Dynamic obstacle trajectory prediction-based auxiliary driving obstacle avoidance method and system, and storage medium

Abstract

The application discloses an auxiliary driving obstacle avoidance method and system based on dynamic obstacle trajectory prediction and a storage medium, and belongs to the technical field of automobile intelligent driving. The auxiliary driving obstacle avoidance method specifically comprises the following steps: establishing a coordinate system with the geometric center of a vehicle as a dynamic coordinate origin, and discretizing a vehicle safety boundary and an obstacle contour into a point set; calculating the shortest distance and direction between the obstacle contour points and the vehicle safety boundary points, obtaining real-time risk distance and obstacle speed information; constructing a dynamic interactive risk field based on the obstacle speed, wherein the risk field diffuses along the direction of the obstacle speed and the diffusion range is determined based on the speed size; constructing a three-dimensional dynamic trajectory graph in combination with the real-time risk distance and its change rate, and delimiting a safety boundary through a risk control function; and when a trajectory point or a trend line crosses the safety boundary, selecting an optimal obstacle avoidance path according to the distribution of the dynamic interactive risk field. Through dynamic risk field planning and trajectory judgment, the application realizes more accurate and forward-looking obstacle avoidance decisions.

Description

Technical Field

[0001] This invention relates to the field of intelligent driving technology, and more specifically, to an assisted driving obstacle avoidance method, system, and storage medium based on dynamic obstacle trajectory prediction. Background Technology

[0002] With the rapid development of automotive intelligent technology, advanced driver assistance systems (ADAS) have become an important feature of modern vehicles. Among them, automatic obstacle avoidance is one of the core technologies for ensuring driving safety. Existing obstacle avoidance solutions have the following shortcomings: 1. Most existing obstacle avoidance methods are based on fixed thresholds or simple time-to-collision (TTC) models, which cannot accurately reflect the spatial relationships between complex contours in reality. Especially in close-range, irregular obstacle scenarios, this simplification leads to significant measurement errors. 2. They lack the ability to accurately judge the trajectory of obstacle points and mainly rely on instantaneous states for decision-making, which cannot effectively cope with sudden situations. 3. Most existing technologies only consider distance or relative speed factors and lack a multi-dimensional comprehensive risk assessment mechanism, making it impossible to make globally optimal decisions in complex multi-obstacle scenarios.

[0003] Therefore, there is an urgent need in this field for an obstacle avoidance solution that can accurately describe the dynamic relationship between vehicles and obstacles and has multi-dimensional risk assessment and forward-looking judgment capabilities. Summary of the Invention

[0004] In order to solve the problems mentioned in the background art, the purpose of this invention is to overcome the shortcomings of the prior art and provide an assisted driving obstacle avoidance method, system and storage medium based on dynamic obstacle trajectory prediction, so as to achieve more accurate and forward-looking obstacle avoidance safety protection in assisted driving.

[0005] To achieve the above objectives, the present invention includes, in one aspect, an assisted driving obstacle avoidance method based on dynamic obstacle trajectory prediction. In the technical solution of the present invention, the assisted driving obstacle avoidance method specifically includes the following steps:

[0006] Step S1: Construct a dynamic coordinate system with the geometric center of the vehicle as the origin of the dynamic coordinate system. Based on the geometric contour of the vehicle, set an assisted driving obstacle avoidance safety boundary around the vehicle body. Uniformly sample the set assisted driving obstacle avoidance safety boundary and discretize it into a boundary point set composed of multiple boundary points. Further obtain the coordinate information of each boundary point in the boundary point set based on the origin of the dynamic coordinate system.

[0007] Step S2: Through the multi-sensor fusion perception module, monitor the surrounding environment of the vehicle in real time, identify and track obstacles, obtain the real-time outline of each obstacle, and uniformly sample the real-time outline of each obstacle to discretize it into a set of outline points composed of multiple outline points. Each set of outline points represents the real-time boundary information of the corresponding obstacle.

