Local path planning method and system for unmanned vehicle based on feasible path

Abstract

The application discloses a local path planning method and system for an unmanned vehicle based on feasible paths, and comprises the following steps: acquiring road information, self-vehicle information and other-vehicle information; determining a static impassable area according to static obstacles, determining a three-section dynamic impassable area according to dynamic obstacles, and obtaining a passable area of the vehicle by removing the static impassable area and the three-section dynamic impassable area from the boundary of the passable area; planning the speed of the vehicle to obtain the real-time speed of the vehicle; calculating various safety distances in the driving process of the vehicle; making a lane-changing decision to determine whether to drive in a lane or to keep driving in the current lane; obtaining the planned path points and smoothing the path to complete the path planning, and the vehicle drives according to the current planned path; and the application has the advantages that the feasibility, timeliness and safety requirements are met simultaneously.

Description

Technical Field

[0001] This invention relates to the field of autonomous driving technology, and more specifically to a method and system for local path planning of unmanned vehicles based on feasible paths. Background Technology

[0002] Autonomous driving technology achieves environmental perception through multi-sensor fusion (LiDAR, cameras, radar, etc.), and autonomous driving is accomplished by a decision-making and planning system (global / local path planning) and a control system. Local path planning is one of the core technologies of autonomous driving.

[0003] The significance of local path planning in autonomous driving lies in generating safe, smooth, and traffic-compliant local trajectories in real time in dynamic and complex environments. This ensures that vehicles can avoid obstacles, maintain comfort, and efficiently reach their destination. Local path planning directly affects the reliability, riding experience, and traffic efficiency of autonomous driving.

[0004] The current challenges in local path planning are: the feasibility of selecting the target location in local planning, the timeliness of vehicle tracking of the original path, and the safety of the vehicle path tracking process.

[0005] 1. Regarding feasibility, existing technologies use the midpoint of the desired lane line as the target location, ignoring the fact that other vehicles may not be in the exact center of the road in reality, and the positions of various obstacles are also random. Therefore, the planned path may not meet actual driving needs. Setting the safety distance too large increases driving costs, while setting it too small lacks safety. For example, the local trajectory planning method and device for intelligent vehicles disclosed in Chinese Patent Publication No. CN 106114507 A uses the center of the lane-changing lane as the target location. If there are small obstacles at the edge of the lane-changing lane that do not affect vehicle passage, this method may judge this situation as not meeting the lane-changing conditions because the safety distance is set too large.

[0006] 2. Regarding timeliness, existing technologies plan at a fixed speed, which does not conform to real-world driving scenarios. In actual driving, unexpected situations can alter the real-time speed, causing the error between the entire path tracking time and the expected time to accumulate and result in poor timeliness. For example, a local path planning method for intelligent vehicles in a structured environment disclosed in Chinese Patent Publication No. CN112362074A directly sets the vehicle speed as Va without planning the speed, thus easily leading to the aforementioned problem of not conforming to real-world driving scenarios.

[0007] 3. Regarding safety, existing technologies plan according to fixed safety distances without considering the speed of the vehicle itself and other vehicles. The required safety distance varies under different conditions, and if the distance is too large, it increases the cost of driving. For example, a vehicle local path planning method and device disclosed in Chinese Patent Publication No. CN 116125987 A selects multiple sets of control points by discrete sampling at fixed distances, ignoring the real-time safety distance issue. When the discrete distance is set too small, there is a risk of collision due to the influence of hydrodynamics; when the discrete distance is set too large, it increases the cost of driving. Summary of the Invention

[0008] The technical problem to be solved by this invention is that the existing local path planning methods for autonomous driving processes cannot simultaneously meet the requirements of feasibility, timeliness and safety.

[0009] This invention solves the above-mentioned technical problems through the following technical means: a local path planning method for autonomous vehicles based on feasible paths, comprising:

[0010] S1. Obtain road information, information about your own vehicle and other vehicles;

[0011] S2. Determine the static impassable area based on static obstacles, and determine the three-segment dynamic impassable area based on dynamic obstacles. The first segment is the position where the dynamic obstacle stops at its maximum deceleration, the second segment is the position reached by the dynamic obstacle traveling at a constant speed within the same time period, and the third segment is the position reached by the dynamic obstacle at its maximum acceleration within the same time period. The area formed between the rear of the vehicle in the first segment and the front of the vehicle in the third segment is the three-segment dynamic impassable area. The same time period refers to the time when the dynamic obstacle stops at its maximum deceleration. Within the boundaries of the passable area, remove the static impassable area and the three-segment dynamic impassable area to obtain the vehicle's passable area.

[0012] S3. Perform speed planning for the vehicle and obtain the vehicle's real-time speed;

[0013] S4. Calculate the various safe distances during the vehicle's travel;

[0014] S5. Based on the information obtained from S1-S4, make a lane-changing decision and decide whether to change lanes or stay in the current lane and follow the vehicle in front.

[0015] S6. Obtain the planned waypoints and smooth the path to complete the path planning. The vehicle then travels according to the currently planned path.

[0016] Furthermore, S2 includes:

[0017] S2.1 Obtain the boundary of the passable area through the road boundary, central median or central divider;

[0018] S2.2. Based on the speed of other vehicles, identify static and dynamic obstacles;

[0019] S2.3. Based on the shape of the static obstacle, which is equivalent to a rectangle parallel to the lane line, determine the static impassable area;

[0020] S2.4. Based on the speed and acceleration / deceleration capabilities of dynamic obstacles, determine the three-segment dynamic impassable area;

[0021] S2.5 Within the boundary of the passable area, remove the static non-passable area and the three-segment dynamic non-passable area. If the distance between the obstacle and the boundary of the passable area is less than the minimum safe distance, then remove the area between the boundary of the passable area and the obstacle, and finally obtain the passable area of ​​the vehicle.

[0022] Furthermore, S3 includes:

[0023] S3.1 Curvature denormalization: the maximum curvature is equivalent to 0, and the minimum curvature is equivalent to 1; set the default minimum vehicle speed V0, the default maximum vehicle speed V0+V1, where V1 is the difference between the default maximum vehicle speed and the default minimum vehicle speed. Given the maximum turning radius Lmax, the minimum turning radius Lmin, and the real-time position curvature of the globally planned vectorized path K, the minimum curvature Kmin=1 / Lmax, and the maximum curvature Kmax=1 / Lmin; denormalize the curvature Knorm=(Kmax-K) / (Kmax-Kmin).

[0024] S3.2, The initial default velocity V = V0 + V1 at each position of the vectorized path in global planning. Knorm;

[0025] S3.3 Given the total length of the path as L, calculate the default travel time T required to travel at the default speed, based on the initial default speed V.

[0026] S3.4 Solving for the real-time velocity VZ=β V, where β is the real-time speed multiplier and β = {T} (L-Lk)} / {L (T-Tk)}, where Tk is the actual travel time and Lk is the actual travel distance. Set the upper and lower limits of VZ to constrain it.

[0027] S3.5. Based on the lane change decision, decelerate, maintain a constant speed, or accelerate during lane change and following.

[0028] Furthermore, S4 includes:

[0029] S4.1. Based on the real-time speed VZ and the maximum deceleration AZ2 of the vehicle, calculate the safe distance LA between the front of the vehicle and the rear of the vehicle in front of it in the same lane. Where T1 is the system decision time; based on the real-time speed VZ of the vehicle, the speed VC of the vehicle in front in the changing lane, and the maximum deceleration AC2 of the vehicle in front in the changing lane, the safe distance LC between the front of the vehicle and the rear of the vehicle in front in the changing lane is calculated. Where T2 is the lane-changing time; based on the real-time speed VZ of the vehicle, the speed VD of the vehicle following in the lane-changing lane, and the maximum acceleration AD1 of the vehicle following in the lane-changing lane, the safe distance LD between the rear of the vehicle and the front of the vehicle following in the lane-changing lane is calculated. ;

[0030] S4.2 Calculate the safe distance perpendicular to the lane lines: Based on the real-time speed VZ of the vehicle and the speeds VX of other vehicles in the left and right lanes, where VX is the speed of the vehicle in front of the vehicle in its lane VA, the speed of the vehicle behind the vehicle in its lane VB, and the speed of the vehicle in front of the vehicle in the lane changing lane VC or the speed of the vehicle behind the vehicle in the lane changing lane VD, calculate the safe distance LS between the vehicle and other vehicles in the left and right lanes perpendicular to the lane lines. Where K1 and K2 are the first time parameter and the second time parameter, respectively.

