CN109733395B - Automatic driving automobile transverse coordination control method based on extendability evaluation - Google Patents

Automatic driving automobile transverse coordination control method based on extendability evaluation Download PDF

Info

Publication number
CN109733395B
CN109733395B CN201811555640.3A CN201811555640A CN109733395B CN 109733395 B CN109733395 B CN 109733395B CN 201811555640 A CN201811555640 A CN 201811555640A CN 109733395 B CN109733395 B CN 109733395B
Authority
CN
China
Prior art keywords
vehicle
road
deviation
center line
delta
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.)
Active
Application number
CN201811555640.3A
Other languages
Chinese (zh)
Other versions
CN109733395A (en
Inventor
蔡英凤
秦顺琪
臧勇
孙晓强
陈特
蔡骏宇
陈龙
江浩斌
唐斌
徐兴
何友国
袁朝春
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.)
Jiangsu University
Original Assignee
Jiangsu University
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 Jiangsu University filed Critical Jiangsu University
Priority to CN201811555640.3A priority Critical patent/CN109733395B/en
Publication of CN109733395A publication Critical patent/CN109733395A/en
Application granted granted Critical
Publication of CN109733395B publication Critical patent/CN109733395B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Landscapes

  • Control Of Driving Devices And Active Controlling Of Vehicle (AREA)
  • Steering Control In Accordance With Driving Conditions (AREA)

Abstract

The invention discloses an automatic driving automobile transverse coordination control method based on the evaluation of the extendability degree, which comprises the following steps of: obtaining the lateral of a vehicle to a road centerlineDeviation to position epHeading deviation of vehicle and road center line
Figure DDA0001911812980000011
Yaw rate of vehicle
Figure DDA0001911812980000012
And the road center line curvature rho is respectively input into the PID feedback controller and the PID feedforward-feedback controller, and the front wheel turning angle delta of the vehicle is correspondingly outputf1And deltaf2Substituting the output value into an evaluation index pre-estimation model to obtain two groups of corresponding vehicle-road pre-estimation state quantities, evaluating the vehicle-road pre-estimation state quantities by adopting a goodness evaluation method, selecting a result with high goodness, and inputting a corresponding front wheel steering angle as an output value into a lower layer controller; the lower layer controller passes the transverse position deviation e of the current vehicle and the road center linepCalculating a correlation function K (S) by using the curvature rho of the center line of the road, and calculating delta by using the correlation function K (S)fAnd coordinating, and outputting the coordinated front wheel rotation angle to a vehicle-road state space equation to obtain the actual front wheel rotation angle.

