Perception shared message generation method oriented to Internet of Vehicles and based on perception precision
By establishing a local autonomous and networked collaboration perception model, comparing the perception accuracy and deciding whether to add SSM messages to be sent, the problem of perceived shared messages in dense traffic environments is solved, and more accurate system state estimation and more efficient Perceive system performance.
Patent Information
- Application Number
- CN202411269192.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-11
- Publication Date
- 2025-02-07
AI Technical Summary
In dense traffic environments, there is a large amount of duplication of target information perceived by V2X users, resulting in redundancy and information overload of perceived shared messages, increasing transmission delays and potentially losing valuable data packets.
By establishing a local autonomous perception model and a networked collaborative perception model, comparing the perception accuracy between the two, using Kalman filtering and Kullback-Leibler divergence to decide whether to add the detection result to the SSM message to be sent, reducing information redundancy and improve the efficiency of the perception system.
More accurate system status estimation is achieved, reducing information redundancy and transmission delay, improving the efficiency and reliability of the vehicle perception system, and avoiding information pollution.
Smart Images
Figure CN119402834A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of vehicle networking, and more specifically, to a method for generating perception sharing messages based on perception accuracy for vehicle networking. Background Art
[0002] Perception sharing is a typical V2X enhanced application. It defines the perception sharing message SSM for perception expansion scenarios. The message can be sent by the on-board unit OBU, the roadside unit RSU and the pedestrian VRU. Vehicles or roadside equipment detect the surrounding traffic environment through their own perception equipment, including other V2X or non-V2X traffic participants and abnormal road conditions, and notify the surrounding V2X facilities of the target information and road traffic events acquired through perception sharing messages, so as to expand the perception range of other V2X vehicles and roadside facilities, assist driving decisions, and thus improve traffic safety and efficiency.
[0003] The process of perception data sharing involves more than just the perception, abstraction, and message encapsulation of target data. In dense traffic environments, V2X users perceive a large amount of target information and transmit it. For users receiving perception data sharing messages, there is obvious duplication of information received from users with similar locations. Therefore, congestion control of perception sharing messages can be performed at the message sending end, that is, when each message is sent, the type, dynamics, and information transmission and reception of each detected target object are used to determine whether to encapsulate its abstract information into the current message.
[0004] As more and more ADAS-equipped vehicles hit the road, the accuracy of detection by each vehicle's own perception equipment varies, as the detection accuracy that can be provided by the vehicle's own perception equipment is also different.
[0005] Since different on-board sensing devices and roadside sensing devices have a common viewing area, they will detect the same traffic objects and thus be encapsulated into SSM messages. It is worth studying how to reduce information redundancy. In addition, the information overload caused by high redundancy will increase transmission delays and even create the risk of losing valuable data packets. Since actual detection and target tracking may not follow the expected movement of the target, it will lead to frequent state changes, generating more SSM messages containing inaccurate detection results to affect the results of autonomous perception. At the same time, the frequent transmission of unreliable data will have a significant negative impact on the channel load. Summary of the invention
[0006] In view of the technical problems existing in the prior art, the present invention provides a method for generating perception sharing messages based on perception accuracy for the Internet of Vehicles to solve the problems raised in the above-mentioned background technology.
[0007] The technical solution of the present invention to solve the above technical problem is as follows: a method for generating a perception sharing message based on perception accuracy for an Internet of Vehicles, comprising the following steps:
[0008] S101. Detect target information data and pre-process it, and establish a local autonomous perception model to track the target information data of the current node;
[0009] S102. Establishing a networked collaborative perception model to track the objects received by the current node through SSM;
[0010] S103. Each node determines whether to add the detection result to the new list of SSM objects to be sent by comparing the tracking accuracy of the local autonomous perception model with the tracking accuracy of the networked collaborative perception model;
[0011] In a preferred embodiment, in S101, the vehicle-mounted unit and the road test unit of the current node are used to detect the data of other traffic participants around the road of the current node and the road abnormality information data of the current node through a radar sensing device and transmit and construct target information data, the target information data is denoised using a Kalman filter, the target is continuously detected and the real-time status of each target is continuously tracked through the radar sensing device, the road abnormality information of the current node is mapped according to the real-time status of each target, and the data of other traffic participants around the road of the current node and the road abnormality information data of the current node are time-aligned.