[0008] Step S3: For each obstacle's contour point set, obtain the value and direction of the shortest distance between each contour point and its nearest boundary point. Further, based on the boundary point's coordinate information and the corresponding shortest distance value and direction, obtain the coordinate information of each contour point. Define the minimum value among the shortest distance values ​​corresponding to all contour points in the contour point set as the real-time risk distance between the obstacle and the vehicle. Further, the value and direction of the shortest distance between the contour point and the boundary point are represented as a vector pointing from the boundary point to the contour point, where the magnitude of the vector represents the value of the shortest distance and the direction of the vector represents the direction of the shortest distance.

[0009] Step S4: For each obstacle's contour point set, further record the coordinate information of each contour point based on the sampling time sequence of the multi-sensor fusion perception module. Calculate the real-time velocity vector of the obstacle based on the change values ​​of the contour point coordinate information and the vehicle's real-time velocity vector, and further obtain the magnitude and direction of the obstacle's real-time velocity. Specifically, calculating the real-time velocity vector of the obstacle based on the change values ​​of the contour point coordinate information and the vehicle's real-time velocity vector includes: calculating the real-time relative velocity vector of the obstacle relative to the vehicle based on the change values ​​of the contour point coordinate information, i.e., obtaining the displacement vector through the change values ​​of the contour point coordinate information, calculating the first derivative of the displacement vector with respect to time to obtain the corresponding real-time relative velocity vector of the obstacle relative to the vehicle; then calculating the real-time velocity vector of the obstacle based on the real-time relative velocity vector of the obstacle relative to the vehicle and the real-time velocity vector of the vehicle, i.e., obtaining the corresponding real-time velocity vector of the obstacle by calculating the vector sum of the real-time relative velocity vector of the obstacle relative to the vehicle and the real-time velocity vector of the vehicle.

[0010] Step S5: Based on the magnitude and direction of the real-time velocity of each obstacle, a dynamic interactive risk field is constructed for each obstacle. The dynamic interactive risk field starts from the real-time outline of each obstacle and spreads along the direction of the real-time velocity of each obstacle. The spread distance is calculated based on the magnitude of the real-time velocity of each obstacle. Furthermore, the spread distance of the dynamic interactive risk field is positively correlated with the magnitude of the real-time velocity of the obstacle, i.e.: ,in Represented as diffusion distance, Represented as real-time speed. Expressed as the diffusion coefficient;

[0011] Step S6: Based on the sampling time sequence records of the real-time risk distance and its first derivative with respect to time by the multi-sensor fusion sensing module, a three-dimensional dynamic obstacle trajectory map is constructed, consisting of the real-time risk distance, the distance change rate, and time. Based on this three-dimensional dynamic obstacle trajectory map, trajectory points representing the instantaneous state and trend lines representing the changing trend are generated. Further, the distance change rate is obtained by calculating the change in real-time risk distance per unit time, using the following formula: ,in Expressed as the rate of change of distance, This represents the real-time risk distance at the current moment. This represents the real-time risk distance at the previous moment. Represented as the sampling time interval;

[0012] Step S7: Define the total system response time based on sensor delay, processor calculation and control response time, and safety margin, and establish an obstacle trajectory risk control function accordingly. This obstacle trajectory risk control function is used to dynamically determine the maximum allowable distance change rate threshold under different real-time risk distances. Further, the specific form of the obstacle trajectory risk control function is as follows: ,in This is expressed as the maximum permissible rate of change of distance. This is expressed as the total system response time. Represented as a risk coefficient;

[0013] Step S8: Project the three-dimensional dynamic obstacle trajectory map onto a two-dimensional plane composed of its real-time risk distance and distance change rate, and define a safety boundary on this two-dimensional plane using the obstacle trajectory risk control function. When the trajectory point or trend line of an obstacle crosses the safety boundary, a collision risk is determined. Further, defining the safety boundary on the two-dimensional plane using the obstacle trajectory risk control function specifically includes: based on the obstacle trajectory risk control function... Calculate the maximum allowable distance change rate threshold under different real-time risk distances, where the range of different real-time risk distances is the sensing distance of the vehicle's multi-sensor fusion perception module; then, based on the calculated maximum allowable distance change rate threshold under different real-time risk distances, define the safety boundary, that is, under different real-time risk distances, the distance change rate cannot exceed the maximum allowable distance change rate threshold for the corresponding real-time risk distance;

[0014] Step S9: When a collision risk is determined, select to perform an obstacle avoidance operation based on the distribution of each obstacle and the dynamic interactive risk field. The obstacle avoidance operation includes longitudinal acceleration adjustment and / or lateral path planning. Further, selecting to perform an obstacle avoidance operation based on the distribution of each obstacle and the dynamic interactive risk field means avoiding collision risks with each obstacle as described in step S8 and avoiding the dynamic interactive risk field corresponding to each obstacle when performing the obstacle avoidance operation.