Description

Automatic driving automobile transverse coordination control method based on extendability evaluation
Technical Field
The invention belongs to the technical field of an automatic driving automobile control system, and particularly relates to a method for transversely coordinating and controlling an automatic driving automobile based on the evaluation of the extendability degree.
Background
As a main research content in the field of intelligent traffic control, intelligent vehicles integrate a variety of modern electronic information technologies. With the increasing demand of the modern society for intellectualization and safety of modern vehicles, intelligent vehicles become a hot problem and a leading edge of technology for competitive research in the traffic field in various countries in the world. Research in the lateral control of vehicles (lane keeping, vehicle lane changing) has received increasing attention as part of the field of intelligent vehicle research.
The vehicle transverse control means that the transverse position deviation and the heading deviation are enabled to be as small as possible through algorithm decision and the action of a bottom layer execution device during the process that the vehicle runs along the expected path, and meanwhile, the vehicle has certain stability and running safety. The currently used transverse control algorithm mainly comprises feedback control, feedforward-feedback control, fuzzy control, a sliding mode control strategy, a single-point preview strategy, model prediction control, optimal control and the like. However, the above control method has many limitations, or the control effect is better under a specific working condition, and the overall control effect is not good under a mixed complex working condition.
The vehicle control is a multi-input multi-output system, and the problem of road environment where the vehicle is located needs to be considered in the vehicle lateral control, so that the control method has many limitations. The method is mainly used for solving the problem that a single control strategy is limited in control area, and based on an extension theory proposed by Chua, the generated goodness evaluation method is a basic method for evaluating the quality of an object including things, strategies, methods and the like in the extension theory. According to the requirements of practical problems, an evaluation standard meeting technical requirements is made, a measuring index is determined, processes of profit and disadvantage and positive and negative change conditions of the processes are reflected, and therefore the quality of an object is evaluated. According to different environments, different control strategies are adopted, and a better control strategy is pertinently adopted according to the goodness evaluation method, so that the automatic driving automobile transverse coordination control method based on the extendability evaluation is designed, and the whole control process can obtain a good control effect.
Disclosure of Invention
An automatic driving automobile transverse coordination control method based on the extendibility evaluation comprises the following two parts: an upper layer controller and a lower layer controller. The upper layer controller selects the controller output value with high goodness through a goodness evaluation method, and the lower layer controller performs coordinated output based on the relevance. The method specifically comprises the following steps:
in the upper controller: obtaining the state quantity of the vehicle and the road, and the transverse position deviation e of the vehicle and the center line of the roadpHeading deviation of vehicle and road center line
Figure BDA0001911812960000011
Yaw rate of vehicle
Figure BDA0001911812960000012
And the road center line curvature rho is respectively input into the PID feedback controller and the PID feedforward-feedback controller as an input value, and the front wheel turning angle delta of the vehicle is correspondingly outputf1And deltaf2Respectively substituting the output values into an evaluation index pre-estimation model to obtain two groups of corresponding vehicle-road pre-estimation state quantities, evaluating the two groups of vehicle-road pre-estimation state quantities by adopting a goodness evaluation method, preferably selecting a result with high goodness, and inputting the output front wheel steering angle with high goodness into a lower layer controller as the output value of an upper layer controller; in the lower level controller: by the transverse positional deviation e of the current vehicle from the road centre linepCalculating a correlation function K (S) according to the curvature rho of the center line of the road, designing a coordination output controller based on the correlation function K (S), and outputting a front wheel corner delta from an upper layer controllerfThe vehicle-road state space equation is established according to a two-degree-of-freedom dynamic model of the vehicle and a path tracking preview deviation model and is used as a control object to verify the effectiveness of the control method.
Further, the vehicle-road state space equation is established according to a vehicle two-degree-of-freedom dynamic model and a path tracking preview deviation model:
Figure BDA0001911812960000021
wherein:
Figure BDA0001911812960000022
further, the evaluation index estimation model is converted from a state space equation:
Figure BDA0001911812960000031
further, the goodness evaluation method includes: firstly, a measure index is determined according to the vehicle-road state
Figure BDA0001911812960000032
And the weight coefficient α of (0,1) is assigned to (α) according to the importance degree of each measuring index1，α2，α3，α4) Wherein, α1+α2+α3+α 41 is ═ 1; secondly, establishing a correlation function K according to the requirements of each measurement index1(x1)，K2(x2)，K3(x3)，K4(x4) Handle the object deltaf1，δf2For each measure MIiIs abbreviated as Ki(δfj) Then each object is deltaf1，δf2About MIiHas a degree of association of Ki＝(Ki(δf1)，Ki(δf2) 1,2,3,4. normalize the above-mentioned association:
Figure BDA0001911812960000033
when i is 1,2,3,4 and j is 1,2, the objects δ are all the samef1，δf2About MIiHas a normalized degree of association of ki＝(ki1，ki2) I is 1,2,3, 4; and then a goodness calculation is performed.
Further, the goodness calculation: let a calculation object ZjFor each measure MI1，MI2，MI3，MI4Has a normalized degree of association of
