Extension and Application of Driving Risk Field Model Considering Vehicle Motion Constraints

Autonomous vehicle trajectory planning method based on the driving risk field is one of the current research hotspots. Currently, mainstream driving risk field models (DRFM) rarely consider basic factors such as vehicle geometry and heading angle, and ignore the impact of driving behavior characteristics on the driving risk field, the accuracy of safety assessment needs to be improved. Therefore, this paper proposes an extended driving risk field model (EDRFM) that considers geometric dimensions, heading angle, and driving behavior characteristics, and provides the basis for determining geometric dimensions and shape parameters, expanding the applicability of the DRFM. On this basis, EDRFM is applied to vehicle trajectory planning, and it is found that the target trajectory (denoted as P1) obtained using EDRFM is not achievable and the vehicle is unstable. This paper further proposes an arc correction method considering vehicle dynamics constraints to correct P1 to get P2. To ensure that the trajectory planning results meet the driving stability and motion of the vehicle. The results show that the EDRFM model considering arc correction can effectively complete vehicle trajectory planning tasks in complex scenes, and ensure the feasibility of trajectory planning and vehicle driving stability.


I. INTRODUCTION
With the rapid development of intelligent networked communication technology, vehicle-to-vehicle communication (V2X) technology has become possible.Using V2X technology, vehicles can interact with surrounding traffic participants in data and information.How to accurately measure complex road traffic environments and reasonably plan vehicle movement trajectory is of great significance for building a safe, efficient, and intelligent transportation network [1], [2].
There are many existing trajectory planning methods, as shown in Table 1, which lists the advantages and disadvantages of some of the current main trajectory planning algorithms.However, when applied to vehicle trajectory planning, more methods are based on vehicle kinematics and dynamics theory, and the expression of vehicle driving safety mainly relies on the vehicle's speed, acceleration, The associate editor coordinating the review of this manuscript and approving it for publication was Wei Wei .yaw rate, and relative motion state of the two vehicles [3], [4].These methods are difficult to reflect the impact of various traffic factors on driving safety, and it is difficult to reflect the interaction and dynamic changes of driver behavior characteristics, traffic environment, and vehicle status.It cannot provide an accurate judgment basis for driving decisions and vehicle control in complex driving environments.
To achieve more accurate vehicle trajectory planning, some researchers have proposed trajectory planning methods based on Artificial Potential Field (APF) theory and applied them to the field of mobile robots.The theory uses a mathematical model similar to a ''physical field'' to quantify the driving risks faced by mobile robots [5], [6].Since then, with the widespread application of ''field'' thinking in the decisionmaking and planning process of mobile robots, some scholars have applied it to traffic vehicles, established potential energy fields, and used APF to model car following behavior and design driving assistance systems.Reference [7] applies APF theory to lane-keeping systems and derives vehicles' lateral motion control rate based on the Lyapunov function, achieving the lane-keeping function.However, this method is only applicable to lane-keeping conditions.Reference [8] proposes a collaborative trajectory planning algorithm based on the concept of elastic order.This method uses APF theory to establish a driving risk field formed by road centerline and road boundaries and is applied to vehicle retention and collision avoidance systems.Reference [9] proposes a vehicle following model based on APF theory, which treats the vehicle as an independent unit charge in the potential energy field.The interaction between various factors on the vehicle is expressed using repulsive force and gravitational force, which better realizes the vehicle's motion following.Reference [10] believes that the natural potential energy fields formed by roads and vehicles exist in the subjective consciousness of drivers.When driving a vehicle, drivers always cross the potential energy field along the lowest point of the field.This theory conforms to the basic driving behavior and judgment mechanism of drivers but ignores the impact of the risk characteristics of drivers on driving safety.Reference [11] demonstrates the objectivity and universality of potential energy fields in transportation from both macro and micro perspectives.Reference [12] proposed a machine learning-based lane-changing intention prediction and autonomous vehicle controller.On the basis of predicting the vehicle's lane-changing intention, a car-following controller was designed using the conditional artificial potential field method to ensure vehicle safety.Reference [13] established a unified model of the driving risk field of man-vehicle road closed-loop systems according to different object mathematics.Reference [14] will use trigonometric functions and exponential functions on obstacles to construct a threedimensional virtual hazard field for roads, and generate the desired collision avoidance trajectory for vehicles based on the gradient descent method.Reference [15] introduces an ellipse correction formula to compress safety parameters in different directions of the road using different weights, making the risk field model more realistic.With the gradual improvement of risk field theory based on APF, vehicle kinematics constraints have gradually been taken into account in the model.Reference [16] has constructed a risk field model that considers obstacles, road boundaries, and tire dynamics.Reference [17] analyzes the relative motion state between vehicles, establishes a safety field, determines the braking deceleration of vehicles, and applies it to braking scenarios.
In summary, the APF-based driving risk field model has been widely used in vehicle trajectory planning, but through analysis of existing driving risk field models, it can be seen that its modeling theory has certain shortcomings: (1) The impact of vehicle travel direction on vehicle collision risk is less analyzed.The collision risk in the direction of vehicle travel is much higher than that in the rear and side.Currently, APF-based risk field construction theory has less consideration for this aspect.
(2) When constructing a driving safety field based on APF, it is obvious that the impact of overall vehicle dimensions on the risk field is not considered.Large trucks pose a higher risk than a small car.
(3) The degree of risk generated by drivers with different driving behavior characteristics is inconsistent, and the risk generated by aggressive drivers is significantly higher than that of cautious drivers.
(4) Inadequate consideration has been given to the driving stability of the vehicle itself, with only a few articles suggesting that vehicle motion needs to meet the motion limit determined by road adhesion conditions.
This study combines vehicle dynamics restrictions with an EDRFM based on APF to undertake vehicle trajectory planning.FIGURE 1 depicts the article's general technical flow.Firstly, the EDRFM model including potential energy field, kinetic energy field and behavior field was developed.Then, the minimum turning radius R min was determined through the vehicle dynamics model, road adhesion conditions, and Ackermann steering mechanism model.Finally, the trajectory P1 calculated based on the EDRFM model was smoothed with arc smoothing to obtain the trajectory P2.

