Vehicle active safety control method and device based on constraint degree in information perception scenario

Abstract

The present invention discloses a vehicle active safety control method and device based on constraint degree in an information perception scenario, and the method steps include: S01. Establishing a coordinate system for different types of road nodes and performing regional subdivision to establish a road node database; S02. Perceiving and capturing road, vehicle and traffic flow information and storing it; S03. Calculating the constraint degree and training to form a constraint degree prediction model; S04. When the detected vehicle enters a preset road node, triggering a safety warning and obtaining the ID number information of the current vehicle; S05. Obtaining the real-time monitoring data of the current vehicle and the surrounding environment information; S06. Calculating the constraint degree level of the current vehicle and predicting the vehicle risk state; S07. Controlling the acquisition of the corresponding active vehicle control strategy release according to the currently predicted constraint degree level; S08. Updating the database and the vehicle safety warning model. The present invention has the advantages of simple implementation method, high control accuracy and efficiency, safety and reliability, etc.

Description

Technical Field

[0001] The present invention relates to the field of vehicle safety control technology, and in particular to a constraint-based vehicle active safety control method and device in an information perception scenario. Background Art

[0002] Complex road nodes are prone to traffic accidents. The main reason for the frequent accidents is that the complex road environment leads to mixed driving behaviors. Vehicles usually change lanes and turn frequently at complex road nodes. If active vehicle safety control can be achieved, the risk can be predicted when the vehicle has an accident risk, and active vehicle control can be carried out, which can effectively reduce the possibility of accidents and improve vehicle driving safety. To achieve vehicle safety control, the key is to predict the risk of vehicle accidents. The risk of vehicle accidents can be predicted based on driving behavior. The methods for predicting the risk of vehicle accidents based on driving behavior in the prior art are mainly implemented in the following ways:

[0003] 1. A method of using historical data extracted from road nodes to establish an accident risk prediction model. For example, patent application CN112949999 discloses a high-speed traffic accident risk warning method, which extracts minute-level traffic flow and average speed from fixed nodes on the highway to obtain a regional vehicle collision index; uses Gaussian mixture model GMM for clustering to divide the high-speed accident risk level; uses Bayesian deep learning to extract features, predicts the minute-level traffic flow and average speed of future target nodes, calculates the regional vehicle collision index, and provides safety warnings based on the risk level. However, this type of method can only represent the historical state by using historical data modeling. There will be certain errors in using this historical state to represent the state of future nodes. Therefore, this type of method cannot accurately predict the real-time traffic accident risk of vehicles, and only considers influencing factors such as traffic flow and average speed, making it difficult to accurately predict the driving risk of a single vehicle.

[0004] 2. A method of using historical data to establish a real-time accident risk prediction model for conventional road environments. For example, patent application CN104732075 discloses a real-time prediction method for urban road traffic accident risks. This method extracts the geometric linear data of the research section, the historical traffic flow basic data n minutes before the traffic accident, and the historical weather condition data to establish a real-time prediction model for urban road traffic accidents based on Poisson distribution. After calibrating the model using historical data, it is necessary to calculate the level and distribution probability of the real-time traffic flow characteristics and weather condition data after converting them into categorical variables to obtain traffic accident analysis and prediction results. This type of method is only applicable to risk prediction of conventional roads, and is difficult to apply to traffic accident risk prediction of complex nodes. The model only considers traffic flow characteristics and weather factors, and it is difficult to accurately reflect the potential risks of vehicles. At the same time, it is also necessary to perform a large amount of preprocessing on the real-time acquired data to obtain the prediction results, and the system computing power requirements are also high, resulting in low prediction efficiency.

[0005] In summary, the vehicle risk prediction methods in the existing technology are usually applicable to conventional road environments, but not to accident risk prediction at complex nodes. They usually take into account conventional traffic flow characteristics, weather and other influencing factors, and are only applicable to accident risk prediction under simple terrain and general driving behavior. They cannot accurately predict the driving risk of a single vehicle. In addition, the method of using historical data to establish a model will also have the problem of poor real-time performance, and it is impossible to fully utilize real-time vehicle collaborative data to predict vehicle accident risks in real time, making it difficult to quickly and accurately achieve vehicle safety management and control. Summary of the invention

[0006] The technical problem to be solved by the present invention is: in response to the technical problems existing in the prior art, the present invention provides a vehicle active safety control method, control method and device based on constraint degree in an information perception scenario with simple implementation method, high prediction accuracy and efficiency, and safety and reliability.

