CN109540159A - A kind of quick complete automatic Pilot method for planning track - Google Patents
A kind of quick complete automatic Pilot method for planning track Download PDFInfo
- Publication number
- CN109540159A CN109540159A CN201811183196.7A CN201811183196A CN109540159A CN 109540159 A CN109540159 A CN 109540159A CN 201811183196 A CN201811183196 A CN 201811183196A CN 109540159 A CN109540159 A CN 109540159A
- Authority
- CN
- China
- Prior art keywords
- path
- planning
- point
- curvature
- driving behavior
- 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
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
- G01C21/3446—Details of route searching algorithms, e.g. Dijkstra, A*, arc-flags, using precalculated routes
Abstract
The present invention relates to a kind of quick complete automatic Pilot method for planning track, comprising the following steps: 1) foundation of planning path;2) collision detection is carried out to basic planning driving path, carries out step 3), collides if it does not exist, then whether judgement basis planning driving path reaches target point, if so, step 4) is carried out, if it is not, then carrying out step 3);3) after obtaining smooth active path using sampling/search-path layout method, step 4) is carried out;4) segmentation speed planning is carried out after the corresponding curvature extremum value of to acquisite approachs, obtains track;5) collision detection is carried out to track in time-domain and spatial domain, if there is collision, then judge whether vehicle is greater than 2 times of minimum braking distances at a distance from barrier, if so, carrying out speed weight-normality draws simultaneously return step 4), if not, then return step 1), path is finally generated into track with velocity composition and is exported, compared with prior art, the present invention has many advantages, such as to promote the real-time of trajectory planning, steady reliable.
Description
Technical field
The present invention relates to the trajectory planning fields of the automatic driving vehicle under urban structure road, more particularly, to one kind
Quickly complete automatic Pilot method for planning track.
Background technique
With the development of society and the continuous promotion of living standards of the people, domestic car ownership is from 2000 to 2017
It increases sharply 13.5 times between 17 years of year, the following car ownership estimation is at 400,000,000.However, severe energy crisis, height are negative increasingly
The increasingly increase that the traffic pressure of lotus and society require traffic safety accelerates the paces of unmanned technology landing.Make
For one of the core of unmanned technology, trajectory planning strategy must the real-time of boosting algorithm kept away with responding the variation of environment
Exempt to collide, causes the unnecessary person and property loss, meanwhile, the completeness of tactful also necessary boosting algorithm, to adapt to
The variation of traffic improves comfort, reduces traffic congestion.
Traditional automatic Pilot trajectory planning strategy is general to consider using the method for sampling (such as A*) or searching method (ratio
Such as RRT) it is this kind of have analytic completeness or a probability completeness algorithm, but this kind of algorithm is there are the process of blind search,
Solution procedure extremely time-consuming and the track generated are a section broken lines, and curvature is simultaneously discontinuous, and vehicle is difficult to carry out;Also it adopts
With this kind of planing method for having rapidity such as Dubins curve, Reeds-Shepp curve, spline curve, but this kind of algorithm
Completeness is not had, when environment is relative complex, the ability that solution obtains feasible solution sharply declines, and time-consuming is also on constantly
It rises.In addition, the result of trajectory planning has to comply with traffic rules under urban structure road conditions, meet the row of driver
For characteristic.
Therefore, how to provide a kind of parking strategy to solve the above problems and system is that those skilled in the art are urgently to be resolved
The problem of.
Summary of the invention
It is an object of the present invention to overcome the above-mentioned drawbacks of the prior art and provide it is a kind of quickly it is complete from
Dynamic driving locus planing method.
The purpose of the present invention can be achieved through the following technical solutions:
A kind of quick complete automatic Pilot method for planning track, comprising the following steps:
1) foundation of planning path obtains the basic planning driving path of imitative driving behavior;
2) collision detection is carried out to basic planning driving path, collided if it exists, then recorded collision and occur in basic planning driving path
Position, carry out step 3), collide if it does not exist, then whether judgement basis planning driving path reaches target point, if so, being walked
It is rapid 4), if it is not, then carrying out step 3);
3) after obtaining smooth active path using sampling/search-path layout method, step 4) is carried out;
4) curvature curve of speed planning method to acquisite approachs is used, and is divided after the corresponding curvature extremum value of to acquisite approachs
Section speed planning, obtains track;
5) collision detection is carried out to track in time-domain and spatial domain, if collisionless, judge vehicle and barrier away from
From whether 2 times of minimum braking distances are greater than, if so, carrying out speed weight-normality draws simultaneously return step 4), if it is not, then return step
1) path finally, is generated into track with velocity composition and is exported.
The step 1) specifically includes the following steps:
11) vehicle key parameter, including vehicle structure parameter, vehicle actuator performance parameter and roadway characteristic ginseng are obtained
Number, wherein vehicle structure parameter includes wheelbase, wheelspan, vehicle commander, vehicle width, car weight and centroid position, and vehicle executes actuator performance
Parameter includes that the max speed, minimum speed, peak acceleration and minimum acceleration, roadway characteristic parameter include coefficient of road adhesion
And surface roughness;
12) according to vehicle key parameter calculate curvature limit value, the curvature limit value be taken as minimum turning radius limit value curvature,
The minimum value of the maximum curvature three of the maximum curvature and coefficient of road adhesion limit value of lateral maximum side acceleration limit value;
13) high-comfort trajectory planning is judged whether to according to user demand or adaptive decision-making:
131) when comfort is of less demanding, i.e., when road restricted speed is not more than 20km/h, kept straight on respectively, lane-change/
And the imitative driving behavior path clustering point in road, turning and u-turn four kinds of driving behaviors guiding is sought, then with B-spline curves or
Person's Bezier is fitted control point, finally obtains smooth basic planning driving path;
132) when comfort requires high, i.e., when road restricted speed is greater than 20km/h, turned and reversed end for end respectively two
The path clustering point of kind driving behavior guiding is sought, and is then fitted control point with helix segment and circular arc path segment, finally
Obtain smooth basic planning driving path
The step 3) specifically includes the following steps:
31) basic planning driving path is cut, obtains active path;
32) by active path or effective node updates to path tree, the path tree is dynamic kd tree, tree
Combined normalized Euclidean distance and accumulation heading angle deviation are realized in the position adjustment of interior joint;
33) judge whether sampling/search point spread number reaches given threshold, if so, step 37) is carried out, if
It is no, then carry out step 34);
34) sampling method or search method expanding node are selected;
35) collision detection is carried out to expanding node and its with the connecting line of father node, collided if it exists, then the node is not deposited
Enter in path tree, and return step 33);
36) judge whether node expands to target point, if it is not, then return step 32), if so, carrying out 37);
37) effective node in path tree is taken out, control point is intended with B-spline curves or Bezier
It closes, obtains smooth active path.
In the step 131), imitative driving behavior path clustering point is obtained according to curvature limit value, control point meet with
Lower constraint:
Wherein, Euclidean distance of the L between two control points, α angle between control point, klimitFor curvature limit value.