II. EDRFM
The driving risk field model mainly consists of the potential energy field generated by stationary objects in the traffic environment, the kinetic energy field generated by moving objects, and the behavioral field generated by driving behavior.The factors that pose risks to driving can be divided into three categories [15]: (1) Potential energy field E R : A physical field that represents the impact of stationary objects on driving safety, including road boundaries, lane lines, stationary vehicles, roadblocks, and so on.(2) Kinetic energy field E v : A physical field that represents the impact of a moving object on driving safety, depending on the position, speed, acceleration, mass, size, and heading angle of the moving object.(3) Behavior field E D : A physical field that represents the impact of driver behavior characteristics on driving safety.Including drivers' physiological and psychological factors, driving.
A. POTENTIAL ENERGY FIELD

1) ROAD POTENTIAL ENERGY FIELD
To prevent vehicles from exiting the road boundary and crossing the opposite lane during driving, it is necessary to establish a road potential energy field.During driving on a two-way road, the vehicle has the highest risk level on the leftmost and rightmost sides of the road in the direction, and the vehicle has the lowest risk coefficient height when driving near the center line.According to the different degrees of danger on the road, a function with different trends in the degree of danger is used to construct the road boundary repulsion potential field.When vehicles are on both sides of the road, they belong to a road with a high degree of danger.
An exponential function with a fast-changing trend is used to establish the boundary potential field; When the vehicle is in the middle of the road, a trigonometric function with a smoother trend of change is used to establish the boundary function.According to the above description, the potential field function of the road boundary is as follows: where, γ 1 , γ 2 represents the potential energy field gain coefficients of the road boundary and the road centerline, L represents the road width, x l , x r represents the x-coordinate of the left road center and the x-coordinate of the right road center, respectively.The established road boundary potential field is shown in FIGURE 2.

