Network connection cooperative driving automobile curve collision early warning method and system

Abstract

The invention belongs to the technical field of automatic driving, and particularly relates to a network connection cooperative driving automobile curve collision early warning method and system, and the method comprises the steps that a roadside RSU collects vehicle state information and curve environment information, and uploads the vehicle state information and the curve environment information to an edge cloud; the edge cloud carries out fusion analysis on the input data, and a Bi-LSTM network is adopted to predict a driving track of a non-networked automobile at a curve; an intelligent networked automobile driving risk field model is established, and the curve driving collision risk is evaluated in real time in combination with a surrounding automobile prediction track; calculating a collision risk threshold value by utilizing a Monte Carlo rule and combining human driving data, and performing curve collision early warning; the intelligent L2-level intelligent automobile environment perception is enabled through network connection, the curve driving collision early warning model is established, the problems of missing report and false report of the collision early warning system can be effectively reduced, and therefore the driving safety and efficiency are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of autonomous driving, and specifically relates to a curve collision warning method and system for a networked cooperatively driven vehicle. Background Art

[0002] With the rapid development of intelligent network technology, the intelligence of automobiles is constantly improving. More and more vehicles interact with the surrounding environment through V2X (Vehicle-to-Everything) technology, which greatly improves the safety and driving efficiency of road traffic. However, the complexity of the curve scene makes traditional vehicle sensors (such as radars and cameras) limited by the perception range and the influence of the external environment, and there are perception blind spots, resulting in poor performance of the collision warning system in such scenes.

[0003] Although existing curve collision warning systems can provide a certain degree of safety assurance, these systems usually rely on the vehicle's own sensing equipment and cannot obtain real-time driving information of other vehicles, limiting the accuracy and reliability of the warning. In contrast, the vehicle-road cooperative system based on intelligent network technology can overcome these limitations through real-time information sharing, providing more comprehensive environmental perception and more accurate risk assessment.

[0004] In recent years, with the introduction of artificial intelligence technology, especially the application of advanced algorithms such as bidirectional long short-term memory networks (Bi-LSTM), networked collaborative driving technology has made significant progress in risk prediction and assessment in complex scenarios. These technologies can accurately predict the driving trajectory of non-networked vehicles on curves, and combined with the driving risk field model of intelligent networked vehicles, significantly improve driving safety and efficiency. However, the existing technology still has certain limitations. Specifically, the safety and efficiency of the existing network-enabled intelligent L2-level smart car environmental perception driving are not high enough, and the collision warning on curves is not reliable enough, and there are problems of missed reports and false alarms. Summary of the invention

[0005] In order to solve the above problems existing in the prior art, the present invention proposes a curve collision warning method and system for networked cooperative driving vehicles.

[0006] In one aspect, the present invention provides a method for warning a collision on a curve of a networked cooperatively driven vehicle, comprising the following steps:

[0007] S1, collect roadside background vehicle status information and curve environment information, and upload to EC;

[0008] S2 and EC fuse the input data and use the Bi-LSTM network to predict the trajectory of non-connected cars on the curve;

[0009] S3. Establish a driving risk field model for intelligent connected vehicles, combine the predicted trajectories of surrounding vehicles, and evaluate the collision risk of driving on curves in real time;

[0010] S4. Use the Monte Carlo algorithm and human driving data to calculate the collision risk threshold and issue a curve collision warning.

[0011] On the other hand, the present invention provides a networked cooperative driving car curve collision warning system, the system comprising: a state information acquisition unit, a trajectory prediction unit, a risk assessment module, and a collision warning module;

[0012] The state information acquisition unit is used to obtain the vehicle state information, surrounding vehicle state information, and road environment information;

[0013] The trajectory prediction unit predicts the driving trajectory of the non-connected vehicles around the vehicle according to the input information of the information collection unit;

[0014] The risk assessment module calculates the risk of collision on a curve of the vehicle in real time based on the vehicle information, surrounding vehicle information and road environment information;

[0015] The collision avoidance warning module compares the real-time collision risk of the vehicle at the curve with a built-in collision risk threshold, and issues a collision risk warning when the risk exceeds the threshold.

