An automatic driving path planning method based on tire trajectory band preview

Abstract

This invention relates to the field of intelligent driving and path planning technology, and particularly to an autonomous driving path planning method based on tire trajectory strip pre-aiming. The method includes acquiring vehicle control parameters and constructing tire trajectory strips for each tire accordingly; collecting road environment information in real time through an onboard perception module to construct a two-dimensional grid map for foreign object detection on each tire trajectory strip; classifying the detected foreign objects by category and hazard level, and constructing a hazard intensity grid map; constructing a trajectory strip mask for each tire based on the two-dimensional grid map, and calculating the tire safety risk index for each tire in conjunction with the hazard intensity grid map; and performing path risk assessment and optimization based on the tire safety risk index. This invention, by constructing tire trajectory strips and introducing a pre-aiming mechanism, refines safety considerations to the tire level, which is in direct contact with the road surface, thereby improving tire lifespan and vehicle driving safety.

Description

Technical Field

[0001] This invention relates to the field of intelligent driving and path planning technology, and in particular to an autonomous driving path planning method based on tire trajectory with pre-aiming. Background Technology

[0002] With the rapid development of autonomous driving technology, local path planning, as a core component of vehicle dynamic control, needs to generate safe, smooth driving trajectories that meet vehicle dynamic constraints in real time under complex road environments. Especially in unstructured scenarios such as urban and rural roads and construction zones, sharp objects and protrusions on the road surface can significantly reduce vehicle safety and ride comfort. Current mainstream research and practice mainly focus on macro-path planning and lateral trajectory tracking based on preview control, while refined modeling and processing of the dynamic interaction between tire-level trajectories and poor road surfaces are still lacking.

[0003] Existing technical solutions mainly fall into two categories: one is the path-velocity decoupling planning method based on the Frenet coordinate system. This method samples the lateral and longitudinal trajectories in the Frenet coordinate system established by the global reference line to generate a set of candidate trajectories, and then evaluates and optimizes the trajectories by combining an obstacle cost map and multi-objective cost functions such as safety and comfort. The other is the local obstacle avoidance planning method based on real-time perception. This method relies on multi-sensor fusion to construct an environmental cost map and performs online rolling optimization that satisfies vehicle dynamics constraints in the original coordinate system. However, these methods generally suffer from a key drawback: the planning process usually uses the vehicle's center of gravity as the control point, failing to fully consider the safety interaction between the tire contact area (i.e., the tire track) and local road hazards.

[0004] When vehicles travel in areas with sharp objects, bumps, potholes, gravel, or other road debris, failure to effectively detect and avoid these dangerous areas can easily lead to tire damage or even vehicle instability. Therefore, there is an urgent need to develop a local path planning method that integrates real-time identification of road hazards with geometric constraints on tire trajectories. Summary of the Invention

[0005] To address the above problems, this invention provides an autonomous driving path planning method based on tire trajectory with anticipation, comprising the following steps:

[0006] S1. Obtain vehicle control parameters and construct tire trajectory strips for each tire accordingly;

[0007] S2. Real-time road environment information of the vehicle's movement direction is collected through the vehicle-mounted sensing module, and a two-dimensional grid map is constructed to detect foreign objects in each tire track.

[0008] S3. Classify the detected foreign objects according to their categories and hazard levels, and construct a hazard intensity grid map;

[0009] S4. Construct a track mask for each tire based on a two-dimensional grid map, and calculate the tire safety risk index for each tire by combining it with the hazard intensity grid map;

[0010] S5. Conduct route risk assessment and optimization based on tire safety risk indicators.

