WO2019042295A1 - 一种无人驾驶路径规划方法、***和装置 - Google Patents

一种无人驾驶路径规划方法、***和装置 Download PDF

Info

Publication number
WO2019042295A1
WO2019042295A1 PCT/CN2018/102811 CN2018102811W WO2019042295A1 WO 2019042295 A1 WO2019042295 A1 WO 2019042295A1 CN 2018102811 W CN2018102811 W CN 2018102811W WO 2019042295 A1 WO2019042295 A1 WO 2019042295A1
Authority
WO
WIPO (PCT)
Prior art keywords
path
sub
vehicle
collision
current
Prior art date
Application number
PCT/CN2018/102811
Other languages
English (en)
French (fr)
Inventor
涂强
赖健明
李鹏
陈盛军
肖志光
Original Assignee
广州小鹏汽车科技有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 广州小鹏汽车科技有限公司 filed Critical 广州小鹏汽车科技有限公司
Publication of WO2019042295A1 publication Critical patent/WO2019042295A1/zh
Priority to US16/732,231 priority Critical patent/US11460311B2/en

Links

Images

Classifications

    • GPHYSICS
    • G01MEASURING; TESTING
    • G01CMEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/20Instruments for performing navigational calculations
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B60VEHICLES IN GENERAL
    • B60WCONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W60/00Drive control systems specially adapted for autonomous road vehicles
    • B60W60/001Planning or execution of driving tasks
    • B60W60/0011Planning or execution of driving tasks involving control alternatives for a single driving scenario, e.g. planning several paths to avoid obstacles
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B62LAND VEHICLES FOR TRAVELLING OTHERWISE THAN ON RAILS
    • B62DMOTOR VEHICLES; TRAILERS
    • B62D15/00Steering not otherwise provided for
    • B62D15/02Steering position indicators ; Steering position determination; Steering aids
    • B62D15/025Active steering aids, e.g. helping the driver by actively influencing the steering system after environment evaluation
    • GPHYSICS
    • G01MEASURING; TESTING
    • G01CMEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/26Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network
    • G01C21/34Route searching; Route guidance
    • G01C21/3446Details of route searching algorithms, e.g. Dijkstra, A*, arc-flags, using precalculated routes
    • GPHYSICS
    • G05CONTROLLING; REGULATING
    • G05DSYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
    • G05D1/00Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
    • G05D1/02Control of position or course in two dimensions
    • G05D1/021Control of position or course in two dimensions specially adapted to land vehicles
    • G05D1/0212Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory
    • G05D1/0217Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory in accordance with energy consumption, time reduction or distance reduction criteria
    • GPHYSICS
    • G05CONTROLLING; REGULATING
    • G05DSYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
    • G05D1/00Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
    • G05D1/02Control of position or course in two dimensions
    • G05D1/021Control of position or course in two dimensions specially adapted to land vehicles
    • G05D1/0231Control of position or course in two dimensions specially adapted to land vehicles using optical position detecting means
    • G05D1/0238Control of position or course in two dimensions specially adapted to land vehicles using optical position detecting means using obstacle or wall sensors
    • GPHYSICS
    • G05CONTROLLING; REGULATING
    • G05DSYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
    • G05D1/00Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
    • G05D1/02Control of position or course in two dimensions
    • G05D1/021Control of position or course in two dimensions specially adapted to land vehicles
    • G05D1/0287Control of position or course in two dimensions specially adapted to land vehicles involving a plurality of land vehicles, e.g. fleet or convoy travelling
    • G05D1/0289Control of position or course in two dimensions specially adapted to land vehicles involving a plurality of land vehicles, e.g. fleet or convoy travelling with means for avoiding collisions between vehicles
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00Scenes; Scene-specific elements
    • G06V20/50Context or environment of the image
    • G06V20/56Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00Scenes; Scene-specific elements
    • G06V20/50Context or environment of the image
    • G06V20/56Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
    • G06V20/58Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B60VEHICLES IN GENERAL
    • B60WCONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W2520/00Input parameters relating to overall vehicle dynamics
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B60VEHICLES IN GENERAL
    • B60WCONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W2552/00Input parameters relating to infrastructure
    • B60W2552/30Road curve radius
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B60VEHICLES IN GENERAL
    • B60WCONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W2552/00Input parameters relating to infrastructure
    • B60W2552/53Road markings, e.g. lane marker or crosswalk
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B60VEHICLES IN GENERAL
    • B60WCONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W2554/00Input parameters relating to objects
    • B60W2554/40Dynamic objects, e.g. animals, windblown objects
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B60VEHICLES IN GENERAL
    • B60WCONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W2556/00Input parameters relating to data
    • B60W2556/45External transmission of data to or from the vehicle
    • B60W2556/50External transmission of data to or from the vehicle of positioning data, e.g. GPS [Global Positioning System] data