[0016] Beneficial effects of the present invention:

[0017] The present invention enables intelligent L2-level smart car environmental perception through networking, establishes a collision warning model for driving on a curve, thereby improving driving safety and efficiency; by fusing multi-source data, the Bi-LSTM network model is used to accurately predict the driving trajectory of non-networked cars on curves, and combined with the driving risk field model of intelligent networked cars, the collision risk of driving on curves is evaluated in real time; the collision risk threshold is calculated by the Monte Carlo method combined with human driving data, making the collision warning more reliable, and can effectively reduce the problems of missed reports and false alarms of the collision warning system. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A flow chart of a method for early warning of a curve collision of a networked cooperatively driven vehicle according to the present invention;

[0019] Figure 2 A schematic diagram of a curve warning for a networked cooperative driving car curve collision warning according to the present invention;

[0020] Figure 3 A schematic diagram of the system structure of the networked cooperative driving car curve collision warning of the present invention;

[0021] Figure 4This is the system technology roadmap for the curve collision warning of the networked cooperative driving car of the present invention. DETAILED DESCRIPTION

[0022] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0023] like Figure 1 As shown, the present invention provides a method for warning a collision on a curve of a networked cooperatively driven vehicle, comprising the following steps:

[0024] S1. The roadside RSU collects vehicle status information and curve environment information and uploads it to the edge cloud (EC);

[0025] like Figure 1 In S1, in this step, the roadside unit (RSU) is installed at the entrance of the curve or other key locations, and uses its equipped multi-sensor system, such as high-precision GPS, lidar, millimeter-wave radar and high-definition cameras, to continuously monitor vehicles passing through the curve and their surrounding environment information. Vehicle status information includes but is not limited to the vehicle's position, speed, steering wheel angle, heading angle, acceleration, vehicle size and its lateral position relative to the lane. Curve environment information covers the lane curvature, road width, road conditions, real-time weather information (such as the slippery road surface caused by rain and snow) and lane identification information (such as lane lines and signs, etc.). By integrating multi-sensor data, the RSU can generate comprehensive environmental perception data and transmit this data to the EC with low latency through the 5G communication network for processing.

[0026] S2 and EC fuse and analyze the input data, and use the bidirectional long short-term memory network (Bi-LSTM) model to predict the driving trajectory of non-connected cars on the curve;

[0027] like Figure 1 In S2, when EC fuses and analyzes the data uploaded by RSU, a unified data set is formed through time series alignment, spatial mapping and feature extraction. After that, the Bi-LSTM model is used to predict the driving trajectory of non-connected cars at the curve. The Bi-LSTM model captures more time series features by independently processing forward and reverse time series data, thereby improving the accuracy of the prediction. The model forms a more reliable trajectory prediction by fusing the forward and backward feature vectors.

[0028] After completing the data fusion, the edge cloud uses the Bi-LSTM model to process the fused data. The Bi-LSTM model consists of two independent LSTM networks, one for processing forward time series data and the other for processing reverse time series data. Through this bidirectional processing, more time series features can be captured and the accuracy of trajectory prediction can be enhanced.

[0029] Specifically, the processing steps of the Bi-LSTM model are as follows:

[0030] Step 1: Forward LSTM processes the sequence from x1 to x t , and generates a hidden state at each time step t

[0031] Step 2: Reverse LSTM processing sequence from x t to x1, and generates a hidden state at each time step t

[0032] Step 3: For each time step t, the feature vectors extracted by the forward and backward LSTM are and spliced ​​together to form a single feature vector h t ;

[0033] Using weighted sum to merge feature vectors can be expressed as:

[0034]

[0035] Among them, α is a weight parameter (usually between 0 and 1) that is used to control the contribution of the forward and reverse feature vectors when merging.

[0036] These feature vectors are input into a hidden layer for deeper feature extraction; the expression is:

[0037] f=W·h t +c

[0038] a=σ(f)

[0039] Where W is the weight matrix; c is the bias term; σ is the activation function Sigmoid.

[0040] S3. Establish a driving risk field model for intelligent connected vehicles, combine the predicted trajectories of surrounding vehicles, and evaluate the collision risk of driving on curves in real time;

[0041] like Figure 1In S3, based on the non-networked vehicle’s driving trajectory predicted by the aforementioned Bi-LSTM model, the system further establishes a driving risk field model for intelligent networked vehicles. The driving risk field model uses a Gaussian function to quantify the collision risk of vehicles at different locations. Model parameters such as height and width are dynamically adjusted by factors such as vehicle speed and road curvature to accurately reflect the current driving risk. By combining the predicted trajectory of non-networked vehicles with environmental information, the system can dynamically adjust the parameters of the Gaussian function to generate accurate risk field maps. These maps combine the vehicle information with the trajectory of surrounding vehicles, providing a basis for real-time collision risk assessment.