[0012] Furthermore, based on the Kalman filter, a local autonomous perception model is constructed, and a probability model is established to characterize the perception accuracy of the radar perception device. Based on the visible cross section and the relative distance between the reference vehicle and the detected target, the measurement accuracy of the probability model is determined by adding noise to the true position of the target. The specific formula is:
[0013] D O (c,d)=(σ x ,σ y )(x,y)
[0014] Where c represents the cross section, d represents the distance, and D O () indicates that the value returned by the function is a tuple, which is used to detect the variance of the current node in the Cartesian coordinate system. x ,σ y) represents the value of the true ground position at (x, y) as the average value in the normal distribution, and randomly selects a value in the normal distribution as the detection value of the radar perception device of the current node. The Kalman filter is used to adjust the radar perception device detection value and the previous vehicle trajectory by continuously applying the dynamic vehicle model. The historical information and the new detection target information data are integrated by adopting the prediction and correction mechanism. When the detection update is obtained at each time step k, the detection operation updates the Kalman filter. The whole update process is divided into a prediction step and an update step.
[0015] Furthermore, based on the Kalman filter, the error between the detection value of the radar sensing device and the vehicle trajectory is continuously adjusted by using the vehicle motion model. The state vector in the prediction step is predicted based on the previous state and the matrix used to describe the vehicle motion model. The specific formula is:
[0016] x k =x k-1 *F k-1
[0017] Among them, x k represents the state vector, x k-1 represents the previous state vector, F k-1 Represents the matrix describing the vehicle motion model, updates the error covariance matrix and adds noise. The specific formula is:
[0018]
[0019] Among them, P k represents the error covariance matrix, represents the transposed matrix of the vehicle motion model described previously, Q k-1 Represents noise. After completing the prediction step, the update step is automatically started. The detection value based on the radar perception device is corrected by calculating the Kalman gain, and the current state is updated based on the detection value and prediction value of the radar perception device of the current node, and the covariance matrix of the process is updated again.
[0020] In a preferred embodiment, in S102, a networked collaborative perception model is established to track a list of target information data received by the current node through SSM, and the networked collaborative perception model uses V2X communication to track the target information data.
[0021] In a preferred embodiment, in S103, each node determines whether to add the detection result to the list of target information data received by the new SSM to be sent by comparing the tracking accuracy of the local autonomous perception model with the tracking accuracy of the networked cooperative perception model. The threshold is defined as 1. When the tracking of the local autonomous perception model is less than the threshold, the tracking is considered valuable and the next step is performed. Otherwise, the detection result is not considered to be included in the list of target information data received by the SSM to be sent next time. The next step calculates the relative entropy between the tracking accuracy of the local autonomous perception model and the tracking accuracy of the networked cooperative perception model through Kullback-Leibler divergence. The specific calculation formula of the Kullback-Leibler divergence is:
[0022]
[0023] Among them, D KL (N LEM ||N V2X ) represents the Kullback-Leibler divergence of the probability distribution of the local autonomous perception model and the networked collaborative perception model, P LEM , P V2X They respectively represent the tracking accuracy of the local autonomous perception model and the networked collaborative perception model, μ1 and μ2 respectively represent the mean state vector of the local autonomous perception model and the networked collaborative perception model, and the calculation result is compared with the set threshold γ, where the threshold γ is defined as 3. When the relative entropy is high enough and greater than the set threshold γ, it proves that the target information data contains rich information and is not highly repetitive, and is added to the list of target information data received by the next SSM to be sent. Otherwise, the detection result will not be considered to be included in the list of target information data received by the next SSM to be sent.