[0011] The beneficial effects of this invention are:

[0012] This invention refines safety considerations down to the tire layer, which is in direct contact with the road surface, by constructing a tire track and introducing a pre-aiming mechanism, thereby improving tire lifespan and vehicle driving safety. Furthermore, this invention achieves accurate differentiation of different threat levels by finely classifying hazardous materials (such as sharp objects, protrusions, potholes, and surface deposits) and assigning them dynamic hazard level parameters. Attached Figure Description

[0013] Figure 1 This is a flowchart of an autonomous driving path planning method based on tire trajectory with pre-aiming, according to the present invention. Detailed Implementation

[0014] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0015] Current path planning for autonomous vehicles typically uses the vehicle's center of gravity or circumbound bounding box as the basic unit of consideration, primarily focusing on macroscopic collision avoidance with obstacles, such as collision risks between the vehicle and pedestrians or other vehicles. However, for tire-level microscopic damage risks that affect driving safety and ride comfort (such as nails, broken glass, or potholes on the road), current methods still have significant limitations at the perception and decision-making levels. Specifically, at the perception level, conventional obstacle detection technologies struggle to effectively distinguish between large obstacles that pose a collision threat to the vehicle body and small road hazards that pose a risk of running over the tires. At the decision-making level, path planning methods based on the entire vehicle's envelope cannot accurately assess and optimize the independent risks faced by each tire. This may result in a planned path that avoids large potholes but causes individual tires to run over sharp objects, leading to serious safety accidents such as tire blowouts.

[0016] Therefore, this invention aims to solve the following technical problem: how to implement a high-precision path planning method oriented towards tire-level stability and safety, enabling autonomous driving systems to proactively identify and avoid local road surface risks that may damage tires, much like an experienced driver. To achieve the above objective, this invention proposes an autonomous driving path planning method based on tire trajectory with pre-aiming, such as... Figure 1 As shown, it specifically includes:

[0017] S1. Obtain vehicle control parameters and construct tire trajectory strips for each tire accordingly.

[0018] In some embodiments, step S1 includes:

[0019] S11. Obtain the vehicle center coordinates and construct the motion trajectory of each tire based on the vehicle control parameters; the vehicle control parameters include the vehicle wheelbase, vehicle track width, and vehicle heading angle.

[0020] For example, in the vehicle coordinate system (with the vehicle's front facing the positive X-axis), the coordinates of the vehicle center (x) are obtained. i ,y i Simultaneously set the vehicle wheelbase L, vehicle track W, and vehicle heading angle θ. i The trajectory of the left front wheel is obtained as follows:

[0021]

[0022]

[0023] Among them, (x) lf ,y lf This represents the center coordinates of the left front wheel. The design for the right front wheel, left rear wheel, and right rear wheel is similar, requiring only minor modifications. , The positive and negative relationships before.

[0024] S12. For each tire, obtain the coordinates of the tire's contact center with the ground at different time points based on the motion trajectory, and smoothly connect these coordinates in time order to obtain the tire trajectory;

[0025] S13. Extend each tire track to the left and right by half the width of the tire to obtain the tire track band.

[0026] Tire track is used to characterize the actual contact area between the tire and the road surface, and is an important basis for subsequent safety risk assessment.

[0027] S2. Real-time road environment information of the vehicle's movement direction is collected through the vehicle-mounted sensing module, and a two-dimensional grid map is constructed based on this to detect foreign objects in each tire track.

[0028] In some embodiments, step S2 includes:

[0029] S21. Extract the geometric equations of road boundaries (curb, guardrail) and lane lines (solid lines, dashed lines) based on road environment information, and construct the closure set D of the drivable area in the vehicle coordinate system accordingly. road , represented as:

[0030]

[0031] In the formula, f represents the spatial coordinates of a point in the vehicle body coordinate system. left (x) represents the geometric equation of the left boundary of the road, f right (x) represents the geometric equation of the right boundary of the road; x pre This indicates the aiming distance, which defines the longitudinal boundary of the drivable area. In lane line constraint strategies, solid lane lines and physical boundaries (i.e., road boundaries) are subject to hard constraints, prohibiting vehicles from crossing the lines; while dashed lane lines are subject to soft constraints, allowing temporary crossing in emergency situations.

[0032] S22. At the pre-aiming distance x pre Within the specified range, the closure set of the drivable area is discretized into a two-dimensional raster map, with each raster cell having an area of... It is used to uniformly describe road space, tire track strips, and hazardous material distribution information, providing a unified calculation framework for subsequent overlap calculations.

