A vehicle lane-changing obstacle avoidance trajectory planning method integrating DMPs and APF

Abstract

This invention relates to a vehicle lane-changing obstacle avoidance trajectory planning method integrating DMPs and APF, comprising the following steps: acquiring a large amount of vehicle driving trajectory information from natural drivers through real-vehicle road tests and natural driving datasets; segmenting the vehicle driving trajectory based on vehicle heading angle information to extract the vehicle lane-changing trajectory; learning the lane-changing trajectory using a dynamic motion primitive algorithm and generalizing it into a human-like lane-changing trajectory; constructing a dynamic obstacle potential field considering the characteristics of natural drivers based on the artificial potential field method; and comprehensively applying the dynamic motion primitive algorithm and the artificial potential field method to plan the vehicle lane-changing obstacle avoidance trajectory. Compared with existing technologies, this invention considers the lane-changing operation characteristics of natural drivers, and can quickly plan lane-changing obstacle avoidance trajectories based on the planned human-like lane-changing trajectory, realizing integrated planning of path and speed, and has the advantages of high computational efficiency and strong generalization ability.

Description

Technical Field

[0001] This invention relates to the field of autonomous vehicle trajectory planning technology, and in particular to a vehicle lane-changing and obstacle-avoidance trajectory planning method that integrates DMPs and APF. Background Technology

[0002] Autonomous vehicles have become a hot research topic for automakers worldwide due to their enormous potential in improving traffic safety and efficiency. Lane-changing and obstacle-avoidance trajectory planning is one of its core technologies, and the results directly affect vehicle driving safety and ride comfort. Most studies decouple trajectory planning into path planning and speed planning, assuming that the speed remains constant. Commonly used path planning methods can be broadly categorized into artificial potential field methods, graph search methods, parametric curve methods, numerical optimization methods, and random sampling methods.

[0003] CN114194215A discloses a method and system for intelligent vehicle obstacle avoidance and lane-changing trajectory planning. This method first generates feasible trajectory families using polynomial curves, then selects collision-free trajectory families, and finally designs a cost function and uses an optimization-based method to solve for the optimal obstacle avoidance and lane-changing trajectory. CN114415694A discloses a real-time trajectory planning method and system for autonomous vehicles. It establishes and solves a motion control objective function and problem model based on model predictive control to complete real-time trajectory planning. CN110244713A discloses an intelligent vehicle lane-changing trajectory planning system and method based on the artificial potential field method. It borrows the obstacle repulsion field model from the artificial potential field method, considers the constraints of maximum lateral safe acceleration and road curvature, and plans the lane-changing path. However, the above trajectory planning methods rarely consider the lane-changing operation characteristics of natural drivers and decouple path planning from speed planning, ignoring the coupling relationship between path and vehicle speed.

[0004] In summary, a rapid lane-change obstacle avoidance trajectory planning method that considers the characteristics of natural driving lane changing and can achieve coordinated planning of lane-change paths and speeds still needs further research. Summary of the Invention

[0005] The purpose of this invention is to provide a vehicle lane-changing obstacle avoidance trajectory planning method that integrates DMPs (Dynamic Movement Primitives) and APF (Artificial Potential Field). Based on the anthropomorphic lane-changing trajectory planned by the Dynamic Movement Primitives algorithm, a dynamic obstacle potential field considering the characteristics of a natural driver is introduced to plan the lane-changing obstacle avoidance trajectory, ensuring that the vehicle can avoid obstacles in real time during the lane-changing process. This achieves collaborative planning of path and speed, improving computational efficiency and generalization ability.