[0007] In order to solve the above technical problems, the technical solution proposed by the present invention is:

[0008] A vehicle active safety control method based on constraint degree in an information perception scenario, comprising the following steps:

[0009] S01. Establish a coordinate system for different types of road nodes, and insert multiple sub-nodes within the monitoring range corresponding to each road node to perform regional subdivision and establish a road node database;

[0010] S02. Based on the roadside information sensing device, the road, vehicle and traffic flow information is sensed and captured and stored;

[0011] S03. Selecting historical sample data from the vehicle movement information database according to a preset time period, calculating the degree of constraint of the vehicle at each road node based on the selected historical sample data, and training a constraint prediction model for each road node;

[0012] S04. When the detected vehicle enters the specified range around the preset road node, a safety warning is triggered and the ID number information of the current vehicle is obtained;

[0013] S05. Obtaining real-time monitoring information of the current vehicle and the traffic flow where the current vehicle is located according to the ID number information of the current vehicle, and obtaining real-time movement and surrounding environment information of each vehicle within a specified range around the current road node;

[0014] S06. Calculate the current vehicle's degree of constraint based on the constraint prediction model and the information obtained in step S05, predict the vehicle risk status based on the calculated degree of constraint and issue warning information;

[0015] S07. The control obtains the corresponding pre-configured active vehicle control strategy according to the currently predicted constraint level and publishes it;

[0016] S08. Update the vehicle movement information database and the vehicle safety warning model according to the preset time period.

[0017] Furthermore, the step S01 includes:

[0018] S101. Establishing a coordinate system: establishing a coordinate system for each road node, wherein the origin is determined according to the endpoints along the driving direction of the boundary points of each node area and the x-axis and y-axis are determined according to the driving direction, and a road plane coordinate system for each node is established;

[0019] S102. Area subdivision: Determine the corresponding monitoring range according to the characteristics of each road node, and subdivide the area within the corresponding monitoring range on the x-axis of each node into multiple secondary monitoring areas, wherein the monitoring range includes a vehicle interweaving section and a normal environment section connected to the vehicle interweaving section.

[0020] Furthermore, the constraint degree is calculated according to the following formula:

[0021]

[0022] Where i represents the area, Si is the degree of constraint imposed on the vehicle in the i-th secondary area compared to the first secondary area, the first secondary area is a normal road section, α1 is the weight of the lane change index, and α2 is the weight of the angle index; L i is the lane-changing frequency of vehicles in the i-th area, C iis the average driving angle of the vehicle in the i-th area, and C0 is the minimum angle value of the vehicle in all areas.

[0023] Furthermore, in step S03, characteristic parameters are extracted from historical sample data to calculate the constraint degree of the vehicle at each road node, and a decision tree model is used to perform model training on the constraint degree calculated for each road node, wherein the calculated constraint degree is used as the predicted value, and the real-time vehicle movement information, surrounding traffic flow status and road environment information extracted from the historical sample data are used as independent variables to construct the constraint degree prediction model for each road node.

[0024] Furthermore, the characteristic parameters include any of the following: the vehicle's charging type, the vehicle's real-time position, the vehicle's real-time speed, the vehicle's real-time lane changing situation, the distance between the vehicle's current position and the node end point, the vehicle's current lane, the real-time traffic flow around the vehicle, the degree of real-time traffic flow congestion around the vehicle, the type of road node where the vehicle is located, the width or curvature radius of the road where the vehicle is located, the type of road markings, the vehicle's lane changing frequency in each secondary area, and the vehicle's average driving angle in each secondary area.

[0025] Furthermore, the steps of step S07 include:

[0026] Divide the constraint degree into multiple levels in advance, and configure corresponding vehicle control strategies for different levels of constraint degree;

[0027] Control the vehicle according to the vehicle control strategy corresponding to the currently predicted constraint level;

[0028] If the vehicle is within a specified range around a preset road node, return to step S05 to continuously monitor the current vehicle position and motion state, calculate the constraint degree to predict the vehicle risk state, and control the vehicle according to the vehicle management and control strategy corresponding to the currently predicted constraint degree level until the vehicle leaves the current road node.

[0029] Furthermore, dividing the constraint degree into multiple levels includes:

[0030] The potential risk of vehicle driving is determined according to the level of constraint, where the constraint degree S of each road node i is i ∈[a,b],S i = a, the constraint degree is the smallest and the corresponding vehicle risk is the highest. i = b, the degree of constraint is the largest and the corresponding vehicle risk is the lowest;

[0031] According to the constraint degree calculation results of the historical sample data, the constraint degree distribution value range of the sample data is determined and multiple constraint degree classification thresholds are determined to divide the constraint degree of the vehicle in the secondary area into multiple levels. The secondary area is the area obtained by subdividing the area within the corresponding monitoring range determined by each road node.

[0032] A vehicle active safety control device based on constraint degree in an information perception scenario, comprising:

[0033] A database establishment module is used to establish a coordinate system for different types of road nodes, and insert multiple sub-nodes within the monitoring range corresponding to each road node to perform regional subdivision and establish a road node database;

[0034] Roadside information sensing equipment, used to sense, capture and store road, vehicle and traffic flow information;

[0035] The constraint degree calculation and model training module is used to select historical sample data from the vehicle motion information database according to a preset time period, calculate the constraint degree of the vehicle at each road node based on the selected historical sample data, and train to form a constraint degree prediction model for each road node;

[0036] The warning trigger module is used to trigger a safety warning and obtain the ID number information of the current vehicle when the detected vehicle enters a specified range around a preset road node;

[0037] The information acquisition module is used to obtain the real-time monitoring information of the current vehicle and the traffic flow where the current vehicle is located according to the ID number information of the current vehicle, and to obtain the real-time movement and surrounding environment information of each vehicle within a specified range around the current road node;

[0038] A risk prediction module, used to calculate the current vehicle's restraint level based on the restraint prediction model and the information obtained in the information acquisition module, predict the vehicle's risk status based on the restraint level and issue warning information;

[0039] Active safety control module, used to control the acquisition of the corresponding pre-configured active vehicle control strategy according to the currently predicted constraint level and publish it;

[0040] The updating module is used to update the vehicle motion information database and the vehicle safety warning model according to a preset time period.

