A Method and System for Trajectory Planning of Autonomous Vehicles Based on Dynamic Safety Buffer

Abstract

This invention discloses a method and system for trajectory planning of autonomous vehicles based on dynamic safety buffers. The method first acquires the state information of the autonomous vehicle and surrounding obstacles; then, it establishes state equations and multi-objective cost functions based on the vehicle's kinematic model. The core of the method lies in dynamically calculating the safety buffer distance based on the collision time between the autonomous vehicle and obstacles, and constructing adaptive collision constraints accordingly. Constraints are integrated into the cost function through an exponential obstacle function, and an efficient iterative solution is achieved using a constraint-iterative linear quadratic regulator algorithm, ultimately outputting the optimal trajectory and control sequence. The system includes corresponding modules for environmental perception, state modeling and optimization objectives, dynamic safety buffer calculation, constraint processing and optimization solution, and trajectory output. This invention achieves adaptive adjustment of the safety distance according to the risk level, improving driving efficiency and comfort while ensuring safety, and meeting the real-time requirements of autonomous driving through an efficient optimization framework.

Description

Technical Field

[0001] This invention relates to a method and system for trajectory planning of autonomous vehicles based on dynamic safety buffers, and belongs to the field of autonomous driving trajectory planning technology. Background Technology

[0002] With the rapid development of autonomous driving technology, trajectory planning, as a core component, directly impacts the safety, comfort, and efficiency of vehicle operation. Traditional trajectory planning methods often employ fixed safety thresholds for collision constraints, which frequently exhibit insufficient adaptability in dynamic traffic environments: they are overly conservative in low-risk scenarios, reducing traffic efficiency; while in high-risk scenarios, they may cause safety hazards due to insufficient safety margins.

[0003] In existing technologies, the Constrained Iterative Linear Quadratic Regulator (CILQR) has been widely used in vehicle trajectory optimization. By iteratively linearizing and quadratic approximating the nonlinear system, it can achieve real-time planning while considering vehicle dynamics constraints. However, traditional CILQR methods typically use fixed safety distances in obstacle avoidance constraints, which cannot be adaptively adjusted according to real-time traffic risks, thus limiting their overall performance in complex dynamic scenarios.

[0004] Therefore, how to design a trajectory planning method that can dynamically adjust the safety boundary according to real-time risks while ensuring the real-time performance, smoothness, and safety of the planning has become a technical problem that urgently needs to be solved in the field of autonomous driving. Summary of the Invention

[0005] The technical problem to be solved by this invention is to provide a method and system for trajectory planning of autonomous vehicles based on dynamic safety buffer. By introducing a dynamic safety buffer distance based on time-to-collision (TTC), an adaptive risk level collision constraint is constructed, and the efficient constraint iterative linear quadratic regulator (CILQR) algorithm is used to solve the problem. This achieves the goal of effectively improving the smoothness, comfort and traffic efficiency of the trajectory while ensuring safety.

[0006] To solve the above-mentioned technical problems, the present invention adopts the following technical solution: The trajectory planning method for autonomous vehicles based on dynamic safety buffers includes the following steps: Step S1: Obtain the state information and environmental perception information of the autonomous vehicle. The state information includes position, speed, heading angle, acceleration and steering angle information. The environmental perception information includes the position, speed and size information of surrounding static and dynamic obstacles. Step S2: Based on state information, environmental perception information and vehicle kinematics bicycle model, construct vehicle discrete state transition equations and establish an optimization cost function with trajectory tracking cost, speed cost and control cost as objectives; Step S3: Calculate the collision time based on the relative distance and relative speed between the autonomous vehicle and each obstacle, and dynamically adjust the safe buffer distance between the autonomous vehicle and the obstacle according to the collision time to generate a dynamic safe buffer zone. Step S4: Introduce the dynamic safety buffer into the collision constraint conditions to construct a collision avoidance function containing the dynamic safety buffer; perform constraint transformation on the collision avoidance function using an exponential barrier function, and introduce an optimization cost function to form a constrained comprehensive cost function; Step S5: Based on the constrained iterative linear quadratic regulator algorithm, the discrete state transition equation of the vehicle and the constrained comprehensive cost function are linearized and approximated quadratically to construct a linear quadratic optimal control problem. The optimal control quantity is solved by reverse recursion, and the vehicle state sequence and control sequence are updated in a forward rolling manner to output the optimal vehicle trajectory and corresponding control sequence that meet the requirements of safety, driving efficiency and comfort.

