Intersection accident influence based trajectory planning method for autonomous vehicle

Abstract

The application discloses a kind of intersection accident influence under automatic driving vehicle trajectory planning method, it belongs to automatic driving vehicle trajectory planning control field.The application solves the problem of low traffic efficiency of intersection after accident.The application quantifies the influence of various factors on vehicle by accurately perceiving and analyzing intersection environment after accident, constructing driving risk field by expected trajectory field, accident repulsion field and vehicle motion field, and providing real-time path adjustment guide for automatic driving vehicle by calculating lateral and longitudinal driving force and acceleration, which avoids long-time retention and traffic congestion of vehicle caused by accident, effectively improves vehicle traffic efficiency under intersection accident scene, reduces traffic delay, enhances driving safety, ensures the safety of vehicle passing through complex intersection environment, and reduces the risk of secondary accident.The application method can be applied to automatic driving vehicle trajectory planning under intersection accident influence.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of trajectory planning control of automatic driving vehicles, and particularly relates to a trajectory planning method for automatic driving vehicles under the influence of intersection accidents. BACKGROUND

[0002] At present, various types of vehicle flows cross in urban road intersections, and due to the multiple conflict areas, various traffic accidents are prone to occur, causing traffic congestion. With the increasing development of cities, more and more vehicles, traffic pressure increases, and traffic accidents at urban road intersections also increase. Once the internal space of the intersection is occupied by traffic accidents, the traffic space of the vehicles is compressed, which affects the traffic flow of different directions, and the process of vehicles detouring the obstacles to pass through the intersection may cause disorder, resulting in a decrease in traffic efficiency. At present, the research on the scene after intersection accidents is very limited, especially the lack of effective methods to optimize the traffic behavior of automatic driving vehicles to reduce the negative impact of accidents on vehicle traffic efficiency. Therefore, in order to solve the problem of low traffic efficiency of intersections after accidents, a trajectory planning method for automatic driving vehicles passing through intersections under the influence of accidents is needed. SUMMARY

[0003] The purpose of the present application is to solve the problem of low traffic efficiency of intersections after accidents, and a trajectory planning method for automatic driving vehicles under the influence of intersection accidents is proposed.

[0004] The technical solution adopted by the present application to solve the above technical problems is: a trajectory planning method for automatic driving vehicles under the influence of intersection accidents, which specifically comprises the following steps:

[0005] Step 1: The automatic driving vehicle uses an environmental perception device to perceive the environmental information of the intersection after the accident, and the perceived environmental information of the intersection after the accident includes intersection geometric layout information, accident obstacle information and vehicle motion state information;

[0006] Step 2: Determine whether the accident range and the expected trajectory overlap according to the accident obstacle information;

[0007] If the accident range and the expected trajectory do not overlap, then construct a vehicle motion field model according to the vehicle motion state information and execute step 3;

[0008] If the accident range and the expected trajectory overlap, then construct a vehicle motion field model according to the vehicle motion state information and execute step 4;

[0009] Step 3: Construct an expected trajectory field model according to the intersection geometric layout information, then construct an intersection post-accident driving risk field model according to the expected trajectory field model and the vehicle motion field model, and then execute step 5;

[0010] Step four, after constructing the accident repulsion field model, the intersection accident post-traffic risk field model is constructed according to the accident repulsion field model and the vehicle motion field model, and then step five is executed;

[0011] Step five, trajectory planning of the autonomous vehicle is performed according to the intersection accident post-traffic risk field model.

[0012] The accident post-traffic risk field model is:

[0013]

[0014] Wherein, is the accident post-traffic risk field model, S=0 represents the case that the accident range has no overlap with the expected trajectory, and S=1 represents the case that the accident range has overlap with the expected trajectory.

[0015] Further, the environment perception device includes a camera sensor, a radar sensor and a laser radar sensor.

[0016] Further, the intersection geometric layout information includes the entry lane width, the exit lane number, the exit lane width, the entry and exit lane position, the intersection shape and size of the intersection.

[0017] Further, the accident obstacle information includes the accident position and the area occupied by the accident.