[0006] The objective of this invention can be achieved through the following technical solutions:

[0007] A vehicle lane-changing obstacle avoidance trajectory planning method integrating DMPs and APF includes the following steps:

[0008] Step 1) Obtain a large amount of vehicle trajectory information from natural drivers through real-vehicle road tests and natural driving datasets;

[0009] Step 2) Segment the vehicle's trajectory based on the vehicle's heading angle information and extract the vehicle's lane-changing trajectory;

[0010] Step 3) Use the dynamic motion primitive algorithm to learn the lane-changing trajectory in Step 2) and generalize it into an anthropomorphic lane-changing trajectory;

[0011] Step 4) Construct a dynamic obstacle potential field that considers the characteristics of a natural driver based on the artificial potential field method;

[0012] Step 5) By combining dynamic motion primitives and artificial potential field methods, plan the vehicle's lane-changing obstacle avoidance trajectory.

[0013] Preferably, the vehicle trajectory information of the natural driver includes the vehicle's lateral and longitudinal positions, vehicle speed, vehicle acceleration, and timestamp.

[0014] Preferably, step 2) includes the following steps:

[0015] Step 2-1) Calculate the vehicle heading angle based on the vehicle position parameters;

[0016] Step 2-2) Traverse the heading angles in the forward direction from the lane change point to obtain the lane change endpoint; traverse the heading angles in the reverse direction to obtain the lane change starting point. The lane change point is defined as the intersection of the vehicle trajectory and the lane line.

[0017] Steps 2-3) Extract vehicle lane change trajectory segments based on the lane change start and end points.

[0018] Preferably, the method for calculating the vehicle heading angle is as follows:

[0019]

[0020] Where, θ (t) Let x be the heading angle at time t. (t) Let y be the longitudinal position of the vehicle at time t. (t) Let t be the lateral position of the vehicle at time t.

[0021] Preferably, the dynamic motion primitive algorithm is as follows:

[0022]

[0023] Where τ is the time scaling factor, used to scale the motion duration, trajectory velocity, and acceleration without changing their shape; α z and β z For system parameters, when α z =4β z At this point, the system is a critically damped system; g is the target position; z is the target velocity; z0 is the system's initial position; z, and These represent the system displacement, velocity, and acceleration, respectively; s is the phase variable, a function of time t; and f is the forcing function.

[0024] More preferably, the forcing function f is specifically:

[0025]

[0026] Among them, w i The basis function weights are M; the number of basis functions is M; and ψ(s) is the Gaussian function, specifically:

[0027] ψ i (s)=exp(-h i (sc i ) 2 )

[0028] Among them, h i c is the height of the basis functions; i It is the center of the basis functions.

[0029] More preferably, the phase variable is specifically:

[0030]

[0031] Where, α x >0 is a predefined constant.

[0032] Preferably, step 3) includes the following steps:

[0033] Step 3-1) Calculate the target forcing function based on the lane-changing trajectory, specifically:

[0034]

[0035] Among them, z demo , and These represent the displacement, velocity, and acceleration of the lane-changing trajectory, respectively.

[0036] Step 3-2) Construct a loss function and use the local weighted regression algorithm to solve for the weight values ​​corresponding to the basis functions. Specifically, the loss function is:

[0037]

[0038] The local weighted regression algorithm solves for the weight values ​​corresponding to the basis function as follows:

[0039]

[0040] in,

[0041] Step 3-3) Substitute the lane change start point, lane change end point, and weights of each basis function of the required generalized trajectory into the dynamic motion primitive algorithm to calculate the anthropomorphic lane change trajectory information.

[0042] Preferably, step 4) includes the following steps:

[0043] Step 4-1) Based on the characteristics of natural drivers, the higher the vehicle speed, the higher the vehicle speed relative to the obstacle vehicle, and the larger the range of influence of the obstacle potential field. Based on this, the range of influence of the obstacle potential field is determined as follows:

[0044] Step 4-1-1) Considering the different lateral and longitudinal motion characteristics of the vehicle, an elliptic curve is used to represent the range of influence of the obstacle potential field, specifically:

[0045]

[0046] Where x represents the longitudinal position coordinate, x0 represents the longitudinal position coordinate of the obstacle, y represents the lateral position coordinate, y0 represents the lateral position coordinate of the obstacle, a represents the longitudinal range of the obstacle's potential field, and b represents the lateral range of the obstacle's potential field.