[0041] A computer device includes a processor and a memory, wherein the memory is used to store a computer program, and the processor is used to execute the computer program to execute the above-mentioned vehicle risk prediction method, or execute the above-mentioned management and control method.

[0042] A computer-readable storage medium storing a computer program, wherein the computer program implements the above method when executed.

[0043] Compared with the prior art, the advantages of the present invention are:

[0044] 1. The present invention introduces the concept of constraint degree in mechanical principles for the special environment of complex road nodes. The road nodes are divided to establish a road node database, and then the historical sample data of the vehicle is used to calculate the constraint degree of the vehicle at each road node, and a constraint degree prediction model for each road node is trained. When the vehicle enters a pre-specified complex road node, the real-time information of the vehicle and the communication and the surrounding environment information are obtained, and then the real-time information and the constraint degree prediction model are used to predict the constraint degree of the vehicle in real time, and the potential risk of the vehicle is predicted. This can solve the problem of insufficient consideration of the environmental constraints of complex nodes in traditional risk prediction methods, so that it can be applied to complex road environments, realize effective prediction of vehicle safety conditions for different nodes, and realize real-time and accurate prediction of the risk of traffic accidents of vehicles at complex nodes, effectively ensure vehicle driving safety, and avoid the occurrence of traffic accidents at complex nodes.

[0045] 2. The present invention establishes a correlation between the degree of constraint of driving behavior and the safety of vehicles at complex road nodes. It first uses the degree of constraint of driving behavior to predict the risk of vehicle accidents, and uses the prediction result to determine the degree of constraint of the vehicle. Further, it obtains corresponding active safety decisions based on the predicted degree of constraint for feedback. It can flexibly obtain appropriate active safety management and control strategies for vehicles under different risks, and effectively ensure the driving safety of vehicles in various environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 It is a schematic diagram of the implementation flow of the constraint-based vehicle risk prediction method in the information perception scenario of this embodiment.

[0047] Figure 2 It is a schematic diagram of the effect of establishing a coordinate system for complex road nodes in a specific application embodiment.

[0048] Figure 3 It is a schematic diagram of the effect of regional subdivision in a specific application embodiment of the present invention.

[0049] Figure 4 It is a schematic diagram of the structural principle of a vehicle control device in a specific application embodiment of the present invention. DETAILED DESCRIPTION

[0050] The present invention is further described below in conjunction with the accompanying drawings and specific preferred embodiments, but the protection scope of the present invention is not limited thereby.

[0051] At complex road nodes, the complex road environment leads to mixed driving behaviors. Vehicles usually change lanes and turn frequently at complex road nodes. The characteristic of this in terms of constraint is that the vehicle movement will show a characteristic of reduced constraint, that is, the driving behavior constraint, which can accurately reflect the potential risk of the vehicle to a certain extent. The driving behavior constraint index can fundamentally characterize the impact of the complex environment on the driver, and then effectively reflect the complexity of the environment. The special road environment at complex road nodes has a much different constraint effect on the driver's driving behavior than the conventional road environment. Therefore, the use of driving behavior constraint can effectively realize the prediction of the potential risk of vehicles at complex road nodes.

[0052] The present invention utilizes the above characteristics and introduces the concept of constraint degree in mechanical principles for the special environment of complex road nodes. The road nodes are first divided to establish a road node database, and then the historical sample data of the vehicle is used to calculate the constraint degree (driving behavior constraint degree) of the vehicle at each road node, and a constraint degree prediction model for each road node is trained. When the vehicle enters a pre-specified complex road node, the real-time information of the vehicle and communication and the surrounding environment information are obtained, and then the real-time information and the constraint degree prediction model are used to predict the constraint degree of the vehicle in real time. Based on the constraint degree, the potential risk of the vehicle can be predicted, which can solve the problem of insufficient consideration of complex node environmental constraints in traditional risk prediction methods, so that it can be applied to complex road environments, realize effective prediction of vehicle safety conditions for different nodes, and realize real-time and accurate prediction of the risk of traffic accidents of vehicles at complex nodes, effectively ensure vehicle driving safety, and avoid the occurrence of traffic accidents at complex nodes. At the same time, by establishing the correlation between driving behavior constraints and vehicle safety at complex road nodes, the driving behavior constraints are first used to predict the risk of vehicle accidents, and the vehicle constraint level is determined using the prediction results. Further, the corresponding active safety decision feedback is obtained based on the predicted constraint level. The appropriate vehicle active safety management and control strategies can be flexibly obtained under different risks, effectively ensuring the driving safety of the vehicle in various environments.