[0024] The beneficial effects of the present invention are as follows: a perception model is established based on Kalman filtering to characterize the perception accuracy of radar perception equipment, the perception accuracy is determined according to factors such as distance and degree of occlusion, and the measurement accuracy is simulated by adding noise to reduce the impact of detection on tracking position, providing a more accurate system state estimation, fusing historical information and new data through a prediction and correction mechanism, which is divided into a prediction step and an update step, and correcting the detection value through the Kalman gain to provide a more stable and accurate position estimation, using networked collaboration and V2X communication technology to improve the sharing and tracking of target information data, and improving the efficiency and accuracy of the vehicle perception system, using the Kalman filter to fuse historical information and new detection target information data to ensure the tracking accuracy of the local autonomous perception model, and by comparing the tracking accuracy of the perception model and using the Kullback-Leibler divergence to determine the processing of the detection results, it is helpful to improve the efficiency and reliability of the system in perception and collaboration, and by measuring the local autonomous perception model and the networked collaborative perception model, a detection object with higher accuracy is selected from them to be included in the list of SSM message objects to be sent, which is conducive to avoiding information pollution. BRIEF DESCRIPTION OF THE DRAWINGS
[0025] Figure 1 is a flow chart of the method of the present invention; DETAILED DESCRIPTION
[0026] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of this application.
[0027] In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of this application, "plurality" means two or more, unless otherwise clearly and specifically defined.
[0028] In the description of the present application, the term "for example" is used to mean "used as an example, illustration or description". Any embodiment described as "for example" in the present application is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is given to enable any technician in the field to implement and use the present invention. In the following description, details are listed for the purpose of explanation. It should be understood that a person of ordinary skill in the art can recognize that the present invention can be implemented without using these specific details. In other examples, well-known structures and processes will not be elaborated in detail to avoid unnecessary details to obscure the description of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown, but is consistent with the widest scope consistent with the principles and features disclosed in the present application.
[0029] This embodiment provides Figure 1 A method for generating a perception sharing message based on perception accuracy for an Internet of Vehicles is shown, and specifically comprises the following steps:
[0030] S101. Detect target information data and pre-process it, and establish a local autonomous perception model to track the target information data of the current node;
[0031] Furthermore, the vehicle-mounted unit and road test unit of the current node are used to detect the data of other traffic participants around the road of the current node and the road abnormality information data of the current node through the radar sensing device and transmit and construct the target information data, including but not limited to vehicles, pedestrians, cyclists and other targets, and the target information data is denoised using Kalman filtering to remove interference information present in the collection process so as to improve the quality of the target information data. The radar sensing device is used to continuously detect the target and continuously track the real-time status of each target, and the road abnormality information of the current node is mapped according to the real-time status of each target. The data of other traffic participants around the road of the current node and the road abnormality information data of the current node are time-aligned to ensure that the target information data is in the same time period.
[0032] Furthermore, based on the Kalman filter, a local autonomous perception model is constructed, and a probability model is established to characterize the perception accuracy of the radar perception device. The perception accuracy depends on the distance between the detected target and the current node and the degree of occlusion of the detected target. Based on the visible cross section and the relative distance between the reference vehicle and the detected target, the measurement accuracy of the probability model is determined by adding noise to the true position of the target. The specific formula is:
[0033] D O (c,d)=(σ x ,σ y )(x,y)
[0034] Where c represents the cross section, d represents the distance, and D O () indicates that the value returned by the function is a tuple, which is used to detect the variance of the current node in the Cartesian coordinate system. x ,σ y ) represents the value of the true ground position at (x, y) as the average value in the normal distribution, and randomly selects a value in the normal distribution as the detection value of the radar perception device of the current node, wherein the normal distribution has the characteristic of no memory, so that the obtained detection point is highly discrete, so as to better perceive and track the state of the detection vehicle. The Kalman filter is applied to weaken the influence of the current detection on the tracking position, and the Kalman filter is used to adjust the detection value of the radar perception device and the previous vehicle trajectory by continuously applying the dynamic vehicle model, so as to weaken the influence of the current detection on the tracking position and provide a more accurate estimate of the system state. The historical information and the new detection target information data are integrated by adopting the prediction and correction mechanism. When the detection obtained at each time step k is updated, the detection operation updates the Kalman filter. The whole update process is divided into a prediction step and an update step. The prediction step process is to obtain a prediction estimate of the next node according to the state of the current node system, and the correction step process combines the detection value and the prediction estimate to obtain the optimal estimate of the system.