Definitions

  • the present application relates to the field of automatic driving, and more particularly to an unmanned path planning method, system and apparatus.
  • Auto-driving technology also known as auto-driving technology
  • auto-driving technology has great potential in travel safety, energy conservation and environmental protection. It is considered to be an effective way to solve traffic congestion, reduce traffic accidents and improve environmental pollution.
  • auto-driving technology has received extensive attention and has become one of the main directions for future car development.
  • an automatic driving system includes the following modules: an environment sensing module, a path planning module, a control execution module, and a human-machine interface module.
  • the above four modules are essential to the entire automatic driving system, directly affecting the intelligentization of the system.
  • Level. The path planning module is responsible for planning the lateral movement of the vehicle (ie, the vehicle or the controlled vehicle) to ensure the safety, comfort and stability of the self-driving vehicle. It is indispensable and essential for the automatic driving system. Link.
  • a good autopilot path planning module needs to comprehensively consider the vehicle's non-holonomic constraints, the optimality of the generated path and the adaptability to different traffic scenarios. Therefore, path planning is the key direction of autonomous driving technology research.
  • the current automotive driverless path planning methods mainly include:
  • a reactive path planning technique proposed by the US patent application US 2016/0313133 A1 the input of which is the obstacle information and lane information detected by the environment sensing module, and the state information of the vehicle. It performs the Dickeni triangulation on the environment model established by the environment perception module to generate a series of virtual nodes. Each virtual node is a vertex of a triangle, and the line between each two vertices is a triangle edge, and then Through the method of graph search and combined with the search conditions, a path that satisfies the requirements is obtained.
  • the disadvantage of this method is that the environment model needs to be preprocessed to obtain a large number of virtual nodes, which requires a large amount of storage space.
  • the path obtained through the search does not necessarily meet the requirements of the vehicle's non-holonomic constraints, and the obtained path is required. Make corrections.
  • the first object of the present application is to provide an unmanned path planning method that has low requirements on a storage space, can meet the requirements of a vehicle's non-holonomic constraints, and has good adaptability and scalability.
  • the second object of the present application is to provide an unmanned path planning system that has low requirements on storage space, can meet the requirements of non-complete constraints of the vehicle, has good adaptability and good scalability.
  • the third object of the present application is to provide an unmanned path planning device that has low requirements on a storage space, can meet the requirements of a vehicle's non-holonomic constraints, and has good adaptability and scalability.
  • An unmanned path planning method includes the following steps:
  • Obtaining environment sensing information and vehicle positioning and navigation information wherein the environment sensing information includes obstacle information, roadside information, and lane line information, and the vehicle positioning and navigation information includes a vehicle pose and a target route;
  • a partial path of the vehicle is obtained based on the result of the sub-path search.
  • the step of generating a sub-path according to the environment sensing information and the vehicle positioning and navigation information to obtain a candidate sub-path satisfying the vehicle constraint is specifically:
  • the step of performing a sub-path search on the non-collision candidate sub-path by using the A* search algorithm includes:
  • the evaluation factor of the sub-path evaluation function includes a lateral acceleration corresponding to the sub-path, a curvature, a curvature change, a cumulative distance, a sub-path end-point angle and a combination of any one or any combination of the difference in the target heading angle and the heuristic distance;
  • the A* search algorithm is used to find the sub-path with the optimal evaluation function value and the corresponding parent path information from the non-collision candidate sub-path;
  • the sub-path with the optimal evaluation function value and the corresponding parent path information are output as the result of the sub-path search.
  • step of calculating an evaluation function value of each sub-path in the non-collision candidate sub-path includes:
  • step of calculating an evaluation function value of each sub-path in the non-collision candidate sub-path further includes the following steps:
  • h(s) is the heuristic distance of the current subpath, k a , k ⁇ k , k k , k d , k ⁇ and k h are a(s), ⁇ k(s), k(s), d, respectively. Weight coefficients of s), ⁇ (s) and h(s).
  • the step of calculating a difference between a heading angle of each of the sub-paths in the non-collision candidate sub-path and a target heading angle includes:
  • step of calculating a target heading angle of each of the sub-paths in the non-collision candidate sub-path includes:
  • the dichotomy is used to calculate the shortest distance from the current point to the first 3 spline segments and the second 3 spline segments;
  • the heading angle of the point on the target path corresponding to the shortest distance from the current point to the target path is selected as the target heading angle ⁇ route of the sub-path.
  • the step of calculating the shortest distance between the current point to the first 3 times spline curve segment and the second 3rd spline curve segment includes:
  • Step 1 The given 3 times spline segment is divided into n segments according to the given dividing parameter t, and n+1 first class nodes are obtained, wherein the given 3 times spline segment is the first 3 a secondary spline segment or a second 3 spline segment;
  • Step 2 calculating the shortest distance from the current point to the n+1 first type nodes and the node Ptemp0 corresponding to the shortest distance;
  • Step 3 dividing the node Ptemp0 corresponding to the shortest distance into two segments in the n segment according to the parameter t, and obtaining n+1 second class nodes;
  • Step 4 calculating the shortest distance dtemp of the current point to n+1 second type nodes and the node Ptemp corresponding to the shortest distance dtemp;
  • An unmanned path planning system includes:
  • the information acquiring module is configured to acquire environment sensing information and vehicle positioning and navigation information, where the environment sensing information includes obstacle information, roadside information, and lane line information, and the vehicle positioning and navigation information includes a vehicle pose and a target route;
  • a subpath generating module configured to perform subpath generation according to the environment sensing information and the vehicle positioning and navigation information, to obtain a candidate subpath satisfying the vehicle constraint;
  • a collision detection module configured to perform collision detection on a candidate sub-path satisfying a vehicle constraint to obtain a candidate sub-path without collision
  • the sub-path search module is configured to perform sub-path search on the non-collision candidate sub-path by using an A* search algorithm
  • the local path generation module is configured to obtain a partial path of the vehicle according to the result of the sub-path search.
  • An unmanned path planning device includes:
  • a memory configured to store a program
  • a processor configured to execute the program for:
  • Obtaining environment sensing information and vehicle positioning and navigation information wherein the environment sensing information includes obstacle information, roadside information, and lane line information, and the vehicle positioning and navigation information includes a vehicle pose and a target route;
  • a partial path of the vehicle is obtained based on the result of the sub-path search.
  • the beneficial effects of the method of the present application are: obtaining an unmanned path through environment sensing information and vehicle positioning and navigation information acquisition, sub-path generation, collision detection, sub-path search, and partial path generation, and no need to pre-predict the environment model.
  • a large number of virtual nodes are processed, and the storage space requirement is small.
  • the candidate sub-paths satisfying the vehicle constraints are obtained by performing sub-path generation, and the A* search algorithm is used in the sub-path search to make the planned unmanned path. It can better meet the requirements of vehicle non-integrity constraints;
  • the acquired environmental perception information includes obstacle information, roadside information and lane line information.
  • Vehicle positioning and navigation information includes vehicle pose and target route, and comprehensive obstacles and lane changes are adopted.
  • the path, lane, vehicle pose and target route information are used for path planning, so that the generated driverless route is no longer limited by the road profile or the lane change path, and the adaptability and scalability are better.
  • the beneficial effects of the system of the present application include: an information acquisition module, a sub-path generation module, a collision detection module, a sub-path search module, and a partial path generation module, through environment sensing information, vehicle positioning and navigation information acquisition, sub-path generation, collision Detection, sub-path search and local path generation to obtain the unmanned path, no need to pre-process the environment model to obtain a large number of virtual nodes, the storage space requirements are small; through the sub-path generation to obtain the candidate for vehicle constraints
  • the path and the A* search algorithm are used in the sub-path search, so that the planned unmanned path can better meet the requirements of vehicle non-integrity constraints; the acquired environmental perception information includes obstacle information, roadside information and lanes.
  • Line information, vehicle positioning and navigation information including vehicle pose and target route, comprehensively using obstacles, lane change routes, lanes, vehicle poses and target route information for path planning, so that the generated driverless route is no longer subject to Road profile or lane change path limitations, adaptability and scalability it is good.
  • the device of the present application has the beneficial effects of including a memory and a processor in which unmanned driving is obtained by environment sensing information and vehicle positioning and navigation information acquisition, sub-path generation, collision detection, sub-path search, and partial path generation.
  • the path no longer needs to preprocess the environment model to obtain a large number of virtual nodes, and the storage space requirements are small; the sub-path generation is used to obtain the candidate sub-paths satisfying the vehicle constraints, and the A* search is used in the sub-path search.
  • the algorithm makes the planned unmanned path more satisfying the requirements of vehicle non-integrity constraints;
  • the acquired environmental perception information includes obstacle information, roadside information and lane line information, and vehicle positioning and navigation information includes vehicle pose and target
  • the route adopts obstacles, lane change routes, lanes, vehicle poses and target route information to make route planning, so that the generated driverless route is no longer restricted by road profile or lane change route, adaptability and expandability. Better sex.
  • FIG. 1 is a flow chart of a method for unmanned driving path planning according to the present application
  • FIG. 2 is a structural block diagram of an automatic driving system according to an embodiment of the present application.
  • FIG. 3 is an algorithm framework diagram of a partial path planning module according to an embodiment of the present application.
  • FIG. 4 is a schematic diagram of a path, an obstacle, and a vehicle outline according to an embodiment of the present application
  • FIG. 5 is a schematic diagram of a current pose and a sub-path of a vehicle under different curvatures according to Embodiment 1 of the present application;
  • FIG. 6 is a flowchart of a sub-path search according to an embodiment of the present application.
  • FIG. 7 is a flow chart of calculating a shortest distance from a current point P to a target path according to Embodiment 1 of the present application;
  • FIG. 9 is a result of a path planning result according to Embodiment 1 of the present application.
  • an unmanned path planning method includes the following steps:
  • Obtaining environment sensing information and vehicle positioning and navigation information wherein the environment sensing information includes obstacle information, roadside information, and lane line information, and the vehicle positioning and navigation information includes a vehicle pose and a target route;
  • a partial path of the vehicle is obtained based on the result of the sub-path search.
  • the collision detection is to eliminate the sub-paths that may collide, and to ensure that the resulting local path has no collision safety.
  • the application can adopt the concept of path contour to perform obstacle collision detection, that is, the boundary contour formed by the vehicle along the path from the starting position to the end position as a whole, and the collision detection is performed by a geometric method.
  • the partial path of the vehicle is the result of unmanned path planning.
  • the step of generating a sub-path according to the environment sensing information and the vehicle positioning and navigation information to obtain a candidate sub-path satisfying the vehicle constraint is specifically:
  • the step of performing a sub-path search on the non-collision candidate sub-path by using the A* search algorithm includes:
  • the evaluation factor of the sub-path evaluation function includes a lateral acceleration corresponding to the sub-path, a curvature, a curvature change, a cumulative distance, a sub-path end-point angle and a combination of any one or any combination of the difference in the target heading angle and the heuristic distance;
  • the A* search algorithm is used to find the sub-path with the optimal evaluation function value and the corresponding parent path information from the non-collision candidate sub-path;
  • the sub-path with the optimal evaluation function value and the corresponding parent path information are output as the result of the sub-path search.
  • the application adds evaluation items such as the curvature change amount and the heading angle difference in the objective function of the search algorithm (ie, the evaluation function), which ensures that the final planned path is as smooth as possible and reduces overshoot as much as possible; by adding the side in the evaluation function
  • the evaluation of the acceleration ensures that the final planned path can meet the passenger comfort requirements as much as possible.
  • the step of calculating an evaluation function value of each sub-path in the non-collision candidate sub-path includes:
  • the step of calculating an evaluation function value of each sub-path in the non-collision candidate sub-path further includes the following steps:
  • h(s) is the heuristic distance of the current subpath, k a , k ⁇ k , k k , k d , k ⁇ and k h are a(s), ⁇ k(s), k(s), d, respectively. Weight coefficients of s), ⁇ (s) and h(s).
  • route is the target path given by the global path plan, and the route is included in the vehicle navigation and positioning information.
  • the step of calculating a difference between a heading angle of each of the sub-paths in the non-collision candidate sub-path and a target heading angle includes:
  • the step of calculating a target heading angle of each of the sub-paths in the non-collision candidate sub-path includes:
  • the dichotomy is used to calculate the shortest distance from the current point to the first 3 spline segments and the second 3 spline segments;
  • the heading angle of the point on the target path corresponding to the shortest distance from the current point to the target path is selected as the target heading angle ⁇ route of the sub-path.
  • the target path is the vehicle center line, and the vehicle center line is described by 8 segments of 3 times spline curve.
  • Selecting the relatively small value of the shortest distance from the current point to the first 3 spline segments and the second 3 spline segments as the shortest distance from the current point to the target path means: if the current point is The shortest distance of the first 3 spline segments is smaller than the shortest distance from the current point to the second 3 spline segments, and the shortest distance from the current point to the target path is equal to the shortest distance from the current point to the first 3 spline segments. distance.
  • the step of calculating the shortest distance between the current point and the first 3 times of the spline segment and the second 3rd spline segment includes:
  • Step 1 The given 3 times spline segment is divided into n segments according to the given dividing parameter t, and n+1 first class nodes are obtained, wherein the given 3 times spline segment is the first 3 a secondary spline segment or a second 3 spline segment;
  • Step 2 calculating the shortest distance from the current point to the n+1 first type nodes and the node Ptemp0 corresponding to the shortest distance;
  • Step 3 dividing the node Ptemp0 corresponding to the shortest distance into two segments in the n segment according to the parameter t, and obtaining n+1 second class nodes;
  • Step 4 calculating the shortest distance dtemp of the current point to n+1 second type nodes and the node Ptemp corresponding to the shortest distance dtemp;
  • an unmanned path planning system including:
  • the information acquiring module is configured to acquire environment sensing information and vehicle positioning and navigation information, where the environment sensing information includes obstacle information, roadside information, and lane line information, and the vehicle positioning and navigation information includes a vehicle pose and a target route;
  • a subpath generating module configured to perform subpath generation according to the environment sensing information and the vehicle positioning and navigation information, to obtain a candidate subpath satisfying the vehicle constraint;
  • a collision detection module configured to perform collision detection on a candidate sub-path satisfying a vehicle constraint to obtain a candidate sub-path without collision
  • the sub-path search module is configured to perform sub-path search on the non-collision candidate sub-path by using an A* search algorithm
  • the local path generation module is configured to obtain a partial path of the vehicle according to the result of the sub-path search.
  • an unmanned path planning device including:
  • a memory configured to store a program
  • a processor configured to execute the program for:
  • Obtaining environment awareness information and vehicle positioning and navigation information wherein the environment awareness information includes obstacle information, roadside information, and lane line information, and the vehicle positioning and navigation information includes a vehicle pose and a target route;
  • a partial path of the vehicle is obtained based on the result of the sub-path search.
  • the block diagram of the automatic driving system of this embodiment is shown in FIG. 2, wherein the partial path planning module performs local path planning according to the environmental model information provided by the sensing information fusion module, the vehicle pose information and the global route provided by the vehicle navigation and positioning module, The result of the planning is output to the path following module.
  • the unmanned path planning method of the present application is mainly implemented by a partial path planning module, and thus other modules of the automatic driving system are not described again.
  • the algorithm block diagram of the local path planning module of the present application is shown in FIG. 3.
  • the input of the local path planning module is environment sensing information and vehicle positioning and navigation information
  • the output is a partial path.
  • the environmental information mainly includes obstacle information, roadside information and lane line information
  • the vehicle positioning and navigation information mainly includes the vehicle pose (x v , y v , ⁇ v ) and the target route.
  • the target route is the lane centerline, which is described by an 8-segment 3-time spline curve.
  • the partial path planning module of the present application includes four modules: subpath generation, subpath collision detection, subpath search, and path generation.
  • any obstacle obj is represented by a line segment which is described by two endpoints A(x 1 , y 1 ) and B(x 2 , y 2 ).
  • the vehicle contour is simplified into a rectangle, and the specific collision detection parameters mainly include three parameters: len length, width vehicle width and h0 rear suspension length.
  • the present application also performs a certain expansion swell of the shape parameter of the vehicle to ensure a certain safety distance between the real vehicle and the obstacle. Therefore, in the present application, the vehicle profile is represented by four parameters: len, width, h0, and swell.
  • a schematic diagram of the vehicle profile, sub-paths and obstacles of the present application is shown in FIG.
  • the sub-path generation module is configured to predict the vehicle state and generate a series of vehicle-constrained candidate sub-paths according to the current pose of the vehicle.
  • any subpath is defined by Five parameters are described, so when When necessary, according to the control variable method, different sub-paths can be obtained corresponding to different parameters r.
  • a set of discrete sub-path sets can be obtained under the action of a set of discrete curvatures k.
  • Figure 5 shows a schematic diagram of the current pose and sub-path of the vehicle at different curvatures k.
  • k is also constrained to:
  • the sub-path collision detection process is a very important part of the local path planning of the present application, and is used to ensure that the final local path result has no collision.
  • the patent application adopts the concept of path contour to perform obstacle collision detection, that is, the boundary contour formed by the vehicle along the path from the starting position to the end position as a whole, and the collision detection is performed by a geometric method.
  • the sub-path search sub-module of the present application evaluates the sub-paths in the OPEN set (ie, the open sequence) by constructing an evaluation function, and the comprehensive evaluation value is the lowest ( That is, the subpath whose evaluation function value is the smallest is saved in the close sequence and used for the parent path of the subpath generation module.
  • the evaluation function (ie, the objective function) is used to evaluate the sub-paths to ensure the comfort, efficiency and optimality of the final path.
  • the evaluation function is composed of six parts, including the lateral acceleration corresponding to the sub-path, the curvature, the curvature change, the cumulative distance, the difference between the sub-path end-point angle and the target heading angle, and the heuristic distance.
  • lateral acceleration is one of the indicators that best reflects the ride comfort of the car. Excessive lateral acceleration is unfavorable for comfort.
  • the present application calculates the lateral acceleration a of each sub-path by the following formula:
  • arctan[(l f +l r ) ⁇ k]
  • M is the mass of the vehicle
  • v is the speed of the vehicle
  • k f and k r are the tire lateral stiffness of the front and rear wheels
  • l f and l r are the distance from the front and rear axles to the center of gravity of the vehicle
  • k is the curvature.
  • the state in which the steering wheel angle (ie, curvature) input is smaller has priority in the selection without collision.
  • the path should be as smooth as possible while ensuring no collision safety. Therefore, in order to ensure excessive smoothness between adjacent two sub-paths, the present application adds a limitation on the curvature variation of adjacent two sub-paths in the evaluation function, and ensures that the path of the final planned storage is as smooth as possible.
  • the cumulative distance represents the distance from the initial vehicle state to the current vehicle state.
  • the cumulative distance of the sub-path can be obtained by calculating the sum of the cumulative distances of the current sub-subpath and its corresponding parent path, and the calculation method can be as follows Formula to represent:
  • d(parent_subpath) represents the cumulative distance of the parent path
  • ds(Current_subpath) represents the path length of the current subpath
  • the heading angle difference is used to limit the final planned path to ensure that there is no collision safety, and there is no large heading fluctuation along the target path as much as possible to ensure smooth path.
  • the sub-path end-point angle ⁇ end is calculated as:
  • ⁇ 0 is the heading angle corresponding to the starting point of the sub-path
  • ds is the length of the sub-path
  • r is the turning radius of the sub-path
  • the target heading angle ⁇ route of the present application is determined by calculating the heading angle of the point on the target path corresponding to the shortest distance dmin of the target path, and the calculation algorithm flow is shown in FIG. 7 .
  • the heading angle difference ⁇ is calculated by the angle formula of the vector, and its calculation formula is as follows:
  • the heuristic distance is used to indicate the distance of the current subpath from the target.
  • the heuristic distance h is represented by the shortest distance from the end point of the current sub-path to the target path, and the calculation formula is as follows:
  • S end is the end point coordinate of the current sub-path
  • route is the target path given by the global path plan
  • d min (S end , route) is the shortest distance from the end point of the current sub-path to the target path.
  • s is the current sub-path
  • a(s) is the lateral acceleration
  • k(s) is the curvature
  • ⁇ k(s) is the curvature change
  • d(s) is the cumulative distance
  • ⁇ (s) is the heading angle difference
  • h(s) is the heuristic distance.
  • the sub-path search module of the present application adopts the A* algorithm as the sub-path search algorithm, the search flow is as shown in FIG. 6, and the search result is shown in FIG. 9.
  • the result of the sub-path search of the present application satisfies No collision, smooth, as far as possible along the target path.
  • the path generation sub-module When the search termination condition is satisfied, the path generation sub-module generates a current desired vehicle travel route (ie, a local path of the vehicle) according to the search result of the sub-path search module and the parent-child relationship between the sub-paths.
  • a current desired vehicle travel route ie, a local path of the vehicle
  • the result of the path planning of the present application is shown in Fig. 9.
  • the local path planning result in the case of not considering the obstacle is a straight line segment
  • the local path planning result in the case of considering the obstacle is a curved curved section. It can be seen from the path shown in FIG.
  • the result of the sub-path search of the present application satisfies the comprehensive evaluation requirement of the evaluation function, and achieves: first, the search result has no collision; second, the search result satisfies the curvature change requirement, and the path is smooth; The search results satisfy the requirement of traveling along a local path as much as possible, and there is no large overshoot of the heading angle.
  • the search algorithm can ensure that the curvature mutation at the intersection of two adjacent sub-paths is as small as possible, and the output of the present application is sent to the subsequent path tracking module for execution. And in the path tracking module, the change of the excessive curvature is processed, so as to ensure the stable running of the vehicle, and also ensure that the deviation between the actual driving trajectory of the vehicle and the partial path outputted by the present application reaches a small value. In order to achieve the safety of the car driving.
  • Line segments are used to express obstacles, and straight arcs are used to express paths, which facilitates storage calculation and construction of mathematical models;