In the step 131),
It is made of according to the straight trip path that driving behavior is oriented to n1 sections of control straightways, straightway meets following constraint:
Wherein, lend-start1For planning terminal to the Euclidean distance of planning starting point, li1For the i-th 1 sections of length of straigh line, αi1The
Angle between the control point of i1 sections of straightways;
According to driving behavior determining lane-change/and path is made of n2 sections of control straightways, and straightway satisfaction is following about
Beam:
Wherein, lend-start2For planning terminal to the Euclidean distance of planning starting point, li2For the i-th 2 sections of length of straigh line, αi2The
Angle between the control point of i2 sections of straightways;
It is made of according to the turning path that driving behavior determines n3 sections of control straightways, straightway meets following constraint:
Not turn constraints of piggybacking:
The turn constraints of left piggybacking:
The turn constraints of upward piggybacking:
Wherein, θj3It is starting point course angle, θg3It is terminal course angle, xg3、yg3For terminal point coordinate, xj3、yj3For starting point coordinate,
li3For the i-th 3 sections of length of straigh line, αi3Angle between the control point of the i-th 3 sections of straightways;
It is made of according to the u-turn path that driving behavior determines n4 sections of control straightways, straightway meets following constraint:
Not turn constraints of piggybacking:
The turn constraints of left piggybacking:
The turn constraints of right piggybacking:
The turn constraints of the equal piggybacking in left and right:
Wherein, θj4It is starting point course angle, θg4It is terminal course angle, xg4、yg4For terminal point coordinate, xj4、yj4For starting point coordinate,
li4For the i-th 4 sections of length of straigh line, αi4Angle between the control point of the i-th 4 sections of straightways.
In the step 132), one section of circular arc and two sections of spiral shells are used according to the improved turning of driving behavior and u-turn path
Spin line composition meets following constraint:
First helix:
a1s2+b1s+c1=k1
c1=kstart
Article 2 helix:
a2s2+b2s+c2=k2
c2=kend
Circular curve:
x1=Rx-Rsinθ1
Y=Ry+Rcosθ1
x2=Rx-Rsinθ2
y2=Ry+Rsinθ2
Wherein, s is path length, s1For first segment helical length, s2For second segment helical length, a1、b1、c1It is
One section spiral line characteristic parameter to be asked, a2、b2、c2For the second section spiral line characteristic parameter to be asked, k1For the first section spiral line song
Rate, k2For the second section spiral line curvature, kstartFor starting point curvature, kendFor terminal curvature, θiFor starting point course, θendFor terminal boat
To θ1For the first section spiral line terminal course, θ2For the second section spiral line starting point course, R is arc radius, Rx, RyFor circular arc circle
Heart coordinate, x1, y1Circular arc starting point coordinate, x2, y2The integral of circular arc terminal point coordinate, spin line is solved using Simpson formula, to reduce
Computation complexity, boosting algorithm real-time.
In the step 31), path is cut specifically:
The point of impingement arrived by collision detection traces back to nearest one along basic path and does not collide control point, this is not touched
Control point is hit as sampling/search-path layout starting point, and this does not collide control point between the starting point of basic path planning
The active path that smooth paths are obtained as cutting.
In the step 3), sampling/search-path layout switching mode includes that user sets and by trajectory planning
Strategy according to environment complexity carry out it is adaptively selected, when in environment exist it is larger compared with multi-obstacle avoidance or environmental map
When, quickly extended using RRT, when exist in environment relatively simple or environmental map it is smaller when, inspired using A*
Formula search.
In the step 34),
Sampling method be improve RRT method, point spread mode include be forced to, driving behavior skewed popularity extension and with
Machine extends three kinds of modes, and three kinds of specific adaptive selection methods of extended mode are as follows:
When environment is single and success rate is higher than given threshold, using being forced to;In the success being repeatedly forced to
When rate is less than given threshold, then extended using driving behavior skewed popularity;When success rate is lower larger with environment complexity, use
Standard extension mode;
Described being forced to is directly to be stored in target point as the node of extension in path tree, and carry out collision inspection
It surveys, verifies the validity of expanding node, effectively then retain, then delete in vain, the driving behavior skewed popularity extension root first
According to driving behavior, by the fan-shaped mark of extension field, and relay node collection, junction node collection and target section are generated in position
Point set, seed point will be chosen out of sector mark domain using gaussian random form, and node is stored in path tree, and is collided
Detection, verifies the validity of expanding node, effectively then retains, then delete in vain, and the standard extension is gone according to driving first
The regular polygon that extension field rectangle or rectangle splice to be marked, and generate relay node collection, road in position
Mouth node collection and destination node collection, seed point will be chosen out of mark domain using standard random basis, and node deposit path tree is worked as
In, and collision detection is carried out, the validity of expanding node is verified, is effectively then retained, is then deleted in vain.
Search method includes the search of driving behavior skewed popularity, steering angle constraint using A* method, point spread mode is improved
Search and three kinds of modes of standard search, three kinds of specific adaptive selection methods of extended mode are as follows:
When environment is single and success rate is higher than given threshold, the extension searched for using driving behavior skewed popularity;More
When the success rate of secondary driving behavior skewed popularity search is less than given threshold, then the extension of steering angle Constrain Searching is used;In success
When rate is lower and environment complexity is larger, using the extended mode of standard search.
The extension of the driving behavior skewed popularity search by the fan-shaped mark of extension field, and generates in position
Relay node collection, junction node collection and destination node collection, point spread will meet the minimum turning radius constraint of vehicle, and totally 220
The direction of a path base carries out, and in the node deposit path tree of extension, and carries out collision detection, verifies the effective of expanding node
Property, effectively then retain, then delete in vain, extension field rectangle or rectangle are spelled in the extension of the steering angle Constrain Searching
The regular polygon mark connect, and relay node collection, junction node collection and destination node collection are generated in position, node expands
Exhibition constrains the minimum turning radius for meeting vehicle, and the direction of totally 220 path bases carries out, and the node deposit path tree of extension is worked as
In, and collision detection is carried out, the validity of expanding node is verified, is effectively then retained, is then deleted in vain, the standard search
Extension, the regular polygon that extension field rectangle or rectangle are spliced mark, and generate in position relay node collection,
Junction node collection and destination node collection, point spread will carry out totally along the direction of vertical and longitudinal 20 path bases, extension
Node is stored in path tree, and carries out collision detection, is verified the validity of expanding node, is effectively then retained, then deletes in vain.
In the step 4), the planning of curvature extremum value segmentation speed carries out even between two point with extreme curvature in basic path
The speed planning of speed, extreme point use the acceleration and deceleration speed planning by road surface attachment constraint and actuator performance constraint outside,
Or point-to-point transmission carries out at the uniform velocity before and after four point with extreme curvature in basic path and sampling/search path
Speed planning, extreme point is outside using the acceleration and deceleration speed planning by road surface attachment constraint and actuator performance constraint.
Sampling/search-path layout switching mode can be user's setting, can also transfer to trajectory planning
Strategy is adaptively selected according to the complexity progress of environment, when existing in the environment compared with multi-obstacle avoidance or larger environmental map,
It is quickly extended using RRT;In the environment in the presence of when relatively simple or environmental map is smaller, heuristic search is carried out using A*
Rope.
In the step 5), time-domain and spatial domain collision detection are with speed representation time-domain, and path represents space
Three-dimensional system of coordinate is constructed in domain, and respectively using the investigative range of sensor, road boundary as spatial domain bound, with road speed limit,
Pavement conditions speed limit and vehicle actuator performance speed limit are time-domain bound;
In the step 5), decision of replanning is specially that whether in safety can plan domain to the point of impingement and collision time
On the basis of interior, whether combining environmental complexity and traffic rules decision pay the utmost attention to speed planning out.
Compared with prior art, the invention has the following advantages that
The present invention provides a kind of method for planning track of driving behavior guiding, are suitable for urban structure road, reduce
The blindness of planning process, reduces the time-consuming of trajectory planning, the track of generation meets the behavioral trait for the person of sailing, automatic Pilot
Process is more steady comfortable.
One, the Path Generation that the present invention plans considers dynamics of vehicle constraint, vehicle actuator performance constraints and road surface
Condition, using B-spline or bezier curve smooth track, the trajectory tortuosity of generation is continuous, and enforceability is strong.
