Curve trajectory planning and tracking method and system for vehicle intelligent cruise control

Abstract

The invention provides a curve track planning and tracking method and system for vehicle intelligent cruise control, and the method comprises the steps: obtaining the curvature information of a road within a preset distance in front of a vehicle based on a navigation map and vehicle positioning information; planning a smooth reference vehicle speed curve according to the curvature information and a preset transverse acceleration threshold value; on the basis of lane line information, a dynamic reference track is generated through an optimization algorithm in combination with a reference vehicle speed curve; a curve tangent strategy is coded in the optimization algorithm, so that the generated track deviates to the outer side of the curve before entering the curve, is close to the inner side of the curve at the curve center and returns to the center when going out of the curve; a model predictive controller is utilized, the dynamic reference trajectory is used as a tracking target to solve the optimal control quantity in a rolling mode, steering and driving / braking instructions are output, the vehicle is controlled to track the dynamic reference trajectory accurately, and safe, comfortable, efficient and personified smooth trajectory tracking of the vehicle in the curve scene is achieved. And the problem of internal and external cutting is fundamentally solved.

Description

Technical Field

[0001] This invention relates to the field of intelligent driving technology, specifically to a method and system for curve trajectory planning and tracking for intelligent cruise control of vehicles. Background Technology

[0002] With the rapid development of intelligent driving technology, Advanced Driver Assistance Systems (ADAS) are increasingly being used in vehicles. CNOA (Cruise Navigation & Overtaking Assist), as an important intelligent navigation driving assistance function, is mainly applied in structured road scenarios such as highways and expressways. It can achieve point-to-point autonomous driving assistance, including automatic cruise control, following the vehicle in front, autonomous lane changing, and entering / exiting ramps. In curve driving conditions, the core technology of this function lies in accurate path planning and stable tracking.

[0003] However, existing CNOA-based vehicles generally suffer from poor trajectory control when navigating curves, primarily manifested as "inside / outside shear" phenomena: Inward tangent phenomenon: When a vehicle makes a steering operation too early or at too large a steering angle when entering a curve, the actual driving trajectory deviates to the inside of the curve, which may cause the vehicle to run over the inside lane line. Outward tangent phenomenon: When a vehicle enters a curve, it understeers or the steering timing is delayed, causing the driving trajectory to deviate to the outside of the curve. This may not only encroach on the outer lane line, but may even encroach on the adjacent lane in complex curves.

[0004] These trajectory deviations pose multiple hazards: First, frequent trajectory corrections cause vehicles to wobble, severely impacting ride comfort; second, driving over the lines creates significant safety risks and increases the risk of traffic accidents; and third, unstable cornering strategies force vehicles to reduce speed, affecting road traffic efficiency.

[0005] Current common solutions often employ a tracking strategy based on a fixed lane centerline, combined with traditional PID control methods. These methods have significant limitations: firstly, they lack the ability to anticipate curve curvature, making proactive speed planning impossible; secondly, the control strategy relies on error feedback, only allowing for lag correction after trajectory deviation, making it difficult to fundamentally avoid trajectory fluctuations.

[0006] Therefore, developing a control method that can proactively predict and globally optimize to achieve smooth and stable cornering has become an urgent need to improve the performance of CNOA systems. Summary of the Invention

[0007] This invention aims to overcome the shortcomings of the prior art and provide a method and system for curve trajectory planning and tracking for intelligent cruise control of vehicles. It overcomes the phenomenon of vehicle trajectory "inside and outside cutting" caused by passively tracking the fixed lane centerline and lagging control in the existing CNOA system in curve scenarios, thereby improving the safety, comfort and traffic efficiency of curve driving.