[0053] like Figure 1 As shown, the specific steps of the vehicle risk prediction method based on constraint degree in the information perception scenario of this embodiment include:

[0054] S01. Establish a coordinate system for different types of road nodes, and insert multiple sub-nodes within the monitoring range corresponding to each road node to perform regional subdivision and establish a road node database;

[0055] S02. Based on the roadside information sensing device, the road, vehicle and traffic flow information is sensed and captured and stored;

[0056] S03. Selecting historical sample data from the vehicle movement information database according to a preset time period, calculating the degree of constraint of the vehicle at each road node based on the selected historical sample data, and training a constraint prediction model for each road node;

[0057] S04. When the detected vehicle enters the specified range around the preset road node, a safety warning is triggered and the ID number information of the current vehicle is obtained;

[0058] S05. Obtaining real-time monitoring information of the current vehicle and the traffic flow where the current vehicle is located according to the ID number information of the current vehicle, and obtaining real-time movement and surrounding environment information of each vehicle within a specified range around the current road node;

[0059] S06. Calculate the current vehicle's constraint degree based on the constraint degree prediction model and the information obtained in step S05, predict the vehicle risk state based on the calculated constraint degree and issue warning information;

[0060] S07. The control obtains the corresponding pre-configured active vehicle control strategy according to the currently predicted constraint level and publishes it;

[0061] S08. Update the vehicle movement information database and the vehicle safety warning model according to the preset time period.

[0062] In this embodiment, step S01 establishes a coordinate system for different types of complex road nodes and performs regional segmentation to establish a complete complex road node database. The specific steps include:

[0063] S101. Establishing a coordinate system: establishing a coordinate system for each road node, wherein the origin is determined according to the endpoints along the driving direction of the boundary points of each node area and the x-axis and y-axis are determined according to the driving direction, and a road plane coordinate system for each node is established;

[0064] S102. Area subdivision: Determine the corresponding monitoring range according to the characteristics of each road node, and subdivide the area within the corresponding monitoring range on the x-axis of each node into multiple secondary monitoring areas. The monitoring range includes a vehicle interweaving section and a normal environment section connected to the vehicle interweaving section.

[0065] In the above step S101, the south end point of the node area boundary point along the driving direction can be taken as the origin, the driving direction is taken as the x-axis, and the driving direction section direction is taken as the y-axis to establish a complete road plane coordinate system on each road node. The specific construction method of the coordinate system can of course be configured according to actual needs.

[0066] In the above step S102, the complete monitoring range of each road node can be determined specifically according to the characteristics of the complex road node, and the monitoring range includes two parts: 1. Road sections where vehicles are prone to interweaving (vehicle interweaving sections); 2. A section of normal environment connected to area 1 (vehicle interweaving sections) in the upstream direction of driving (conventional sections where vehicles are not prone to interweaving). Multiple points are uniformly inserted in the complex road node area within the corresponding monitoring range on the x-coordinate axis of each road node, so as to subdivide the complete monitoring range of the node into multiple secondary monitoring areas of uniform length.

[0067] In a specific application embodiment, assuming that the established complex road node database is recorded as a complex road node set N = {n1, n2, ...}, the specific steps of establishing the coordinate system of node n1 and subdividing the area are:

[0068] (1) Establishing a coordinate system: Establish a coordinate system for the complex road node n1, taking the southern end point along the driving direction (the most southwestern end point when the driving direction is west) of the boundary point of the node n1 region as the origin, the driving direction as the x-axis, and the driving direction cross-section direction as the y-axis, to establish a complete road plane coordinate system on the node n1. The effect of establishing a coordinate system for a complex road node in a specific application embodiment is as follows: Figure 2 shown.

[0069] (2) Regional subdivision: First, determine the complete monitoring range of node n1 according to the characteristics of the complex road node. Evenly insert i-1 points in the complex road node area within the corresponding monitoring range on the x-coordinate axis of node n1, and subdivide the complete monitoring range of node n1 into i secondary monitoring areas of uniform length. The secondary areas are numbered in descending order according to the x-coordinate as d1, d2, ...d i , let the secondary region set be D = {d1, d2…d i}, where d1 is the regular main road section corresponding to the vehicle interweaving section, then the complex road node with a length of L is divided into i nodes with a length of The secondary region of . Let O be the origin of the coordinate system, and its coordinates are (0,0). Then the left boundary of the nth region along the x-axis is The right boundary is

[0070] In a specific application embodiment, Figure 3 This is a schematic diagram of the initial lanes and toll channels in the toll station diversion area divided in the above manner. The numbers in the figure represent the serial numbers of the divided areas.

[0071] In this embodiment, the constraint degree is calculated according to the following formula:

[0072]

[0073] Where i represents the area, Si is the degree of constraint imposed on the vehicle in the i-th secondary area compared to the first secondary area (i.e., normal environment section), α1 is the weight of the lane change index, and α2 is the weight of the angle index; L i is the lane-changing frequency of vehicles in the i-th area, C i is the average driving angle of the vehicle in the i-th area, and C0 is the minimum angle value of the vehicle in all areas.

