Space-time corridor based intelligent connected vehicle trajectory planning method combining real-time aggressiveness

Abstract

The application relates to the technical field of automatic driving and discloses an intelligent networked vehicle trajectory planning method based on a space-time corridor combined with real-time aggressiveness. The intelligent networked vehicle trajectory planning method considers other vehicles in a dynamic traffic environment and their predicted trajectories by constructing a space-time corridor, and real-time driving behavior information of the other vehicles is integrated into trajectory planning of a host vehicle. On this basis, a 7th-degree polynomial method based on sampling is used to generate a path, and a sequence quadratic programming optimization technology with soft constraints is combined to balance safety, efficiency and comfort of the trajectory. In this way, the application can effectively predict and respond to aggressive changes in the traffic environment, adjust the trajectory in real time, ensure safety and comfort during vehicle driving, so as to handle complex urban traffic environments, ensure that the host vehicle can dynamically adjust its driving path under different driving scenarios to respond to changes of surrounding vehicles and obstacles, and provide more efficient and stable driving experience.

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving technology, and in particular to a trajectory planning method for intelligent connected vehicles based on space-time corridors combined with real-time aggressiveness. Background Art

[0002] In recent years, the rapid development of autonomous driving technology has driven the advancement of intelligent transportation, with the core goal of improving driving safety, convenience, and road efficiency. Some features of autonomous driving technology have been widely implemented in mass-produced vehicles, using intelligent systems to assist drivers and reduce the incidence of traffic accidents caused by human error. Lane changing, a highly dynamic and time-sensitive driving task, is a high-incidence scenario for traffic accidents and a key challenge facing autonomous driving technology. During a lane change, autonomous vehicles must not only comprehensively assess the traffic risks associated with the current and target lanes but also consider the dynamic uncertainties of the surrounding environment to plan a safe, feasible, and efficient trajectory. Traditional lane-changing methods typically divide the process into two phases: decision-making and motion planning. The decision phase assesses potential collision risks during the lane change, while the motion planning phase generates the vehicle's future trajectory within the planned timeframe. However, existing lane-changing algorithms often suffer from issues such as inconsistency between decision-making and motion planning, insufficient real-time performance, and difficulty handling complex traffic scenarios. These issues significantly limit the applicability of autonomous driving technology in real-world traffic environments.

[0003] In existing technologies, lane-changing safety assessment methods based on risk indicators, spatiotemporal path search, or reverse reachability sets are often used in the decision-making phase. However, some methods fail to adequately consider lane-changing feasibility and reachability, leading to inconsistencies between lane-changing decisions and motion planning. Even analysis methods based on reachability sets involve complex mathematical operations and impose a high computational burden. In the field of motion planning, existing methods primarily generate candidate lateral and longitudinal motion plans using predefined mathematical functions or sampling-based approaches, and then evaluate these plans using optimization methods. However, sampling-based motion planning methods are often limited by predefined motion patterns and lack flexibility in adapting to diverse dynamic traffic scenarios. While trajectory planning methods based on numerical optimization can generate high-quality continuous solutions, their computational complexity increases significantly with increasing state dimensions, making them difficult to meet the real-time requirements of practical applications. To address these issues, researchers have proposed a hybrid planning method that combines sampling and optimization. By incorporating risk assessment mechanisms based on spatiotemporal corridors and dynamic environments, this method enhances the feasibility analysis capabilities of lane-changing decisions while maintaining both flexibility and computational efficiency in trajectory planning. However, current research still has shortcomings in dynamic trajectory generation, diverse motion pattern representation, and feasibility exploration of the solution space. There is an urgent need for a hybrid planning method that can effectively combine sampling and optimization techniques to achieve safe and efficient lane changing operations for autonomous vehicles in complex and dynamic traffic environments. Summary of the Invention

[0004] In order to solve the above technical problems, the present invention provides a trajectory planning method for intelligent connected vehicles based on spatiotemporal corridors combined with real-time aggressiveness. The technical solution is as follows:

[0005] S1. Establish a spatial set bounded by the road edge as the representation area of ​​the ego vehicle reachable set. Iterate the ego vehicle reachable set using the step size planned by the ego vehicle algorithm. The specific expression is as follows:

[0006] V={P L ,P R ,P L,i ,P R,j ,T plan}

[0007]

[0008] a min ≤a(t)≤a max δ min ≤δ(t)≤δ max

[0009] Among them, V is the space set, P L ,P R ,P L,i ,PR,j ,T plan denote the left line of the lane where the ego vehicle is located, the right line of the lane where the ego vehicle is located, the i-th left edge line of the lane where the ego vehicle is located, the j-th right edge line of the lane where the ego vehicle is located, and the time step of the spatial set (0.5s). R(x0,t) is the reachable set of the ego vehicle. X(t), x(t), y(t), θ(t), v(t), u(t), a(t), δ(t) denote the state vector of the vehicle at time t, the x-coordinate position of the vehicle in the global coordinate system at time t, the y-coordinate position of the vehicle in the global coordinate system at time t, the heading angle of the vehicle at time t, the speed of the vehicle at time t, the control input vector, the acceleration of the vehicle at time t, and the steering angle of the vehicle at time t, respectively.

[0010] S2. Based on the vehicle's kinematic formula, the ego vehicle reachable set is iteratively constructed with the step size of the ego vehicle trajectory planning algorithm to achieve the initial trajectory planning of the ego vehicle in dynamic uncertain scenarios. The upper and lower bounds of the constructed reachable set are expressed as follows:

[0011]

[0012] r∈[0,1]

[0013]

[0014] r∈[0,1]

[0015] where y up and y down They represent the longitudinal positions of the upper and lower boundaries of the vehicle in trajectory planning after a given time interval Δt. up and x down where y0 and x0 represent the upper and lower lateral positions of the vehicle in trajectory planning after a given time interval Δt. y0 and x0 represent the initial longitudinal and lateral positions of the vehicle, respectively. and Represent the initial longitudinal and lateral velocities of the vehicle, respectively. x,max and v x,min Respectively represent the maximum and minimum lateral speeds allowed for the vehicle, which are used to calculate the effective longitudinal acceleration to ensure that the vehicle does not exceed the lateral speed limit when turning. y,max and a y,min Respectively represent the maximum and minimum longitudinal acceleration allowed for the vehicle. x,max and a x,min Respectively represent the maximum and minimum lateral acceleration allowed for the vehicle. y,max,eff and a y,min,effThe values ​​represent the vehicle's maximum and minimum effective longitudinal acceleration, respectively, after accounting for the vehicle's speed limit. These values ​​are used to fully account for the vehicle's lateral acceleration limit during cornering, ensuring the vehicle does not exceed its physical limitations. r is the acceleration coupling distribution coefficient, which adjusts the vehicle's lateral and longitudinal acceleration limits during cornering. When r approaches 0, the vehicle barely turns. Increasing r indicates a greater curvature during cornering, which reduces the effective longitudinal acceleration.

[0016] S3. Obtain the obstacle vehicle action set based on the obstacle vehicle driving information. The specific expression is as follows:

[0017] Γ={K m ,K a ,K d ,K ha ,K hd ,K cl ,K cr}

[0018] Among them, K m To maintain speed, K a is accelerated and K a ≤1.25m / s 2 , K d is deceleration and K d ≥-1m / s 2 , K ha For rapid acceleration and K ha ≤2.5m / s 2 , K hd For rapid deceleration and K hd ≥-2m / s 2 , K cl and K cr Respectively represent lateral movement to the left and right and the lateral speed |v x |≤2m / s.

