Vehicle collision intelligent early warning system and method based on data analysis

Abstract

The invention relates to the technical field of vehicle early warning, and discloses a vehicle collision intelligent early warning system and method based on data analysis, and the system comprises a vehicle data collection module which is used for collecting current vehicle state data and vehicle surrounding environment data; the data processing module is used for processing and analyzing the collected current vehicle state data and vehicle surrounding environment data to obtain a vehicle collision probability coefficient; the collision analysis module is used for predicting vehicles which may collide according to the vehicle collision probability coefficient; and the warning and early warning module is used for warning the prediction result in real time, reminding a driver and informing the vehicle which may collide. According to the invention, the data of the vehicle and the data of the target vehicle are acquired through the vehicle data acquisition module, and the ground slippery condition, the traffic condition of the current road section and the vehicle type of the vehicle are all included in an analysis range, so that the calculated vehicle collision probability coefficient is more accurate, and the early warning accuracy of vehicle collision is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle early warning, and in particular to a vehicle collision intelligent early warning system and method based on data analysis. Background Art

[0002] With the rapid development of the automobile industry, the number of cars in the world continues to grow, the road traffic volume has increased dramatically, and the road traffic safety issue has become increasingly prominent and has become a global focus. According to statistics, frequent traffic accidents have not only caused a large number of casualties, but also caused huge economic losses and social burdens.

[0003] In this context, improving the safety performance of automobiles and preventing traffic accidents have become important issues that need to be addressed in the automotive industry and related technical fields. Traditional vehicle safety systems mainly rely on passive safety devices, such as seat belts and airbags. Although these devices can reduce occupant injuries after an accident, they cannot fundamentally prevent accidents.

[0004] In recent years, with the rapid development of sensor technology, computer technology, and communication technology, intelligent vehicle collision warning systems have emerged. The system integrates multiple sensors such as radar, LiDAR, and cameras, and combines advanced algorithm analysis to monitor the vehicle's surroundings in real time, identify potential dangerous targets, predict collision risks in advance, and issue warnings to drivers through sound, vision, and other methods, and even automatically take emergency braking and other measures in some cases, thereby effectively avoiding or reducing the occurrence of collision accidents.

[0005] However, some vehicle collision intelligent warning systems only predict the vehicle's driving trajectory and provide collision warnings based on whether there is an intersection between the current vehicle and the target vehicle's driving trajectory. However, they ignore factors such as the current vehicle's model and slippery ground during actual driving, causing the collision warning system to make incorrect judgments or miss reports, affecting the user experience and the actual safety of the vehicle. Summary of the invention

[0006] The purpose of the present invention is to provide a vehicle collision intelligent warning system and method based on data analysis, so as to solve the technical problem that some vehicle collision intelligent warning systems only predict the vehicle's driving trajectory and perform collision warning according to whether there is an intersection between the driving trajectories of the current vehicle and the target vehicle, but ignore factors such as the current vehicle model and slippery ground during actual driving, resulting in misjudgment or omission of the collision warning system, affecting the user experience and the actual vehicle safety.

[0007] The purpose of the present invention can be achieved through the following technical solutions:

[0008] A vehicle collision intelligent warning system based on data analysis, comprising:

[0009] Vehicle data collection module, used to collect current vehicle status data and vehicle surrounding environment data;

[0010] The data processing module processes and analyzes the collected current vehicle status data and vehicle surrounding environment data to obtain the vehicle collision probability coefficient;

[0011] A collision analysis module, used to predict vehicles that may collide based on vehicle collision probability coefficients;

[0012] The warning module warns the predicted results in real time and reminds the driver by audio, video or text, informing the vehicle that may collide.

[0013] As a further technical solution, the current vehicle status data is collected by on-board sensors, including the current vehicle position, speed, acceleration, and heading;

[0014] The vehicle's surrounding environment data is collected through radar, lidar and cameras, including the position, speed and direction of other vehicles.