[0074] After establishing the road node coordinate system and obtaining the results of the regional segmentation, the historical sample data is calculated and processed based on the coordinate system and the results. Specifically, computer video recognition technology can be used to extract vehicle micro-trajectory information to further extract effective feature parameters for constraint calculation. After the road node database is established, in step S02 of this embodiment, the road, vehicle and traffic flow information are captured and stored in real time based on the roadside information perception equipment to obtain a road, traffic flow and vehicle motion information database. In step S03 of this embodiment, the constraint degree of the vehicle at each road node is calculated by extracting feature parameters from the historical sample data, and the constraint degree calculated for each road node is trained using a decision tree model, wherein the calculated constraint degree is used as the predicted value, and the real-time vehicle motion information, surrounding traffic flow status and road environment information extracted from the historical sample data are used as independent variables to construct a constraint degree prediction model for each road node.

[0075] The above-mentioned characteristic parameters specifically include the vehicle's charging type, the vehicle's real-time position, the vehicle's real-time speed, the vehicle's real-time lane changing situation, the distance between the vehicle's current position and the node end point, the vehicle's current lane, the real-time traffic flow around the vehicle, the degree of real-time traffic flow congestion around the vehicle, the type of road node where the vehicle is located, the width or curvature radius of the road where the vehicle is located, the type of road markings, the vehicle's lane changing frequency in each secondary area, the vehicle's average driving angle in each secondary area, etc. The selection of characteristic parameters can be specifically selected according to actual needs.

[0076] In a specific application embodiment, after extracting relevant parameters of historical samples according to a certain time period, a decision tree model is adopted, with the constraint degree calculated in the above formula (1) as the prediction value (i.e., the dependent variable), and parameters such as vehicle real-time motion information, surrounding traffic flow status, and road environment information extracted from historical samples as independent variables, to perform model training and verify the model results, and finally construct a constraint degree level prediction model for each complex road node.

[0077] In a specific application embodiment, the above historical sample data can be specifically extracted cyclically with a monitoring cycle T. For example, a week is a monitoring cycle T=7 days. The historical sample data is 5 small segments of data extracted within the previous day of this monitoring cycle, which are one segment each during the morning and evening peaks, morning and afternoon flat peaks, and nighttime. The length of each small segment of data is not less than 30 minutes. In a specific application embodiment, the selected historical sample data are: the instantaneous speed of the vehicle when entering the toll station diversion area is 50.78km / h; the overall average speed of the vehicle in area 1 is 42.50km / h; the speed of the vehicle when it is about to enter the toll channel is 23.57km / h, and the overall average speed in area 8 is 22.28km / h; the average speed of the vehicle is 33.79km / h; the average driving time of the vehicle through the complex node is 20.23s. The driving angle characteristics of the vehicle extracted in the specific application embodiment are shown in Table 1, and the lane change behavior characteristics of the vehicle are shown in Table 2.

[0078] Table 1 Driving angle characteristics

[0079]

[0080] Table 2 Vehicle lane-changing behavior characteristics

[0081]

[0082] In a specific application embodiment, a decision tree model can be established using a MATLAB box, and a cross-validation method is selected to verify the model. The decision tree model fitting accuracy is 84.9%, the model AUC = 0.90, and the true positive rate (ie, sensitivity) is 90%, that is, the model is highly authentic.

[0083] In the above step S04, when the real-time monitoring of the safety status of complex road nodes is turned on, a real-time vehicle ID is established in the form of a number for the vehicle entering the monitoring area, and the vehicle IDs at the same node in the same time period cannot be repeated.

[0084] In the above step S05, real-time monitoring data of vehicles and traffic flows are obtained through road monitoring equipment. The real-time monitoring data specifically includes speed data collected by radar speedometers, vehicle trajectory data collected by high-definition cameras, location data obtained by high-precision positioning devices, vehicle type data collected by charging methods, etc., which can be selected according to actual needs. The real-time motion parameters of the vehicle and the surrounding environment parameters specifically include the independent variables required to construct the constraint prediction model in step S03. The calculation can be completed by edge computing and cloud computing to ensure that the calculation speed meets the requirements of safety warning. For example, the simple motion parameters of the vehicle are edge calculated by the road side unit (Road Side Unit), and the complex motion parameters are cloud-assisted calculated by the highway management platform.

[0085] In a specific application embodiment, when the above method of the present invention is used to realize vehicle risk prediction, first, a coordinate system is established for different types of complex road nodes, and regional subdivision is performed to establish a complete complex road node database; then, sample data is selected from the vehicle motion information database according to a certain time period to construct a constraint prediction model; when a vehicle enters a complex road node, the road detection equipment detects the vehicle and triggers the safety warning function of the complex road node, and the vehicle obtains a unique vehicle ID; then real-time vehicle motion information and environmental information are obtained, and real-time monitoring data of vehicles and traffic flows are obtained through road monitoring equipment, and real-time motion and surrounding environmental information of each vehicle is calculated; based on the obtained relevant data and the established constraint prediction model, the driving behavior constraint of the vehicle is calculated, and the size of the constraint corresponds to the level of risk of the vehicle. The smaller the constraint, the higher the potential risk, and the vehicle risk prediction can be realized.