[0031] Furthermore, S5 includes:

[0032] S5.1 Detect whether there are other vehicles within the forward detection range LF and the rear detection range LR of the vehicle itself;

[0033] S5.2 Obtain the distance SA between the vehicle and the vehicle in front of it in its lane, the distance SC between the vehicle and the vehicle in front of it in its lane-changing lane, and the distance SD between the vehicle and the vehicle behind it in its lane-changing lane.

[0034] S5.3 If SA > LF, maintain the original speed in the lane, which is the planned real-time speed VZ; if there is a vehicle A in front of the vehicle in the lane within the forward detection distance LF, and LA < SA ≤ LF, maintain the constant speed in the lane, which is the vehicle's real-time speed VZ when vehicle A in front of the vehicle in the lane was first detected; if SA ≤ LA, compare the vehicle's real-time speed VZ with the speed VA of the vehicle in front of the vehicle in the lane. Then maintain lane following, with a speed equal to the speed VA of the vehicle in front in the same lane, where K3 and K4 are the first and second proportional coefficients, respectively, and K3 < 1, K4 > 1; if If so, maintain the original speed in the lane, which is the planned real-time speed VZ; if Then prepare to change lanes;

[0035] S5.4 Lane Change Direction Decision: Determine the width of the narrowest passable area on both the left and right sides of the vehicle, and choose to change lanes to the wider lane; if the front and rear safety distance to the wider lane is insufficient, choose to change lanes to the narrower lane; if the front and rear safety distance to the narrower lane is insufficient, maintain lane following.

[0036] S5.5 If there are no other vehicles within the forward detection distance LF and backward detection distance LR of the lane changing lane (i.e., SC > LF and SD > LR), then the vehicle changes lanes at its original speed, which is the planned real-time speed VZ. If LC < SC ≤ LF and SD > LR, then the vehicle changes lanes at an accelerated speed, which is K5. VZ, where K5 is the third proportionality coefficient and K5 > 1; if LD < SD ≤ LR and SC > LF, then change lanes and accelerate to a speed of K6. VZ, where K6 is the fourth proportionality coefficient and K6 > 1; if LC < SC ≤ LF and LD < SD ≤ LR, then change lanes and accelerate to a speed of K7. VZ, where K7 is the fifth proportional coefficient and K7>1; if SC≤LC, or SD≤LD, then maintain lane following and drive at the speed of the vehicle in front, where K5<K6<K7.

[0037] Furthermore, S5.2 includes:

[0038] Define the X-axis as the direction parallel to the lane lines, the Y-axis as the direction perpendicular to the lane lines, and the origin as the geometric center of the vehicle's starting position. Calculate the distance between the vehicle and the vehicle A in front of it in the same lane. ,in, Let X be the geometric center of the vehicle ahead in the lane. The X-axis coordinate of the vehicle's geometric center. The length of the vehicle in front of you in your lane. The length of the vehicle; the distance between the vehicle and the vehicle in front (C) in the lane where the vehicle changes lanes. ,in, The X-axis coordinate of the geometric center of the vehicle ahead in the lane-changing lane is given. The distance between the vehicle in front of it in the lane changing lane and the vehicle behind it in the lane changing lane is D. ,in, The X-axis coordinate of the geometric center of the vehicle after changing lanes. This refers to the length of the vehicle following the lane change.

[0039] Furthermore, S6 includes:

[0040] S6.1 In a lane-changing scenario, if the speed of vehicle A in the preceding lane is zero, the coordinates of two feature points are obtained within the passable area as (X2, Y2) and (X3, Y3), where X2 is the X-axis coordinate of vehicle A in the preceding lane at the geometric center of the static impassable area, and Y2 is the Y-axis coordinate of vehicle A in the preceding lane at the geometric center of the static impassable area relative to the lane-changing direction. The sum of, For the width of the vehicle, X3 is the width of the vehicle in front of the vehicle in the lane; X3 is the X-axis coordinate of the front position of vehicle A in the static non-passable area; Y3 is the Y-axis coordinate of the center position of the lane changing lane.

[0041] S6.2 In lane-changing scenarios, if the speed of vehicle A in the preceding lane is not zero, the coordinates of three feature points are obtained within the passable area as (X1', Y1'), (X2', Y2'), and (X3', Y3'). Here, X1' is the X-axis coordinate of the rear of vehicle A in the first segment of the three-segment dynamic non-passable area, and Y1' is the Y-axis coordinate of vehicle A at the geometric center of the three-segment dynamic non-passable area relative to the lane-changing direction. The sum of X and Y; X2' is the X-axis coordinate of the geometric center of vehicle A in the second segment of the three-segment dynamic no-passing zone, and Y2' is the Y-axis coordinate of the geometric center of vehicle A in the three-segment dynamic no-passing zone relative to the lane-changing direction. X3' is the X-axis coordinate of the front of vehicle A in the three-segment dynamic non-passable area, and Y3' is the Y-axis coordinate of the center position of the lane-changing lane.

[0042] S6.3. Use cubic spline interpolation to make the feature points continuous, obtain the local path trajectory, and complete the path planning.

[0043] Furthermore, this method is executed once every 10ms according to the complete process of S1-S6 to plan the local path trajectory under the current environmental state.

[0044] This invention also provides a local path planning system for autonomous vehicles based on feasible paths, comprising:

[0045] The data acquisition module is used to acquire road information, information about the vehicle itself, and information about other vehicles.

[0046] The passable area determination module is used to determine static impassable areas based on static obstacles and three-segment dynamic impassable areas based on dynamic obstacles. The first segment is the position where the dynamic obstacle stops at its maximum deceleration, the second segment is the position reached by the dynamic obstacle traveling at a constant speed within the same time period, and the third segment is the position reached by the dynamic obstacle at its maximum acceleration within the same time period. The area formed between the rear of the vehicle in the first segment and the front of the vehicle in the third segment is the three-segment dynamic impassable area. The same time period refers to the time when the dynamic obstacle stops at its maximum deceleration. Within the boundaries of the passable area, the static impassable area and the three-segment dynamic impassable area are removed to obtain the vehicle's passable area.

[0047] The speed planning module is used to plan the vehicle's speed and obtain the vehicle's real-time speed.

[0048] The safe distance calculation module is used to calculate various safe distances during vehicle travel.

[0049] The lane change decision module is used to make lane change decisions based on the information obtained from the data acquisition module and the safe distance calculation module, deciding whether to change lanes or stay in the current lane and follow the vehicle.

[0050] The path planning module is used to obtain planned waypoints and smooth the path to complete the path planning. The vehicle then travels according to the currently planned path.

[0051] Furthermore, the passage area determination module is also used for:

[0052] S2.1 Obtain the boundary of the passable area through the road boundary, central median or central divider;

[0053] S2.2. Based on the speed of other vehicles, identify static and dynamic obstacles;

[0054] S2.3. Based on the shape of the static obstacle, which is equivalent to a rectangle parallel to the lane line, determine the static impassable area;

[0055] S2.4. Based on the speed and acceleration / deceleration capabilities of dynamic obstacles, determine the three-segment dynamic impassable area;

[0056] S2.5 Within the boundary of the passable area, remove the static non-passable area and the three-segment dynamic non-passable area. If the distance between the obstacle and the boundary of the passable area is less than the minimum safe distance, then remove the area between the boundary of the passable area and the obstacle, and finally obtain the passable area of ​​the vehicle.

[0057] Furthermore, the speed planning module is also used for:

[0058] S3.1 Curvature denormalization: the maximum curvature is equivalent to 0, and the minimum curvature is equivalent to 1; set the default minimum vehicle speed V0, the default maximum vehicle speed V0+V1, where V1 is the difference between the default maximum vehicle speed and the default minimum vehicle speed. Given the maximum turning radius Lmax, the minimum turning radius Lmin, and the real-time position curvature of the globally planned vectorized path K, the minimum curvature Kmin=1 / Lmax, and the maximum curvature Kmax=1 / Lmin; denormalize the curvature Knorm=(Kmax-K) / (Kmax-Kmin).

[0059] S3.2, The initial default velocity V = V0 + V1 at each position of the vectorized path in global planning. Knorm;

[0060] S3.3 Given the total length of the path as L, calculate the default travel time T required to travel at the default speed, based on the initial default speed V.

[0061] S3.4 Solving for the real-time velocity VZ=β V, where β is the real-time speed multiplier and β = {T} (L-Lk)} / {L (T-Tk)}, where Tk is the actual travel time and Lk is the actual travel distance. Set the upper and lower limits of VZ to constrain it.