Landscapes

  • Engineering & Computer Science (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • Physics & Mathematics (AREA)
  • Automation & Control Theory (AREA)
  • General Physics & Mathematics (AREA)
  • Aviation & Aerospace Engineering (AREA)
  • Transportation (AREA)
  • Mechanical Engineering (AREA)
  • Human Computer Interaction (AREA)
  • Theoretical Computer Science (AREA)
  • Multimedia (AREA)
  • Electromagnetism (AREA)
  • Chemical & Material Sciences (AREA)
  • Combustion & Propulsion (AREA)
  • Navigation (AREA)
  • Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)
  • Traffic Control Systems (AREA)

Abstract

提供了一种无人驾驶路径规划方法、***和装置,其中该方法包括:获取环境感知信息和车辆定位与导航信息,其中,环境感知信息包括障碍物信息、路沿信息和车道线信息,车辆定位与导航信息包括车辆位姿和目标路线;根据环境感知信息和车辆定位与导航信息进行子路径生成,得到满足车辆约束的候选子路径;对满足车辆约束的候选子路径进行碰撞检测,得到无碰撞的候选子路径;采用A*搜索算法对无碰撞的候选子路径进行子路径搜索;根据子路径搜索的结果得到车辆的局部路径。该无人驾驶路径规划方法、***和装置具有对存储空间的要求小,能满足车辆非完整约束的要求,适应性好和可扩展性好的优点,可广泛应用于自动驾驶领域。

