The present invention relates to the technical field related to vehicle collision recognition, and in particular to a vehicle collision intelligent recognition method, system, terminal equipment and storage medium. Background Art

[0002] Vehicle collision warning technology mainly relies on sensors such as radar, LiDAR, and cameras, combined with computer vision algorithms, radar data fusion and other technologies to perform obstacle detection and collision prediction. Although it can effectively identify static and dynamic obstacles to a certain extent, it still has certain limitations when dealing with complex road environments, especially dynamic changes in traffic flow and vehicle speed differences when driving at high speeds. In addition, existing collision recognition systems often focus on static distance judgment or speed limit recognition, and lack accurate analysis and comprehensive evaluation of vehicle dynamic characteristics, resulting in insufficient response in some emergency situations and the inability to make accurate decisions in a short period of time.

[0003] In summary, the prior art has a technical problem of being unable to accurately assess collision risks when driving at high speeds or in dense traffic or when vehicle speeds vary greatly. Summary of the invention

[0004] The present application provides a vehicle collision intelligent identification method, system, terminal device and storage medium, aiming to solve the technical problem in the prior art that the collision risk cannot be accurately assessed when driving at high speed or in dense traffic or with large vehicle speed differences.

[0005] In view of the above problems, the technical solution to implement this application is:

[0006] In a first aspect, the present application provides a method for intelligently identifying vehicle collisions, wherein the method includes: marking M speed limit marking points within a collision identification area, wherein the M speed limit marking points include U maximum speed limit marking points and V minimum speed limit marking points, and M=U+V; after a target vehicle enters the collision identification area, identifying the lane in which the target vehicle is located and obtaining the vehicle's motion state; based on the U maximum speed limit marking points among the M speed limit marking points, using a bidirectional linked list architecture in conjunction with a first sliding time window, configuring a first speed monitoring mechanism; based on the V minimum speed limit marking points among the M speed limit marking points, using a bidirectional linked list architecture in conjunction with a second sliding time window, configuring a second speed monitoring mechanism; through the vehicle's motion state, establishing a collision risk identification model according to the first sliding time window and the first speed monitoring mechanism, the second sliding time window and the second speed monitoring mechanism; based on the collision risk identification model, sending the collision risk identification result to the target vehicle.

[0007] In a second aspect, the present application provides a vehicle collision intelligent identification system, wherein the system includes: a marking module, used to mark M speed limit marking points in a collision identification area, wherein the M speed limit marking points include U maximum speed limit marking points and V minimum speed limit marking points, and M=U+V; a data recording module, used to identify the lane in which the target vehicle is located and obtain the vehicle motion state after the target vehicle enters the collision identification area; a first configuration module, used to configure a first speed monitoring mechanism based on the U maximum speed limit marking points among the M speed limit marking points, using a bidirectional linked list architecture in conjunction with a first sliding time window; a second configuration module, used to configure a second speed monitoring mechanism based on the V minimum speed limit marking points among the M speed limit marking points, using a bidirectional linked list architecture in conjunction with a second sliding time window; a model building module, used to establish a collision risk identification model based on the vehicle motion state, according to the first sliding time window and the first speed monitoring mechanism, and the second sliding time window and the second speed monitoring mechanism; a risk identification module, used to send the collision risk identification result to the target vehicle based on the collision risk identification model.

[0008] In a third aspect of the present application, a terminal device is provided, which includes: a processor; a memory for storing instructions executable by the processor; wherein the processor is used to execute a vehicle collision intelligent identification method provided in the present application.

[0009] In a fourth aspect of the present application, a storage medium is provided, on which a computer program is stored, wherein when the computer program is executed by a processor, the steps of the vehicle collision intelligent identification method described in any one of the first aspects are implemented.

[0010] In summary, one or more technical solutions provided in the present application solve the technical problem that it is impossible to accurately assess the risk of collision when driving at high speed or in dense traffic or with large speed differences. It realizes the establishment of a vehicle's motion state simulation model when driving at high speed or in dense traffic, and combines the acceleration and angular velocity data obtained by the IMU sensor to accurately extract the dynamic characteristics of the target vehicle, which can more comprehensively and accurately analyze the vehicle's motion state, thereby more effectively assessing the technical effect of the risk of collision. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 A flow chart of a vehicle collision intelligent identification method is provided for this application.

[0012] Figure 2 A structural schematic diagram of a vehicle collision intelligent identification system is provided for this application.

[0013] Figure 3 A structural diagram of a terminal device is provided for this application.