[0062] Furthermore, the safe distance calculation module is also used for:

[0063] S4.1. Based on the real-time speed VZ and the maximum deceleration AZ2 of the vehicle, calculate the safe distance LA between the front of the vehicle and the rear of the vehicle in front of it in the same lane. Where T1 is the system decision time; based on the real-time speed VZ of the vehicle, the speed VC of the vehicle in front in the changing lane, and the maximum deceleration AC2 of the vehicle in front in the changing lane, the safe distance LC between the front of the vehicle and the rear of the vehicle in front in the changing lane is calculated. Where T2 is the lane-changing time; based on the real-time speed VZ of the vehicle, the speed VD of the vehicle following in the lane-changing lane, and the maximum acceleration AD1 of the vehicle following in the lane-changing lane, the safe distance LD between the rear of the vehicle and the front of the vehicle following in the lane-changing lane is calculated. ;

[0064] S4.2 Calculate the safe distance perpendicular to the lane lines: Based on the real-time speed VZ of the vehicle and the speeds VX of other vehicles in the left and right lanes, where VX is the speed of the vehicle in front of the vehicle in its lane VA, the speed of the vehicle behind the vehicle in its lane VB, and the speed of the vehicle in front of the vehicle in the lane changing lane VC or the speed of the vehicle behind the vehicle in the lane changing lane VD, calculate the safe distance LS between the vehicle and other vehicles in the left and right lanes perpendicular to the lane lines. Where K1 and K2 are the first time parameter and the second time parameter, respectively.

[0065] Furthermore, the lane-changing decision module is also used for:

[0066] S5.1 Detect whether there are other vehicles within the forward detection range LF and the rear detection range LR of the vehicle itself;

[0067] S5.2 Obtain the distance SA between the vehicle and the vehicle in front of it in its lane, the distance SC between the vehicle and the vehicle in front of it in its lane-changing lane, and the distance SD between the vehicle and the vehicle behind it in its lane-changing lane.

[0068] S5.3 If SA > LF, maintain the original speed in the lane, which is the planned real-time speed VZ; if there is a vehicle A in front of the vehicle in the lane within the forward detection distance LF, and LA < SA ≤ LF, maintain the constant speed in the lane, which is the vehicle's real-time speed VZ when vehicle A in front of the vehicle in the lane was first detected; if SA ≤ LA, compare the vehicle's real-time speed VZ with the speed VA of the vehicle in front of the vehicle in the lane. Then maintain lane following, with a speed equal to the speed VA of the vehicle in front in the same lane, where K3 and K4 are the first and second proportional coefficients, respectively, and K3 < 1, K4 > 1; if If so, maintain the original speed in the lane, which is the planned real-time speed VZ; if Then prepare to change lanes;

[0069] S5.4 Lane Change Direction Decision: Determine the width of the narrowest passable area on both the left and right sides of the vehicle, and choose to change lanes to the wider lane; if the front and rear safety distance to the wider lane is insufficient, choose to change lanes to the narrower lane; if the front and rear safety distance to the narrower lane is insufficient, maintain lane following.

[0070] S5.5 If there are no other vehicles within the forward detection distance LF and backward detection distance LR of the lane changing lane (i.e., SC > LF and SD > LR), then the vehicle changes lanes at its original speed, which is the planned real-time speed VZ. If LC < SC ≤ LF and SD > LR, then the vehicle changes lanes at an accelerated speed, which is K5. VZ, where K5 is the third proportionality coefficient and K5 > 1; if LD < SD ≤ LR and SC > LF, then change lanes and accelerate to a speed of K6. VZ, where K6 is the fourth proportionality coefficient and K6 > 1; if LC < SC ≤ LF and LD < SD ≤ LR, then change lanes and accelerate to a speed of K7. VZ, where K7 is the fifth proportional coefficient and K7>1; if SC≤LC, or SD≤LD, then maintain lane following and drive at the speed of the vehicle in front, where K5<K6<K7.

[0071] Furthermore, S5.2 includes:

[0072] Define the X-axis as the direction parallel to the lane lines, the Y-axis as the direction perpendicular to the lane lines, and the origin as the geometric center of the vehicle's starting position. Calculate the distance between the vehicle and the vehicle A in front of it in the same lane. ,in, Let X be the geometric center of the vehicle ahead in the lane. The X-axis coordinate of the vehicle's geometric center. The length of the vehicle in front of you in your lane. The length of the vehicle; the distance between the vehicle and the vehicle in front (C) in the lane where the vehicle changes lanes. ,in, The X-axis coordinate of the geometric center of the vehicle ahead in the lane-changing lane is given. The distance between the vehicle in front of it in the lane changing lane and the vehicle behind it in the lane changing lane is D. ,in, The X-axis coordinate of the geometric center of the vehicle after changing lanes. This refers to the length of the vehicle following the lane change.

[0073] Furthermore, the path planning module is also used for:

[0074] S6.1 In a lane-changing scenario, if the speed of vehicle A in the preceding lane is zero, the coordinates of two feature points are obtained within the passable area as (X2, Y2) and (X3, Y3), where X2 is the X-axis coordinate of vehicle A in the preceding lane at the geometric center of the static impassable area, and Y2 is the Y-axis coordinate of vehicle A in the preceding lane at the geometric center of the static impassable area relative to the lane-changing direction. The sum of, For the width of the vehicle, X3 is the width of the vehicle in front of the vehicle in the lane; X3 is the X-axis coordinate of the front position of vehicle A in the static non-passable area; Y3 is the Y-axis coordinate of the center position of the lane changing lane.

[0075] S6.2 In lane-changing scenarios, if the speed of vehicle A in the preceding lane is not zero, the coordinates of three feature points are obtained within the passable area as (X1', Y1'), (X2', Y2'), and (X3', Y3'). Here, X1' is the X-axis coordinate of the rear of vehicle A in the first segment of the three-segment dynamic non-passable area, and Y1' is the Y-axis coordinate of vehicle A at the geometric center of the three-segment dynamic non-passable area relative to the lane-changing direction. The sum of X and Y; X2' is the X-axis coordinate of the geometric center of vehicle A in the second segment of the three-segment dynamic no-passing zone, and Y2' is the Y-axis coordinate of the geometric center of vehicle A in the three-segment dynamic no-passing zone relative to the lane-changing direction. X3' is the X-axis coordinate of the front of vehicle A in the three-segment dynamic non-passable area, and Y3' is the Y-axis coordinate of the center position of the lane-changing lane.

[0076] S6.3. Use cubic spline interpolation to make the feature points continuous, obtain the local path trajectory, and complete the path planning.

[0077] Furthermore, the system executes the complete process from the data acquisition module to the path planning module every 10ms to plan the local path trajectory under the current environmental conditions.

[0078] The advantages of this invention are:

[0079] (1) This invention sets up static impassable areas, three-segment dynamic impassable areas, and finally determines the passable areas. Path planning is performed within the passable areas, and the planned path meets the actual driving needs, thus avoiding the problem of existing technologies directly using the midpoint of the desired lane line as the target location. The solution is highly feasible. At the same time, this invention plans the speed and obtains the real-time speed of the vehicle, avoiding the problem of driving at a fixed speed which does not conform to the driving scenario, and ensuring the timeliness of the planned path. In addition, various safety distances are calculated, improving safety. The overall solution simultaneously meets the requirements of feasibility, timeliness, and safety.

[0080] (2) In traditional methods, planning is carried out according to lane lines. Due to the uncertainty of the actual position of other vehicles, some driving space may be sacrificed. In this invention, the passable area is treated as a whole. Under the premise of complying with road traffic regulations, the most reasonable local path planning is made according to the specific driving environment. This ensures safety and is more in line with human driving habits. In the three-segment dynamic non-passable area, the probability of the speed change of other vehicles follows a normal distribution. The probability of driving at maximum deceleration and maximum acceleration is small. That is, the probability of the first segment non-passable area and the third segment non-passable area is small. Therefore, in the selection of feature points, the position of the first feature point does not need to be too far away, thus reducing the driving cost of the vehicle under the premise of ensuring safety.

[0081] (3) The present invention performs real-time dynamic speed planning, which not only ensures the controllability of the entire planning process time, but also adjusts the speed in real time according to the specific situation to better suit the current state, so as to get closer to the expected tracking time while ensuring safety. Real-time dynamic safety distance planning can obtain the most suitable safety distance under the current environmental conditions; the complete technical solution can adjust various control parameters in real time, solve for the optimal solution under the current environmental conditions, and plan the best running trajectory while ensuring safety.