Figure BDA0001911812960000034
Then object Z is calculatedjThe goodness of the method is as follows:
Figure BDA0001911812960000035
for deltafjComparing the goodness of (A): if it is
Figure BDA0001911812960000036
Then the object deltaf0Is more preferable.
Further, when calculating the correlation function K (S), the characteristic quantity selects the transverse preview deviation epAnd the road center line curvature rho form a characteristic state S (e)p,ρ)。
Further, when the correlation function k(s) is calculated, the interval of the classical domain divided by the two-dimensional extension set is: lateral position deviation ep[-ep1,ep1]Road center line curvature p < - > p1,ρ1](ii) a The extension region interval is: lateral position deviation ep[-ep2,-ep1)∪(ep1,ep2]Road center line curvature p < - > p2,ρ1)∪(ρ1,ρ2](ii) a The non-domain interval is: lateral position deviation ep(-∞,-ep2)∪(ep2, + ∞), road centerline curvature ρ (-infinity, - ρ -2)∪(ρ2,+∞)。
Further, the calculation of the correlation function k(s) requires first converting the extension distances in the two-dimensional extension set into one-dimensional extension distances.
Further, the correlation function is:
Figure BDA0001911812960000041
whereinD(P3,<P5,P2>,<P4,P1>)＝ρ(P3,<P5,P2>)-ρ(P3,<P4,P1>)，ρ(P3,<P4,P1>) And ρ (P)3,<P5,P2>) Are respectively a point P3Extension to classical and extension domains, P3Characteristic quantity (e) of vehicle-road model collected during vehicle drivingpρ) real-time state value, P3Connecting with the origin to intersect the boundary of the classical domain at point P1、P4Intersecting the extension domain boundary at point P2、P5。
Further, the coordinated output controller takes 1-K (S) and K (S) as the output value delta of the previous secondf(t-1)And the current output value deltaf(t)The final output value of the weighting coefficient of (2) is: deltaf＝[1-K(S)]δf(t-1)+K(S)δf(t)
The invention has the beneficial effects that:
(1) determining the degree of the control output quantity which meets the requirements of the measurement indexes through a goodness evaluation method, and calculating the comprehensive goodness of each control output quantity to be evaluated so as to judge the goodness of the control output quantity to be evaluated, so that the unmanned vehicle can transversely control the object: the state quantity of the vehicle-road system is converted from the uncontrollable area to the stable controllable area, so that the restriction of the factors such as vehicle dynamics-road model complexity or limited use condition of the control method is eliminated, and the capabilities of transverse control and track tracking control of the unmanned vehicle are improved.
(2) According to the invention, by means of extension control, the unmanned vehicle can realize optimal control by a goodness evaluation method in two controllers under the complex mixed working condition that the unmanned vehicle passes through a road with small curvature and a curve with large curvature and large corner, so that the whole control process obtains smaller control error.
(3) The invention can obviously improve the shaking condition of controller switching by the coordination output controller based on the weighting coefficient of the correlation function K (S), so that the whole control process is more stable, and the comfort is higher.
Drawings
FIG. 1 is a flow chart of a lateral coordination control method for an autonomous vehicle based on an evaluation of extendability
FIG. 2 is a flow chart of PID feedback control;
FIG. 3 is a flow chart of PID feedforward-feedback control;
FIG. 4 is a two degree of freedom vehicle dynamics model diagram;
FIG. 5 is a diagram of a lateral preview deviation model;
FIG. 6 is a model diagram of course deviation;
FIG. 7 is a flowchart of a goodness evaluation method
FIG. 8 is a two-dimensional extent set diagram;
FIG. 9 is a diagram of a one-dimensional extension set;
Detailed Description
The invention will be further described in the following with reference to the description of the figures and the embodiments, without limiting the scope of the invention thereto.
Fig. 1 is a flow chart of an automatic driving automobile lateral coordination control method based on the evaluation of the extendibility, and the method comprises the following steps of designing an upper layer controller and a lower layer controller:
an upper layer controller: acquiring the vehicle-road state quantity collected by a vehicle sensor and the transverse position deviation e of the vehicle and the road center linepHeading deviation of vehicle and road center line
Figure BDA0001911812960000051
Yaw rate of vehicle
Figure BDA0001911812960000052
And road centerline curvature ρ. Respectively inputting the vehicle-road state quantity as an input value into a PID feedback controller and a PID feedforward-feedback controller, and correspondingly outputting the front wheel turning angle delta of the vehiclef1And deltaf2And the output values are respectively brought into the evaluation index pre-estimation model to obtain two groups of corresponding vehicle-road pre-estimation state quantities, and the two groups of vehicle-road pre-estimation state quantities are evaluated by adopting a goodness evaluation method,preferably, the higher-priority scheme is such that the controller outputs the front wheel steering angle as the output value of the upper controller and inputs the front wheel steering angle to the lower controller.
The lower layer controller: by the transverse positional deviation e of the current vehicle from the road centre linepCalculating the degree of association K (S) of the curvature rho of the center line of the road, designing a coordination output controller based on the degree of association K (S), and outputting a front wheel corner delta from an upper layer controllerfThe vehicle-road state space equation is established according to a two-degree-of-freedom dynamic model of the vehicle and a path tracking preview deviation model and is used as a control object to verify the effectiveness of the control method.