[0007] This invention also proposes an autonomous vehicle trajectory planning system based on dynamic safety buffers, comprising: The environmental perception module is used to acquire the state information and environmental perception information of the autonomous vehicle. The state information includes position, speed, heading angle, acceleration and steering angle information, and the environmental perception information includes the position, speed and size information of surrounding static obstacles and dynamic obstacles. The State Modeling and Optimization Objective Module is used to construct discrete state transition equations for a vehicle based on state information, environmental perception information, and a bicycle kinematics model, and to establish an optimization cost function with trajectory tracking cost, speed cost, and control cost as objectives. The dynamic safety buffer calculation module is used to calculate the collision time based on the relative distance and relative speed between the autonomous vehicle and each obstacle, and dynamically adjust the safety buffer distance between the autonomous vehicle and the obstacle according to the collision time to generate a dynamic safety buffer. The constraint processing module is used to introduce the dynamic safety buffer into the collision constraint conditions, construct a collision avoidance function containing the dynamic safety buffer, perform constraint transformation on the collision avoidance function using an exponential barrier function, and introduce an optimization cost function to form a constrained comprehensive cost function. The optimization solution and trajectory output module is used to linearize and quadratically approximate the vehicle discrete state transition equation and the constrained comprehensive cost function based on the constrained iterative linear quadratic regulator algorithm, construct a linear quadratic optimal control problem, solve the optimal control quantity through backward recursion, and perform forward rolling updates of the vehicle state sequence and control sequence to output the optimal vehicle trajectory and corresponding control sequence that meet the requirements of safety, driving efficiency and comfort.

[0008] Compared with the prior art, the present invention, employing the above technical solution, has the following technical effects: 1. This invention introduces a TTC-based dynamic safety buffer distance, enabling vehicles to adaptively adjust safety boundaries based on the real-time risk level of obstacles. In high-risk scenarios, the buffer distance is automatically increased, reserving more reaction time and space for the planning and control systems, significantly improving safety. In low-risk scenarios, the buffer distance is appropriately reduced to avoid overly conservative trajectories, thus improving road utilization efficiency and traffic flow.

[0009] 2. This invention transforms complex nonlinear geometric collision constraints into differentiable cost terms using an exponential obstacle function, and then uses the CILQR algorithm for iterative solution. CILQR, through a linear quadratic approximation of the nonlinear problem sequence, can efficiently handle state and control constraints, exhibiting a fast convergence speed, thus meeting the requirements of real-time planning for autonomous driving.

[0010] 3. Under a unified optimization framework, this invention simultaneously considers trajectory tracking accuracy, control smoothness, ride comfort, and dynamic safety constraints, and can generate trajectories that achieve a good balance among multiple objectives such as safety, comfort, and efficiency. Attached Figure Description

[0011] Figure 1 This is an overall flowchart of the autonomous vehicle trajectory planning method based on dynamic safety buffer of the present invention; Figure 2 This is a schematic diagram of the vehicle kinematics bicycle model of the present invention; Figure 3 This is a model diagram of the autonomous vehicle and obstacles of the present invention; Figure 4 This is a schematic diagram of trajectory planning for an autonomous vehicle in a same-direction following scenario, as described in this embodiment of the invention. Figure 5 This is a convergence curve of the cost function of the CILQR optimization algorithm in this embodiment of the invention. Detailed Implementation

[0012] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0013] like Figure 1 As shown, this invention provides a trajectory planning method for autonomous vehicles based on dynamic safety buffers, the specific steps of which are as follows: S1. Acquire autonomous vehicle status information and environmental perception information. Autonomous vehicle status information includes, but is not limited to, position, speed, heading angle, acceleration, and steering angle. Environmental perception information includes the position, speed, and size information of surrounding static and dynamic obstacles.

[0014] S2. Based on the vehicle kinematics bicycle model, construct the vehicle discrete state transition equation and establish an optimization cost function with trajectory tracking cost, speed cost and control cost as objectives; The vehicle dynamics are described using a kinematic bicycle model for a given time interval. The nonlinear discrete model of the vehicle can be represented by the following equation: , in, Represents the discrete dynamics equations of the vehicle. Indicates the vehicle is in The state vector of a time step is defined as follows: , and These are represented as the x and y coordinates of the vehicle's center point, respectively. Indicates speed, Indicates the heading angle. Indicates the vehicle is in The control vector for the time step is defined as follows: , and These represent acceleration and steering angle, respectively. and This represents the distance from the center of mass to the front and rear axles.