[0014] Explanation of the reference numerals: marking module M100 , lane recognition module M200 , first configuration module M300 , second configuration module M400 , model building module M500 , risk identification module M600 , processor 21 , memory 22 , input device 23 , output device 24 . DETAILED DESCRIPTION

[0015] Embodiment 1

[0016] The present application is described in detail below with reference to the accompanying drawings. Figure 1 As shown, the present application provides a vehicle collision intelligent identification method, wherein the method comprises:

[0017] S1: Mark M speed limit marking points in the collision identification area, where the M speed limit marking points include U maximum speed limit marking points and V minimum speed limit marking points, and M=U+V; S2: After the target vehicle enters the collision identification area, identify the lane in which the target vehicle is located and obtain the vehicle's motion status.

[0018] Specifically, the collision recognition area refers to a predefined geographical area where the speed and behavior of vehicles will be monitored to identify potential collision risks; speed limit marking points refer to specific locations set up in the area to monitor vehicle speeds, which are divided into maximum speed limit marking points and minimum speed limit marking points, corresponding to the upper and lower limits of the speed respectively; the M in the M speed limit marking points represents the total number of marking points, which consists of U maximum speed limit marking points and V minimum speed limit marking points, that is, M=U+V; the target vehicle refers to the vehicle being monitored in the collision recognition area; lane recognition refers to determining the specific position of the target vehicle on the road; the vehicle's motion state includes parameters such as the vehicle's speed, acceleration, and direction, which are crucial for assessing collision risks.

[0019] Specific implementation steps: M speed limit marking points are marked in the preset collision identification area, including the maximum speed limit marking point and the minimum speed limit marking point, to ensure that the vehicle can be monitored under different speed limits. When the target vehicle enters the collision identification area, the lane in which the vehicle is located is determined by lane recognition technology (including cameras, sensors, etc.); the vehicle's motion state information is collected, and the IMU (Inertial Measurement Unit) sensor integrated in the target vehicle is used. The IMU sensor can provide the vehicle's acceleration and angular velocity data to determine the vehicle's real-time motion state; by identifying the lane and obtaining the vehicle's motion state, the vehicle behavior can be monitored more accurately, and the necessary input data for establishing a collision risk identification model can be provided.

[0020] S3: Based on the U highest speed limit marking points among the M speed limit marking points, use a bidirectional linked list architecture in conjunction with a first sliding time window to configure a first speed monitoring mechanism; S4: Based on the V lowest speed limit marking points among the M speed limit marking points, use a bidirectional linked list architecture in conjunction with a second sliding time window to configure a second speed monitoring mechanism.

[0021] Specifically, a doubly linked list architecture is a data structure that allows data elements to be linked in two directions, that is, each node has a pointer to the previous node and the next node. The doubly linked list architecture makes data insertion and deletion operations more efficient because there is no need to move other elements when inserting or deleting elements like an array; the first sliding time window and the second sliding time window are technologies used for real-time data analysis, which correspond to the maximum speed limit and the minimum speed limit respectively. The sliding time window is a dynamic data structure that slides on time series data to analyze data within a given time range; the first speed monitoring mechanism and the second speed monitoring mechanism are configured based on the first sliding time window and the second sliding time window, and are used to monitor whether the vehicle exceeds the maximum speed limit or is lower than the minimum speed limit.

[0022] Specific implementation steps: Based on U maximum speed limit marking points among M speed limit marking points, a bidirectional linked list architecture is used in conjunction with a first sliding time window to configure a first speed monitoring mechanism. The first speed monitoring mechanism will monitor the speed of the vehicle in real time and compare it with the maximum speed limit. If the speed of the vehicle exceeds the corresponding maximum speed limit, the speeding event will be recorded. Similarly, based on V minimum speed limit marking points among M speed limit marking points, a bidirectional linked list architecture is used in conjunction with a second sliding time window to configure a second speed monitoring mechanism to monitor whether the vehicle is below the minimum speed limit (generally speaking, if it is below the minimum speed limit of the fast lane, it can switch from the fast lane to the slow lane); by using a bidirectional linked list and a sliding time window, real-time data streams can be efficiently processed and analyzed, and speed changes can be quickly responded to, so that speeding or underspeed driving can be identified in a timely manner, and potential collision risks can be more accurately predicted and identified, thereby improving driving safety.

[0023] S5: Establish a collision risk identification model through the vehicle motion state, according to the first sliding time window and the first speed monitoring mechanism, and the second sliding time window and the second speed monitoring mechanism; S6: Send the collision risk identification result to the target vehicle based on the collision risk identification model.