[0047] Step 4-1-2) Determine the longitudinal range of the obstacle potential field based on the collision time (TTC) and the headway (THW):

[0048]

[0049] Among them, v ego v0 is the longitudinal velocity of the vehicle, v0 is the longitudinal velocity of the obstacle vehicle, TTC0 = 4s, THW0 = 2s;

[0050] Step 4-2) Considering the different lateral and longitudinal collision avoidance characteristics of vehicles, establish lateral and longitudinal potential fields respectively, specifically as follows:

[0051]

[0052] Among them, f xmax f ymax These represent the maximum values ​​of the longitudinal and transverse potential field strengths, respectively. rela y relaThese are the longitudinal relative distance and the lateral relative distance, respectively.

[0053] Step 4-3) Establish a Risk Perception (RP) model that includes TTC and THW to represent the degree of danger in the workshop, and adjust the obstacle avoidance potential field strength accordingly:

[0054] Step 4-3-1) Construct a risk perception coefficient model:

[0055]

[0056]

[0057] In the formula, A represents the weight of the steady-state term and B represents the weight of the transient term. When the vehicle and the vehicle in front are traveling at a constant speed, the steady-state term A / THW is mainly used to perceive the safety risks in the workshop. When the relative speed between the vehicle and the vehicle in front is large, the steady-state term A / THW and the transient term B / TTC are used together to perceive the safety risks in the workshop.

[0058] Step 4-3-2) Adjust the obstacle avoidance potential field strength:

[0059]

[0060] Among them, F x F y These represent the final longitudinal and transverse potential field intensities, respectively, while F represents the transverse and longitudinal coupled potential field intensities.

[0061] Preferably, step 5) includes the following steps:

[0062] Step 5-1) Define the vehicle's direction of motion and the obstacle deflection angle, specifically:

[0063] φ=arccos((oe) T v ego / (||oe||·||v ego ||))

[0064] Where o is the obstacle position vector, e is the vehicle position vector, (oe) is the vector between obstacle position o and vehicle position e, and v ego The longitudinal velocity vector of the vehicle;

[0065] Step 5-2) Define the obstacle avoidance coupling term:

[0066]

[0067] Where R represents the rotation about the axis r a =(oe)×v egoRotate by π / 2 to obtain the rotation matrix, where F is the intensity of the horizontal and vertical coupled potential field of the dynamic obstacle potential field determined in step 3).

[0068] Step 5-3) When there are multiple obstacles in the obstacle avoidance scenario, calculate the total potential field strength, specifically:

[0069]

[0070] Where j represents the obstacle number, n represents the number of obstacles, and p j This represents the obstacle avoidance coupling term for the j-th obstacle;

[0071] Step 5-4) Integrating dynamic motion primitives and artificial potential field methods, plan the vehicle's lane-changing obstacle avoidance trajectory, specifically as follows:

[0072]

[0073] Where f represents the learning of known information—existing lane-changing trajectories—and p all This reflects the current vehicle's response to the external environment, which cannot be solved using the concept of learning.

[0074] Compared with the prior art, the present invention has the following beneficial effects:

[0075] (1) High computational efficiency and strong generalization ability: In the process of generating anthropomorphic lane-changing trajectory and lane-changing obstacle avoidance trajectory, this invention only needs iterative calculation and does not involve numerical solution and optimization problems, so the computational efficiency is high; and according to the different target lane-changing positions, anthropomorphic lane-changing trajectory can be quickly generalized and can adapt to multi-obstacle environments to quickly generate lane-changing obstacle avoidance trajectory.