[0086] In step S07 of this embodiment, the specific steps of using the vehicle risk prediction results to implement active vehicle safety control include:

[0087] Divide the constraint degree into multiple levels in advance, and configure corresponding vehicle control strategies for different levels of constraint degree;

[0088] Control the vehicle according to the vehicle control strategy corresponding to the currently predicted constraint level;

[0089] If the vehicle is within the specified range around the preset road node, return to step S05 to continuously monitor the current vehicle position and motion status, calculate the constraint degree to predict the vehicle risk state, and control the vehicle according to the vehicle management and control strategy corresponding to the currently predicted constraint degree level until the vehicle leaves the current road node.

[0090] This embodiment first predicts vehicle risk based on the driving behavior constraint, uses the prediction result to determine the vehicle constraint level, and calls the corresponding control strategy output based on the constraint level, so that appropriate vehicle active safety control strategies can be flexibly obtained under different risks, effectively ensuring the vehicle's driving safety in various environments.

[0091] Specifically, the constraint degree can be divided into several constraint degree levels according to the actual situation, and an active safety management and control decision library can be formulated based on the constraint degree levels, which specifically includes the following steps:

[0092] (1) Determine the level of constraint: The constraint level is used to measure the potential risk of vehicle driving. The smaller the constraint level, the less restricted the vehicle is by road markings, signs, or surrounding traffic. The greater the possibility of the vehicle changing its state of motion. At the same time, the higher the degree of surrounding traffic congestion, the higher the driving risk.

[0093] Let S i ∈[a,b],S i = a, the constraint degree is the smallest, and the driving risk of the vehicle is the highest at this time, while S i = b, the degree of restraint is the largest, and the driving risk of the vehicle is the lowest at this time. Based on the calculation results of the vehicle restraint degree of the sample data, the sample restraint degree distribution range is determined, and the three restraint degree classification thresholds (i.e., the classification judgment critical points) are determined by the percentile method (25%, 50%, 75%) to divide the vehicle's motion restraint degree in the secondary area into four levels: A, B, C, and D. From A to D, the degree of restraint increases from low to high, which means that the potential risk of the vehicle decreases from high to low.

[0094] (2) Formulate an active safety control decision library based on the constraint level: Four levels of control decisions are formulated for the four levels of constraints A, B, C, and D. The content of the hierarchical active safety control decision for the complex road node n1 is shown in Table 4.

[0095] Table 3 Hierarchical active safety control decision of complex road node n1

[0096]

[0097] In a specific application embodiment, the vehicle's restraint level is calculated every 3-5 seconds, and the decision is updated over time and published on a roadside electronic display screen or a smart vehicle-mounted terminal. Further, active safety management and control decision information can be published on a roadside electronic display screen and a smart vehicle-mounted terminal so that users can obtain active safety management and control decision information in real time.

[0098] In a specific application embodiment, when the above method of the present invention is used for vehicle safety control, the degree of constraint is divided into several constraint levels in advance according to the actual situation, and an active safety control decision library is formulated based on the constraint level; after the vehicle's motion constraint level is predicted, the active safety decision corresponding to the current constraint level is obtained based on the active safety control decision library, and the warning information and active safety decision are published to the roadside information display screen and the intelligent vehicle terminal; the driver takes corresponding measures after obtaining the information, the road monitoring equipment monitors the vehicle's position and motion status, and updates the corresponding vehicle's position and motion information in the database, and executes the risk prediction of the vehicle in a loop until the vehicle leaves the complex node. It can be understood that after obtaining the current active safety control strategy of the vehicle, it can also be configured according to actual needs to control the vehicle according to the obtained active safety decision to achieve automatic control.

[0099] The vehicle active safety control device based on constraint degree in the information perception scenario of this embodiment includes:

[0100] A database establishment module is used to establish a coordinate system for different types of road nodes, and insert multiple sub-nodes within the monitoring range corresponding to each road node to perform regional subdivision and establish a road node database;

[0101] Roadside information sensing equipment, used to sense, capture and store road, vehicle and traffic flow information;

[0102] The constraint degree calculation and model training module is used to select historical sample data from the vehicle motion information database according to a preset time period, calculate the constraint degree of the vehicle at each road node based on the selected historical sample data, and train to form a constraint degree prediction model for each road node;

[0103] The warning trigger module is used to trigger a safety warning and obtain the ID number information of the current vehicle when the detected vehicle enters a specified range around a preset road node;

[0104] The information acquisition module is used to obtain the real-time monitoring data of the current vehicle and the traffic flow where the current vehicle is located according to the ID number information of the current vehicle, and to obtain the real-time movement and surrounding environment information of each vehicle within a specified range around the current road node;

[0105] The risk prediction module is used to calculate the current vehicle's restraint level based on the restraint prediction model and the data obtained in the information acquisition module, predict the vehicle's risk status based on the restraint level and issue warning information;

[0106] Active safety control module, used to control the acquisition of the corresponding pre-configured active vehicle control strategy according to the currently predicted constraint level and publish it;

[0107] The updating module is used to update the vehicle motion information database and the vehicle safety warning model according to a preset time period.

