A local path planning method, system and storage medium
By dynamic path planning and historical path matching under the frenet coordinate system, the problem of path planning consistency of driverless vehicles in low-precision positioning environment is solved, and time consistency is achieved without relying on high-precision positioning. Path tracking reduces costs and facilitates technology promotion.
Patent Information
- Application Number
- CN202110119180.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-01-28
- Publication Date
- 2024-12-10
- Estimated Expiration
- 2041-01-28
AI Technical Summary
The existing local path planning methods of unmanned vehicles cannot meet the time consistency requirements in low-precision positioning environments, especially in scenarios where GPS signals are poor, and the high-precision positioning cost is high, which limits the promotion of technology.
Dynamic path planning is carried out under the frenet coordinate system, and paths that meet the time consistency requirements are generated by matching and evaluating historical paths, and paths under the Cartesian coordinate system are output without relying on high-precision positioning to ensure path tracking accuracy.
It achieves time consistency in meeting path tracking without high-precision positioning, reduces the cost of unmanned vehicles, and facilitates the promotion and implementation of technology.
Smart Images
Figure CN114815794B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of unmanned driving technology, and in particular to a local path planning method, a local path planning system and a computer-readable storage medium. Background Art
[0002] The core modules of unmanned vehicles generally include perception, positioning, decision-making and planning, and control modules. The planning module in the decision-making and planning module can generate a path that can be executed by the control module based on the surrounding environment information. In order to achieve a better path tracking effect, the control module puts forward a time consistency requirement for the path output by the planning module. The time consistency requirement means that when there is no major change in the surrounding environment, the path generated in the current planning cycle needs to be consistent with the path generated in the previous planning cycle.
[0003] If local path planning cannot meet good temporal consistency, the actual route of the driverless vehicle will deviate more and more from the path planned at the first moment as time goes by.
[0004] In the prior art, the path generated by the first planning moment is saved, and in the next planning cycle, the position point P of the unmanned vehicle is determined by high-precision positioning. car , point P car Project it onto the planning path of the previous planning cycle to get the position point P match , intercepted from P match The path from the beginning to the end of the previous cycle planning path is output as the path of the current planning cycle.
[0005] The above scheme has high requirements for high-precision positioning. In places such as tunnels where GPS signals are poor, this method fails when high-precision positioning cannot be obtained. In addition, the cost of equipping unmanned vehicles with high-precision positioning is too high, which is not conducive to the promotion of unmanned driving technology. Summary of the invention
[0006] The purpose of the present invention is to address the technical defects in the prior art and to provide a local path planning method, a local path planning system and a computer-readable storage medium that meet time consistency without high-precision positioning.
[0007] The technical solution adopted to achieve the purpose of the present invention is:
[0008] A local path planning method comprises the steps of:
[0009] S1. Determine whether there is a historical path in the previous planning cycle based on the frenet coordinate system. If not, determine the starting point of the path planning as the vehicle position and then proceed to step S2. If yes, proceed to step S4.
[0010] S2. Determine whether the lateral distance between the unmanned vehicle and the reference line is greater than a set threshold. If so, proceed to step S3 for local path planning. Otherwise, output the reference line as the path of the current planning cycle.
[0011] S3. Perform dynamic path planning in the frenet coordinate system to obtain the path in the frenet coordinate system, transform the path coordinates in the frenet coordinate system, obtain the path in the Cartesian coordinate system and output it;
[0012] S4. Match the path to be output that is consistent with the current vehicle position on the historical path of the previous planning cycle, and evaluate whether the path to be output is available according to the preset rules. If so, output the path to be output; if not, determine the path planning starting point on the historical path and proceed to step S3 for re-planning.
[0013] Furthermore, after obtaining the path in the frenet coordinate system, the method further includes saving the obtained planned path and the corresponding reference line in the frenet coordinate system for use in the next planning cycle.
[0014] Preferably, the steps of matching and obtaining the path to be output are as follows:
[0015] The mileage of the unmanned vehicle from the previous planning cycle to the current planning cycle is calculated by the vehicle speedometer, and then the first mileage component of the unmanned vehicle traveling along the reference line is calculated from the mileage. According to the first mileage component, the current position of the unmanned vehicle is matched to the first position P of the historical path of the previous planning cycle on the historical path based on the Frenet coordinate system. match_1 ;
[0016] The angle between the current vehicle heading angle and the reference line is obtained through the inertial navigation module, and the second mileage component of the unmanned vehicle traveling along the reference line is calculated. According to the second mileage component, the second position P of the current unmanned vehicle position matching the historical path of the previous planning cycle is obtained on the historical path based on the Frenet coordinate system. match_2 ;
[0017] Using the first position P match_1 and the second position P match_2 , the final matching point P is obtained by the following formula match_3 ,
[0018] P match_3 =P match_1 +λ(P match_2 -P match_1 ),0<λ<1
[0019] Delete the final matching point P match_3The previous path, the remaining path is the path to be output after matching the historical path.
