V2X message adaptive broadcasting method based on vehicle risk degree

Abstract

According to the V2X message adaptive broadcasting method based on the vehicle risk degree provided by the invention, the risk degree of possible collision between the to-be-tracked vehicle and the neighbor vehicle node in the area range is reflected by constructing the penalty factor, and the transmission interval of the basic safety information BSM is adjusted by calculating the penalty factor of the vehicle. The size of the transmission interval is adjusted according to the change of the condition of the penalty factor, so that the balance between the relieving of V2X channel congestion and the timeliness and accuracy of the message is achieved, and the vehicle in the Internet of Vehicles can adaptively adjust the broadcast period of the BSM message according to the specific service scene, so that the accuracy of the V2X safety message is ensured on the premise of reducing the V2X channel load, and the user experience is improved. And thus, the probability of traffic accidents is reduced.

Description

Technical Field

[0001] The present invention relates to the field of intelligent vehicle control technology, and in particular to a V2X message adaptive broadcasting method based on vehicle danger level. Background Art

[0002] The safety application of V2X communication in the Internet of Vehicles (IoV) is mainly based on the interaction of safety messages. By broadcasting basic safety messages (BSM), vehicles can continuously perceive the road conditions around them. The faster the frequency of sending basic safety messages (BSM), that is, the smaller the transmission interval, the more timely the road vehicle information is updated, and the more accurate the driver can obtain surrounding vehicle information. However, if the transmission interval is small, when the density of surrounding vehicles is large, too many data packets will be sent in the channel, exceeding the channel capacity and causing network congestion; on the contrary, the slower the information transmission frequency, that is, the larger the transmission interval, although the channel load is reduced, the real-time performance is worse. At present, most broadcast models for V2X basic safety messages BSM follow a fixed transmission cycle, usually a transmission cycle of 10HZ. The following are the problems that may be caused by a fixed transmission cycle: In an environment with a high density of road vehicles, if all vehicles send BSM data at a frequency of 10HZ, there will be a high probability of channel blocking, thus affecting the accuracy of data reception; moreover, if in an environment with a high density of vehicles, vehicles with a higher degree of danger also send BSM data at a frequency of 10HZ, then when the channel is blocked, it will seriously affect the driving safety of other vehicles. Summary of the invention

[0003] In order to solve the problem that in the existing vehicle-to-vehicle (V2X) communication technology, the broadcast models of safety messages BSM mostly follow a fixed sending cycle, which poses a traffic safety hazard in an environment with a large road vehicle density, the present invention provides a V2X message adaptive broadcast method based on vehicle danger level, which can adaptively adjust the broadcast cycle of BSM messages according to specific business scenarios, so as to ensure the accuracy of V2X safety messages while reducing the V2X channel load, thereby reducing the probability of traffic accidents.

[0004] The technical solution of the present invention is as follows: a V2X message adaptive broadcast method based on vehicle danger level, characterized in that it includes the following steps: S1: Construct vehicle trajectory prediction network based on neural network; The vehicle trajectory prediction network predicts the movement trajectory of the specified vehicle in the future based on the vehicle operation status data of the specified vehicle in the past time; The vehicle running status data includes: position information, speed information, acceleration information and steering information; S2: construct penalty factor; The penalty factor is used to mark the degree of overlap between two predicted motion trajectories; the smaller the distance between the two paths, the greater the degree of overlap between the paths, and the larger the corresponding penalty factor; ; Where P collision (t) represents the penalty factor corresponding to the specified vehicle at time t; C is a constant, ∈ is the preset collision safety distance threshold; D(t) represents the distance between the predicted motion trajectories of the two vehicles at time t; S3: dividing the value range of the penalty factor according to the preset penalty factor segmentation value to obtain different factor determination ranges, and specifying a corresponding security message sending interval for each factor determination range; S4: record all vehicles in the vehicle network as vehicles to be tracked; Each of the vehicles to be tracked reports the vehicle operation status data of the vehicle in real time; S5: confirming all neighboring vehicles of each vehicle to be tracked; The vehicle running status data of the vehicle to be tracked and its corresponding neighboring vehicle are simultaneously sent to the vehicle trajectory prediction network to perform vehicle running trajectory prediction; The predicted motion trajectory of the vehicle to be tracked is recorded as P1(t), and the predicted motion trajectory of the neighboring vehicle is recorded as: P2(t); S6: Calculate the distance D(t) between the predicted motion trajectory P1(t) of the vehicle to be tracked and the predicted motion trajectory P2(t) of the neighboring vehicle at time t; S7: For each vehicle to be tracked, calculating a corresponding penalty factor; S8: Determine the factor determination range to which the penalty factor of the vehicle to be tracked belongs, recorded as: adjustment determination range; The safety message sending interval of the vehicle to be tracked is adjusted to the safety message sending interval corresponding to the adjustment determination range.