[0082] (4) In terms of feasibility, the existing technology (CN112362074A) only obtains a passable area based on the current state. In terms of timeliness, speed planning only includes two local speed states: maintaining the current speed and stopping. In terms of safety, the safety distance in the local path curve cluster planning is a set safety margin constant D. In terms of feasibility, this application obtains road information, information about the vehicle and other vehicles to obtain static impassable areas and three-segment dynamic impassable areas, and then obtains the passable area based on the road boundary, predicting the environmental state over a period of time, which is more in line with the actual scenario. In terms of timeliness, it includes global speed planning and local speed planning under different decision scenarios. During the entire path tracking process, it can ensure that even if different obstacles appear, the destination can be reached within the specified time. Local speed planning includes maintaining the original lane speed, constant speed, following the vehicle, switching to the original speed, and accelerating, which not only better reflects the real state of driving, but also improves safety. In terms of safety, the safety distance in local path planning uses the real-time speed of the vehicle and the speed of other vehicles in the left and right lanes as variables, which conforms to physical constraints such as fluid dynamics. In addition, the method of obtaining the optimal path in this application is different from the existing technology. The optimal path is obtained directly through feature points that conform to the current state, without the need to generate multiple candidate curve clusters and then calculate the optimal path through a complex cost function, thus reducing the amount of computation. Attached Figure Description

[0083] Figure 1 This is a flowchart of a local path planning method for autonomous vehicles based on feasible paths, as disclosed in an embodiment of the present invention.

[0084] Figure 2 This is a Frenet coordinate system diagram in the local path planning method for autonomous vehicles based on feasible paths disclosed in the embodiments of the present invention;

[0085] Figure 3 This is a schematic diagram of the passable area in the local path planning method for unmanned vehicles based on feasible paths disclosed in the embodiments of the present invention.

[0086] Figure 4 This is a schematic diagram illustrating the distance definition in the local path planning method for autonomous vehicles based on feasible paths disclosed in the embodiments of the present invention;

[0087] Figure 5 This is a schematic diagram of the lane-changing decision-making process in the local path planning method for autonomous vehicles based on feasible paths disclosed in the embodiments of the present invention.

[0088] Figure 6 These are schematic diagrams illustrating different lane-changing scenarios in the local path planning method for autonomous vehicles based on feasible paths disclosed in the embodiments of the present invention.

[0089] Figure 7 This is a schematic diagram of feature points and smooth paths in different scenarios in the local path planning method for autonomous vehicles based on feasible paths disclosed in the embodiments of the present invention. Detailed Implementation

[0090] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, 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.

[0091] Example 1

[0092] like Figure 1 As shown, Embodiment 1 of the present invention provides a local path planning method for autonomous vehicles based on feasible paths, including the following steps:

[0093] S1. Obtain the vectorized path, vehicle, and environment information for the global planning, mainly acquiring road information, information about the vehicle itself, and information about other vehicles; the specific process is as follows:

[0094] S1.1, Road Information

[0095] The vectorized path coordinates of the global planning are transformed into the Frenet coordinate system, where the distance in the direction parallel to the lane line is S and the distance in the direction perpendicular to the lane line is d. Figure 2 This is a schematic diagram of the Frenet coordinate system.

[0096] Vectorized path turning radius R;

[0097] Types of lane dividers and center dividers, road boundaries, central medians, etc.

[0098] The above information acquisition methods are existing technologies, and the information acquisition methods are not improvements of this invention and are not within the scope of protection of this application, so they will not be described in detail. For example, lane line recognition can adopt the lane line recognition method described in the "Lane Line Recognition Model Training Method, Device and Lane Line Recognition Method, Device" published in Chinese Patent Publication No. CN113298050A.

[0099] S1.2, Information about your own vehicle and other vehicles

[0100] Define the lane that each vehicle is currently traveling on as its own lane, and the lane that the vehicle is about to change lanes as its lane changing lane;

[0101] Each vehicle is defined as vehicle Z, vehicle A in front of vehicle A in vehicle A in vehicle A in vehicle B in vehicle B in vehicle B in lane B in lane C in lane B; vehicle D in front of vehicle C in lane B in lane B in lane B.

[0102] Forward detection range LF of the vehicle sensor; Rearward detection range LR of the vehicle sensor.

[0103] The system acquires the geometric center position, shape, speed, speed direction, and maximum acceleration / deceleration of both the user vehicle and other vehicles. The user vehicle's information is obtained from data stored in its memory. Information about other vehicles is acquired using existing technologies, such as the shape and centroid position acquisition method described in Chinese Patent Publication No. CN115268460A, "A Local Path Planning and Guidance Method for Differential Mobile Robots in a Mixed Environment," which calculates speed and speed direction based on the centroid position at different times. Chinese Patent Publication No. CN117689842A describes the shape and centroid position of other vehicles, then calculates speed and speed direction based on the centroid position at different times. CN117889867A describes a method for acquiring the centroid position and speed direction of other vehicles. CN114973123A describes a method for identifying other vehicle models, and the maximum acceleration / deceleration of other vehicles can be obtained from publicly available online information. The solution disclosed in CN 114503174 A describes a method for identifying other vehicle models, and the maximum acceleration and deceleration of other vehicles can be obtained through publicly available information online. Therefore, information about other vehicles can be obtained entirely through the solutions described in the prior art. This invention does not improve upon this part of the technology and does not impose any particular limitations.

[0104] Transform the geometric center positions to the Frenet coordinate system: the geometric center position of the vehicle in front of it (XZ, YZ), the geometric center position of the vehicle in front of it in its lane (XA, YA), the geometric center position of the vehicle behind it in its lane (XB, YB), the geometric center position of the vehicle in front of it in the lane it changes to (XC, YC), and the geometric center position of the vehicle behind it in the lane it changes to (XD, YD).

[0105] The length of the vehicle in its own lane is XLZ, the length of the vehicle in front of it in its own lane is XLA, the length of the vehicle behind it in its own lane is XLB, the length of the vehicle in front of it in the lane it changes to is XLC, and the length of the vehicle behind it in the lane it changes to is XLD.

[0106] The width of the vehicle in front of it is WZ, the width of the vehicle in front of it in its lane is WA, the width of the vehicle behind it in its lane is WB, the width of the vehicle in front of it in the lane it changes to is WC, and the width of the vehicle behind it in the lane it changes to is WD.

[0107] The speed of the vehicle in its own lane is VZ, the speed of the vehicle in front of it in its own lane is VA, the speed of the vehicle behind it in its own lane is VB, the speed of the vehicle in front of it in the lane it changes to is VC, and the speed of the vehicle behind it in the lane it changes to is VD.

[0108] Maximum acceleration of the vehicle AZ1, maximum deceleration of the vehicle AZ2, maximum acceleration of the vehicle in front of the vehicle in the lane AA1, maximum deceleration of the vehicle in front of the vehicle in the lane AA2, maximum acceleration of the vehicle behind the vehicle in the lane AB1, maximum deceleration of the vehicle behind the vehicle in the lane AB2, maximum acceleration of the vehicle in front of the vehicle in the lane changing lane AC1, maximum deceleration of the vehicle in front of the vehicle in the lane changing lane AC2, maximum acceleration of the vehicle behind the vehicle in the lane changing lane AD1, maximum deceleration of the vehicle behind the vehicle in the lane changing lane AD2.

[0109] S2. Based on static obstacles, determine static impassable areas; based on dynamic obstacles, determine three-segment dynamic impassable areas. Within the boundaries of the passable area, remove the static impassable areas and the three-segment dynamic impassable areas to obtain the passable area for the vehicle; Figure 3 As shown, the specific process is as follows:

[0110] S2.1 Obtain the boundary of the passable area through the road boundary, central median or central divider;

[0111] S2.2. Based on the speed of other vehicles, identify static and dynamic obstacles; obstacles with a speed of 0 are static obstacles, and obstacles with a speed other than 0 are dynamic obstacles.

[0112] S2.3. Based on the shape of the static obstacle, which is equivalent to a rectangle parallel to the lane line, determine the static impassable area;

[0113] S2.4. Based on the speed and acceleration / deceleration capabilities of the dynamic obstacle, determine a three-segment dynamic impassable zone. The first segment is the position where the dynamic obstacle stops at its maximum deceleration. The second segment is the position reached by the dynamic obstacle traveling at a constant speed within the same time period. The third segment is the position reached by the dynamic obstacle at its maximum acceleration within the same time period. The area formed between the rear of the vehicle in the first segment and the front of the vehicle in the third segment is the three-segment dynamic impassable zone. The same time period refers to the time it takes for the dynamic obstacle to stop at its maximum deceleration.