[0033] S23. In the vehicle coordinate system, regarding the pre-aiming distance x pre The system tracks the tire tracks of each tire and identifies whether foreign objects exist in the overlapping grid cells. If foreign objects are found, their characteristic data is extracted. This characteristic data includes the physical properties of the foreign objects (such as shape, size, material, and surface characteristics), which are used to classify the foreign objects and determine their hazard level parameters.

[0034] S3. Classify the detected foreign objects according to their categories and hazard levels, and construct a hazard intensity grid map.

[0035] In some embodiments, step S3 includes:

[0036] S31. Use machine learning methods to extract and classify the features of foreign objects to obtain their physical properties.

[0037] Specifically, the categories of foreign objects include: sharp objects (such as nails), protruding objects (such as small stones), potholes (such as local potholes or road collapses), and localized dangerous surfaces (such as icy areas).

[0038] S32. Based on the vehicle's operating status and the physical properties of the foreign object, establish a fuzzy continuous hazard level for the foreign object;

[0039] Specifically, the physical properties of foreign objects include material, size, shape, and surface characteristics; vehicle operating status includes vehicle speed, tire pressure, etc.

[0040] Specifically, based on the vehicle's operating status and the physical properties of the foreign object, a fuzzy continuous hazard level is established for the foreign object, including:

[0041] S321. Different types of foreign objects exhibit different risk levels due to their varying physical properties. For example, sharp foreign objects typically have a higher risk level once they are present; while protruding foreign objects, pits, and locally hazardous surfaces require a comprehensive assessment based on their key physical parameters (such as protrusion height, depth, and coefficient of friction) to determine the degree of risk. Therefore, a basic risk level within the range of [0,1] is first determined based on the physical properties of the foreign object.

[0042] S322. Based on the real-time vehicle operating status (vehicle speed, tire pressure, etc.), a correction coefficient K in the range of [0,1] is obtained. The correction coefficient K is multiplied by the basic hazard level to obtain the fuzzy continuous hazard level.

[0043] Specifically, the value of the correction factor K needs to be determined by comprehensively considering the vehicle's operating conditions and the type of foreign object. For protruding objects, potholes, and sharp objects, the K value should increase accordingly with increasing vehicle speed because increased vehicle speed will exacerbate the impact load and stress concentration on the tires. Conversely, when driving straight on low-friction surfaces such as ice, increasing vehicle speed can actually help reduce tire slippage and friction loss, so the K value can be appropriately reduced. In addition, both excessively high and low tire pressure will lead to deterioration of tire contact characteristics and uneven stress distribution, so the K value needs to be increased accordingly to reflect the potential risks.

[0044] S33. Map the hazard level of foreign objects to their corresponding grid cells to construct a hazard intensity grid map H(k), thereby forming a road hazard distribution model within the target range.

[0045] S4. Construct a track mask for each tire based on a two-dimensional grid map, and calculate the tire safety risk index for each tire by combining it with the hazard intensity grid map.

[0046] In some embodiments, step S4 includes:

[0047] S41. Construct a track mask for each tire based on a two-dimensional grid map, including:

[0048] Construct a trajectory strip mask function M for the j-th tire (j∈{1,2,3,4}). j ; where M j(k) represents the track band mask value of the j-th tire in the k-th grid cell; if the k-th grid cell is located within the tire track band of the j-th tire, then M j (k)=1, otherwise M j (k)=0.

[0049] S42. Overlay the track mask of each tire with the hazard intensity grid map to obtain the overlapping grid map, and calculate the effective hazard intensity overlap area within the aiming distance, expressed as:

[0050]

[0051] In the formula, A jeff This represents the effective danger intensity overlap area of ​​the j-th tire within the pre-aiming distance.

[0052] The effective hazard intensity overlap area comprehensively reflects the degree of spatial overlap and hazard intensity between the tire track and the foreign object, and is an important quantitative indicator of tire safety risk.

[0053] S43. For each tire, calculate the overlap length between the tire and the foreign object in its tire track based on the projection range of its corresponding overlapping grid map in the vehicle's driving direction, and estimate the contact time between the tire and the foreign object in conjunction with the vehicle's driving speed.