[0042] Specifically, the construction of the driving risk field model includes the following steps:

[0043] Step 1: Determine the risk field model parameters. The risk field model uses a two-dimensional Gaussian function to represent the collision risk at different locations. The risk function R(x,y) of the model represents the collision risk at the coordinate point (x,y) and is defined as:

[0044]

[0045] Where R(x,y) is the collision risk function of point (x,y), h represents the height of the Gaussian cross-sectional function, w represents the width of the Gaussian cross-sectional function, and r c Represents the turning radius of the vehicle, (x o ,y o ) represents the coordinates of the center point of the turning radius.

[0046] Step 2: Combine vehicle trajectory with environmental information. The risk field model combines the predicted trajectory of non-connected vehicles, the dynamic information of the current vehicle (such as speed and direction), and the curve environment information (such as road curvature and width). This information is used to dynamically adjust the parameters of the Gaussian function so that the risk field can accurately reflect the current driving risk. In order to better achieve this goal, the system first generates an initial map based on the road environment and obstacles based on the surrounding environment information of the driving scene of the intelligent connected vehicle, and divides the map into multiple grid points.

[0047] Next, these grid points are assigned values ​​and cost calculations based on the obstacle type and orientation. Specifically, the system defines the cost based on the type of obstacle, and combines these costs with the orientation of the obstacle to generate a driving scenario cost map. The calculation of the orientation cost is based on the coordinates of all grid points, the vehicle position, the road width, and the vehicle turning radius. Updating the orientation matrix through the Dirac delta function can further optimize and reflect the orientation cost of the obstacle. Among them, the updated orientation matrix is ​​expressed as:

[0048]

[0049] Among them, Hdir,i is the orientation matrix, which represents the position information of the surrounding objects relative to the vehicle. dir represents the different orientations of the objects relative to the vehicle. i is the initial orientation matrix, which is calculated by combining the coordinates of all points on the current cost map with the vehicle position, road width, and vehicle turning radius. The initial orientation matrix represents the sensor's perception range, which is determined by dividing the grid points. The smaller the perception range, the greater the amount of calculation of the grid points involved.

[0050] The corresponding driving scenario cost Sc is obtained by multiplying the cost of each obstacle type and the cost of the obstacle orientation. The driving scenario cost Sc is expressed as:

[0051]

[0052] Among them, Sc represents the driving scenario cost of the intelligent connected car, dir=1 represents the front, dir=2 represents the back, dir=3 represents the left, dir=4 represents the right, i represents the obstacle type, cost i Represents the definition cost of the i-th obstacle type; when i=1, it represents edge green plants; when i=2, it represents roads; when i=3, it represents obstacle vehicles.

[0053] Multiply the risk field with the cost map of the driving scenario and sum it over all grid points to get the quantified driving risk R k :

[0054]

[0055] Among them, R k Represents the driving collision risk value, and meshgrids represents the set of all grid points.

[0056] Step 3: Real-time risk assessment. By calculating the risk value R(x,y) at different locations, the system can assess the collision risk of the vehicle in real time while the vehicle is driving. This risk assessment not only takes into account the current vehicle position and speed, but also combines the movement trajectory of surrounding vehicles to provide a comprehensive risk assessment result.

[0057] S4. Use the Monte Carlo algorithm and human driving data to calculate the collision risk threshold and issue a curve collision warning.

[0058] like Figure 1In S4, the system uses the Monte Carlo method to simulate various driving scenarios that may be encountered during cornering to assess the collision risk in different scenarios. The Monte Carlo method generates a large number of random sample driving scenarios to fully capture various risk events that may occur in the curve. These scenarios include parameters such as the relative position and speed between the vehicle and surrounding obstacles and other vehicles, thereby generating a wide range of driving scenario data sets.

[0059] To enhance the realism of the simulation results, the system introduces a large amount of real human driving data. This data covers the driving behavior patterns of drivers under different curve conditions, such as deceleration, steering, acceleration and other operations. These behavior data are embedded in the Monte Carlo simulation, making the risk assessment closer to actual human driving behavior, thereby improving the prediction accuracy of real driving risks.