[0005] It is further characterized by: It also includes the following steps: Execute steps S4 to S8 in a loop to adaptively adjust the safety message sending interval of V2X communication of all vehicles in the vehicle network in real time; The vehicle trajectory prediction network is constructed based on the LSTM long short-term memory neural network; In step S2, the distance D(t) between the predicted motion trajectories of the two vehicles is calculated as follows: ; Where D(t) represents the distance between the predicted motion trajectories P1(t) and P2(t) of two adjacent vehicles at time t; The expression method of the predicted motion trajectory is: Assuming that the path of each vehicle's running trajectory is a time series consisting of discrete position points, then: P1(t)=(x1(t),y1(t)); P2(t)=(x2(t),y2(t)); In the formula, x(t) and y(t) represent the horizontal and vertical coordinates of the specified vehicle at time t; In step S3, the penalty factor segmentation values ​​include: 15 and 18; When the penalty factor is <15, the corresponding security message sending interval is 100ms; When 15≤penalty factor<18, the corresponding security message sending interval is 50ms; When the penalty factor is ≥ 18, the corresponding security message sending interval is 20ms.

[0006] The present application provides a V2X message adaptive broadcast method based on vehicle danger level, which constructs a penalty factor to reflect the degree of danger of a possible collision between a vehicle to be tracked and a neighboring vehicle node within the area, calculates the penalty factor of the vehicle and adjusts the transmission interval of its basic safety information BSM, and adjusts the size of the transmission interval according to the penalty factor, so as to achieve a balance between alleviating V2X channel congestion and message timeliness and accuracy, and ensure that vehicles in the Internet of Vehicles can adaptively adjust the broadcast period of BSM messages according to specific business scenarios, so as to ensure the accuracy of V2X safety messages while reducing the V2X channel load, thereby reducing the probability of traffic accidents. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] Figure 1 The figure is a flow chart of a V2X message adaptive broadcasting method based on vehicle danger level. DETAILED DESCRIPTION

[0008] like Figure 1 As shown, the present application includes a V2X message adaptive broadcasting method based on vehicle danger level, which includes the following steps.

[0009] S1: Construct vehicle trajectory prediction network based on neural network; The vehicle trajectory prediction network predicts the movement trajectory of a specified vehicle in the future based on the vehicle operation status data of the specified vehicle in the past time; the vehicle operation status data includes: position information, speed information, acceleration information and steering information.

[0010] The vehicle trajectory prediction network can be constructed based on the neural network model in the prior art, such as: RNN, GUN, CNN, LSTM, etc. In this embodiment, the vehicle trajectory prediction network is constructed based on the LSTM long short-term memory neural network. In the specific implementation, it is necessary to first build an LSTM neural network, and then train the neural network model based on the real-time collected vehicle status data to obtain a trained network model; finally, the model is deployed on the vehicle computing terminal, and the recorded vehicle historical driving status data is input to predict the vehicle's movement trajectory for a period of time in the future.

[0011] S2: Construct penalty factor.

[0012] This method reflects the risk of a possible collision between the vehicle to be tracked and its neighboring vehicle nodes by constructing a penalty factor within the area where the vehicle to be tracked is located. In order to calculate the risk of a possible collision between the vehicle to be tracked and its neighboring vehicle nodes, it is first necessary to predict the trajectories of the vehicle to be tracked and its neighboring vehicle nodes in the future, calculate the overlap between the two trajectories (or multiple trajectories), and then define the vehicle penalty factor HVF.

[0013] The calculation of the penalty factor depends on the overlap between the predicted trajectory of the vehicle to be tracked and the trajectory of the neighboring vehicle node. The higher the overlap between the two predicted paths (i.e., the smaller the distance between the paths), the larger the penalty factor should be. In order to quantify the overlap of the predicted paths, a distance-based penalty factor is designed. In specific implementation, the predicted trajectory of the vehicle is predicted based on the pre-constructed vehicle trajectory prediction network.

[0014] In this method, the predicted trajectory P1(t) of the vehicle to be tracked and the predicted trajectory P2(t) of the neighboring vehicle node can be represented by the position points in the time series. Then: The representation method of the predicted motion trajectory is: Assuming that the path of each vehicle's running trajectory is a time series consisting of discrete position points, then: P1(t)=(x1(t),y1(t)); P2(t)=(x2(t),y2(t)); Where x(t) and y(t) represent the horizontal and vertical coordinates of the specified vehicle at time t.

[0015] If the distance between the two paths is less than a set safety threshold, it can be considered that a collision has occurred. For continuous time points, collisions can be detected by calculating the Euclidean distance between the two paths at each time step. Then: The distance D(t) between the two predicted trajectories is calculated as: ; Where D(t) represents the distance between P1(t) and P2(t) at time t.