[0020] Preferably, in step S4, the evaluation of whether the path to be output is available is to evaluate whether the path to be output collides with obstacles in the current environment and / or whether the end point of the path to be output is located on the current reference line. If there is a collision with obstacles in the current environment and / or the end point of the path to be output is not located on the current reference line, re-planning is performed.
[0021] Preferably, evaluating whether the path to be output collides with an obstacle in the current environment is achieved by the following steps:
[0022] Calculate the minimum distance L between the obstacle and the output path that deviates from the parameter line in the frenet coordinate system min , when any obstacle's L min <L threshold , it indicates that there is a risk of collision with obstacles in the current environment.
[0023] Preferably, evaluating whether the end point of the path to be output is located on the current reference line is achieved by the following steps:
[0024] Project the end point of the path to be output onto the current reference line and obtain the shortest distance D from the end point to the reference line. min , when D min <D threshold , it is considered that the end point of the path to be output is on the current reference line.
[0025] Furthermore, before the step of determining whether there is a historical path of the previous planning cycle in step S1, it also includes a step of determining whether it is the first planning. If it is the first planning, the planning starting point is determined to be the vehicle position, and then step S2 is entered. If not, step S1 is entered to determine whether there is a historical path of the previous planning cycle.
[0026] Preferably, the step of determining the starting point of the path planning on the historical path is as follows;
[0027] Take a section of the path from the history Path, determine whether the path collides with surrounding obstacles. If a collision occurs, reset the planning starting point to point S on the historical path of the previous planning cycle. 0 ; Otherwise, select point S from the historical path of the previous planning cycle 1 As the starting point for path planning;
[0028] Among them, S 0 Point S is the point where the center of the rear axle of the driverless vehicle is matched and projected onto the historical path. 1 Point is the previous point on the history path.
[0029] The present invention also aims to provide a local path planning system, comprising:
[0030] A historical path judgment unit is used to judge whether there is a historical path of the previous planning cycle based on the Frenet coordinate system in the system;
[0031] A planning starting point determination unit is used to determine the planning starting point according to the judgment result of the historical path judgment unit, wherein when there is no historical path of the previous planning cycle, the path planning starting point is determined to be the vehicle position; when there is a historical path of the previous planning cycle, the path planning starting point is determined on the historical path;
[0032] A lateral distance determination unit, used to determine whether the lateral distance between the unmanned vehicle and the reference line is greater than a set threshold;
[0033] A dynamic path planning unit is used to perform dynamic path planning in the Frenet coordinate system when the lateral distance between the unmanned vehicle and the reference line is greater than a set threshold or when the historical path matching evaluation unit evaluates that the output path is unavailable, obtain the path in the Frenet coordinate system, and transform the path coordinates in the Frenet coordinate system to obtain the path in the Cartesian coordinate system;
[0034] A historical path matching and evaluation unit is used to match the historical path of the previous planning cycle to obtain a to-be-output path that is consistent with the current vehicle position, and to evaluate whether the to-be-output path is available according to a preset rule;
[0035] The planned path output unit is used to output the path formed in the current planning cycle to the control module of the unmanned vehicle.
[0036] The present invention also aims to provide a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and the computer program is configured to implement the steps of the local path planning method when called by a processor.
[0037] The present invention uses dynamic programming to perform path planning in the frenet coordinate system to obtain a path (FrenetPath) in the frenet coordinate system, and then converts the FrenetPath into a path (CartesianPath) in the Cartesian coordinate system that can be used by a control module through coordinate transformation, and outputs it to the control module for tracking control, thereby meeting the control module's requirements for time consistency of the path, and does not require high-precision positioning, thereby reducing costs. BRIEF DESCRIPTION OF THE DRAWINGS
[0038] Figure 1 is a flow chart of the local path planning method of the present invention;
[0039] Figure 2A schematic diagram for determining the starting point of the planned path from the historical path;
[0040] Figure 3 Schematic diagram of the difference between the Frenet coordinate system and the Cartesian coordinate system;
[0041] Figure 4 This is a schematic diagram of dynamic path planning based on the Frenet coordinate system. DETAILED DESCRIPTION
[0042] The present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.
