CN108088456A - A kind of automatic driving vehicle local paths planning method with time consistency - Google Patents

A kind of automatic driving vehicle local paths planning method with time consistency Download PDF

Info

Publication number
CN108088456A
CN108088456A CN201711397825.1A CN201711397825A CN108088456A CN 108088456 A CN108088456 A CN 108088456A CN 201711397825 A CN201711397825 A CN 201711397825A CN 108088456 A CN108088456 A CN 108088456A
Authority
CN
China
Prior art keywords
planning
curve
vehicle
point
mrow
Prior art date
Legal status (The legal status 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 status listed.)
Granted
Application number
CN201711397825.1A
Other languages
Chinese (zh)
Other versions
CN108088456B (en
Inventor
段建民
田晓生
马学峥
陈强龙
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing University of Technology
Original Assignee
Beijing University of Technology
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 Beijing University of Technology filed Critical Beijing University of Technology
Priority to CN201711397825.1A priority Critical patent/CN108088456B/en
Publication of CN108088456A publication Critical patent/CN108088456A/en
Application granted granted Critical
Publication of CN108088456B publication Critical patent/CN108088456B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/26—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network
    • G01C21/34—Route searching; Route guidance

Landscapes

  • Engineering & Computer Science (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • Automation & Control Theory (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)
  • Traffic Control Systems (AREA)

Abstract

The present invention disclose a kind of pilotless automobile local paths planning method with time consistency, is path planning curve increase vehicle movement differential constraint first so that vehicle can be exactly to the track of generation into line trace；Secondly, model parameter promotion has been carried out to path planned trajectory curve model, the degree of freedom of bigger has been added for path planning, convenient for the solution optimization problem of mathematics will be converted into the restricted problem of curvature；Finally, there is larger difference for cycle optimum results before and after obtaining local path method using cycle Optimization Solution, the problem of vehicle is caused to be shaken when carrying out real-time route tracking, local paths planning section is divided into weak planning and strong planning two parts by the present invention, weak planning section is obtained by the strong program results of last planning horizon, by the segmentation of planning, that is, ensure the time consistency of program results, and can make local paths planning that there is higher avoidance ability.