[0035] Furthermore, based on the Kalman filter, the error between the detection value of the radar sensing device and the vehicle trajectory is continuously adjusted by using the vehicle motion model. The state vector in the prediction step is predicted based on the previous state and the matrix used to describe the vehicle motion model. The specific formula is:
[0036] x k =x k-1 *F k-1
[0037] Among them, x k represents the state vector, x k-1 represents the previous state vector, F k-1 Represents the matrix describing the vehicle motion model, updates the error covariance matrix and adds noise. The specific formula is:
[0038]
[0039] Among them, P k represents the error covariance matrix, represents the transposed matrix of the vehicle motion model described previously, Q k-1 Represents noise. After completing the prediction step, the update step is automatically started. The detection value based on the radar perception device is corrected by calculating the Kalman gain, and the current state is updated based on the detection value and prediction value of the radar perception device of the current node. The covariance matrix of the process is updated again to provide a more stable and accurate position estimation.
[0040] S102. Establishing a networked collaborative perception model to track the objects received by the current node through SSM;
[0041] Furthermore, a networked collaborative perception model is established to track the list of target information data received by the current node through SSM. The networked collaborative perception model uses V2X communication to track target information data. Taking the vehicle node as an example, the OBU obtains the corresponding object list by receiving the SSM message and uses it as the detected target. Its detection accuracy depends on the distance and degree of occlusion between the detected target and the vehicle, just like the local autonomous perception model. That is, the same calculation method as the autonomous perception model is used and Kalman filtering is used to calculate the tracking accuracy. The network is through V2X, C-V2X, DSRC, 5G / 4G and other transmission methods.
[0042] S103. Each node determines whether to add the detection result to the new list of SSM objects to be sent by comparing the tracking accuracy of the local autonomous perception model with the tracking accuracy of the networked collaborative perception model;
[0043] Furthermore, each node determines whether to add the detection result to the list of target information data received by the new SSM to be sent by comparing the tracking accuracy of the local autonomous perception model with the tracking accuracy of the networked cooperative perception model. For the tracking list of the local autonomous perception model, since the Kalman filter needs to fuse historical information and new detection target information data through a prediction and correction mechanism, and the process covariance matrix P needs to be updated in a timely manner in both the prediction and correction parts, in order to ensure the tracking accuracy of the local autonomous perception model, it is necessary to check the covariance matrix P in a timely manner throughout the process to ensure its working correctness. The threshold is defined as 1. When the tracking of the local autonomous perception model is less than the threshold, the tracking is considered valuable and the next step is performed. Otherwise, the detection result is not considered to be included in the list of target information data received by the SSM to be sent next time. The next step calculates the relative entropy between the tracking accuracy of the local autonomous perception model and the tracking accuracy of the networked cooperative perception model through the Kullback-Leibler divergence. The specific calculation formula of the Kullback-Leibler divergence is:
[0044]
[0045] Among them, D KL (N LEM ||N V2X ) represents the Kullback-Leibler divergence of the probability distribution of the local autonomous perception model and the networked collaborative perception model, P LEM , P V2XThey respectively represent the tracking accuracy of the local autonomous perception model and the networked collaborative perception model, μ1 and μ2 respectively represent the mean state vector of the local autonomous perception model and the networked collaborative perception model, which are used to estimate the similarity of the probability distribution, and compare the calculation result with the set threshold γ, where the threshold γ is defined as 3. When the relative entropy is high enough and greater than the set threshold γ, it proves that the target information data contains rich information and is not highly repetitive, and is added to the list of target information data received by the next SSM to be sent. Otherwise, the detection result will not be considered to be included in the list of target information data received by the next SSM to be sent.
[0046] It should be noted that in the above embodiments, the description of each embodiment has its own emphasis, and for parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.
[0047] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0048] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded computer, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0049] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.
[0050] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.
[0051] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.
[0052] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.
Claims
1. A method for generating perception sharing messages based on perception accuracy for Internet of Vehicles, characterized in that: The specific steps include: S101. Detect target information data and pre-process it, and establish a local autonomous perception model to track the target information data of the current node; S102. Establishing a networked collaborative perception model to track the objects received by the current node through SSM; S103. Each node determines whether to add the detection result to the new list of SSM objects to be sent by comparing the tracking accuracy of the local autonomous perception model with the tracking accuracy of the networked collaborative perception model.