[0114] S2.5 Within the boundary of the passable area, remove the static non-passable area and the three-segment dynamic non-passable area. If the distance between the obstacle and the boundary of the passable area is less than the minimum safe distance, then remove the area between the boundary of the passable area and the obstacle, and finally obtain the passable area of ​​the vehicle.

[0115] S3. Perform speed planning for the vehicle and obtain its real-time speed; the specific process is as follows:

[0116] S3.1 Curvature denormalization: the maximum curvature is equivalent to 0, and the minimum curvature is equivalent to 1; set the default minimum vehicle speed V0, the default maximum vehicle speed V0+V1, where V1 is the difference between the default maximum vehicle speed and the default minimum vehicle speed. Given the maximum turning radius Lmax, the minimum turning radius Lmin, and the real-time position curvature of the globally planned vectorized path K, the minimum curvature Kmin=1 / Lmax, and the maximum curvature Kmax=1 / Lmin; denormalize the curvature Knorm=(Kmax-K) / (Kmax-Kmin).

[0117] S3.2, The initial default velocity V = V0 + V1 at each position of the vectorized path in global planning. Knorm;

[0118] S3.3 Given the total path length L, calculate the default travel time T required to travel at the default speed based on the initial default speed V. The default travel time T is calculated using existing technology, such as by integration. The total path length L is equal to the default travel time T, which is calculated by integrating the initial default speed V. Substituting the total path length L and the initial default speed V into this formula yields the default travel time T.

[0119] S3.4 Solving for the real-time velocity VZ=β V, where β is the real-time speed multiplier and β = {T} (L-Lk)} / {L (T-Tk)}, where Tk is the actual travel time and Lk is the actual travel distance. Set upper and lower limits for VZ and constrain it, VZmin≤VZ≤VZmax.

[0120] S3.5. Based on the lane change decision, decelerate, maintain a constant speed, or accelerate during lane change and following.

[0121] S4. Calculate the various safe distances during the vehicle's journey; the specific process is as follows:

[0122] S4.1 Calculate the safe distance in the direction parallel to the lane lines: Based on the real-time speed VZ and the maximum deceleration AZ2 of the vehicle, calculate the safe distance LA between the front of the vehicle and the rear of the vehicle in front of it in the lane. Where T1 is the system decision time; based on the real-time speed VZ of the vehicle, the speed VC of the vehicle in front in the changing lane, and the maximum deceleration AC2 of the vehicle in front in the changing lane, the safe distance LC between the front of the vehicle and the rear of the vehicle in front in the changing lane is calculated. Where T2 is the lane-changing time; based on the real-time speed VZ of the vehicle, the speed VD of the vehicle following in the lane-changing lane, and the maximum acceleration AD1 of the vehicle following in the lane-changing lane, the safe distance LD between the rear of the vehicle and the front of the vehicle following in the lane-changing lane is calculated. ;

[0123] S4.2 Calculate the safe distance perpendicular to the lane lines: Based on the real-time speed VZ of the vehicle and the speeds VX of other vehicles in the left and right lanes, where VX is the speed of the vehicle in front of the vehicle in its lane VA, the speed of the vehicle behind the vehicle in its lane VB, and the speed of the vehicle in front of the vehicle in the lane changing lane VC or the speed of the vehicle behind the vehicle in the lane changing lane VD, calculate the safe distance LS between the vehicle and other vehicles in the left and right lanes perpendicular to the lane lines. Where K1 and K2 are the first time parameter and the second time parameter, respectively.

[0124] S5. Based on the information obtained from S1-S4, make a lane-changing decision, deciding whether to change lanes or remain in the current lane and follow the vehicle in front; for example... Figures 4 to 6 As shown, the specific process is as follows:

[0125] S5.1 Detect whether there are other vehicles within the forward detection distance LF and the rear detection distance LR of the vehicle's lane and lane changing lane; the forward detection distance LF and the rear detection distance LR are the distances that the vehicle's forward sensor and rear sensor can detect, respectively.

[0126] S5.2 Calculate the distances between the vehicle and other detected vehicles, including the distance SA between the vehicle and vehicle A in front of it in its lane, the distance SC between the vehicle and vehicle C in front of it in its lane-changing lane, and the distance SD between the vehicle and vehicle D behind it in its lane-changing lane. Specifically, the calculation method is as follows: using the direction parallel to the lane lines as the X-axis, the direction perpendicular to the lane lines as the Y-axis, and the geometric center of the vehicle's starting position as the origin, calculate the distance between the vehicle and vehicle A in front of it in its lane. ,in, Let X be the geometric center of the vehicle ahead in the lane. The X-axis coordinate of the vehicle's geometric center. The length of the vehicle in front of you in your lane. The length of the vehicle; the distance between the vehicle and the vehicle in front (C) in the lane where the vehicle changes lanes. ,in, The X-axis coordinate of the geometric center of the vehicle ahead in the lane-changing lane is given. The distance between the vehicle in front of it in the lane changing lane and the vehicle behind it in the lane changing lane is D. ,in, The X-axis coordinate of the geometric center of the vehicle after changing lanes. This refers to the length of the vehicle following the lane change.

[0127] S5.3 If there is no vehicle A in front of the vehicle in the lane within the forward detection distance LF (i.e., SA > LF), then maintain the original speed in the lane at the planned real-time speed VZ; if there is vehicle A in front of the vehicle in the lane within the forward detection distance LF, and LA < SA ≤ LF, then maintain the constant speed in the lane at the vehicle's real-time speed VZ when vehicle A was first detected; if SA ≤ LA, then compare the vehicle's real-time speed VZ with the speed VA of the vehicle in front of the vehicle in the lane. Then maintain lane following, with a speed equal to the speed VA of the vehicle in front in the same lane, where K3 and K4 are the first and second proportional coefficients, respectively, and K3 < 1, K4 > 1; if If so, maintain the original speed in the lane, which is the planned real-time speed VZ; if Then prepare to change lanes;

[0128] S5.4 Lane Change Direction Decision: Determine the width of the narrowest passable area on both the left and right sides of the vehicle, and choose to change lanes to the wider lane; if the front and rear safety distance to the wider lane is insufficient, choose to change lanes to the narrower lane; if the front and rear safety distance to the narrower lane is insufficient, maintain lane following.

[0129] S5.5 If there are no other vehicles within the forward detection distance LF and backward detection distance LR of the lane changing lane (i.e., SC > LF and SD > LR), then the vehicle changes lanes at its original speed, which is the planned real-time speed VZ. If LC < SC ≤ LF and SD > LR, then the vehicle changes lanes at an accelerated speed, which is K5. VZ, where K5 is the third proportionality coefficient and K5 > 1; if LD < SD ≤ LR and SC > LF, then change lanes and accelerate to a speed of K6. VZ, where K6 is the fourth proportionality coefficient and K6 > 1; if LC < SC ≤ LF and LD < SD ≤ LR, then change lanes and accelerate to a speed of K7. VZ, where K7 is the fifth proportional coefficient and K7>1; if SC≤LC, or SD≤LD, then maintain lane following and drive at the speed of the vehicle in front, where K5<K6<K7.

[0130] S6. Obtain the planned waypoints and smooth the path to complete the path planning. The vehicle then travels according to the currently planned path. Figure 7 As shown, the specific process is as follows:

[0131] S6.1 In a lane-changing scenario, if the speed of vehicle A in the preceding lane is zero, the coordinates of two feature points are obtained within the passable area as (X2, Y2) and (X3, Y3), where X2 is the X-axis coordinate of vehicle A in the preceding lane at the geometric center of the static impassable area, and Y2 is the Y-axis coordinate of vehicle A in the preceding lane at the geometric center of the static impassable area relative to the lane-changing direction. The sum of, For the width of the vehicle, X3 is the width of the vehicle in front of the vehicle in the lane; X3 is the X-axis coordinate of the front position of vehicle A in the static non-passable area; Y3 is the Y-axis coordinate of the center position of the lane changing lane.

[0132] S6.2 In lane-changing scenarios, if the speed of vehicle A in the preceding lane is not zero, the coordinates of three feature points are obtained within the passable area as (X1', Y1'), (X2', Y2'), and (X3', Y3'). Here, X1' is the X-axis coordinate of the rear of vehicle A in the first segment of the three-segment dynamic non-passable area, and Y1' is the Y-axis coordinate of vehicle A at the geometric center of the three-segment dynamic non-passable area relative to the lane-changing direction. The sum of X and Y; X2' is the X-axis coordinate of the geometric center of vehicle A in the second segment of the three-segment dynamic no-passing zone, and Y2' is the Y-axis coordinate of the geometric center of vehicle A in the three-segment dynamic no-passing zone relative to the lane-changing direction. X3' is the X-axis coordinate of the front of vehicle A in the three-segment dynamic non-passable area, and Y3' is the Y-axis coordinate of the center position of the lane-changing lane.