[0015] As a further technical solution, the data processing module includes a preliminary prediction unit, and the working process of the preliminary prediction unit is:

[0016] Based on the current vehicle status data and the vehicle's surrounding environment data, the Kalman filter algorithm is used to predict the movement trajectory of the current vehicle and other vehicles in the future;

[0017] The current vehicle's trajectory is marked as C A , and number the motion trajectories of other vehicles as C B1 , C B2 , C B3 , ..., C Bn ;

[0018] The current vehicle's motion trajectory C A The motion trajectories C of other vehicles in sequence B1 , C B2 , C B3 , ..., C Bn Compare them one by one;

[0019] If there is an intersection between the two motion trajectories, it is preliminarily predicted that the current vehicle will collide with the corresponding vehicle;

[0020] Otherwise, it is preliminarily predicted that the current vehicle and the corresponding vehicle will not collide.

[0021] As a further technical solution, the process of determining whether two motion trajectories have an intersection is as follows:

[0022] Superimpose the two motion trajectories. If there is an overlap between the two motion trajectories, mark the position as the intersection point.

[0023] If there is a position where the distance value d1≤d0 exists between the two motion trajectories, then the position is also marked as the intersection point, and d0 is the preset distance threshold.

[0024] As a further technical solution, the process of obtaining the vehicle collision probability coefficient is:

[0025] The vehicle collision probability coefficient corresponding to the vehicle that is initially judged not to collide is set to 0;

[0026] Mark the vehicle that is initially predicted to collide with the vehicle, and calculate the vehicle collision probability coefficient of the marked vehicle;

[0027] The specific method for calculating the vehicle collision probability coefficient of the marked vehicle is:

[0028] First obtain the relative distance Δd between the two cars, and then obtain the relative speed v between the two cars r ;

[0029] Substitute into the formula The collision time TTC is calculated in

[0030] Substitute the collision time TTC into the formula Calculate and obtain the vehicle collision probability coefficient E of the currently marked vehicle;

[0031] Where F is the traffic complexity coefficient, Q is the road slippery coefficient, W is the vehicle model coefficient; α, β, γ, δ are preset weight coefficients.

[0032] As a further technical solution, the traffic complexity coefficient Among them, μ1 and μ2 are preset coefficients, q is the traffic flow of the current road section, and q th is the maximum flow that the current road section can handle, k is the traffic density of the current road section, and k c is the reference traffic density of the current road section.

[0033] As a further technical solution, the road slip coefficient Among them, ρ1, ρ2, ρ3 are preset proportional coefficients, JY is the rainfall per unit time, PS is the drainage efficiency of the current road, T is the temperature, r is the dynamic coefficient, and the process of calculating the dynamic coefficient is:

[0034] Obtain the change curve Q(t of the road slippery coefficient over time in the previous time period and the standard road slippery coefficient Q0;

[0035] Substituting into the formula: The dynamic coefficient r is calculated. If r>0, r takes the value of 1 / 2, otherwise, r takes the value of 1 / 3.

[0036] As a further technical solution, the vehicle model coefficient When the vehicle model is a small car, the value of W is 3; when the vehicle model is a medium-sized car, the value of W is 2; when the vehicle model is a large car, the value of W is 1.

[0037] As a further technical solution, the process of predicting vehicles that may collide based on the vehicle collision probability coefficient is as follows:

[0038] The vehicle collision probability coefficients E of the marked vehicles are sorted in order from large to small, and the vehicle ranked first is predicted as the vehicle that may collide.

[0039] A vehicle collision intelligent warning method based on data analysis, the method is applicable to the above-mentioned vehicle collision intelligent warning system based on data analysis.