[0054] S44. Calculate the tire safety risk index for each tire based on the overlap length and contact time, expressed as follows:

[0055]

[0056]

[0057] In the formula, ω1 and ω2 represent weighting coefficients, and ω1 + ω2 = 1. K' represents the number of grid cells. This indicates the overlapping risk of normalized hazard intensity. This indicates the sequence of hazardous material crushing and the impact risk, in relation to the contact time t. i related. It depends on the type of foreign object and the contact time t i Certainly. For example, the faster the driving speed (i.e., the shorter the contact time) with protruding objects or potholes, the greater the damage to the tires. The higher the value, the more significant the tire damage will be, especially on icy areas on straightaways where the contact time is longer. The larger the value.

[0058] S5. Conduct route risk assessment and optimization based on tire safety risk indicators.

[0059] In some embodiments, step S5 includes:

[0060] S51. Determine whether the tire safety risk index of all tires is less than the preset risk threshold R. th If yes, it indicates that the interaction risk between each tire and the danger zone at the current aiming distance is within a safe and controllable range, and path replanning does not need to be initiated. If not, path replanning is triggered, and step S52 is executed;

[0061] S52. Implement partial replanning for the four-child policy, including:

[0062] Within the target distance, a set of candidate paths is generated using the current path as the seed through a sampling-based method;

[0063] The tire trajectory distribution of each tire is predicted synchronously under each candidate path, forming a set of four tire trajectory bands that satisfy the rigid body kinematic constraints of the vehicle.

[0064] Calculate the cost function J for each candidate path. total , represented as:

[0065]

[0066] In the formula, W safe W smooth W rule These represent the weights of the security layer, smoothing layer, and rule layer, respectively; J safe J smooth J rule These represent the cost terms for the safety layer, smoothness layer, and rule layer, respectively. In calculating the cost function J... total Previously, all cost terms were normalized and mapped to the same order of magnitude.

[0067] The candidate path corresponding to the minimum cost function value is selected as the target path.

[0068] In some embodiments, the cost items are calculated as follows:

[0069] Security layer cost term J safe The formula used to assess the risk of interaction between the tire track strip and the hazardous area is:

[0070]

[0071] In the formula, R j Let λ represent the tire safety risk index of the j-th tire ∈ {1,2,3,4}. j These represent weighting coefficients. 1, 2, 3, and 4 represent the four tire indices: front left, front right, rear left, and rear right, respectively.

[0072] Smooth Layer Cost Item J smoothThis is used to evaluate the dynamic stability, ride comfort, and path geometric smoothness of a vehicle. Since the candidate path consists of a series of discrete sampling points, a discrete summation method is used to calculate the cumulative fluctuation of the candidate path within the aiming time T. The calculation formula is as follows:

[0073]

[0074] In the formula, i represents the discrete sampling point index of the candidate path, and N represents the total number of discrete sampling points; This represents the change in tire vertical load at the i-th discrete sampling point of the candidate path in adjacent time steps. This term is restricted to prevent drastic fluctuations in tire grip caused by sharp steering, thereby ensuring vehicle handling stability. α represents the variance term at the i-th discrete sampling point of the candidate path. i This represents the tire slip angle corresponding to the i-th discrete sampling point of the candidate path. This represents the average tire slip angle. This value is used to penalize excessive sideslip tendency and ensure that the vehicle does not become unstable during obstacle avoidance. This represents the rate of change of path curvature at the i-th discrete sampling point of the candidate path. This term is restricted to ensure that the generated trajectory is geometrically continuous and smooth, and to avoid passenger discomfort caused by "sudden steering". These are the internal normalized weighting coefficients for each smoothness indicator.

[0075] Rule layer cost term J rule It is used to constrain vehicles to comply with road traffic rules and physical boundary restrictions, and consists of two parts: hard constraints and soft constraints.