[0018] Further, the vehicle motion state information includes the position of the autonomous vehicle in the intersection, the driving direction of the autonomous vehicle, the speed of the autonomous vehicle, the mass of the autonomous vehicle, the position of the front adjacent vehicle in the intersection, the driving direction of the front adjacent vehicle, the speed of the front adjacent vehicle and the distance between the autonomous vehicle and the front adjacent vehicle.

[0019] Further, the expected trajectory field model is constructed according to the geometric layout information of the intersection where the traffic accident occurs:

[0020]

[0021] Wherein, is the expected trajectory field model; λ lat is the normal attraction coefficient; d exp is the vertical distance of the spatial point (x, y) from the expected trajectory; W is the trajectory bandwidth; λ des is the end point coefficient, which attracts the vehicle to drive towards the end point; l exit is the trajectory length from the vertical point of the spatial point (x, y) on the expected trajectory to the end point of the expected trajectory; represents the expected direction of the autonomous vehicle at time t.

[0022] Further, the expected direction of the autonomous vehicle at time t is:

[0023]

[0024] wherein x0(t) represents the current position of the autonomous vehicle, x d (t) represents the desired position of the autonomous vehicle.

[0025] Further, the vehicle motion field model is constructed according to the vehicle motion state information, specifically:

[0026]

[0027] wherein λ con is a repulsion coefficient of the motion repulsion; λ ita is an obstacle repulsion coefficient; is the velocity vector of the autonomous vehicle at time t; Δv (t) is the difference between the speed of the autonomous vehicle and the speed of the front adjacent vehicle at time t; d k is the distance between the autonomous vehicle and the front adjacent vehicle; λ gui is an attraction coefficient of the motion attraction; L is a safety distance in the intersection; is the included angle between the driving direction vector of the autonomous vehicle and the front adjacent vehicle;

[0028]

[0029] wherein v (t) is the speed of the autonomous vehicle at time t; s0 is a static safety distance; T is a desired headway; α0 is the maximum acceleration of the autonomous vehicle; d is the comfortable deceleration of the autonomous vehicle.

[0030] Further, the accident repulsion field model is:

[0031]

[0032] wherein λ ita is an obstacle repulsion coefficient; d ita represents the closest distance between the spatial point (x, y) and the obstacle edge; d exit represents the straight-line distance between the spatial point (x, y) and the end point; λ itades is an end point coefficient of the accident repulsion field.

[0033] Further, the specific process of step five is:

[0034] Let the start time of the nth time step be t, and then calculate the driving risk field of the autonomous vehicle at (x, y) in the intersection at time t according to the post-accident driving risk field model, and then calculate the lateral field driving force and the longitudinal field driving force obtained by the autonomous vehicle:

[0035]

[0036] Among them, F X(n) F represents the lateral driving force of the autonomous vehicle within the nth time step. Y(n) The longitudinal driving force of the autonomous vehicle within the nth time step;

[0037] Based on the mass, lateral driving force, and longitudinal driving force of the autonomous vehicle, calculate the lateral and longitudinal accelerations at the nth time step:

[0038]

[0039] Among them, a X(n) For the lateral acceleration of an autonomous vehicle, a Y(n) Let M be the longitudinal acceleration of the autonomous vehicle, and M be the mass of the autonomous vehicle.

[0040] By default, the autonomous vehicle accelerates uniformly along the X and Y axes within a short time step. The autonomous vehicle will avoid obstacles and reach the endpoint of its trajectory in a timely manner, completing the passage through intersections with obstacles, while minimizing deviation from the intended path.

[0041] The trajectory planning result for the autonomous vehicle is as follows:

[0042]

[0043] Among them, (X) (n) ,Y (n) ) represents the position coordinates of the autonomous vehicle at the start of the nth time step; t′ represents the length of a time step, (X) (n+1) ,Y (n+1) () represents the position coordinates of the autonomous vehicle at the beginning of the (n+1)th time step; and These are the X-axis and Y-axis velocities of the autonomous vehicle at the (n+1)th time step, respectively. and These are the X-axis and Y-axis velocities of the autonomous vehicle at the nth time step, respectively.