[0043] The local path planning method of the present invention is based on the reference line of the current unmanned vehicle. The overall planning process is as follows: Figure 1 shown.
[0044] Step 1: Determine whether it is the first planning. If yes, go to step 2; if not, go to step 3;
[0045] Step 2: The planning starting point is the position of the vehicle, with coordinates (x, y) = (0.0, 0.0); calculate the lateral distance δl between the unmanned vehicle and the reference line. If δl>threshold, local path planning is required, and go to step 4; if not, output the reference line as the path of the current planning cycle;
[0046] Step 3: Determine whether the path of the previous cycle is empty. If it is empty, go to step 2; if it is not empty, go to step 5;
[0047] Step 4: Execute dynamic path planning to obtain FrenetPath based on the frenet coordinate system, and save FrenetPath and the corresponding reference line as input for the next planning cycle; perform coordinate transformation on FrenetPath to obtain CartesianPath in the Cartesian coordinate system, and output it to the control module;
[0048] Step 5: Historical path matching, convert the historical path into the path of the current cycle, and evaluate based on the path to estimate whether the path meets the non-collision requirements of the current environment. If not, redetermine the planning starting point, re-plan, and go to step 4; if satisfied, output the path.
[0049] The lateral distance δl between the unmanned vehicle and the reference line refers to the deviation distance between the center of the rear axle of the unmanned vehicle and the reference line, such as Figure 2 shown.
[0050] In the above technical solution, the reference line of the current unmanned vehicle is fitted by the perception module in combination with the current environment (lane line information, road boundary information), and the reference line parameter equation is set as:
[0051] y=f(x)=ax 3 +bx 2 +cx+d
[0052] The origin of the coordinate system is located at the center point of the rear axle of the unmanned vehicle, the front of the unmanned vehicle is the positive x direction, the left side is the positive y direction, and a, b, c, and d are fitting parameters.
[0053] In the present invention, there is a difference in determining the planning starting point according to the planning attributes and whether there is a historical path in the previous cycle: if the current planning cycle is the first planning or not the first planning and there is no historical path in the previous planning cycle, the planning starting point is set to the vehicle position P xy =(0.0, 0.0), P sl =(s 0 , l 0 ,dl,ddl),s 0 , l 0 P xy The sl coordinate obtained by projecting onto the reference line is dl=ddl=0.0.
[0054] If it is not the first planning cycle and the path of the previous planning cycle is empty, local path planning is required, then the path from the center of the rear axle of the unmanned vehicle (such as Figure 2 Point C shown in the figure matches point S on the historical path 0 , to the planning starting point S 1 The length is:
[0055] L start =a 1 v+a 2
[0056] Among them, v is the actual speed of the current vehicle, a 1 is the response delay of the unmanned driving system, a 2 is a constant term, which can be set to the distance from the center of the rear axle to the front of the unmanned vehicle.
[0057] like Figure 2 As shown, S 0 Point is the point where the center of the rear axle of the driverless vehicle matches the historical path. Taken from the historical track The path of this section is judged whether the path collides with the surrounding obstacles. If a collision occurs, the planning starting point is reset to point S on the path of the previous planning cycle. 0; Otherwise, select point S from the previous frame cycle trajectory 1 As a starting point for planning.
[0058] In the present invention, the above-mentioned dynamic path planning is based on the frenet coordinate system for path planning, and the FrenetPath based on the frenet coordinate system is obtained through planning, and the coordinates of the FrenetPath are transformed to obtain the path CartesianPath in the Cartesian coordinate system.
[0059] Among them, the difference between the Frenet coordinate system and the Cartesian coordinate system Figure 3 As shown, the xy coordinate system in the figure is a Cartesian coordinate system, the sl coordinate system is a Frenet coordinate system, the s direction is along the reference line, l is the normal direction of the reference line, the left side of the reference line is positive, and the right side is negative.