Description

一种无人驾驶路径规划方法、***和装置
相关申请的交叉引用
本申请要求于2017年08月31日提交中国专利局的申请号为CN201710770252.6、名称为“一种无人驾驶路径规划方法、***和装置”的中国专利申请的优先权,其全部内容通过引用结合在本申请中。
技术领域
本申请涉及自动驾驶领域,尤其是一种无人驾驶路径规划方法、***和装置。
背景技术
汽车自动驾驶技术(也称为汽车无人驾驶技术)在出行安全、节能环保等方面存在巨大潜力,被认为是解决交通拥堵、降低交通事故和改善环境污染的有效途径。在最近一段时间内,汽车自动驾驶技术得到了广泛关注,成为未来汽车发展的主要方向之一。
一般而言,一个自动驾驶***包括以下几个模块:环境感知模块、路径规划模块、控制执行模块和人机界面模块,上述四个模块对整个自动驾驶***至关重要,直接影响***的智能化水平。路径规划模块负责对本车(即自车,或者被控制的车)的横向运动进行规划,以保障自动驾驶汽车的安全性、舒适性和稳定性,是自动驾驶***不可或缺和至关重要的环节。
一个好的自动驾驶路径规划模块需要综合考虑车辆非完整约束、生成路径的最优性以及针对不同交通场景的适应能力,因此路径规划是自动驾驶技术研究的重点方向。
目前的汽车无人驾驶路径规划方法主要包括:
(1)美国发明专利US 2016/0313133A1提出的一种反应式路径规划方法(reactive path planning technique),该方法的输入为环境感知模块检测到的障碍物信息和车道信息,以及本车的状态信息;其将环境感知模块建立的环境模型进行狄克尼三角剖分,生成一系列的虚拟节点,每一个虚拟节点都是三角形的顶点,每两个顶点之间的连线为三角形的边,然后通过图搜索的方式,并结合搜索条件,得到一条满足要求的路径。该方法的缺点是需要首先对环境模型进行预处理得到大量的虚拟节点,这对存储空间的要求较大,另外通过搜索得到的路径并不一定满足车辆非完整约束的要求,需要对得到的路径进行修正。
(2)美国发明专利US2016/0129907A1提出的一种适应于高速公路等结构化道路的无人驾驶路径规划方法,该方法根据全局路径规划的结果和当前车辆位姿信息,对道路中心进行偏移,得到一系列生成虚拟目标和对应的速度轮廓,并根据上述虚拟目标和速度轮廓,对虚拟目标进行样条拟合,得到满足道路轮廓的候选路径,然后根据环境信息和目标函数,选择满足要求的一条最优路径作为最终路径。该方法的缺点是只能生成跟随道路轮廓的路 径,适应性较差。
(3)美国发明专利US9428187B2提出的一种自动驾驶换道路径规划方法,该方法将换道路径通过5次多项式的形式进行描述,并通过对本车的横纵向运动状态进行预测来确定多项式拟合的目标点。该方法的不足之处是只能对换道路径进行路径规划,可扩展性较差。
发明内容
为解决上述技术问题,本申请的第一目的在于:提供一种对存储空间的要求小,能满足车辆非完整约束的要求,适应性好和可扩展性好的,无人驾驶路径规划方法。
本申请的第二目的在于:提供一种对存储空间的要求小,能满足车辆非完整约束的要求,适应性好和可扩展性好的,无人驾驶路径规划***。
本申请的第三目的在于:提供一种对存储空间的要求小,能满足车辆非完整约束的要求,适应性好和可扩展性好的,无人驾驶路径规划装置。
本申请所采取的第一技术方案是:
一种无人驾驶路径规划方法,包括以下步骤:
获取环境感知信息和车辆定位与导航信息,其中,环境感知信息包括障碍物信息、路沿信息和车道线信息,车辆定位与导航信息包括车辆位姿和目标路线;
根据环境感知信息和车辆定位与导航信息进行子路径生成,得到满足车辆约束的候选子路径;
对满足车辆约束的候选子路径进行碰撞检测,得到无碰撞的候选子路径;
采用A*搜索算法对无碰撞的候选子路径进行子路径搜索;
根据子路径搜索的结果得到车辆的局部路径。
进一步,所述根据环境感知信息和车辆定位与导航信息进行子路径生成,得到满足车辆约束的候选子路径这一步骤,具体为:
根据环境感知信息和车辆定位与导航信息,采用控制变量法得到一组离散的曲率k对应的子路径集合,其中,k=1/r,且曲率k满足的约束为:-k max≤k i≤k max,r为路径对应的转弯半径,k i为该组离散的曲率中的第i个曲率,k max为给定的最大曲率。
进一步,所述采用A*搜索算法对无碰撞的候选子路径进行子路径搜索这一步骤,具体包括:
计算无碰撞的候选子路径中每一条子路径的评价函数值,其中,子路径的评价函数的评价因子包括子路径对应的侧向加速度、曲率、曲率变化、累积距离、子路径终点航向角与目标航向角的差值以及启发式距离中的任意一个或任意几个的组合;
根据计算的评价函数值采用A*搜索算法从无碰撞的候选子路径中寻找出评价函数值最优的子路径以及对应的父路径信息;
将评价函数值最优的子路径以及对应的父路径信息作为子路径搜索的结果进行输出。
进一步,所述计算无碰撞的候选子路径中每一条子路径的评价函数值这一步骤,具体包括:
计算无碰撞的候选子路径中每一条子路径的侧向加速度,所述子路径的侧向加速度a计算公式为:
Figure PCTCN2018102811-appb-000001
其中,v为车速,δ=arctan[(l f+l r)·k[,M 为本车质量,
Figure PCTCN2018102811-appb-000002
k为曲率,k f和k r分别为本车前后轮的轮胎侧偏刚度,l f和l r分别为本车前后轴到车辆重心的距离;
计算无碰撞的候选子路径中每一条子路径的曲率;
计算无碰撞的候选子路径中每一条子路径的曲率变化量;
计算无碰撞的候选子路径中每一条子路径的累积距离,所述子路径的累积距离计算公式为:d(Current_subpath)=d(parent_subpath)+ds(Current_subpath),其中,d(Current_subpath)为当前子路径的累积距离,d(parent_subpath)表示当前子路径对应的父路径的累积距离,ds(Current_subpath)表示当前子路径的路径长度;
计算无碰撞的候选子路径中每一条子路径终点航向角与目标航向角的差值,所述子路径终点航向角与目标航向角的差值Δθ计算公式为:Δθ=ar cos[cos(θ end)·cos(θ route)+sin(θ end)·sin(θ route)],其中,θ end为子路径终点航向角,θ route为目标航向角;
计算无碰撞的候选子路径中每一条子路径的启发距离,所述子路径的启发距离h计算公式为:h=d min(S end,route),其中,S end为当前子路径的终点坐标,route为全局路径规划给出的目标路径,d min(S end,route)为当前子路径的终点到目标路径的最短距离。
进一步,所述计算无碰撞的候选子路径中每一条子路径的评价函数值这一步骤,还包括以下步骤:
根据子路径的侧向加速度、曲率、曲率变化量、累积距离、子路径终点航向角与目标航向角的差值和启发距离计算无碰撞的候选子路径中每一条子路径的评价函数值,所述子路径的评价函数值计算公式为:f(s)=k a·a(s)+k Δk·Δk(s)+k k·k(s)+k d·d(s)+k ΔθΔθ(s)+k h·h(s),其中,s为当前子路径,f(s)为当前子路径的评价函数值,a(s)为当前子路径的侧向加速度,k(s)为当前子路径的曲率,Δk(s)为当前子路径的曲率变化量,d(s)为当前子路径的累积距离,Δθ(s)为当前子路径终点航向角与目标航向角的差值,h(s)为当前子路径的启发距离,k a、k Δk、k k、k d、k Δθ和k h分别为a(s)、Δk(s)、k(s)、d(s)、Δθ(s)和h(s)的权重系数。
进一步,所述计算无碰撞的候选子路径中每一条子路径终点航向角与目标航向角的差值这一步骤,具体包括:
计算无碰撞的候选子路径中每一条子路径终点航向角,所述子路径终点航向角θ end计 算公式为:
Figure PCTCN2018102811-appb-000003
其中,θ 0为子路径起点对应的航向角,ds为子路径的长度,r为子路径的转弯半径;
计算无碰撞的候选子路径中每一条子路径的目标航向角;
根据计算的子路径终点航向角和目标航向角计算无碰撞的候选子路径中每一条子路径终点航向角与目标航向角的差值Δθ。
进一步,所述计算无碰撞的候选子路径中每一条子路径的目标航向角这一步骤,具体包括:
选取无碰撞的候选子路径中每一条子路径的终点作为当前点;
计算当前点到目标路径的8段3次样条曲线段端点的最短距离,并将当前点到目标路径的8段3次样条曲线段端点的最短距离记为第一最短距离;
选择第一最短距离对应的端点所对应的前后两段3次样条曲线段,并将选择的前后两段3次样条曲线段分别记为第一3次样条曲线段和第二3次样条曲线段;
采用二分法分别计算当前点到第一3次样条曲线段和第二3次样条曲线段的最短距离;
从当前点到第一3次样条曲线段和第二3次样条曲线段的最短距离中选出二者的相对较小值作为当前点到目标路径的最短距离;
选取当前点到目标路径的最短距离对应的目标路径上的点的航向角作为子路径的目标航向角θ route
进一步,所述计算当前点到第一3次样条曲线段和第二3次样条曲线段的最短距离这一步骤,具体包括:
步骤一:将给定的3次样条曲线段按照给定的划分参数t划分为n段,得到n+1个第一类节点,其中,给定的3次样条曲线段为第一3次样条曲线段或第二3次样条曲线段;
步骤二:计算当前点到n+1个第一类节点的最短距离以及最短距离对应的节点Ptemp0;
步骤三:将最短距离对应的节点Ptemp0在n段中的相邻两段按照参数t划分为n段,得到n+1个第二类节点;
步骤四:计算当前点到n+1个第二类节点的最短距离dtemp以及最短距离dtemp对应的节点Ptemp;
步骤五:判断最短距离dtemp是否小于设定的最小距离阈值,若是,则以dtemp作为当前点到给定的3次样条曲线段的最短距离进行输出;反之,则令Ptemp0=Ptemp并返回步骤三。
本申请所采取的第二技术方案是:
一种无人驾驶路径规划***,包括:
信息获取模块,配置成获取环境感知信息和车辆定位与导航信息,其中,环境感知信息包括障碍物信息、路沿信息和车道线信息,车辆定位与导航信息包括车辆位姿和目标路线;
子路径生成模块,配置成根据环境感知信息和车辆定位与导航信息进行子路径生成,得到满足车辆约束的候选子路径;
碰撞检测模块,配置成对满足车辆约束的候选子路径进行碰撞检测,得到无碰撞的候选子路径;
子路径搜索模块,配置成采用A*搜索算法对无碰撞的候选子路径进行子路径搜索;
局部路径生成模块,配置成根据子路径搜索的结果得到车辆的局部路径。
本申请所采取的第三技术方案是:
一种无人驾驶路径规划装置,包括:
存储器,配置成存放程序;
处理器,配置成执行所述程序以用于:
获取环境感知信息和车辆定位与导航信息,其中,环境感知信息包括障碍物信息、路沿信息和车道线信息,车辆定位与导航信息包括车辆位姿和目标路线;
根据环境感知信息和车辆定位与导航信息进行子路径生成,得到满足车辆约束的候选子路径;
对满足车辆约束的候选子路径进行碰撞检测,得到无碰撞的候选子路径;
采用A*搜索算法对无碰撞的候选子路径进行子路径搜索;
根据子路径搜索的结果得到车辆的局部路径。
本申请的方法的有益效果是:通过环境感知信息和车辆定位与导航信息获取、子路径生成、碰撞检测、子路径搜索和局部路径生成来得到无人驾驶路径,不再需要对环境模型进行预处理得到大量的虚拟节点,对存储空间的要求小;通过进行子路径生成得到满足车辆约束的候选子路径,并在进行子路径搜索时采用了A*搜索算法,使得规划出的无人驾驶路径更能满足车辆非完整性约束的要求;获取的环境感知信息包括障碍物信息、路沿信息和车道线信息,车辆定位与导航信息包括车辆位姿和目标路线,综合采用了障碍物、换道路径、车道、车辆位姿和目标路线信息来进行路径规划,使得生成的无人驾驶路径不再受道路轮廓或换道路径的限制,适应性和可扩展性更好。
本申请的***的有益效果是:包括信息获取模块、子路径生成模块、碰撞检测模块、子路径搜索模块和局部路径生成模块,通过环境感知信息和车辆定位与导航信息获取、子路径生成、碰撞检测、子路径搜索和局部路径生成来得到无人驾驶路径,不再需要对环境模型进行预处理得到大量的虚拟节点,对存储空间的要求小;通过进行子路径生成得到满足车辆约束的候选子路径,并在进行子路径搜索时采用了A*搜索算法,使得规划出的无人驾驶路径更能满足车辆非完整性约束的要求;获取的环境感知信息包括障碍物信息、路沿信息和车道线信息,车辆定位与导航信息包括车辆位姿和目标路线,综合采用了障碍物、换道路径、车道、车辆位姿和目标路线信息来进行路径规划,使得生成的无人驾驶路径不再受道路轮廓或换道路径的限制,适应性和可扩展性更好。
本申请的装置的有益效果是:包括存储器和处理器,在处理器中通过环境感知信息和车辆定位与导航信息获取、子路径生成、碰撞检测、子路径搜索和局部路径生成来得到无人驾驶路径,不再需要对环境模型进行预处理得到大量的虚拟节点,对存储空间的要求小;通过进行子路径生成得到满足车辆约束的候选子路径,并在进行子路径搜索时采用了A*搜索算法,使得规划出的无人驾驶路径更能满足车辆非完整性约束的要求;获取的环境感知信息包括障碍物信息、路沿信息和车道线信息,车辆定位与导航信息包括车辆位姿和目标路线,综合采用了障碍物、换道路径、车道、车辆位姿和目标路线信息来进行路径规划,使得生成的无人驾驶路径不再受道路轮廓或换道路径的限制,适应性和可扩展性更好。
附图说明
图1为本申请一种无人驾驶路径规划方法的流程图;
图2为本申请实施例一自动驾驶***的结构框图;
图3为本申请实施例一局部路径规划模块的算法框架图;
图4为本申请实施例一路径、障碍物和车辆轮廓示意图;
图5为本申请实施例一在不同曲率下车辆当前位姿与子路径的示意图;
图6为本申请实施例一子路径搜索流程图;
图7为本申请实施例一当前点P到目标路径最短距离的计算流程图;
图8为本申请实施例一当前点P到给定的三次样条曲线段的最短距离的计算流程图;
图9为本申请实施例一的路径规划结果图。
具体实施方式
参照图1,一种无人驾驶路径规划方法,包括以下步骤:
获取环境感知信息和车辆定位与导航信息,其中,环境感知信息包括障碍物信息、路沿信息和车道线信息,车辆定位与导航信息包括车辆位姿和目标路线;
根据环境感知信息和车辆定位与导航信息进行子路径生成,得到满足车辆约束的候选子路径;
对满足车辆约束的候选子路径进行碰撞检测,得到无碰撞的候选子路径;
采用A*搜索算法对无碰撞的候选子路径进行子路径搜索;
根据子路径搜索的结果得到车辆的局部路径。
其中,碰撞检测是为了剔除可能发生碰撞的子路径,保证最终得到的局部路径无碰撞安全。本申请可采用路径轮廓的概念来进行障碍物碰撞检测,即将车辆沿着路径从起点位姿行驶至终点位姿所形成的边界轮廓作为一个整体,并通过几何的方法进行碰撞检测。车辆的局部路径即为无人驾驶路径规划的结果。
进一步作为优选的实施方式,所述根据环境感知信息和车辆定位与导航信息进行子路径生成,得到满足车辆约束的候选子路径这一步骤,具体为:
根据环境感知信息和车辆定位与导航信息,采用控制变量法得到一组离散的曲率k对应的子路径集合,其中,k=1/r,且曲率k满足的约束为:-k max≤k i≤k max,r为路径对应的转弯半径,k i为该组离散的曲率中的第i个曲率,k max为给定的最大曲率。
进一步作为优选的实施方式,所述采用A*搜索算法对无碰撞的候选子路径进行子路径搜索这一步骤,具体包括:
计算无碰撞的候选子路径中每一条子路径的评价函数值,其中,子路径的评价函数的评价因子包括子路径对应的侧向加速度、曲率、曲率变化、累积距离、子路径终点航向角与目标航向角的差值以及启发式距离中的任意一个或任意几个的组合;
根据计算的评价函数值采用A*搜索算法从无碰撞的候选子路径中寻找出评价函数值最优的子路径以及对应的父路径信息;
将评价函数值最优的子路径以及对应的父路径信息作为子路径搜索的结果进行输出。
本申请在搜索算法的目标函数(即评价函数)中增加了曲率变化量和航向角差值等评价项,保证了最终规划路径尽可能平顺和尽可能减少超调;通过在评价函数中增加侧向加速度的评价项,保证了最终规划路径能尽可能满足乘客舒适性的要求。
进一步作为优选的实施方式,所述计算无碰撞的候选子路径中每一条子路径的评价函数值这一步骤,具体包括:
计算无碰撞的候选子路径中每一条子路径的侧向加速度,所述子路径的侧向加速度a计算公式为:
Figure PCTCN2018102811-appb-000004
其中,v为车速,δ=arc tan[(l f+l r)·k],M为本车质量,
Figure PCTCN2018102811-appb-000005
k为曲率,k f和k r分别为本车前后轮的轮胎侧偏刚度,l f和l r分别为本车前后轴到车辆重心的距离;