Two, the present invention devises a kind of basic paths planning method of imitative driving behavior, and driving behavior is divided into directly
Row, lane-change/and road, turning and u-turn, keep planning process more succinct, reduce the complexity of algorithm, and real-time is high.
Three, the present invention devises a kind of improvement sampling planning and improves searching method, by being formed with basic Path Method
Complementation, makes method for planning track have completeness, and trajectory planning result is reliable and stable.
Four, the present invention devises a kind of speed planning method of curvature extremum value segmentation, and this method and path planning decouple, real
The modularization of existing path planning and speed planning, meanwhile, the speed planning of anterior-posterior approach has portability, velocity planning algorithm
Complexity is low, strong real-time.
Five, the present invention is in trajectory planning specific implementation process, it is contemplated that the complexity of environment, vehicle-state, driving
Member's demand and algorithm success rate are provided with multiple self-adaptive links, adaptation of the strategy to environment and operator demand to algorithm
Ability is strong.
Detailed description of the invention
Fig. 1 is the automatic Pilot trajectory planning strategic process figure for having rapidity and completeness of the invention;
Fig. 2 is trajectory planning strategy of the invention with path that driving behavior is guiding and the specific reality that changes machine RRT and merge
Apply flow chart;
Fig. 3 is the specific implementation that trajectory planning strategy of the invention is merged with the path that driving behavior is guiding with improvement A*
Flow chart;
Fig. 4 is the path clustering point and smoothed out path schematic diagram of lane-change of the invention/and road driving behavior guiding.Its
In, figure (4a) is the control straightway schematic diagram of the path clustering point construction of lane-change/and road driving behavior guiding, and figure (4b) is to change
The control straightway of road/and the path clustering point construction of road driving behavior guiding is through the smoothed out schematic diagram of B-spline curves;
Fig. 5 is the path clustering point and smoothed out path schematic diagram that a kind of turning driving behavior of the invention is oriented to.Its
In, figure (5a) is the control straightway schematic diagram of the not path clustering point construction of piggybacking turning driving behavior guiding, and figure (5b) is
The control straightway of the path clustering point construction of piggybacking turning driving behavior guiding is not through the smoothed out schematic diagram of B-spline curves;
Fig. 6 is the path clustering point and smoothed out path schematic diagram that a kind of turning driving behavior of the invention is oriented to.Its
In, figure (6a) is the control straightway schematic diagram of the path clustering point construction of left piggybacking turning driving behavior guiding, and figure (6b) is
The control straightway of the path clustering point construction of left piggybacking turning driving behavior guiding is through the smoothed out schematic diagram of B-spline curves;
Fig. 7 is the path clustering point and smoothed out path schematic diagram that a kind of turning driving behavior of the invention is oriented to.Its
In, figure (7a) is the control straightway schematic diagram of the path clustering point construction of upper piggybacking turning driving behavior guiding, and figure (7b) is
The control straightway of the path clustering point construction of upper piggybacking turning driving behavior guiding is through the smoothed out schematic diagram of B-spline curves;
Fig. 8 is the path clustering point and smoothed out path schematic diagram that a kind of u-turn driving behavior of the invention is oriented to.Its
In, figure (8a) is that piggybacking does not reverse end for end the control straightway schematic diagram that the path clustering point of driving behavior guiding constructs, and figure (8b) is
The control straightway of the path clustering point construction of piggybacking u-turn driving behavior guiding is not through the smoothed out schematic diagram of B-spline curves;
Fig. 9 is the path clustering point and smoothed out path schematic diagram that a kind of u-turn driving behavior of the invention is oriented to.Its
In, figure (9a) is the control straightway schematic diagram for the path clustering point construction that left piggybacking reverses end for end driving behavior guiding, and figure (9b) is
The control straightway of the path clustering point construction of left piggybacking u-turn driving behavior guiding is through the smoothed out schematic diagram of B-spline curves;
Figure 10 is the path clustering point and smoothed out path schematic diagram that a kind of u-turn driving behavior of the invention is oriented to.Its
In, figure (10a) is the control straightway schematic diagram for the path clustering point construction that right piggybacking reverses end for end driving behavior guiding, is schemed (10b)
The control straightway of the path clustering point construction of driving behavior guiding is reversed end for end through the smoothed out signal of B-spline curves for right piggybacking
Figure;
Figure 11 is the path clustering point and smoothed out path schematic diagram that a kind of u-turn driving behavior of the invention is oriented to;Its
In, figure (11a) is the control straightway schematic diagram for the path clustering point construction that the equal piggybacking in left and right reverses end for end driving behavior guiding, figure
(11b) is that the control straightway for the path clustering point construction that the equal piggybacking in left and right reverses end for end driving behavior guiding is smooth through B-spline curves
Schematic diagram afterwards;
Figure 12 is the turning of a modification of the present invention and the smooth paths schematic diagram of u-turn driving behavior guiding.Wherein,
Scheme the smooth paths schematic diagram that (12a) is improved turning driving behavior guiding, figure (12b) is that improved u-turn driving behavior is led
To smooth paths schematic diagram;
Figure 13 is lane-change/of the invention and road driving behavior is that the path being oriented to cuts to obtain the schematic diagram of active path;
Figure 14 is straight trip and lane-change/of the invention and road driving behavior is three kinds of samplings/search space schematic diagram of guiding.
Wherein, figure (14a) is straight trip and lane-change/and road driving behavior schematic diagram of a scenario, and figure (14b) is straight trip and lane-change/and road drives
The Stochastic propagation of behavior/standard search schematic diagram, figure (14c) is straight trip and lane-change/and the skewed popularity of road driving behavior extends/turns
Angle Constrain Searching/skewed popularity searches for schematic diagram, and figure (14d) be that straight trip and lane-change/simultaneously road driving behavior is forced to schematic diagram;
Figure 15 is the environment schematic that turning driving behavior of the invention is guiding.
Figure 16 is a kind of sampling/search space schematic diagram that turning driving behavior of the invention is guiding;
Figure 17 is a kind of sampling/search space schematic diagram that turning driving behavior of the invention is guiding;
Figure 18 is the environment schematic that u-turn driving behavior of the invention is guiding;
Figure 19 is a kind of sampling/search space schematic diagram that u-turn driving behavior of the invention is guiding;
Figure 20 is a kind of sampling/search space schematic diagram that u-turn driving behavior of the invention is guiding;
Figure 21 is a kind of point spread schematic diagram of the invention;
Figure 22 is a kind of point spread schematic diagram of the invention;
Figure 23 is a kind of point spread schematic diagram of the invention;
Figure 24 is a kind of path curvatures schematic diagram of the invention;
Figure 25 is a kind of speed planning curve synoptic diagram of the invention;
Figure 26 is a kind of path curvatures schematic diagram of the invention;
Figure 27 is a kind of speed planning curve synoptic diagram of the invention;
Figure 28 is a kind of speed planning curve synoptic diagram of the invention.
Specific embodiment
The present invention is described in detail with specific embodiment below in conjunction with the accompanying drawings.