[0060] When planning a path based on the Frenet coordinate system, the obstacle is projected onto the reference line to obtain the boundary information of the obstacle relative to the reference line: boundary = (s min ,s max , l min , l max ),like Figure 4 The path generated by path planning should avoid collision with obstacles as much as possible, so the result of path planning is required to satisfy l in each s direction. i =f(s i ) <l min or i =f(s i )>l max (Condition 1), in order to meet the above constraints such as non-collision and being as close to the reference path as possible, the idea of dynamic planning is introduced for path planning, which is called dynamic path planning. The specific implementation steps are as follows:
[0061] Step S1: Generate Figure 4 The intersection of each column and each row in the SL grid is called a node, such as Figure 4 As shown in the circle, there is only one node in the first column, which is called the first node. Figure 4 Node shown start , which is the starting point of planning.
[0062] Among them, the S direction range can be set to [0, 100], the sampling interval between every two S directions can be set to Δs=4*v+10, v is the actual speed of the unmanned vehicle, the L direction range is set according to the passable width, generally can be set to [-1.75, 1.75], and the sampling interval between every two L directions is Δl=0.583.
[0063] It should be noted that the above SL grid division method can also be flexibly divided according to the SL boundaries occupied by obstacles, and is not limited to the above embodiments.
[0064] Step S2: Calculate the cost value of each node. The cost of each node includes the cost of the node itself and the cost of the previous node connected to the node, that is,
[0065] Cost ni =Cost ni_self +Cost ni_pre
[0066] Cost ni_self =Cost path +Cost obstacle ,
[0067] Cost path It represents the smoothness of the path and the degree of its closeness to the reference line.
[0068]
[0069] Where v is the actual speed of the driverless vehicle, Cost obstacle It represents the distance from the obstacle. When the node cannot meet the above condition 1, Cost obstacle is infinite; when condition 1 is met,
[0070]
[0071] Among them, l obstacle is the minimum distance between the path and the obstacle (such as Figure 4 shown), w 1 , w 2 , w 3 , w 4 Represents the cost calculation weight.
[0072] Among them, each node can be connected to multiple nodes in the previous column (called the forward nodes of the node), and the cost value of each connection is calculated according to the above cost calculation method. After the calculation is completed, the connection with the smallest cost is selected, and the corresponding cost value and forward node are stored in the node.
[0073] Step S3: Select the node with the smallest cost from the last column of the SL grid, and then find the node in the previous column based on the forward node stored in the node, and find the corresponding nodes in turn until the first node is found.
[0074] Step S4: The set of all nodes found in step S3 is the path planning feature points of the current planning cycle, such as Figure 4 The black solid dots shown {(s 0 , l 0 ),...,(s i , l i ),...,(s n , l n )}), the feature points are interpolated by cubic spline, the first-order derivative of the left endpoint is constrained to be the first-order derivative dl of the first node, and the first-order derivative of the right endpoint is constrained to be 0, and the path based on the Frenet coordinates is obtained;
[0075]
[0076] In the formula, a 01 -a 04 , a i1 -a i4 , a n1 -a n4 , represents the interpolation coefficient.
[0077] S5: Convert l=f(s) from the Frenet coordinate system to the Cartesian coordinate system to obtain the path y=f(x) in the Cartesian coordinate system.
[0078] Then, the path FrenetPath in the frenet coordinate system and the corresponding reference line are saved as the input of the next planning cycle; the path FrenetPath in the frenet coordinate system is transformed into the coordinates to obtain the path CartesianPath in the Cartesian coordinate system, which is output to the control module and used as an executable path by the control module.
[0079] The path of the previous planning cycle is the path FrenetPath in the Frenet coordinate system, which is called the historical path (such as Figure 2 The starting point shown is S start , the end point is S end ), is a function of l with respect to s, that is
[0080] l=f(s), (1)
[0081] At the same time, the arc length L corresponding to each s is determined starting from the starting point of the path arc The mathematical relationship is
[0082]
[0083] In the formula, ds represents the differential of s.
[0084] In addition, as s changes, the angle between the vehicle heading angle θ and the reference line also changes, that is,
[0085] θ(s)=arctan(f(s)′), (3)
[0086] Therefore, in the present invention, historical path matching is achieved by projecting the current vehicle position on the historical path, and the steps are as follows:
[0087] Step 1: The mileage of the unmanned vehicle from the last planning cycle to the current planning cycle can be calculated by the vehicle wheel speed meter. The mileage can be considered as the length of the arc length of the unmanned vehicle along the path of the last cycle. The mileage component s of the unmanned vehicle along the reference line direction can be obtained by formula 2. Substituting it into formula 1, the first position P of the current unmanned vehicle position matching the path of the last cycle can be obtained. match_1 .