[0019] S4. Based on the real-time driving style of the obstacle vehicle, the overall cost function is constructed as the price function. The specific expression is as follows:

[0020] Longitudinal factors, including acceleration and deceleration intensity:

[0021]

[0022] Lateral factors, specifically lane change intensity:

[0023]

[0024] Speed ​​fluctuation, including speed standard deviation and jerk:

[0025]

[0026] Safety time interval considering minimum safety distance:

[0027]

[0028] Overall cost function design:

[0029] J=J long +J lat +J speed +J safe

[0030] Where a(t) is the acceleration of the obstacle vehicle at time t, a d (t) is the lateral acceleration of the obstacle vehicle at time t, d(t) is the lateral displacement of the obstacle vehicle at time t, σ v is the standard deviation of the obstacle vehicle’s velocity, j(t) is the jerk of the obstacle vehicle at time t, T min is the minimum reaction time of the obstacle vehicle driver and T min =1.5s, T safe is the safety time interval, t r is the reaction time of the autonomous driving system, J long is the cost function of the longitudinal factor, J lat is the cost function of the horizontal factor, J speed is the cost function of speed fluctuation, J safe is the cost function of the safety time interval considering the minimum safety distance, J is the overall cost function, and λ1,λ2,...,λ7 are the undetermined weight coefficients of the overall cost function.

[0031] S5. Establish a three-degree-of-freedom vehicle dynamics model for the vehicle. The specific expression is as follows:

[0032]

[0033] where v x (t) and v y (t) represents the change of velocity in the x and y directions with time, r(t) represents the change of rotation velocity around the z axis with time, Φ(t) represents the change of angle or angular displacement with time, F x represents the force in the x direction acting on the vehicle, F y represents the force in the y direction acting on the vehicle, m represents the mass of the vehicle, M z represents the moment in the z direction acting on the vehicle, I z represents the vehicle's moment of inertia around the z-axis.

[0034] S6. Construct the global trajectory of the vehicle based on polynomial sampling. The specific expression is as follows:

[0035] P(t)=a0+a1t+a2t2 +a3t 3 +a4t 4 +a5t 5 +a6t 6 +a7t 7

[0036] Where P(x) is the position at time t, a0 is the initial position, a1 is the velocity-related coefficient, a2 is the acceleration-related coefficient, a3 is the jerk-related coefficient, and a4, ..., a7 are control coefficients for higher-order rates of change.

[0037] S7. Establish a sequential quadratic programming with soft constraints for the ego vehicle to optimize trajectory and speed. The specific expression is as follows:

[0038]

[0039] where v k is the velocity at the kth moment, v ref,k is the target speed of path planning, a k is the acceleration at the kth moment, a ref,k is the target acceleration of path planning, a max is the maximum acceleration, a min is the minimum acceleration, w1, w2 are weight factors to balance the optimization objectives of speed and acceleration, w3, w4 are penalty factors used to control the penalty intensity for exceeding the acceleration range, and N is the number of steps in the prediction domain.

[0040] S8. Ensure the vehicle trajectory tracking control accuracy based on MPC rolling optimization. The specific expression is as follows:

[0041]

[0042] Where R is the weighted matrix of the control input, Q is the weighted matrix of the state error, and Q f is the weighted matrix of the terminal state error, x ref,k is the expected state of the reference trajectory at step k.

[0043] Beneficial effects of the present invention

[0044] Compared with the prior art, the present invention has the following significant advantages:

[0045] 1. Improved accuracy: Joint optimization of time and path ensures optimal overall coordination between path and time, avoiding conflicts or inconsistencies that may arise when path and time are planned separately.

[0046] 2. Optimization performance: The use of trajectory planning and multi-constraint optimization strategies can generate the optimal trajectory while satisfying multiple constraints (such as dynamic constraints and control input constraints), significantly improving the efficiency and stability of system operation.