Embodiment
Fig. 1 is the flow chart of the automatic Pilot trajectory planning strategy for having rapidity and completeness of the invention, Fig. 2 and figure
3 be the flow chart for having rapidity from two kinds of different realizations of the automatic Pilot trajectory planning strategy of completeness of the invention.It should
Method specific steps include:
Step 1: as shown in Figure 1, obtaining vehicle key parameter, including vehicle structure parameter, vehicle actuator performance parameter
With roadway characteristic parameter.Wherein, vehicle structure parameter includes wheelbase, wheelspan, vehicle commander, vehicle width, car weight, centroid position etc., vehicle
Executing actuator performance parameter includes the max speed, minimum speed, peak acceleration, minimum acceleration etc., roadway characteristic parameter
Including coefficient of road adhesion, surface roughness etc.;
Step 2: as shown in Figure 1, calculating curvature limit value, { the minimum turning half of curvature limit value=min according to vehicle key parameter
The curvature of diameter limit value, the maximum curvature of lateral maximum side acceleration limit value, the maximum curvature of coefficient of road adhesion limit value };
Step 3: as shown in Figure 1, according to the demand of user, or it is adaptive for being changed etc. with current vehicle speed, road curvature
Condition chooses whether to need to carry out high-comfort trajectory planning, is to be transferred to step 31, is otherwise transferred to step 32:
Step 31: kept straight on respectively, lane-change/and road, turning and u-turn four kinds of driving behaviors guiding path clustering point
It seeks;
Step 311: control point being calculated according to the curvature limit value of step 2, the constraint equation that control point needs meet isWherein Euclidean distance of the L between two control points, α angle between control point, klimitFor song
Rate limit value.It is transferred to step 312, step 313, step 314 or step 315 respectively according to driving behavior decision;
Step 312: straight line path is formed using the control straightway of n1 segment length, and straightway needs the constraint equation met
Are as follows:
Wherein, lend-start1It is the Euclidean distance for planning terminal to planning starting point, li1For the i-th 1 sections of length of straigh line, αi1The
Angle between the control point of i1 sections of straightways.
Step 313: as shown in figure 4, lane-change/simultaneously for path using the control straightway composition of n2 segment length, straightway need to
The constraint equation to be met are as follows:
Wherein, lend-start2It is the Euclidean distance for planning terminal to planning starting point, li2For the i-th 2 sections of length of straigh line, αi2The
Angle between the control point of i2 sections of straightways.
Step 314: as shown in Figure 5-Figure 7, turning path is formed using the control straightway of n3 segment length, and straightway needs
The constraint equation of satisfaction are as follows:
Not turn constraints of piggybacking:
The turn constraints of left piggybacking:
The turn constraints of upward piggybacking:
Wherein, θj3It is starting point course angle, θg3It is terminal course angle, xg3、yg3For terminal point coordinate, xj3、yj3For starting point coordinate,
li3For the i-th 3 sections of length of straigh line, αi3Angle between the control point of the i-th 3 sections of straightways;
Step 315: as shown in figures s-11, u-turn path is formed using the control straightway of n segment length, and straightway needs
The constraint equation of satisfaction are as follows:
The not u-turn constraint of piggybacking:
The u-turn of left piggybacking constrains:
The u-turn of right piggybacking constrains:
The u-turn of the equal piggybacking in left and right constrains:
Wherein, θj4It is starting point course angle, θg4It is terminal course angle, xg4、yg4For terminal point coordinate, xj4、yj4For starting point coordinate,
li4For the i-th 4 sections of length of straigh line, αi4Angle between the control point of the i-th 4 sections of straightways.
Step 32: the step is transferred to when comfort requires high, according to behaviour decision making as a result, being turned and being reversed end for end respectively
The path clustering point of two kinds of driving behaviors guiding is sought, and is then fitted control point with helix segment and circular arc path segment, most
Smooth basic planning driving path is obtained eventually;
Step 321: as shown in figure 12, improved turning and u-turn path are formed using one section of circular arc and two section spiral lines,
The constraint equation for needing to meet are as follows:
First helix
a1s2+b1s+c1=k1
c1=kstart
Article 2 helix
a2s2+b2s+c2=k2
c2=kend
Circular curve
x1=Rx-Rsinθ1
Y=Ry+Rcosθ1
x2=Rx-Rsinθ2
y2=Ry+Rsinθ2
Wherein, s is path length, s1For first segment helical length, s2For second segment helical length, a1, b1, c1It is
One section spiral line characteristic parameter to be asked, a2, b2, c2For the second section spiral line characteristic parameter to be asked, k1For the first section spiral line song
Rate, k2For the second section spiral line curvature, kstartFor starting point curvature, kendFor terminal curvature, i is starting point course, and end is terminal boat
To 1 is the first section spiral line terminal course, and 2 be the second section spiral line starting point course, and R is arc radius, Rx, RyFor the circular arc center of circle
Coordinate, x1, y1Circular arc starting point coordinate, x2, y2Circular arc terminal point coordinate.
Step 322: the integral of helix is solved using Simpson formula, and to reduce computation complexity, boosting algorithm is real-time
Property.
Step 4: collision detection being carried out to basic planning driving path, it is determined whether there is collision, if there is collision, record collision
Occur to then branch to step 6 in the position of basic planning driving path.;
Step 5: whether judgement basis planning driving path reaches target point, if reaching target point, jumps to step 13;
Step 6: as shown in figure 13, basic planning driving path being cut, active path is obtained.It is by touching that path, which is cut,
The point of impingement detected is hit, nearest one is traced back to along basic path and does not collide control point, this does not collide control point conduct and adopts
Sample/search-path layout starting point, and this does not collide control point to the smooth paths between the starting point of basic path planning as sanction
The active path cut;
Step 7: in obtained active path or effective node updates to path tree.Path tree is one dynamic
Combined normalized Euclidean distance and accumulation heading angle deviation are realized in kd tree, the position adjustment for setting interior joint;
Step 8: judging whether sampling/search point spread number reaches given threshold, be to jump to step 12;
Step 9: selection sampling method or search method expanding node.Sampling/search-path layout switching mode can be
User's setting can also transfer to the strategy of trajectory planning to carry out according to the complexity of environment adaptively selected, exist in the environment
It when compared with multi-obstacle avoidance or larger environmental map, is quickly extended using RRT, executes step 91;In the environment exist compared with
When simple or environmental map is smaller, heuristic search is carried out using A*, jumps to step 92;
Step 91: sampling method is the RRT improved, and point spread mode is forced to, driving behavior is biased to
Property extension, three kinds of modes of Stochastic propagation:
The RRT point spread mode improved selects to be adaptively success rate, the environment complexity of integration node extension
And random seed probability element, when environment is single and success rate is higher than certain threshold value, using being forced to;Repeatedly forcing
When the small Mr. Yu's threshold value of the success rate of extension, then extended using driving behavior skewed popularity;Success rate is lower and environment complexity compared with
When big, using standard extension mode.
It is forced to be directly to be stored in target point as the node of extension in path tree, and carry out collision detection, test
The validity for demonstrate,proving expanding node, effectively then retains, then deletes in vain.
The extension of driving behavior skewed popularity is first according to driving behavior, by the fan-shaped mark of extension field, such as Figure 14, Figure 16 and figure
Shown in 19, and generate relay node collection, junction node collection and destination node collection in position, seed point will using Gauss with
Machine form is chosen out of sector mark domain, and node is stored in path tree, and carries out collision detection, verifies the effective of expanding node
Property, effectively then retain, then deletes in vain.
According to driving behavior, the regular polygon that extension field rectangle or rectangle splice is marked first for standard extension,
As shown in Figure 14, Figure 17 and Figure 20, and relay node collection, junction node collection and destination node collection are generated in position, kind
Son point will be chosen out of mark domain using standard random basis, and node is stored in path tree, and carries out collision detection, and verifying is expanded
The validity for opening up node, effectively then retains, then deletes in vain.