Description

A kind of automatic driving vehicle local paths planning method with time consistency
Technical field
The invention belongs to intelligent vehicle local paths planning Related Research Domains, are related to a kind of nothing with time consistency People drives vehicle local paths planning method.
Technical background
According to statistics, China's domestic automobile sales volume in 2016 has reached 2802.8 ten thousand, and 13.7% was added compared to 2015, Continue to hold a post or title global sales first within continuous 8 years.By the end of 2015 end of the years, national total mileage has also reached 457.73 ten thousand kilometers, Add 14.5 ten thousand kilometers on year-on-year basis.The increase of car ownership and total mileage of highway causes traffic accident to take place frequently, traffic congestion feelings Condition is serious.For expert to the analysis shows of urban transportation data, human factor is the main reason for causing urban traffic accident.Mirror In this, countries in the world give intelligent transportation system (Intelligent Transportation System) and fully weigh Depending on putting into substantial amounts of manpower and materials to wherein.Among this, intelligent vehicle (Intelligent Vehicles) is as intelligence The key components of traffic system play the effect that can not be ignored.So-called intelligent vehicle, exactly gather multiple technologies in (such as environment sensing, integrated navigation, artificial intelligence, machine vision, automatic technology) all over the body there is perception environment and planning to determine The intellectualizing system of plan ability.Intelligent vehicle detects vehicle-periphery information by environment sensing, and according to decision rule system Corresponding demand for control is determined in controlling, the operations such as acceleration and deceleration, steering is performed by vehicle executing agency, to reach vehicle safety row The purpose sailed is avoided because the factors such as driving technology, awareness of safety of driver cause traffic accident.
Pilotless automobile control system can be divided into planning layer and execution level by the difference of responsible function, and planning layer is main Complete the generation of destination path, the planning of planning and local path including global path.Wherein, local paths planning refers to Under global path known case, automatic driving vehicle obtains surrounding in real time running, by radar and camera includes obstacle Then it is collisionless to obtain a safety by sector planning algorithm for the environmental information of the information such as size, shape and the position of object The process of optimal path.
The method that current local paths planning algorithm is generally based on coordinates measurement and Path selection.This method has real It is strong with property, the advantages of computational complexity is low, and method is flexible, but it is not bound in curve generation the kinematics parameters pair of vehicle The curve of generation is constrained, and causes the local path that final choice goes out that cannot fundamentally ensure the accuracy of path trace, And due to the use of be cycle Optimization Solution mode, the program results of current planning horizon and next planning horizon can Time consistency can be lost there are larger difference, cause the path trace output for being likely to be obtained step, cause vehicle body not Surely, traffic safety is influenced.
The content of the invention
Time consistency is from an economic concept, and Time Inconsistency refers in t1Moment is planned to obtain tnMoment Optimal plan p, in t2Moment has not been tnIts concept is introduced by the optimal plan at moment, Beijing Institute of Technology doctor Jiang Yan Intelligent vehicle local paths planning field.
For the current more automatic driving vehicle local paths planning side based on coordinates measurement and Path selection used The curve cooked up existing for method lack vehicle kinematics constraint, the accuracy that can not ensure vehicle route tracking and security, And it obtains the poor program results that is easy to cause of sector planning Path Method cycle uniformity by Optimization Solution and step shadow occurs A kind of the problem of ringing vehicle safe driving, it is proposed that automatic driving vehicle local paths planning method with time consistency. Using the accurate performance and traveling that this method can improve the environmental adaptability of pilotless automobile, destination path tracks Security performance.
To achieve the above object, the present invention adopts the following technical scheme that：
A kind of automatic driving vehicle local paths planning method with time consistency, step include as follows：
Step 1, judge attribute planning horizon, determine the planning distance of multistage；
Step 2, automatic driving vehicle position is determined
For any one section of global path Eij, P is current vehicle position, vectorRepresent the course of vehicle, θ1Represent vehicle Course and EijAngle, θ2Represent Pj and EijAngle, set θ190 ° of ＜ and θ290 ° of ＜, calculation procedure is as follows：
1) in G (N, E), E1=argmind (P, Eij),(Eij∈E,|θ190 ° of ＜, | θ290 ° of ＜ |),
d(P,Eij)=min | PQ |, Q ∈ E }
2) in set E1In：E2=argmaxf (P, Eij),(Eij∈E1)
Wherein,
3) in search f (P, Eij) maximum side, judge | O1O2| the ρ of ＞=2minWhether satisfaction.
Wherein,If not meeting, E is checked2Next f (P, Eij), until condition satisfaction, finally, obtain The E arrivedijIt is side when search starts；
In above formula, G (N, E) be vehicle coordinate point and the confirmable two dimensional surface of global path point, wherein N pick up the car after Axis center point, E are global object path point sequence.
Step 3, local paths planning
The distance of strong, weak planning part is determined by step 1, current vehicle planning horizon phase is determined by step 2 For the reference position in global object path, local paths planning link will complete determining and advising by force for weak planning section path Determining for the starting point in path is drawn, then using curve systematic function, obtains the strong planning curve of cluster；By connecting weak rule Tracing and strong planning curve, finally obtain cluster local paths planning curve.
Step 4, Path selection
By step 3, the planning curve of cluster wheeled has been cooked up in front of driving vehicle.Step 4, according to environment sensing The Obstacle Position dimension information that system provides makes choice planning curve, obtains the local path rule of current planning horizon Check off fruit；
Step 5, sector planning path is transferred
The sector planning path curve obtained by step 4 is the point of series of discrete, the discrete point obtained to planning layer Handled, to realize the smooth docking of planning layer and key-course, using the mode of curve matching, by the use of 5 order polynomials as Matched curve, form are as follows：
Y=a0+a1x+a2x2+a3x3+a4x4+a5x5
In formula, a0,a1,a2,a3,a4,a5The respectively polynomial constant term of matched curve ordinate of orthogonal axes, first order, secondary Item, cubic term, four times to, five item coefficients to be asked；X is abscissa；Y is the ordinate value fitted.
As preferred：Local paths planning is divided into weak planning part and strong planning part, weak planning part was from last The local paths planning of planning horizon as a result, and plan that the update of path planning is responsible in part by force, by dividing strong, weak planning Section reaches keep before and after period planning uniformity ensure the real-time of path planning again, which corresponds to step 1 and is specially： In the beginning of each planning horizon, it is first determined whether current planning horizon is to plan for the first time, if first planning, then will Weak planning distance d=0, it is strong to plan distance D=V*T, plan distance R=d+D；It is planned if not first, then weak planning distance D=V2/ 2a+BV+A, it is strong to plan distance D=V*T, plan distance R=d+D.Wherein, V is the obstacle measured by vehicle body radar Object closing speed；The safe avoidance time of avoidance decision-making is carried out when T is artificial driving；A is the maximum acceleration of vehicle braking Degree；B is the delay time at stop of brake control subsystem；A is the safety stop distance of the emergency braking of setting.
Preferably, curve generating portion introduces vehicle kinematics equation in step 3, so as to get planning curve meet The ride characteristic of vehicle is handled by carrying out liter dimension to curve model, and formation curve has been carried out to include maximum curvature, entirety Smoothness, the optimization for planning length of curve, step 3 are specially：
Step 3.1, planning starting point determines for the first time
Step 3.1.1, it is first to plan that starting point determines