[0047] 3. Enhanced adaptability: Based on controller optimization design and model predictive control (MPC) algorithm, the present invention can achieve real-time control of complex systems in dynamic environments and has strong environmental adaptability and robustness. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 Schematic diagram of the process of the intelligent connected vehicle trajectory planning method in the present invention;

[0049] Figure 2 Schematic diagram of the reachable set of vehicles in the present invention;

[0050] Figure 3 Schematic diagram of the space-time corridor combining real-time aggression in the present invention;

[0051] Figure 4 This is a schematic diagram of the kinetic modeling in the present invention;

[0052] Figure 5 This is a schematic diagram of the optimal speed curve in the ST diagram of the present invention;

[0053] Figure 6 This is a graph showing the curvature change of the optimal lane-changing path for active obstacle avoidance of the vehicle in the present invention; DETAILED DESCRIPTION

[0054] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0055] See also Figure 1 ,A trajectory planning method for an intelligent connected vehicle based on spatiotemporal corridor and real-time aggression, including the following steps:

[0056] S1. Establish a spatial set bounded by the road edge as the representation area of ​​the ego vehicle reachable set. Iterate and establish the reachable set based on the step size of the ego vehicle trajectory planning algorithm. The specific expression is as follows:

[0057] V={P L ,P R ,P L,i ,P R,j ,T plan}

[0058]

[0059] a min ≤a(t)≤a max δ min ≤δ(t)≤δ max

[0060] Among them, V is the space set, P L ,P R ,P L,i ,P R,j ,T plan where x0, t, y0, t, a0, t, and δ0 represent the vehicle’s state vector at time t, the vehicle’s x-coordinate position in the global coordinate system at time t, the vehicle’s y-coordinate position in the global coordinate system at time t, the vehicle’s heading angle at time t, the vehicle’s speed at time t, the control input vector, the vehicle’s acceleration at time t, and the vehicle’s steering angle at time t.

[0061] S2. Based on the vehicle's kinematic formula, the vehicle's reachable set is iteratively constructed using the step size of the ego vehicle trajectory planning algorithm to achieve the initial ego vehicle trajectory planning in dynamic uncertain scenarios. Figure 2 , the specific expressions of the upper and lower boundaries of the constructed reachable set are as follows:

[0062]

[0063] r∈[0,1]

[0064]

[0065] r∈[0,1]

[0066] where y up and y down They represent the longitudinal positions of the upper and lower boundaries of the vehicle in trajectory planning after a given time interval Δt. up and x down where y0 and x0 represent the upper and lower lateral positions of the vehicle in trajectory planning after a given time interval Δt. y0 and x0 represent the initial longitudinal and lateral positions of the vehicle, respectively. and Represent the initial longitudinal and lateral velocities of the vehicle, respectively. x,max and v x,minRespectively represent the maximum and minimum lateral speeds allowed for the vehicle, which are used to calculate the effective longitudinal acceleration to ensure that the vehicle does not exceed the lateral speed limit when turning. y,max and a y,min Respectively represent the maximum and minimum longitudinal acceleration allowed for the vehicle. x,max and a x,min Respectively represent the maximum and minimum lateral acceleration allowed for the vehicle. y,max,eff and a y,min,eff The values ​​represent the vehicle's maximum and minimum effective longitudinal acceleration, respectively, after accounting for the vehicle's speed limit. These values ​​are used to fully account for the vehicle's lateral acceleration limit during cornering, ensuring the vehicle does not exceed its physical limitations. r is the acceleration coupling distribution coefficient, which adjusts the vehicle's lateral and longitudinal acceleration limits during cornering. When r approaches 0, the vehicle barely turns. Increasing r indicates a greater curvature during cornering, which reduces the effective longitudinal acceleration.

[0067] By calculating the upper and lower boundaries of the reachable set, we can determine the vehicle’s position in T plan The range of the drivable area within a certain time period can help the detection system detect whether the vehicle will cross the safety boundary in the future and determine whether there is a risk of collision.

[0068] S3. Obtain the obstacle vehicle action set based on the obstacle vehicle driving information. The specific expression is as follows:

[0069] Γ={K m ,K a ,K d ,K ha ,K hd ,K cl ,K cr}

[0070] Among them, K m To maintain speed, K a is accelerated and K a ≤1.25m / s 2 , K d is deceleration and K d ≥-1m / s 2 , K ha For rapid acceleration and K ha ≤2.5m / s 2 , K hd For rapid deceleration and K hd ≥-2m / s 2 , K cl and K cr Respectively represent lateral movement to the left and right and the lateral speed |v x |≤2m / s.