[0040] Beneficial effects of the present invention:

[0041] (1) The present invention collects data of the vehicle and the target vehicle through the vehicle data collection module, and includes the slippery ground conditions, the traffic conditions of the current road section and the vehicle model in the analysis scope, so that the calculated vehicle collision probability coefficient is more accurate, which plays a role in improving the accuracy of vehicle collision warning and reducing the possibility of false alarms and missed alarms;

[0042] (2) The present invention processes the position, speed, acceleration, heading of the current vehicle, the position, speed and direction of other vehicles and other data through a data processing module, and predicts the motion trajectory of the vehicle and the target vehicle in the future through a Kalman filter algorithm. By superimposing the motion trajectory of the vehicle and the motion trajectory of the target vehicle, it is possible to quickly determine whether there is an intersection or a place with a close position. If there is, it means that the vehicle has the possibility of colliding with the vehicle. If not, it obviously indicates that the vehicle will not collide with the vehicle. However, since the vehicle is driving on the road, there may be more than one target vehicle around, so the target vehicles with the possibility of collision are marked, and then the vehicle collision probability coefficient corresponding to each target vehicle is calculated, and then sorted after calculation. The first ranked one has the highest collision probability. Finally, the warning module warns the predicted result in real time, and reminds the driver in the form of audio, video or text to inform the vehicle that may collide.

[0043] (3) The present invention uses a preliminary prediction unit to perform preliminary screening of vehicles around the vehicle, thereby avoiding invalid vehicle collision probability coefficients for these vehicles that are initially judged to be safe, thereby improving the overall operating efficiency; and for vehicles that are initially judged to have the possibility of collision, the vehicle collision probability coefficient of the currently marked vehicle is calculated through a formula, and the traffic complexity coefficient, road slippery coefficient, vehicle model coefficient and collision time are all included in it, and a correlation model is established to achieve accurate calculation of vehicle collision probability, providing an accurate data source for subsequent warnings and early warnings. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The present invention will be further described below in conjunction with the accompanying drawings.

[0045] Figure 1 It is a system structure block diagram of the present invention. DETAILED DESCRIPTION

[0046] 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.

[0047] See also Figure 1 As shown, the present invention is a vehicle collision intelligent warning system based on data analysis, comprising:

[0048] Vehicle data collection module, used to collect current vehicle status data and vehicle surrounding environment data;

[0049] The data processing module processes and analyzes the collected current vehicle status data and vehicle surrounding environment data to obtain the vehicle collision probability coefficient;

[0050] The collision analysis module is used to predict vehicles that may collide based on the vehicle collision probability coefficient. The process of predicting vehicles that may collide based on the vehicle collision probability coefficient is as follows:

[0051] The vehicle collision probability coefficients E of the marked vehicles are sorted in descending order, and the vehicle ranked first is predicted as a vehicle that may collide;

[0052] The warning module warns the predicted results in real time and reminds the driver by audio, video or text, informing the vehicle that may collide.

[0053] The current vehicle status data is collected by the vehicle-mounted sensor, including the current vehicle position, speed, acceleration, and heading; the process and principle of the above vehicle-mounted sensor to collect data are directly selected from the existing technology and will not be elaborated here;

[0054] The vehicle's surrounding environment data is collected through radar, lidar and cameras, including the position, speed and direction of other vehicles.

[0055] The data processing module includes a preliminary prediction unit, and the working process of the preliminary prediction unit is as follows:

[0056] According to the current vehicle state data and the vehicle surrounding environment data, the Kalman filter algorithm is used to predict the movement trajectory of the current vehicle and other vehicles in the future. It should be noted that the Kalman filter algorithm is a direct selection of the existing technology and will not be elaborated here.

[0057] The current vehicle's trajectory is marked as C A , and number the motion trajectories of other vehicles as C B1 , C B2 , C B3 , ..., C Bn ;

[0058] The current vehicle's motion trajectory C A The motion trajectories C of other vehicles in sequence B1 , C B2 , C B3 , ..., C Bn Compare them one by one;

[0059] If there is an intersection between the two motion trajectories, it is preliminarily predicted that the current vehicle will collide with the corresponding vehicle;

[0060] Otherwise, it is preliminarily predicted that the current vehicle and the corresponding vehicle will not collide.