Weak planning distance d=0 is set, it is strong to plan distance D=VT, song is planned by force using vehicle geometric center O points as cluster The starting point of line plans the pose of the origin of curve using the pose at vehicle current time as the cluster；Wherein, V is to pass through vehicle body radar The barrier closing speed measured；The safe avoidance time of avoidance decision-making is carried out when T is artificial driving；
Step 3.1.2, it is first to plan that terminal determines
Using O points as the center of circle, drawn and justified as radius using D=VT, which can have at least one intersection point with global path, with Intersection point close to global path terminal takes aim at point O in advance as strong planning1, global path is in O1Tangential direction be O1The boat of point vehicle To along the normal direction in course according to certain lateral separation in definite several object pose points O2,O3,O4,O5, pose point Course is all consistent, and makes the pre- direction for taking aim at circumference and close terminal one side intersection point global path tangent line in global path intersection point, The intersection point both sides are distributed the distance between pose point and the quantity of pose point is determined by road environment and barrier size, by The starting point posture information of this strong path planning when can determine to plan for the first time and several terminal posture informations；
Step 3.2, non-first planning starting point determines
Step 3.2.1 is strong to plan that curve starting point determines
Using vehicle current period pose point O as the center of circle, with d=V2/ 2a+BV+A is radius, intercepts a upper planning horizon Weak program results of the local paths planning result as current planning horizon, Oq1For the planning of the last planning horizon of interception As a result, with q1The normal in point vehicle target course does straight line, plans last planning horizon set of curves in q if cutting respectively2, q3,q4,q5, planned the point of contact q on set of curves last planning horizon1,q2,q3,q4,q5Just the strong planning as current planning horizon The origin of curve；
Step 3.2.2 is strong to plan that End of Curve determines
It using vehicle current period pose point O as the center of circle, is drawn and justified as radius using R=D+d, if circumference and separate of global path The intersection point of point is Q1, with Q1The normal of tangential direction at the global path of place, along normal direction according to different lateral positions Distance sets several strong planning terminals；
Step 3.3, planning curve is solved
Step 3.3.1, programme
In first planning horizon, weak planning distance is zero, is only the current pose of vehicle into beginning-of-line pose, terminal pose For O1,O2,O3,O4,O5The corresponding pose of point, it is therefore desirable to cook up corresponding OO1,OO2,OO3,OO4,OO5Five curves；At it Planning horizon afterwards, weak program results are obtained by the part program results for intercepting the previous cycle, need not re-start curve Generation, and the strong planning region outside weak program results then needs to re-start curve generation, by strong planning starting point q1,q2,q3, q4,q5The posture information at place and strong planning terminal O1,O2,O3,O4,O5Posture information pass through Curve derivative algorithm, you can obtain q1O1,q2O2,q3O3,q4O4,q5O5Strong planning set of curves.
Step 3.3.2 plans Curve derivative algorithm
The purpose of planning curve seeks to obtain the smoothed curve of a connection planning beginning and end, which meets Point and terminal pose condition；Wherein, vehicle posture information includes the x and y-axis coordinate under global coordinate system of vehicle, vehicle Course angle θ, the front wheel slip angle δ of vehicle, it is assumed that planning starting point be A, planning terminal be B, then A points pose be expressed as (xA,yA, θA,δA), B point poses are expressed as (xB,yB,θB,δB)；
In order to which the curve cooked up is enable to be followed exactly by vehicle, traveling curve will also meet the motion of vehicle Constraint：
In above formula, x is the transverse axis coordinate of vehicle；Y is the ordinate of orthogonal axes of vehicle；θ is vehicle course angle；δ is vehicle front-wheel Drift angle；V is vehicle rear axle center spot speed；L is vehicle wheelbase.
Meeting the curve of the vehicle movement differential equation can describe to the polynomial equation of dimensionless group u：X=x (u), y=y (u), u ∈ [0,1].
The solution of 3 definite order polynomial equations can be obtained by boundary condition and the vehicle movement differential equation, so as to obtain Plan that the parametric equation form of curve is as follows：
X (u)=x0+x1u+x2u2+x3u3
Y (u)=y0+y1u+y2u2+y3u3
Parameter promotion is carried out to polynomial equation, introduces additional adjustment parameter ε1,ε2,ε3,ε4, obtain parametric equation such as Under：
X (u)=x0+x1u+x2u2+x3u3+ε1u4+ε2u5
Y (u)=y0+y1u+y2u2+y3u3+ε3u4+ε4u5
By increasing parameter, full scale equation can be converted into display equation of the coefficient to adjustment parameter, by being carried out to it Optimization Solution can obtain the parametric equation parameter for meeting constraints, consider maximum curvature, the smoothness of planning curve With the optimization purpose of length, it is as follows to set Optimization Solution function：
Wherein, K be Optimization Solution functional value, K1,K2,K3Respectively plan the power of curve maximum curvature, smoothness and length Weight coefficient, k are the curvature of planning each discrete point of curve, and s is planning length of curve.
Preferably, step 4 is specially：The formation curve that barrier collision may occur or scratch is judged, is adopted Dangerous path is judged with sliding scale,
It suppose there is AP1, AP2, AP3Three rules and regulations tracing, before carrying out judgement of slide, it is necessary first to be carried out to planning curve excellent Change sequence, then carry out judgement of slide according to sequence；
The target function that optimal sequencing is carried out to planning curve is as follows：
Wherein, K be curve optimal index value, K1,K2,K3Planning curve maximum curvature, smoothness and the length respectively obtained The weight coefficient of degree, k are the curvature of planning each discrete point of curve, and s is planning length of curve,
It, can be to cooking up the set of curves come by the optimal sequence to suboptimum, afterwards by planning curve target function When carrying out slide deciding, judged successively to suboptimum curve from optimal curve, so judge first obtained safe nothing Impact curve can be exported as last sector planning path,
If by planning that the result that curve optimal sequencing obtains is AP3Better than AP2Better than AP1, then just from AP3It carries out Slide deciding,
The geometric center of vehicle is put into AP first3First discrete point on, then judge vehicle Geometry edge with barrier Hinder the situation of object edge discrete point, if obstacles borders discrete point not in vehicle Geometry edge line, by vehicle combination center Point is put into next discrete point position, continues to judge；If there are obstacles borders discrete point in vehicle Geometry edge line if sentence It is set to uneasy full curve, continues next rules and regulations tracing judgement after deleting this curve, it is bent until obtaining a safe collisionless Line terminates；If the entire planning set of curves of traversal, can not find the collisionless wheeled curve of a safety, then send parking and refer to Order.
Compared with the conventional method, the present invention has the following advantages：
1) method for proposing local paths planning segmentation planning, preferably solves local paths planning ability and planning As a result the problem of time consistency；
2) local coordinates measurement curve is constrained using vehicle differential motion differential condition, and promoted by parameter Method improves the performance for cooking up curve；
3) carried out using the method for polynomial curve fitting discrete between local paths planning layer and path trace layer Point data is transferred, and ensure that modularization and the standardization of control system.
Description of the drawings
Fig. 1 is method flow diagram according to the present invention
The location point and tracing deviation of Fig. 2 vehicles on the digital map
Fig. 3 vehicle location schematic diagrames
Fig. 4 Rule of judgment schematic diagrames
The first local paths planning planning horizon schematic diagrames of Fig. 5
The non-first local paths planning planning horizon schematic diagrames of Fig. 6
Fig. 7 coordinates measurement algorithm simulating figures
Fig. 8 routing resource schematic diagrames
Fig. 9 local paths planning allomeric function simulation result figures
3D effect figure of Figure 10 present invention under CarSim vehicle simulated environment
Figure 11 time consistency simulated effect figures
Specific embodiment