[0024] Specifically, the collision risk identification model refers to an algorithm or system that integrates the vehicle's motion status, speed monitoring mechanism and time window technology, which is used to evaluate and predict the risk of vehicle collision. The collision risk identification model will analyze the vehicle's motion data, including speed, acceleration and direction, to determine the relationship with the speed limit, and then determine whether the vehicle is in a high-risk state; sending it to the target vehicle means transmitting the collision risk identification results to the target vehicle in real time, and through the vehicle's internal information system, such as the on-board display or warning system, to remind the driver of potential collision risks.

[0025] Specific implementation steps: Collect and analyze the vehicle's motion state data, including acceleration and angular velocity information obtained through IMU sensors, and speed information obtained through lane identification and speed limit marking points; combine the first sliding time window and the first speed monitoring mechanism (for the maximum speed limit), as well as the second sliding time window and the second speed monitoring mechanism (for the minimum speed limit) to establish a collision risk identification model. The collision risk identification model evaluates the vehicle's driving state in real time and compares it with the preset safety parameters to determine whether the vehicle is in a high-risk state; comprehensively evaluate the collision risk and determine whether the vehicle is likely to collide. Once a high-risk state is identified, it is immediately sent to the target vehicle so that the driver can take necessary risk avoidance measures, such as slowing down, changing lanes or taking emergency braking. Real-time risk identification and risk reminders can significantly improve driving safety and reduce the occurrence of traffic accidents.

[0026] Furthermore, the lane in which the target vehicle is located is identified and the vehicle motion state is obtained. The method of the present application also includes:

[0027] A three-axis acceleration sensor is connected to collect acceleration monitoring data; a three-axis gyroscope is connected to collect angular velocity monitoring data; based on the acceleration monitoring data and the angular velocity monitoring data, the dynamic characteristics of the target vehicle are extracted; and a motion state simulation model is established, wherein the motion state simulation model is used to match the vehicle motion state according to the dynamic characteristics of the target vehicle.

[0028] Specifically, a three-axis accelerometer is a sensor that can measure acceleration on three spatial axes (usually X, Y, and Z axes). It can detect changes in the vehicle's acceleration in all directions, including linear acceleration and acceleration when turning; a three-axis gyroscope is a sensor that measures angular velocity and can detect the vehicle's rotation speed around three axes; acceleration monitoring data and angular velocity monitoring data refer to data collected by a three-axis accelerometer and a three-axis gyroscope, and are the basis for dynamic characteristics analysis; dynamic characteristics refer to the physical characteristics exhibited by a vehicle during motion, such as acceleration, deceleration, steering angle, etc.; a motion state simulation model is a computer model that simulates the actual motion state of a vehicle and can predict the vehicle's behavior based on its dynamic characteristics.

[0029] Specific implementation steps: connect to the three-axis acceleration sensor and three-axis gyroscope on the vehicle, the three-axis acceleration sensor and the three-axis gyroscope are responsible for collecting the acceleration and angular velocity data of the vehicle on three spatial axes respectively; based on the acceleration monitoring data and the angular velocity monitoring data, extract the dynamic characteristics of the vehicle, including the acceleration change, steering angle, speed change, etc. of the vehicle; establish a motion state simulation model, the motion state simulation model will simulate the motion state of the vehicle according to the dynamic characteristics of the vehicle to match the actual vehicle motion.

[0030] In the above steps, by real-time monitoring and analysis of the vehicle's acceleration and angular velocity, the vehicle's actual motion state can be accurately understood. The motion state simulation model can predict the vehicle's behavior in different situations, thereby more effectively identifying potential collision risks and issuing warnings to the driver in a timely manner, thereby improving the accuracy and reliability of collision warnings and helping to reduce false alarms and missed alarms.

[0031] Furthermore, based on the acceleration monitoring data and the angular velocity monitoring data, the dynamic characteristics of the target vehicle are extracted, and the method of the present application further includes:

[0032] The acceleration monitoring data is low-pass filtered to obtain first initial processing data; the angular velocity monitoring data is low-pass filtered to obtain second initial processing data; complementary filtering is performed based on the first initial processing data and the second initial processing data to determine the dynamic characteristics of the target vehicle.