[0133] S6.3 Smooth Path: Cubic spline interpolation is used to make the feature points continuous, obtaining the local path trajectory and completing the path planning. The specific process of cubic spline interpolation is as follows: When the speed of the vehicle ahead A in the lane is not zero, three feature points are selected, plus the coordinates of the vehicle Z at this moment, and the coordinates of the four points are known; it is also known that the direction of the vehicle Z at this moment and the direction of the feature point (X3', Y3') are both parallel to the lane line. A piecewise cubic polynomial is constructed to force the function values, first derivatives and second derivatives of all internal points and boundary points to be continuous, and the directions of the start and end points are used as the clamping boundary conditions; the derivatives of the intermediate points are solved through the global equation system to obtain a highly smooth curve.

[0134] Given that the speed of vehicle A in front of the vehicular vehicle is zero, two feature points were selected, along with the coordinates of the vehicular vehicle Z at this moment, giving the coordinates of these three points. It is also known that the direction of vehicular vehicle Z at this moment and the direction of the feature point (X3, Y3) are both parallel to the lane line. A piecewise cubic polynomial is constructed, forcing the function values, first derivatives, and second derivatives of all internal and boundary points to be continuous. The directions of the start and end points are used as clamping boundary conditions. The derivatives at intermediate points are solved through a global system of equations, resulting in a highly smooth curve. This balances abrupt changes in centrifugal acceleration, improving driving comfort. Cubic spline interpolation is existing technology and will not be discussed further here.

[0135] It should be noted that, according to the system's operating cycle, this invention executes the method once every 10ms according to the complete process of S1-S6 to plan the local path trajectory under the current environmental state.

[0136] Through the above technical solutions, this invention sets static impassable areas, three-segment dynamic impassable areas, and finally determines the passable areas. Path planning is performed within the passable areas, and the planned paths meet actual driving needs, thus avoiding the problem of existing technologies directly using the midpoint of the desired lane line as the target location. This solution demonstrates strong feasibility. Simultaneously, this invention plans speed by acquiring the vehicle's real-time speed, avoiding the problem of driving at a fixed speed which is inconsistent with the driving scenario, ensuring the timeliness of the planned path. Furthermore, it calculates various safety distances, improving safety. The overall solution simultaneously meets the requirements of feasibility, timeliness, and safety.

[0137] Example 2

[0138] Based on Embodiment 1, Embodiment 2 of the present invention also provides a local path planning system for autonomous vehicles based on feasible paths, including:

[0139] The data acquisition module is used to acquire road information, information about the vehicle itself, and information about other vehicles.

[0140] The passable area determination module is used to determine static impassable areas based on static obstacles and three-segment dynamic impassable areas based on dynamic obstacles. The first segment is the position where the dynamic obstacle stops at its maximum deceleration, the second segment is the position reached by the dynamic obstacle traveling at a constant speed within the same time period, and the third segment is the position reached by the dynamic obstacle at its maximum acceleration within the same time period. The area formed between the rear of the vehicle in the first segment and the front of the vehicle in the third segment is the three-segment dynamic impassable area. The same time period refers to the time when the dynamic obstacle stops at its maximum deceleration. Within the boundaries of the passable area, the static impassable area and the three-segment dynamic impassable area are removed to obtain the vehicle's passable area.

[0141] The speed planning module is used to plan the vehicle's speed and obtain the vehicle's real-time speed.

[0142] The safe distance calculation module is used to calculate various safe distances during vehicle travel.

[0143] The lane change decision module is used to make lane change decisions based on the information obtained from the data acquisition module and the safe distance calculation module, deciding whether to change lanes or stay in the current lane and follow the vehicle.

[0144] The path planning module is used to obtain planned waypoints and smooth the path to complete the path planning. The vehicle then travels according to the currently planned path.

[0145] Specifically, the passage area determination module is also used for:

[0146] S2.1 Obtain the boundary of the passable area through the road boundary, central median or central divider;

[0147] S2.2. Based on the speed of other vehicles, identify static and dynamic obstacles;

[0148] S2.3. Based on the shape of the static obstacle, which is equivalent to a rectangle parallel to the lane line, determine the static impassable area;

[0149] S2.4. Based on the speed and acceleration / deceleration capabilities of dynamic obstacles, determine the three-segment dynamic impassable area;

[0150] S2.5 Within the boundary of the passable area, remove the static non-passable area and the three-segment dynamic non-passable area. If the distance between the obstacle and the boundary of the passable area is less than the minimum safe distance, then remove the area between the boundary of the passable area and the obstacle, and finally obtain the passable area of ​​the vehicle.

[0151] Specifically, the speed planning module is also used for:

[0152] S3.1 Curvature denormalization: the maximum curvature is equivalent to 0, and the minimum curvature is equivalent to 1; set the default minimum vehicle speed V0, the default maximum vehicle speed V0+V1, where V1 is the difference between the default maximum vehicle speed and the default minimum vehicle speed. Given the maximum turning radius Lmax, the minimum turning radius Lmin, and the real-time position curvature of the globally planned vectorized path K, the minimum curvature Kmin=1 / Lmax, and the maximum curvature Kmax=1 / Lmin; denormalize the curvature Knorm=(Kmax-K) / (Kmax-Kmin).

[0153] S3.2, The initial default velocity V = V0 + V1 at each position of the vectorized path in global planning. Knorm;

[0154] S3.3 Given the total length of the path as L, calculate the default travel time T required to travel at the default speed, based on the initial default speed V.

[0155] S3.4 Solving for the real-time velocity VZ=β V, where β is the real-time speed multiplier and β = {T} (L-Lk)} / {L (T-Tk)}, where Tk is the actual travel time and Lk is the actual travel distance. Set the upper and lower limits of VZ to constrain it.

[0156] Specifically, the safe distance calculation module is also used for:

[0157] S4.1. Based on the real-time speed VZ and the maximum deceleration AZ2 of the vehicle, calculate the safe distance LA between the front of the vehicle and the rear of the vehicle in front of it in the same lane. Where T1 is the system decision time; based on the real-time speed VZ of the vehicle, the speed VC of the vehicle in front in the changing lane, and the maximum deceleration AC2 of the vehicle in front in the changing lane, the safe distance LC between the front of the vehicle and the rear of the vehicle in front in the changing lane is calculated. Where T2 is the lane-changing time; based on the real-time speed VZ of the vehicle, the speed VD of the vehicle following in the lane-changing lane, and the maximum acceleration AD1 of the vehicle following in the lane-changing lane, the safe distance LD between the rear of the vehicle and the front of the vehicle following in the lane-changing lane is calculated. ;

[0158] S4.2 Calculate the safe distance perpendicular to the lane lines: Based on the real-time speed VZ of the vehicle and the speeds VX of other vehicles in the left and right lanes, where VX is the speed of the vehicle in front of the vehicle in its lane VA, the speed of the vehicle behind the vehicle in its lane VB, and the speed of the vehicle in front of the vehicle in the lane changing lane VC or the speed of the vehicle behind the vehicle in the lane changing lane VD, calculate the safe distance LS between the vehicle and other vehicles in the left and right lanes perpendicular to the lane lines. Where K1 and K2 are the first time parameter and the second time parameter, respectively.

[0159] More specifically, the lane-changing decision module is also used for:

[0160] S5.1 Detect whether there are other vehicles within the forward detection range LF and the rear detection range LR of the vehicle itself;

[0161] S5.2 Obtain the distance SA between the vehicle and the vehicle in front of it in its lane, the distance SC between the vehicle and the vehicle in front of it in its lane-changing lane, and the distance SD between the vehicle and the vehicle behind it in its lane-changing lane.

[0162] S5.3 If SA > LF, maintain the original speed in the lane, which is the planned real-time speed VZ; if there is a vehicle A in front of the vehicle in the lane within the forward detection distance LF, and LA < SA ≤ LF, maintain the constant speed in the lane, which is the vehicle's real-time speed VZ when vehicle A in front of the vehicle in the lane was first detected; if SA ≤ LA, compare the vehicle's real-time speed VZ with the speed VA of the vehicle in front of the vehicle in the lane. Then maintain lane following, with a speed equal to the speed VA of the vehicle in front in the same lane, where K3 and K4 are the first and second proportional coefficients, respectively, and K3 < 1, K4 > 1; if If so, maintain the original speed in the lane, which is the planned real-time speed VZ; if Then prepare to change lanes;

[0163] S5.4 Lane Change Direction Decision: Determine the width of the narrowest passable area on both the left and right sides of the vehicle, and choose to change lanes to the wider lane; if the front and rear safety distance to the wider lane is insufficient, choose to change lanes to the narrower lane; if the front and rear safety distance to the narrower lane is insufficient, maintain lane following.