[0016] The penalty factor is used to mark the degree of overlap between two predicted paths; the smaller the distance between the two paths, the greater the degree of overlap between the paths, and the larger the corresponding penalty factor; ; Where P collision (t) represents the penalty factor corresponding to the specified vehicle at time t; C is a constant, ∈ is the preset collision safety distance threshold; D(t) represents the distance between the two predicted paths at time t.

[0017] When D(t) is close to or less than ∈, it indicates that there is a collision risk, and the penalty factor P collision (t) will increase rapidly, then the V2X message sending time interval can be reduced for the vehicle to be tracked in a dangerous scenario. In this embodiment, the collision safety distance threshold ∈ is 30m. In actual application, the value can be dynamically adjusted according to the vehicle speed range. For example, when the average vehicle speed is 50KM / h, ∈ is 50m.

[0018] S3: In this method, different security message BSM transmission levels are divided according to the penalty factor. The specific method is: According to the preset penalty factor segmentation value, the value range of the penalty factor is divided to obtain different factor determination ranges, and a corresponding security message sending interval is specified for each factor determination range.

[0019] The penalty factor segmentation value can be specifically set according to the actual vehicle networking environment. In this embodiment, the penalty factor segmentation values ​​include: 15 and 18. At present, the V2X related standards require that the minimum service transmission interval be 20ms. The default interval for sending BSM messages is 100ms, that is, the sending frequency is 10HZ. In order to solve the problem of insufficient channel resources, the transmission interval range can be considered to be 20 milliseconds to 1000 milliseconds. Since the minimum message transmission interval required for V2X data interaction is 20ms, the message transmission intervals of the vehicles are set to 100ms, 50ms and 20ms respectively.

[0020] In this embodiment, based on the two segmentation values ​​of 15 and 18, three different specified factor determination ranges are obtained, as follows: When the penalty factor is <15, the corresponding security message sending interval is 100ms; When 15≤penalty factor<18, the corresponding security message sending interval is 50ms; When the penalty factor is ≥ 18, the corresponding security message sending interval is 20ms.

[0021] In this method, the V2X message adaptive broadcast model based on vehicle trajectory prediction introduces a penalty factor (HVF) of the vehicle to be tracked, and different BSM transmission levels are divided accordingly. When the penalty factor is high, the BSM message sending frequency of the vehicle to be tracked is increased, that is, the BSM transmission interval of the vehicle is shortened; when the penalty factor is low, the BSM message sending frequency of the vehicle is reduced, that is, the BSM transmission interval is extended.

[0022] This method adjusts the transmission interval of its basic safety information BSM by calculating the penalty factor of the vehicle, and adjusts the size of the transmission interval according to the change of the penalty factor, so as to achieve a balance between alleviating V2X channel congestion and message timeliness and accuracy. According to the range of the penalty factor set in this method, the safety message sending interval of each vehicle is adaptively adjusted. In an environment with a high density of road vehicles, if ordinary vehicles (vehicles with low risk) reduce the frequency of sending BSM messages at this time, the probability of channel blocking will be greatly reduced, thereby improving the accuracy of received data and thus improving driving safety.

[0023] S4: record all vehicles in the vehicle network as vehicles to be tracked; Each vehicle to be tracked reports its own vehicle operation status data in real time.

[0024] S5: confirm all neighboring vehicles of each vehicle to be tracked; The vehicle running status data of the vehicle to be tracked and its corresponding neighbor vehicles are simultaneously sent to the vehicle trajectory prediction network to predict the vehicle running trajectory; The predicted motion trajectory of the vehicle to be tracked is recorded as P1(t), and the predicted motion trajectory of the neighboring vehicle is recorded as: P2(t).

[0025] Neighbor vehicles in the Internet of Vehicles refer to other vehicles that are spatially close to the current vehicle in the Internet of Vehicles environment. These vehicles exchange information through wireless communication technology and together form an interactive wireless network. The definition of neighbor vehicles is not only based on geographical proximity, but also involves vehicles within the communication range to ensure that information can be effectively transmitted and shared. Internet of Vehicles technology enables communication between vehicles (V2V), vehicles and infrastructure (V2I), vehicles and pedestrians (V2P), etc., thereby providing auxiliary information for vehicle driving and intelligent traffic management. The identification and positioning of neighbor vehicles is an important part of the Internet of Vehicles technology. The identification of neighbor vehicles in this application is implemented based on the existing Internet of Vehicles technology.

[0026] S6: Calculate the distance D(t) between the predicted motion trajectory P1(t) of the vehicle to be tracked at time t and the predicted motion trajectory P2(t) of the neighboring vehicle.

[0027] S7: For each vehicle to be tracked, calculate its corresponding penalty factor.