[0033] Specifically, low-pass filtering is a signal processing technology that allows low-frequency signals to pass through while suppressing high-frequency signals. In the context of vehicle motion data, low-pass filtering can be used to remove noise and unnecessary high-frequency fluctuations to obtain smoother and more accurate data; the first initial processed data and the second initial processed data refer to the acceleration data and angular velocity data after low-pass filtering, respectively; complementary filtering is a filtering technology that combines low-pass filtering and high-pass filtering. It improves the accuracy of data by combining the outputs of the two filters. In the analysis of vehicle dynamic characteristics, complementary filtering can be used to more accurately estimate the motion state of the vehicle.

[0034] Specific implementation steps: collect acceleration monitoring data from a three-axis acceleration sensor and angular velocity monitoring data from a three-axis gyroscope; perform low-pass filtering on the acceleration monitoring data and the angular velocity monitoring data to remove noise and unnecessary high-frequency fluctuations in the data, thereby obtaining smoother and more accurate acceleration and angular velocity data, and the data processed by low-pass filtering are referred to as first initial processing data and second initial processing data; perform complementary filtering based on the first initial processing data and the second initial processing data, and then combine the outputs of low-pass filtering and high-pass filtering to more accurately estimate the dynamic characteristics of the vehicle, such as the vehicle's speed, acceleration and steering angle.

[0035] In the above steps, through low-pass filtering and complementary filtering, noise and unnecessary fluctuations in the data can be more effectively removed, so as to obtain more accurate vehicle motion state information, provide a precise description of the actual motion state of the vehicle, thereby improving the accuracy of the data and more effectively identifying potential collision risks.

[0036] Furthermore, complementary filtering is performed based on the first initial processing data and the second initial processing data to determine the dynamic characteristics of the target vehicle. The method of the present application also includes:

[0037] Based on the first initial processing data, the first state variable and the first covariance matrix are initialized; the state transfer function and the observation judgment logic are set, and the IMU data is used for state prediction; at the same time, based on the first state variable and the first covariance matrix, a first complementary filtering loop is established through an updating step.

[0038] Specifically, the first state variable refers to the variable used to describe the current state of the vehicle in the vehicle motion state estimation, such as position, velocity, acceleration, etc.; the first covariance matrix is ​​a statistical matrix used to describe the uncertainty of state variables and the correlation between them; the state transfer function is a mathematical function used to describe the change of vehicle state over time; the observation judgment logic refers to the logic of updating the state variables based on sensor data; IMU data refers to data from the inertial measurement unit (IMU), including acceleration and angular velocity data; the first complementary filtering loop refers to a process of continuously updating state variables and covariance matrices to improve the accuracy of state estimation.

[0039] Specific implementation steps: Initialize the first state variable and the first covariance matrix based on the first initial processing data (i.e., the acceleration data processed by low-pass filtering). Initializing the first state variable and the first covariance matrix provides a starting point for subsequent state estimation; set the state transfer function and the observation judgment logic, and use the IMU data for state prediction. The state transfer function will predict the state at the next moment based on the physical model of the vehicle, and the observation judgment logic will update the state prediction based on the actual IMU data; based on the first state variable and the first covariance matrix, establish the first complementary filtering loop through the updating step. The first complementary filtering loop will continuously update the state variables and the covariance matrix based on the new IMU data, thereby improving the accuracy of state estimation.

[0040] In the above steps, by initializing the state variables and covariance matrix, and using the state transfer function and observation judgment logic to continuously update the state prediction, the vehicle's motion state can be tracked more accurately.

[0041] Furthermore, the present application method also includes:

[0042] Initialize the second state variable and the second covariance matrix based on the second initial processing data; calculate the complementary gain value based on the second state variable and the second covariance matrix; use the complementary gain value to update the estimated value of the state variable and the covariance matrix of the prediction error, and formulate a second complementary filtering loop.

[0043] Specifically, the second initial processed data refers to the angular velocity data processed by low-pass filtering, which is obtained from the three-axis gyroscope and is used to capture the rotational dynamics of the vehicle; the second state variable and the second covariance matrix are another set of variables and matrices in state estimation, similar to the first state variable and the first covariance matrix, but they are based on angular velocity data; the complementary gain value is a parameter used in complementary filtering, which determines the mixing ratio of the low-pass filter and high-pass filter outputs to optimize the accuracy of the state estimation; the second complementary filtering loop refers to the continuous update process based on the second state variable and the covariance matrix, which is used to continuously optimize the estimation of the vehicle's motion state.

[0044] Specific implementation steps: Initialize the second state variables and the second covariance matrix based on the second initial processing data (i.e., the angular velocity data after low-pass filtering), and initialize the second state variables and the second covariance matrix to describe the rotation state of the vehicle and the related uncertainty; calculate the complementary gain value based on the second state variables and the second covariance matrix, and the complementary gain value will determine how to combine the acceleration data (low-frequency information) and the angular velocity data (high-frequency information) to obtain a more accurate state estimate; use the calculated complementary gain value to update the estimated value of the state variable and the covariance matrix of the prediction error, thereby formulating a second complementary filtering loop, and the second complementary filtering loop will be continuously iterated to update the vehicle's motion state estimate in real time.

[0045] In the above steps, by initializing the second state variable and covariance matrix and calculating the complementary gain value to update the state estimate, the angular velocity change of the vehicle is captured more accurately. The second complementary filtering loop is established to continuously optimize the state estimate, provide more reliable input data for the collision risk identification model, and enhance the prediction ability and safety of the entire system.

[0046] Furthermore, the state transfer function and observation judgment logic are set, and the state prediction is performed using IMU data. The method of the present application also includes:

[0047] Will Defined as the predicted state of the ith floating point at time step k, for a state vector of dimension n, it satisfies: ;in, is the mean of the predicted states, is the covariance matrix of the predicted state, The i-th floating point in the covariance weight, is the covariance matrix of the process noise.

[0048] Specifically, the i-th floating point refers to the value used to represent the vehicle state in the state vector, including the acceleration (ACCx, ACCy, ACCz) and angular velocity (GyroX, GyroY, GyroZ) corresponding to the IMU data; time step k refers to the k-th time point in the time series; predicted state refers to the estimate of the vehicle state at the current or future time based on the state transfer function and previous observation data; the state vector is a vector containing all state variables, which is used to describe the motion state of the vehicle; the covariance matrix is a matrix describing the uncertainty of state variables and the correlation between them; covariance weights is an element in the covariance matrix that represents the uncertainty of a particular state variable; process noise It refers to the random error introduced in the process of state transfer.

[0049] Specific implementation steps: Defined as the predicted state of the ith floating point at time step k, for a state vector of dimension n, it satisfies: ; By defining the state vector, mean, covariance matrix and process noise, the vehicle's motion state can be described more accurately, providing a mathematical basis for the prediction of the vehicle's motion state. In addition, the motion state simulation model and collision risk identification model rely on accurate vehicle motion data to predict potential collision risks. A mathematical model is established for the prediction of the vehicle's motion state, thereby improving the accuracy of state prediction, more effectively identifying potential collision risks, and issuing timely warnings to the driver.

[0050] Furthermore, according to the first sliding time window and the first speed monitoring mechanism, the second sliding time window and the second speed monitoring mechanism, a collision risk identification model is established, and the method of the present application also includes:

[0051] A bidirectional linked list architecture is defined, where the size of each linked list is N, which is used to store ACCy values ​​and ACCz values ​​respectively; based on the bidirectional linked list architecture, a circular buffer method is adopted to formulate a circular storage protocol; under the circular storage protocol, a node structure is configured, wherein each node contains two floating-point numbers and pointers to the front and rear nodes; based on the node structure, a pointer to the front node is used to traverse forward to set the front collision risk network layer according to the first sliding time window and the first speed monitoring mechanism; based on the node structure, a pointer to the rear node is used to traverse backward to set the rear collision risk network layer according to the second sliding time window and the second speed monitoring mechanism; the front collision risk network layer and the rear collision risk network layer are connected to determine the collision risk identification model.

[0052] Specifically, the bidirectional linked list architecture is a data structure in which each node contains data and two pointers, pointing to the previous and next nodes respectively, allowing forward or backward traversal from any node; the ACCy value and the ACCz value represent the acceleration values ​​in the y-axis and z-axis directions respectively; the ring buffer method is a special buffer management method, in which data is stored in a fixed-size circular buffer, and when the buffer is full, new data will overwrite the old data; the circular storage protocol refers to the storage and access rules of data in the ring buffer; the node structure refers to the specific structure of each node in the linked list, including the stored data and pointers; the front collision risk network layer and the rear collision risk network layer are logical layers based on the vehicle motion state and speed monitoring mechanism for evaluating the front and rear collision risks of the vehicle.

[0053] Specific implementation steps: define a bidirectional linked list architecture, where the size of each linked list is N, which is used to store ACCy values ​​and ACCz values ​​respectively. The bidirectional linked list architecture allows the acceleration data to be stored and accessed in both directions. A ring buffer method is adopted, and a circular storage protocol is formulated, which means that when the linked list reaches its maximum capacity N, the new data will overwrite the oldest data to form a continuously updated data stream. Under the circular storage protocol, a node structure is configured, where each node contains two floating-point numbers (used to store ACCy values ​​and ACCz values) and pointers to the previous and next nodes.