[0076]

[0077] In the formula, This indicates a penalty for violating hard boundaries (a veto item) when the tire track touches the physical boundaries of the road (such as curbs or guardrails). This candidate path is directly eliminated. obs,i This represents the Euclidean distance between the tire track and the nearest soft boundary at the i-th discrete sampling point of the candidate path. A potential field is established using an inverse proportional function. The closer the distance, the higher the cost, guiding the vehicle to drive in the center of the safe area. This is the lane centerline tracking error term, used to guide the vehicle to stay centered in the lane. i This represents the lateral coordinates of the vehicle's centroid at the i-th discrete sampling point on the candidate path; y ref,i This represents the lateral position coordinates of the lane centerline at the i-th discrete sampling point of the candidate path; To prevent tiny positive numbers with a denominator of zero. soft k centerThis represents the normalized weight coefficients within the rule layer.

[0078] Specifically, d obs,i It is obtained by calculating the Euclidean distance between the four tire track strips and their nearest soft boundary element (the dashed lane line on the road), and then selecting the minimum Euclidean distance.

[0079] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An autonomous driving path planning method based on tire trajectory with pre-aiming, characterized in that, Includes the following steps: S1. Obtain vehicle control parameters and construct tire trajectory strips for each tire accordingly; S2. Real-time road environment information of the vehicle's movement direction is collected through the vehicle-mounted sensing module, and a two-dimensional grid map is constructed to detect foreign objects in each tire track. S3. Classify the detected foreign objects according to their categories and hazard levels, and construct a hazard intensity grid map; S4. Construct a track mask for each tire based on a two-dimensional grid map, and calculate the tire safety risk index for each tire by combining it with the hazard intensity grid map; S5. Conduct route risk assessment and optimization based on tire safety risk indicators.

2. The autonomous driving path planning method based on tire trajectory with pre-aiming according to claim 1, characterized in that, Step S1 includes: S11. Obtain the vehicle center coordinates and construct the motion trajectory of each tire based on the vehicle control parameters; the vehicle control parameters include the vehicle wheelbase, vehicle track width, and vehicle heading angle; S12. For each tire, obtain the coordinates of the tire's contact center with the ground at different time points based on the motion trajectory, and smoothly connect these coordinates in time order to obtain the tire trajectory; S13. Extend each tire track to the left and right by half the width of the tire to obtain the tire track band.

3. The autonomous driving path planning method based on tire trajectory with pre-aiming according to claim 1, characterized in that, Step S2 includes: S21. Extract the road boundary geometric equations and lane line equations based on the road environment information, and construct the drivable area closure set D in the vehicle coordinate system accordingly. road , is represented as: ， In the formula, f represents the spatial coordinates of a point in the vehicle body coordinate system. left (x) represents the geometric equation of the left boundary of the road, f right (x) represents the geometric equation of the right boundary of the road, x pre Indicates the aiming distance; S22. At the pre-aiming distance x pre Within the specified range, the closure set of drivable areas is discretized into a two-dimensional raster map, where the area of ​​each raster cell is... ; S23. In the vehicle coordinate system, regarding the pre-aiming distance x pre The system identifies the tire track bands of each tire and determines whether there are foreign objects in the grid cells that overlap with these tire track bands. If foreign objects are found, the system extracts their feature data.

4. The autonomous driving path planning method based on tire trajectory with pre-aiming as described in claim 1, characterized in that, Step S3 includes: S31. Use machine learning methods to extract and classify the features of foreign objects to obtain their physical properties; S32. Based on the vehicle's operating status and the physical properties of the foreign object, establish a fuzzy continuous hazard level for the foreign object; S33. Map the hazard level of foreign objects to their corresponding grid cells to construct a hazard intensity grid map. .

5. The autonomous driving path planning method based on tire trajectory with pre-aiming according to claim 4, characterized in that, Based on the vehicle's operating status and the physical properties of the foreign object, a fuzzy continuous hazard level is established for the foreign object, including: S321. Different types of foreign objects present different risk levels due to their differences in physical properties. A basic hazard level in the range of [0,1] is determined based on the physical properties of the foreign object. S322. Based on the vehicle's operating status, obtain a correction coefficient K in the range [0,1]. Multiply the correction coefficient K by the basic hazard level to obtain the fuzzy continuous hazard level.