2) STATIC OBSTACLE POTENTIAL ENERGY FIELD
To improve DRFM, this subsection incorporates vehicle dimensions into the static potential energy field.First, the vehicle shape is simplified to a rectangular shape, and then the static potential energy field of the vehicle is described using an elliptic equation.
The potential energy function can be represented by a twodimensional normal distribution function, with the function expression as follows: To simplify, let ρ = 0, then Formula (4) can be simplified as follows: The formula represents the gain coefficient of the static barrier potential energy field and ρ represents the correlation coefficient between the transverse and longitudinal directions.Considering that the exponential function must be greater than 0, a minimal positive number E( 0 According to the condition that the potential energy of static obstacles at each point on the obstacle avoidance trajectory is equal, we can further obtain it δ 1 . FIGURE 3 (a) and FIGURE 3 (b) are schematic diagrams for solving δ 1 and δ 2 .
When the heading angle ϕ of a static obstacle is not 0 • , it is necessary to perform a coordinate transformation based on formula (5).Assuming that the original coordinate is (x, y, E R2 ), the changed coordinate expression is (x ′ , y ′ , E ′ R2 ), and the transformation formula is as follows: According to formulas (2), (5), and ( 8), the comprehensive potential energy field formed by multiple stationary objects on the road can be obtained, such as formula (9).where E i R2 represents the static potential energy field generated by the i th obstacle, and n represents the number of obstacles.
The potential energy field vectors formed by lane lines, road boundaries, and stationary obstacles in the road are superimposed to obtain a comprehensive potential energy field.As shown in FIGURE 4, the distribution of the potential energy field can be easily seen.

B. KINETIC ENERGY FIELD
The kinetic energy field represents the physical field of the impact of a moving object on driving safety, which depends on the position, speed, acceleration, mass, size, and heading angle of the moving object.
Kinetic Energy Field E v : During vehicle driving, in addition to static obstacles and road boundaries, there are also more obstacles such as moving vehicles and pedestrians.When calculating the kinetic energy field of moving 109956 VOLUME 11, 2023 Authorized licensed use limited to the terms of the applicable license agreement with IEEE.Restrictions apply.obstacles, the calculation is as follows: 0 other (10) where, v represents the gain coefficient of the kinetic energy field, which depends on the mass parameter of the moving object.The greater the mass, the greater the gain coefficient k v .v is the vehicle speed, and θ is the velocity direction angle of the moving object.The kinetic energy field distribution diagram of a traffic vehicle calculated according to formula (10) is shown in FIGURE 5.

C. BEHAVIOR FIELD
The behavior field refers to the quantification of the impact of drivers' behavioral characteristics on driving safety, including drivers' physiological and psychological factors, driving skills, and driving style.Under the same conditions, the driving risk caused by aggressive drivers is higher than that caused by conservative drivers; Drivers with low driving skills often have a larger behavioral field than drivers with high driving skills.In the model, the behavioral field formed by drivers can be represented by the product of the driver risk factors that characterize the impact of driver characteristics and the kinetic energy field formed by their driving vehicles.
where, K D represents the driver risk factor vector.K D is decomposed into x-direction and y-direction is as follows: The expression for decomposing the kinetic energy field E v into x-direction and y-direction is as follows: K x D represents the driver's risk factor in the x-direction, indicating the degree of longitudinal speed fluctuation of the driver.The greater the fluctuation, the greater the K x D , and the greater risk generated; The K x D solution is shown below.
where M represents the number of data points collected in rolling mode, v represents the average speed, v max represents the maximum speed, v min represents the minimum speed, ȳ represents the average speed, y max represents the maximum speed, and y min represents the minimum speed.According to the above formulas ( 14) and ( 15), taking the US NGsim database as an example, the database obtained a public database of vehicle type, length, position, speed, acceleration, and other data through video capture in 2005.This article selects the data of three vehicles (recorded as Vehicle1, Vehicle2, and Vehicle3) during a certain distance, as shown in FIGURE 6.
According to the data of three vehicles in FIGURE 6, the K x D , K y D coefficients of the three vehicles can be obtained (recorded as D , respectively), and their behavior field distribution can be plotted using formulas (11), (12), and ( 13), as shown in FIGURE 7.

D. EDRFM MODEL SIMULATION 1) OBSTACLE VEHICLE RISK FIELD IN TYPICAL SCENARIOS
Different vehicle motion states can have an impact on the distribution of the driving risk field.In FIGURE 8   There is a static obstacle with a length of 5m and a width of 2m, and its safety field distribution presents an ellipse.The long and short axes of the ellipse depend on δ 1 and δ 2 .

b: HORIZONTAL MOVEMENT SCENE
A moving obstacle with a length of 4m and a width of 1.8m, with a speed of 8m/s and a velocity direction angle of 0 • .The right side of the vehicle's center of mass is the direction of travel, and the resulting risk field value is significantly higher than the left side of the vehicle's center of mass.This is because the vehicle's speed to the right will collide with objects on the right.The reason for the lower left side is that  it is away from objects on the left, the risk field generated on the left area is significantly smaller.

c: STEERING MOVEMENT SCENE
A moving obstacle with a length of 4m and a width of 1.8m, with a speed of 8m/s and a speed direction angle of 45 • .The value of the risk field generated in the front of the vehicle traveling direction is significantly higher than that in the rear of the vehicle traveling direction.