[0008] The technical solution of the present invention to solve the above-mentioned technical problems is as follows: In a first aspect, the present invention provides a method for curve trajectory planning and tracking for intelligent cruise control of vehicles, comprising the following steps: S1. Curvature Pre-aiming and Cooperative Speed ​​Planning: Based on the navigation map and vehicle positioning information, obtain the road curvature information within a predetermined distance in front of the vehicle; and plan a smooth reference speed curve according to the curvature information and a preset lateral acceleration threshold. S2. Dynamic Optimal Trajectory Generation: Based on lane line information and combined with the reference vehicle speed curve, a dynamic reference trajectory is generated through an optimization algorithm; wherein, the optimization algorithm is encoded with a curve tangent strategy so that the generated trajectory deviates from the outside of the curve before entering the curve, approaches the inside of the curve at the apex, and returns to the center when exiting the curve. S3. Model Predictive Tracking Control: Using a model predictive controller, with the dynamic reference trajectory as the tracking target, the optimal control quantity is solved in a rolling manner, and steering and driving / braking commands are output to control the vehicle to accurately track the dynamic reference trajectory.

[0009] Based on the above scheme, step S1 further includes the following steps: The reference vehicle speed curve planning in step S1 specifically includes the following steps: S11. Curvature Calculation: Extract the shape point sequence of the center line of the lane ahead from the navigation map, and optimize the curvature accuracy and eliminate curvature abrupt fluctuations through curve fitting and curvature calculation formula to obtain continuous and smooth road curvature data. S12. Determination of Basic Speed ​​Limit: Based on the formula

[0010] Calculate the reference vehicle speed at path point s, where, The lateral acceleration threshold is... Let be the road curvature at point s; S13. Speed ​​curve smoothing: The initial reference vehicle speed sequence is smoothed. The smoothing process includes using fifth-order polynomial fitting or Gaussian filtering to generate a reference vehicle speed curve with smooth acceleration and deceleration transitions.

[0011] Furthermore, the smoothing process in step S13 also includes a dynamic optimization process: based on the model predictive control framework, the deceleration start point and acceleration are dynamically adjusted according to the real-time vehicle status and the trend of curvature change ahead.

[0012] Furthermore, step S2 specifically includes the following steps: S21. Constructing a reference line: Based on the lane centerline, the reference line heading is corrected by combining the curve tangent strategy to form a smooth reference line; S22. Coordinate system transformation: Project the vehicle state onto the reference line and transform it to the Frenet coordinate system to decouple the vehicle motion into longitudinal motion along the reference line and lateral motion perpendicular to the reference line. S23. Trajectory Planning: Perform trajectory planning for longitudinal and lateral motion respectively to generate longitudinal and lateral trajectories; S24. Trajectory Fusion and Filtering: The longitudinal and lateral trajectories are fused into a complete trajectory in Cartesian space, and the optimal trajectory is selected by evaluating the cost function.

[0013] Furthermore, the cost function of the optimization algorithm in step S2 includes at least: a centerline cost term for bringing the vehicle trajectory closer to the lane centerline, a comfort cost term for penalizing the rate of change of trajectory curvature to ensure smoothness, and a safety cost term for preventing the vehicle trajectory from approaching the lane boundary.

[0014] Furthermore, step S23 specifically includes: Using the vehicle's speed as the initial state, the longitudinal termination state is sampled according to the cruise, following, or parking modes. The initial state and the sampled termination state are fitted using a fifth-order polynomial to generate multiple longitudinal trajectories. Lateral offset is sampled around the lane centerline, and the lateral velocity and acceleration at the lateral termination state are set to zero. A fourth-order polynomial is used to fit and generate the lateral trajectory.

[0015] Furthermore, step S3 specifically includes the following steps: S31. Predictive Model Construction: Use the vehicle dynamics model as the predictive model to predict the state of the vehicle over a future period of time. S32, Rolling Optimization Solution: In each control cycle, an optimization problem is solved to find a sequence of optimal front wheel steering angles and acceleration control quantities; wherein, the objective of the optimization problem is to make the predicted vehicle trajectory accurately track the dynamic reference trajectory generated in step S2, while satisfying vehicle dynamics constraints; S33. Feedforward control implementation: By utilizing the future curve curvature information contained in the dynamic reference trajectory, steering action is applied in advance to achieve smooth steering control.