[0054] Utilizing the node structure, using a pointer pointing to the front node, according to the first sliding time window and the first speed monitoring mechanism, traversing the linked list forward, setting the front collision risk network layer, the front collision risk network layer is responsible for evaluating the collision risk of the front of the vehicle; similarly, using a pointer pointing to the rear node, according to the second sliding time window and the second speed monitoring mechanism, traversing the linked list backward, setting the rear collision risk network layer, the rear collision risk network layer is responsible for evaluating the collision risk of the rear of the vehicle; connecting the front collision risk network layer and the rear collision risk network layer to form a complete collision risk identification model.

[0055] In the above steps, by using a bidirectional linked list and a circular buffer, data can be quickly accessed and updated while maintaining data continuity. The setting of the front and rear collision risk network layers enables the system to evaluate the collision risks of the front and rear of the vehicle respectively, and connecting the front collision risk network layer and the rear collision risk network layer forms a collision risk identification model. In addition, by clarifying the collision detection of front collision and rear collision, the collision risk identification model can be effectively integrated with the safety system, which has a certain degree of scalability and can adapt to more application scenarios.

[0056] In summary, the beneficial effects of the embodiments of the present application are:

[0057] The present application marks M speed limit points in the collision identification area, and identifies the lane and motion status when the target vehicle enters; uses a bidirectional linked list and a sliding time window technology to configure speed monitoring mechanisms for U maximum speed limit points and V minimum speed limit points respectively; builds a collision risk identification model in combination with the vehicle status, and sends the collision risk results to the target vehicle in real time, thereby achieving the establishment of a vehicle motion state simulation model under high-speed driving or dense traffic conditions, and accurately extracts the dynamic characteristics of the target vehicle in combination with the acceleration and angular velocity data obtained by the IMU sensor, which can more comprehensively and accurately analyze the vehicle's motion state, thereby more effectively evaluating the technical effect of the collision risk.

[0058] Embodiment 2

[0059] Based on the same inventive concept as the vehicle collision intelligent identification method in the aforementioned embodiment, Figure 2As shown, an embodiment of the present application provides a vehicle collision intelligent identification system, wherein the system includes:

[0060] The marking module M100 is used to mark M speed limit marking points in the collision identification area, wherein the M speed limit marking points include U maximum speed limit marking points and V minimum speed limit marking points, and M=U+V;

[0061] The lane recognition module M200 is used to identify the lane where the target vehicle is located and obtain the vehicle movement state after the target vehicle enters the collision recognition area;

[0062] A first configuration module M300 is used to configure a first speed monitoring mechanism based on U maximum speed limit marking points among the M speed limit marking points by using a bidirectional linked list architecture in conjunction with a first sliding time window;

[0063] A second configuration module M400 is used to configure a second speed monitoring mechanism based on V minimum speed limit marking points among the M speed limit marking points using a bidirectional linked list architecture in conjunction with a second sliding time window;

[0064] A model building module M500 is used to build a collision risk identification model according to the vehicle motion state, the first sliding time window and the first speed monitoring mechanism, and the second sliding time window and the second speed monitoring mechanism;

[0065] The risk identification module M600 is used to send a collision risk identification result to the target vehicle based on the collision risk identification model.

[0066] Furthermore, the lane recognition module M200 is also used to execute the following method:

[0067] Connect the three-axis acceleration sensor to collect acceleration monitoring data;

[0068] Connect to the three-axis gyroscope to collect angular velocity monitoring data;

[0069] Extracting the dynamic characteristics of the target vehicle based on the acceleration monitoring data and the angular velocity monitoring data;

[0070] A motion state simulation model is established, wherein the motion state simulation model is used to match the vehicle motion state according to the dynamic characteristics of the target vehicle.

[0071] Furthermore, the lane recognition module M200 is also used to execute the following method:

[0072] Performing low-pass filtering on the acceleration monitoring data to obtain first initial processing data;

[0073] Performing low-pass filtering on the angular velocity monitoring data to obtain second initial processing data;

[0074] Complementary filtering is performed based on the first initial processing data and the second initial processing data to determine the dynamic characteristics of the target vehicle.

[0075] Furthermore, the lane recognition module M200 is also used to execute the following method:

[0076] Initializing a first state variable and a first covariance matrix based on the first initial processed data;

[0077] Set the state transfer function and observation judgment logic, and use IMU data for state prediction;

[0078] At the same time, based on the first state variable and the first covariance matrix, a first complementary filtering loop is established through an updating step.