[0108] The above modules can be implemented in a separate manner or in an integrated manner, and the functions of each module can also be implemented in different independent modules. Figure 4 As shown, the vehicle active safety control device of this embodiment specifically includes:

[0109] The vehicle ID establishment module is used to establish an ID number for each vehicle in the node in real time;

[0110] Information storage module, used to store real-time dynamic data acquired by monitoring equipment and static data such as basic data of complex road nodes and basic data of vehicles;

[0111] The real-time traffic information acquisition module is used to acquire vehicle driving data in real time. The vehicle driving data includes the vehicle's charging type, the vehicle's real-time position, the vehicle's real-time speed, the vehicle's real-time lane change situation, the distance between the vehicle's current position and the node end point, the vehicle's current lane, the vehicle's surrounding real-time traffic flow, the degree of real-time traffic flow congestion around the vehicle, the type of road node where the vehicle is located, the width or curvature radius of the road where the vehicle is located, the type of road markings, and other information;

[0112] The restraint level calculation and active safety decision-making module is used to calculate the restraint level of the vehicle at different positions in real time based on relevant algorithms and data, and make active safety decisions based on the restraint level;

[0113] The information publishing module is used to publish active safety decision information on the roadside electronic display screen and transmit it to the vehicle-mounted information transmission device at the same time.

[0114] Furthermore, it also includes an information dynamic update module for uploading the acquired updated vehicle driving data to the information system.

[0115] The above-mentioned traffic information acquisition module specifically includes vehicle movement information sensing equipment and / or a GPS positioning system at complex road nodes; the information dynamic update module includes an Internet information system and / or a vehicle information sharing platform.

[0116] In this embodiment, the vehicle risk prediction device based on the driving behavior constraint degree and the vehicle risk prediction method based on the driving behavior constraint degree are in one-to-one correspondence, and the control device and the control method are in one-to-one correspondence, which will not be described one by one here.

[0117] The computer device of this embodiment includes a processor and a memory, wherein the memory is used to store a computer program, and the processor is used to execute the computer program to perform the above method, or to perform the above method.

[0118] This embodiment also provides a computer-readable storage medium storing a computer program, and the computer program implements the above method when executed.

[0119] The above is only a preferred embodiment of the present invention, and does not limit the present invention in any form. Although the present invention has been disclosed as a preferred embodiment, it is not intended to limit the present invention. Therefore, any simple modification, equivalent change and modification made to the above embodiment according to the technical essence of the present invention without departing from the content of the technical solution of the present invention shall fall within the scope of protection of the technical solution of the present invention.

Claims

1. A vehicle active safety control method based on constraint degree in an information perception scenario, characterized in that the steps include: S01. Establish a coordinate system for different types of road nodes, and insert multiple sub-nodes within the monitoring range corresponding to each road node to perform regional subdivision and establish a road node database; S02. Based on the roadside information sensing device, the road, vehicle and traffic flow information is captured and stored; S03. Selecting historical sample data from the vehicle movement information database according to a preset time period, calculating the degree of constraint of the vehicle at each road node based on the selected historical sample data, and training a constraint prediction model for each road node; The constraint degree is calculated according to the following formula: In the formula, i represents the region, is the degree of constraint imposed on the vehicle in the ith secondary area compared to the first secondary area, where the first secondary area is a normal road section, is the weight of the lane-changing index, is the weight of the angle indicator; is the lane-changing frequency of vehicles in the i-th zone, is the average driving angle of vehicles in the i-th area, is the minimum angle value of the vehicle in all areas; S04. When the detected vehicle enters the specified range around the preset road node, a safety warning is triggered and the ID number information of the current vehicle is obtained; S05. Obtaining real-time monitoring information of the current vehicle and the traffic flow where the current vehicle is located according to the ID number information of the current vehicle, and obtaining real-time movement and surrounding environment information of each vehicle within a specified range around the current road node; S06. Calculate the current vehicle's degree of constraint based on the constraint prediction model and the information obtained in step S05, predict the vehicle risk status based on the calculated degree of constraint and issue warning information; S07. Divide the constraint degree into multiple levels in advance, and configure corresponding vehicle control strategies for different levels of constraint degree, and control to obtain and publish the corresponding pre-configured active vehicle control strategy according to the currently predicted constraint degree level; said dividing the constraint degree into multiple levels includes: The potential risk of vehicle driving is determined according to the level of constraint, where the constraint degree S of each road node i is i ∈[a,b],S i = a, the constraint degree is the smallest and the corresponding vehicle risk is the highest. i = b, the degree of constraint is the largest and the corresponding vehicle risk is the lowest; According to the constraint degree calculation results of the historical sample data, the constraint degree distribution value range of the sample data is determined and multiple constraint degree classification thresholds are determined to divide the constraint degree of the vehicle in the secondary area into multiple levels, and the secondary area is an area obtained by subdividing the area within the corresponding monitoring range determined by each road node; S08. Update the vehicle movement information database and the vehicle safety warning model according to the preset time period.