[0016] Secondly, the present invention also provides a curve trajectory planning and tracking system for intelligent cruise control of vehicles, comprising: The perception and positioning module is configured to acquire the vehicle's own status, surrounding environment information, and high-precision positioning data. The curvature prediction and vehicle speed planning module is communicatively connected to the perception and positioning module and is configured to obtain the curvature of the road ahead based on the navigation map and plan a reference vehicle speed curve that meets the requirements. The dynamic trajectory planning module is communicatively connected to the curvature prediction and vehicle speed planning module, and is configured to generate a dynamic reference trajectory based on lane line information and the reference vehicle speed curve by using an optimization algorithm with a curve tangent strategy. The model prediction control module is communicatively connected to the dynamic trajectory planning module and is configured to solve and output steering and driving / braking commands based on the dynamic reference trajectory to control the vehicle to stably follow the trajectory.

[0017] Thirdly, the present invention also provides an electronic device, including at least one processor and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the at least one processor implements the curve trajectory planning and tracking method for intelligent cruise control of a vehicle as described in the first aspect by executing the instructions stored in the memory.

[0018] Fourthly, the present invention also provides a computer-readable storage medium storing at least one instruction or at least one program, wherein the at least one instruction or at least one program is loaded and executed by a processor to implement the curve trajectory planning and tracking method for intelligent cruise control of a vehicle as described in the first aspect.

[0019] Compared with existing technologies, the technical solution provided by this invention systematically solves the internal and external tangent problems in cornering cruise by combining two core ideas: "curvature-velocity coordination" and "dynamic reference path planning," along with the powerful tracking capabilities of Model Predictive Control (MPC). The fundamental difference from traditional methods lies in the fact that this method is an active, feedforward, and globally optimized solution. It does not correct errors after they occur, but rather plans a trajectory that fundamentally avoids internal and external tangent phenomena before errors occur. Specifically, this invention achieves the following beneficial effects: 1. Significantly improved safety: Through the dynamic trajectory planning and precise tracking control, the vehicle can always be kept stably controlled in the center of the lane, effectively avoiding crossing the line and eliminating the safety risks caused by inward or outward lane cutting.

[0020] 2. Significantly improved ride comfort: By generating a smooth and stable trajectory and achieving smooth acceleration and deceleration based on curvature-speed coordination, the lateral and longitudinal impact when the vehicle is cornering is significantly reduced, fundamentally overcoming the "dragon-like" swaying caused by traditional PID control, making the ride experience more comfortable.

[0021] 3. Improved traffic efficiency: While ensuring safety and comfort, optimized trajectory and speed planning enable vehicles to pass through curves at a higher average speed, thus improving traffic efficiency.

[0022] 4. Highly human-like driving behavior: By simulating the "outside-inside-outside" cornering strategy of excellent human drivers, a trajectory that is more in line with human driving habits is generated, which significantly enhances passengers' trust in the intelligent driving system.

[0023] In summary, this system can significantly improve the overall performance of CNOA in curve scenarios, making intelligent driving safer, more comfortable, more efficient, and more human-like. Attached Figure Description

[0024] Figure 1 A schematic diagram of a curve trajectory planning and tracking method for intelligent cruise control of vehicles provided in an embodiment of the present invention; Figure 2 A schematic diagram of a curve trajectory planning and tracking system for intelligent cruise control of a vehicle provided in an embodiment of the present invention; Figure 3 A schematic diagram of the hardware structure of a possible electronic device provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of the hardware structure of a possible computer-readable storage medium provided in an embodiment of the present invention. Detailed Implementation

[0025] The implementation methods of this solution will be described in further detail below. Obviously, the described embodiments are only a part of the embodiments of this solution, and not an exhaustive list of all embodiments. It should be noted that, unless otherwise specified, the embodiments and features in the embodiments of this solution can be combined with each other.