Fig. 2 shows a PID feedback control flow chart based on preview deviation, which is suitable for a small road curvature, and the lateral control implementation only needs to track an upper expected track under a small rotation angle, and mainly needs to solve the problem of minimizing the lateral position deviation and the lateral deviation as much as possible. Fig. 3 shows a control flow chart of a PID feedforward-feedback controller based on road curvature, which is suitable for a small turning radius of a vehicle, a large road curvature, and a need of a quick response of the vehicle, and ensures that the vehicle can turn wheels to a required turning angle in time, so that the front wheel turning angle of the vehicle quickly responds, the response speed of a control system to small curvature path interference is increased, and the hysteresis and the volatility of the vehicle direction control are reduced.
The present embodiment employs a two degree of freedom vehicle dynamics model that considers only lateral motion along the vehicle y-axis and yaw motion about the z-axis, and assumes the following:
(1) neglecting the air resistance suffered by the automobile;
(2) assuming that the road surface on which the automobile runs is horizontal, the road resistance of the ground to the automobile is 0;
(3) assuming that the steering angles of two front wheels and two rear wheels of the automobile are respectively equal and the camber angle of the wheels is 0;
(4) neglecting the action of the steering system in the steering process of the vehicle, directly taking the wheel rotation angle of the front wheel as the control input of the system;
(5) assuming that the suspension of the automobile is rigid, the body does not have up-and-down motion perpendicular to the ground, front-and-back pitching and body rolling, and the motion of the body is parallel to the ground;
(6) neglecting the left and right offset of the vehicle mass during the turning, the supporting forces of the ground in the vertical direction received by the left and right wheels are considered to be equal.
Fig. 4 is a schematic diagram of a two-degree-of-freedom dynamic model of a vehicle. The resultant force sigma F along the y axis can be obtained according to the Newton's second law theoremy,iEquilibrium equation and resultant moment sigma M around z-axiszThe equilibrium equation:
Figure BDA0001911812960000061
wherein m is the vehicle mass (kg); v. ofx、vyThe longitudinal speed and the transverse speed (m.s) of the vehicle-1)；
Figure BDA0001911812960000062
Is the vehicle heading angle (rad); deltafIs the vehicle front wheel corner (rad); i iszIs the moment of inertia (kg · m) of the vehicle around the z-axis2) (ii) a a. b is the distance between the front axle and the rear axle of the vehicle respectively; k is a radical of1、k2Lateral deflection stiffness (N. rad) of front and rear wheel tires, respectively-1)；ωrYaw rate (rad · s) for vehicle-1)。
In the track tracking process of the vehicle, the motion process comprises the translational motion and the rotational motion of the vehicle, and the path tracking preview deviation model comprises a transverse preview deviation model and a heading deviation model, which are respectively shown in fig. 3 and fig. 4.
In fig. 5, A, B are two front and rear end points of the vehicle, point C is the center of mass of the vehicle, point D is the preview point, points a and b are the front axle distance and the rear axle distance (m) of the vehicle, L is the preview distance (m), epIs the transverse deviation (m) of the preview point, e is the transverse position deviation (m) at the mass center of the vehicle,
Figure BDA0001911812960000063
for vehicle and road centreCourse deviation (rad), c of the line1、c2Respectively, the lateral position deviation (m) between the vehicle front end and the desired path.
From the geometric relationship, one can obtain:
Figure BDA0001911812960000071
then
Figure BDA0001911812960000072
From the graph can also be derived:
Figure BDA0001911812960000073
then
Figure BDA0001911812960000074
Namely, it is
Figure BDA0001911812960000075
Due to the fact that
Figure BDA0001911812960000076
Therefore, it is not only easy to use
Figure BDA0001911812960000077
To epThe derivation can be:
Figure BDA0001911812960000078
in the context of figure 6, it is shown,
Figure BDA0001911812960000079
the course deviation of the vehicle and the road center line;
Figure BDA00019118129600000710
is the vehicle heading angle;
Figure BDA00019118129600000711
is the included angle between the tangent line of the center line of the road and the abscissa of the ground. From the geometrical relationship it can be found that:
Figure BDA00019118129600000712
Figure BDA00019118129600000713
wherein the longitudinal speed v of the vehiclex(m·s-1) Constant, road center line curvature ρ (m)-1) Is the reciprocal of the radius of the road centerline circle, a known quantity.
Thus is provided with
Figure BDA00019118129600000714
The change rate of the transverse position deviation between the centroid position and the road center line in the vehicle track tracking process is as follows:
Figure BDA00019118129600000715
due to the fact that
Figure BDA0001911812960000081
Therefore, it is not only easy to use
Figure BDA0001911812960000082
By substituting formula (13) for formula (8)
Figure BDA0001911812960000083
Simplified lateral speed v of vehicle state quantityyLateral direction of the moving bodyAcceleration of a vehicle
Figure BDA0001911812960000084
Yaw rate of vehicle
Figure BDA0001911812960000085
Yaw angular acceleration of vehicle
Figure BDA0001911812960000086
And the relation with the parameters in the path tracking preview deviation model, namely the path tracking preview deviation model:
Figure BDA0001911812960000087
combining the path tracking preview deviation model with the vehicle two-degree-of-freedom dynamic model to form a vehicle-road preview deviation model, and selecting ep、
Figure BDA0001911812960000088
For the state quantities of the state space equation, the state space equation of the vehicle-road model can be obtained
Figure BDA0001911812960000089
Figure BDA00019118129600000810
Wherein:
Figure BDA0001911812960000091
the estimation model of the evaluation index obtained from the formulas (4), (5), (10) and (15) is as follows
Figure BDA0001911812960000092
The goodness evaluation method is mainly used for selecting the controller output value with higher goodness, and as shown in a flow chart of the goodness evaluation method in fig. 7, the goodness evaluation method specifically comprises the following steps:
1) determining metrics
To evaluate the quality of an object, a measure must first be specified. The quality is relative to a certain standard. An object may be advantageous with respect to certain metrics and may be disadvantageous with respect to other metrics. Thus, assessing the goodness of an object must reflect the degree of pros and cons and their possible variations. The method needs to work out an evaluation standard meeting the requirement according to the requirement of the actual problem and determine a measurement index
Figure BDA0001911812960000093
2) Determining a weight coefficient
Evaluating a controller output value deltafj(j-1, 2) excellence or disadvantage
Figure BDA0001911812960000094
Each measure MIiThe weight is given to each of the metrics, and the degree of importance of each metric is expressed by a weight coefficient, and the value of (0,1) is given to each metric. The weight coefficient is recorded as
α＝(α1，α2，α3，α4) (18)
Wherein, α1+α2+α3+α4＝1.
3) Establishing a correlation function and calculating the correlation degree
Measurement index
Figure BDA0001911812960000101
The weight coefficient is assigned α ═ α1，α2，α3，α4) According to the requirements of various metrics, the object delta is setf1，δf2For each measure MIiIs abbreviated as Ki(δfj) Then each object is deltaf1，δf2About MIiHas a degree of association of
Ki＝(Ki(δf1)，Ki(δf2))，i＝1,2,3,4. (19)
Normalizing the relevance:
Figure BDA0001911812960000102
then each object deltaf1，δf2About MIiHas a normalized degree of association of ki＝(ki1，ki2)，i＝1,2,3,4 (21)
4) Calculate goodness
Object ZjFor each measure MI1，MI2，MI3，MI4Has a normalized degree of association of
Figure BDA0001911812960000103
Object ZjHas a goodness of
Figure BDA0001911812960000104
5) Scheme with higher selection goodness
For deltafjComparing the goodness of (A): if it is
Figure BDA0001911812960000105
Then the object deltaf0Is preferably, deltaf0It is the value that the upper level controller inputs to the lower level controller.
The goodness evaluation table is shown in the table:
Figure BDA0001911812960000111
because the output value of the controller selected by the goodness evaluation method in the last second is not the same as the currently selected controller, the output values of the two times of the optimization are probably greatly different, the steering of the front wheels is not coordinated, and the shaking condition can occur, so that the controller is coordinated to output, and the whole vehicle-road system is more stable, 1-K (S), K (S) are adopted as the output value delta in the last secondf(t-1)And the current output value deltaf(t)The final output value of the weighting coefficient of (2) is:
δf＝[1-K(S)]δf(t-1)+K(S)δf(t)(24)
the calculation procedure for the correlation function k(s) is as follows:
1) feature quantity extraction
The main index for evaluating the transverse control of the automatic driving vehicle is the position deviation between the vehicle and the central line of the road, the design of the whole control system strategy is closely related to the road condition, the index reflecting the basic road condition is the curvature of the road, in addition, the transverse control of the automatic driving vehicle is established on the parameter of the curvature of the road by adopting the basis of feedforward-feedback control, therefore, the characteristic quantity selects the transverse preview deviation epAnd the road center line curvature rho form a characteristic state S (e)p,ρ)。
2) Extension set partitioning
As shown in fig. 7, a two-dimensional extension set is established, and the extension domain preview lateral deviation e is determinedpMaximum allowable range of [ -e [ - ]p2,ep2]Maximum allowable range (-rho) of curvature rho of road centerline in extended area2,ρ2]. Aiming at the classical domain, because the adopted control strategy is the traditional PID feedback control and has limited control capability, when a vehicle tracks the central line track of a road with a large corner, the satisfactory control effect cannot be achieved, and therefore, the lateral deviation e of the classical domainpMaximum allowable range [ -e [ ]p1,ep1]The maximum allowable range of the road curvature ρ is [ - ρ [ ]1,ρ1]. Therefore, in the two-dimensional extension set, the interval of the classical domain is: lateral position deviation ep[-ep1,ep1]Road center line curvature p < - > p1,ρ1](ii) a The extension region interval is: lateral position deviation ep[-ep2,-ep1)∪(ep1,ep2]Road center line curvature p < - > p2,ρ1)∪(ρ1,ρ2](ii) a The non-domain interval is: lateral position deviation ep(-∞,-ep2)∪(ep2, + ∞), road centerline curvature ρ (-infinity, - ρ -2)∪(ρ2,+∞)。
3) Calculation of correlation function
The characteristic quantities selected by the traditional extension controller are deviation and deviation differential, and the embodiment selects the preview transverse deviation epAnd the road curvature p is used as a characteristic quantity, and an extension distance is calculated to determine a correlation function value, and a two-dimensional extension set is shown in fig. 8.
In the two-dimensional extension set, the origin (0,0) is the optimal point of the feature state. Suppose there is a point P in the extensibility domain3，P3Connecting the origin with P for the current state of the vehicle-road system3Point, obtain P3Shortest distance | OP toward optimum point (0,0)3L. The straight line of the line segment intersects with the boundary of the classical domain at P1And P4Point, cross extension domain boundary at P2And P5And (4) point. In guarantee of P3Under the precondition that the distance approaching the origin is shortest, P can be determined according to the intersection points3Closest distance to the extendible domain, the classical domain.
In the one-dimensional extension set, the extension distance is substantially the minimum distance from a point to a boundary of an interval, and the extension distance in the two-dimensional extension set can be converted into the one-dimensional extension distance according to the principle, as shown in fig. 9. P3The extension distances from the point to the classical domain and the extension domain are respectively rho (P)3,<P4,P1>) And ρ (P)3,<P5,P2>) The solution is as follows:
Figure BDA0001911812960000121
Figure BDA0001911812960000122
the correlation function may then be determined as:
Figure BDA0001911812960000123
wherein: d (P)3,<P5,P2>,<P4,P1>)＝ρ(P3,<P5,P2>)-ρ(P3,<P4,P1>)。
The above-listed detailed description is only a specific description of a possible embodiment of the present invention, and they are not intended to limit the scope of the present invention, and equivalent embodiments or modifications made without departing from the technical spirit of the present invention should be included in the scope of the present invention.

Claims (9)

1. An automatic driving automobile transverse coordination control method based on the evaluation of the extendability degree is characterized in that an upper layer controller and a lower layer controller are designed; the upper-layer controller selects a vehicle front wheel steering angle output value with high goodness through a goodness evaluation method, the lower-layer controller coordinates the front wheel steering angle output by the upper-layer controller based on the relevance degree, and outputs the front wheel steering angle after coordination to a control object;
an upper layer controller: obtaining the state quantity of the vehicle and the road, and the transverse position deviation e of the vehicle and the center line of the roadpHeading deviation of vehicle and road center line
Figure FDA0002376620530000011
Yaw rate of vehicle
Figure FDA0002376620530000012
And the road center line curvature rho is respectively input into the PID feedback controller and the PID feedforward-feedback controller as an input value, and the front wheel turning angle delta of the vehicle is correspondingly outputf1And deltaf2And respectively substituting the output values into an evaluation index pre-estimation model to obtain two groups of corresponding vehicle-road pre-estimation state quantities, evaluating the two groups of vehicle-road pre-estimation state quantities by adopting a goodness evaluation method, selecting a result with a high goodness value, and inputting the corresponding front wheel steering angle as the output value of the upper layer controller into the lower layer controller.
2. The automatic driving automobile transverse coordination control method based on the popularity evaluation as claimed in claim 1, characterized in that the control method is implemented specifically as follows:
the lower layer controller:by the transverse positional deviation e of the current vehicle from the road centre linepAnd calculating a correlation function K (S) according to the curvature rho of the center line of the road, and utilizing the correlation function K (S) to output a front wheel corner delta from an upper layer controllerfAnd coordinating, and outputting the front wheel rotation angle after coordination to a vehicle-road state space equation.
3. The automatic driving automobile transverse coordination control method based on the extendibility evaluation as claimed in claim 2, characterized in that the vehicle-road state space equation is established according to a vehicle two-degree-of-freedom dynamic model and a path tracking preview deviation model to obtain:
Figure FDA0002376620530000013
wherein:
Figure FDA0002376620530000021
m is vehicle mass (kg); v. ofx、vyThe longitudinal speed and the transverse speed (m.s) of the vehicle-1)；
Figure FDA0002376620530000022
Is the vehicle heading angle (rad); deltafIs the vehicle front wheel corner (rad); i iszIs the moment of inertia (kg · m) of the vehicle around the z-axis2) (ii) a a. b is the distance between the front axle and the rear axle of the vehicle respectively; k is a radical of1、k2Lateral deflection stiffness (N. rad) of front and rear wheel tires, respectively-1)；ωrYaw rate (rad · s) for vehicle-1)；
epIs the lateral deviation of the preview point,
Figure FDA0002376620530000023
is epThe first derivative of (a) is,
Figure FDA0002376620530000024
is epThe second derivative of (a) is,
Figure FDA0002376620530000025
is the heading deviation of the vehicle from the center line of the road,
Figure FDA0002376620530000026
the first derivative and the second derivative are respectively, and L is the pre-aiming distance.
4. The method of claim 3, wherein the two-degree-of-freedom dynamic model of the vehicle is:
Figure FDA0002376620530000031
wherein m is the vehicle mass (kg); v. ofx、vyThe longitudinal speed and the transverse speed (m.s) of the vehicle-1)；
Figure FDA0002376620530000032
Is the vehicle heading angle (rad); deltafIs the vehicle front wheel corner (rad); i iszIs the moment of inertia (kg · m) of the vehicle around the z-axis2) (ii) a a. b is the distance between the front axle and the rear axle of the vehicle respectively; k is a radical of1、k2Lateral deflection stiffness (N. rad) of front and rear wheel tires, respectively-1)；ωrYaw rate (rad · s) for vehicle-1)；
The path tracking preview deviation model:
Figure FDA0002376620530000033
wherein v isyLateral speed, v, being vehicle state quantityxThe longitudinal speed of the vehicle is set by the speed of the vehicle,
Figure FDA0002376620530000034
for vehicles and roadsCourse deviation of the road center line, L is the pre-aiming distance, rho is the curvature of the road center line,
Figure FDA0002376620530000035
is composed of
Figure FDA0002376620530000036
The second derivative of (a) is,
Figure FDA0002376620530000037
is v isyThe first derivative of (a).
5. The automatic driving automobile transverse coordination control method based on the extendability evaluation as claimed in claim 3, wherein the evaluation index estimation model is obtained by converting a vehicle-road state space equation:
Figure FDA0002376620530000038
Figure FDA0002376620530000039
is the angle between the tangent line of the center line of the road and the abscissa of the earth, c1、c2Respectively, the lateral position deviation between the front end of the vehicle and the desired path.
6. The method of claim 3, wherein the goodness evaluation method comprises:
firstly, a measure index is determined according to the vehicle-road state
Figure FDA0002376620530000041
And the weight coefficient α of (0,1) is assigned to (α) according to the importance degree of each measuring index1，α2，α3，α4) Wherein, α1+α2+α3+α4＝1；
Secondly, establishing a correlation function K according to the requirements of each measurement index1(x1)，K2(x2)，K3(x3)，K4(x4) Handle the object deltaf1，δf2For each measure MIiIs abbreviated as Ki(δfj) Then each object is deltaf1，δf2About MIiHas a degree of association of
Ki＝(Ki(δf1)，Ki(δf2) 1,2,3,4. normalize the above-mentioned association:
Figure FDA0002376620530000042
then each object deltaf1，δf2About MIiHas a normalized degree of association of ki＝(ki1，ki2) I is 1,2,3, 4; and then a goodness calculation is performed.
7. The method for lateral coordination control of an autonomous vehicle based on goodness evaluation as claimed in claim 6, wherein the goodness calculation method comprises:
let a calculation object ZjFor each measure MI1，MI2，MI3，MI4Has a normalized degree of association of
Figure FDA0002376620530000043
Goodness calculation object ZjThe goodness of the method is as follows:
Figure FDA0002376620530000044
for deltafjComparing the goodness of (A): if it is
Figure FDA0002376620530000045
Then the object deltaf0Is more preferable.
8. The method of claim 7, wherein the output-coordinated front wheel steering angle is:
δf＝[1-K(S)]δf(t-1)+K(S)δf(t)。
9. the method for laterally coordinating and controlling an autonomous vehicle according to claim 8, wherein the correlation function k(s) is calculated as follows:
1) feature quantity extraction
Selecting the characteristic quantity as a transverse preview deviation epAnd the road center line curvature rho form a characteristic state S (e)p,ρ)；
2) Extension set partitioning
Establishing a two-dimensional extension set and determining the extension domain preview transverse deviation epMaximum allowable range of [ -e [ - ]p2,ep2]Maximum allowable range (-rho) of curvature rho of road centerline in extended area2,ρ2]For the classical domain, the classical domain lateral deviation epMaximum allowable range [ -e [ ]p1,ep1]The maximum allowable range of the road curvature ρ is [ - ρ [ ]1,ρ1](ii) a Therefore, in the two-dimensional extension set, the interval of the classical domain is set as: lateral position deviation ep[-ep1,ep1]Road center line curvature p < - > p1,ρ1](ii) a The extension region interval is: lateral position deviation ep[-ep2,-ep1)∪(ep1,ep2]Road center line curvature p < - > p2,ρ1)∪(ρ1,ρ2](ii) a The non-domain interval is: lateral position deviation ep(-∞,-ep2)∪(ep2, + ∞), road centerline curvature ρ (-infinity, - ρ -2)∪(ρ2,+∞)；
3) Calculation of correlation function
Selecting a pre-aiming transverse deviation epAnd road curvature rho is used as a characteristic quantity, and an extension distance is calculated to determine a correlation function value; in the two-dimensional extension set, the origin (0,0) is TecOptimum points for the eigenstates, assuming that there is a point P in the extension field3，P3Connecting the origin with P for the current state of the vehicle-road system3Point, obtain P3The shortest distance to the optimum point (0,0) is the line segment | OP3The line where the line segment is intersected with the boundary of the classical domain at P1And P4Point, cross extension domain boundary at P2And P5Point; in guarantee of P3Under the precondition that the distance approaching the origin is shortest, P is determined by utilizing the intersection points3The closest distance to the extension domain and the classical domain;
in the one-dimensional extension set, the extension distances in the two-dimensional extension set are converted into one-dimensional extension distances, P3The extension distances from the point to the classical domain and the extension domain are respectively rho (P)3,<P4,P1>) And ρ (P)3,<P5,P2>) The solution is as follows:
Figure FDA0002376620530000051
Figure FDA0002376620530000061
the correlation function is then determined to be:
Figure FDA0002376620530000062
wherein: d (P)3,<P5,P2>,<P4,P1>)＝ρ(P3,<P5,P2>)-ρ(P3,<P4,P1>)。
CN201811555640.3A 2018-12-19 2018-12-19 Automatic driving automobile transverse coordination control method based on extendability evaluation Active CN109733395B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201811555640.3A CN109733395B (en) 2018-12-19 2018-12-19 Automatic driving automobile transverse coordination control method based on extendability evaluation