Step 92: search method is the A* improved, and point spread mode has the search of driving behavior skewed popularity, turns to
Three kinds of angle Constrain Searching, standard search modes:
The A* point spread mode improved selects to be that adaptively, the accumulative success rate of integration node extension, environment are multiple
Miscellaneous degree element, when environment is single and success rate is higher than certain threshold value, the extension searched for using driving behavior skewed popularity;Multiple
When the small Mr. Yu's threshold value of success rate of driving behavior skewed popularity search, then the extension of steering angle Constrain Searching is used;Success rate compared with
When low and environment complexity is larger, using the extended mode of standard search.
The extension of driving behavior skewed popularity search, by the fan-shaped mark of extension field, as shown in Figure 14, Figure 16 and Figure 19,
And it generates relay node collection, junction node collection and destination node collection, point spread in position and will meet the minimum of vehicle
Turning radius constraint, the direction of totally 220 path bases carries out, and as shown in Figure 21-Figure 23, the node deposit path tree of extension is worked as
In, and collision detection is carried out, the validity of expanding node is verified, is effectively then retained, is then deleted in vain.
The regular polygon that extension field rectangle or rectangle splice is marked, is such as schemed by the extension of steering angle Constrain Searching
Shown in 14- Figure 20, and generation relay node collection, junction node collection and destination node collection, point spread will expire in position
The minimum turning radius of sufficient vehicle constrains, and the direction of totally 220 path bases carries out, and as shown in Figure 21-Figure 23, the node of extension is deposited
Enter in path tree, and carry out collision detection, verifies the validity of expanding node, effectively then retain, then delete in vain.
The extension of standard search marks the regular polygon that extension field rectangle or rectangle splice, such as Figure 14-Figure 20
It is shown, and generate relay node collection, junction node collection and destination node collection in position, point spread will along vertical and
The direction of longitudinal totally 20 path bases carries out, and as shown in figure 21, in the node deposit path tree of extension, and carries out collision inspection
It surveys, verifies the validity of expanding node, effectively then retain, then delete in vain.
Step 10: carrying out collision detection to expanding node and its with the connecting line of father node, when there is collision, the node is not
It is stored in path tree, and jumps back to step 8;
Step 11: judging whether node expands to target point, if it is not, jumping to step 7;
Step 12: taking out effective node in path tree, control point is carried out with B-spline curves or Bezier
Fitting, finally obtains smooth active path;
Step 13: the curvature curve of solution path, and the curvature extremum value (maximum and minimum) in path is found, then
Segmentation speed planning is carried out, track is obtained:
As shown in Figure 24-Figure 25, the planning of curvature extremum value segmentation speed will be between two point with extreme curvature in basic path
Speed planning at the uniform velocity is carried out, is advised outside extreme point using the acceleration and deceleration speed constrained by road surface attachment constraint and actuator performance
It draws;
As shown in Figure 26-Figure 28, the planning of curvature extremum value segmentation speed will be in basic path and sampling/search path
The front and back point-to-point transmission of four point with extreme curvature carries out speed planning at the uniform velocity, constrains and executes using by road surface attachment outside extreme point
The acceleration and deceleration speed planning of device performance constraints.
Step 14: collision detection being carried out to track in time-domain and spatial domain and jumps to step 16 if collisionless.
Time-domain and spatial domain collision detection are with speed representation time-domain, and path represents spatial domain and constructs three-dimensional system of coordinate, and respectively
Using the investigative range of sensor, road boundary as spatial domain bound, with road speed limit, pavement conditions speed limit and vehicle actuator
Performance speed limit is time-domain bound;
Step 15: carry out decision of replanning, if carry out speed planning, jump back to step 13, if carry out without
Speed weight-normality is drawn, then jumps back to step 2.Decision of replanning be on the basis of step 14 to the point of impingement and collision time whether
On the basis of safety can be planned in domain, whether combining environmental complexity and traffic rules decision pay the utmost attention to speed planning out.;
Step 16: by path and velocity group biosynthesis locus, and exporting.
The present invention provides a kind of automatic Pilot trajectory planning strategies for having rapidity and completeness, and strategy is with driver
Behavior is guiding, by basic Path Planning, sampling/search-path layout strategy, collision detection strategy, speed planning strategy
With five part organic assembling of weight planning strategy.Basic Path Planning and sampling/search-path layout strategy are to drive row
To realize the acceleration of algorithm for guiding, while the organic assembling of strategy has ensured the completeness of algorithm, it is ensured that feelings existing for path
Under condition, which can export effective track.Wherein basic Path Planning is firstly the need of acquisition key parameter (vehicle structure
Parameter, vehicle actuator performance parameter and roadway characteristic parameter);Secondly the calculating of path curvatures limit value is carried out;Then according to comfortable
Property require choose driving behavior guiding basic path planning or improved driving behavior guiding basic path planning;Most
Basic planning driving path is exported afterwards.Sampling/search-path layout strategy is to be difficult to meet demand (collision in basic Path Planning
Detection does not pass through or path does not reach target point) when replenishment strategy, first basic planning driving path is cut until effective
Path simultaneously updates path tree;Secondly, carrying out sampling/search node extension;Then, it successively carries out collision detection and stops detection
(reaching target point or the limitation of planning number);Finally, being carried out to live part path smooth.Speed planning strategy uses curvature
The planning of extreme value segmentation speed.Decision of replanning strategy is to carry out path replanning or speed weight-normality stroke to decision.The present invention
By driving behavior to guide basic path planning, and incorporate sampling/search planning strategy, significant increase trajectory planning it is real-time
The completeness of property and algorithm, the track of generation meet human driver's characteristic, and automatic Pilot process is more steadily and reliably.
The above is only a preferred embodiment of the present invention, it should be pointed out that: for the ordinary skill people of the art
For member, various improvements and modifications may be made without departing from the principle of the present invention, these improvements and modifications are also answered
It is considered as protection scope of the present invention.
Claims (10)
1. a kind of quick complete automatic Pilot method for planning track, which comprises the following steps:
1) foundation of planning path obtains the basic planning driving path of imitative driving behavior;
2) collision detection is carried out to basic planning driving path, collided if it exists, then recorded collision and occur in the position of basic planning driving path
It sets, carries out step 3), collide if it does not exist, then whether judgement basis planning driving path reaches target point, if so, carrying out step
4), if it is not, then carrying out step 3);
3) after obtaining smooth active path using sampling/search-path layout method, step 4) is carried out;
4) curvature curve of speed planning method to acquisite approachs is used, and carries out segmentation speed after the corresponding curvature extremum value of to acquisite approachs
Metric is drawn, and track is obtained;
5) collision detection is carried out to track in time-domain and spatial domain and judges that vehicle is at a distance from barrier if there is collision
It is no to be greater than 2 times of minimum braking distances, if so, carrying out speed weight-normality draws simultaneously return step 4), if it is not, then return step 1), most
Path generates track with velocity composition and exports at last.
2. a kind of quick complete automatic Pilot method for planning track according to claim 1, which is characterized in that described
Step 1) specifically includes the following steps:
11) vehicle key parameter, including vehicle structure parameter, vehicle actuator performance parameter and roadway characteristic parameter are obtained,
In, vehicle structure parameter includes wheelbase, wheelspan, vehicle commander, vehicle width, car weight and centroid position, and vehicle executes actuator performance parameter
It include coefficient of road adhesion and road including the max speed, minimum speed, peak acceleration and minimum acceleration, roadway characteristic parameter
Surface roughness;
12) curvature limit value is calculated according to vehicle key parameter, which is taken as the curvature, lateral of minimum turning radius limit value
The minimum value of the maximum curvature three of the maximum curvature and coefficient of road adhesion limit value of maximum side acceleration limit value;
13) high-comfort trajectory planning is judged whether to according to user demand or adaptive decision-making:
131) when comfort is of less demanding, i.e., when road restricted speed is not more than 20km/h, kept straight on respectively, lane-change/and road,
The imitative driving behavior path clustering point of turning and u-turn four kinds of driving behaviors guiding is sought, then with B-spline curves or shellfish
Sai Er curve is fitted control point, finally obtains smooth basic planning driving path;
132) it when comfort requires high, i.e., when road restricted speed is greater than 20km/h, is turned respectively and first two is adjusted to drive
The path clustering point for sailing behavior guidance is sought, and is then fitted control point with helix segment and circular arc path segment, is finally obtained
Smooth basic planning driving path.
3. a kind of quick complete automatic Pilot method for planning track according to claim 1, which is characterized in that described
Step 3) specifically includes the following steps:
31) basic planning driving path is cut, obtains active path;
32) by active path or effective node updates to path tree, the path tree is dynamic kd tree;
33) judge whether sampling/search point spread number reaches given threshold, if so, step 37) is carried out, if it is not, then
Carry out step 34);
34) sampling method or search method expanding node are selected;
35) collision detection is carried out to expanding node and its with the connecting line of father node, collided if it exists, then the node is not stored in road
In diameter tree, and return step 33);
36) judge whether node expands to target point, if it is not, then return step 32), if so, carrying out 37);
37) effective node in path tree is taken out, control point is fitted with B-spline curves or Bezier, is obtained
To smooth active path.
4. a kind of quick complete automatic Pilot method for planning track according to claim 2, which is characterized in that described
In step 131), imitative driving behavior path clustering point is obtained according to curvature limit value, control point meets following constraint:
Wherein, Euclidean distance of the L between two control points, α angle between control point, klimitFor curvature limit value.
5. a kind of quick complete automatic Pilot method for planning track according to claim 4, which is characterized in that described
In step 131),
It is made of according to the straight trip path that driving behavior is oriented to n1 sections of control straightways, straightway meets following constraint:
Wherein, lend-start1For planning terminal to the Euclidean distance of planning starting point, li1For the i-th 1 sections of length of straigh line, αi1The i-th 1 sections
Angle between the control point of straightway;
According to driving behavior determining lane-change/and path is made of n2 sections of control straightways, and straightway meets following constraint:
Wherein, lend-start2For planning terminal to the Euclidean distance of planning starting point, li2For the i-th 2 sections of length of straigh line, αi2The i-th 2 sections
Angle between the control point of straightway;
It is made of according to the turning path that driving behavior determines n3 sections of control straightways, straightway meets following constraint:
Not turn constraints of piggybacking:
The turn constraints of left piggybacking:
The turn constraints of upward piggybacking:
Wherein, θj3It is starting point course angle, θg3It is terminal course angle, xg3、yg3For terminal point coordinate, xj3、yj3For starting point coordinate, li3For
The i-th 3 sections of length of straigh line, αi3Angle between the control point of the i-th 3 sections of straightways;
It is made of according to the u-turn path that driving behavior determines n4 sections of control straightways, straightway meets following constraint:
Not turn constraints of piggybacking:
The turn constraints of left piggybacking:
The turn constraints of right piggybacking:
The turn constraints of the equal piggybacking in left and right:
Wherein, θj4It is starting point course angle, θg4It is terminal course angle, xg4、yg4For terminal point coordinate, xj4、yj4For starting point coordinate, li4For
The i-th 4 sections of length of straigh line, αi4Angle between the control point of the i-th 4 sections of straightways.
6. a kind of quick complete automatic Pilot method for planning track according to claim 4, which is characterized in that described
In step 132), is formed, met using one section of circular arc and two section spiral lines according to the improved turning of driving behavior and u-turn path
It constrains below:
First helix:
a1s2+b1s+c1=k1
c1=kstart
Article 2 helix:
a2s2+b2s+c2=k2
c2=kemd
Circular curve:
x1=Rx-Rsinθ1
y=Ry+Rcosθ1
x2=Rs-Rsinθ2
y2=Ry+Rsinθ2
Wherein, s is path length, s1For first segment helical length, s2For second segment helical length, a1、b1、c1For first segment
Helix characteristic parameter to be asked, a2、b2、c2For the second section spiral line characteristic parameter to be asked, k1For the first section spiral line curvature, k2For
Second section spiral line curvature, kstartFor starting point curvature, kendFor terminal curvature,iFor starting point course,endFor terminal course,1It is
One section spiral line terminal course,2For the second section spiral line starting point course, R is arc radius, Rx, RyFor center coordinate of arc, x1,
y1Circular arc starting point coordinate, x2, y2Circular arc terminal point coordinate.
7. a kind of quick complete automatic Pilot method for planning track according to claim 3, which is characterized in that described
In step 31), path is cut specifically:
The point of impingement arrived by collision detection traces back to nearest one along basic path and does not collide control point, this does not collide control
System point as sampling/search-path layout starting point, and this do not collide control point to basis path planning starting point between it is smooth
The active path that path is obtained as cutting.
8. a kind of quick complete automatic Pilot method for planning track according to claim 3, which is characterized in that described
In step 3), sampling/search-path layout switching mode includes user's setting and the strategy foundation environment by trajectory planning
Complexity carry out adaptively selected, when existing in environment compared with multi-obstacle avoidance or larger environmental map, carried out using RRT fast
The extension of speed, when exist in environment relatively simple or environmental map it is smaller when, heuristic search is carried out using A*.
9. a kind of quick complete automatic Pilot method for planning track according to claim 8, which is characterized in that described
In step 34),
Sampling method is to improve RRT method, and point spread mode is including being forced to, driving behavior skewed popularity extends and expands at random
Three kinds of modes are opened up, three kinds of specific adaptive selection methods of extended mode are as follows:
When environment is single and success rate is higher than given threshold, using being forced to;It is small in the success rate being repeatedly forced to
When given threshold, then extended using driving behavior skewed popularity;When success rate is lower larger with environment complexity, using standard
Extended mode;
Search method includes the search of driving behavior skewed popularity, steering angle Constrain Searching using A* method, point spread mode is improved
With three kinds of modes of standard search, three kinds of specific adaptive selection methods of extended mode are as follows:
When environment is single and success rate is higher than given threshold, the extension searched for using driving behavior skewed popularity;Repeatedly driving
When sailing the success rate of behavior skewed popularity search less than given threshold, then the extension of steering angle Constrain Searching is used;Success rate compared with
When low and environment complexity is larger, using the extended mode of standard search.