2. The vehicle active safety control method based on constraint degree in information perception scenario according to claim 1 is characterized in that: The step S01 comprises: S101. Establishing a coordinate system: establishing a coordinate system for each road node, wherein the origin is determined according to the endpoints along the driving direction of the boundary points of each node area and the x-axis and y-axis are determined according to the driving direction, and a road plane coordinate system for each node is established; S102. Area subdivision: Determine the corresponding monitoring range according to the characteristics of each road node, and subdivide the area within the corresponding monitoring range on the x-axis of each node into multiple secondary monitoring areas, wherein the monitoring range includes a vehicle interweaving section and a normal environment section connected to the vehicle interweaving section.

3. The vehicle active safety control method based on constraint degree in information perception scenario according to claim 1 is characterized in that: In the step S03, characteristic parameters are extracted from the historical sample data to calculate the constraint degree of the vehicle at each road node, and the constraint degree calculated for each road node is trained using a decision tree model, wherein the calculated constraint degree is used as the predicted value, and the real-time vehicle movement information, surrounding traffic flow status and road environment information extracted from the historical sample data are used as independent variables to construct the constraint degree prediction model for each road node.

4. The vehicle active safety control method based on constraint degree in information perception scenario according to claim 3 is characterized in that: The characteristic parameters include the vehicle's charging type, the vehicle's real-time position, the vehicle's real-time speed, the vehicle's real-time lane changing situation, the distance between the vehicle's current position and the node end point, the vehicle's current lane, the real-time traffic flow around the vehicle, the degree of real-time traffic flow congestion around the vehicle, the type of road node where the vehicle is located, the width or curvature radius of the road where the vehicle is located, the type of road markings, the vehicle's lane changing frequency in each secondary area, and the vehicle's average driving angle in each secondary area.

5. The vehicle active safety control method based on constraint degree in information perception scenario according to claim 1 is characterized in that: The steps of step S07 include: Divide the constraint degree into multiple levels in advance, and configure corresponding vehicle control strategies for different levels of constraint degree; Control the vehicle according to the vehicle control strategy corresponding to the currently predicted constraint level; If the vehicle is within a specified range around a preset road node, return to step S05 to continuously monitor the current vehicle position and motion state, calculate the constraint degree to predict the vehicle risk state, and control the vehicle according to the vehicle management and control strategy corresponding to the currently predicted constraint degree level until the vehicle leaves the current road node.

6. A vehicle active safety control device based on constraint degree in an information perception scenario, characterized in that: include: A database establishment module is used to establish a coordinate system for different types of road nodes, and insert multiple sub-nodes within the monitoring range corresponding to each road node to perform regional subdivision and establish a road node database; Roadside information sensing equipment, used to sense, capture and store road, vehicle and traffic flow information; The constraint degree calculation and model training module is used to select historical sample data from the vehicle motion information database according to a preset time period, calculate the constraint degree of the vehicle at each road node based on the selected historical sample data, and train to form a constraint degree prediction model for each road node; The constraint degree is calculated according to the following formula: In the formula, i represents the region, is the degree of constraint imposed on the vehicle in the ith secondary area compared to the first secondary area, where the first secondary area is a normal road section, is the weight of the lane-changing index, is the weight of the angle indicator; is the lane-changing frequency of vehicles in the i-th zone, is the average driving angle of vehicles in the i-th area, is the minimum angle value of the vehicle in all areas; The warning trigger module is used to trigger a safety warning and obtain the ID number information of the current vehicle when the detected vehicle enters a specified range around a preset road node; The information acquisition module is used to obtain the real-time monitoring information of the current vehicle and the traffic flow where the current vehicle is located according to the ID number information of the current vehicle, and to obtain the real-time movement and surrounding environment information of each vehicle within a specified range around the current road node; The risk prediction module is used to pre-classify the degree of constraint into multiple levels, configure corresponding vehicle control strategies for different levels of constraint, calculate the constraint level of the current vehicle according to the constraint prediction model and the information obtained in the information acquisition module, predict the vehicle risk state according to the constraint level and issue warning information; the classification of the constraint into multiple levels includes: The potential risk of vehicle driving is determined according to the level of constraint, where the constraint degree S of each road node i is i ∈[a,b],S i = a, the constraint degree is the smallest and the corresponding vehicle risk is the highest. i = b, the degree of constraint is the largest and the corresponding vehicle risk is the lowest; According to the constraint degree calculation results of the historical sample data, the constraint degree distribution value range of the sample data is determined and multiple constraint degree classification thresholds are determined to divide the constraint degree of the vehicle in the secondary area into multiple levels, and the secondary area is an area obtained by subdividing the area within the corresponding monitoring range determined by each road node; Active safety control module, used to control the acquisition of the corresponding pre-configured active vehicle control strategy according to the currently predicted constraint level and publish it; The updating module is used to update the vehicle motion information database and the vehicle safety warning model according to a preset time period.

7. A computer device comprising a processor and a memory, wherein the memory is used to store a computer program, wherein: The processor is configured to execute the computer program to perform the method according to any one of claims 1 to 5.

Images