[0015] Another aspect of the present invention includes an assisted driving obstacle avoidance system based on dynamic obstacle trajectory prediction, used to implement the assisted driving obstacle avoidance method based on dynamic obstacle trajectory prediction as described above. In the technical solution of the present invention, the assisted driving obstacle avoidance system includes a multi-sensor fusion perception module, a data processing module, a risk assessment module, and a control execution module. The multi-sensor fusion perception module includes one or more of millimeter-wave radar, a camera, lidar, and ultrasonic radar, used to monitor the vehicle's surrounding environment in real time, identify and track obstacles, and obtain real-time shape contour information of each obstacle. The data processing module is based on the perception data from the multi-sensor fusion perception module. The information is processed, analyzed, and calculated in steps S1 to S9 of the assisted driving obstacle avoidance method based on dynamic obstacle trajectory prediction as described above. Based on the data processing, analysis, and calculation results of the data processing module, the risk judgment module performs the collision risk determination in step S8 of the assisted driving obstacle avoidance method based on dynamic obstacle trajectory prediction as described above. Based on the collision risk determination results of the risk judgment module and the data processing, analysis, and calculation results of the data processing module, the control execution module performs the obstacle avoidance operation in step S9 of the assisted driving obstacle avoidance method based on dynamic obstacle trajectory prediction as described above.

[0016] Another aspect of the present invention includes a computer-readable storage medium having a computer program stored thereon. In the technical solution of the present invention, when the computer program is executed by a processor, it implements an assisted driving obstacle avoidance method based on dynamic obstacle trajectory prediction as described above.

[0017] Beneficial Effects: In summary, this invention provides an assisted driving obstacle avoidance method, system, and storage medium based on dynamic obstacle trajectory prediction. Compared with existing assisted driving obstacle avoidance technologies, this invention avoids the errors of traditional simplified point models by using boundary discretization for precise modeling, significantly improving obstacle avoidance control accuracy. Simultaneously, this invention uses a dynamic risk field setting to assess risks during the obstacle avoidance interaction process, identifying potential risks in advance and improving the ability to predict obstacle avoidance risks. Furthermore, this invention combines real-time risk distance and distance change rate to comprehensively judge the obstacle's movement trend, optimizing obstacle avoidance decisions. In the obstacle trajectory risk control function model of this invention, the risk coefficient can be adjusted according to dynamic factors such as real-time application scenarios and real-time vehicle and personnel conditions to improve the control tolerance for obstacle avoidance safety risks.

[0018] Other features and advantages of the present invention will be set forth in the following description. Attached Figure Description

[0019] To more clearly illustrate the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a schematic flowchart of an assisted driving obstacle avoidance method based on dynamic obstacle trajectory prediction according to an embodiment of the present invention;

[0021] Figure 2 This is a schematic diagram of a three-dimensional dynamic obstacle trajectory according to an embodiment of the present invention;

[0022] Figure 3 This is a two-dimensional projection and safety boundary diagram of a three-dimensional dynamic obstacle trajectory map according to an embodiment of the present invention;

[0023] Figure 4 This is a schematic diagram of the dynamic interactive risk field distribution of an obstacle according to an embodiment of the present invention;

[0024] In the diagram: L represents the trend line; P represents the trajectory point; and S represents the dynamic interactive risk field. Detailed Implementation

[0025] To make the objectives, features, and advantages of this invention more apparent and understandable, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described below are only a part of the embodiments of this invention, and not all of them. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0026] To address the problems mentioned in the background art, this embodiment provides an assisted driving obstacle avoidance method, system, and storage medium based on dynamic obstacle trajectory prediction, so as to achieve more accurate and forward-looking obstacle avoidance safety protection in assisted driving.