10. a kind of quick complete automatic Pilot method for planning track according to claim 1, which is characterized in that described
Step 4) in, curvature extremum value segmentation speed planning carried out between two point with extreme curvature in basic path at the uniform velocity speed rule
It draws, the acceleration and deceleration speed planning constrained by road surface attachment constraint and actuator performance is used outside extreme point,
Or point-to-point transmission carries out speed at the uniform velocity before and after four point with extreme curvature in basic path and sampling/search path
Planning, extreme point is outside using the acceleration and deceleration speed planning by road surface attachment constraint and actuator performance constraint.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811183196.7A CN109540159B (en) | 2018-10-11 | 2018-10-11 | Rapid and complete automatic driving track planning method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811183196.7A CN109540159B (en) | 2018-10-11 | 2018-10-11 | Rapid and complete automatic driving track planning method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109540159A true CN109540159A (en) | 2019-03-29 |
CN109540159B CN109540159B (en) | 2020-11-27 |
Family
ID=65843846
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811183196.7A Active CN109540159B (en) | 2018-10-11 | 2018-10-11 | Rapid and complete automatic driving track planning method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109540159B (en) |
Cited By (23)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109835336A (en) * | 2019-02-19 | 2019-06-04 | 北京工业大学 | A kind of belt restraining square-wave-shaped speed planing method based on fuzzy algorithmic approach |
CN110146883A (en) * | 2019-05-17 | 2019-08-20 | 华东师范大学 | The same period accelerated based on minimum and more satellite platform MT-InSAR three-dimensional deformation decomposition methods of different phase |
CN110187706A (en) * | 2019-05-28 | 2019-08-30 | 上海钛米机器人科技有限公司 | A kind of speed planning method, apparatus, electronic equipment and storage medium |
CN110298131A (en) * | 2019-07-05 | 2019-10-01 | 西南交通大学 | Automatic Pilot lane-change decision model method for building up under a kind of mixing driving environment |
CN110530390A (en) * | 2019-09-16 | 2019-12-03 | 哈尔滨工程大学 | A kind of non-particle vehicle path planning method under narrow environment |
CN110703754A (en) * | 2019-10-17 | 2020-01-17 | 南京航空航天大学 | Path and speed highly-coupled trajectory planning method for automatic driving vehicle |
CN110861650A (en) * | 2019-11-21 | 2020-03-06 | 驭势科技(北京)有限公司 | Vehicle path planning method and device, vehicle-mounted equipment and storage medium |
CN111504340A (en) * | 2020-05-22 | 2020-08-07 | 北京汽车研究总院有限公司 | Vehicle path planning method and device and vehicle |
CN111497827A (en) * | 2020-02-17 | 2020-08-07 | 湖北亿咖通科技有限公司 | Automatic parking method, device, medium and equipment |
CN111645677A (en) * | 2020-05-20 | 2020-09-11 | 吉林大学 | Vehicle braking and steering coordinated control emergency anti-collision system and control method |
CN111750859A (en) * | 2020-05-29 | 2020-10-09 | 广州极飞科技有限公司 | Transition path planning method and related device |
CN111879307A (en) * | 2020-06-22 | 2020-11-03 | 国网河北省电力有限公司信息通信分公司 | Vehicle path planning method based on vehicle body parameters and engineering construction information |
CN112461240A (en) * | 2020-11-11 | 2021-03-09 | 武汉理工大学 | Unmanned aerial vehicle obstacle avoidance path planning method and system |
CN112462785A (en) * | 2020-12-04 | 2021-03-09 | 厦门大学 | Mobile robot path planning method and device and storage medium |
CN112506176A (en) * | 2019-08-26 | 2021-03-16 | 上海汽车集团股份有限公司 | Path planning method and device |
CN112632734A (en) * | 2020-12-29 | 2021-04-09 | 武汉中海庭数据技术有限公司 | High-precision map road connectivity test method and system |
CN112945254A (en) * | 2021-01-21 | 2021-06-11 | 西北工业大学 | Unmanned vehicle curvature continuous path planning method based on rapid expansion random tree |
WO2021120200A1 (en) * | 2019-12-20 | 2021-06-24 | Baidu.Com Times Technology (Beijing) Co., Ltd. | Spline curve and spiral curve based reference line smoothing method |
CN113320547A (en) * | 2021-07-15 | 2021-08-31 | 广州小鹏自动驾驶科技有限公司 | Path detection method and device and automobile |
CN113525510A (en) * | 2020-04-15 | 2021-10-22 | 百度(美国)有限责任公司 | System and method for automatically returning steering of an autonomously driven vehicle to a central location |
CN114089742A (en) * | 2021-10-25 | 2022-02-25 | 广东嘉腾机器人自动化有限公司 | AGV traveling speed control method and device based on path curvature and medium |
WO2022152283A1 (en) * | 2021-01-15 | 2022-07-21 | 北京理工大学 | Local path planning method for unmanned vehicle |
CN117533354A (en) * | 2023-12-28 | 2024-02-09 | 安徽蔚来智驾科技有限公司 | Track generation method, driving control method, storage medium and intelligent device |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104615889A (en) * | 2015-02-09 | 2015-05-13 | 武汉大学 | Intelligent vehicle path tracking method and system based on clothoid following |
US20150353082A1 (en) * | 2014-06-05 | 2015-12-10 | Carnegie Mellon University | Unified motion planning algorithm for autonomous driving vehicle in obstacle avoidance maneuver |
CN106873600A (en) * | 2017-03-31 | 2017-06-20 | 深圳市靖洲科技有限公司 | It is a kind of towards the local obstacle-avoiding route planning method without person bicycle |
CN107063280A (en) * | 2017-03-24 | 2017-08-18 | 重庆邮电大学 | A kind of intelligent vehicle path planning system and method based on control sampling |
WO2017153896A1 (en) * | 2016-03-07 | 2017-09-14 | Effidence | Motor-driven autonomous robot with obstacle anticipation |
CN108196536A (en) * | 2017-12-21 | 2018-06-22 | 同济大学 | A kind of improved unmanned vehicle rapidly-exploring random tree paths planning method |
CN108229730A (en) * | 2017-12-19 | 2018-06-29 | 同济大学 | A kind of automatic driving vehicle orbit generation method based on fuzzy reward |
-
2018
- 2018-10-11 CN CN201811183196.7A patent/CN109540159B/en active Active
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20150353082A1 (en) * | 2014-06-05 | 2015-12-10 | Carnegie Mellon University | Unified motion planning algorithm for autonomous driving vehicle in obstacle avoidance maneuver |
CN104615889A (en) * | 2015-02-09 | 2015-05-13 | 武汉大学 | Intelligent vehicle path tracking method and system based on clothoid following |
WO2017153896A1 (en) * | 2016-03-07 | 2017-09-14 | Effidence | Motor-driven autonomous robot with obstacle anticipation |
CN107063280A (en) * | 2017-03-24 | 2017-08-18 | 重庆邮电大学 | A kind of intelligent vehicle path planning system and method based on control sampling |
CN106873600A (en) * | 2017-03-31 | 2017-06-20 | 深圳市靖洲科技有限公司 | It is a kind of towards the local obstacle-avoiding route planning method without person bicycle |
CN108229730A (en) * | 2017-12-19 | 2018-06-29 | 同济大学 | A kind of automatic driving vehicle orbit generation method based on fuzzy reward |
CN108196536A (en) * | 2017-12-21 | 2018-06-22 | 同济大学 | A kind of improved unmanned vehicle rapidly-exploring random tree paths planning method |
Non-Patent Citations (2)
Title |
---|
LIANG MA 等: "Efficient Sampling-Based Motion Planning for On-Road Autonomous Driving", 《IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS》 * |
李奕姗 等: "基于三次螺旋线的无人车路径计算与仿真", 《佳木斯大学学报( 自然科学版)》 * |
Cited By (36)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109835336A (en) * | 2019-02-19 | 2019-06-04 | 北京工业大学 | A kind of belt restraining square-wave-shaped speed planing method based on fuzzy algorithmic approach |