The present invention provides a kind of automatic driving vehicle local paths planning method with time consistency, including walking as follows Suddenly：
Step 1, judge attribute planning horizon, determine the planning distance of multistage
In the beginning of each planning horizon, it is necessary first to determine whether current planning horizon is to plan for the first time, if first Secondary planning, then it is strong to plan distance D=V*T by weak planning distance d=0, plan distance R=d+D；It is planned if not first, Then weak planning distance d=V2/ 2a+BV+A, it is strong to plan distance D=V*T, plan distance R=d+D.Wherein, V is to pass through vehicle body The barrier closing speed that radar measures；The safe avoidance time of avoidance decision-making is carried out when T is artificial driving；A is vehicle The peak acceleration of braking；B is the delay time at stop of brake control subsystem；A be setting emergency braking safety stop away from From.Generally take T=3~5s.
Step 2, automatic driving vehicle position is determined
It needs to determine position of the automatic driving vehicle compared with global path point first in the starting stage of path planning.By It being constantly kept in motion in vehicle, vehicle position information is constantly to change, therefore in each carry out planning horizon office Also position of the vehicle compared with global path is redefined during portion's path planning.
Location point of the current car body in global path is obtained by vehicle posture information, as shown in Figure 2.
I and i+1 therein represent global discreet paths point sequence number, (Cx,Cy) it is the current global coordinate system coordinate of vehicle, (Lx,Ly) it is location point of the vehicle in global path, ε is tracking range deviation, and δ is heading angle deviation.
As shown in figure 3, for any one section of global path Eij, P is current vehicle position, vectorRepresent the boat of vehicle To θ1Represent vehicle course and EijAngle, θ2Represent Pj and EijAngle, with reference to vehicle actual conditions, limit θ here1＜ 90 ° and θ290 ° of ＜.Calculation procedure is as follows：
4) in G (N, E), E1=argmind (P, Eij),(Eij∈E,|θ190 ° of ＜, | θ290 ° of ＜ |), d (P, Eij)= min{|PQ|,Q∈E}
5) in set E1In：E2=argmaxf (P, Eij),(Eij∈E1)
Wherein,
6) in search f (P, Eij) maximum side, judge | O1O2| the ρ of ＞=2minWhether satisfaction.
Wherein,If not meeting, E is checked2Next f (P, Eij), until condition meet, finally, Obtained EijIt is side when search starts, the condition of judgement is as shown in Figure 4.
Step 3, local paths planning
Step 3.1, planning starting point determines for the first time
Step 3.1.1, it is first to plan that starting point determines
As shown in fig. 1, in planning local path for the first time, since there is no the connection with previous result planning horizon Sex chromosome mosaicism, therefore strong program results need to be only obtained, therefore weak planning distance d=0 is set, it is strong to plan distance D=VT.Such as Shown in Fig. 5, plan the starting point of curve by force using a few center O points of vehicle as cluster at this time, using the pose at vehicle current time as The cluster plans the pose of the origin of curve.
Step 3.1.2, it is first to plan that terminal determines
Using O points as the center of circle, drawn and justified as radius using D=VT, which can have at least one intersection point with global path, with Intersection point close to global path terminal takes aim at point O in advance as strong planning1, global path is in O1Tangential direction be O1The boat of point vehicle To along the normal direction in course according to certain lateral separation in several definite object pose points, such as O in Fig. 52,O3,O4, O5Shown in pose point, pose point course is all consistent, and circumference and close terminal one side intersection point in global path intersection point are taken aim to be pre- The direction of global path tangent line.The quantity of the distance between intersection point both sides distribution pose point and pose point is by road environment It is determined with barrier size, when barrier size is larger, horizontal pose point spacing is larger, global path both sides wheeled area When domain is wider, pose point can set more more.
The starting point posture information of strong path planning and several terminal posture informations when thus can determine to plan for the first time.
Step 3.2, non-first planning starting point determines
Step 3.2.1 is strong to plan that curve starting point determines
First planning horizon is removed, when each new planning horizon arrives, the sector planning curve vehicle of last planning horizon Sub-fraction is only travelled, most program results vehicle is not run through.Therefore, in order to improve program results Time consistency, improve the utilization rate of program results, reduce algorithm consumption, using vehicle current period pose point O as the center of circle, With d=V2/ 2a+BV+A is radius, intercepts the local paths planning result of a upper planning horizon as current planning horizon Weak program results.Concrete operations are as shown in fig. 6, Oq1For the program results of the last planning horizon of interception, with q1Point vehicle target The normal in course does straight line, plans last planning horizon set of curves in q if cutting respectively2,q3,q4,q5, last planning horizon Plan the point of contact q on set of curves1,q2,q3,q4,q5Just the strong planning origin of curve as current planning horizon.
Step 3.2.2 is strong to plan that End of Curve determines
It using vehicle current period pose point O as the center of circle, is drawn and justified as radius using R=D+d, if circumference and separate of global path The intersection point of point is Q1, with Q1The normal of tangential direction at the global path of place, along normal direction according to different lateral positions Distance sets several strong planning terminals, such as O in Fig. 61,O2,O3,O4,O5It is shown.
Step 3.3, planning curve is solved
Step 3.3.1, programme
As shown in figure 5, in first planning horizon, weak planning distance is zero, is only vehicle present bit into beginning-of-line pose Appearance, terminal pose are O1,O2,O3,O4,O5The corresponding pose of point.Therefore need to cook up corresponding OO1,OO2,OO3,OO4,OO5Five Curve；As shown in fig. 6, in planning horizon afterwards, weak program results is obtained by the part program results for intercepting the previous cycle It arrives, curve generation need not be re-started, and the strong planning region outside weak program results then needs to re-start curve generation, By strong planning starting point q1,q2,q3,q4,q5The posture information at place and strong planning terminal O1,O2,O3,O4,O5Posture information lead to Cross Curve derivative algorithm, you can obtain q1O1,q2O2,q3O3,q4O4,q5O5Strong planning set of curves.
Step 3.3.2 plans Curve derivative algorithm
The purpose of planning curve seeks to obtain the smoothed curve of a connection planning beginning and end, which meets Point and terminal pose condition.Wherein, vehicle posture information includes the x and y-axis coordinate under global coordinate system of vehicle, vehicle Course angle θ, the front wheel slip angle δ of vehicle.Assuming that planning starting point is A, planning terminal is B, then A points pose is expressed as (xA,yA, θA,δA), B point poses are expressed as (xB,yB,θB,δB)。
In order to which the curve cooked up is enable to be followed exactly by vehicle, traveling curve will also meet the motion of vehicle Constraint：
Meeting the curve of the vehicle movement differential equation can describe to the polynomial equation of dimensionless group u：X=x (u), y=y (u), u ∈ [0,1].
The solution of 3 definite order polynomial equations can be obtained by boundary condition and the vehicle movement differential equation.So as to obtain Plan that the parametric equation form of curve is as follows：
X (u)=x0+x1u+x2u2+x3u3
Y (u)=y0+y1u+y2u2+y3u3
However, 3 order polynomial equations that the parameter so obtained uniquely determines, which can not well solve, plans that curve is excellent The problem of change, can not be constrained planning maximum curvature, smoothness and the length of curve.Therefore, here to polynomial equation Parameter promotion is carried out, introduces additional adjustment parameter ε1,ε2,ε3,ε4, it is as follows to obtain parametric equation：
X (u)=x0+x1u+x2u2+x3u3+ε1u4+ε2u5
Y (u)=y0+y1u+y2u2+y3u3+ε3u4+ε4u5
By increasing parameter, full scale equation can be converted into display equation of the coefficient to adjustment parameter, by being carried out to it Optimization Solution can obtain the parametric equation parameter for meeting constraints, consider maximum curvature, the smoothness of planning curve With the optimization purpose of length, it is as follows that Optimization Solution function is set here：