[0027] In this embodiment, an assisted driving obstacle avoidance method based on dynamic obstacle trajectory prediction is provided. Figure 1 This is a schematic diagram of a driver assistance obstacle avoidance method based on dynamic obstacle trajectory prediction according to an embodiment of the present invention, as shown below. Figure 1 As shown, the obstacle avoidance method for assisted driving specifically includes steps S1-S9.

[0028] like Figure 1 As shown, in step S1, a dynamic coordinate system is first constructed with the geometric center of the vehicle as the origin of the dynamic coordinate system. That is, the geometric center of the vehicle is always used as the origin of the coordinate system during driving. At the same time, an auxiliary driving obstacle avoidance safety boundary, i.e. a collision safety boundary, is set around the vehicle body based on the geometric contour of the vehicle. The set auxiliary driving obstacle avoidance safety boundary is uniformly sampled and discretized into a boundary point set composed of multiple boundary points. Furthermore, the coordinate information of each boundary point in the boundary point set is obtained based on the origin of the dynamic coordinate system. That is, the coordinate information of the corresponding boundary point is calculated based on the position of each boundary point relative to the origin of the dynamic coordinate system.

[0029] Specifically, this can be represented as a vector pointing from the origin of the dynamic coordinate system to each boundary point. The magnitude of the vector represents the distance of the corresponding boundary point from the origin of the dynamic coordinate system, the direction of the vector represents the direction of that distance, and the coordinates of the vector represent the coordinates of the corresponding boundary point. In other words, the geometric contour of the vehicle is accurately modeled through boundary discretization.

[0030] Specifically, in step S2, the multi-sensor fusion perception module monitors the surrounding environment of the vehicle in real time, identifies and tracks obstacles, obtains the real-time outline of each obstacle, and uniformly samples the real-time outline of each obstacle, discretizing it into a set of outline points composed of multiple outline points. Each set of outline points represents the real-time boundary information of the corresponding obstacle. Similarly, the geometric outline of each obstacle is accurately modeled through boundary discretization.

[0031] Here, the geometric contours of the obstacles that are discretized here refer to the contours of the overall obstacle contours that are closest to the vehicle, which can be perceived by the multi-sensor fusion perception module.

[0032] Specifically, in step S3, for each obstacle's contour point set, the value and direction of the shortest distance between each contour point and its nearest boundary point are obtained. Further, based on the boundary point's coordinate information and the corresponding shortest distance value and direction, the coordinate information of each contour point is obtained, and the minimum value among the shortest distance values ​​corresponding to all contour points in the contour point set is defined as the real-time risk distance between the obstacle and the vehicle.

[0033] The shortest distance between the contour point and the boundary point can be represented by a vector pointing from the boundary point to the contour point. The magnitude of the vector represents the value of the shortest distance, and the direction of the vector represents the direction of the shortest distance. The real-time risk distance is selected as the minimum value among the shortest distance values, specifically representing the nearest collision distance between the corresponding obstacle and the vehicle. Specifically, this nearest collision distance, i.e., the real-time risk distance, changes in real time. That is, the minimum value among the shortest distance values ​​corresponding to all contour points in the contour point set is selected as the real-time risk distance. It can also be seen that not only does the real-time risk distance change, but the contour point used to determine the shortest distance and thus the real-time risk distance also changes in real time. That is, the nearest collision contour point for the obstacle collision also changes in real time. This real-time changing collision distance and collision point fully reflects the uncertainty and random variability of various obstacles in assisted driving.

[0034] Specifically, in step S4, for each obstacle's contour point set, the coordinate information of each contour point is further recorded based on the sampling time sequence of the multi-sensor fusion perception module. The real-time velocity vector of the obstacle is calculated based on the change value of the coordinate information of each contour point and the real-time velocity vector of the vehicle, and the magnitude and direction of the obstacle's real-time velocity are further obtained.