[0076] (2) The characteristics of natural drivers are fully considered: The present invention obtains lane change trajectory samples based on natural driver driving data, and the generalized lane change trajectory obtained therefrom is more in line with the operating characteristics of natural drivers; in addition, the characteristics of natural drivers are also fully considered in the process of establishing dynamic obstacle potential field based on artificial potential field method. Attached Figure Description

[0077] Figure 1 This is a flowchart of the method of the present invention;

[0078] Figure 2 A schematic diagram of extracted lane-change trajectory sample information of natural drivers;

[0079] Figure 3 This is a schematic diagram of the lane-changing trajectory learning.

[0080] Figure 4 This is a schematic diagram of the generalized lane-changing trajectory.

[0081] Figure 5A schematic diagram of the potential field construction for a dynamic obstacle;

[0082] Figure 6 Schematic diagram of vehicle speed and potential force coupling

[0083] Figure 7 A schematic diagram of the obstacle avoidance trajectory planning for lane changing. Detailed Implementation

[0084] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. These embodiments are based on the technical solution of the present invention and provide detailed implementation methods and specific operating procedures. However, the scope of protection of the present invention is not limited to the following embodiments.

[0085] This embodiment provides a vehicle lane-changing obstacle avoidance trajectory planning method that integrates DMPs and APF, such as Figure 1 As shown, it includes the following steps:

[0086] Step 1) Obtain a large amount of vehicle trajectory information from natural drivers through real-vehicle road tests and natural driving datasets.

[0087] In this embodiment, the vehicle trajectory information of the natural driver includes the vehicle's lateral and longitudinal positions, vehicle speed, vehicle acceleration, and timestamp.

[0088] Step 2) Segment the vehicle's trajectory based on the vehicle's heading angle information and extract the vehicle's lane-changing trajectory.

[0089] Specifically, it includes the following steps:

[0090] Step 2-1) Calculate the vehicle heading angle based on the vehicle position parameters:

[0091]

[0092] Where, θ (t) Let x be the heading angle at time t. (t) Let y be the longitudinal position of the vehicle at time t. (t) Let t be the lateral position of the vehicle at time t.

[0093] Step 2-2) Traverse the heading angles in the forward direction from the lane change point to obtain the lane change endpoint; traverse the heading angles in the reverse direction to obtain the lane change starting point. The lane change point is defined as the intersection of the vehicle trajectory and the lane line.

[0094] Steps 2-3) Based on the start and end points of the lane change, extract the vehicle lane change trajectory segment information, such as... Figure 2 As shown.

[0095] Step 3) Use the dynamic motion primitive algorithm to learn the lane-changing trajectory in Step 2), such as... Figure 3 As shown, and generalized to anthropomorphic lane-changing trajectories, such as... Figure 4 As shown.

[0096] The dynamic motion primitive algorithm is as follows:

[0097]

[0098] Where τ is the time scaling factor, used to scale the motion duration, trajectory velocity, and acceleration without changing their shape; α z and β z For system parameters, when α z =4β z At this point, the system is a critically damped system; g is the target position; z is the target velocity; z0 is the system's initial position; z, and These represent the system displacement, velocity, and acceleration, respectively; s is the phase variable, a function of time t; and f is the forcing function.

[0099] In this embodiment, the forcing function f is specifically:

[0100]

[0101] Among them, w i The basis function weights are M; the number of basis functions is M; and ψ(s) is the Gaussian function, specifically:

[0102] ψ i (s)=exp(-h i (sc i ) 2 )

[0103] Among them, h i c is the height of the basis functions; i It is the center of the basis functions.

[0104] The phase variables are specifically:

[0105]

[0106] Where, α x >0 is a predefined constant.

[0107] Specifically, step 3) includes the following steps:

[0108] Step 3-1) Calculate the target forcing function based on the lane-changing trajectory, specifically:

[0109]

[0110] Among them, z demo , and These represent the displacement, velocity, and acceleration of the lane-changing trajectory, respectively.

[0111] Step 3-2) Construct the loss function and use the local weighted regression algorithm to solve for the weight values ​​corresponding to the basis functions. The specific loss function is as follows:

[0112]

[0113] The local weighted regression algorithm solves for the weight values ​​corresponding to the basis functions as follows:

[0114]

[0115] in,

[0116] Step 3-3) Substitute the lane change start point, lane change end point, and weights of each basis function of the required generalized trajectory into the dynamic motion primitive algorithm to calculate the anthropomorphic lane change trajectory information.

[0117] Step 4) Construct a dynamic obstacle potential field that considers the characteristics of a natural driver based on the artificial potential field method, such as... Figure 5 As shown.

[0118] Specifically, it includes the following steps:

[0119] Step 4-1) Based on the characteristics of natural drivers, the higher the vehicle speed, the higher the vehicle speed relative to the obstacle vehicle, and the larger the range of influence of the obstacle potential field. Based on this, the range of influence of the obstacle potential field is determined as follows:

[0120] Step 4-1-1) Considering the different lateral and longitudinal motion characteristics of the vehicle, an elliptic curve is used to represent the range of influence of the obstacle potential field, specifically:

[0121]

[0122] Where x represents the longitudinal position coordinate, x0 represents the longitudinal position coordinate of the obstacle, y represents the lateral position coordinate, y0 represents the lateral position coordinate of the obstacle, a represents the longitudinal range of the obstacle's potential field, and b represents the lateral range of the obstacle's potential field.

[0123] Step 4-1-2) Determine the longitudinal range of the obstacle potential field based on the collision time (TTC) and the headway (THW):

[0124]

[0125] Among them, v ego v0 is the longitudinal velocity of the vehicle, and v0 is the longitudinal velocity of the obstacle vehicle. TTC0 = 4s, THW0 = 2s.

[0126] Step 4-2) Considering the different lateral and longitudinal collision avoidance characteristics of vehicles, establish lateral and longitudinal potential fields respectively, specifically as follows:

[0127]

[0128] Among them, f xmax f ymax These represent the maximum values ​​of the longitudinal and transverse potential field strengths, respectively. rela y rela These are the longitudinal relative distance and the lateral relative distance, respectively.

[0129] Step 4-3) Establish a Risk Perception (RP) model that includes TTC and THW to represent the degree of danger in the workshop, and adjust the obstacle avoidance potential field strength accordingly:

[0130] Step 4-3-1) Construct a risk perception coefficient model:

[0131]

[0132]

[0133] In the formula, A represents the weight of the steady-state term and B represents the weight of the transient term. When the vehicle and the vehicle in front are traveling at a constant speed, the steady-state term A / THW is mainly used to perceive the safety risks in the workshop. When the relative speed between the vehicle and the vehicle in front is large, the steady-state term A / THW and the transient term B / TTC are used together to perceive the safety risks in the workshop.

[0134] Step 4-3-2) Adjust the obstacle avoidance potential field strength:

[0135]

[0136] Among them, F x F y These represent the final longitudinal and transverse potential field intensities, respectively, while F represents the transverse and longitudinal coupled potential field intensities.

[0137] Step 5) By combining dynamic motion primitives and artificial potential field methods, the vehicle lane-changing obstacle avoidance trajectory is planned to ensure that the vehicle can avoid obstacles in real time during the lane-changing process.

[0138] Specifically, it includes the following steps:

[0139] Step 5-1) Define the vehicle's direction of motion and the obstacle deflection angle, such as... Figure 6 As shown, specifically:

[0140] φ=arccos((oe) T v ego / (||oe||·||v ego||))

[0141] Where o is the obstacle position vector, e is the vehicle position vector, (oe) is the vector between obstacle position o and vehicle position e, and v ego This is the longitudinal velocity vector of the vehicle.

[0142] Step 5-2) Define the obstacle avoidance coupling term:

[0143]

[0144] Where R represents the rotation about the axis r a =(oe)×v ego Rotate by π / 2 to obtain the rotation matrix, where F is the intensity of the horizontal and vertical coupled potential field of the dynamic obstacle potential field determined in step 3).