Wherein, K be Optimization Solution functional value, K1,K2,K3Respectively plan the power of curve maximum curvature, smoothness and length Weight coefficient, k are the curvature of planning each discrete point of curve, and s is planning length of curve.
Under simulated environment, curve generation effect is as shown in Figure 7.
Step 4, Path selection
As shown in fig. 7, by step 3, the planning curve of cluster wheeled has been cooked up in driving vehicle front, in order to ensure The sector planning path vehicle finally obtained can reach the collisionless purpose of safety in motion, it is also necessary to according to environment sense Know that the Obstacle Position dimension information that system provides makes choice planning curve, so as to obtain the part of current planning horizon Route programming result.Step 4.1, dangerous path judges
Local paths planning algorithm based on grid often relies on the method for corrosion expansion to barrier collision may occur Or the formation curve scratched is judged, the present invention judges dangerous path using sliding scale.
As shown in figure 8, it is assumed that there is AP1, AP2, AP3Three rules and regulations tracing, before carrying out judgement of slide, it is necessary first to right Planning curve optimizes sequence, then carries out judgement of slide according to sequence.
Step 4.1.1 plans curve optimal sequencing
Different from the Moving Objects that robot etc. slowly moves, vehicle has the attribute of high-speed motion, it is therefore necessary to consider The computation complexity of planning algorithm is reduced as far as possible.Therefore curve optimal sequencing must be carried out before judging planning curve, So after judging to obtain optimal planning curve, without judging again remaining planning curve, local path is improved Planning efficiency.
The target function that optimal sequencing is carried out to planning curve is as follows：
Wherein, K be curve optimal index value, K1,K2,K3Planning curve maximum curvature, smoothness and the length respectively obtained The weight coefficient of degree, k are the curvature of planning each discrete point of curve, and s is planning length of curve.
The corresponding optimal index value of curve is smaller, represents that the performance of curve is higher.
It, can be to cooking up the set of curves come by the optimal sequence to suboptimum, afterwards by planning curve target function When carrying out slide deciding, judged successively to suboptimum curve from optimal curve, so judge first obtained safe nothing Impact curve can be exported as last sector planning path, reduce the number for judging operation.
Step 4.1.2, slide deciding
If as shown in figure 8, by planning that the result that curve optimal sequencing obtains is AP3Better than AP2Better than AP1, then just From AP3Carry out slide deciding.
The geometric center of vehicle is put into AP first3First discrete point on, then judge vehicle Geometry edge with barrier Hinder the situation of object edge discrete point, if obstacles borders discrete point not in vehicle Geometry edge line, by vehicle combination center Point is put into next discrete point position, continues to judge；If there are obstacles borders discrete point in vehicle Geometry edge line if sentence It is set to uneasy full curve, continues next rules and regulations tracing judgement after deleting this curve, it is bent until obtaining a safe collisionless Line terminates；If the entire planning set of curves of traversal, can not find the collisionless wheeled curve of a safety, then send parking and refer to Order.
In order to reduce planning algorithm computation complexity, vehicle geometric shape is set as with vehicle geometric center here For the center of circle, using vehicle diagonal distance half as the circumference of radius.
Step 5, sector planning path is transferred
By step 4, obtained sector planning path curve is the point of series of discrete, when planning distance is longer, rule Tracing discrete point quantity is more considerable.If these discrete points are transferred directly to key-course, can cause to occupy a large amount of Passing interface, be unfavorable for the Standardized Design of control system.In addition, planning horizon and controlling cycle are not consistent, control Preparative layer carries out control using the discrete point of planning layer can not obtain optimal effect.It is therefore desirable to what planning layer was obtained Discrete point is handled, to realize the smooth docking of planning layer and key-course.Used here as the mode of curve matching, utilize 5 times For multinomial as matched curve, form is as follows：
Y=a0+a1x+a2x2+a3x3+a4x4+a5x5
In formula, a=[a0,a1,a2,a3,a4,a5] it is parameter to be asked.
Passing through the method for fitting, it would be desirable to the discrete programming path point of transmission is changed into the problem of transferring polynomial parameters, Reduce the quantity of passing interface between planning module and control module.
Fig. 9 is the analogue system that this method is built in Matlab/Simulink and vehicle dynamics simulation software Carsim The situation of middle operation.Wherein, transverse axis be the x directions of motion, the longitudinal axis be the y directions of motion, O points be vehicle starting point, P1P2It is red bent Line is Global motion planning path, and blue curve is the local paths planning of each planning horizon as a result, green is vehicle actual travel Track, ABCD are the barrier set.As shown in Figure 9, in start time, vehicle deviates global path, therefore local path is advised Draw unit guiding vehicle tracking global path；When an obstacle is detected, the safe collisionless of local paths planning unit planning department Local path, get around barrier；When vehicle is after barrier, it is complete that local paths planning unit guiding vehicle continues tracking Office path.
The 3D effect figure that Figure 10 realizes for institute's extracting method of the present invention under CarSim vehicle simulated environment.Automatic driving vehicle The destination path of script is the straight line positioned at right-hand lane, but traveling process always encounters one and is parked among road Disabled vehicle, the local paths planning unit of automatic driving vehicle utilize institute's extracting method of the present invention, have cooked up safe collisionless Wheeled path, as seen from Figure 10, vehicle can cross safely barrier and continue to track global path without collision.
Figure 11 is time consistency simulated effect figure in method used herein, wherein transverse axis cycle in order to control, each Controlling cycle T=0.05s, the longitudinal axis are the steering wheel angle value resolved.As shown in Figure 11, initial time vehicle deviates mesh Path is marked, two kinds of paths planning methods can adjust vehicle tracking destination path, and without the vehicle of time consistency planing method Start to detect barrier in or so 140 planning horizons, although be made that can with the planning of avoiding barrier, due to The local path curve and its front and rear period planning result that each period planning comes out do not have time consistency, cause entirely to keep away During barrier, the steering wheel angle value of output is controlled acute variation occur and cause that vehicle body is caused to shake.And it uses and has been carried herein The avoidance planning process of time consistency method due to each adjacent planning horizon program results have time consistency, because The steering wheel angle value exported during this final path trace has flatness, ensure that the stability of automatic driving vehicle traveling And security.
Complex chart 9, Figure 10 and Figure 11 understand that pilotless automobile local paths planning method proposed by the invention has Preferable path planning effects and higher algorithm feasibility.
The pilotless automobile local paths planning method with time consistency that the present invention is carried is first planning road Diametal curve increases vehicle movement differential constraint so that vehicle can be exactly to the track of generation into line trace；Next, in order to It avoids the path curve curvature that vehicle is cooked up when driving in fair speed excessive, path planned trajectory curve model is carried out Model parameter promotion adds the degree of freedom of bigger for path planning, convenient for will be converted into mathematics to the restricted problem of curvature Solution optimization problem；Have finally, for cycle optimum results before and after obtaining local path method using cycle Optimization Solution Larger difference, the problem of causing to shake during vehicle progress real-time route tracking, local paths planning section is divided by the present invention Weak planning and strong planning two parts, weak planning section are obtained by the strong program results of last planning horizon, by the segmentation of planning, Ensure the time consistency of program results, and can make local paths planning that there is higher avoidance ability.