2) DRIVING RISK FIELDS FOR DIFFERENT VEHICLE SIZES
Next, simulate the impact of vehicle size on the driving risk field.Assuming that at a certain moment, the speed of three obstacle vehicles is 8m/s, FIGURE 9 shows the distribution of the driving risk field of three different size obstacle vehicles.As can be seen from FIGURE 9, at the same speed, the size of an obstacle vehicle will affect the distribution of the driving risk field: the wider the vehicle, the higher the lateral driving risk, and the greater the impact range; the longer the vehicle, the higher the longitudinal driving risk, and the greater the impact range.

3) DRIVING RISK FIELDS IN COMPLEX TRAFFIC SCENARIOS
This section simulates a more complex full-road traffic scenario where multiple vehicles coexist.Figure 10 shows a one-way two-lane traffic situation, with the longitudinal axis of the road being the x-axis and the lateral axis of the road being the y-axis.At a certain moment: Vehicle A (4m long and 1.6m wide) is located at (−5,6) m, with a heading angle of 0 • and a speed of 15m/s; Vehicle B (5m long and 1.8m wide) is located at (20,6) m, with a heading angle of 6 • and a speed of 10m/s; Vehicle C (4.8m long and 1.75m wide) is located at (14,2) m, with a heading angle of 45 • and a speed of 15m/s; Vehicle D (6m long and 1.8m wide) is located at (35,2) m, with a heading angle of 0 • and a speed of 20m/s; Vehicle E (length: 5m, width: 1.8m) is located at (35,5.8) m, heading angle: 0 • , speed: 16m/s; The comprehensive driving risk field under this scenario is shown in Figure 10.The comprehensive risk field is the superposition of the potential energy field and the total risk field generated by five vehicles.In the figure, the driving safety degree of each location on the road can be seen.On this basis, vehicle behavior decisionmaking and trajectory planning can be carried out to avoid high-risk areas and improve driving safety and road traffic efficiency.

III. EDRFM-BASED TRAJECTORY PLANNING AND TRAJECTORY MODIFICATION CONSIDERING VEHICLE MOTION CONSTRAINTS
In theory, trajectory planning based on the EDRFM can meet the requirements of vehicle obstacle avoidance, ensuring that the vehicle will not collide with obstacles.However, on the one hand, due to the constraints of its own shape, size, and wheel rotation angle, the curvature radius of its motion trajectory must be greater than the limit turning radius determined by the contour and wheel rotation angle during the actual vehicle movement; On the other hand, when the vehicle speed is high and the wheel angle is large, it is easy to occur side slip or tail-flick on the front axle, which seriously affects the driving safety of vehicle occupants.
Trajectory planning based solely on EDRFM cannot ensure the realizability of the planned trajectory and the requirements for the vehicle's driving stability.Therefore, it is necessary to modify the trajectory planning results based on EDRFM.Correction method: A. CALCULATE VEHICLE DYNAMICS CONSTRAINTS Calculate the vehicle limit turning radius determined by vehicle dynamics, road adhesion conditions, and vehicle Ackermann steering mechanism.
The center of mass sideslip angle β and yaw rate ω are key state parameters that describe the driving stability of a vehicle.When β is very small, the ω calculated by the 2degree-of-freedom vehicle model is stable, so the reference yaw rate ω ref can be calculated from the 2-degree-of-freedom vehicle model.The 2-degree-of-freedom vehicle model is shown in FIGURE 11 and its dynamic equation is derived as shown in Formula (16).
where: δ f -front wheel angle; l r is the distance from the center of mass of the vehicle to the front and rear axles, k 2 is the lateral stiffness of the front and rear tires of the vehicle;α 2 refers to the front and rear wheel side slip angles; v is the lateral velocity; u is the longitudinal speed; β is the centroid sideslip angle; ω r is the yaw rate, I z -Moment of inertia around the axis.When the vehicle runs stably, it also tends to a constant value of ω r , with v and ωr both being zero.Substitute it into Equation ( 16) to obtain the steady state ω r as follows: Calculate the yaw rate determined by the wheel angle limit using Equation (17).
where, δ max represents the maximum front wheel angle, and ω max 1 calculated from ( 18) is the limit under ideal conditions, which needs to be corrected according to road adhesion conditions.Under the tire adhesion limit, the lateral acceleration a y meets: wherein a y can be expressed as: In the formula, the second item is very small, and according to the literature, the first item in usually accounts for 15% [18].Therefore, according to formulas (19) and ( 20), the reference yaw rate ω max ref can be obtained: In summary, the ultimate yaw rate ω max : When the vehicle is stable, the yaw rate and turning radius approximately meet the formula: Therefore, according to Formula ( 22) and ( 23), it is preliminarily determined that the turning radius limit considering vehicle driving stability should meet the following requirements: When the vehicle is moving very slowly, there is a kinematic condition between the inner and outer wheels that allows them to turn slip-free [19].The condition is called the Ackerman condition and is expressed by: where, are the steering angle of the inner wheel and the outer wheel's steer angle?The inner and outer wheels are defined based on the turning center O.The distance between the steer axes of the steerable wheels is called the track and is shown by W .The distance between the front and rear axles is called the wheelbase and is shown by L. According to FIGURE 12, it can be obtained: Combining ( 25) and ( 26), the ultimate turning radius R min of the vehicle can be obtained: By synthesizing formulas (24) and ( 27), the R min determined by vehicle dynamics, road adhesion conditions, and vehicle Ackermann steering mechanism can be obtained: B. ARC SMOOTHING Perform arc smoothing correction on the trajectory planning curve based on the driving risk field, with the arc radius of R min .The correction method is shown in FIGURE 13.