[0145] Step 5-3) When there are multiple obstacles in the obstacle avoidance scenario, calculate the total potential field strength, specifically:

[0146]

[0147] Where j represents the obstacle number, n represents the number of obstacles, and p j Let j represent the obstacle avoidance coupling term for the j-th obstacle.

[0148] Step 5-4) Integrating dynamic motion primitives and artificial potential field methods, plan the vehicle's lane-changing obstacle avoidance trajectory, such as... Figure 7 As shown, specifically:

[0149]

[0150] Where f represents the learning of known information—existing lane-changing trajectories—and p all This reflects the current vehicle's response to the external environment, which cannot be solved using the concept of learning.

[0151] The preferred embodiments of the present invention have been described in detail above. It should be understood that those skilled in the art can make numerous modifications and variations based on the concept of the present invention without creative effort. Therefore, all technical solutions that can be obtained by those skilled in the art based on the concept of the present invention through logical analysis, reasoning, or limited experimentation on the basis of existing technology should be within the scope of protection defined by the claims.

Claims

1. A vehicle lane-changing obstacle avoidance trajectory planning method integrating DMPs and APF, characterized in that, Includes the following steps: Step 1) Obtain a large amount of vehicle trajectory information from natural drivers through real-vehicle road tests and natural driving datasets; Step 2) Segment the vehicle's trajectory based on the vehicle's heading angle information and extract the vehicle's lane-changing trajectory; Step 3) Use the dynamic motion primitive algorithm to learn the lane-changing trajectory in Step 2) and generalize it into an anthropomorphic lane-changing trajectory; Step 4) Construct a dynamic obstacle potential field that considers the characteristics of a natural driver based on the artificial potential field method; Step 5) By comprehensively applying dynamic motion primitives and artificial potential field methods, plan the vehicle's lane-changing obstacle avoidance trajectory; Step 5) includes the following steps: Step 5-1) Define the vehicle's direction of motion and the obstacle deflection angle, specifically: in, The obstacle position vector, For vehicle position vector, Location of the obstacle With vehicle position Vectors between, The longitudinal velocity vector of the vehicle; Step 5-2) Define the obstacle avoidance coupling term: in, Indicates rotation around the axis Rotation Obtain the rotation matrix. The intensity of the lateral and longitudinal coupled potential field of the dynamic obstacle potential field determined in step 3); Step 5-3) When there are multiple obstacles in the obstacle avoidance scenario, calculate the total potential field strength, specifically: in, Indicates the obstacle number. Indicates the number of obstacles. Indicates the first One obstacle avoidance coupling term; Step 5-4) Integrating dynamic motion primitives and artificial potential field methods, plan the vehicle's lane-changing obstacle avoidance trajectory, specifically as follows: in, It is a time scaling factor used to scale the duration of motion, trajectory velocity, and acceleration without changing their shape; and For system parameters, when At this time, the system is a critically damped system; For the target location; For target speed; This is the system's starting position; , and These are the system displacement, velocity, and acceleration, respectively. Let be the phase variable and be time. The function; It is a forced function.

2. The vehicle lane-changing obstacle avoidance trajectory planning method integrating DMPs and APF according to claim 1, characterized in that, The vehicle trajectory information of the natural driver includes the vehicle's lateral and longitudinal positions, vehicle speed, vehicle acceleration, and timestamp.

3. The vehicle lane-changing obstacle avoidance trajectory planning method integrating DMPs and APF according to claim 1, characterized in that, Step 2) includes the following steps: Step 2-1) Calculate the vehicle heading angle based on the vehicle position parameters; Step 2-2) Traverse the heading angles in the forward direction from the lane change point to obtain the lane change endpoint; traverse the heading angles in the reverse direction to obtain the lane change starting point. The lane change point is defined as the intersection of the vehicle trajectory and the lane line. Steps 2-3) Extract vehicle lane change trajectory segments based on the lane change start and end points.