[0028] S8: Determine the factor determination range to which the penalty factor of the vehicle to be tracked belongs, recorded as: adjustment determination range; The data transmission interval of the vehicle to be tracked is adjusted to the safety message transmission interval corresponding to the adjustment determination range. That is, when the penalty factor HVF is less than the penalty factor threshold of 15, the vehicle's safety message BSM transmission interval is 100ms; when the penalty factor HVF is greater than or equal to the penalty factor threshold of 15 and less than 18, the safety message transmission interval is adjusted to 50ms; when the penalty factor HVF is greater than or equal to the penalty factor threshold of 18, its safety message transmission interval is 20ms.

[0029] Specifically, when the method is used in the Internet of Vehicles, steps S4 to S8 are executed in a loop to adaptively adjust the V2X communication of vehicles in the Internet of Vehicles in real time to ensure the safe operation of all vehicles.

[0030] After using the technical solution of the present invention, a penalty factor is used to describe the degree of danger of the road scene where the vehicle to be tracked is located at this time. According to the penalty privacy, the transmission period of the V2X basic safety message BSM is adaptively adjusted according to the vehicle danger degree obtained by the vehicle trajectory prediction. That is, when the vehicle danger degree is high, the frequency of sending BSM messages is increased; when the vehicle danger degree is low, the frequency of sending BSM messages is reduced, thereby achieving the improvement of the accuracy of receiving V2X key BSM data while reducing the probability of channel blocking.

Claims

1. A V2X message adaptive broadcasting method based on vehicle danger level, characterized in that: It includes the following steps: S1: Construct vehicle trajectory prediction network based on neural network; The vehicle trajectory prediction network predicts the movement trajectory of the specified vehicle in the future based on the vehicle operation status data of the specified vehicle in the past time; The vehicle running status data includes: position information, speed information, acceleration information and steering information; S2: construct penalty factor; The penalty factor is used to mark the degree of overlap between two predicted motion trajectories; the smaller the distance between the two paths, the greater the degree of overlap between the paths, and the larger the corresponding penalty factor; ； Where P collision (t) represents the penalty factor corresponding to the specified vehicle at time t; C is a constant, ∈ is the preset collision safety distance threshold; D(t) represents the distance between the predicted motion trajectories of the two vehicles at time t; S3: dividing the value range of the penalty factor according to the preset penalty factor segmentation value to obtain different factor determination ranges, and specifying a corresponding security message sending interval for each factor determination range; S4: record all vehicles in the vehicle network as vehicles to be tracked; Each of the vehicles to be tracked reports the vehicle operation status data of the vehicle in real time; S5: confirming all neighboring vehicles of each vehicle to be tracked; The vehicle running status data of the vehicle to be tracked and its corresponding neighboring vehicle are simultaneously sent to the vehicle trajectory prediction network to perform vehicle running trajectory prediction; The predicted motion trajectory of the vehicle to be tracked is recorded as P1(t), and the predicted motion trajectory of the neighboring vehicle is recorded as: P2(t); S6: Calculate the distance D(t) between the predicted motion trajectory P1(t) of the vehicle to be tracked and the predicted motion trajectory P2(t) of the neighboring vehicle at time t; S7: For each vehicle to be tracked, calculating a corresponding penalty factor; S8: Determine the factor determination range to which the penalty factor of the vehicle to be tracked belongs, recorded as: adjustment determination range; The safety message sending interval of the vehicle to be tracked is adjusted to the safety message sending interval corresponding to the adjustment determination range.

2. According to claim 1, a V2X message adaptive broadcast method based on vehicle risk level is characterized by: It also includes the following steps: Steps S4 to S8 are executed in a loop to adaptively adjust the safety message sending interval of the V2X communication of all vehicles in the vehicle network in real time.

3. According to claim 1, a V2X message adaptive broadcast method based on vehicle risk level is characterized by: The vehicle trajectory prediction network is constructed based on the LSTM long short-term memory neural network.

4. According to claim 1, a V2X message adaptive broadcast method based on vehicle risk level is characterized by: In step S2, the distance D(t) between the predicted motion trajectories of the two vehicles is calculated as follows: ； Where D(t) represents the distance between the predicted motion trajectories P1(t) and P2(t) of two adjacent vehicles at time t; The expression method of the predicted motion trajectory is: Assuming that the path of each vehicle's running trajectory is a time series consisting of discrete position points, then: P1(t)=(x1(t),y1(t)); P2(t)=(x2(t),y2(t)); Where x(t) and y(t) represent the horizontal and vertical coordinates of the specified vehicle at time t.

5. The method for adaptively broadcasting V2X messages based on vehicle risk level according to claim 1, characterized in that: In step S3, the penalty factor segmentation values ​​include: 15 and 18; When the penalty factor is <15, the corresponding security message sending interval is 100ms; When 15≤penalty factor<18, the corresponding security message sending interval is 50ms; When the penalty factor is ≥ 18, the corresponding security message sending interval is 20ms.

Images