[0015] The optimization cost function J is expressed as: , in, This represents the terminal time step in the prediction time domain. , These respectively indicate the vehicle's position in the [missing information]. State vectors of time step and terminal time step , The first Reference trajectory points for time step and terminal time step. Denotes the weighted L2 norm with M as the weight matrix. All are weight matrices, used to penalize tracking error, control vector, and terminal state error, respectively. The weight matrix M corresponding to the weighted 2-norm in the formula takes the values ​​Q, R, and P in sequence.

[0016] S3. Calculate the collision time TTC based on the relative distance and relative speed between the autonomous vehicle and each obstacle, and dynamically adjust the safe buffer distance between the autonomous vehicle and the obstacle according to the TTC to generate a dynamic safety buffer zone. The formula for calculating the collision time TTC is: , , , in, and These represent the following vehicle (autonomous vehicle). and the car in front In the The speed of the time step and They represent the following vehicles respectively. and the car in front In the The position of the time step Indicates the length of the vehicle.

[0017] The dynamic safety buffer distance is calculated as follows: a preset high TTC threshold is used. and low TTC threshold ,when When using a reduction factor Reduce the baseline safety buffer distance ;when When using the magnification factor Increase ;when In this case, the safety buffer distance is dynamically adjusted through linear interpolation, thereby achieving adaptive safety distance adjustment based on risk level. The specific calculation formula is as follows: , in, As a baseline safety distance, To reduce the factor, This is the magnification factor.

[0018] S4. Introduce the dynamic safety buffer into the collision constraint to construct a collision avoidance inequality constraint that includes the dynamic safety buffer; use an exponential barrier function to transform the inequality constraint and introduce it into the optimization cost function to form a constrained comprehensive cost function. Obstacles are modeled as ellipses, and autonomous vehicles are modeled as double-circle models. This is combined with a dynamic safety buffer distance. Construct a collision constraint function that represents the relationship between the reference point of the autonomous vehicle and the elliptical boundary of the obstacle. The expression is: , in, For the coordinates of the reference point of the autonomous vehicle, The coordinates of the center of the elliptical model; The formulas for calculating the major axis a and minor axis b of an ellipse are: , , Where a and b represent the major and minor axes of the elliptical obstacle, respectively. and These represent the length and width of the obstacle vehicle, respectively. This indicates the dynamic safe distance between autonomous vehicles and vehicles with obstacles. This represents the radius of the reference point for autonomous vehicles, which can be approximated as half the width of the autonomous vehicle.

[0019] The collision avoidance inequality constraint is An exponential barrier function is used. The inequality constraint is transformed and introduced into the optimization cost function to form a constrained comprehensive cost function. ,in and This is an adjustable parameter used to control the strictness of the constraints.

[0020] S5. Based on the constrained iterative linear quadratic regulator CILQR algorithm, the vehicle state equation and comprehensive cost function are iteratively linearized and quadratically approximated to construct a linear quadratic problem; the optimal control increment is solved by reverse recursion, and the vehicle state and control sequence are updated forward until the preset convergence condition is met, and the optimal vehicle trajectory and corresponding control input sequence that meet the requirements of safety, driving efficiency and comfort are output. In each iteration, the system dynamics equations are... and comprehensive cost function A first-order Taylor expansion and a second-order approximation are performed near the current trajectory to transform it into a standard LQ problem under a linear dynamic system. The optimal control law is calculated recursively in reverse order to obtain the optimal control gain matrix and feedback term, thereby obtaining the optimal control increment. Forward simulation is then performed to update the state and control sequence. This process is iterated until the state and control updates are less than a set threshold or the maximum number of iterations is reached, ultimately outputting the converged optimal trajectory and control sequence.