[0060] Based on the scenarios and risk data generated by Monte Carlo simulation, the system statistically analyzes all possible risk values ​​to form a risk distribution model. By analyzing the model, the system determines a reasonable collision risk threshold. This threshold is used to determine whether a collision warning needs to be triggered when the risk value calculated in real time exceeds the threshold during actual driving.

[0061] During the driving process, the system calculates the current driving risk value in real time and compares it with the preset collision risk threshold. When the real-time calculated risk value exceeds the set threshold, the system will immediately issue a collision warning signal to avoid possible collisions.

[0062] In this embodiment, through multi-sensor data fusion and processing, the system obtains the vehicle's speed, position, steering wheel angle, heading angle, acceleration and other vehicle status information in real time, and also obtains the speed, position, historical trajectory of neighboring vehicles, and environmental information such as the road curvature, width, and road conditions of the curve. These data are transmitted to the EC via the 5G network, and the Bi-LSTM model is used to predict the driving trajectory of non-networked vehicles.

[0063] In this embodiment, the system establishes a driving risk field model through a Gaussian function and dynamically adjusts the model parameters according to real-time data to accurately assess the collision risk of the vehicle in the curve. A vehicle curve model can be constructed based on the current driving parameters, and the corresponding turning radius of the vehicle when turning in the current curve can be analyzed according to the vehicle curve model. The vehicle curve model can refer to Figure 2 When the risk value exceeds the set threshold, the system will immediately issue a warning signal to avoid potential collision accidents. The system's rapid response mechanism significantly improves the vehicle's driving safety in complex curve scenarios.

[0064] A method and system for warning collision on a curve of a networked cooperatively driven vehicle, such as Figure 3 As shown, the system includes: a state information acquisition unit, a trajectory prediction unit, a risk assessment module, and a collision warning module;

[0065] The status information acquisition unit is used to obtain the vehicle status information, surrounding vehicle status information, and road environment information; the status information acquisition unit uses a variety of sensors such as high-precision GPS, lidar, millimeter-wave radar and camera to collect vehicle speed, position, direction and other information in real time, and simultaneously obtains dynamic data of neighboring vehicles and environmental information such as the road curvature, width, and road conditions of the curve.

[0066] The trajectory prediction unit predicts the driving trajectories of non-connected vehicles around the vehicle based on the input information of the information collection unit; the trajectory prediction unit can identify potential collision risks in advance and provide important prediction data support for subsequent risk assessment.

[0067] The risk assessment module calculates the risk of collision on a curve of the vehicle in real time based on the vehicle information, surrounding vehicle information and road environment information; the risk assessment module ensures the comprehensiveness and accuracy of the risk calculation and provides the system with real-time risk status information.

[0068] The collision avoidance warning module compares the real-time collision risk of the vehicle at a curve with a built-in collision risk threshold, and issues a collision risk warning when the risk exceeds the threshold; the collision avoidance warning module effectively improves the safety of the vehicle when driving on a curve through real-time monitoring and rapid response.

[0069] In this embodiment, the system includes at least: an RSU for collecting driving data of non-connected vehicles; an edge cloud processing module for fusing and preprocessing the input data, and using the Bi-LSTM model to predict the driving trajectory of non-connected vehicles on curves; a driving risk field model for combining the driving risk field model of intelligent connected vehicles to evaluate the collision risk of driving on curves in real time; a collision warning system for calculating the collision risk threshold through the Monte Carlo method combined with human driving data, and issuing a collision warning when the risk value calculated in real time exceeds the set threshold.

[0070] In a specific embodiment, the RSU uses a variety of sensors such as cameras, lidars, and high-precision GPS to comprehensively collect driving data (such as speed, position, steering wheel angle) and surrounding environment information (such as road curvature, lane ID) of non-connected vehicles. These data are transmitted to the edge cloud in real time through the 5G communication network. The edge cloud fuses and preprocesses the received data and uses the Bi-LSTM model to predict the vehicle's driving trajectory. The predicted trajectory data will be further input into the driving risk field model to dynamically evaluate the collision risk in the curve to ensure that the evaluation at each moment is based on the latest vehicle motion prediction results. When the real-time risk value exceeds the preset threshold, the system immediately sends a warning signal through the 5G network. The system adopts a modular design to support future upgrades or integration with other advanced driver assistance systems, thereby continuously improving driving safety.

[0071] The system implementation of the present invention is the same as the method implementation.