CN110146883A (en) * | 2019-05-17 | 2019-08-20 | 华东师范大学 | The same period accelerated based on minimum and more satellite platform MT-InSAR three-dimensional deformation decomposition methods of different phase |
CN110146883B (en) * | 2019-05-17 | 2022-04-05 | 华东师范大学 | Simultaneous and asynchronous multi-satellite platform MT-InSAR three-dimensional deformation decomposition method based on minimum acceleration |
CN110187706A (en) * | 2019-05-28 | 2019-08-30 | 上海钛米机器人科技有限公司 | A kind of speed planning method, apparatus, electronic equipment and storage medium |
WO2020237890A1 (en) * | 2019-05-28 | 2020-12-03 | 上海钛米机器人科技有限公司 | Speed planning method and apparatus, electronic device and storage medium |
JP7345210B2 (en) | 2019-05-28 | 2023-09-15 | 上海▲タイ▼米機器人股▲フン▼有限公司 | Velocity planning methods, devices, electronic equipment and storage media |
CN110298131A (en) * | 2019-07-05 | 2019-10-01 | 西南交通大学 | Automatic Pilot lane-change decision model method for building up under a kind of mixing driving environment |
CN110298131B (en) * | 2019-07-05 | 2021-07-13 | 西南交通大学 | Method for establishing automatic driving lane change decision model in hybrid driving environment |
CN112506176A (en) * | 2019-08-26 | 2021-03-16 | 上海汽车集团股份有限公司 | Path planning method and device |
CN110530390A (en) * | 2019-09-16 | 2019-12-03 | 哈尔滨工程大学 | A kind of non-particle vehicle path planning method under narrow environment |
CN110703754A (en) * | 2019-10-17 | 2020-01-17 | 南京航空航天大学 | Path and speed highly-coupled trajectory planning method for automatic driving vehicle |
CN110703754B (en) * | 2019-10-17 | 2021-07-09 | 南京航空航天大学 | Path and speed highly-coupled trajectory planning method for automatic driving vehicle |
CN110861650A (en) * | 2019-11-21 | 2020-03-06 | 驭势科技(北京)有限公司 | Vehicle path planning method and device, vehicle-mounted equipment and storage medium |
CN110861650B (en) * | 2019-11-21 | 2021-04-16 | 驭势科技(北京)有限公司 | Vehicle path planning method and device, vehicle-mounted equipment and storage medium |
CN113286983A (en) * | 2019-12-20 | 2021-08-20 | 百度时代网络技术(北京)有限公司 | Reference line smoothing method based on spline curve and spiral curve |
WO2021120200A1 (en) * | 2019-12-20 | 2021-06-24 | Baidu.Com Times Technology (Beijing) Co., Ltd. | Spline curve and spiral curve based reference line smoothing method |
CN111497827A (en) * | 2020-02-17 | 2020-08-07 | 湖北亿咖通科技有限公司 | Automatic parking method, device, medium and equipment |
CN111497827B (en) * | 2020-02-17 | 2021-07-27 | 湖北亿咖通科技有限公司 | Automatic parking method, device, medium and equipment |
CN113525510A (en) * | 2020-04-15 | 2021-10-22 | 百度(美国)有限责任公司 | System and method for automatically returning steering of an autonomously driven vehicle to a central location |
CN111645677A (en) * | 2020-05-20 | 2020-09-11 | 吉林大学 | Vehicle braking and steering coordinated control emergency anti-collision system and control method |
CN111504340A (en) * | 2020-05-22 | 2020-08-07 | 北京汽车研究总院有限公司 | Vehicle path planning method and device and vehicle |
CN111504340B (en) * | 2020-05-22 | 2022-04-29 | 北京汽车研究总院有限公司 | Vehicle path planning method and device and vehicle |
CN111750859A (en) * | 2020-05-29 | 2020-10-09 | 广州极飞科技有限公司 | Transition path planning method and related device |
CN111879307A (en) * | 2020-06-22 | 2020-11-03 | 国网河北省电力有限公司信息通信分公司 | Vehicle path planning method based on vehicle body parameters and engineering construction information |
CN112461240A (en) * | 2020-11-11 | 2021-03-09 | 武汉理工大学 | Unmanned aerial vehicle obstacle avoidance path planning method and system |
CN112462785B (en) * | 2020-12-04 | 2022-06-03 | 厦门大学 | Mobile robot path planning method and device and storage medium |
CN112462785A (en) * | 2020-12-04 | 2021-03-09 | 厦门大学 | Mobile robot path planning method and device and storage medium |
CN112632734A (en) * | 2020-12-29 | 2021-04-09 | 武汉中海庭数据技术有限公司 | High-precision map road connectivity test method and system |
WO2022152283A1 (en) * | 2021-01-15 | 2022-07-21 | 北京理工大学 | Local path planning method for unmanned vehicle |
CN112945254A (en) * | 2021-01-21 | 2021-06-11 | 西北工业大学 | Unmanned vehicle curvature continuous path planning method based on rapid expansion random tree |
CN113320547A (en) * | 2021-07-15 | 2021-08-31 | 广州小鹏自动驾驶科技有限公司 | Path detection method and device and automobile |
CN113320547B (en) * | 2021-07-15 | 2023-08-25 | 广州小鹏自动驾驶科技有限公司 | Path detection method and device and automobile |
CN114089742A (en) * | 2021-10-25 | 2022-02-25 | 广东嘉腾机器人自动化有限公司 | AGV traveling speed control method and device based on path curvature and medium |
CN114089742B (en) * | 2021-10-25 | 2023-09-26 | 广东嘉腾机器人自动化有限公司 | AGV running speed control method, device and medium based on path curvature |
CN117533354A (en) * | 2023-12-28 | 2024-02-09 | 安徽蔚来智驾科技有限公司 | Track generation method, driving control method, storage medium and intelligent device |
CN117533354B (en) * | 2023-12-28 | 2024-04-02 | 安徽蔚来智驾科技有限公司 | Track generation method, driving control method, storage medium and intelligent device |
Also Published As
Publication number | Publication date |
---|---|
CN109540159B (en) | 2020-11-27 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109540159A (en) | A kind of quick complete automatic Pilot method for planning track | |
CN110136481B (en) | Parking strategy based on deep reinforcement learning | |
CN110297494B (en) | Decision-making method and system for lane change of automatic driving vehicle based on rolling game | |
Zhang et al. | Reinforcement learning-based motion planning for automatic parking system | |
KR102306939B1 (en) | Method and device for short-term path planning of autonomous driving through information fusion by using v2x communication and image processing | |
AU2005325706B2 (en) | Point-to-point path planning | |
Zhao et al. | Dynamic motion planning for autonomous vehicle in unknown environments | |
CN107702716A (en) | A kind of unmanned paths planning method, system and device | |
CN108820039A (en) | A kind of automatic driving vehicle bend crosswise joint system and method | |
CN109579854B (en) | Unmanned vehicle obstacle avoidance method based on fast expansion random tree | |
US11499834B2 (en) | Aligning road information for navigation | |
CN108256233A (en) | Intelligent vehicle trajectory planning and tracking and system based on driver style | |
CN107063280A (en) | A kind of intelligent vehicle path planning system and method based on control sampling | |
CN110345957A (en) | Vehicle route identification | |
CN109324620A (en) | The dynamic trajectory planing method for carrying out avoidance based on lane line parallel offset and overtaking other vehicles | |
CN104298239A (en) | Enhanced map learning path planning method for indoor mobile robot | |
CN110189547A (en) | A kind of obstacle detection method, device and vehicle | |
CN110254422B (en) | Automobile obstacle avoidance method based on multi-target reinforcement learning and Bezier curve | |
CN112947406A (en) | FLOYD and Astar-based hybrid path planning method | |
CN113701780B (en) | Real-time obstacle avoidance planning method based on A star algorithm | |
CN112829747A (en) | Driving behavior decision method and device and storage medium | |
CN112683290A (en) | Vehicle track planning method, electronic equipment and computer readable storage medium | |
CN114564016A (en) | Navigation obstacle avoidance control method, system and model combining path planning and reinforcement learning | |
CN109283843A (en) | A kind of lane-change method for planning track merged based on multinomial with particle swarm algorithm | |
CN110509923A (en) | Automatic Pilot paths planning method, system, computer readable storage medium and vehicle |
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 |