[0035] Specifically, the calculation of the obstacle's real-time velocity vector based on the changes in the coordinates of each contour point and the vehicle's real-time velocity vector includes: calculating the obstacle's real-time relative velocity vector relative to the vehicle based on the changes in the coordinates of each contour point, i.e., obtaining the displacement vector through the changes in the coordinates of each contour point, and calculating the first derivative of the displacement vector with respect to time to obtain the corresponding obstacle's real-time relative velocity vector relative to the vehicle. Specifically, when the obstacle deflects, the changes in the coordinates of its contour points may be different, with the contour point with the largest change value being used to calculate the displacement vector and real-time relative velocity vector; and then calculating the obstacle's real-time velocity vector based on the obstacle's real-time relative velocity vector relative to the vehicle and the vehicle's real-time velocity vector, i.e., obtaining the obstacle's real-time velocity vector by calculating the vector sum of the obstacle's real-time relative velocity vector relative to the vehicle and the vehicle's real-time velocity vector.

[0036] Specifically, in step S5, a dynamic interactive risk field is constructed for each obstacle based on the magnitude and direction of its real-time velocity. The dynamic interactive risk field starts from the real-time outline of each obstacle and spreads along the direction of the real-time velocity of each obstacle. The spread distance is calculated based on the magnitude of the real-time velocity of each obstacle.

[0037] Figure 4 This is a schematic diagram of the dynamic interactive risk field distribution of an obstacle according to an embodiment of the present invention, such as... Figure 4 As shown, the dynamic interactive risk field S of each obstacle deflects synchronously when the obstacle's position changes, and always uses the direction of the obstacle's real-time velocity as the direction of the diffusion of the dynamic interactive risk field S. Figure 4 The dynamic interactive risk field S shown is a brief fan-shaped region within the deflection time during the deflection process.

[0038] Specifically, the diffusion distance of the dynamic interactive risk field S is positively correlated with the magnitude of the real-time velocity of the obstacle, that is: ,in Represented as diffusion distance, Represented as real-time speed. It is expressed as the diffusion coefficient, which means that the greater the real-time speed of the obstacle, the longer the diffusion distance of its corresponding dynamic interactive risk field S.

[0039] Specifically, in step S6, Figure 2 This is a schematic diagram of a three-dimensional dynamic obstacle trajectory according to an embodiment of the present invention, such as... Figure 2 As shown, based on the sampling time sequence records of the multi-sensor fusion sensing module, the real-time risk distance and its first derivative with respect to time are recorded to construct a system with real-time risk distance as the basis. Distance change rate Please refer to the three-dimensional dynamic obstacle trajectory diagram composed of time t. Figure 2 In this three-dimensional dynamic obstacle trajectory diagram, the X-axis represents the real-time risk distance. Therefore, the origin of the X-axis coordinate system is the maximum sensing distance of the multi-sensor fusion sensing module, meaning the origin of the X-axis coordinate system represents the maximum real-time risk distance. The value is not 0, and the real-time risk distance is along the X-axis. The values ​​gradually decrease to 0, with the Y-axis representing the rate of change of distance. Real-time risk distance as the obstacle approaches the vehicle Decrease corresponds to the rate of change of distance. The value is negative, and the Z-axis is the time t-axis, representing the sampling timing of the multi-sensor fusion sensing module;

[0040] Specifically, based on this three-dimensional dynamic obstacle trajectory map, trajectory points P representing the instantaneous state and trend lines L representing the changing trend are generated; where trajectory point P refers to the real-time risk distance at any specific sampling moment. and the corresponding rate of change of distance It can be projected onto the real-time risk distance. Distance change rate Let P be a discrete data point on a two-dimensional plane with coordinate axes. Each trajectory point P precisely reflects the relative motion state between the vehicle and the obstacle at that instant. The trend line L is a continuous curve generated by linear fitting based on a series of historical trajectory points P over a period of time. This trend line L reveals the risk status, i.e., the real-time risk distance. and distance change rate The dynamic direction and inertia that evolve over time are used for forward-looking risk assessment;

[0041] Specifically, the rate of change of distance Calculate the real-time risk distance per unit time. The change value is obtained by the following formula: ,in Expressed as the rate of change of distance, This represents the real-time risk distance at the current moment. This represents the real-time risk distance at the previous moment. Represented as the sampling time interval, where the distance change rate is... The difference between the real-time relative velocity and that of obstacles lies in the rate of change of distance. What is more relevant is the real-time relative velocity of the nearest collision profile point, which itself changes in real time. This real-time changing collision point reflects the uncertainty and random variability of obstacles in assisted driving, representing the collision risk. Therefore, the distance change rate... Synchronization implies a risk of collision.