[0026] Example 1 This invention provides a method for curve trajectory planning and tracking for intelligent cruise control of vehicles, such as... Figure 1 As shown, it includes the following steps: Step S1, Curvature Pre-aiming and Cooperative Speed ​​Planning: Based on the navigation map and vehicle positioning information, obtain the road curvature information within a predetermined distance in front of the vehicle; according to the curvature information and a preset lateral acceleration threshold, plan a smooth reference speed curve.

[0027] Specifically, step S1 includes the following sub-steps: Step S11, Curvature Calculation: Extract the lane centerline shape point sequence of the path 150-300 meters ahead from the navigation map (SD MAP). These shape points are precise data collected by sensors through LiDAR, cameras, and other devices, and then thinned and optimized. After vehicle positioning, the lane centerline shape point sequence of the road segment ahead is retrieved from the map based on its own position, serving as the basic data source for curvature calculation. The curvature calculation is performed in the following way: The initial curvature is calculated using basic algorithms: two efficient algorithms are commonly used. One is the three-point method, which selects three consecutive shape points to determine the arc, and calculates the radius of the arc by the cross product of vectors and the modulus. The curvature is the reciprocal of the radius. The other is the bisection method optimization, which uses the bisection method to continuously select new intermediate points and recombine them for calculation if the arc determined by the three points cannot fit more shape points. This continues until the curvature matches the actual curvature of the road.

[0028] Curve fitting optimizes curvature accuracy: Perform least squares fitting on the shape point sequence to obtain a smooth fitting curve, and then substitute it into the curvature formula to calculate the curvature of each point on the curve, ensuring that the calculation result fits the real curvature of the road.

[0029] Step S12, Curvature Smoothing Optimization: The calculated curvature sequence is filtered, and low-pass filtering and other methods are used to eliminate abrupt fluctuations in adjacent curvatures and remove outliers. For locations such as the beginning and end of the road that cannot be calculated using three points, reasonable curvature values ​​of adjacent points are reused, and finally, a continuous and stable curvature sequence of the road ahead is output.

[0030] Step S13, Determining the Basic Speed ​​Limit: Calculate the reference speed at point s on the path according to the formula, where s is the preset comfortable lateral acceleration threshold (e.g., 0.3g-0.4g). When determining the basic speed limit based on curvature, select the point with the largest curvature in the path as the anchor point, and use the speed limit at this point as the core benchmark to avoid excessive centrifugal force when cornering.

[0031] The following is an example of the correspondence between radius and vehicle speed (unit: radius in meters, vehicle speed in km / h): Step S14, Speed ​​Curve Smoothing: Smoothing the initial reference vehicle speed sequence, including: Use a smooth curve to fit the speed transition: First, segment the path and fit it with Euler spirals, etc., so that the curvature changes linearly with the arc length, providing a basis for smooth speed; then use a fifth-order polynomial to fit the speed curve, for example, when entering a curve, reduce the current speed to the curve limit with a comfortable deceleration, and when exiting the curve, restore the speed to the cruising speed with a comfortable acceleration, ensuring that there are no sudden changes in acceleration and deceleration.

[0032] The optimized algorithm ensures dynamic comfort: It employs quadratic programming optimization, incorporating a rate of change of speed term into the cost function to balance speed limit accuracy and speed smoothness, while simultaneously constraining speed and acceleration to not exceed vehicle dynamic limits. Furthermore, it utilizes data-driven learning to collect lateral acceleration data from different drivers during cornering online, correcting the speed parameters under corresponding curvatures.

[0033] Combined with predictive control for dynamic adjustment: By using model predictive control, the curvature changes of the subsequent path can be predicted in advance, and the speed can be adjusted in advance (such as starting deceleration in advance when the curvature ahead is gradually increasing). At the same time, the error between the actual driving speed and the planned speed is fed back in real time, and the acceleration is dynamically fine-tuned.

[0034] Step S2, Dynamic Optimal Trajectory Generation: Based on lane line information and combined with the reference vehicle speed curve, a dynamic reference trajectory is generated through an optimization algorithm; wherein, the optimization algorithm encodes a curve tangent strategy so that the generated trajectory deviates from the outside of the curve before entering the curve, approaches the inside of the curve at the apex, and returns to the center when exiting the curve.

[0035] Specifically, step S2 includes the following sub-steps: Step S21: Construct a reference line: Based on the lane centerline, combine the curve tangent strategy to correct the local heading of the reference line, ensuring that the reference line conforms to the curve geometry and is smooth.

[0036] Step S22, Coordinate System Transformation: Project the vehicle state (position, vehicle speed, etc.) onto the reference line and transform it to the Frenet coordinate system. Decouple the vehicle motion into longitudinal motion along the reference line and lateral motion perpendicular to the reference line, reducing planning complexity.

[0037] Step S23, Trajectory Planning: Using the Frenet coordinate system and Lattice planning scheme, combined with polynomial fitting and multi-stage optimization, trajectory planning is performed for both longitudinal and lateral motions. Longitudinal trajectory planning: Using the vehicle's speed as the initial state, sample the longitudinal termination state in three modes: cruise, following, and parking (for example, in cruise mode, sample constant speeds at different time points, and the termination acceleration is zero); then use a fifth-order polynomial to fit the initial speed and the sampled termination state to generate multiple longitudinal trajectories, while controlling parameters such as acceleration to ensure smoothness.

[0038] Lateral trajectory planning: Sample the lateral offset around the lane centerline, and set the termination state to have zero lateral velocity and acceleration to ensure that the vehicle eventually fits the centerline; generate the lateral trajectory by fitting the initial lateral state and the sampling termination state with a fourth-order polynomial to avoid abrupt turns where the trajectory is tangent to or tangent to the lane boundary.

[0039] Step S24, Trajectory Fusion and Filtering: Match the parameters of the longitudinal and lateral trajectories and fuse them into a complete trajectory in Cartesian space. Then, evaluate each trajectory using a cost function, which includes: Centerline logic: Encourage vehicles to stay close to the centerline of the lane.

[0040] Comfort logic: penalize the rate of change of curvature of the trajectory to ensure a smooth trajectory.

[0041] Safety logic: Penalize vehicles that get too close to the lane boundary.

[0042] Efficiency logic: Encourage driving at higher speeds.

[0043] The system filters out collision-free trajectories that meet vehicle dynamics constraints, and finally outputs a smooth and optimal driving trajectory. The cornering tangent strategy is encoded in the cost function, allowing the vehicle to deviate slightly to the outside of the corner before entering the corner, move slightly closer to the inside of the corner at the apex, and smoothly return to the center when exiting the corner, similar to an "outside-inside-outside" line, in order to reduce the curvature of the actual driving trajectory.

[0044] Step S3, Model Predictive Tracking Control: Using the Model Predictive Controller (MPC), with the dynamic reference trajectory as the tracking target, the optimal control quantity is solved in a rolling manner, and steering and driving / braking commands are output to control the vehicle to accurately track the dynamic reference trajectory.

[0045] Specifically, step S3 includes the following sub-steps: Step S31, Prediction Model Construction: Use a vehicle dynamics model (such as a single-vehicle model) as a prediction model to predict the state of the vehicle in the future.

[0046] Step S32, Rolling Optimization Solution: In each control cycle, MPC solves an optimization problem to find a series of optimal front wheel steering angles and acceleration control values, so that the predicted trajectory is as close as possible to the optimal trajectory generated by the planner, while satisfying vehicle dynamics constraints.

[0047] Step S33, Feedforward and Anti-interference Control: By utilizing the future curve curvature information contained in the dynamic reference trajectory, steering actions are applied in advance to achieve smooth steering similar to that of a human (feedforward control); at the same time, it can compensate for trajectory deviations caused by crosswinds and uneven road surfaces in real time (anti-interference), preventing the vehicle from exhibiting inside or outside shear. Throughout the process, the controller will make fine adjustments in real time based on vehicle status feedback to ensure the robustness of control.

[0048] It should be noted that this method systematically solves the inside-outside tangent problem in curve cruise by combining two core ideas: curvature-velocity co-planning and dynamic reference path planning, along with the tracking capability of MPC. Compared with traditional methods (such as tracking a fixed lane centerline and using PID control), this method is proactive, feedforward, and globally optimized. It does not correct errors after they occur, but plans a trajectory to avoid errors before they occur.

[0049] As an example of this embodiment, combined with Figure 2 The process is as follows: 1. Perception and localization stage: The system perceives lane lines and other vehicles through sensor fusion (camera, radar, lidar); combined with SD MAP and GPS / IMU, it achieves sub-meter level positioning and obtains attitude information such as the vehicle's current yaw angle and lateral speed.

[0050] 2. Curvature-velocity co-planning stage: The curvature prediction module obtains the curvature of the path ahead from the SD MAP, and generates a curvature sequence after calculation and optimization; the velocity planner generates a comfortable velocity curve based on the curvature to achieve smooth deceleration.

[0051] 3. Trajectory Generation Stage: The trajectory planner integrates lane centerline, speed information, and curve tangent strategy to generate a smooth trajectory without internal or external tangency in the Frenet coordinate system.

[0052] 4. Control Execution Phase: The lateral MPC controller calculates the front wheel steering angle and executes the EPS command; the longitudinal PID / MPC controller controls the throttle and brakes to ensure the vehicle accurately tracks the trajectory.

[0053] Example 2 This embodiment provides a curve trajectory planning and tracking system for implementing the method of Embodiment 1 above. The system includes: Perception and positioning module: configured to acquire vehicle status, surrounding environment information and high-precision positioning data. Specifically, it is configured to perceive lane lines and other vehicles through sensor fusion, and combine SD MAP and GPS / IMU to achieve sub-meter level positioning and acquire vehicle attitude information (such as yaw angle and lateral speed).

[0054] Curvature prediction and speed planning module: It communicates with the perception and positioning module and is configured to obtain the curvature of the road ahead based on the navigation map and plan a reference speed curve that meets the requirements. Specifically, it is configured to extract the curvature of the road ahead from the SD MAP, and after calculation and filtering optimization, plan a reference speed curve that meets the comfort requirements (including curvature prediction, speed planning and smoothing).

[0055] Dynamic trajectory planning module: It communicates with the curvature prediction and vehicle speed planning module and is configured to generate a dynamic reference trajectory based on lane line information and reference vehicle speed curve by encoding an optimization algorithm with a curve tangent strategy (such as Frenet coordinate system + Lattice planning).

[0056] Model Predictive Control Module: Communicates with the dynamic trajectory planning module and is configured to solve and output steering and drive / braking commands based on the dynamic reference trajectory to control the vehicle to stably follow the trajectory. It also has feedforward and anti-interference functions.

[0057] Example 3 See Figure 3 , Figure 3 This is a schematic diagram illustrating an embodiment of the electronic device provided in this invention. For example... Figure 3 As shown, this embodiment of the invention provides an electronic device 300, including a memory 310, a processor 320, and a computer program 311 stored in the memory 310 and executable on the processor 320. When the processor 320 executes the computer program 311, it performs the following steps: S1. Curvature Pre-aiming and Cooperative Speed ​​Planning: Based on the navigation map and vehicle positioning information, obtain the road curvature information within a predetermined distance in front of the vehicle; and plan a smooth reference speed curve according to the curvature information and a preset lateral acceleration threshold. S2. Dynamic Optimal Trajectory Generation: Based on lane line information and combined with the reference vehicle speed curve, a dynamic reference trajectory is generated through an optimization algorithm; wherein, the optimization algorithm is encoded with a curve tangent strategy so that the generated trajectory deviates from the outside of the curve before entering the curve, approaches the inside of the curve at the apex, and returns to the center when exiting the curve. S3. Model Predictive Tracking Control: Using a model predictive controller, with the dynamic reference trajectory as the tracking target, the optimal control quantity is solved in a rolling manner, and steering and driving / braking commands are output to control the vehicle to accurately track the dynamic reference trajectory.