[0079] Furthermore, the lane recognition module M200 is also used to execute the following method:

[0080] Initializing a second state variable and a second covariance matrix based on the second initial processed data;

[0081] Calculating a complementary gain value based on the second state variable and the second covariance matrix;

[0082] The complementary gain values ​​are used to update the estimated values ​​of the state variables and the covariance matrix of the prediction error, and a second complementary filtering loop is formulated.

[0083] Furthermore, the lane recognition module M200 is also used to execute the following method:

[0084] Will Defined as the predicted state of the ith floating point at time step k, for a state vector of dimension n, it satisfies: ;

[0085] in, is the mean of the predicted states, is the covariance matrix of the predicted state, The i-th floating point in the covariance weight, is the covariance matrix of the process noise.

[0086] Furthermore, the model building module M500 is used to execute the following method:

[0087] Define a bidirectional linked list architecture, where the size of each linked list is N, which is used to store ACCy values ​​and ACCz values ​​respectively;

[0088] Based on the bidirectional linked list architecture, a circular storage protocol is proposed using a ring buffer method;

[0089] Under the circular storage protocol, a node structure is configured, wherein each node contains two floating point numbers and pointers to the previous and next nodes;

[0090] Based on the node structure, using a pointer pointing to a previous node, and according to the first sliding time window and the first speed monitoring mechanism, traverse forward to set a vehicle head collision risk network layer;

[0091] Based on the node structure, using a pointer pointing to a rear node, and according to the second sliding time window and the second speed monitoring mechanism, traversing backward to set a rear-end collision risk network layer;

[0092] The frontal collision risk network layer and the rear collision risk network layer are connected to determine the collision risk identification model.

[0093] Embodiment 3

[0094] like Figure 3 , which is a schematic diagram of the structure of a terminal device provided in Example 2 of the present application, and is a block diagram of an exemplary terminal device suitable for implementing an implementation mode of the present invention. Figure 3 The terminal device shown is only an example and should not limit the functions and scope of use of the embodiments of the present invention. Figure 3 As shown, the terminal device includes a processor 21, a memory 22, an input device 23 and an output device 24; the number of processors 21 in a terminal device can be one or more. Figure 3 Taking a processor 21 as an example, the processor 21, memory 22, input device 23 and output device 24 in a terminal device can be connected through a bus or other means. Figure 3 The example of connecting through bus is taken in the following.

[0095] The memory 22 is a computer-readable storage medium that can be used to store software programs, computer executable programs and modules, such as program instructions / modules corresponding to a vehicle collision intelligent identification method in an embodiment of the present application. The processor 21 executes various functional applications and data processing of the computer device by running the software programs, instructions and modules stored in the memory 22, that is, realizing the above-mentioned vehicle collision intelligent identification method.

[0096] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0097] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

[0098] This specification and the drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.

Claims

1. A vehicle collision intelligent identification method, characterized in that: The method comprises: Marking M speed limit marking points in the collision identification area, wherein the M speed limit marking points include U maximum speed limit marking points and V minimum speed limit marking points, and M=U+V; After the target vehicle enters the collision identification area, identifying the lane where the target vehicle is located and obtaining the vehicle motion state; Based on the U highest speed limit marking points among the M speed limit marking points, a first speed monitoring mechanism is configured using a bidirectional linked list architecture in conjunction with a first sliding time window; based on the V lowest speed limit marking points among the M speed limit marking points, a second speed monitoring mechanism is configured using a bidirectional linked list architecture in conjunction with a second sliding time window; Among them, the bidirectional linked list architecture is a data structure that allows data elements to be linked in two directions, and each node has a pointer to the previous node and the next node; the first sliding time window and the second sliding time window are used for real-time data analysis, corresponding to the maximum speed limit and the minimum speed limit respectively, and the sliding time window is a dynamic data structure that slides on the time series data and analyzes the data within a given time range; the first speed monitoring mechanism and the second speed monitoring mechanism monitor whether the vehicle exceeds the maximum speed limit or is lower than the minimum speed limit; By means of the vehicle motion state, according to the first sliding time window and the first speed monitoring mechanism, and the second sliding time window and the second speed monitoring mechanism, a collision risk identification model is established, including: Define a bidirectional linked list architecture, where the size of each linked list is N, and is used to store ACCy and ACCz values ​​respectively; ACCy and ACCz values ​​represent the acceleration values ​​in the y-axis and z-axis directions respectively; Based on the bidirectional linked list architecture, a circular storage protocol is proposed using a ring buffer method; Under the circular storage protocol, a node structure is configured, wherein each node of a bidirectional linked list contains two floating point numbers for storing ACCy value, ACCz value and pointers to the previous and next nodes respectively; Based on the node structure, using a pointer pointing to a previous node, and according to the first sliding time window and the first speed monitoring mechanism, traverse forward to set a vehicle head collision risk network layer; Based on the node structure, using a pointer pointing to a rear node, and according to the second sliding time window and the second speed monitoring mechanism, traversing backward to set a rear-end collision risk network layer; Connecting the front collision risk network layer and the rear collision risk network layer to determine the collision risk identification model; Based on the collision risk identification model, a collision risk identification result is sent to the target vehicle.