[0042] Specifically, in step S7, the total system response time is defined based on sensor delay, processor calculation and control response time, and safety margin, and an obstacle trajectory risk control function is established accordingly. The specific form of the obstacle trajectory risk control function is as follows: ,in This is expressed as the maximum permissible rate of change of distance. This is represented as real-time risk distance. This is expressed as the total system response time. Represented as a risk coefficient, this obstacle trajectory risk control function is used to dynamically determine the risk at different real-time risk distances. The maximum allowable rate of change threshold Real-time risk distance The smaller the maximum allowable rate of change threshold The lower the risk level, the better, especially considering the ability to dynamically adjust the risk coefficient based on real-time application scenarios, real-time vehicle and personnel conditions, and other factors. This is to improve the control and tolerance of obstacle avoidance safety risks.

[0043] Specifically, in step S8, Figure 3 This is a two-dimensional projection and safety boundary diagram of a three-dimensional dynamic obstacle trajectory map according to an embodiment of the present invention, such as... Figure 3 As shown, the three-dimensional dynamic obstacle trajectory map is projected onto its real-time risk distance. and distance change rate On the two-dimensional plane, a safety boundary is defined by an obstacle trajectory risk control function. When the trajectory point P or trend line L of an obstacle crosses the safety boundary, a collision risk is determined.

[0044] For details, please continue reading. Figure 3 Delineating safety boundaries on the two-dimensional plane using the obstacle trajectory risk control function specifically includes: based on the obstacle trajectory risk control function Calculate at different real-time risk distances The maximum allowable rate of change threshold Different real-time risk distances The range is the sensing distance of the vehicle's multi-sensor fusion sensing module, that is, as... Figure 3 The distance from the origin of the X-axis to the function is shown. The range of intersections with the X-axis; and then based on the calculated real-time risk distances. The maximum allowable rate of change threshold Delineate safety boundaries, that is, at different real-time risk distances. Below, the rate of change of distance It cannot exceed the corresponding real-time risk distance. Maximum allowable rate of change threshold ,refer to Figure 3 As shown, due to the rate of change of distance as the obstacle approaches the vehicle... Since it is a negative value, it is represented here as the rate of change of distance. The absolute value cannot exceed the corresponding real-time risk distance. Maximum allowable rate of change threshold The absolute value of the distance, i.e., the rate of change of distance. It needs to be controlled in such a way Figure 3 The function shown Above the line.

[0045] Specifically, in step S9, when a collision risk is determined, obstacle avoidance operation is selected to be performed based on the distribution of each obstacle and the dynamic interactive risk field. The obstacle avoidance operation includes longitudinal acceleration adjustment and / or lateral path planning.

[0046] Among them, selecting to perform obstacle avoidance operation based on the distribution status of each obstacle and the dynamic interactive risk field means avoiding collision risks with each obstacle as in step S8 and avoiding the dynamic interactive risk field corresponding to each obstacle when performing obstacle avoidance operation.

[0047] This embodiment also provides an assisted driving obstacle avoidance system based on dynamic obstacle trajectory prediction, used to implement the assisted driving obstacle avoidance method based on dynamic obstacle trajectory prediction as described above. The assisted driving obstacle avoidance system includes:

[0048] The multi-sensor fusion perception module includes one or more of millimeter-wave radar, camera, lidar and ultrasonic radar, used to monitor the vehicle's surrounding environment in real time, identify and track obstacles, and obtain real-time shape contour information of each obstacle;

[0049] The data processing module, based on the perception information from the multi-sensor fusion perception module, performs data processing, analysis, and calculation in steps S1 to S9 of the assisted driving obstacle avoidance method based on dynamic obstacle trajectory prediction described above.

[0050] The risk assessment module, based on the data processing analysis and calculation results of the data processing module, performs the collision risk assessment in step S8 of the assisted driving obstacle avoidance method based on dynamic obstacle trajectory prediction described above.