[0164] S5.5 If there are no other vehicles within the forward detection distance LF and backward detection distance LR of the lane changing lane (i.e., SC > LF and SD > LR), then the vehicle changes lanes at its original speed, which is the planned real-time speed VZ. If LC < SC ≤ LF and SD > LR, then the vehicle changes lanes at an accelerated speed, which is K5. VZ, where K5 is the third proportionality coefficient and K5 > 1; if LD < SD ≤ LR and SC > LF, then change lanes and accelerate to a speed of K6. VZ, where K6 is the fourth proportionality coefficient and K6 > 1; if LC < SC ≤ LF and LD < SD ≤ LR, then change lanes and accelerate to a speed of K7. VZ, where K7 is the fifth proportional coefficient and K7>1; if SC≤LC, or SD≤LD, then maintain lane following and drive at the speed of the vehicle in front, where K5<K6<K7.

[0165] More specifically, S5.2 includes:

[0166] Define the X-axis as the direction parallel to the lane lines, the Y-axis as the direction perpendicular to the lane lines, and the origin as the geometric center of the vehicle's starting position. Calculate the distance between the vehicle and the vehicle A in front of it in the same lane. ,in, Let X be the geometric center of the vehicle ahead in the lane. The X-axis coordinate of the vehicle's geometric center. The length of the vehicle in front of you in your lane. The length of the vehicle; the distance between the vehicle and the vehicle in front (C) in the lane where the vehicle changes lanes. ,in, The X-axis coordinate of the geometric center of the vehicle ahead in the lane-changing lane is given. The distance between the vehicle in front of it in the lane changing lane and the vehicle behind it in the lane changing lane is D. ,in, The X-axis coordinate of the geometric center of the vehicle after changing lanes. This refers to the length of the vehicle following the lane change.

[0167] More specifically, the path planning module is also used for:

[0168] S6.1 In a lane-changing scenario, if the speed of vehicle A in the preceding lane is zero, the coordinates of two feature points are obtained within the passable area as (X2, Y2) and (X3, Y3), where X2 is the X-axis coordinate of vehicle A in the preceding lane at the geometric center of the static impassable area, and Y2 is the Y-axis coordinate of vehicle A in the preceding lane at the geometric center of the static impassable area relative to the lane-changing direction. The sum of, For the width of the vehicle, X3 is the width of the vehicle in front of the vehicle in the lane; X3 is the X-axis coordinate of the front position of vehicle A in the static non-passable area; Y3 is the Y-axis coordinate of the center position of the lane changing lane.

[0169] S6.2 In lane-changing scenarios, if the speed of vehicle A in the preceding lane is not zero, the coordinates of three feature points are obtained within the passable area as (X1', Y1'), (X2', Y2'), and (X3', Y3'). Here, X1' is the X-axis coordinate of the rear of vehicle A in the first segment of the three-segment dynamic non-passable area, and Y1' is the Y-axis coordinate of vehicle A at the geometric center of the three-segment dynamic non-passable area relative to the lane-changing direction. The sum of X and Y; X2' is the X-axis coordinate of the geometric center of vehicle A in the second segment of the three-segment dynamic no-passing zone, and Y2' is the Y-axis coordinate of the geometric center of vehicle A in the three-segment dynamic no-passing zone relative to the lane-changing direction. X3' is the X-axis coordinate of the front of vehicle A in the three-segment dynamic non-passable area, and Y3' is the Y-axis coordinate of the center position of the lane-changing lane.

[0170] S6.3. Use cubic spline interpolation to make the feature points continuous, obtain the local path trajectory, and complete the path planning.

[0171] Specifically, the system executes the complete process from the data acquisition module to the path planning module every 10ms to plan the local path trajectory under the current environmental conditions.

[0172] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. 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. Such 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 local path planning method for autonomous vehicles based on feasible paths, characterized in that, include: S1. Obtain road information, information about your own vehicle and other vehicles; S2. Determine the static impassable area based on static obstacles, and determine the three-segment dynamic impassable area based on dynamic obstacles. The first segment is the position where the dynamic obstacle stops at its maximum deceleration, the second segment is the position reached by the dynamic obstacle traveling at a constant speed within the same time period, and the third segment is the position reached by the dynamic obstacle at its maximum acceleration within the same time period. The area formed between the rear of the vehicle in the first segment and the front of the vehicle in the third segment is the three-segment dynamic impassable area. The same time period refers to the time when the dynamic obstacle stops at its maximum deceleration. Within the boundaries of the passable area, remove the static impassable area and the three-segment dynamic impassable area to obtain the vehicle's passable area. S3. Perform speed planning for the vehicle and obtain its real-time speed; S3 includes: S3.1 Curvature denormalization: the maximum curvature is equivalent to 0, and the minimum curvature is equivalent to 1; set the default minimum vehicle speed V0, the default maximum vehicle speed V0+V1, where V1 is the difference between the default maximum vehicle speed and the default minimum vehicle speed. Given the maximum turning radius Lmax, the minimum turning radius Lmin, and the real-time position curvature of the globally planned vectorized path K, the minimum curvature Kmin=1 / Lmax, and the maximum curvature Kmax=1 / Lmin; denormalize the curvature Knorm=(Kmax-K) / (Kmax-Kmin). S3.2, The initial default velocity V = V0 + V1 at each position of the vectorized path in global planning. Knorm; S3.3 Given the total length of the path as L, calculate the default travel time T required to travel at the default speed, based on the initial default speed V. S3.4 Solving for the real-time velocity VZ=β V, where β is the real-time speed multiplier and β = {T} (L-Lk)} / {L (T-Tk)}, where Tk is the actual driving time and Lk is the actual driving distance. Set the upper and lower limits of VZ to constrain it; S4. Calculate the various safe distances during the vehicle's travel; S5. Based on the information obtained from S1-S4, make a lane-changing decision and decide whether to change lanes or stay in the current lane and follow the vehicle in front. S6. Obtain the planned waypoints and smooth the path to complete the path planning. The vehicle then travels according to the currently planned path.

2. The method for local path planning of unmanned vehicles based on feasible paths according to claim 1, characterized in that, S2 include: S2.1 Obtain the boundary of the passable area through the road boundary, central median or central divider; S2.

2. Based on the speed of other vehicles, identify static and dynamic obstacles; S2.

3. Based on the shape of the static obstacle, which is equivalent to a rectangle parallel to the lane line, determine the static impassable area; S2.

4. Based on the speed and acceleration / deceleration capabilities of dynamic obstacles, determine the three-segment dynamic impassable area; S2.5 Within the boundary of the passable area, remove the static non-passable area and the three-segment dynamic non-passable area. If the distance between the obstacle and the boundary of the passable area is less than the minimum safe distance, then remove the area between the boundary of the passable area and the obstacle, and finally obtain the passable area of ​​the vehicle.

3. The method for local path planning of unmanned vehicles based on feasible paths according to claim 1, characterized in that, S4 include: S4.

1. Based on the real-time speed VZ and the maximum deceleration AZ2 of the vehicle, calculate the safe distance LA between the front of the vehicle and the rear of the vehicle in front of it in the same lane. Where T1 is the system decision time; based on the real-time speed VZ of the vehicle, the speed VC of the vehicle in front in the changing lane, and the maximum deceleration AC2 of the vehicle in front in the changing lane, the safe distance LC between the front of the vehicle and the rear of the vehicle in front in the changing lane is calculated. Where T2 is the lane-changing time; based on the real-time speed VZ of the vehicle, the speed VD of the vehicle following in the lane-changing lane, and the maximum acceleration AD1 of the vehicle following in the lane-changing lane, the safe distance LD between the rear of the vehicle and the front of the vehicle following in the lane-changing lane is calculated. ; S4.2 Calculate the safe distance perpendicular to the lane lines: Based on the real-time speed VZ of the vehicle and the speeds VX of other vehicles in the left and right lanes, where VX is the speed of the vehicle in front of the vehicle in its lane VA, the speed of the vehicle behind the vehicle in its lane VB, and the speed of the vehicle in front of the vehicle in the lane changing lane VC or the speed of the vehicle behind the vehicle in the lane changing lane VD, calculate the safe distance LS between the vehicle and other vehicles in the left and right lanes perpendicular to the lane lines. Where K1 and K2 are the first time parameter and the second time parameter, respectively.