Claims (4)

1. a kind of automatic driving vehicle local paths planning method with time consistency, it is characterised in that：Its step includes It is as follows：
Step 1, judge attribute planning horizon, determine the planning distance of multistage；
Step 2, automatic driving vehicle position is determined
For any one section of global path Eij, P is current vehicle position, vectorRepresent the course of vehicle, θ1Represent vehicle boat To with EijAngle, θ2Represent Pj and EijAngle, set θ190 ° of ＜ and θ290 ° of ＜, calculation procedure is as follows：
1) in G (N, E), E1=argmind (P, Eij),(Eij∈E,|θ190 ° of ＜, | θ290 ° of ＜ |),
d(P,Eij)=min | PQ |, Q ∈ E }
2) in set E1In：E2=argmaxf (P, Eij),(Eij∈E1)
Wherein,
3) in search f (P, Eij) maximum side, judge | O1O2| the ρ of ＞=2minWhether satisfaction.
Wherein,If not meeting, E is checked2Next f (P, Eij), until condition satisfaction, finally, obtain EijIt is side when search starts；
In above formula, G (N, E) is taken for vehicle coordinate point and the confirmable two dimensional surface of global path point, wherein N in vehicle rear axle Heart point, E are global object path point sequence.
Step 3, local paths planning
By step 1 determine it is strong, it is weak planning part distance, by step 2 determine current vehicle planning horizon compared with The reference position in global object path, local paths planning link will complete the definite of weak planning section path and plan road by force The starting point in footpath determines, then using curve systematic function, obtains the strong planning curve of cluster；By connecting weak planning curve With strong planning curve, cluster local paths planning curve is finally obtained.
Step 4, Path selection
By step 3, the planning curve of cluster wheeled has been cooked up in front of driving vehicle.Step 4, according to context aware systems The Obstacle Position dimension information provided makes choice planning curve, obtains the local paths planning knot of current planning horizon Fruit；
Step 5, sector planning path is transferred
The sector planning path curve obtained by step 4 is the point of series of discrete, and the discrete point that planning layer obtains is carried out Processing, to realize the smooth docking of planning layer and key-course, using the mode of curve matching, by the use of 5 order polynomials as being fitted Curve, form are as follows：
Y=a0+a1x+a2x2+a3x3+a4x4+a5x5
In formula, a0,a1,a2,a3,a4,a5The respectively polynomial constant term of matched curve ordinate of orthogonal axes, first order, quadratic term, three Secondary item, four times to, five item coefficients to be asked；X is abscissa；Y is the ordinate value fitted.
2. a kind of automatic driving vehicle local paths planning method with time consistency as described in claim 1, special Sign is：Local paths planning is divided into weak planning part and strong planning part, weak planning part is from last planning horizon Local paths planning as a result, and plan that the update of path planning is responsible in part by force, reach and keep by dividing strong, weak planning section Front and rear period planning uniformity ensures the real-time of path planning again, which corresponds to step 1 and be specially：In each planning horizon Beginning, it is first determined whether current planning horizon is to plan for the first time, if first planning, then by weak planning distance d=0, Strong planning distance D=V*T, plans distance R=d+D；It is if not first planning, then weak to plan distance d=V2/ 2a+BV+A, by force It plans distance D=V*T, plans distance R=d+D.Wherein, V is the barrier closing speed measured by vehicle body radar；T is people To carry out the safe avoidance time of avoidance decision-making during driving；A is the peak acceleration of vehicle braking；B is brake control The delay time at stop of system；A is the safety stop distance of the emergency braking of setting.
3. a kind of automatic driving vehicle local paths planning method with time consistency as described in claim 1, special Sign is：Curve generating portion introduces vehicle kinematics equation in step 3, so as to get planning curve meet the traveling of vehicle Characteristic is handled by carrying out liter dimension to curve model, and formation curve has been carried out to include maximum curvature, whole smoothness, planning The optimization of length of curve, step 3 are specially：
Step 3.1, planning starting point determines for the first time
Step 3.1.1, it is first to plan that starting point determines
Weak planning distance d=0 is set, it is strong to plan distance D=VT, curve is planned by force using vehicle geometric center O points as cluster Starting point plans the pose of the origin of curve using the pose at vehicle current time as the cluster；Wherein, V is to be measured by vehicle body radar Barrier closing speed；The safe avoidance time of avoidance decision-making is carried out when T is artificial driving；
Step 3.1.2, it is first to plan that terminal determines
Using O points as the center of circle, drawn and justified as radius using D=VT, which there can be at least one intersection point with global path, with close The intersection point of global path terminal takes aim at point O in advance as strong planning1, global path is in O1Tangential direction be O1The course of point vehicle, edge The normal direction in course is according to certain lateral separation in definite several object pose points O2,O3,O4,O5, pose point course is all It is consistent, makes the pre- direction for taking aim at circumference and close terminal one side intersection point global path tangent line in global path intersection point, the intersection point The quantity of the distance between both sides distribution pose point and pose point is determined by road environment and barrier size, thus can Determine the starting point posture information and several terminal posture informations of strong path planning during first planning；
Step 3.2, non-first planning starting point determines
Step 3.2.1 is strong to plan that curve starting point determines
Using vehicle current period pose point O as the center of circle, with d=V2/ 2a+BV+A is radius, intercepts the part of a upper planning horizon Weak program results of the route programming result as current planning horizon, Oq1For the program results of the last planning horizon of interception, with q1The normal in point vehicle target course does straight line, plans last planning horizon set of curves in q if cutting respectively2,q3,q4,q5, on Point of contact q on a planning planning horizon set of curves1,q2,q3,q4,q5Just the strong planning origin of curve as current planning horizon；
Step 3.2.2 is strong to plan that End of Curve determines
It using vehicle current period pose point O as the center of circle, is drawn and justified as radius using R=D+d, if circumference and global path are away from starting point Intersection point is Q1, with Q1The normal of tangential direction at the global path of place is set along normal direction according to different lateral position distances Put several strong planning terminals；
Step 3.3, planning curve is solved
Step 3.3.1, programme
In first planning horizon, weak planning distance is zero, is only the current pose of vehicle into beginning-of-line pose, terminal pose is O1, O2,O3,O4,O5The corresponding pose of point, it is therefore desirable to cook up corresponding OO1,OO2,OO3,OO4,OO5Five curves；In rule afterwards It draws the cycle, weak program results is obtained by the part program results for intercepting the previous cycle, need not re-start curve generation, and Strong planning region outside weak program results then needs to re-start curve generation, by strong planning starting point q1,q2,q3,q4,q5Place Posture information and strong planning terminal O1,O2,O3,O4,O5Posture information pass through Curve derivative algorithm, you can obtain q1O1,q2O2, q3O3,q4O4,q5O5Strong planning set of curves.
Step 3.3.2 plans Curve derivative algorithm
The purpose of planning curve seeks to obtain the smoothed curve of a connection planning beginning and end, the curve meet starting point and Terminal pose condition；Wherein, vehicle posture information includes the x and y-axis coordinate under global coordinate system of vehicle, the course of vehicle Angle θ, the front wheel slip angle δ of vehicle, it is assumed that planning starting point is A, and planning terminal is B, then A points pose is expressed as (xA,yA,θA,δA),B Point pose is expressed as (xB,yB,θB,δB)；
In order to which the curve cooked up is enable to be followed exactly by vehicle, traveling curve will also meet the motion of vehicle about Beam：
<mrow> <mover> <mi>x</mi> <mo>&amp;CenterDot;</mo> </mover> <mo>=</mo> <mi>v</mi> <mo>&amp;times;</mo> <mi>c</mi> <mi>o</mi> <mi>s</mi> <mrow> <mo>(</mo> <mi>&amp;theta;</mi> <mo>)</mo> </mrow> </mrow>
<mrow> <mover> <mi>y</mi> <mo>&amp;CenterDot;</mo> </mover> <mo>=</mo> <mi>v</mi> <mo>&amp;times;</mo> <mi>s</mi> <mi>i</mi> <mi>n</mi> <mrow> <mo>(</mo> <mi>&amp;theta;</mi> <mo>)</mo> </mrow> </mrow>
<mrow> <mover> <mi>&amp;theta;</mi> <mo>&amp;CenterDot;</mo> </mover> <mo>=</mo> <mi>v</mi> <mo>&amp;times;</mo> <mfrac> <mrow> <mi>t</mi> <mi>a</mi> <mi>n</mi> <mrow> <mo>(</mo> <mi>&amp;delta;</mi> <mo>)</mo> </mrow> </mrow> <mi>L</mi> </mfrac> </mrow>
In above formula, x is the transverse axis coordinate of vehicle；Y is the ordinate of orthogonal axes of vehicle；θ is vehicle course angle；δ is vehicle front wheel slip angle； V is vehicle rear axle center spot speed；L is vehicle wheelbase.
Meeting the curve of the vehicle movement differential equation can describe to the polynomial equation of dimensionless group u：X=x (u), y =y (u), u ∈ [0,1].
The solution of 3 definite order polynomial equations can be obtained by boundary condition and the vehicle movement differential equation, so as to be planned The parametric equation form of curve is as follows：
X (u)=x0+x1u+x2u2+x3u3
Y (u)=y0+y1u+y2u2+y3u3
Parameter promotion is carried out to polynomial equation, introduces additional adjustment parameter ε1,ε2,ε3,ε4, it is as follows to obtain parametric equation：
X (u)=x0+x1u+x2u2+x3u3+ε1u4+ε2u5
Y (u)=y0+y1u+y2u2+y3u3+ε3u4+ε4u5
By increasing parameter, full scale equation can be converted into display equation of the coefficient to adjustment parameter, by being optimized to it Solution can obtain the parametric equation parameter for meeting constraints, consider maximum curvature, smoothness and the length of planning curve The optimization purpose of degree, setting Optimization Solution function are as follows：
<mrow> <mi>K</mi> <mo>=</mo> <msub> <mi>K</mi> <mn>1</mn> </msub> <mi>m</mi> <mi>a</mi> <mi>x</mi> <mrow> <mo>(</mo> <mi>k</mi> <mo>)</mo> </mrow> <mo>+</mo> <msub> <mi>K</mi> <mn>2</mn> </msub> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>i</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <msub> <mi>k</mi> <mi>i</mi> </msub> <mo>+</mo> <msub> <mi>K</mi> <mn>3</mn> </msub> <mi>s</mi> </mrow>
Wherein, K be Optimization Solution functional value, K1,K2,K3Respectively plan the weight system of curve maximum curvature, smoothness and length Number, k are the curvature of planning each discrete point of curve, and s is planning length of curve.
4. a kind of automatic driving vehicle local paths planning method with time consistency as described in claim 1, special Sign is：Step 4 is specially：The formation curve that barrier collision may occur or scratch is judged, using sliding scale pair Dangerous path judged,
It suppose there is AP1, AP2, AP3Three rules and regulations tracing, before carrying out judgement of slide, it is necessary first to which row is optimized to planning curve Then sequence carries out judgement of slide according to sequence；
The target function that optimal sequencing is carried out to planning curve is as follows：
<mrow> <mi>K</mi> <mo>=</mo> <msub> <mi>K</mi> <mn>1</mn> </msub> <mi>m</mi> <mi>a</mi> <mi>x</mi> <mrow> <mo>(</mo> <mi>k</mi> <mo>)</mo> </mrow> <mo>+</mo> <msub> <mi>K</mi> <mn>2</mn> </msub> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>i</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <msub> <mi>k</mi> <mi>i</mi> </msub> <mo>+</mo> <msub> <mi>K</mi> <mn>3</mn> </msub> <mi>s</mi> </mrow>
Wherein, K be curve optimal index value, K1,K2,K3Planning curve maximum curvature, smoothness and the length respectively obtained Weight coefficient, k are the curvature of planning each discrete point of curve, and s is planning length of curve,
By planning curve target function, can be carried out to cooking up the set of curves come by the optimal sequence to suboptimum afterwards During slide deciding, judged successively to suboptimum curve from optimal curve, so judge first obtained safe collisionless Curve can be exported as last sector planning path,
If by planning that the result that curve optimal sequencing obtains is AP3Better than AP2Better than AP1, then just from AP3Slide and sentence It is fixed,
The geometric center of vehicle is put into AP first3First discrete point on, then judge vehicle Geometry edge and barrier side Along the situation of discrete point, if vehicle combination central point is put by obstacles borders discrete point not in vehicle Geometry edge line Next discrete point position, continues to judge；If there are obstacles borders discrete point in vehicle Geometry edge line if be determined as not Safety curve continues next rules and regulations tracing judgement after deleting this curve, is terminated until obtaining a safe collisionless curve； If the entire planning set of curves of traversal, can not find the collisionless wheeled curve of a safety, then send cutoff command.
CN201711397825.1A 2017-12-21 2017-12-21 Unmanned vehicle local path planning method with time consistency Active CN108088456B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201711397825.1A CN108088456B (en) 2017-12-21 2017-12-21 Unmanned vehicle local path planning method with time consistency