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201811555640.3A CN109733395B (en) 2018-12-19 2018-12-19 Automatic driving automobile transverse coordination control method based on extendability evaluation

Publications (2)

Publication Number Publication Date
CN109733395A CN109733395A (en) 2019-05-10
CN109733395B true CN109733395B (en) 2020-06-09

Family

ID=66360629

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201811555640.3A Active CN109733395B (en) 2018-12-19 2018-12-19 Automatic driving automobile transverse coordination control method based on extendability evaluation

Country Status (1)

Country Link
CN (1) CN109733395B (en)

Families Citing this family (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110262509B (en) * 2019-07-10 2022-06-28 百度在线网络技术（北京）有限公司 Automatic vehicle driving method and device
CN110733505A (en) * 2019-10-18 2020-01-31 上海格陆博实业有限公司 Control strategy of automobile lane keeping control systems
CN110615003B (en) * 2019-10-30 2020-09-18 吉林大学 Cruise control system based on strategy gradient online learning algorithm and design method
CN113104037B (en) * 2019-12-24 2022-08-30 浙江吉利汽车研究院有限公司 Method and system for determining steering angle of vehicle steering wheel
CN111674386B (en) * 2020-05-15 2021-06-15 武汉科技大学 Vehicle coordination control method and device based on AFS and DYC and storage medium
CN111806445A (en) * 2020-05-29 2020-10-23 北汽福田汽车股份有限公司 Vehicle transverse control method and device, medium, equipment and vehicle
CN111806444A (en) * 2020-05-29 2020-10-23 北汽福田汽车股份有限公司 Vehicle transverse control method and device, medium, equipment and vehicle
CN111703436B (en) * 2020-08-18 2020-11-24 北京主线科技有限公司 Control method and device for automatically driving vehicle
CN112092825B (en) * 2020-08-31 2022-04-29 英博超算(南京)科技有限公司 Lane keeping method based on machine learning
CN112537303B (en) * 2020-12-14 2022-02-18 英博超算(南京)科技有限公司 Intelligent vehicle lane centering keeping method
CN113032880B (en) * 2021-03-26 2022-06-14 武汉理工大学 Method for evaluating design rationality of circular curve road section

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102360365A (en) * 2011-10-08 2012-02-22 广东工业大学 Method for generating extension strategy for screening clothing on internet
JP6035207B2 (en) * 2013-06-14 2016-11-30 日立オートモティブシステムズ株式会社 Vehicle control system
DE102016221574A1 (en) * 2016-11-03 2018-05-03 Bayerische Motoren Werke Aktiengesellschaft A vehicle system and method for automatically performing a lane-change maneuver with functionality for determining the release of a temporary page strip
CN107600176B (en) * 2017-08-29 2019-06-28 江苏大学 A kind of intelligent vehicle active steering control method based on extension control theory
CN108216231B (en) * 2018-01-12 2019-07-19 合肥工业大学 One kind can open up united deviation auxiliary control method based on steering and braking

Also Published As

Publication number Publication date
CN109733395A (en) 2019-05-10

Similar Documents

Publication Publication Date Title
CN109733395B (en) Automatic driving automobile transverse coordination control method based on extendability evaluation
CN108732921B (en) Transverse extension preview switching control method for automatic driving automobile
CN109131325B (en) Lane keeping control method for three-dimensional extension preview switching of intelligent driving automobile
CN111610780B (en) Automatic driving vehicle path tracking control method and device
CN110001637B (en) Unmanned vehicle path tracking control device and control method based on multipoint tracking
CN112733270B (en) System and method for predicting vehicle running track and evaluating risk degree of track deviation
CN105329238B (en) A kind of autonomous driving vehicle lane-change control method based on monocular vision
Dominguez et al. Comparison of lateral controllers for autonomous vehicle: Experimental results
CN109606368B (en) Intelligent automobile extension vehicle speed self-adaptive change trajectory tracking control method
CN110036353A (en) For the self-adaptation control method and system in the surface car of trace, especially in automatic Pilot scene
CN109606363B (en) Multi-state feedback intelligent automobile extension lane keeping control method
CN105867377A (en) Automatic navigation control method of agricultural machine
CN109664884B (en) Extension self-adaptive lane keeping control method under variable vehicle speed
CN108919837B (en) Second-order sliding mode control method of automatic driving vehicle based on visual dynamics
CN111717207A (en) Cooperative steering control method considering human-vehicle conflict
CN110641465A (en) Lane keeping system and method based on vehicle speed
CN113291323B (en) Automatic driving path tracking control method and system for rhombic vehicle and vehicle
CN112109732A (en) Intelligent driving self-adaptive curve pre-aiming method
CN109292018A (en) Four-wheel steering Trajectory Tracking Control method based on coaxial-type wheel leg structure
CN113211438B (en) Wheel type robot control method and system based on pretightening distance self-adaption
CN107992039B (en) Trajectory planning method based on flow field in dynamic environment
CN116198522A (en) Unmanned mining card transverse and vertical coupling hierarchical control method for complex mining area working conditions
Shen et al. Steering control strategy guide by two preview vision cues
CN110362088A (en) A kind of tracking control system and method suitable for unmanned straddle carrier
CN115963836A (en) Path tracking and vehicle body posture cooperative control method

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