[0071] S4. Based on the real-time driving style of the obstacle vehicle, the overall cost function is constructed as the price function. The specific expression is as follows:

[0072] Longitudinal factors, including acceleration and deceleration intensity:

[0073]

[0074] Lateral factors, specifically lane change intensity:

[0075]

[0076] Speed ​​fluctuation, including speed standard deviation and jerk:

[0077]

[0078] Safety time interval considering minimum safety distance:

[0079]

[0080] Overall cost function design:

[0081] J=J long +J lat +J speed +J safe

[0082] Where a(t) is the acceleration of the obstacle vehicle at time t, a d (t) is the lateral acceleration of the obstacle vehicle at time t, d(t) is the lateral displacement of the obstacle vehicle at time t, σ v is the standard deviation of the obstacle vehicle’s velocity, j(t) is the jerk of the obstacle vehicle at time t, T min is the minimum reaction time of the obstacle vehicle driver and T min =1.5s, T safe is the safety time interval, t r is the reaction time of the autonomous driving system, J long is the cost function of the longitudinal factor, J lat is the cost function of the horizontal factor, J speed is the cost function of speed fluctuation, J safe is the cost function of the safety time interval considering the minimum safety distance, J is the overall cost function, and λ1,λ2,...,λ7 are the undetermined weight coefficients of the overall cost function.

[0083] In this method, the speed standard deviation is calculated as:

[0084]

[0085] Where N is the number of steps in the prediction time domain, is the average speed of the urban road driving mode.

[0086] In this method, the derivation process of the safety time interval formula is:

[0087]

[0088] where d min The minimum safe distance.

[0089] By establishing an overall cost function, it is possible to track and calculate the real-time aggressiveness of different obstacle vehicles in the spatial set, and combine it with the reachable set to form a spatiotemporal corridor based on the real-time aggressiveness. This method considers other vehicles in a dynamic traffic environment and their predicted trajectories by constructing a spatiotemporal corridor, and integrates their driving behavior information into the trajectory planning of the main vehicle in real time. Figure 3 .

[0090] S5. Establish a three-degree-of-freedom vehicle dynamics model. Figure 4 , its specific expression is as follows:

[0091]

[0092] where v x (t) and v y (t) represents the change of velocity in the x and y directions with time, r(t) represents the change of rotation velocity around the z axis with time, Φ(t) represents the change of angle or angular displacement with time, F x represents the force in the x direction acting on the vehicle, F y represents the force in the y direction acting on the vehicle, m represents the mass of the vehicle, M z represents the moment in the z direction acting on the vehicle, I z represents the vehicle's moment of inertia around the z-axis.

[0093] S6. The ego vehicle constructs a global trajectory based on polynomial sampling. The specific expression is as follows:

[0094] P(t)=a0+a1t+a2t 2 +a3t 3 +a4t 4 +a5t 5 +a6t 6 +a7t 7

[0095] Where P(x) is the position at time t, a0 is the initial position, a1 is the velocity-related coefficient, a2 is the acceleration-related coefficient, a3 is the jerk-related coefficient, and a4, ..., a7 are control coefficients for higher-order rates of change.

[0096] Polynomial constraints

[0097]

[0098] where x s ,x e are the initial and end positions of the trajectory planning, v s ,v e are the initial and terminal velocities of the trajectory planning, a s ,a e are the initial and terminal accelerations of the trajectory planning, j s ,j e are the initial and terminal accelerations of the trajectory planning,

[0099] S7. Establish a sequential quadratic program with soft constraints for optimizing trajectory and speed. Figure 5 , Figure 6 , its specific expression is as follows:

[0100]

[0101] where v k is the velocity at the kth moment, v ref,k is the target speed of path planning, a k is the acceleration at the kth moment, a ref,k is the target acceleration of path planning, a max is the maximum acceleration, a min is the minimum acceleration, w1, w2 are weight factors to balance the optimization objectives of speed and acceleration, w3, w4 are penalty factors used to control the penalty intensity for exceeding the acceleration range, and N is the number of steps in the prediction domain.

[0102] S8. Establishing an MPC-based rolling optimization to ensure the vehicle trajectory tracking control accuracy is characterized by the following specific expression:

[0103]

[0104] Where R is the weighted matrix of the control input, Q is the weighted matrix of the state error, and Q f is the weighted matrix of the terminal state error, x ref,i is the expected state of the reference trajectory at step i.

[0105] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A trajectory planning method for intelligent connected vehicles based on spatiotemporal corridors combined with real-time aggressiveness, characterized in that: The following steps are involved: S1. Establish a spatial set bounded by the road edge as the representation area of ​​the vehicle's reachable set; S2. Based on the vehicle's kinematic formula, iteratively construct the ego vehicle reachable set using the step size of the ego vehicle trajectory planning algorithm; S3, calculating and obtaining an obstacle vehicle action set based on the obstacle vehicle driving information; S4, judging the real-time driving style based on the obstacle vehicle's actions and constructing an overall cost function as a price function; The obstacle vehicle action includes longitudinal factors, lateral factors, speed fluctuations, and safety time intervals considering the minimum safety distance; The cost function of the longitudinal factor specifically includes the acceleration and deceleration intensity, and its expression is as follows: The cost function of the lateral factor specifically includes the lane change intensity, which is expressed as follows: The cost function of speed fluctuation specifically includes speed standard deviation and acceleration, and its expression is as follows: The cost function of the safety time interval considering the minimum safety distance is expressed as follows: The overall cost function is designed as: J=J long +J lat +J speed +J safe Where a(t) is the acceleration of the obstacle vehicle at time t, a d (t) is the lateral acceleration of the obstacle vehicle at time t, d(t) is the lateral displacement of the obstacle vehicle at time t, σ v is the standard deviation of the obstacle vehicle’s velocity, j(t) is the jerk of the obstacle vehicle at time t, T min is the minimum reaction time of the obstacle vehicle driver, T safe is the safety time interval, t r is the reaction time of the autonomous driving system, J long is the cost function of the longitudinal factor, J lat is the cost function of the horizontal factor, J speed is the cost function of speed fluctuation, J safe is the cost function of the safety time interval considering the minimum safety distance, J is the overall cost function, and λ1,λ2,...,λ7 are the undetermined weight coefficients of the overall cost function; S5. Establish a three-degree-of-freedom vehicle dynamics model for the vehicle; S6, constructing the global trajectory of the ego vehicle based on polynomial sampling; S7. Establish a sequential quadratic program with soft constraints for the ego vehicle to optimize trajectory and speed; The expression of the sequential quadratic programming of the ego vehicle with soft constraints is as follows: where v k is the velocity at the kth moment, v ref,k is the target speed of path planning, a k is the acceleration at the kth moment, a ref,k is the target acceleration of path planning, a max is the maximum acceleration, a min is the minimum acceleration, w1, w2 are weight factors to balance the optimization objectives of speed and acceleration, w3, w4 are penalty factors to control the penalty intensity for exceeding the acceleration range, and N is the number of steps in the prediction domain; S8, ensure trajectory tracking control accuracy based on MPC rolling optimization.

2. The intelligent connected vehicle trajectory planning method according to claim 1, characterized in that: The expressions of the spatial set V bounded by the road edge and the vehicle reachable set R(x0,t) are as follows: V={P L ,P R ,P L,i ,P R,j ,T plan } Among them, V is the space set, P L ,P R ,P L,i ,P R,j ,T plan denote the left line of the lane where the ego vehicle is located, the right line of the lane where the ego vehicle is located, the i-th left edge line of the lane where the ego vehicle is located, the j-th right edge line of the lane where the ego vehicle is located, and the time step of the spatial set, respectively. R(x0,t) is the reachable set of the ego vehicle, X(t), x(t), y(t), θ(t), v(t), u(t), a(t), δ(t) denote the state vector of the vehicle at time t, the x-coordinate position of the vehicle in the global coordinate system at time t, the y-coordinate position of the vehicle in the global coordinate system at time t, the heading angle of the vehicle at time t, the speed of the vehicle at time t, the control input vector, the acceleration of the vehicle at time t, and the steering angle of the vehicle at time t, respectively.