[0044] The beneficial effects of this invention are:

[0045] This invention accurately senses and analyzes the intersection environment after an accident, constructs a driving risk field through the expected trajectory field, accident rejection field, and vehicle motion field, quantifies the impact of various factors on vehicles, and calculates the lateral and longitudinal driving forces and accelerations to provide real-time path adjustment guidance for autonomous vehicles. This avoids prolonged vehicle delays and traffic congestion caused by accidents, effectively improves vehicle throughput in intersection accident scenarios, reduces traffic delays, enhances driving safety, ensures safe passage of vehicles in complex intersection environments, and reduces the risk of secondary accidents.

[0046] The method proposed in this invention can dynamically respond to different types of accidents and the resulting environmental changes, exhibiting strong adaptability. For intersections with different geometric layouts and different types of traffic accidents, the method of this invention can perform effective trajectory planning, demonstrating good robustness and versatility. Attached Figure Description

[0047] Figure 1 This is a flowchart of an autonomous vehicle trajectory planning method under the influence of intersection accidents according to the present invention. Detailed Implementation

[0048] Specific implementation method one: Combining Figure 1 This embodiment describes a trajectory planning method for autonomous vehicles under the influence of intersection accidents. The method specifically includes the following steps:

[0049] Step 1: Autonomous vehicles use environmental perception devices to perceive environmental information at the intersection after an accident. The environmental perception devices used include, but are not limited to, camera sensors, radar sensors, and lidar sensors. The environmental perception technologies used include, but are not limited to, VTX technology.

[0050] The perceived post-accident intersection environmental information includes intersection geometry, accident obstacle information, and vehicle motion status information.

[0051] The specific geometric layout information of an intersection includes the width of the entrance lanes, the number of exit lanes, the width of the exit lanes, the location of the entrance and exit lanes, and the shape and size of the intersection. The number of entrance and exit lanes determines the number of track strips, and the shape of the track strips (curvature, radius, etc.) is related to the location of the entrance and exit lanes and the shape and size of the intersection.

[0052] Accident obstacle information specifically includes the location of the accident and the area it occupies.

[0053] The vehicle motion status information specifically includes the autonomous vehicle's position within the intersection, its direction of travel, its speed, its mass, the position of the adjacent vehicle in front within the intersection, its direction of travel, its speed, and the distance between the autonomous vehicle and the adjacent vehicle in front.

[0054] Step Two: In the absence of a traffic accident at the intersection, based on the intersection's size and layout and vehicle traffic habits, a short, rule-compliant driving trajectory can be used to constrain vehicle behavior. This trajectory is called the desired trajectory. When vehicles travel along the desired trajectory, individual differences will create a trajectory range, or trajectory band. Vehicles will travel to the endpoint of the trajectory within the range formed by the desired trajectory band. Therefore, when a traffic accident occurs at the intersection, the first thing to determine is whether the vehicle can travel to the endpoint according to the desired trajectory.

[0055] Determine whether the accident area overlaps with the expected trajectory based on the accident obstacle information;

[0056] If the accident area does not overlap with the expected trajectory, proceed to step three;

[0057] If the accident area overlaps with the expected trajectory, proceed to step four;

[0058] Step 3: For cases where the accident area does not overlap with the expected trajectories of each entrance / exit lane, i.e., the accident area does not affect vehicles in each entrance lane from traveling along the expected trajectory to the target exit lane, construct the expected trajectory field model based on the geometric layout information of the intersection where the traffic accident occurred:

[0059]

[0060] Where, λ lat It is the normal gravitational coefficient; d exp λ is the perpendicular distance from the spatial point (x, y) to the desired trajectory; W is the width of the trajectory band, which is generally taken as 3.5, but can also be determined based on the width of the entrance / exit lanes; des It is the destination coefficient, attracting vehicles to travel towards the destination; exit Let be the length of the trajectory from the perpendicular point of the spatial point (x,y) to the end point of the desired trajectory; This represents the desired direction of the autonomous vehicle at time t;

[0061]

[0062] Where x0(t) represents the current position of the autonomous vehicle, x d (t) represents the desired position of the autonomous vehicle, and ||·|| represents the 2-norm.