[0061] In this embodiment, in order to solve the technical problems in the background technology, the vehicle data collection module is first used to collect the data of the vehicle and the target vehicle, and the slippery ground conditions, the traffic conditions of the current road section and the vehicle model are all included in the analysis scope, so that the calculated vehicle collision probability coefficient is more accurate, which plays a role in improving vehicle collision warning and reducing the possibility of false alarms and missed alarms; specifically, the data processing module processes the position, speed, acceleration, heading of the current vehicle, the position, speed and direction of other vehicles, and the Kalman filter algorithm is used to predict the movement trajectory of the vehicle and the target vehicle in the future. The motion trajectories of the target vehicles are superimposed, so that it is possible to quickly determine whether there is an intersection or a place with a close position. If there is, it means that the vehicle has the possibility of colliding with the vehicle. If not, it obviously indicates that the vehicle will not collide with the vehicle. However, since the vehicle is driving on the road, there may be more than one target vehicle around it. Therefore, the target vehicles with the possibility of collision are marked, and then the vehicle collision probability coefficient corresponding to each target vehicle is calculated. After calculation, they are sorted. The first one in the sorting indicates that the collision probability is the highest. Finally, the warning module will warn the predicted results in real time, and remind the driver in the form of audio, video or text to inform the vehicle that may collide.

[0062] The process of judging whether two motion trajectories have an intersection is as follows:

[0063] Superimpose the two motion trajectories. If there is an overlap between the two motion trajectories, mark the position as the intersection point.

[0064] If there is a position where the distance value d1≤d0 exists between the two motion trajectories, then the position is also marked as the intersection point, and d0 is the preset distance threshold.

[0065] In this example, in order to prevent the omission of intersections, after the movement trajectories of the vehicles are superimposed, not only are the overlapping positions marked as intersections, but the distance values ​​between the two vehicles at each moment in the future are also collected. Once the collected distance value is less than the distance threshold, it means that there is a possibility of intersection between the two vehicles at this position, so it is also marked as an intersection to improve the accuracy of the preliminary prediction unit and avoid omissions in subsequent warnings of possible collision vehicles as much as possible.

[0066] The process of obtaining the vehicle collision probability coefficient is:

[0067] The vehicle collision probability coefficient corresponding to the vehicle that is initially judged not to collide is set to 0;

[0068] Mark the vehicle that is initially predicted to collide with the vehicle, and calculate the vehicle collision probability coefficient of the marked vehicle;

[0069] The specific method for calculating the vehicle collision probability coefficient of the marked vehicle is:

[0070] First obtain the relative distance Δd between the two cars, and then obtain the relative speed v between the two cars r ;

[0071] Substitute into the formula The collision time TTC is calculated in

[0072] Substitute the collision time TTC into the formula Calculate and obtain the vehicle collision probability coefficient E of the currently marked vehicle;

[0073] Among them, F is the traffic complexity coefficient, Q is the road slippery coefficient, and W is the vehicle model coefficient; α, β, γ, and δ are preset weight coefficients, which are comprehensively formulated based on historical data and experience data.

[0074] In this embodiment, a calculation method for obtaining a vehicle collision probability coefficient E is provided. First, the vehicles around the vehicle are preliminarily screened by the preliminary prediction unit, and the vehicle collision probability coefficient E is directly set to 0, thereby avoiding the invalid vehicle collision probability coefficient E for the vehicles preliminarily judged to be safe again, thereby improving the overall operation efficiency; and for the vehicles preliminarily judged to have the possibility of collision, the formula The vehicle collision probability coefficient E of the currently marked vehicle is calculated. It can be seen from the above formula that the traffic complexity coefficient, road slippery coefficient, vehicle model coefficient and collision time are all included, and a correlation model is established to achieve accurate calculation of the vehicle collision probability, providing an accurate data source for subsequent warnings and early warnings.

[0075] The traffic complexity coefficient Among them, μ1 and μ2 are preset coefficients, q is the traffic flow of the current road section, and q th is the maximum flow that the current road section can handle, k is the traffic density of the current road section, and k c is the reference traffic density of the current road section.