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201711397825.1A CN108088456B (en) 2017-12-21 2017-12-21 Unmanned vehicle local path planning method with time consistency

Publications (2)

Publication Number Publication Date
CN108088456A true CN108088456A (en) 2018-05-29
CN108088456B CN108088456B (en) 2021-07-16

Family

ID=62178072

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201711397825.1A Active CN108088456B (en) 2017-12-21 2017-12-21 Unmanned vehicle local path planning method with time consistency

Country Status (1)

Country Link
CN (1) CN108088456B (en)

Cited By (37)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108897319A (en) * 2018-06-23 2018-11-27 广州市柯乐名迪电子科技有限公司 Automatic Pilot method, apparatus and system based on digital geometry figure
CN109556623A (en) * 2018-12-04 2019-04-02 中国兵器装备集团自动化研究所 Fusion antenna algorithm and the iterative path planning algorithm for cutting line-plot method
CN109557912A (en) * 2018-10-11 2019-04-02 同济大学 A kind of decision rule method of automatic Pilot job that requires special skills vehicle
CN109733411A (en) * 2019-02-20 2019-05-10 苏州风图智能科技有限公司 A kind of method for controlling driving speed and device
CN109855637A (en) * 2018-12-24 2019-06-07 北京新能源汽车股份有限公司 A kind of automatic Pilot paths planning method, device and the equipment of vehicle
CN110262488A (en) * 2019-06-18 2019-09-20 重庆长安汽车股份有限公司 Local paths planning method, system and the computer readable storage medium of automatic Pilot
CN110928290A (en) * 2019-03-06 2020-03-27 纽劢科技(上海)有限公司 Quintic curve path planning method and lane center line maintenance method
CN111006667A (en) * 2019-12-09 2020-04-14 东风商用车有限公司 Automatic driving track generation system under high-speed scene
CN111174797A (en) * 2020-01-16 2020-05-19 湖南大学 Closed area global path planning method
CN111289008A (en) * 2020-04-28 2020-06-16 南京维思科汽车科技有限公司 Local path planning algorithm for unmanned vehicle
CN111615618A (en) * 2018-12-26 2020-09-01 百度时代网络技术（北京）有限公司 Polynomial fitting based reference line smoothing method for high speed planning of autonomous vehicles
CN111679667A (en) * 2020-05-20 2020-09-18 东南大学 Path and vehicle speed collaborative planning method for unmanned racing vehicle
CN111721309A (en) * 2019-03-19 2020-09-29 上海汽车集团股份有限公司 Path planning method and device
WO2020191712A1 (en) * 2019-03-28 2020-10-01 Baidu.Com Times Technology (Beijing) Co., Ltd. A map-less and camera-based lane markings sampling method for level-3 autonomous driving vehicles
CN111998864A (en) * 2020-08-11 2020-11-27 东风柳州汽车有限公司 Unmanned vehicle local path planning method, device, equipment and storage medium
CN112082557A (en) * 2020-09-14 2020-12-15 哈尔滨工程大学 UUV submarine topography tracking path rolling generation method based on Bessel fitting
CN112092810A (en) * 2020-09-24 2020-12-18 上海汽车集团股份有限公司 Vehicle parking-out method and device and electronic equipment
CN112254727A (en) * 2020-09-23 2021-01-22 锐捷网络股份有限公司 TEB-based path planning method and device
CN112504288A (en) * 2020-07-22 2021-03-16 北京智行者科技有限公司 Local path planning method based on dynamic planning
CN112540602A (en) * 2019-11-27 2021-03-23 深圳优地科技有限公司 Robot local path optimization method and robot
CN112666945A (en) * 2020-12-18 2021-04-16 广东嘉腾机器人自动化有限公司 AGV curve path optimization method
CN112783144A (en) * 2019-10-22 2021-05-11 舜宇光学（浙江）研究院有限公司 Path generation method, path planning method, system and equipment
CN112859864A (en) * 2021-01-15 2021-05-28 大连海事大学 Unmanned ship-oriented geometric path planning method
CN112880700A (en) * 2021-02-26 2021-06-01 重庆智行者信息科技有限公司 Local path planning method and device for pivot steering vehicle
CN113286732A (en) * 2018-12-17 2021-08-20 Zf汽车德国有限公司 Control system and control method for hybrid method for determining possible trajectories of a motor vehicle
CN113324554A (en) * 2021-05-28 2021-08-31 江铃汽车股份有限公司 Automatic driving route planning method and device, readable storage medium and electronic equipment
CN113406960A (en) * 2021-07-08 2021-09-17 浙江大学 Real-time path planning and control method for ground steering of agricultural unmanned vehicle
CN113515111A (en) * 2020-03-25 2021-10-19 郑州宇通客车股份有限公司 Vehicle obstacle avoidance path planning method and device
CN113825978A (en) * 2019-04-12 2021-12-21 赛峰电子与防务公司 Method for defining a path
CN114185352A (en) * 2021-12-08 2022-03-15 东风悦享科技有限公司 High-precision high-real-time automatic driving local path planning method
CN114237229A (en) * 2021-11-26 2022-03-25 青岛德智汽车科技有限公司 Unstructured road operation vehicle path planning method based on empirical path fitting
CN114310941A (en) * 2021-12-21 2022-04-12 哈尔滨工业大学芜湖机器人产业技术研究院 Robot path generation method for hub wheel hole deburring machining
CN114370874A (en) * 2020-10-15 2022-04-19 郑州宇通客车股份有限公司 Vehicle, and vehicle path planning method and device
CN114459486A (en) * 2022-02-24 2022-05-10 深圳市优必选科技股份有限公司 Robot and channel navigation method, device and storage medium thereof
CN114995465A (en) * 2022-08-02 2022-09-02 北京理工大学 Multi-unmanned vehicle motion planning method and system considering vehicle motion capability
US11648936B2 (en) 2019-10-09 2023-05-16 Apollo Intelligent Driving Technology (Beiiing) Co., Ltd. Method and apparatus for controlling vehicle
CN116663939A (en) * 2023-07-31 2023-08-29 北京理工大学 Unmanned vehicle path planning scene and task complexity evaluation method and system

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102047076A (en) * 2008-05-30 2011-05-04 阿特拉斯·科普柯凿岩设备有限公司 Method and arrangement for calculating a conformity between a representation of an environment and said environment
US20130090823A1 (en) * 2009-08-18 2013-04-11 Palo Alto Research Center Incorporated Model based method to assess road curvature effect on travel time and comfort for route planning
CN107121980A (en) * 2017-03-17 2017-09-01 北京理工大学 A kind of automatic driving vehicle paths planning method based on virtual constraint
CN107246876A (en) * 2017-07-31 2017-10-13 中北智杰科技（北京）有限公司 A kind of method and system of pilotless automobile autonomous positioning and map structuring

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102047076A (en) * 2008-05-30 2011-05-04 阿特拉斯·科普柯凿岩设备有限公司 Method and arrangement for calculating a conformity between a representation of an environment and said environment
US20130090823A1 (en) * 2009-08-18 2013-04-11 Palo Alto Research Center Incorporated Model based method to assess road curvature effect on travel time and comfort for route planning
CN107121980A (en) * 2017-03-17 2017-09-01 北京理工大学 A kind of automatic driving vehicle paths planning method based on virtual constraint
CN107246876A (en) * 2017-07-31 2017-10-13 中北智杰科技（北京）有限公司 A kind of method and system of pilotless automobile autonomous positioning and map structuring

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
KUHNE F ET AL.: "Model Predictive Control of a Mobile Robot Using Linearization", 《PRODEEDINGS OF MECHATRONICS AND ROBOTOTIC》 *
姜岩等: "无人驾驶车辆局部路径规划的时间一致性与鲁棒性研究", 《自动化学报》 *