C. TRACK PLANNING SIMULATION VERIFICATION
To verify the effectiveness of the trajectory planning algorithm, this paper sets the road adhesion coefficient to 109960 VOLUME 11, 2023 Authorized licensed use limited to the terms of the applicable license agreement with IEEE.Restrictions apply.The trajectory planning results for the 1s, 2s, 3s, 5.5s, 7s, and 13s are selected respectively.In FIGURE 14, P1 represents the trajectory planning results based on EDRFM, and P2 represents the trajectory planning results after further considering vehicle dynamics constraints based on EDRFM.The distribution of the driving risk field at each time is in FIGURE 15.From FIGURE 14 and FIGURE 15, it can be seen that vehicles can automatically avoid obstacles based on the risk field distribution of traffic scenarios, ensuring driving safety, avoiding areas with high driving risks at every moment, and selecting low-risk areas to drive, in line with the basic driving behavior of drivers.In the enlarged image of E1, it can be seen that there is a significant difference between the P1 trajectory and the P2 trajectory near x = 70m, mainly because the P1 trajectory obtained without considering vehicle dynamics constraints can cause a large side slip of the vehicle.The trajectory P2 obtained by solving the dynamics constraints in this paper has a significantly larger curvature than P1, ensuring that the vehicle does not experience a serious side slip, ensuring the rationality of the planning; Similarly, from the enlarged view of E2, it can be seen that between x = 90m, the curvature of the P2 trajectory is significantly greater than that of P1, ensuring the feasibility of the vehicle's motion trajectory and driving safety.

IV. CONCLUSION
Aiming at the decision-making problem of the safe driving trajectory of the autonomous vehicle, this paper innovatively proposes the EDRFM based on the analysis of the shortcomings of existing DRFM models.The model introduces geometric dimensions, heading angle, and driving behavior characteristics information, and provides a basis for determining geometric dimensions and shape parameters, expanding the accuracy and applicability of the driving risk field model.Based on typical traffic scenarios, the distribution of driving risk fields under different vehicle motion states and sizes are obtained, verifying the effectiveness of the proposed model in vehicle driving safety assessment.On this basis, the EDRFM model is applied to solve the vehicle trajectory planning problem.Aiming at the problem that the target trajectory P1 obtained from the EDRFM model may be impossible to achieve and the vehicle may become unstable, this paper further proposes a trajectory arc correction method considering vehicle dynamics constraints.Based on the analysis of the vehicle limit turning radius R min determined by vehicle dynamics, road adhesion conditions, and vehicle Ackermann steering mechanism by establishing a vehicle two-degree of freedom model, a circular arc smoothing algorithm is further used to correct P1 to ensure that the planned trajectory meets the vehicle's driving stability and feasibility.In the trajectory planning simulation verification, based on the real driving data of NGsim, this paper selects a complex multi-vehicle coexistence full road traffic scenario and performs trajectory planning for one of the vehicles.The results show that the EDRFM model considering vehicle dynamics constraints can effectively deal with the problems of unrealizable trajectory and vehicle instability caused by excessive local curvature, which proves that the trajectory planning method in this paper has a good application range.

FIGURE 1 .
FIGURE 1.The technical route of this article.
) is added to the expression.L, W represents the length and width of the vehicle, respectively, β represents the gain coefficient of obstacle potential energy field, δ 1 , δ 2 represent control coefficients for vehicle length and width, respectively, and are used to control the impact range of the risk field in both the horizontal and vertical directions.Their values have an important impact on vehicle trajectory planning.Reasonably determine the values of δ 1 and δ 2 is of great significance.When there are static obstacles in front of the vehicle, the vehicle usually performs lane change and obstacle avoidance operations, and the obstacle avoidance trajectory generally conforms to a quintic polynomial curve.Therefore, this article determines and is δ 2 based on the lane change quintic polynomial curve.Assuming that the starting area position of the vehicle for obstacle avoidance is A 1 = (x 1 , y 1 ), the position for completing the lane change back to the original lane is B 2 = (x 1 , y 2 ), and the vertex position of the lane change is C 3 = (x 2 , y 0 ), during the obstacle avoidance process, the longitudinal driving distance is d = |y 2 − y 0 | = |y 1 − y 0 |, and the lateral driving distance is b = |x 2 −x 1 |.During the obstacle avoidance process, the potential energy field generated by the static obstacle on the obstacle avoidance trajectory curve should tend to zero, so it can be preliminarily determined:

FIGURE 3 .
FIGURE 3. Schematic diagram for determining the potential energy field parameters δ 1 and static obstacles.

FIGURE 5 .
FIGURE 5. Distribution diagram of the kinetic energy field of traffic vehicles.

D
represents the driver's risk factor in the y-direction, indicating the lateral fluctuation degree of the driver during driving.The greater the fluctuation, the greater the K y D , and the greater the risk generated.

FIGURE 6 .
FIGURE 6. Lateral position and speed of 3 vehicles.

FIGURE 7 .
FIGURE 7. Distribution diagram of the behavior field of three vehicles.

FIGURE 8 .
FIGURE 8. Obstacle vehicle risk field in typical scenarios.

FIGURE 9 .
FIGURE 9. Distribution of vehicle risk field with different obstacle sizes.

FIGURE 10 .
FIGURE 10.Comprehensive risk field under complex traffic scenarios.

FIGURE 12 .
FIGURE 12.A front-wheel-steering vehicle and the Ackerman condition.

FIGURE 13 .
FIGURE 13.Schematic diagram of trajectory correction considering vehicle dynamics constraints.

FIGURE 14 .
FIGURE 14. Track planning results in complex traffic scenarios.

FIGURE 15 .
FIGURE 15.Distribution of driving risk field at various times under complex traffic scenarios.
NING SUN received the master's degree in vehicle engineering from the School of Automotive Engineering, Jilin University, in 2016, where he is currently pursuing the Ph.D. degree in the Research Group of Academician Guo Konghui and Prof. Xu Nan with the State Key Laboratory of Automotive Simulation and Control.His research interests include autonomous driving technology and the field of intelligent vehicle decision-making and planning.MIN HU received the B.S. degree in automobile service engineering from Inner Mongolia University, Hohhot, China, in 2020, and M.S. degree in automotive engineering from Jilin University, Changchun, China.She is currently pursuing the Ph.D. degree.Her current research interests include vehicle dynamics and stability control and autonomous vehicle motion planning and control.ZEYANG ZHANG received the master's degree in automotive engineering from the Wuhan University of Technology, Wuhan, China, in 2016.He is currently a High-Energy Research and Development Manager with the Technical Center, Dongfeng Motor Group Corporation Ltd., Wuhan.His research interests include vehicle dynamics and control, in-wheel motor, and project management.HANG DU received the B.S. degree in automotive engineering from Jilin University, Changchun, China, in 2020.He is currently pursuing the M.S. degree.His current research interest includes intelligent vehicle decision-making and planning.SHENGXUAN ZHAO received the B.S. degree from the College of Mechanical and Electronic Engineering, China University of Petroleum, Qingdao, China, in 2022.He is currently pursuing the M.S. degree.His current research interests include trajectory prediction and intelligent vehicle decision-making and planning.

TABLE 1 .
Comparison of typical trajectory planning algorithms.