计算无碰撞的候选子路径中每一条子路径的曲率;
计算无碰撞的候选子路径中每一条子路径的曲率变化量;
计算无碰撞的候选子路径中每一条子路径的累积距离,所述子路径的累积距离计算公式为:d(Current_subpath)=d(parent_subpath)+ds(Current_subpath),其中,d(Current_subpath)为当前子路径的累积距离,d(parent_subpath)表示当前子路径对应的父路径的累积距离,ds(Current_subpath)表示当前子路径的路径长度;
计算无碰撞的候选子路径中每一条子路径终点航向角与目标航向角的差值,所述子路径终点航向角与目标航向角的差值Δθ计算公式为:Δθ=ar cos[cos(θ end)·cos(θ route)+sin(θ end)·sin(θ route)],其中,θ end为子路径终点航向角,θ route为目标航向角;
计算无碰撞的候选子路径中每一条子路径的启发距离,所述子路径的启发距离h计算公式为:h=d min(S end,route),其中,S end为当前子路径的终点坐标,route为全局路径规划给出的目标路径,d min(S end,route)为当前子路径的终点到目标路径的最短距离。
进一步作为优选的实施方式,所述计算无碰撞的候选子路径中每一条子路径的评价函数值这一步骤,还包括以下步骤:
根据子路径的侧向加速度、曲率、曲率变化量、累积距离、子路径终点航向角与目标航向角的差值和启发距离计算无碰撞的候选子路径中每一条子路径的评价函数值,所述子路径的评价函数值计算公式为:f(s)=k a·a(s)+k Δk·Δk(s)+k k·k(s)+k d·d(s)+k ΔθΔθ(s)+k h·h(s),其中,s为当前子路径,f(s)为当前子路径的评价函数值,a(s)为当前子路径的侧向加速度,k(s)为当前子路径的曲率,Δk(s)为当前子路径的曲率变化量,d(s)为当前子路径的累积距离,Δθ(s)为当前子路径终点航向角与目标航向角的差值,h(s)为当前子路径的启发距离,k a、k Δk、k k、k d、k Δθ和k h分别为a(s)、Δk(s)、k(s)、d(s)、Δθ(s)和h(s)的权重系数。
其中,route为全局路径规划给出的目标路径,route包含在车辆导航与定位信息中。
进一步作为优选的实施方式,所述计算无碰撞的候选子路径中每一条子路径终点航向角与目标航向角的差值这一步骤,具体包括:
计算无碰撞的候选子路径中每一条子路径终点航向角,所述子路径终点航向角θ end计算公式为:
Figure PCTCN2018102811-appb-000006
其中,θ 0为子路径起点对应的航向角,ds为子路径的长度,r为子路径的转弯半径;
计算无碰撞的候选子路径中每一条子路径的目标航向角;
根据计算的子路径终点航向角和目标航向角计算无碰撞的候选子路径中每一条子路径终点航向角与目标航向角的差值Δθ。
进一步作为优选的实施方式,所述计算无碰撞的候选子路径中每一条子路径的目标航 向角这一步骤,具体包括:
选取无碰撞的候选子路径中每一条子路径的终点作为当前点;
计算当前点到目标路径的8段3次样条曲线段端点的最短距离,并将当前点到目标路径的8段3次样条曲线段端点的最短距离记为第一最短距离;
选择第一最短距离对应的端点所对应的前后两段3次样条曲线段,并将选择的前后两段3次样条曲线段分别记为第一3次样条曲线段和第二3次样条曲线段;
采用二分法分别计算当前点到第一3次样条曲线段和第二3次样条曲线段的最短距离;
从当前点到第一3次样条曲线段和第二3次样条曲线段的最短距离中选出二者的相对较小值作为当前点到目标路径的最短距离;
选取当前点到目标路径的最短距离对应的目标路径上的点的航向角作为子路径的目标航向角θ route
其中,目标路径为车辆中心线,车辆中心线通过8段3次样条曲线进行描述。从当前点到第一3次样条曲线段和第二3次样条曲线段的最短距离中选出二者的相对较小值作为当前点到目标路径的最短距离是指:若当前点到第一3次样条曲线段的最短距离小于当前点到第二3次样条曲线段的最短距离,则当前点到目标路径的最短距离等于当前点到第一3次样条曲线段的最短距离。
进一步作为优选的实施方式,所述计算当前点到第一3次样条曲线段和第二3次样条曲线段的最短距离这一步骤,具体包括:
步骤一:将给定的3次样条曲线段按照给定的划分参数t划分为n段,得到n+1个第一类节点,其中,给定的3次样条曲线段为第一3次样条曲线段或第二3次样条曲线段;
步骤二:计算当前点到n+1个第一类节点的最短距离以及最短距离对应的节点Ptemp0;
步骤三:将最短距离对应的节点Ptemp0在n段中的相邻两段按照参数t划分为n段,得到n+1个第二类节点;
步骤四:计算当前点到n+1个第二类节点的最短距离dtemp以及最短距离dtemp对应的节点Ptemp;
步骤五:判断最短距离dtemp是否小于设定的最小距离阈值,若是,则以dtemp作为当前点到给定的3次样条曲线段的最短距离进行输出;反之,则令Ptemp0=Ptemp并返回步骤三。
其中,令Ptemp0=Ptemp并返回步骤三,是指返回最短距离dtemp对应的节点Ptemp,使得Ptemp0=Ptemp,然后重复上述步骤(即重复执行步骤三以及步骤四这两个步骤)直至最短距离dtemp小于设定的最小距离阈值为止。
与图1的方法相对应,本申请还提供了一种无人驾驶路径规划***,包括:
信息获取模块,配置成获取环境感知信息和车辆定位与导航信息,其中,环境感知信息包括障碍物信息、路沿信息和车道线信息,车辆定位与导航信息包括车辆位姿和目标路线;
子路径生成模块,配置成根据环境感知信息和车辆定位与导航信息进行子路径生成,得到满足车辆约束的候选子路径;
碰撞检测模块,配置成对满足车辆约束的候选子路径进行碰撞检测,得到无碰撞的候选子路径;
子路径搜索模块,配置成采用A*搜索算法对无碰撞的候选子路径进行子路径搜索;
局部路径生成模块,配置成根据子路径搜索的结果得到车辆的局部路径。
与图1的方法相对应,本申请还提供了一种无人驾驶路径规划装置,包括:
存储器,配置成存放程序;
处理器,配置成执行所述程序以用于:
获取环境感知信息和车辆定位与导航信息,其中,环境感知信息包括障碍物信息、路 沿信息和车道线信息,车辆定位与导航信息包括车辆位姿和目标路线;
根据环境感知信息和车辆定位与导航信息进行子路径生成,得到满足车辆约束的候选子路径;
对满足车辆约束的候选子路径进行碰撞检测,得到无碰撞的候选子路径;
采用A*搜索算法对无碰撞的候选子路径进行子路径搜索;
根据子路径搜索的结果得到车辆的局部路径。
下面结合说明书附图和具体实施例对本申请作进一步解释和说明。
实施例一
针对现有技术对存储空间的要求大,不能满足车辆非完整约束的要求,适应性差和可扩展性差的问题,本申请专门为自动驾驶***设计了一种新的无人驾驶路径规划方法、***和装置。
本实施例的自动驾驶***框图如图2所示,其中,局部路径规划模块根据感知信息融合模块提供的环境模型信息、车辆导航与定位模块提供的车辆位姿信息和全局路线进行局部路径规划,并将规划的结果输出给路径跟随模块。本申请的无人驾驶路径规划方法主要由局部路径规划模块来实现,因此自动驾驶***的其它模块不再赘述。
本申请局部路径规划模块的算法框图如图3所示,从图3中可以看出,局部路径规划模块的输入为环境感知信息和车辆定位与导航信息,输出为局部路径。其中,环境信息主要包括障碍物信息、路沿信息和车道线信息;车辆定位与导航信息主要包括车辆位姿(x v,y vv)和目标路线。在本申请中,目标路线为车道中心线,通过8段3次样条曲线进行描述。
从图3可以看出,本申请的局部路径规划模块包括子路径生成、子路径碰撞检测、子路径搜索和路径生成四个模块。在本申请中,子路径subpath通过
Figure PCTCN2018102811-appb-000007
五个参数表示,其中,
Figure PCTCN2018102811-appb-000008
为子路径对应的车辆起点位姿,ds为路径长度,ds为正,表示车辆向前行驶,ds为负,表示车辆向后行驶;r为路径对应的转弯半径,r为正,表示车左转,r为负,表示车辆右转,当r=0时,表示车辆直线行驶。在本申请中,任意一个障碍物obj由一条线段表示,该线段通过两个端点A(x 1,y 1)和B(x 2,y 2)进行描述。本申请在碰撞检测过程中将车辆轮廓简化为一个长方形,具体碰撞检测参数主要包括len车长、width车宽和h0后悬长三个参数。而要使车辆安全地在规划的路径上行驶,必须使得其外轮廓与障碍物保持一定的距离。为此,本申请在碰撞检测时,还将车辆的外形参数进行一定的膨胀swell,以保证真实车辆与障碍物存在一定的安全距离。因此在本申请中,车辆轮廓由len、width、h0和swell四个参数表示。本申请的车辆轮廓、子路径和障碍物示意图如图4所示。
下面详细介绍本申请的局部路径规划算法中各组成模块的实现方法。
(1)子路径生成
子路径生成模块用于对车辆状态进行预测,并根据车辆当前位姿生成一系列满足车辆约束候选子路径。
前面已经定义,任意一条子路径由
Figure PCTCN2018102811-appb-000009
五个参数进行描述,因此当
Figure PCTCN2018102811-appb-000010
一定时,根据控制变量方法,对应不同的参数r,可以得到不同的子路径。本申请可通过变量曲率k来生成子路径,其中,k和r满足:k=1/r。具体来说,就是在一组离散的曲率k的作用下,可以得到一组离散的子路径集合。图5给出了在不同曲率k下,车辆当前位姿和子路径的示意图。
由于车辆转弯半径的约束,在子路径生成模块中,还需将k约束为:
-k max≤k i≤k max
(2)子路径碰撞检测
子路径碰撞检测过程是本申请局部路径规划中非常重要的环节,用于保证最终的局部路径结果无碰撞。本专利申请采用路径轮廓的概念来进行障碍物碰撞检测,即将车辆沿着路径从起点位姿行驶至终点位姿所形成的边界轮廓作为一个整体,并通过几何的方法进行碰撞检测。
(3)子路径搜索
如图6所示,基于A*搜索算法的思想和规则,本申请的子路径搜索子模块通过构建一个评价函数对OPEN集合(即open序列)中的子路径进行评价,将综合评价值最低(即评价函数值最小)的子路径保存到close序列中并用于子路径生成模块的父路径。
1)评价函数
评价函数(即目标函数)用于对子路径进行评价,保证最终路径的舒适性、效率和最优性。在本申请中,评价函数由6部分构成,包括子路径对应的侧向加速度、曲率、曲率变化、累积距离、子路径终点航向角与目标航向角的差值以及启发式距离。
这6部分的详细说明如下:
a)侧向加速度
对于乘客而言,侧向加速度是最能体现汽车乘坐舒适性的指标之一,侧向加速度过大对舒适性而言是不利的。结合二自由度车辆模型,本申请通过如下公式来计算每一条子路径的侧向加速度a:
Figure PCTCN2018102811-appb-000011
其中δ=arctan[(l f+l r)·k],
Figure PCTCN2018102811-appb-000012
M为本车质量,v是车速,k f和k r是前后轮的轮胎侧偏刚度,l f和l r为前后轴到车辆重心的距离,k为曲率。
b)曲率
在自动驾驶过程中,一个非常重要的控制目标就是在保证安全的前提下使得横向输入尽可能的小,以保证车辆稳定性和舒适性。因此状态(即子路径)搜索过程中,在保证无碰撞的前提下,转向盘转角(即曲率)输入越小的状态具有优先被选择权。
c)曲率变化
在自动驾驶过程中,在保证无碰撞安全的前提下,路径应尽可能保证平滑。因此为保证相邻两段子路径之间的过度更平滑,本申请在评价函数中增加了对相邻两段子路径曲率变化量的限制,保证了最终规划储的路径尽可能平滑。
d)累积距离
累积距离表示由初始车辆状态到当前车辆状态的距离。在本申请中,由于每一段子路径的长度ds是已知的,因此可以通过计算当前段子路径及其对应的父路径的累积距离的和来得到子路径的累积距离,其计算方法可以通过如下公式来表示:
d(Current_subpath)=d(parent_subpath)+ds(Current_subpath)
其中,d(parent_subpath)表示父路径的累积距离,ds(Current_subpath)表示当前子路径的路径长度。
e)子路径终点航向角与目标航向角的差值
该航向角差值用于限制最终规划出的路径在保证无碰撞安全的前提下,尽可能沿着目标路径而不存在大的航向波动,以保证路径的平顺。在本申请中,子路径终点航向角θ end计算公式为:
Figure PCTCN2018102811-appb-000013
其中,θ 0为子路径起点对应的航向角,ds为子路径的长度,r为子路径的转弯半径。
而本申请的目标航向角θ route则通过计算子路径终点到目标路径最短距离dmin对应的目标路径上的点的航向角来确定,其计算算法流程如图7所示。
而对于给定的一段三次样条曲线来说,其可采用二分法的思想来进行近似求解出子路径终点到该段三次样条曲线的最短距离Pmin,具体算法流程如图8所示。图8中,d0为设定的最小距离阈值。通过图7和图8的两个流程图,最终求得当前点P到目标路径的最短距离以及该最短距离对应的目标路径上的点P route,进而通过点P route的航向角确定目标航向角θ route
而航向角差值Δθ则通过向量的夹角公式进行计算,其计算公式如下:
Δθ=ar cos[cos(θ end)·cos(θ route)+sin(θ end)·sin(θ route)]
f)启发距离
启发距离用于表示当前子路径距离目标的距离。在本申请中,启发距离h用当前子路径的终点到目标路径的最短距离来表示,计算公式如下:
h=d min(S end,route)
其中,S end为当前子路径的终点坐标,route为全局路径规划给出的目标路径,d min(S end,route)为当前子路径的终点到目标路径的最短距离。
综上可以得到,本申请的评价函数f(s)的表达式为:
f(s)=k a·a(s)+k Δk·Δk(s)+k k·k(s)+k d·d(s)+k ΔθΔθ(s)+k h·h(s)
其中,s为当前子路径,a(s)为侧向加速度,k(s)为曲率,Δk(s)为曲率变化,d(s)为累积距离,Δθ(s)为航向角差值,h(s)为启发距离,在上述公式中,每一项评价指标前面都对应一个权重系数,该对应的权重系数可以根据实际情况进行修正。
2)搜索
本申请的子路径搜索模块采用A*算法作为子路径搜索算法,搜索流程如图6所示,搜索结果如图9所示,从图9中可以看出,本申请子路径搜索的结果满足了无碰撞、平滑、尽量沿目标路径行驶等要求。
(4)局部路径生成
当满足搜索终止条件时,路径生成子模块会根据上述的子路径搜索模块的搜索结果以及各子路径之间的父子关系,生成当前期望的本车行驶路径(即本车的局部路径)。本申请的路径规划结果如图9所示,在不考虑障碍物情况下的局部路径规划结果为直线线段,在 考虑障碍物情况下的局部路径规划结果为弯曲的曲线段。从图9显示的路径可以看出,本申请子路径搜索的结果满足了评价函数的综合评价要求,做到了:一,搜索结果无碰撞;二,搜索结果满足曲率变化要求,路径平顺;三,搜索结果满足尽量沿局部路径行驶,航向角无大超调的要求。
由于在子路径搜索算法中增加了曲率变化的评价指标,本搜索算法能够保证相邻两段子路径交接处曲率突变尽可能小,同时,本申请的输出结果会发送给后续的路径跟踪模块进行执行,并在路径跟踪模块中会对过度处曲率的变化进行处理,从而在保证车辆稳定行驶的同时,也可以保证车辆实际行驶轨迹与本申请输出的局部路径之间的偏差达到一个较小值,以实现本车行驶的安全性。
与现有技术相比,本申请具有以下优点:
(1)采用了线段来表达障碍物,直线圆弧来表达路径,便于存储计算以及数学模型的构建;
(2)采用了圆弧段子路径作为状态,并采用A*算法作为搜索算法,使得规划出的路径更能满足车辆非完整性约束的要求;
(3)通过碰撞检测剔除可能发生碰撞的子路径,保证最终规划出的路径无碰撞安全;
(4)在搜索算法的目标函数中增加了曲率变化量和航向角差值等评价项,保证最终规划出的路径尽可能平顺和尽可能减少超调;
(5)通过在评价函数中增加侧向加速度的评价项,保证最终规划出的路径能尽可能满足乘客舒适性的要求;
(6)综合采用了障碍物、换道路径、车道、车辆位姿和目标路线信息来进行路径规划,实现了避障、换道和车道保持的功能,可以尽可能减少决策模块的状态切换。
以上是对本申请的较佳实施进行了具体说明,但本申请并不限于所述实施例,熟悉本领域的技术人员在不违背本申请精神的前提下还可做作出种种的等同变形或替换,这些等同的变形或替换均包含在本申请权利要求所限定的范围内。