2. A vehicle collision intelligent identification method as claimed in claim 1, characterized in that: Identifying the lane in which the target vehicle is located and obtaining the vehicle motion state, the method further includes: Connect the three-axis acceleration sensor to collect acceleration monitoring data; Connect to the three-axis gyroscope to collect angular velocity monitoring data; Extracting the dynamic characteristics of the target vehicle based on the acceleration monitoring data and the angular velocity monitoring data; A motion state simulation model is established, wherein the motion state simulation model is used to match the vehicle motion state according to the dynamic characteristics of the target vehicle.

3. A vehicle collision intelligent identification method as claimed in claim 2, characterized in that: Based on the acceleration monitoring data and the angular velocity monitoring data, the dynamic characteristics of the target vehicle are extracted, and the method includes: Performing low-pass filtering on the acceleration monitoring data to obtain first initial processing data; Performing low-pass filtering on the angular velocity monitoring data to obtain second initial processing data; Complementary filtering is performed based on the first initial processing data and the second initial processing data to determine the dynamic characteristics of the target vehicle.

4. A vehicle collision intelligent identification method as claimed in claim 3, characterized in that: Complementary filtering is performed based on the first initial processing data and the second initial processing data to determine the dynamic characteristics of the target vehicle, the method comprising: Initializing a first state variable and a first covariance matrix based on the first initial processed data; Set the state transfer function and observation judgment logic, and use IMU data for state prediction; At the same time, based on the first state variable and the first covariance matrix, a first complementary filtering loop is established through an updating step.

5. A vehicle collision intelligent identification method as claimed in claim 4, characterized in that: Initializing a second state variable and a second covariance matrix based on the second initial processed data; Calculating a complementary gain value based on the second state variable and the second covariance matrix; The complementary gain values ​​are used to update the estimated values ​​of the state variables and the covariance matrix of the prediction error, and a second complementary filtering loop is formulated.

6. A vehicle collision intelligent identification method as claimed in claim 5, characterized in that: The state transfer function and observation judgment logic are set, and the state prediction is performed using IMU data. The method includes: Will Defined as the predicted state of the ith floating point at time step k, for a state vector of dimension n, it satisfies: ; in, is the mean of the predicted states, is the covariance matrix of the predicted state, The i-th floating point in the covariance weight, is the covariance matrix of the process noise.

7. A vehicle collision intelligent identification system, characterized in that: A vehicle collision intelligent identification method for implementing any one of claims 1 to 6, the system comprising: A marking module, used to mark M speed limit marking points in the collision identification area, wherein the M speed limit marking points include U maximum speed limit marking points and V minimum speed limit marking points, and M=U+V; A data recording module, used to identify the lane where the target vehicle is located and obtain the vehicle movement state after the target vehicle enters the collision identification area; A first configuration module is used to configure a first speed monitoring mechanism based on U maximum speed limit marking points among the M speed limit marking points by using a bidirectional linked list architecture in conjunction with a first sliding time window; A second configuration module is used to configure a second speed monitoring mechanism based on V lowest speed limit marking points among the M speed limit marking points using a bidirectional linked list architecture in conjunction with a second sliding time window; A model building module, used to build a collision risk identification model according to the vehicle motion state, the first sliding time window and the first speed monitoring mechanism, and the second sliding time window and the second speed monitoring mechanism; The risk identification module is used to send a collision risk identification result to the target vehicle based on the collision risk identification model.

8. A terminal device, characterized in that: include: A memory for storing executable instructions; The processor is used to implement a vehicle collision intelligent identification method according to any one of claims 1 to 6 when executing the executable instructions stored in the memory.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, a vehicle collision intelligent identification method as described in any one of claims 1 to 6 is implemented.

Images