[0063] When autonomous vehicles navigate intersections, they are not only repelled by obstacles and attracted by the endpoint of their trajectory, but also must employ strategies to maintain their safety boundaries to avoid collisions with surrounding vehicles. Therefore, the motion of adjacent vehicles in front also affects the vehicle's motion in real time; their directions, speeds, speed differences, and distances all influence the vehicle's movement. Due to the influence of adjacent vehicles, there is a critical distance at which the direction of force changes; this distance is defined as the vehicle's safe distance. Because of the complex movement of vehicles in intersections, there are safe distances in all directions, and the closed area enclosed by these safe distance points is the safe zone. Within the safe zone, adjacent vehicles exert a repulsive force on the vehicle; outside the safe zone, they exert an attractive force. When the distance between the vehicle and adjacent vehicles is at the safe distance, the vehicle is neither affected by attraction nor repulsion. The safe distance is not a fixed value but an instantaneous state parameter. Therefore, the motion of adjacent vehicles also affects the vehicle's motion in real time, and the safe distance between them is related to parameters such as their directions, speeds, speed differences, and distances.

[0064] Specifically, the vehicle motion field model, constructed based on the safe distance formed by the motion states of the autonomous vehicle and the adjacent vehicle in front, and the mutual attraction and repulsion effects experienced by the vehicle within and outside the safe distance, is as follows:

[0065]

[0066] Where, λ con λ is the repulsion coefficient of the repulsive force due to motion; ita The barrier rejection coefficient; Let Δv be the velocity vector of the autonomous vehicle at time t; (t) d represents the difference between the speed of the autonomous vehicle and the speed of the adjacent vehicle at time t; k λ represents the distance between the autonomous vehicle and the adjacent vehicle in front; gui is the attraction coefficient due to kinetic attraction; L is the safe distance within the intersection; It is the angle between the autonomous vehicle and the driving direction vector of the adjacent vehicle in front.

[0067]

[0068] Among them, v (t) Let be the speed of the autonomous vehicle at time t; s0 be the safe distance from the stationary vehicle, typically 2m; T be the desired headway, typically 1.1s; and α0 be the maximum acceleration of the vehicle, typically 1m / s². 2 d represents the vehicle's comfortable deceleration, taken as 1.5 m / s². 2 .

[0069] Then, construct the intersection post-accident driving risk field model based on the expected trajectory field model and the vehicle motion field model, and then execute step five;

[0070] Step 4: Construct an accident exclusion field model, and then construct a post-accident traffic risk field model at the intersection based on the accident exclusion field model and the vehicle motion field model. Then proceed to Step 5.

[0071] For cases where the accident area overlaps with the expected trajectories of each entrance / exit lane, i.e., the accident area affects the travel of vehicles in that entrance lane to the target exit lane, an accident repulsion field model is constructed based on the degree to which the behavior of autonomous vehicles is affected by accidents within the intersection:

[0072]

[0073] Where, λ ita d is the barrier rejection coefficient; ita d represents the closest distance from a spatial point (x, y) to the edge of an obstacle; exit λ represents the straight-line distance from a point (x, y) in space to the endpoint; itades It is the endpoint coefficient of the accident repulsion field;

[0074] Combining steps three and four, the post-accident driving risk field model can be obtained as follows:

[0075]

[0076] in, It is a post-accident driving risk field model. S=0 indicates that the accident area and the expected trajectory do not overlap, and S=1 indicates that the accident area and the expected trajectory overlap.