[0058] Example 4 See Figure 4 , Figure 4 This is a schematic diagram illustrating an embodiment of a computer-readable storage medium provided by the present invention. (See diagram below.) Figure 4 As shown, this embodiment provides a computer-readable storage medium 400 on which a computer program 311 is stored. When the computer program 311 is executed by a processor, it performs the following steps: S1. Curvature Pre-aiming and Cooperative Speed ​​Planning: Based on the navigation map and vehicle positioning information, obtain the road curvature information within a predetermined distance in front of the vehicle; and plan a smooth reference speed curve according to the curvature information and a preset lateral acceleration threshold. S2. Dynamic Optimal Trajectory Generation: Based on lane line information and combined with the reference vehicle speed curve, a dynamic reference trajectory is generated through an optimization algorithm; wherein, the optimization algorithm is encoded with a curve tangent strategy so that the generated trajectory deviates from the outside of the curve before entering the curve, approaches the inside of the curve at the apex, and returns to the center when exiting the curve. S3. Model Predictive Tracking Control: Using a model predictive controller, with the dynamic reference trajectory as the tracking target, the optimal control quantity is solved in a rolling manner, and steering and driving / braking commands are output to control the vehicle to accurately track the dynamic reference trajectory.

[0059] Based on the above embodiments, it can be seen that this invention achieves global optimization of cornering control by integrating curvature-velocity co-planning, dynamic trajectory generation, and model predictive control. Specific beneficial effects include: by organically combining curvature-velocity co-planning, dynamic trajectory optimization, and model predictive control, this invention achieves smooth trajectory tracking in cornering scenarios that is safe (avoiding the risk of crossing the line), comfortable (eliminating the "dragon-like" phenomenon), efficient (maintaining a high passing speed), and human-like (simulating human driving strategies), fundamentally solving the internal and external tangency problems existing in traditional methods. This invention is not a simple improvement on existing technologies, but rather provides an active, feedforward solution through methodological innovation, providing core technical support for the application of intelligent driving systems in cornering scenarios.

[0060] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for curve trajectory planning and tracking for intelligent cruise control of vehicles, characterized in that, Includes the following steps: S1. Curvature Pre-aiming and Cooperative Speed ​​Planning: Based on the navigation map and vehicle positioning information, obtain the road curvature information within a predetermined distance in front of the vehicle; and plan a smooth reference speed curve according to the curvature information and a preset lateral acceleration threshold. S2. Dynamic Optimal Trajectory Generation: Based on lane line information and combined with the reference vehicle speed curve, a dynamic reference trajectory is generated through an optimization algorithm; wherein, the optimization algorithm is encoded with a curve tangent strategy so that the generated trajectory deviates from the outside of the curve before entering the curve, approaches the inside of the curve at the apex, and returns to the center when exiting the curve. S3. Model Predictive Tracking Control: Using a model predictive controller, with the dynamic reference trajectory as the tracking target, the optimal control quantity is solved in a rolling manner, and steering and driving / braking commands are output to control the vehicle to accurately track the dynamic reference trajectory.

2. The method according to claim 1, characterized in that, The reference vehicle speed curve planning in step S1 specifically includes the following steps: S11. Curvature Calculation: Extract the shape point sequence of the center line of the lane ahead from the navigation map, and optimize the curvature accuracy and eliminate curvature abrupt fluctuations through curve fitting and curvature calculation formula to obtain continuous and smooth road curvature data. S12. Determination of Basic Speed ​​Limit: Based on the formula Calculate the reference vehicle speed at path point s, where, The lateral acceleration threshold is... Let be the road curvature at point s; S13. Speed ​​curve smoothing: The initial reference vehicle speed sequence is smoothed. The smoothing process includes using fifth-order polynomial fitting or Gaussian filtering to generate a reference vehicle speed curve with smooth acceleration and deceleration transitions.