[0051] The control execution module, based on the collision risk assessment result of the risk assessment module and the data processing analysis and calculation result of the data processing module, executes the obstacle avoidance operation in step S9 of the assisted driving obstacle avoidance method based on dynamic obstacle trajectory prediction described above.

[0052] This embodiment also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements an assisted driving obstacle avoidance method based on dynamic obstacle trajectory prediction as described above.

[0053] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A driver assistance obstacle avoidance method based on dynamic obstacle trajectory prediction, characterized in that, Specifically, the steps include the following: Step S1: Construct a dynamic coordinate system with the geometric center of the vehicle as the origin of the dynamic coordinate system. Based on the geometric contour of the vehicle, set an assisted driving obstacle avoidance safety boundary around the vehicle body. Uniformly sample the set assisted driving obstacle avoidance safety boundary and discretize it into a boundary point set composed of multiple boundary points. Further obtain the coordinate information of each boundary point in the boundary point set based on the origin of the dynamic coordinate system. Step S2: Through the multi-sensor fusion perception module, monitor the surrounding environment of the vehicle in real time, identify and track obstacles, obtain the real-time outline of each obstacle, and uniformly sample the real-time outline of each obstacle to discretize it into a set of outline points composed of multiple outline points. Each set of outline points represents the real-time boundary information of the corresponding obstacle. Step S3: For each obstacle's contour point set, obtain the value and direction of the shortest distance between each contour point and its nearest boundary point. Further, based on the boundary point's coordinate information and the corresponding shortest distance value and direction, obtain the coordinate information of each contour point. Define the minimum value among the shortest distance values ​​corresponding to all contour points in the contour point set as the real-time risk distance between the obstacle and the vehicle. Step S4: For each obstacle's contour point set, further record the coordinate information of each contour point based on the sampling time sequence of the multi-sensor fusion perception module. Calculate the real-time velocity vector of the obstacle based on the change value of each contour point's coordinate information and the vehicle's real-time velocity vector, and further obtain the magnitude and direction of the obstacle's real-time velocity. Step S5: Based on the magnitude and direction of the real-time speed of each obstacle, construct a dynamic interactive risk field for each obstacle. The dynamic interactive risk field starts from the real-time outline of each obstacle and spreads along the direction of the real-time speed of each obstacle. The spread distance is calculated based on the magnitude of the real-time speed of each obstacle. Step S6: Based on the sampling time sequence recording of the multi-sensor fusion perception module, the real-time risk distance and its first derivative with respect to time are recorded to construct a three-dimensional dynamic obstacle trajectory map consisting of real-time risk distance, distance change rate and time, and based on the three-dimensional dynamic obstacle trajectory map, trajectory points representing instantaneous state and trend lines representing change trend are generated; Step S7: Define the total system response time based on sensor delay, processor calculation and control response time and safety margin, and establish an obstacle trajectory risk control function accordingly. This obstacle trajectory risk control function is used to dynamically determine the maximum allowable distance change rate threshold under different real-time risk distances. Step S8: Project the three-dimensional dynamic obstacle trajectory map onto a two-dimensional plane composed of its real-time risk distance and distance change rate, and define a safety boundary on the two-dimensional plane using the obstacle trajectory risk control function. When the trajectory point or trend line of the obstacle crosses the safety boundary, it is determined that there is a collision risk. Step S9: When a collision risk is determined, select to perform obstacle avoidance operation based on the distribution of each obstacle and the dynamic interactive risk field. The obstacle avoidance operation includes longitudinal acceleration adjustment and / or lateral path planning.

2. The assisted driving obstacle avoidance method based on dynamic obstacle trajectory prediction according to claim 1, characterized in that, In step S3, the value and direction of the shortest distance between the contour point and the boundary point are represented as a vector pointing from the boundary point to the contour point. The magnitude of the vector is represented as the value of the shortest distance, and the direction of the vector is represented as the direction of the shortest distance.