3. The intelligent connected vehicle trajectory planning method according to claim 1, characterized in that: The expressions for the upper and lower bounds of the self-car reachable set are as follows: where y up and y down They represent the longitudinal positions of the upper and lower boundaries of the vehicle in trajectory planning after a given time interval Δt; up and x down They represent the lateral positions of the upper and lower boundaries of the vehicle in trajectory planning after a given time interval Δt; y0 and x0 represent the initial longitudinal and lateral positions of the vehicle, respectively. and They represent the initial longitudinal and lateral velocities of the vehicle respectively; v x,max and v x,min Respectively represent the maximum and minimum lateral speeds allowed for the vehicle, which are used to calculate the effective longitudinal acceleration to ensure that the vehicle does not exceed the lateral speed limit when turning; y,max and a y,min Respectively represent the maximum and minimum longitudinal acceleration allowed for the vehicle; a x,max and a x,min Respectively represent the maximum and minimum lateral acceleration allowed for the vehicle; a y,max,eff and a y,min,eff They represent the effective maximum and minimum longitudinal accelerations of the vehicle after taking into account the vehicle speed limit, respectively. They are used to fully consider the lateral acceleration limit of the vehicle when turning and ensure that the vehicle does not exceed its physical limitations. r is the acceleration coupling distribution coefficient, which is used to adjust the lateral and longitudinal acceleration limits of the vehicle when turning. When r is close to 0, the vehicle hardly turns. When r increases, it means that the vehicle's turning curvature becomes larger, which can reduce the effective longitudinal acceleration.

4. The intelligent connected vehicle trajectory planning method according to claim 1, characterized in that: The expression of the obstacle vehicle action set Γ is as follows: Γ={K m ,K a ,K d ,K ha ,K hd ,K cl ,K cr } Among them, K m To maintain speed, K a is accelerated and K a ≤1.25m / s 2 , K d is deceleration and K d ≥-1m / s 2 , K ha For rapid acceleration and K ha ≤2.5m / s 2 , K hd For rapid deceleration and K hd ≥-2m / s 2 , K cl and K cr Respectively represent lateral movement to the left and right and the lateral speed |v x |≤2m / s.

5. The intelligent connected vehicle trajectory planning method according to claim 1, characterized in that: The expression of the three-degree-of-freedom vehicle dynamics model is as follows: where v x (t) and v y (t) represents the change of velocity in the x and y directions with time, r(t) represents the change of rotation velocity around the z axis with time, Φ(t) represents the change of angle or angular displacement with time, F x represents the force in the x direction acting on the vehicle, F y represents the force in the y direction acting on the vehicle, m represents the mass of the vehicle, M z represents the moment in the z direction acting on the vehicle, I z represents the vehicle's moment of inertia around the z-axis.

6. The intelligent connected vehicle trajectory planning method according to claim 1, characterized in that: The expression of the global trajectory of the ego vehicle is as follows: P(t)=a0+a1t+a2t 2 +a3t 3 +a4t 4 +a5t 5 +a6t 6 +a7t 7 Where P(t) is the position at time t, a0 is the initial position, a1 is the velocity-related coefficient, a2 is the acceleration-related coefficient, a3 is the jerk-related coefficient, and a4, ..., a7 are control coefficients for higher-order rates of change.

7. The intelligent connected vehicle trajectory planning method according to claim 1, characterized in that: The expression based on MPC rolling optimization is as follows: Where R is the weighted matrix of the control input, Q is the weighted matrix of the state error, and Q f is the weighted matrix of the terminal state error, x ref,k is the expected state of the reference trajectory at step i.

Images