Claims (10)

  1. 一种无人驾驶路径规划方法,其特征在于:包括以下步骤:
    获取环境感知信息和车辆定位与导航信息,其中,环境感知信息包括障碍物信息、路沿信息和车道线信息,车辆定位与导航信息包括车辆位姿和目标路线;
    根据环境感知信息和车辆定位与导航信息进行子路径生成,得到满足车辆约束的候选子路径;
    对满足车辆约束的候选子路径进行碰撞检测,得到无碰撞的候选子路径;
    采用A*搜索算法对无碰撞的候选子路径进行子路径搜索;
    根据子路径搜索的结果得到车辆的局部路径。
  2. 根据权利要求1所述的一种无人驾驶路径规划方法,其特征在于:所述根据环境感知信息和车辆定位与导航信息进行子路径生成,得到满足车辆约束的候选子路径这一步骤,具体为:
    根据环境感知信息和车辆定位与导航信息,采用控制变量法得到一组离散的曲率k对应的子路径集合,其中,k=1/r,且曲率k满足的约束为:-k max≤k i≤k max,r为路径对应的转弯半径,k i为该组离散的曲率中的第i个曲率,k max为给定的最大曲率。
  3. 根据权利要求1所述的一种无人驾驶路径规划方法,其特征在于:所述采用A*搜索算法对无碰撞的候选子路径进行子路径搜索这一步骤,具体包括:
    计算无碰撞的候选子路径中每一条子路径的评价函数值,其中,子路径的评价函数的评价因子包括子路径对应的侧向加速度、曲率、曲率变化、累积距离、子路径终点航向角与目标航向角的差值以及启发式距离中的任意一个或任意几个的组合;
    根据计算的评价函数值采用A*搜索算法从无碰撞的候选子路径中寻找出评价函数值最优的子路径以及对应的父路径信息;
    将评价函数值最优的子路径以及对应的父路径信息作为子路径搜索的结果进行输出。
  4. 根据权利要求3所述的一种无人驾驶路径规划方法,其特征在于:所述计算无碰撞的候选子路径中每一条子路径的评价函数值这一步骤,具体包括:
    计算无碰撞的候选子路径中每一条子路径的侧向加速度,所述子路径的侧向加速度a计算公式为:
    Figure PCTCN2018102811-appb-100001
    其中,v为车速,δ=arc tan[(l f+l r)·k],M 为本车质量,
    Figure PCTCN2018102811-appb-100002
    k为曲率,k f和k r分别为本车前后轮的轮胎侧偏刚度,l f和l r分别为本车前后轴到车辆重心的距离;
    计算无碰撞的候选子路径中每一条子路径的曲率;
    计算无碰撞的候选子路径中每一条子路径的曲率变化量;
    计算无碰撞的候选子路径中每一条子路径的累积距离,所述子路径的累积距离计算公式为:d(Current_subpath)=d(parent_subpath)+ds(Current_subpath),其中,d(Current_subpath)为当前子路径的累积距离,d(parent_subpath)表示当前子路径对应的父路径的累积距离,ds(Current_subpath)表示当前子路径的路径长度;
    计算无碰撞的候选子路径中每一条子路径终点航向角与目标航向角的差值,所述子路径终点航向角与目标航向角的差值Δθ计算公式为:Δθ=ar cos[cos(θ end)·cos(θ route)+sin(θ end)·sin(θ route)],其中,θ end为子路径终点航向角,θ route为目标航向角;
    计算无碰撞的候选子路径中每一条子路径的启发距离,所述子路径的启发距离h计算公式为:h=d min(S end,route),其中,S end为当前子路径的终点坐标,route为全局路径规划给出的目标路径,d min(S end,route)为当前子路径的终点到目标路径的最短距离。
  5. 根据权利要求4所述的一种无人驾驶路径规划方法,其特征在于:所述计算无碰撞的候选子路径中每一条子路径的评价函数值这一步骤,还包括以下步骤:
    根据子路径的侧向加速度、曲率、曲率变化量、累积距离、子路径终点航向角与目标航向角的差值和启发距离计算无碰撞的候选子路径中每一条子路径的评价函数值,所述子路径的评价函数值计算公式为:f(s)=k a·a(s)+k Δk·Δk(s)+k k·k(s)+k d·d(s)+k ΔθΔθ(s)+k h·h(s),其中,s为当前子路径,f(s)为当前子路径的评价函数值,a(s)为当前子路径的侧向加速度,k(s)为当前子路径的曲率,Δk(s)为当前子路径的曲率变化量,d(s)为当前子路径的累积距离,Δθ(s)为当前子路径终点航向角与目标航向角的差值,h(s)为当前子路径的启发距离,k a、k Δk、k k、k d、k Δθ和k h分别为a(s)、Δk(s)、k(s)、d(s)、Δθ(s)和h(s)的权重系数。
  6. 根据权利要求4或5所述的一种无人驾驶路径规划方法,其特征在于:所述计算无碰撞的候选子路径中每一条子路径终点航向角与目标航向角的差值这一步骤,具体包括:
    计算无碰撞的候选子路径中每一条子路径终点航向角,所述子路径终点航向角θ end计算公式为:
    Figure PCTCN2018102811-appb-100003
    其中,θ 0为子路径起点对应的航向角,ds为子路径的长度,r为子路径的转弯半径;
    计算无碰撞的候选子路径中每一条子路径的目标航向角;
    根据计算的子路径终点航向角和目标航向角计算无碰撞的候选子路径中每一条子路径终点航向角与目标航向角的差值Δθ。
  7. 根据权利要求6所述的一种无人驾驶路径规划方法,其特征在于:所述计算无碰撞的候选子路径中每一条子路径的目标航向角这一步骤,具体包括:
    选取无碰撞的候选子路径中每一条子路径的终点作为当前点;
    计算当前点到目标路径的8段3次样条曲线段端点的最短距离,并将当前点到目标路径的8段3次样条曲线段端点的最短距离记为第一最短距离;
    选择第一最短距离对应的端点所对应的前后两段3次样条曲线段,并将选择的前后两段3次样条曲线段分别记为第一3次样条曲线段和第二3次样条曲线段;
    采用二分法分别计算当前点到第一3次样条曲线段和第二3次样条曲线段的最短距离;
    从当前点到第一3次样条曲线段和第二3次样条曲线段的最短距离中选出二者的相对较小值作为当前点到目标路径的最短距离;
    选取当前点到目标路径的最短距离对应的目标路径上的点的航向角作为子路径的目标航向角θ route
  8. 根据权利要求7所述的一种无人驾驶路径规划方法,其特征在于:所述计算当前点到第一3次样条曲线段和第二3次样条曲线段的最短距离这一步骤,具体包括:
    步骤一:将给定的3次样条曲线段按照给定的划分参数t划分为n段,得到n+1个第一类节点,其中,给定的3次样条曲线段为第一3次样条曲线段或第二3次样条曲线段;
    步骤二:计算当前点到n+1个第一类节点的最短距离以及最短距离对应的节点Ptemp0;
    步骤三:将最短距离对应的节点Ptemp0在n段中的相邻两段按照参数t划分为n段,得到n+1个第二类节点;
    步骤四:计算当前点到n+1个第二类节点的最短距离dtemp以及最短距离dtemp对应的节点Ptemp;
    步骤五:判断最短距离dtemp是否小于设定的最小距离阈值,若是,则以dtemp作为当前点到给定的3次样条曲线段的最短距离进行输出;反之,则令Ptemp0=Ptemp并返回步骤三。
  9. 一种无人驾驶路径规划***,其特征在于:包括:
    信息获取模块,配置成获取环境感知信息和车辆定位与导航信息,其中,环境感知信息包括障碍物信息、路沿信息和车道线信息,车辆定位与导航信息包括车辆位姿和目标路线;
    子路径生成模块,配置成根据环境感知信息和车辆定位与导航信息进行子路径生成,得到满足车辆约束的候选子路径;
    碰撞检测模块,配置成对满足车辆约束的候选子路径进行碰撞检测,得到无碰撞的候选子路径;
    子路径搜索模块,配置成采用A*搜索算法对无碰撞的候选子路径进行子路径搜索;
    局部路径生成模块,配置成根据子路径搜索的结果得到车辆的局部路径。
  10. 一种无人驾驶路径规划装置,其特征在于:包括:
    存储器,配置成存放程序;
    处理器,配置成执行所述程序以用于:
    获取环境感知信息和车辆定位与导航信息,其中,环境感知信息包括障碍物信息、路沿信息和车道线信息,车辆定位与导航信息包括车辆位姿和目标路线;
    根据环境感知信息和车辆定位与导航信息进行子路径生成,得到满足车辆约束的候选子路径;
    对满足车辆约束的候选子路径进行碰撞检测,得到无碰撞的候选子路径;
    采用A*搜索算法对无碰撞的候选子路径进行子路径搜索;
    根据子路径搜索的结果得到车辆的局部路径。
PCT/CN2018/102811 2017-08-31 2018-08-29 一种无人驾驶路径规划方法、***和装置 WO2019042295A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US16/732,231 US11460311B2 (en) 2017-08-31 2019-12-31 Path planning method, system and device for autonomous driving

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201710770252.6A CN107702716B (zh) 2017-08-31 2017-08-31 一种无人驾驶路径规划方法、***和装置
CN201710770252.6 2017-08-31

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US16/732,231 Continuation US11460311B2 (en) 2017-08-31 2019-12-31 Path planning method, system and device for autonomous driving

Publications (1)

Publication Number Publication Date
WO2019042295A1 true WO2019042295A1 (zh) 2019-03-07

Family

ID=61170418

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/102811 WO2019042295A1 (zh) 2017-08-31 2018-08-29 一种无人驾驶路径规划方法、***和装置

Country Status (3)

Country Link
US (1) US11460311B2 (zh)
CN (1) CN107702716B (zh)
WO (1) WO2019042295A1 (zh)