4. A vehicle lane-changing obstacle avoidance trajectory planning method integrating DMPs and APF according to claim 1 or 3, characterized in that, The method for calculating the vehicle's heading angle is as follows: in, for The heading angle at any moment, for The longitudinal position of the vehicle at any given moment. for The lateral position of the vehicle at any given moment.

5. The vehicle lane-changing obstacle avoidance trajectory planning method integrating DMPs and APF according to claim 1, characterized in that, The dynamic motion primitive algorithm is specifically as follows: in, It is a time scaling factor used to scale the duration of motion, trajectory velocity, and acceleration without changing their shape; and For system parameters, when At this time, the system is a critically damped system; For the target location; For target speed; This is the system's starting position; , and These are the system displacement, velocity, and acceleration, respectively. Let be the phase variable and be time. The function; It is a forced function.

6. The vehicle lane-changing obstacle avoidance trajectory planning method integrating DMPs and APF according to claim 5, characterized in that, The forced function Specifically: in, The basis function weights; The number of basis functions; The Gaussian function is as follows: in, The height of the basis functions; It is the center of the basis functions.

7. The vehicle lane-changing obstacle avoidance trajectory planning method integrating DMPs and APF according to claim 5, characterized in that, The phase variable is specifically: in, These are predefined constants.

8. The vehicle lane-changing obstacle avoidance trajectory planning method integrating DMPs and APF according to claim 6, characterized in that, Step 3) includes the following steps: Step 3-1) Calculate the target forcing function based on the lane-changing trajectory, specifically: in, , and These represent the displacement, velocity, and acceleration of the lane-changing trajectory, respectively. Step 3-2) Construct a loss function and use the local weighted regression algorithm to solve for the weight values ​​corresponding to the basis functions. Specifically, the loss function is: The local weighted regression algorithm solves for the weight values ​​corresponding to the basis function as follows: in, , , Step 3-3) Substitute the lane change start point, lane change end point, and weights of each basis function of the required generalized trajectory into the dynamic motion primitive algorithm to calculate the anthropomorphic lane change trajectory information.

9. The vehicle lane-changing obstacle avoidance trajectory planning method integrating DMPs and APF according to claim 1, characterized in that, Step 4) includes the following steps: Step 4-1) Based on the characteristics of natural drivers, the higher the vehicle speed, the higher the vehicle speed relative to the obstacle vehicle, and the larger the range of influence of the obstacle potential field. Based on this, the range of influence of the obstacle potential field is determined as follows: Step 4-1-1) Considering the different lateral and longitudinal motion characteristics of the vehicle, an elliptic curve is used to represent the range of influence of the obstacle potential field, specifically: in, Indicates the vertical position coordinates. Indicates the longitudinal position coordinates of the obstacle. Indicates the horizontal position coordinates. Indicates the lateral position coordinates of the obstacle. Indicates the longitudinal range of the potential field of the obstacle. Indicates the lateral range of the potential field of an obstacle; Step 4-1-2) Determine the longitudinal range of the obstacle potential field based on the collision time (TTC) and the headway (THW): in, For the longitudinal speed of the vehicle, The longitudinal speed of the obstacle vehicle; Step 4-2) Considering the different lateral and longitudinal collision avoidance characteristics of vehicles, establish lateral and longitudinal potential fields respectively, specifically: in, , These represent the maximum values ​​of the longitudinal and transverse potential field strengths, respectively. , These are the longitudinal relative distance and the lateral relative distance, respectively. Step 4-3) Establish a risk perception coefficient model that includes TTC and THW to represent the degree of danger in the workshop, and adjust the obstacle avoidance potential field strength, specifically as follows: Step 4-3-1) Construct a risk perception coefficient model: In the formula, A The weights represent the steady-state terms. B Represents the weights of the transient terms; Step 4-3-2) Adjust the obstacle avoidance potential field strength: in, , These represent the final longitudinal and transverse potential field intensities, respectively. The intensity of the horizontal and vertical coupled potential field.

Images