[0021] Based on the above-mentioned autonomous vehicle trajectory planning method, the present invention also provides an autonomous vehicle trajectory planning system based on dynamic safety buffer, comprising: The environmental perception module is used to acquire the state information and environmental perception information of the autonomous vehicle. The state information includes position, speed, heading angle, acceleration and steering angle information, and the environmental perception information includes the position, speed and size information of surrounding static obstacles and dynamic obstacles. The State Modeling and Optimization Objective Module is used to construct discrete state transition equations for a vehicle based on state information, environmental perception information, and a bicycle kinematics model, and to establish an optimization cost function with trajectory tracking cost, speed cost, and control cost as objectives. The dynamic safety buffer calculation module is used to calculate the collision time based on the relative distance and relative speed between the autonomous vehicle and each obstacle, and dynamically adjust the safety buffer distance between the autonomous vehicle and the obstacle according to the collision time to generate a dynamic safety buffer. The constraint processing module is used to introduce the dynamic safety buffer into the collision constraint conditions, construct a collision avoidance function containing the dynamic safety buffer, perform constraint transformation on the collision avoidance function using an exponential barrier function, and introduce an optimization cost function to form a constrained comprehensive cost function. The optimization solution and trajectory output module is used to linearize and quadratically approximate the vehicle discrete state transition equation and the constrained comprehensive cost function based on the constrained iterative linear quadratic regulator algorithm, construct a linear quadratic optimal control problem, solve the optimal control quantity through backward recursion, and perform forward rolling updates of the vehicle state sequence and control sequence to output the optimal vehicle trajectory and corresponding control sequence that meet the requirements of safety, driving efficiency and comfort.

[0022] The following example illustrates the trajectory planning of an autonomous vehicle on an urban road scenario involving vehicles traveling in the same direction ahead.

[0023] Step S1: Information Acquisition The current status of autonomous vehicles is obtained through onboard GPS, IMU, wheel speed sensors, etc., including location. ,speed Heading angle acceleration and steering angle By fusing sensors such as LiDAR, cameras, and millimeter-wave radar, the system obtains the status of target vehicles ahead, including their location. ,speed ,length ,width .

[0024] Step S2: Model and Cost Function Establishment The aforementioned kinematic bicycle model is used as the prediction model. The kinematic bicycle model is as follows: Figure 2 As shown. Discrete time step. Predicting the time domain Set a reference trajectory. The centerline of the lane. Design the weight matrix. , , To balance tracking accuracy with control consumption.

[0025] Step S3: Calculation of dynamic safety buffer distance Calculate the relative speed between the autonomous vehicle and the vehicle in front. and relative distance Calculate TTC using the aforementioned TTC calculation formula; set parameters, including the baseline safety distance. threshold , ,coefficient , Based on the current TTC value, the current dynamic safety buffer distance is calculated by substituting it into the aforementioned piecewise function. For example, if the current Then, by linear interpolation, we can obtain... .

[0026] Step S4: Collision Constraint Construction and Transformation like Figure 3 As shown, the surrounding vehicles (SVs) are expanded into an elliptical obstacle. The radius of the reference point for the autonomous vehicle (EV) is taken as... Calculate the ellipse parameters to obtain... , Establish a distance function from the autonomous vehicle's reference point to the center of the ellipse, and construct a constraint inequality that ensures the point lies outside the ellipse. Choose the exponential barrier function. The inequality constraint is transformed into a penalty term and added to the comprehensive cost function.

[0027] Step S5: CILQR optimization solution (1) Initialization: Starting from the current state, generate an initial trajectory using an LQR or a simple controller that does not consider obstacle constraints. .

[0028] (2) Iterative optimization: a. Linearization and quadratic approximation: In the i-th iteration, along the current trajectory For system dynamics f and comprehensive cost function Perform a first-order Taylor expansion and a second-order approximation; b. Reverse recursion: from the terminal Initially, the linear quadratic optimal control problem equations were solved in reverse, yielding a series of gain matrices. and feedforward terms ; c. Forward Roll: Starting from the initial state, use the calculated optimal control law. Perform forward simulation to update the state and control sequences and obtain the new trajectory. ; d. Convergence Criterion: Calculate the difference between the old and new trajectories. If the difference is less than the set tolerance or the maximum number of iterations is reached, stop iterating; otherwise, let... Return to step a.

[0029] (3) Output: The optimal control sequence obtained in the last iteration and state trajectory Output. Typically, the first control variable... It is applied to the vehicle actuator and the entire process is repeated in the next planning cycle.

[0030] Through the above process, the vehicle can safely follow the vehicle in front while adjusting the following distance based on the real-time TTC (Traffic Traction Center). When the TTC is low, the planned trajectory will increase lateral deviation or brake earlier, reflecting a greater safety margin; when the TTC is high, the planned trajectory will be closer to the reference lane lines, pursuing driving efficiency.

[0031] Figure 4 This demonstrates the trajectory planning performance of an autonomous vehicle in a same-direction following scenario. The autonomous vehicle smoothly travels around the lane centerline, maintaining its lane position while achieving safe avoidance of target vehicles ahead through adaptive lateral offset, verifying the trajectory safety and smoothness of the algorithm under dynamic obstacle constraints.