Cited By (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110487279A (zh) * 2019-08-27 2019-11-22 东南大学 一种基于改进a*算法的路径规划方法
CN111536987A (zh) * 2020-05-19 2020-08-14 安庆师范大学 一种用于大面积垃圾清扫的车辆路径规划方法
CN111737389A (zh) * 2020-06-12 2020-10-02 北京百度网讯科技有限公司 曲线优化方法、装置、设备和介质
CN111815160A (zh) * 2020-07-07 2020-10-23 清华大学 一种多维度节点连接评估方法
CN111898804A (zh) * 2020-07-10 2020-11-06 东风商用车有限公司 局部规划的道路自适应采样方法
CN112000090A (zh) * 2020-07-03 2020-11-27 中国北方车辆研究所 一种用于非结构化道路的轨迹队列管理方法
CN112612267A (zh) * 2020-12-08 2021-04-06 苏州挚途科技有限公司 自动驾驶的路径规划方法和装置
CN113286985A (zh) * 2020-09-17 2021-08-20 华为技术有限公司 一种路径规划方法和路径规划装置
CN113335311A (zh) * 2021-07-22 2021-09-03 中国第一汽车股份有限公司 一种车辆碰撞检测方法、装置、车辆及存储介质
CN113447022A (zh) * 2020-03-24 2021-09-28 百德机械股份有限公司 工件检测的路径规划方法及其***
CN113830079A (zh) * 2021-10-19 2021-12-24 同济大学 任意起始位姿的连续曲率泊车路径在线规划方法及***
CN113885525A (zh) * 2021-10-30 2022-01-04 重庆长安汽车股份有限公司 一种自动驾驶车辆脱困的路径规划方法、***、车辆及存储介质
CN114228748A (zh) * 2022-02-10 2022-03-25 重庆大学 一种基于几何路径生成的类人化自动驾驶轨迹规划方法

Families Citing this family (95)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107702716B (zh) * 2017-08-31 2021-04-13 广州小鹏汽车科技有限公司 一种无人驾驶路径规划方法、***和装置
JP6897597B2 (ja) * 2018-02-16 2021-06-30 トヨタ自動車株式会社 駐車支援装置
CN108519773B (zh) * 2018-03-07 2020-01-14 西安交通大学 一种结构化环境下无人驾驶车辆的路径规划方法
CN110307853B (zh) * 2018-03-27 2021-11-02 北京京东尚科信息技术有限公司 一种导航控制方法和装置
JP7046740B2 (ja) * 2018-07-02 2022-04-04 日立Astemo株式会社 予測制御装置
CN108951374B (zh) * 2018-07-10 2020-12-15 北京艾尔动力科技有限公司 碾压工程车的控制方法与装置
US10970929B2 (en) * 2018-07-16 2021-04-06 Occipital, Inc. Boundary detection using vision-based feature mapping
CN108958258B (zh) * 2018-07-25 2021-06-25 吉林大学 一种无人车的轨迹跟随控制方法、控制***及相关装置
CN108931981A (zh) * 2018-08-14 2018-12-04 汽-大众汽车有限公司 一种自动驾驶车辆的路径规划方法
CN109263639B (zh) * 2018-08-24 2020-09-01 武汉理工大学 基于状态栅格法的驾驶路径规划方法
US11181921B2 (en) * 2018-09-14 2021-11-23 Huawei Technologies Co., Ltd. System and method for hierarchical planning in autonomous vehicles
DE102018129572A1 (de) * 2018-11-23 2020-05-28 Bayerische Motoren Werke Aktiengesellschaft Fahrassistenzsystem für ein automatisiert fahrendes Fahrzeug und Verfahren zum Führen eines automatisiert fahrenden Fahrzeugs
CN109631899B (zh) * 2018-12-18 2022-05-06 仲恺农业工程学院 一种用于环境感知的移动机器人的路径优化方法及***
EP3700793A4 (en) * 2018-12-26 2020-09-09 Baidu.com Times Technology (Beijing) Co., Ltd. CORNER NEGOTIATION PROCEDURE FOR AUTONOMOUS VEHICLES WITHOUT MAP AND LOCALIZATION
CN109597317B (zh) * 2018-12-26 2022-03-18 广州小鹏汽车科技有限公司 一种基于自学习的车辆自动驾驶方法、***及电子设备
US11180160B2 (en) * 2018-12-26 2021-11-23 Baidu Usa Llc Spiral curve based vertical parking planner system for autonomous driving vehicles
CN109671286A (zh) * 2018-12-26 2019-04-23 东软睿驰汽车技术(沈阳)有限公司 一种道路信息和道路交通信息的处理方法和装置
CN109708656A (zh) * 2019-01-24 2019-05-03 爱驰汽车有限公司 基于实时路况的路线规划方法、***、设备及存储介质
CN110928290B (zh) * 2019-03-06 2023-05-23 纽劢科技(上海)有限公司 五次曲线路径规划方法及车道中心线维护方法
CN110045730B (zh) * 2019-03-20 2022-07-12 文远知行有限公司 路径规划的方法、装置、计算机设备和存储介质
CN111721311A (zh) * 2019-03-22 2020-09-29 北京京东尚科信息技术有限公司 碰撞验证方法和装置
CN110045732B (zh) * 2019-03-28 2022-06-07 暨南大学 一种启发式路径规划方法
CN109917794B (zh) * 2019-04-18 2022-02-18 北京智行者科技有限公司 全局路径规划方法及装置
CN110220528A (zh) * 2019-06-10 2019-09-10 福州大学 一种基于a星算法的自动驾驶无人车双向动态路径规划方法
CN112148811B (zh) * 2019-06-26 2023-01-10 陕西汽车集团股份有限公司 一种车载gps轨迹路径压缩方法
CN110341711B (zh) * 2019-07-06 2021-10-22 深圳数翔科技有限公司 一种基于码头环境的行驶轨迹生成***及方法
CN110262508B (zh) * 2019-07-06 2024-05-31 广州卡尔动力科技有限公司 应用于封闭场地无人驾驶货运车辆上的自动引导***及方法
CN110471412A (zh) * 2019-08-01 2019-11-19 广州小鹏汽车科技有限公司 一种评估路径跟踪算法性能的方法及评估***
CN112506176B (zh) * 2019-08-26 2024-05-28 上海汽车集团股份有限公司 路径规划方法及装置
US20210064031A1 (en) * 2019-08-28 2021-03-04 Zenuity Ab Path planning for autonomous and semi-autonomous vehicles
CN112622924B (zh) * 2019-09-24 2023-06-09 北京百度网讯科技有限公司 一种驾驶规划方法及装置、车辆
CN110703754B (zh) * 2019-10-17 2021-07-09 南京航空航天大学 一种自动驾驶车辆路径与速度高度耦合的轨迹规划方法
WO2021142799A1 (zh) * 2020-01-17 2021-07-22 华为技术有限公司 路径选择方法和路径选择装置
US11427184B2 (en) * 2020-03-11 2022-08-30 Black Sesame Technologies Inc. Breadth first search-depth first search three dimensional rapid exploring random tree search with physical constraints
CN111426330B (zh) * 2020-03-24 2022-03-15 江苏徐工工程机械研究院有限公司 路径生成方法和设备、无人化运输***和存储介质
CN111536985B (zh) * 2020-04-02 2022-07-01 东软睿驰汽车技术(沈阳)有限公司 规划路线的评价方法、装置、电子设备及存储介质
CN111591307B (zh) * 2020-04-15 2021-10-01 毫末智行科技有限公司 一种避障轨迹规划方法、***及车辆
CN112356851B (zh) * 2020-04-22 2021-11-05 青岛慧拓智能机器有限公司 一种无人驾驶矿车骑坑规划方法
CN111750886B (zh) * 2020-05-12 2022-06-28 重庆兰德适普信息科技有限公司 局部路径规划方法及装置
CN111750859B (zh) * 2020-05-29 2021-11-05 广州极飞科技股份有限公司 过渡路径规划方法及相关装置
CN111638717B (zh) * 2020-06-06 2023-11-07 浙江科钛机器人股份有限公司 一种分布式自主机器人交通协调机制的设计方法
CN111619588B (zh) * 2020-06-08 2022-02-01 三一汽车制造有限公司 行驶方向控制***、方法和无人驾驶装置
CN111694356B (zh) * 2020-06-10 2023-06-27 上海高仙自动化科技发展有限公司 一种行驶控制方法、装置、电子设备及存储介质
CN111750887B (zh) * 2020-06-11 2023-11-21 上海交通大学 降低事故严重程度的无人驾驶车辆轨迹规划方法及***
CN111879307A (zh) * 2020-06-22 2020-11-03 国网河北省电力有限公司信息通信分公司 一种基于车身参数及工程建设信息的车辆路径规划方法
CN111857148B (zh) * 2020-07-28 2022-04-29 湖南大学 一种非结构化道路车辆路径规划方法
CN112269965B (zh) * 2020-08-10 2024-04-05 中国北方车辆研究所 一种非完整约束条件下的连续曲率路径优化方法
CN112085960A (zh) * 2020-09-21 2020-12-15 北京百度网讯科技有限公司 车路协同信息处理方法、装置、设备及自动驾驶车辆
CN112158198B (zh) * 2020-09-30 2021-09-03 重庆长安汽车股份有限公司 L3级自动驾驶的横向感知安全驾驶控制方法、***及车辆
CN114312822A (zh) * 2020-09-30 2022-04-12 奥迪股份公司 自动驾驶控制方法、自动驾驶控制***、介质及车辆
CN114370874B (zh) * 2020-10-15 2023-08-25 宇通客车股份有限公司 一种车辆、车辆路径规划方法及装置
CN112526988B (zh) * 2020-10-30 2022-04-22 西安交通大学 一种自主移动机器人及其路径导航和路径规划方法、***
CN112577506B (zh) * 2020-10-30 2023-04-07 上汽大众汽车有限公司 一种自动驾驶局部路径规划方法和***
CN114435389B (zh) * 2020-11-02 2024-01-30 上海汽车集团股份有限公司 一种车辆控制方法、装置及车辆
WO2022116935A1 (zh) * 2020-12-03 2022-06-09 广州极飞科技股份有限公司 路径范围确定方法及装置、路径规划方法及装置
CN112519782A (zh) * 2020-12-08 2021-03-19 英博超算(南京)科技有限公司 车辆自动变道方法、装置、车辆以及计算机可读存储介质
CN112578796B (zh) * 2020-12-17 2022-09-06 武汉中海庭数据技术有限公司 一种基于曲率约束的引导线生成方法及装置
FR3118669A1 (fr) * 2021-01-04 2022-07-08 Psa Automobiles Sa Procédé et dispositif de détermination d’une trajectoire d’un véhicule autonome
CN112783211A (zh) * 2021-01-06 2021-05-11 中国人民解放军陆军装甲兵学院 一种基于剖分理论的无人机与地面装甲编队协同控制方法
CN112945254B (zh) * 2021-01-21 2022-08-02 西北工业大学 一种基于快速拓展随机树的无人车曲率连续路径规划方法
CN114852099A (zh) * 2021-02-03 2022-08-05 郑州宇通客车股份有限公司 机动车换道行为的预测方法
CN112964267B (zh) * 2021-02-07 2022-12-27 舜宇光学(浙江)研究院有限公司 自主避障导航方法及其***和设备
CN112880700B (zh) * 2021-02-26 2024-04-16 北京智行者科技股份有限公司 原地转向车辆的局部路径规划方法及装置
CN112964271B (zh) * 2021-03-15 2023-03-31 西安交通大学 一种面向多场景的自动驾驶规划方法及***
JP7306419B2 (ja) * 2021-03-26 2023-07-11 いすゞ自動車株式会社 運転制御装置
CN112965500B (zh) * 2021-03-27 2022-07-05 同济大学 一种过必经点集且有额外硬约束的路径规划方法及设备
CN112985445B (zh) * 2021-04-20 2021-08-13 速度时空信息科技股份有限公司 基于高精地图的车道级精度实时性运动规划方法
CN113256013B (zh) * 2021-06-02 2022-09-13 北京理工大学 一种在环境约束下的车辆路径智能搜索方法
CN113085850B (zh) * 2021-06-08 2021-09-21 新石器慧通(北京)科技有限公司 车辆避障方法、装置、电子设备及存储介质
CN113433937B (zh) * 2021-06-08 2023-05-16 杭州未名信科科技有限公司 基于启发式探索的分层导航避障***、分层导航避障方法
CN113306573B (zh) * 2021-06-23 2022-08-16 湖南大学 一种自动驾驶车辆学习型路径跟踪预测控制方法
CN113320543B (zh) * 2021-06-29 2024-03-22 东软睿驰汽车技术(沈阳)有限公司 行车方法、装置、车辆及存储介质
CN113255091B (zh) * 2021-06-29 2021-10-26 中移(上海)信息通信科技有限公司 路径拓扑关系的建立方法、装置及相关设备
CN113295169B (zh) * 2021-07-01 2023-12-15 上海仙工智能科技有限公司 一种自动行驶装置的路径规划方法
CN113479105A (zh) * 2021-07-20 2021-10-08 钟求明 一种基于自动驾驶车辆的智能充电方法及智能充电站
CN113548039B (zh) * 2021-07-29 2023-01-17 上海欧菲智能车联科技有限公司 自动泊车方法、自动泊车装置、车辆及存储介质
CN113997951B (zh) * 2021-08-17 2024-06-04 北京百度网讯科技有限公司 车辆的行驶轨迹的确定方法、装置及自动驾驶车辆
CN113741463B (zh) * 2021-09-07 2023-11-10 北京理工大学 一种无人地面机动平台的定点定向停车控制方法及***
CN113819917A (zh) * 2021-09-16 2021-12-21 广西综合交通大数据研究院 自动驾驶路径规划方法、装置、设备及存储介质
US11920944B2 (en) * 2021-10-19 2024-03-05 Capital One Services, Llc Dynamically generating scenery for a virtual reality driving session based on route information
CN113985872A (zh) * 2021-10-25 2022-01-28 广东嘉腾机器人自动化有限公司 基于视觉检测的agv取货路径规划方法、装置及介质
CN114220282B (zh) * 2021-11-15 2022-12-02 三一专用汽车有限责任公司 车辆调度控制方法、装置及电子设备
CN114194210A (zh) * 2021-11-19 2022-03-18 岚图汽车科技有限公司 一种高速公路领航***、控制方法及其设备
CN114162140B (zh) * 2021-12-08 2023-08-01 武汉中海庭数据技术有限公司 一种最优车道匹配方法及***
CN114379584B (zh) * 2021-12-17 2023-12-22 阿波罗智联(北京)科技有限公司 车辆控制方法、装置、车辆、设备和存储介质
CN114527758A (zh) * 2022-02-24 2022-05-24 北京百度网讯科技有限公司 路径规划方法及装置、设备、介质和产品
CN114594778A (zh) * 2022-03-17 2022-06-07 华侨大学 一种基于二分法和路径融合的全局避障路径规划方法
CN115311883A (zh) * 2022-08-09 2022-11-08 杭州中威电子股份有限公司 一种基于机器学习的智慧停车***及其方法
CN115588185B (zh) * 2022-11-15 2023-03-14 广汽埃安新能源汽车股份有限公司 行驶路线生成方法、装置、电子设备和计算机可读介质
CN115657685B (zh) * 2022-12-12 2023-05-12 中国南方电网有限责任公司超高压输电公司广州局 水上避障方法、装置、计算机设备和存储介质
CN115639827B (zh) * 2022-12-22 2023-03-21 华北科技学院(中国煤矿安全技术培训中心) 一种结构化环境下机器人路径规划方法和***
CN116124162B (zh) * 2022-12-28 2024-03-26 北京理工大学 一种基于高精地图的园区小车导航方法
CN116414139B (zh) * 2023-06-07 2023-08-11 成都理工大学 基于A-Star算法的移动机器人复杂路径规划方法
CN117382656B (zh) * 2023-12-11 2024-02-06 天津所托瑞安汽车科技有限公司 节油驾驶引导方法、装置、终端及存储介质
CN117705123B (zh) * 2024-02-01 2024-04-09 戴盟(深圳)机器人科技有限公司 一种轨迹规划方法、装置、设备及存储介质