[0077] Step 5: Perform trajectory planning for autonomous vehicles based on the post-intersection traffic risk field model;

[0078] Specifically, the start time of the nth time step is denoted as t. After calculating the driving risk field of the autonomous vehicle at (x,y) within the intersection at time t based on the post-accident driving risk field model, the lateral driving force and longitudinal driving force obtained by the autonomous vehicle are then calculated:

[0079]

[0080] Among them, F X(n) F represents the lateral driving force of the autonomous vehicle within the nth time step. Y(n) The longitudinal driving force of the autonomous vehicle within the nth time step;

[0081] Based on the mass, lateral driving force, and longitudinal driving force of the autonomous vehicle, calculate the lateral and longitudinal accelerations at the nth time step:

[0082]

[0083] Among them, a X(n) For the lateral acceleration of an autonomous vehicle, a Y(n) Let M be the longitudinal acceleration of the autonomous vehicle, and M be the mass of the autonomous vehicle.

[0084] By default, the autonomous vehicle accelerates uniformly along the X and Y axes within a short time step. The autonomous vehicle will avoid obstacles and reach the endpoint of the trajectory in a timely manner while minimizing deviation, thus completing the passage through intersections with obstacles.

[0085] The trajectory planning result for the autonomous vehicle is as follows:

[0086]

[0087] Among them, (X) (n) ,Y (n) ) represents the coordinates of the autonomous vehicle at the start of the nth time step; t′ represents the length of a time step, (X) (n+1) ,Y (n+1) () represents the coordinates of the autonomous vehicle at the beginning of the (n+1)th time step; and These are the X-axis and Y-axis velocities of the autonomous vehicle at the (n+1)th time step, respectively. and These are the X-axis and Y-axis velocities of the autonomous vehicle at the nth time step, respectively.

[0088] Experimental Section

[0089] The specific intersection parameters are shown in Table 1:

[0090] Table 1 Intersection Parameters

[0091]

[0092] Using the simulation parameters shown in Table 2, the obstacle avoidance path planning effect of vehicles at intersections after a traffic accident can be obtained based on this model.

[0093] Table 2 Simulation Parameter Design

[0094]

[0095] This invention constructs a desired trajectory field, an accident exclusion field, and a moving vehicle field based on the acquired post-accident intersection environmental information. It then combines this with the post-accident traffic risk field at the intersection to perform trajectory planning, obtaining a safe obstacle avoidance path for vehicles within the intersection where the accident occurred. This invention accurately and precisely represents each element of the post-accident intersection traffic environment, models the impact of each element on vehicle behavior, and enables dynamic adjustment of the driving state and path of autonomous vehicles when passing through an accident-affected intersection.

[0096] The above examples of the present invention are merely illustrative of the computational model and process of the present invention, and are not intended to limit the implementation of the present invention. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is impossible to exhaustively list all possible implementations here. Any obvious variations or modifications derived from the technical solutions of the present invention are still within the scope of protection of the present invention.

Claims

1. A trajectory planning method for autonomous vehicles under the influence of intersection accidents, characterized in that, The method specifically includes the following steps: Step 1: The autonomous vehicle uses environmental perception equipment to perceive the environmental information of the intersection after the accident. The perceived environmental information of the intersection after the accident includes the geometric layout information of the intersection, the accident obstacle information, and the vehicle motion status information. Step 2: Determine whether the accident area overlaps with the expected trajectory based on the accident obstacle information; If the accident area does not overlap with the expected trajectory, then proceed to step three after constructing a vehicle motion field model based on the vehicle motion state information. The construction of the vehicle motion field model based on the vehicle motion state information specifically includes: in, The repulsion coefficient is the coefficient of motion repulsion. The barrier rejection coefficient; For autonomous vehicles The velocity vector at any given moment; The speed of the autonomous vehicle and the speed of the vehicle in front. The difference in velocity magnitude at any given moment; The distance between the autonomous vehicle and the adjacent vehicle in front; The attraction coefficient is the force of attraction due to motion. Safety distance within the intersection; The angle between the autonomous vehicle and the driving direction vector of the adjacent vehicle in front; Indicates that autonomous vehicles are in The expected direction at any given moment; in, For autonomous vehicles The magnitude of the velocity at any given moment; This is a safe distance when stationary. To achieve the desired headway, This represents the maximum acceleration of an autonomous vehicle. To improve the comfort of autonomous vehicles; If the accident area overlaps with the expected trajectory, then proceed to step four after constructing a vehicle motion field model based on the vehicle motion state information. Step 3: After constructing the desired trajectory field model based on the intersection geometric layout information, construct the intersection post-accident driving risk field model based on the desired trajectory field model and the vehicle motion field model, and then execute Step 5. Step 4: After constructing the accident exclusion field model, construct the intersection post-accident driving risk field model based on the accident exclusion field model and the vehicle motion field model, and then proceed to Step 5. Step 5: Plan the trajectory of the autonomous vehicle based on the post-intersection traffic risk field model.