6. The autonomous driving path planning method based on tire trajectory with pre-aiming according to claim 1, characterized in that, Step S4 includes: S41. Construct a track mask for each tire based on a two-dimensional grid map, including: Construct a trajectory band mask function M for the j-th tire. j Among them, M j (k) represents the track band mask value of the j-th tire in the k-th grid cell; if the k-th grid cell is located within the tire track band of the j-th tire, then M j (k)=1, otherwise M j (k)=0; S42. Overlay the track mask of each tire with the hazard intensity grid map to obtain the overlapping grid map, and calculate the effective hazard intensity overlap area within the aiming distance, expressed as: ， In the formula, A jeff This represents the effective danger intensity overlap area of ​​the j-th tire within the pre-aiming distance; S43. For each tire, calculate the overlap length between the tire and the foreign object in its tire track based on the projection range of its corresponding overlapping grid map in the vehicle's driving direction, and estimate the contact time between the tire and the foreign object in conjunction with the vehicle's driving speed. S44. Calculate the tire safety risk index for each tire based on the overlap length and contact time, expressed as follows: ， ， In the formula, ω1 and ω2 represent weighting coefficients, ω1 + ω2 = 1; K' represents the number of grid cells. This indicates the overlapping risk of normalized hazard intensity. This indicates the sequence of hazardous material crushing and the impact risk, in relation to the contact time t. i related.

7. The autonomous driving path planning method based on tire trajectory with pre-aiming according to claim 1, characterized in that, Step S5 includes: S51. Determine whether the tire safety risk index of all tires is less than the preset risk threshold R. th If yes, then there is no need to start path replanning; if not, then path replanning is triggered, and step S52 is executed. S52. Implement partial replanning for the four-child policy, including: Within the target distance, a set of candidate paths is generated using the current path as the seed through a sampling-based method; The tire trajectory distribution of each tire is predicted synchronously under each candidate path, forming a set of four tire trajectory bands that satisfy the rigid body kinematic constraints of the vehicle. Calculate the cost function J for each candidate path. total , is represented as: ， In the formula, W safe W smooth W rule These represent the weights of the security layer, smoothing layer, and rule layer, respectively; J safe J represents the safety layer cost term, used to assess the risk of interaction between the tire track strip and the danger zone; smooth J represents the ride comfort layer cost term, used to evaluate the vehicle's dynamic stability, ride comfort, and path geometric smoothness; rule This represents the cost term at the rule layer, used to constrain vehicles to comply with road traffic rules and physical boundary restrictions; The candidate path corresponding to the minimum cost function value is selected as the target path.

8. The autonomous driving path planning method based on tire trajectory with pre-aiming according to claim 7, characterized in that, Security layer cost term J safe The calculation formula: ， In the formula, R j λ represents the tire safety risk index of the j-th tire. j Indicates the weighting coefficient; Smooth Layer Cost Item J smooth The calculation formula: ， In the formula, i represents the discrete sampling point index of the candidate path, and N represents the total number of discrete sampling points; α represents the change in tire vertical load at the i-th discrete sampling point of the candidate path in adjacent time steps. i This represents the tire slip angle corresponding to the i-th discrete sampling point of the candidate path. This represents the average tire slip angle. This represents the rate of change of path curvature at the i-th discrete sampling point of the candidate path. Indicates the internal normalized weighting coefficient; Rule layer cost term J rule Calculation formula: ， In the formula, This indicates a penalty for a hard boundary violation, which occurs when a tire track touches the physical boundary of the road. The candidate path is directly eliminated; d obs,i This represents the Euclidean distance from the tire track to the nearest soft boundary at the i-th discrete sampling point of the candidate path; y i This represents the lateral coordinates of the vehicle's centroid at the i-th discrete sampling point on the candidate path; y ref,i This represents the lateral position coordinates of the lane centerline at the i-th discrete sampling point of the candidate path; k represents a small positive number. soft k center This represents the normalized weighting coefficient.

Images