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102591332A (zh) * 2011-01-13 2012-07-18 同济大学 用于无人驾驶汽车局部路径规划的装置及方法
KR20150084143A (ko) * 2014-01-13 2015-07-22 한양대학교 산학협력단 다양한 주행 환경 기반의 차량 자율 주행 장치 및 그 경로 계획 방법
CN106598055A (zh) * 2017-01-19 2017-04-26 北京智行者科技有限公司 一种智能车局部路径规划方法及其装置、车辆
CN107702716A (zh) * 2017-08-31 2018-02-16 广州小鹏汽车科技有限公司 一种无人驾驶路径规划方法、***和装置

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9734455B2 (en) * 2015-11-04 2017-08-15 Zoox, Inc. Automated extraction of semantic information to enhance incremental mapping modifications for robotic vehicles
US10223380B2 (en) * 2016-03-23 2019-03-05 Here Global B.V. Map updates from a connected vehicle fleet
US9857795B2 (en) * 2016-03-24 2018-01-02 Honda Motor Co., Ltd. System and method for trajectory planning for unexpected pedestrians
US20190086226A1 (en) * 2016-03-30 2019-03-21 Mitsubishi Electric Corporation Travel plan generation device, travel plan generation method, and computer readable recording medium

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102591332A (zh) * 2011-01-13 2012-07-18 同济大学 用于无人驾驶汽车局部路径规划的装置及方法
KR20150084143A (ko) * 2014-01-13 2015-07-22 한양대학교 산학협력단 다양한 주행 환경 기반의 차량 자율 주행 장치 및 그 경로 계획 방법
CN106598055A (zh) * 2017-01-19 2017-04-26 北京智行者科技有限公司 一种智能车局部路径规划方法及其装置、车辆
CN107702716A (zh) * 2017-08-31 2018-02-16 广州小鹏汽车科技有限公司 一种无人驾驶路径规划方法、***和装置

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
FANG, YANJUN ET AL.: "Study of the Path Planning of Unmanned Vehicles Based on the Multi-curvature Fitting-model", AUTOMATION & INSTRUMENTATION, 15 March 2012 (2012-03-15), pages 1 - 5, ISSN: 1001-9944 *
JIANG, YAN ET AL.: "System Design of self-driving in Simplified Urban Environments", JOURNAL OF MECHANICAL ENGINEERING, vol. 48, no. 20, 20 October 2012 (2012-10-20), pages 103 - 112, ISSN: 0577-6686 *
ZHANG, . XUEFEI ET AL.: "A Method of Local Path Planning and Trajectory Generation for ALV", PROCEEDINGS OF SAE- CHINA CONGRESS 2007, 30 August 2007 (2007-08-30), pages 1225 - 1228 *
ZHOU, LIANG: "The Research on Mobile Robot Path Planning in Irregular Obstacle Environment", INFORMATION SCIENCE & TECHNOLOGY, CHINA MASTER'S THESES FULL-TEXT DATABASE, 15 June 2013 (2013-06-15), ISSN: 1674-0246 *

Cited By (21)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110487279B (zh) * 2019-08-27 2022-12-13 东南大学 一种基于改进a*算法的路径规划方法
CN110487279A (zh) * 2019-08-27 2019-11-22 东南大学 一种基于改进a*算法的路径规划方法
CN113447022A (zh) * 2020-03-24 2021-09-28 百德机械股份有限公司 工件检测的路径规划方法及其***
CN111536987A (zh) * 2020-05-19 2020-08-14 安庆师范大学 一种用于大面积垃圾清扫的车辆路径规划方法
CN111536987B (zh) * 2020-05-19 2023-03-28 安庆师范大学 一种用于大面积垃圾清扫的车辆路径规划方法
CN111737389A (zh) * 2020-06-12 2020-10-02 北京百度网讯科技有限公司 曲线优化方法、装置、设备和介质
CN111737389B (zh) * 2020-06-12 2023-07-25 北京百度网讯科技有限公司 曲线优化方法、装置、设备和介质
CN112000090A (zh) * 2020-07-03 2020-11-27 中国北方车辆研究所 一种用于非结构化道路的轨迹队列管理方法
CN112000090B (zh) * 2020-07-03 2023-06-27 中国北方车辆研究所 一种用于非结构化道路的轨迹队列管理方法
CN111815160B (zh) * 2020-07-07 2022-05-24 清华大学 一种基于越野环境态势场模型的行车风险评估方法
CN111815160A (zh) * 2020-07-07 2020-10-23 清华大学 一种多维度节点连接评估方法
CN111898804B (zh) * 2020-07-10 2021-11-23 东风商用车有限公司 局部规划的道路自适应采样方法
CN111898804A (zh) * 2020-07-10 2020-11-06 东风商用车有限公司 局部规划的道路自适应采样方法
CN113286985A (zh) * 2020-09-17 2021-08-20 华为技术有限公司 一种路径规划方法和路径规划装置
CN112612267A (zh) * 2020-12-08 2021-04-06 苏州挚途科技有限公司 自动驾驶的路径规划方法和装置
CN113335311B (zh) * 2021-07-22 2022-09-23 中国第一汽车股份有限公司 一种车辆碰撞检测方法、装置、车辆及存储介质
CN113335311A (zh) * 2021-07-22 2021-09-03 中国第一汽车股份有限公司 一种车辆碰撞检测方法、装置、车辆及存储介质
CN113830079A (zh) * 2021-10-19 2021-12-24 同济大学 任意起始位姿的连续曲率泊车路径在线规划方法及***
CN113830079B (zh) * 2021-10-19 2023-09-01 同济大学 任意起始位姿的连续曲率泊车路径在线规划方法及***
CN113885525A (zh) * 2021-10-30 2022-01-04 重庆长安汽车股份有限公司 一种自动驾驶车辆脱困的路径规划方法、***、车辆及存储介质
CN114228748A (zh) * 2022-02-10 2022-03-25 重庆大学 一种基于几何路径生成的类人化自动驾驶轨迹规划方法

Also Published As

Publication number Publication date
CN107702716A (zh) 2018-02-16
US20200149906A1 (en) 2020-05-14
US11460311B2 (en) 2022-10-04
CN107702716B (zh) 2021-04-13

Similar Documents

Publication Publication Date Title
WO2019042295A1 (zh) 一种无人驾驶路径规划方法、***和装置
CN114234998B (zh) 基于语义道路地图的无人驾驶多目标点轨迹并行规划方法
CN112673234B (zh) 路径规划方法和路径规划装置
CN109976329B (zh) 一种车辆避障换道路径的规划方法
US9428187B2 (en) Lane change path planning algorithm for autonomous driving vehicle
WO2018176593A1 (zh) 一种面向无人自行车的局部避障路径规划方法
CN110361013B (zh) 一种用于车辆模型的路径规划***及方法
CN113916246A (zh) 一种无人驾驶避障路径规划方法和***
CN108088456A (zh) 一种具有时间一致性的无人驾驶车辆局部路径规划方法
US20150353082A1 (en) Unified motion planning algorithm for autonomous driving vehicle in obstacle avoidance maneuver
CN110954122B (zh) 高速场景下的自动驾驶轨迹生成方法
CN109987092A (zh) 一种车辆避障换道时机的确定方法及避障换道的控制方法
Lan et al. Continuous curvature path planning for semi-autonomous vehicle maneuvers using RRT
Zheng et al. RRT based path planning for autonomous parking of vehicle
CN109270933A (zh) 基于圆锥曲线的无人驾驶避障方法、装置、设备及介质
CN111806467A (zh) 一种基于车辆行驶规律的变速动态换道轨迹规划方法
CN113721637B (zh) 智能车动态避障路径连续规划方法、***及存储介质
US11586209B2 (en) Differential dynamic programming (DDP) based planning architecture for autonomous driving vehicles
CN111896004A (zh) 一种狭窄通道车辆轨迹规划方法及***
CN113515111B (zh) 一种车辆避障路径规划方法及装置
Zhang et al. Smooth path and velocity planning under 3D path constraints for car-like vehicles
Lattarulo et al. Real-time trajectory planning method based on n-order curve optimization
Park et al. Path generation algorithm based on crash point prediction for lane changing of autonomous vehicles
Nagasaka et al. Towards safe, smooth, and stable path planning for on-road autonomous driving under uncertainty
Daniel et al. Navigation-based constrained trajectory generation for advanced driver assistance systems

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 18850115

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 18850115

Country of ref document: EP

Kind code of ref document: A1