[0032] Figure 5 The curve shows the convergence of the cost function of the CILQR optimization algorithm. The cost function value decreases rapidly in the early stages of iteration and stabilizes after only about 4 iterations, eventually converging to a low level. This demonstrates that the optimization method has the characteristics of fast convergence speed and good stability, which can meet the real-time requirements of autonomous driving trajectory planning.

[0033] Based on the same inventive concept, embodiments of this application provide a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the aforementioned autonomous vehicle trajectory planning method based on dynamic safety buffer.

[0034] Based on the same inventive concept, embodiments of this application provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the aforementioned autonomous vehicle trajectory planning method based on dynamic safety buffers.

[0035] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0036] This invention is described with reference to flowchart illustrations of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each step in the flowchart and combinations thereof can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing device, generate instructions for implementing the processes... Figure 1 A device for a function specified in one or more processes.

[0037] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 The function specified in one or more processes.

[0038] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 Steps of a specified function in one or more processes.

[0039] The above embodiments are merely illustrative of the technical concept of the present invention and should not be construed as limiting the scope of protection of the present invention. Any modifications made to the technical solutions based on the technical concept proposed in this invention shall fall within the scope of protection of this invention.

Claims

1. A trajectory planning method for autonomous vehicles based on dynamic safety buffers, characterized in that, Includes the following steps: Step S1: Obtain the state information and environmental perception information of the autonomous vehicle. The state information includes position, speed, heading angle, acceleration and steering angle information. The environmental perception information includes the position, speed and size information of surrounding static and dynamic obstacles. Step S2: Based on state information, environmental perception information and vehicle kinematics bicycle model, construct vehicle discrete state transition equations and establish an optimization cost function with trajectory tracking cost, speed cost and control cost as objectives; Step S3: Calculate the collision time based on the relative distance and relative speed between the autonomous vehicle and each obstacle, and dynamically adjust the safe buffer distance between the autonomous vehicle and the obstacle according to the collision time to generate a dynamic safe buffer zone. In step S3, a high collision time threshold is preset. and low collision time threshold ,when When, a reduction factor is used to reduce the baseline safety buffer distance; when When, the reference safety buffer distance is increased by using an amplification factor; when At that time, the safety buffer distance is dynamically adjusted through linear interpolation, thereby achieving adaptive safety distance adjustment based on the risk level; the dynamic adjustment formula for the safety buffer distance is as follows: ， in, To provide a safe buffer distance between autonomous vehicles and obstacles, As a baseline safety buffer distance, To reduce the factor, , This is the magnification factor. , The collision time; Step S4: Introduce the dynamic safety buffer into the collision constraint conditions to construct a collision avoidance function containing the dynamic safety buffer; perform constraint transformation on the collision avoidance function using an exponential barrier function, and introduce an optimization cost function to form a constrained comprehensive cost function; In step S4, an ellipse model is used to represent the obstacle, and a double-circle model is used to represent the autonomous vehicle. A collision constraint function is constructed based on geometric relationships to characterize the relationship between the autonomous vehicle's reference point and the elliptical boundary of the obstacle, as shown in the following formula: ， in, Represents the collision constraint function. , For the coordinates of the reference point of the autonomous vehicle, The coordinates of the center of the obstacle ellipse; and Let represent the major and minor axes of the obstacle ellipse model, respectively, as shown in the following formula: ， ， in, and These represent the length and width of the obstacle, respectively. Indicates the radius of the reference point for autonomous vehicles; Exponential barrier function The expression is: ， in, All parameters are adjustable. and ; This is a collision constraint function used to characterize the spatial constraint relationship between autonomous vehicles and obstacles; Will Introducing an optimization cost function Form a constrained comprehensive cost function : ； Step S5: Based on the constrained iterative linear quadratic regulator algorithm, the discrete state transition equation of the vehicle and the constrained comprehensive cost function are linearized and approximated quadratically to construct a linear quadratic optimal control problem. The optimal control quantity is solved by reverse recursion, and the vehicle state sequence and control sequence are updated in a forward rolling manner to output the optimal vehicle trajectory and corresponding control sequence that meet the requirements of safety, driving efficiency and comfort.

2. The autonomous vehicle trajectory planning method based on dynamic safety buffer according to claim 1, characterized in that, In step S2, a kinematic bicycle model is used to describe the vehicle dynamics for a given time step interval. The discrete state transition equation for the vehicle is expressed as follows: ， in, Represents the discrete dynamics equations of the vehicle. , These respectively indicate the vehicle's position in the [missing information]. The state vector at each time step , and These are represented as the x and y coordinates of the vehicle's center point, respectively. Indicates speed, Indicates the heading angle. Indicates the vehicle is in The control vector for the time step is defined as follows: , and These represent acceleration and steering angle, respectively. and These are the distances from the vehicle's center of gravity to the front and rear axles, respectively.

3. The autonomous vehicle trajectory planning method based on dynamic safety buffer according to claim 1, characterized in that, In step S2, the optimization cost function is expressed as: ， in, This represents the optimization cost function. This represents the terminal time step in the prediction time domain. , These respectively indicate the vehicle's position in the [missing information]. State vectors of time step and terminal time step , The first Reference trajectory points for time step and terminal time step. Denotes the weighted L2 norm with M as the weight matrix. All are weight matrices, used to penalize tracking error, control vector, and terminal state error, respectively. The weight matrix M corresponding to the weighted 2-norm in the formula takes the values ​​Q, R, and P in sequence.

4. The autonomous vehicle trajectory planning method based on dynamic safety buffer according to claim 1, characterized in that, In step S3, the formula for calculating the collision time is as follows: ， ， ， in, This indicates that at time step k, the autonomous vehicle n is in close proximity to the vehicle in front. The collision time; and These represent autonomous vehicles. and the car in front In the The position of the time step Indicates the length of an autonomous vehicle. and These represent autonomous vehicles. and the car in front In the The speed of the time step.

5. The autonomous vehicle trajectory planning method based on dynamic safety buffer according to claim 1, characterized in that, In step S5, during trajectory planning at each time step, an initial trajectory is generated using a linear quadratic regulator that does not consider obstacle constraints, followed by iterative optimization. During each iteration, the vehicle's discrete state transition equation is adjusted along the initial trajectory. and constrained synthesis cost function A linear quadratic optimal control problem is constructed by performing a first-order Taylor expansion and a second-order approximation. The optimal control gain matrix and feedforward term are calculated by backpropagation to obtain the optimal control quantity. Then, forward simulation is performed to update the state sequence and control sequence to obtain a new trajectory. The iteration terminates when the difference between the new trajectory and the initial trajectory is less than a set threshold or the maximum number of iterations is reached, and the converged optimal trajectory and optimal control sequence are output. The first control quantity in the optimal control sequence is applied to the autonomous vehicle to execute the control of the current time step, and the above trajectory planning process is repeated in the next time step.

6. A system applied to the trajectory planning method for autonomous vehicles based on dynamic safety buffers as described in any one of claims 1-5, characterized in that, include: The environmental perception module is used to acquire the state information and environmental perception information of the autonomous vehicle. The state information includes position, speed, heading angle, acceleration and steering angle information, and the environmental perception information includes the position, speed and size information of surrounding static obstacles and dynamic obstacles. The State Modeling and Optimization Objective Module is used to construct discrete state transition equations for a vehicle based on state information, environmental perception information, and a bicycle kinematics model, and to establish an optimization cost function with trajectory tracking cost, speed cost, and control cost as objectives. The dynamic safety buffer calculation module is used to calculate the collision time based on the relative distance and relative speed between the autonomous vehicle and each obstacle, and dynamically adjust the safety buffer distance between the autonomous vehicle and the obstacle according to the collision time to generate a dynamic safety buffer. The constraint processing module is used to introduce the dynamic safety buffer into the collision constraint conditions, construct a collision avoidance function containing the dynamic safety buffer, perform constraint transformation on the collision avoidance function using an exponential barrier function, and introduce an optimization cost function to form a constrained comprehensive cost function. The optimization solution and trajectory output module is used to linearize and quadratically approximate the vehicle discrete state transition equation and the constrained comprehensive cost function based on the constrained iterative linear quadratic regulator algorithm, construct a linear quadratic optimal control problem, solve the optimal control quantity through backward recursion, and perform forward rolling updates of the vehicle state sequence and control sequence to output the optimal vehicle trajectory and corresponding control sequence that meet the requirements of safety, driving efficiency and comfort.

7. A computer device comprising a memory, a processor, and a computer program stored in the memory and capable of running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the autonomous vehicle trajectory planning method based on dynamic safety buffer as described in any one of claims 1 to 5.

8. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the autonomous vehicle trajectory planning method based on dynamic safety buffer as described in any one of claims 1 to 5.

Images