[0076] This embodiment provides a specific calculation method for a complex traffic system. First, the traffic flow q and the traffic density k of the current road section are obtained through road monitoring cameras and special detection instruments. Then, based on historical data analysis, the maximum flow q that the current road section can handle is obtained. th and the reference traffic density k of the current road section c , and finally substitute into the formula The traffic complexity system F is calculated in the above formula. It can be shown that the larger the ratio of traffic density to reference traffic density, the greater the current road traffic density is and the higher the traffic complexity is. The larger the ratio of the current road section's traffic flow to the maximum flow that the current road section can handle, the closer the current road traffic flow is to saturation and the higher the traffic complexity is. However, since the road traffic density has a greater impact on the traffic complexity, assigning μ2 in the form of an exponent can make the calculation of the traffic complexity coefficient more accurate.

[0077] The road slip coefficient Among them, ρ1, ρ2, ρ3 are preset proportional coefficients, which are selected and determined based on empirical data, JY is the rainfall per unit time, PS is the drainage efficiency of the current road, T is the temperature, r is the dynamic coefficient, and the process of calculating the dynamic coefficient is:

[0078] Obtain the change curve Q(t) of the road slippery coefficient over time in the previous time period and the standard road slippery coefficient Q0;

[0079] Substituting into the formula: In the above example, the dynamic coefficient r is calculated. If r>0, r is 1 / 2, otherwise, r is 1 / 3. In this embodiment, a method for obtaining the road slippery coefficient is provided. Specifically, the rainfall per unit time JY, the drainage efficiency PS of the current road and the temperature T are obtained based on the detection instrument, and then substituted into the formula The road slippery coefficient Q is calculated in ; it can be seen from the formula that the higher the current road drainage efficiency, the less water on the road, the lower the road humidity, the smaller the road slippery coefficient Q, the higher T, the faster the road dries, and therefore the smaller the road slippery coefficient Q. However, the greater the rainfall, the higher the road humidity, and the greater the road slippery coefficient Q. Through the above method, a model for evaluating the road slippery degree is established for rainfall, temperature and road drainage efficiency, which plays a role in rapid evaluation and calculation with high accuracy; at the same time, in order to make dynamic adjustments, the formula is used Calculations are performed to change the road dynamic coefficient according to the cumulative change of the road slippery coefficient, making the calculation of the road slippery coefficient more accurate; the larger the road slippery coefficient Q, the more slippery the road is and the more likely a vehicle collision will occur.

[0080] The vehicle model coefficient When the vehicle model is a small car, the value of W is 3; when the vehicle model is a medium-sized car, the value of W is 2; when the vehicle model is a large car, the value of W is 1.

[0081] In this embodiment, the size of the vehicle can reflect the driver's emergency control ability of the vehicle. Obviously, the larger the vehicle size, the higher the control difficulty in the event of a collision, so the smaller the value. On the contrary, the smaller the vehicle size, the lower the control difficulty in the event of a collision, so the larger the value. The above scheme can effectively evaluate the influence of the vehicle model on the overall vehicle collision probability coefficient, thereby improving the accuracy of the final vehicle collision probability assessment.

[0082] A vehicle collision intelligent warning method based on data analysis, the method is applicable to the above-mentioned vehicle collision intelligent warning system based on data analysis.

[0083] It should be noted that the calculation formulas and various parameters involved in the calculation in the present invention have been dimensionally processed in advance, and the process of dimensionless processing is well known in the industry and will not be described here.

[0084] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A vehicle collision intelligent warning system based on data analysis, characterized in that: include: Vehicle data collection module, used to collect current vehicle status data and vehicle surrounding environment data; The data processing module processes and analyzes the collected current vehicle status data and vehicle surrounding environment data to obtain the vehicle collision probability coefficient; A collision analysis module, used to predict vehicles that may collide based on vehicle collision probability coefficients; The warning module warns the predicted results in real time and reminds the driver by audio, video or text, informing the vehicle that may collide.

2. The vehicle collision intelligent warning system based on data analysis according to claim 1 is characterized in that: The current vehicle status data is collected through vehicle-mounted sensors, including the current vehicle position, speed, acceleration, and heading; The vehicle's surrounding environment data is collected through radar, lidar and cameras, including the position, speed and direction of other vehicles.

3. The vehicle collision intelligent warning system based on data analysis according to claim 2 is characterized in that: The data processing module includes a preliminary prediction unit, and the working process of the preliminary prediction unit is as follows: Based on the current vehicle status data and the vehicle's surrounding environment data, the Kalman filter algorithm is used to predict the movement trajectory of the current vehicle and other vehicles in the future; The current vehicle's trajectory is marked as C A , and number the motion trajectories of other vehicles as C B1 , C B2 , C B3 , ..., C Bn ; The current vehicle's motion trajectory C A The motion trajectories C of other vehicles in sequence B1 , C B2 , C B3 , ..., C Bn Compare them one by one; If there is an intersection between the two motion trajectories, it is preliminarily predicted that the current vehicle will collide with the corresponding vehicle; Otherwise, it is preliminarily predicted that the current vehicle and the corresponding vehicle will not collide.

4. The vehicle collision intelligent warning system based on data analysis according to claim 3 is characterized in that: The process of judging whether two motion trajectories have an intersection is as follows: Superimpose the two motion trajectories. If there is an overlap between the two motion trajectories, mark the position as the intersection point. If there is a position where the distance value d1≤d0 exists between the two motion trajectories, then the position is also marked as the intersection point, and d0 is the preset distance threshold.

5. The vehicle collision intelligent warning system based on data analysis according to claim 3 is characterized in that: The process of obtaining the vehicle collision probability coefficient is: The vehicle collision probability coefficient corresponding to the vehicle that is initially judged not to collide is set to 0; Mark the vehicle that is initially predicted to collide with the vehicle, and calculate the vehicle collision probability coefficient of the marked vehicle; The specific method for calculating the vehicle collision probability coefficient of the marked vehicle is: First obtain the relative distance Δd between the two cars, and then obtain the relative speed v between the two cars r ; Substitute into the formula The collision time TTC is calculated in Then substitute the collision time TTC into the formula Calculate and obtain the vehicle collision probability coefficient E of the currently marked vehicle; Where F is the traffic complexity coefficient, Q is the road slippery coefficient, W is the vehicle model coefficient; α, β, γ, δ are preset weight coefficients.

6. The vehicle collision intelligent warning system based on data analysis according to claim 5 is characterized in that: The traffic complexity coefficient Among them, μ1 and μ2 are preset coefficients, q is the traffic flow of the current road section, and q th is the maximum flow that the current road section can handle, k is the traffic density of the current road section, and k c is the reference traffic density of the current road section.

7. The vehicle collision intelligent warning system based on data analysis according to claim 5 is characterized in that: The road slip coefficient Among them, ρ1, ρ2, ρ3 are preset proportional coefficients, JY is the rainfall per unit time, PS is the drainage efficiency of the current road, T is the temperature, r is the dynamic coefficient, and the process of calculating the dynamic coefficient is: Obtain the change curve Q(t of the road slippery coefficient over time in the previous time period and the standard road slippery coefficient Q0; Substituting into the formula: The dynamic coefficient r is calculated. If r>0, r takes the value of 1 / 2, otherwise, r takes the value of 1 / 3.

8. The vehicle collision intelligent warning system based on data analysis according to claim 5 is characterized in that: The vehicle model coefficient When the vehicle model is a small car, the value of W is 3; when the vehicle model is a medium-sized car, the value of W is 2; when the vehicle model is a large car, the value of W is 1.

9. The vehicle collision intelligent warning system based on data analysis according to claim 1 is characterized in that: The process of predicting vehicles that may collide based on the vehicle collision probability coefficient is as follows: The vehicle collision probability coefficients E of the marked vehicles are sorted in order from large to small, and the vehicle ranked first is predicted as the vehicle that may collide.

10. A vehicle collision intelligent warning method based on data analysis, characterized in that: The method is applicable to the vehicle collision intelligent warning system based on data analysis as described in any one of claims 1 to 9.

Images