3. The method according to claim 2, characterized in that, The smoothing process in step S13 also includes a dynamic optimization process: based on the model predictive control framework, the deceleration start point and acceleration are dynamically adjusted according to the real-time vehicle status and the trend of curvature change.

4. The method according to claim 1, characterized in that, Step S2 specifically includes the following steps: S21. Constructing a reference line: Based on the lane centerline, the reference line heading is corrected by combining the curve tangent strategy to form a smooth reference line; S22. Coordinate system transformation: Project the vehicle state onto the reference line and transform it to the Frenet coordinate system to decouple the vehicle motion into longitudinal motion along the reference line and lateral motion perpendicular to the reference line. S23. Trajectory Planning: Perform trajectory planning for longitudinal and lateral motion respectively to generate longitudinal and lateral trajectories; S24. Trajectory Fusion and Filtering: The longitudinal and lateral trajectories are fused into a complete trajectory in Cartesian space, and the optimal trajectory is selected by evaluating the cost function.

5. The method according to claim 4, characterized in that, The cost function of the optimization algorithm in step S2 includes at least: a centerline cost term for bringing the vehicle trajectory closer to the lane centerline, a comfort cost term for penalizing the rate of change of trajectory curvature to ensure smoothness, and a safety cost term for preventing the vehicle trajectory from approaching the lane boundary.

6. The method according to claim 4, characterized in that, Step S23 specifically includes: Starting with the vehicle's speed as the initial state, the longitudinal termination state is sampled according to cruise, following, or parking modes. The initial state and the sampled termination state are fitted using a fifth-order polynomial to generate multiple longitudinal trajectories. The lateral offset is sampled around the centerline of the lane. The lateral velocity and acceleration at the lateral termination state are set to zero. A fourth-order polynomial is used to fit and generate the lateral trajectory.

7. The method according to claim 1, characterized in that, Step S3 specifically includes the following steps: S31. Predictive Model Construction: Use the vehicle dynamics model as the predictive model to predict the state of the vehicle over a future period of time. S32, Rolling Optimization Solution: In each control cycle, an optimization problem is solved to find a sequence of optimal front wheel steering angles and acceleration control quantities; wherein, the objective of the optimization problem is to make the predicted vehicle trajectory accurately track the dynamic reference trajectory generated in step S2, while satisfying vehicle dynamics constraints; S33. Feedforward control implementation: By utilizing the future curve curvature information contained in the dynamic reference trajectory, steering action is applied in advance to achieve smooth steering control.

8. A curve trajectory planning and tracking system for intelligent cruise control of vehicles, used to implement the method of any one of claims 1 to 7, characterized in that, include: The perception and positioning module is configured to acquire the vehicle's own status, surrounding environment information, and high-precision positioning data. The curvature prediction and vehicle speed planning module is communicatively connected to the perception and positioning module and is configured to obtain the curvature of the road ahead based on the navigation map and plan a reference vehicle speed curve that meets the requirements. The dynamic trajectory planning module is communicatively connected to the curvature prediction and vehicle speed planning module, and is configured to generate a dynamic reference trajectory based on lane line information and the reference vehicle speed curve by using an optimization algorithm with a curve tangent strategy. The model prediction control module is communicatively connected to the dynamic trajectory planning module and is configured to solve and output steering and driving / braking commands based on the dynamic reference trajectory to control the vehicle to stably follow the trajectory.

9. An electronic device, comprising at least one processor and a memory communicatively connected to said at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, which executes the instructions stored in the memory to implement the curve trajectory planning and tracking method for intelligent cruise control of a vehicle as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing at least one instruction or at least one program, said at least one instruction or at least one program being loaded and executed by a processor to implement the curve trajectory planning and tracking method for intelligent cruise control of a vehicle as described in any one of claims 1 to 7.

Images