[0072] A person skilled in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing the relevant hardware through a program, and the program can be stored in a computer-readable storage medium, which can include: ROM, RAM, disk or CD, etc.

[0073] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for warning collision on a curve of a networked cooperatively driven vehicle, characterized in that: include: S1, the roadside RSU collects vehicle status information and curve environment information, and uploads it to EC, where EC is the edge cloud server; S2 and EC fuse the input data and use the Bi-LSTM model to predict the trajectory of non-connected cars on the curve; S3. Establish a driving risk field model for intelligent connected vehicles, combine the predicted trajectories of surrounding vehicles, and evaluate the collision risk of driving on curves in real time; S4. Use the Monte Carlo algorithm and human driving data to calculate the collision risk threshold and issue a curve collision warning.

2. A method for warning a curve collision of a networked cooperatively driven vehicle according to claim 1, characterized in that: The information collected by the roadside RSU includes: the vehicle status information, the surrounding vehicle status information and the curve environment information; The vehicle status information includes: vehicle position, vehicle steering wheel angle, vehicle heading angle, vehicle speed, and vehicle size; The state information of the surrounding vehicles includes: vehicle position, vehicle speed, vehicle historical trajectory, and vehicle size; The curve environment information includes: lane curvature and lane ID.

3. The method for early warning of a curve collision of a networked cooperatively driven vehicle according to claim 1, characterized in that: EC fuses the input data, including: The EC data is fused to determine the input information of the non-connected car trajectory prediction model.

4. The method for early warning of a curve collision of a networked cooperatively driven vehicle according to claim 1, characterized in that: The Bi-LSTM model includes two independent LSTM networks, which are used to process the forward and reverse information of the input sequence respectively, so as to predict the driving trajectory of the non-networked car at the curve. The forward LSTM processes the sequence from x1 to x t , and generates a hidden state h at each time step t t fwd ; Reverse LSTM processes the sequence from x t to x1, and generates a hidden state at each time step t For each time step t, the hidden states generated by the forward and backward LSTMs are and spliced ​​together to form a single feature vector h t .

5. A method for warning a curve collision of a networked cooperatively driven vehicle according to claim 4, characterized in that: The hidden states generated by the forward and backward LSTM and spliced ​​together to form a single feature vector h t , include: Among them, α represents the contribution weight parameter used to control the forward and reverse feature vectors when merging.

6. The method for early warning of a curve collision of a networked cooperatively driven vehicle according to claim 1, characterized in that: Establish a risk field model for intelligent connected vehicles, including: Where R(x,y) is the collision risk function of point (x,y), h represents the height of the Gaussian cross-sectional function, w represents the width of the Gaussian cross-sectional function, and r c Represents the turning radius of the vehicle, (x o ,y o ) represents the coordinates of the center point of the turning radius.

7. The method for warning a curve collision of a networked cooperatively driven vehicle according to claim 1, characterized in that: Using the Monte Carlo method and combining human driving data to calculate the collision risk threshold, a curve collision warning is provided, including: The Monte Carlo method is used to calculate the collision risk threshold, which is then corrected using human driving data. The threshold is then compared with the real-time collision risk to complete the curve collision warning.

8. The method for early warning of a curve collision of a networked cooperatively driven vehicle according to claim 7, characterized in that: The Monte Carlo method is used to calculate the collision risk threshold, including: Among them, R k represents the driving collision risk value, meshgrids represents the grid point set, Sc represents the driving scenario cost of the intelligent connected vehicle, and R(x,y) is the collision risk function of the point (x,y).

9. A networked cooperative driving vehicle curve collision warning system, the system is used to execute any one of the networked cooperative driving vehicle curve collision warning methods according to claims 1 to 8, characterized in that: include: Status information acquisition unit, trajectory prediction unit, risk assessment module, and collision warning module; The state information acquisition unit is used to obtain the vehicle state information, surrounding vehicle state information, and road environment information; The trajectory prediction unit predicts the driving trajectory of the non-connected vehicles around the vehicle according to the input information of the information collection unit; The risk assessment module calculates the risk of collision on a curve of the vehicle in real time based on the vehicle information, surrounding vehicle information and road environment information; The collision avoidance warning module compares the real-time collision risk of the vehicle at the curve with a built-in collision risk threshold, and issues a collision risk warning when the risk exceeds the threshold.

Images