3. The assisted driving obstacle avoidance method based on dynamic obstacle trajectory prediction according to claim 1, characterized in that, In step S4, calculating the real-time velocity vector of the obstacle based on the changes in the coordinates of each contour point and the real-time velocity vector of the vehicle specifically includes: The real-time relative velocity vector of the obstacle relative to the vehicle is calculated based on the change value of the coordinate information of each contour point. That is, the displacement vector is obtained by the change value of the coordinate information of each contour point. The first derivative of the displacement vector with respect to time is calculated to obtain the corresponding real-time relative velocity vector of the obstacle relative to the vehicle. The real-time velocity vector of the obstacle is calculated based on the real-time relative velocity vector of the obstacle relative to the vehicle and the real-time velocity vector of the vehicle. In other words, the real-time velocity vector of the obstacle is obtained by calculating the vector sum of the real-time relative velocity vector of the obstacle to the vehicle and the real-time velocity vector of the vehicle.

4. The assisted driving obstacle avoidance method based on dynamic obstacle trajectory prediction according to claim 1, characterized in that, In step S5, the diffusion distance of the dynamic interactive risk field is positively correlated with the real-time velocity of the obstacle, that is: ,in Represented as diffusion distance, Represented as real-time speed. It is expressed as the diffusion coefficient.

5. The assisted driving obstacle avoidance method based on dynamic obstacle trajectory prediction according to claim 1, characterized in that, In step S6, the distance change rate is obtained by calculating the change in real-time risk distance per unit time, and the calculation formula is: ,in Expressed as the rate of change of distance, This represents the real-time risk distance at the current moment. This represents the real-time risk distance at the previous moment. This represents the sampling time interval.

6. The assisted driving obstacle avoidance method based on dynamic obstacle trajectory prediction according to claim 5, characterized in that, In step S7, the specific form of the obstacle trajectory risk control function is as follows: ,in This is expressed as the maximum permissible rate of change of distance. This is expressed as the total system response time. This is expressed as a risk coefficient.

7. The assisted driving obstacle avoidance method based on dynamic obstacle trajectory prediction according to claim 6, characterized in that, In step S8, defining the safety boundary on the two-dimensional plane using the obstacle trajectory risk control function specifically includes: Risk control function based on obstacle trajectory Calculate the maximum allowable distance change rate threshold under different real-time risk distances, where the range of different real-time risk distances is the sensing distance of the vehicle's multi-sensor fusion perception module; The safety boundary is defined based on the maximum allowable rate of change threshold for different real-time risk distances. That is, the rate of change of distance cannot exceed the maximum allowable rate of change threshold for the corresponding real-time risk distance.

8. The assisted driving obstacle avoidance method based on dynamic obstacle trajectory prediction according to claim 1, characterized in that, In step S9, selecting to perform obstacle avoidance operation based on the distribution of each obstacle and the dynamic interactive risk field means avoiding collision risks with each obstacle as described in step S8 and avoiding the dynamic interactive risk field corresponding to each obstacle when performing obstacle avoidance operation.

9. A driver assistance system based on dynamic obstacle trajectory prediction, used to implement the driver assistance method based on dynamic obstacle trajectory prediction as described in any one of claims 1-8, characterized in that, include: The multi-sensor fusion perception module includes one or more of millimeter-wave radar, camera, lidar and ultrasonic radar, used to monitor the vehicle's surrounding environment in real time, identify and track obstacles, and obtain real-time shape contour information of each obstacle; The data processing module, based on the perception information of the multi-sensor fusion perception module, performs the data processing, analysis and calculation in steps S1 to S9 of the assisted driving obstacle avoidance method based on dynamic obstacle trajectory prediction as described in any one of claims 1-8. The risk assessment module, based on the data processing analysis and calculation results of the data processing module, performs the collision risk assessment in step S8 of the assisted driving obstacle avoidance method based on dynamic obstacle trajectory prediction as described in any one of claims 1-8. The control execution module, based on the collision risk assessment result of the risk assessment module and the data processing analysis and calculation result of the data processing module, executes the obstacle avoidance operation in step S9 of the assisted driving obstacle avoidance method based on dynamic obstacle trajectory prediction as described in any one of claims 1-8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements an assisted driving obstacle avoidance method based on dynamic obstacle trajectory prediction as described in any one of claims 1 to 8.

Images