2. The method for trajectory planning of an autonomous vehicle under the influence of an intersection accident as described in claim 1, characterized in that, The intersection geometry information includes the width of the entrance lanes, the number of exit lanes, the width of the exit lanes, the location of the entrance and exit lanes, and the shape and size of the intersection.

3. The method for trajectory planning of an autonomous vehicle under the influence of an intersection accident as described in claim 2, characterized in that, The accident obstacle information includes the accident location and the area occupied by the accident.

4. The method for trajectory planning of an autonomous vehicle under the influence of an intersection accident as described in claim 3, characterized in that, The vehicle motion status information includes the autonomous vehicle's position within the intersection, its driving direction, speed, mass, the position of the adjacent vehicle in front within the intersection, its driving direction, speed, and the distance between the autonomous vehicle and the adjacent vehicle.

5. The method for trajectory planning of an autonomous vehicle under the influence of an intersection accident as described in claim 4, characterized in that, The desired trajectory field model is constructed based on the geometric layout information of the intersection where the traffic accident occurred: in, It is the desired trajectory field model; It is the normal gravitational coefficient; It is a spatial point The vertical distance from the desired trajectory; The width of the trajectory band; It is the endpoint coefficient; For spatial points The length of the trajectory from the perpendicular point on the desired trajectory to the end point of the desired trajectory.

6. The method for trajectory planning of an autonomous vehicle under the influence of an intersection accident as described in claim 5, characterized in that, The autonomous vehicle The expected direction at any given time is: in, Indicates the current location of the autonomous vehicle. Indicates the desired location of the autonomous vehicle. This represents the 2-norm.

7. The method for trajectory planning of an autonomous vehicle under the influence of an intersection accident as described in claim 6, characterized in that, The accident repulsion field model is as follows: in, For accident repulsion field model, The barrier rejection coefficient; Representing a spatial point The closest distance to the edge of the obstacle; Representing a spatial point The straight-line distance from the destination; It is the endpoint coefficient of the accident repulsion field.

8. The method for trajectory planning of an autonomous vehicle under the influence of an intersection accident as described in claim 7, characterized in that, The post-accident driving risk field model is as follows: in, It is a post-accident driving risk field model. S=0 indicates that the accident area and the expected trajectory do not overlap, and S=1 indicates that the accident area and the expected trajectory overlap.

9. The method for trajectory planning of an autonomous vehicle under the influence of an intersection accident as described in claim 8, characterized in that, The specific process of step five is as follows: The first The start time of each time step is denoted as Calculated based on the post-accident driving risk field model Autonomous vehicles at intersections After determining the driving risk field, the lateral and longitudinal driving forces obtained by the autonomous vehicle are then calculated: , in, For the first Lateral driving force of autonomous vehicles within a time step For the first The longitudinal driving force of an autonomous vehicle within a time step; Based on the mass, lateral driving force, and longitudinal driving force of the autonomous vehicle, the calculation is performed on the... Lateral and longitudinal accelerations within a time step: , in, For the lateral acceleration of autonomous vehicles, For the longitudinal acceleration of autonomous vehicles, For the quality of autonomous vehicles; The trajectory planning result for the autonomous vehicle is as follows: , , in, In the first The position coordinates of the autonomous vehicle at the start of each time step; The length of a time step In the first The position coordinates of the autonomous vehicle at the start of each time step; and For autonomous vehicles in the X-axis and Y-axis velocities at each time step; and For autonomous vehicles in the X-axis and Y-axis velocities at each time step.

Images