[0088] Step 2: The angle between the current vehicle heading angle and the reference line can be obtained through inertial navigation. The mileage component s of the unmanned vehicle traveling along the reference line can be calculated by formula 3. Substituting it into formula 1, the second position P of the current unmanned vehicle position matching the previous cycle path can be obtained. match_2 .
[0089] Step 3: Comprehensive consideration of P match_1 and P match_2 , get the final matching point P match_3 (like Figure 2 Point S shown 0 shown),
[0090] P match_3 =P match_1 +λ(P match_2 -P match_1 ), 0<λ<1
[0091] The value of λ can be adjusted according to the actual tracking effect of the control module. match_3 The previous path, the remaining path is the path obtained after matching the historical path, such as Figure 2 The S shown 0 Point to the end point S end point.
[0092] In the present invention, after completing the historical path matching, the historical path evaluation is also performed to determine whether the path formed by the matching is available and whether it meets the security requirements. If it does not meet the requirements, it is re-planned. If it meets the requirements, it is output for use. The evaluation method is as follows:
[0093] First, calculate the minimum distance L between all obstacles and the current path that deviates from the parameter line in the Frenet coordinate system. min , when any obstacle's L min <L thresholdWhen the L of all obstacles is min ≥L threshold When , determine whether the end of the path is on the current reference line: project the end point of the path onto the current reference line and obtain the shortest distance D from the point to the reference line. min , when D min <D threshold , it is considered that the path at the previous moment can be spliced and extended with the current reference path, and there is no need to re-plan the local path. Otherwise, local path re-planning is required.
[0094] The present invention also provides a local path planning system, including a historical path judgment unit, which is used to judge whether there is a historical path of the previous planning cycle based on the frenet coordinate system in the system; a planning starting point determination unit, which is used to determine the planning starting point according to the judgment result of the judgment unit, wherein when there is no historical path of the previous planning cycle, the path planning starting point is determined to be the vehicle position; when there is a historical path of the previous planning cycle, the path planning starting point is determined on the historical path; a lateral distance judgment unit, which is used to judge whether the lateral distance between the unmanned vehicle and the reference line is greater than a set threshold; a dynamic path planning unit, which is used to perform dynamic path planning in the frenet coordinate system when the lateral distance between the unmanned vehicle and the reference line is greater than the set threshold or when the historical path matching evaluation unit evaluates that the path to be output is unavailable, and to obtain a path in the frenet coordinate system, and to transform the path coordinates in the frenet coordinate system to obtain a path in the Cartesian coordinate system; a historical path matching evaluation unit, which is used to match the path to be output that is consistent with the current vehicle position on the historical path of the previous planning cycle, and to evaluate whether the path to be output is available according to preset rules; a planning path output unit, which is used to output the formed path of the current planning cycle to the control module of the unmanned vehicle.
[0095] For the specific technical solution and detailed description of the above-mentioned local path planning system, please refer to the description of the aforementioned local path planning method, which will not be described in detail again.
[0096] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and the computer program is configured to implement the steps of the local path planning method when called by a processor.
[0097] Without relying on high-precision positioning, the path generated by the present invention can meet the time consistency requirement; the dynamic path planning method can cover most urban road driving scenarios, and has both high completeness and flexibility; and because it does not rely on high-precision positioning, it can reduce the cost of unmanned vehicles and facilitate the promotion and implementation of unmanned driving technology.
[0098] The above is only a preferred embodiment of the present invention. It should be pointed out that, for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.
Claims
1. A local path planning method, characterized in that: Includes steps: S1. Determine whether there is a historical path in the previous planning cycle based on the frenet coordinate system. If not, determine the starting point of the path planning as the vehicle position and then proceed to step S2. If yes, proceed to step S4. S2. Determine whether the lateral distance between the unmanned vehicle and the reference line is greater than a set threshold. If so, proceed to step S3 for local path planning. Otherwise, output the reference line as the path of the current planning cycle. S3. Perform dynamic path planning in the frenet coordinate system to obtain the path in the frenet coordinate system, transform the path coordinates in the frenet coordinate system, obtain the path in the Cartesian coordinate system and output it; S4. Match the output path consistent with the current vehicle position on the historical path of the previous planning cycle, evaluate whether the output path is available according to the preset rules, and if so, output the output path. If not, determine the path planning starting point on the historical path and proceed to step S3 for re-planning; The steps of matching and obtaining the path to be output are as follows: The mileage of the unmanned vehicle from the previous planning cycle to the current planning cycle is calculated by the vehicle speedometer, and then the first mileage component of the unmanned vehicle traveling along the reference line is calculated from the mileage. According to the first mileage component, the current position of the unmanned vehicle is matched to the first position P of the historical path of the previous planning cycle on the historical path based on the Frenet coordinate system. match_1 ; The angle between the current vehicle heading angle and the reference line is obtained through the inertial navigation module, and the second mileage component of the unmanned vehicle traveling along the reference line is calculated. According to the second mileage component, the second position P of the current unmanned vehicle position matching the historical path of the previous planning cycle is obtained on the historical path based on the Frenet coordinate system. match_2 ; Using the first position P match_1 and the second position P match_2 , the final matching point P is obtained by the following formula match_3 , P match_3 =P match_1 +λ(P match_2 -P match_1 ),0<λ<1 Delete the final matching point P match_3 The previous path, the remaining path is the path to be output after matching the historical path.