4. The local path planning method for unmanned vehicles based on feasible paths according to claim 3, characterized in that, S5 include: S5.1 Detect whether there are other vehicles within the forward detection range LF and the rear detection range LR of the vehicle itself; S5.2 Obtain the distance SA between the vehicle and the vehicle in front of it in its lane, the distance SC between the vehicle and the vehicle in front of it in its lane-changing lane, and the distance SD between the vehicle and the vehicle behind it in its lane-changing lane. S5.3 If SA > LF, maintain the original speed in the lane, which is the planned real-time speed VZ; if there is a vehicle A in front of the vehicle in the lane within the forward detection distance LF, and LA < SA ≤ LF, maintain the constant speed in the lane, which is the vehicle's real-time speed VZ when vehicle A in front of the vehicle in the lane was first detected; if SA ≤ LA, compare the vehicle's real-time speed VZ with the speed VA of the vehicle in front of the vehicle in the lane. Then maintain lane following, with a speed equal to the speed VA of the vehicle in front in the same lane, where K3 and K4 are the first and second proportional coefficients, respectively, and K3 < 1, K4 > 1; if If so, maintain the original speed in the lane, which is the planned real-time speed VZ; if Then prepare to change lanes; S5.4 Lane Change Direction Decision: Determine the width of the narrowest passable area on both the left and right sides of the vehicle, and choose to change lanes to the wider lane; if the front and rear safety distance to the wider lane is insufficient, choose to change lanes to the narrower lane; if the front and rear safety distance to the narrower lane is insufficient, maintain lane following. S5.5 If there are no other vehicles within the forward detection distance LF and backward detection distance LR of the lane changing lane (i.e., SC > LF and SD > LR), then the vehicle changes lanes at its original speed, which is the planned real-time speed VZ. If LC < SC ≤ LF and SD > LR, then the vehicle changes lanes at an accelerated speed, which is K5. VZ, where K5 is the third proportionality coefficient and K5 > 1; if LD < SD ≤ LR and SC > LF, then change lanes and accelerate to a speed of K6. VZ, where K6 is the fourth proportionality coefficient and K6 > 1; if LC < SC ≤ LF and LD < SD ≤ LR, then change lanes and accelerate to a speed of K7. VZ, where K7 is the fifth proportional coefficient and K7>1; if SC≤LC, or SD≤LD, then maintain lane following and drive at the speed of the vehicle in front, where K5<K6<K7.

5. The local path planning method for unmanned vehicles based on feasible paths according to claim 4, characterized in that, S5.2 includes: Define the X-axis as the direction parallel to the lane lines, the Y-axis as the direction perpendicular to the lane lines, and the origin as the geometric center of the vehicle's starting position. Calculate the distance between the vehicle and the vehicle A in front of it in the same lane. ,in, Let X be the geometric center of the vehicle in front in the lane. The X-axis coordinate of the vehicle's geometric center. The length of the vehicle in front of you in your lane. The length of the vehicle; the distance between the vehicle and the vehicle in front (C) in the lane where the vehicle changes lanes. ,in, The X-axis coordinate of the geometric center of the vehicle ahead in the lane-changing lane is given. The distance between the vehicle in front of it in the lane changing lane; the distance between the vehicle and the vehicle behind it in the lane changing lane (D). ,in, The X-axis coordinate of the geometric center of the vehicle after changing lanes. This refers to the length of the vehicle following the lane change.

6. The method for local path planning of unmanned vehicles based on feasible paths according to claim 4, characterized in that, S6 include: S6.1 In a lane-changing scenario, if the speed of vehicle A in the preceding lane is zero, the coordinates of two feature points are obtained within the passable area as (X2, Y2) and (X3, Y3), where X2 is the X-axis coordinate of vehicle A in the preceding lane at the geometric center of the static impassable area, and Y2 is the Y-axis coordinate of vehicle A in the preceding lane at the geometric center of the static impassable area relative to the lane-changing direction. The sum of, For the width of the vehicle, X3 is the width of the vehicle in front of the vehicle in the lane; X3 is the X-axis coordinate of the front position of vehicle A in the static non-passable area; Y3 is the Y-axis coordinate of the center position of the lane changing lane. S6.2 In lane-changing scenarios, if the speed of vehicle A in the preceding lane is not zero, the coordinates of three feature points are obtained within the passable area as (X1', Y1'), (X2', Y2'), and (X3', Y3'). Here, X1' is the X-axis coordinate of the rear of vehicle A in the first segment of the three-segment dynamic non-passable area, and Y1' is the Y-axis coordinate of vehicle A at the geometric center of the three-segment dynamic non-passable area relative to the lane-changing direction. The sum of X and Y; X2' is the X-axis coordinate of the geometric center of vehicle A in the second segment of the three-segment dynamic no-passing zone, and Y2' is the Y-axis coordinate of the geometric center of vehicle A in the three-segment dynamic no-passing zone relative to the lane-changing direction. X3' is the X-axis coordinate of the front of vehicle A in the three-segment dynamic non-passable area, and Y3' is the Y-axis coordinate of the center position of the lane-changing lane. S6.

3. Use cubic spline interpolation to make the feature points continuous, obtain the local path trajectory, and complete the path planning.

7. The method for local path planning of unmanned vehicles based on feasible paths according to claim 1, characterized in that, This method is executed once every 10ms according to the complete process of S1-S6 to plan the local path trajectory under the current environment state.

8. A local path planning system for autonomous vehicles based on feasible paths, characterized in that, include: The data acquisition module is used to acquire road information, information about the vehicle itself, and information about other vehicles. The passable area determination module is used to determine static impassable areas based on static obstacles and three-segment dynamic impassable areas based on dynamic obstacles. The first segment is the position where the dynamic obstacle stops at its maximum deceleration, the second segment is the position reached by the dynamic obstacle traveling at a constant speed within the same time period, and the third segment is the position reached by the dynamic obstacle at its maximum acceleration within the same time period. The area formed between the rear of the vehicle in the first segment and the front of the vehicle in the third segment is the three-segment dynamic impassable area. The same time period refers to the time when the dynamic obstacle stops at its maximum deceleration. Within the boundaries of the passable area, the static impassable area and the three-segment dynamic impassable area are removed to obtain the vehicle's passable area. The speed planning module is used to plan the vehicle's speed and obtain the vehicle's real-time speed; the speed planning module is also used for: S3.1 Curvature denormalization: the maximum curvature is equivalent to 0, and the minimum curvature is equivalent to 1; set the default minimum vehicle speed V0, the default maximum vehicle speed V0+V1, where V1 is the difference between the default maximum vehicle speed and the default minimum vehicle speed. Given the maximum turning radius Lmax, the minimum turning radius Lmin, and the real-time position curvature of the globally planned vectorized path K, the minimum curvature Kmin=1 / Lmax, and the maximum curvature Kmax=1 / Lmin; denormalize the curvature Knorm=(Kmax-K) / (Kmax-Kmin). S3.2, The initial default velocity V = V0 + V1 at each position of the vectorized path in global planning. Knorm; S3.3 Given the total length of the path as L, calculate the default travel time T required to travel at the default speed, based on the initial default speed V. S3.4 Solving for the real-time velocity VZ=β V, where β is the real-time speed multiplier and β = {T} (L-Lk)} / {L (T-Tk)}, where Tk is the actual driving time and Lk is the actual driving distance. Set the upper and lower limits of VZ to constrain it; The safe distance calculation module is used to calculate various safe distances during vehicle travel. The lane change decision module is used to make lane change decisions based on the information obtained from the data acquisition module and the safe distance calculation module, deciding whether to change lanes or stay in the current lane and follow the vehicle. The path planning module is used to obtain planned waypoints and smooth the path to complete the path planning. The vehicle then travels according to the currently planned path.

9. The local path planning system for unmanned vehicles based on feasible paths according to claim 8, characterized in that, The access area determination module is also used for: S2.1 Obtain the boundary of the passable area through the road boundary, central median or central divider; S2.

2. Based on the speed of other vehicles, identify static and dynamic obstacles; S2.

3. Based on the shape of the static obstacle, which is equivalent to a rectangle parallel to the lane line, determine the static impassable area; S2.

4. Based on the speed and acceleration / deceleration capabilities of dynamic obstacles, determine the three-segment dynamic impassable area; S2.5 Within the boundary of the passable area, remove the static non-passable area and the three-segment dynamic non-passable area. If the distance between the obstacle and the boundary of the passable area is less than the minimum safe distance, then remove the area between the boundary of the passable area and the obstacle, and finally obtain the passable area of ​​the vehicle.

Images