2. The method for generating perception sharing messages based on perception accuracy for Internet of Vehicles according to claim 1, characterized in that: In the S101, the specific steps of detecting the target information data and preprocessing are: using the vehicle-mounted unit and the road test unit of the current node to detect the data of other traffic participants around the road of the current node and the road abnormality information data of the current node through the radar sensing device and transmit and construct the target information data, using Kalman filtering to denoise the target information data, continuously detecting the target through the radar sensing device and continuously tracking the real-time status of each target, mapping the road abnormality information of the current node according to the real-time status of each target, and time-aligning the data of other traffic participants around the road of the current node and the road abnormality information data of the current node.
3. The method for generating perception sharing messages based on perception accuracy for Internet of Vehicles according to claim 1, characterized in that: In the S101, a local autonomous perception model is constructed based on the Kalman filter, a probability model is established to characterize the perception accuracy capability of the radar perception device, and based on the visible cross section and the relative distance between the reference vehicle and the detected target, the measurement accuracy of the probability model is determined by adding noise to the true position of the target, and a value is randomly selected from the normal distribution as the detection value of the radar perception device of the current node. The Kalman filter is used to adjust the detection value of the radar perception device and the previous vehicle trajectory by continuously applying the dynamic vehicle model, and the historical information and the new detection target information data are fused by adopting a prediction and correction mechanism. When the detection update is obtained at each time step k, the detection operation updates the Kalman filter.
4. The method for generating perception sharing messages based on perception accuracy for Internet of Vehicles according to claim 3, characterized in that: The entire Kalman filter update process is divided into a prediction step and an update step. In the prediction step, the state vector is predicted based on the previous state and the matrix used to describe the vehicle motion model, the error covariance matrix is updated and noise is added. After the prediction step is completed, the update step is automatically started. The detection value based on the radar perception device is corrected by calculating the Kalman gain, and the current state is updated based on the radar perception device detection value and predicted value of the current node, and the covariance matrix of the process is updated again.
5. The method for generating perception sharing messages based on perception accuracy for Internet of Vehicles according to claim 4, characterized in that: The specific prediction formula is: x k =x k-1 *F k-1 Among them, x k represents the state vector, x k-1 represents the previous state vector, F k-1 represents the matrix describing the vehicle motion model, and the specific formula for adding noise is: Among them, P k represents the error covariance matrix, represents the transposed matrix of the vehicle motion model described previously, Q k-1 Indicates noise.
6. The method for generating perception sharing messages based on perception accuracy for Internet of Vehicles according to claim 1, characterized in that: In S102, a networked collaborative perception model is established to track a list of target information data received by the current node through the SSM, and the networked collaborative perception model uses V2X communication to track the target information data.
7. The method for generating perception sharing messages based on perception accuracy for Internet of Vehicles according to claim 6, characterized in that: The networking is through V2X, C-V2X, DSRC, 5G / 4G and other transmission methods.
8. The method for generating perception sharing messages based on perception accuracy for Internet of Vehicles according to claim 1, characterized in that: In S103, each node determines whether to add the detection result to the list of target information data received by the new SSM to be sent by comparing the tracking accuracy of the local autonomous perception model with the tracking accuracy of the networked collaborative perception model, calculates the relative entropy between the tracking accuracy of the local autonomous perception model and the tracking accuracy of the networked collaborative perception model through Kullback-Leibler divergence, and compares the calculation result with the set threshold.
9. The method for generating perception sharing messages based on perception accuracy for Internet of Vehicles according to claim 8, characterized in that: The specific calculation formula of the Kullback-Leibler divergence is: Among them, D KL (N LEM ||N V2X ) represents the Kullback-Leibler divergence of the probability distribution of the local autonomous perception model and the networked collaborative perception model, P LEM , P V2X They represent the tracking accuracy of the local autonomous perception model and the networked collaborative perception model respectively, μ1 and μ2 represent the mean state vectors of the local autonomous perception model and the networked collaborative perception model respectively.