2. The local path planning method according to claim 1, characterized in that: In step S3, after obtaining the path in the frenet coordinate system, the step of saving the obtained planned path and the corresponding reference line in the frenet coordinate system for use in the next planning cycle is also included.
3. The local path planning method according to claim 1, characterized in that: In step S4, whether the path to be output is available is evaluated, that is, whether the path to be output collides with obstacles in the current environment and / or whether the end point of the path to be output is located on the current reference line. If there is a collision with obstacles in the current environment and / or the end point of the path to be output is not located on the current reference line, re-planning is performed.
4. The local path planning method according to claim 3, characterized in that: The evaluation of whether the output path collides with obstacles in the current environment is achieved through the following steps: Calculate the minimum distance L between the obstacle and the output path that deviates from the parameter line in the frenet coordinate system min , when any obstacle's L min <L threshold , it indicates that there is a risk of collision with obstacles in the current environment.
5. The local path planning method according to claim 3, characterized in that: The evaluation of whether the end point of the path to be output is located on the current reference line is achieved by the following steps: Project the end point of the path to be output onto the current reference line and obtain the shortest distance D from the end point to the reference line. min , when D min <D threshold , it is considered that the end point of the path to be output is on the current reference line.
6. The local path planning method according to claim 1, characterized in that: Before the step of determining whether there is a historical path of the previous planning cycle in step S1, it also includes a step of determining whether it is the first planning. If it is the first planning, the planning starting point is determined to be the vehicle position, and then step S2 is entered. If not, step S1 is entered to determine whether there is a historical path of the previous planning cycle.
7. The local path planning method according to claim 1, characterized in that: The steps of determining the starting point of path planning on the historical path are as follows; Take a section of the path from the history Path, determine whether the path collides with surrounding obstacles. If a collision occurs, reset the planning starting point to point S0 on the historical path of the previous planning cycle; Otherwise, select point S1 from the historical path of the previous planning cycle as the starting point of path planning; Among them, point S0 is the point where the center of the rear axle of the unmanned vehicle is matched and projected onto the historical path, and point S1 is the point in front on the historical path.
8. A local path planning system, characterized in that: The local path planning method according to any one of claims 1 to 7 is implemented, comprising: A historical path judgment unit is used to judge whether there is a historical path of the previous planning cycle based on the Frenet coordinate system in the system; A planning starting point determination unit is used to determine the planning starting point according to the judgment result of the historical path judgment unit, wherein when there is no historical path of the previous planning cycle, the path planning starting point is determined to be the vehicle position; when there is a historical path of the previous planning cycle, the path planning starting point is determined on the historical path; A lateral distance determination unit, used to determine whether the lateral distance between the unmanned vehicle and the reference line is greater than a set threshold; A dynamic path planning unit is used to perform dynamic path planning in the Frenet coordinate system when the lateral distance between the unmanned vehicle and the reference line is greater than a set threshold or when the historical path matching evaluation unit evaluates that the output path is unavailable, obtain the path in the Frenet coordinate system, and transform the path coordinates in the Frenet coordinate system to obtain the path in the Cartesian coordinate system; A historical path matching and evaluation unit is used to match the historical path of the previous planning cycle to obtain a to-be-output path that is consistent with the current vehicle position, and evaluate whether the to-be-output path is available according to preset rules; The planned path output unit is used to output the path formed in the current planning cycle to the control module of the unmanned vehicle.
9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and the computer program is configured to implement the steps of the local path planning method according to any one of claims 1 to 7 when called by a processor.