Cited By (61)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108897319A (en) * 2018-06-23 2018-11-27 广州市柯乐名迪电子科技有限公司 Automatic Pilot method, apparatus and system based on digital geometry figure
CN109557912B (en) * 2018-10-11 2020-07-28 同济大学 Decision planning method for automatically driving special operation vehicle
CN109557912A (en) * 2018-10-11 2019-04-02 同济大学 A kind of decision rule method of automatic Pilot job that requires special skills vehicle
CN109556623A (en) * 2018-12-04 2019-04-02 中国兵器装备集团自动化研究所 Fusion antenna algorithm and the iterative path planning algorithm for cutting line-plot method
CN109556623B (en) * 2018-12-04 2022-07-29 中国兵器装备集团自动化研究所 Iterative path planning algorithm fusing whisker algorithm and tangent graph method
CN113286732A (en) * 2018-12-17 2021-08-20 Zf汽车德国有限公司 Control system and control method for hybrid method for determining possible trajectories of a motor vehicle
CN109855637A (en) * 2018-12-24 2019-06-07 北京新能源汽车股份有限公司 A kind of automatic Pilot paths planning method, device and the equipment of vehicle
CN111615618B (en) * 2018-12-26 2023-08-29 百度时代网络技术（北京）有限公司 Polynomial fitting-based reference line smoothing method for high-speed planning of autonomous vehicles
CN111615618A (en) * 2018-12-26 2020-09-01 百度时代网络技术（北京）有限公司 Polynomial fitting based reference line smoothing method for high speed planning of autonomous vehicles
US11993259B2 (en) 2019-02-20 2024-05-28 Venti Technologies Corporation Vehicle speed control method and apparatus
CN109733411B (en) * 2019-02-20 2020-10-13 苏州风图智能科技有限公司 Vehicle speed control method and device
CN109733411A (en) * 2019-02-20 2019-05-10 苏州风图智能科技有限公司 A kind of method for controlling driving speed and device
CN110928290B (en) * 2019-03-06 2023-05-23 纽劢科技(上海)有限公司 Quintic curve path planning method and lane center line maintenance method
CN110928290A (en) * 2019-03-06 2020-03-27 纽劢科技(上海)有限公司 Quintic curve path planning method and lane center line maintenance method
CN111721309A (en) * 2019-03-19 2020-09-29 上海汽车集团股份有限公司 Path planning method and device
CN112041637B (en) * 2019-03-28 2021-11-19 百度时代网络技术（北京）有限公司 Map-less and camera-based lane marker sampling method for 3-level autonomous vehicles
US11267476B2 (en) 2019-03-28 2022-03-08 Baidu Usa Llc Map-less and camera-based lane markings sampling method for level-3 autonomous driving vehicles
CN112041637A (en) * 2019-03-28 2020-12-04 百度时代网络技术（北京）有限公司 Map-less and camera-based lane marker sampling method for 3-level autonomous vehicles
WO2020191712A1 (en) * 2019-03-28 2020-10-01 Baidu.Com Times Technology (Beijing) Co., Ltd. A map-less and camera-based lane markings sampling method for level-3 autonomous driving vehicles
CN113825978B (en) * 2019-04-12 2022-08-26 赛峰电子与防务公司 Method and device for defining path and storage device
CN113825978A (en) * 2019-04-12 2021-12-21 赛峰电子与防务公司 Method for defining a path
CN110262488A (en) * 2019-06-18 2019-09-20 重庆长安汽车股份有限公司 Local paths planning method, system and the computer readable storage medium of automatic Pilot
CN110262488B (en) * 2019-06-18 2021-11-30 重庆长安汽车股份有限公司 Automatic driving local path planning method, system and computer readable storage medium
US11648936B2 (en) 2019-10-09 2023-05-16 Apollo Intelligent Driving Technology (Beiiing) Co., Ltd. Method and apparatus for controlling vehicle
CN112783144A (en) * 2019-10-22 2021-05-11 舜宇光学（浙江）研究院有限公司 Path generation method, path planning method, system and equipment
CN112783144B (en) * 2019-10-22 2023-09-29 舜宇光学（浙江）研究院有限公司 Path generation method, path planning method, system and equipment thereof
CN112540602A (en) * 2019-11-27 2021-03-23 深圳优地科技有限公司 Robot local path optimization method and robot
CN111006667B (en) * 2019-12-09 2021-07-06 东风商用车有限公司 Automatic driving track generation system under high-speed scene
CN111006667A (en) * 2019-12-09 2020-04-14 东风商用车有限公司 Automatic driving track generation system under high-speed scene
CN111174797A (en) * 2020-01-16 2020-05-19 湖南大学 Closed area global path planning method
CN113515111B (en) * 2020-03-25 2023-08-25 宇通客车股份有限公司 Vehicle obstacle avoidance path planning method and device
CN113515111A (en) * 2020-03-25 2021-10-19 郑州宇通客车股份有限公司 Vehicle obstacle avoidance path planning method and device
CN111289008A (en) * 2020-04-28 2020-06-16 南京维思科汽车科技有限公司 Local path planning algorithm for unmanned vehicle
CN111679667B (en) * 2020-05-20 2022-09-02 东南大学 Path and vehicle speed collaborative planning method for unmanned racing vehicle
CN111679667A (en) * 2020-05-20 2020-09-18 东南大学 Path and vehicle speed collaborative planning method for unmanned racing vehicle
CN112504288B (en) * 2020-07-22 2022-06-28 北京智行者科技有限公司 Local path planning method based on dynamic planning
CN112504288A (en) * 2020-07-22 2021-03-16 北京智行者科技有限公司 Local path planning method based on dynamic planning
CN111998864B (en) * 2020-08-11 2023-11-07 东风柳州汽车有限公司 Unmanned vehicle local path planning method, device, equipment and storage medium
CN111998864A (en) * 2020-08-11 2020-11-27 东风柳州汽车有限公司 Unmanned vehicle local path planning method, device, equipment and storage medium
CN112082557A (en) * 2020-09-14 2020-12-15 哈尔滨工程大学 UUV submarine topography tracking path rolling generation method based on Bessel fitting
CN112254727A (en) * 2020-09-23 2021-01-22 锐捷网络股份有限公司 TEB-based path planning method and device
CN112092810A (en) * 2020-09-24 2020-12-18 上海汽车集团股份有限公司 Vehicle parking-out method and device and electronic equipment
CN114370874A (en) * 2020-10-15 2022-04-19 郑州宇通客车股份有限公司 Vehicle, and vehicle path planning method and device
CN114370874B (en) * 2020-10-15 2023-08-25 宇通客车股份有限公司 Vehicle, vehicle path planning method and device
CN112666945A (en) * 2020-12-18 2021-04-16 广东嘉腾机器人自动化有限公司 AGV curve path optimization method
CN112859864A (en) * 2021-01-15 2021-05-28 大连海事大学 Unmanned ship-oriented geometric path planning method
CN112880700B (en) * 2021-02-26 2024-04-16 北京智行者科技股份有限公司 Local path planning method and device for in-situ steering vehicle
CN112880700A (en) * 2021-02-26 2021-06-01 重庆智行者信息科技有限公司 Local path planning method and device for pivot steering vehicle
CN113324554A (en) * 2021-05-28 2021-08-31 江铃汽车股份有限公司 Automatic driving route planning method and device, readable storage medium and electronic equipment
CN113324554B (en) * 2021-05-28 2023-12-29 江铃汽车股份有限公司 Automatic driving route planning method and device, readable storage medium and electronic equipment
CN113406960A (en) * 2021-07-08 2021-09-17 浙江大学 Real-time path planning and control method for ground steering of agricultural unmanned vehicle
CN114237229B (en) * 2021-11-26 2023-06-13 青岛德智汽车科技有限公司 Unstructured road work vehicle path planning method based on empirical path fitting
CN114237229A (en) * 2021-11-26 2022-03-25 青岛德智汽车科技有限公司 Unstructured road operation vehicle path planning method based on empirical path fitting
CN114185352A (en) * 2021-12-08 2022-03-15 东风悦享科技有限公司 High-precision high-real-time automatic driving local path planning method
CN114185352B (en) * 2021-12-08 2024-01-19 东风悦享科技有限公司 High-precision high-real-time automatic driving local path planning method
CN114310941B (en) * 2021-12-21 2023-10-20 长三角哈特机器人产业技术研究院 Robot path generation method for hub wheel hole deburring
CN114310941A (en) * 2021-12-21 2022-04-12 哈尔滨工业大学芜湖机器人产业技术研究院 Robot path generation method for hub wheel hole deburring machining
CN114459486A (en) * 2022-02-24 2022-05-10 深圳市优必选科技股份有限公司 Robot and channel navigation method, device and storage medium thereof
CN114995465A (en) * 2022-08-02 2022-09-02 北京理工大学 Multi-unmanned vehicle motion planning method and system considering vehicle motion capability
CN116663939B (en) * 2023-07-31 2023-10-17 北京理工大学 Unmanned vehicle path planning scene and task complexity evaluation method and system
CN116663939A (en) * 2023-07-31 2023-08-29 北京理工大学 Unmanned vehicle path planning scene and task complexity evaluation method and system

Also Published As

Publication number Publication date
CN108088456B (en) 2021-07-16

Similar Documents

Publication Publication Date Title
CN108088456A (en) A kind of automatic driving vehicle local paths planning method with time consistency
CN111123952B (en) Trajectory planning method and device
US11561542B2 (en) Safety and comfort constraints for navigation
CN107702716B (en) Unmanned driving path planning method, system and device
US20210078562A1 (en) Planning for unknown objects by an autonomous vehicle
US10281920B2 (en) Planning for unknown objects by an autonomous vehicle
CN110531770B (en) RRT path planning method and system based on improvement
US10234864B2 (en) Planning for unknown objects by an autonomous vehicle
CN109059944B (en) Motion planning method based on driving habit learning
CN109084798A (en) Network issues the paths planning method at the control point with road attribute
US11499834B2 (en) Aligning road information for navigation
CN110304074A (en) A kind of hybrid type driving method based on stratification state machine
CN104897168A (en) Intelligent vehicle path search method and system based on road risk assessment
CN104537834A (en) Intersection identification and intersection trajectory planning method for intelligent vehicle in urban road running process
CN112101120B (en) Map model based on automatic driving application scene and application method thereof
CN113619603B (en) Method for planning turning track of double-stage automatic driving vehicle
Park et al. Path generation algorithm based on crash point prediction for lane changing of autonomous vehicles
CN117109620A (en) Automatic driving path planning method based on interaction of vehicle behaviors and environment
CN115140048A (en) Automatic driving behavior decision and trajectory planning model and method
CN114830202A (en) Planning for unknown objects by autonomous vehicles
CN115246394A (en) Intelligent vehicle overtaking obstacle avoidance method
GALLAZZI Clothoid-SLAM: simultaneous localization and mapping of line markings for autonomous driving HD maps

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant