PATH-CONTROLLING MODULE, ASSOCIATED PATH-CONTROLLING DEVICE AND ASSOCIATED METHOD

Abstract

A motor-vehicle path-controlling module is arranged to model the path of the vehicle during a change in traffic lane by a Bezier curve relating a value of a parameter to a value of a lateral deviation of the vehicle from the center of a traffic lane and to a value of a time-dependent variable representative of the variation in the change of path; determine a setpoint state vector of a closed feedback loop of a path-controlling device, the loop being designed to control the motor vehicle so that it follows the path modelled by the Bezier curve, the vector being determined on the basis of the lateral deviation, of the time-dependent variable and of the parameter, and transmit the setpoint state vector to the input of the loop.

Claims

1-10. (canceled)

11. A path-control method for a motor vehicle, comprising: modeling a path of the vehicle during a lane change using a Bézier curve linking a value of a parameter to a value of a lateral deviation of the vehicle with respect to a center of a lane and to a value of a time-dependent variable representative of a variation in a change in the path; determining a setpoint state vector of a path-control device based on the lateral deviation, of the time-dependent variable and of the parameter, and steering the vehicle using the path-control device based on the state variable so that the vehicle follows the modeled path.

12. The method as claimed in claim 11, wherein the determining of the setpoint state vector comprises: a) determining a value of the parameter so that each increment of a time counter of the path-control device is comprised between two consecutive time-dependent variables; b) determining two consecutive lateral deviations based on the value of the parameter; c) determining an intermediate lateral deviation by interpolating between the two consecutive lateral deviations; and d) determining setpoint values comprising a setpoint lateral deviation, a setpoint yaw velocity, a setpoint heading angle, and a setpoint lateral velocity, the setpoint values being determined based on the intermediate lateral deviation, of the two consecutive time-dependent variables and of the two consecutive lateral deviations, wherein the setpoint state vector comprises the setpoint lateral deviation and the setpoint values.

13. The method as claimed in claim 12, further comprising determining a corrective radius of curvature based on the two consecutive time-dependent variables, of the two consecutive lateral deviations and of the longitudinal velocity of the vehicle, and transmitting the corrective radius of curvature to the path-control device.

14. The method as claimed in claim 13, wherein when the change in the path is triggered, the time counter is started, and when the value reached by the counter is greater than or equal to the value of the time-dependent variable associated with the control point of the Bézier curve comprising the time-dependent variable of greatest value, the counter is stopped to indicate the end of the change in the path.

15. The method as claimed in claim 14, comprising repeating iterations of steps a), b), c), d) and repeating iterations of the determining of the corrective radius of curvature for each increment of the counter, two consecutive increments being separated by a constant duration.

16. A path-control module for a motor vehicle, the module being configured to: model a path of the vehicle during a lane change using a Bézier curve linking a value of a parameter to a value of a lateral deviation of the vehicle with respect to a center of a lane and to a value of a time-dependent variable representative of a variation in a change in the path; determine a setpoint state vector of a closed control loop of a path-control device, the loop being configured to steer the motor vehicle so that the motor vehicle follows the path modeled by the Bézier curve, the vector being determined based on the lateral deviation, of the time-dependent variable and of the parameter, and transmit the setpoint state vector as input to the loop.

17. The module as claimed in claim 16, the module being configured to: e) determine a value of the parameter so that each increment instant of a time counter of the path-control device is comprised between two consecutive time-dependent variables; f) determine two consecutive lateral deviations based on the value of the parameter; g) determine an intermediate lateral deviation by interpolating between the two consecutive lateral deviations; and h) determine setpoint values comprising a setpoint lateral deviation, a setpoint yaw velocity, a setpoint heading angle, and a setpoint lateral velocity, the setpoint values being determined based on the intermediate lateral deviation, of the two consecutive time-dependent variables and of the two consecutive lateral deviations, wherein the setpoint state vector comprises the setpoint lateral deviation and the setpoint values.

18. The module as claimed in claim 17, the module being further configured to determine a corrective radius of curvature based on the two consecutive time-dependent variables, of the two consecutive lateral deviations and of the longitudinal velocity of the vehicle, and to transmit the corrective radius of curvature to the path-control device.

19. The module as claimed in claim 16, the module being configured to start the time counter when a change in path is triggered and to stop the counter to indicate the end of the change in path when the value reached by the counter is greater than or equal to the value of the time-dependent variable associated with the control point of the Bézier curve comprising the time-dependent variable of greatest value.

20. A path-control device comprising: the module as claimed in claim 16; and a closed control loop configured to receive as a setpoint the setpoint state vector determined by the path-control module, the control loop being further configured to steer the motor vehicle so that the motor vehicle follows the path modeled by the Bézier curve.

Description

[0058] Further objects, features and advantages of the invention will become apparent from reading the following description, given solely by way of nonlimiting example, and made with reference to the attached drawings in which:

[0059] FIG. 1 which has already been mentioned, schematically illustrates a device for the real-time control of the path of a vehicle according to the prior art;

[0060] FIG. 2 schematically illustrates one embodiment of a vehicle path-control device according to the invention;

[0061] FIG. 3 and

[0062] FIG. 4 illustrate an example of the modeling of a path of the vehicle 1 according to the invention;

[0063] FIG. 5 illustrates one implementation of the vehicle path-control device according to the invention; and

[0064] FIG. 6 depicts the variation with respect to time of the setpoint lateral deviation, of the lateral deviation and of the angle of the wheels following action of the control device according to the invention.

[0065] One embodiment of a path-control device 5 for the vehicle 1 according to one aspect of the invention has been schematically depicted in FIG. 2. Those elements of the device 5 that are identical to the elements that make up the device DISP of FIG. 1 bear the same references.

[0066] Once again, there are the closed control loop Br comprising the controller device 2 for generating the control signal Ust, the observer 3 for the real-time generation of the estimated state vector {circumflex over (ξ)} for the following of a straight lane by the vehicle on the basis of the current measurements vector η and the predictor module 4 for generating the second steering lock command Uff.

[0067] The command Ust is equal to the value of an angle δ.sub.rect for the following of a straight rectilinear path.

[0068] The device 5 further comprises a path-control module 6 generating the setpoint state vector during a lane change.

[0069] The control module 6 models the path of the vehicle 1 during a lane change using a Bézier curve P linking a value of a parameter S to a value of the lateral deviation yl of the vehicle with respect to the center of the vehicle's current lane and to a value of a time-dependent variable x representative of the variation in change of path. The control module 6 determines the setpoint state vector ξ* on the basis of the lateral deviation yl, of the time-dependent variable x and of the parameter S, and transmits the setpoint state vector ξ* as input to the control loop Br.

[0070] FIGS. 3 and 4 illustrate an example of the modeling by the module 6 of a path of the vehicle 1 driving at the center C1 of a first traffic lane VC1 and moving sideways to the center C2 of a second traffic lane VC2 adjacent to the first lane VC1.

[0071] Although in this example the vehicle 1 moves sideways to its left in the direction of forward travel, the vehicle 1 may, as an alternative, move sideways to its right.

[0072] The curve TRAJ represents the path of the vehicle 1 to move from the first traffic lane VC1 to the second lane VC2.

[0073] R(O, X, Y) denotes a frame of reference linked to the lane VC1 with origin O on the center C1 of the lane VC1 before the vehicle 1 begins to move sideways, X is oriented in the direction of travel of the vehicle 1 and Y is positioned so that the frame of reference R is orthogonal.

[0074] The path of the vehicle 1 is determined in a straight line, the predictor module 4 determining the second steering lock command Uff on the basis of the curvature yff of the road, of the velocity V of the vehicle and of the characteristics of the vehicle 1, notably the mass of the vehicle and the stiffness of the tires.

[0075] More particular reference is made to FIG. 4 which illustrates an example of the modeling of the path TRAJ using a Bézier curve P in the frame of reference R(O, X, Y).

[0076] The Bézier curve P comprises four control points P.sub.0, P.sub.1, P.sub.2 and P.sub.3 with respective coordinates (P.sub.ox; P.sub.oy), (P.sub.1x; P.sub.1y)/(P.sub.2x; P.sub.ty) and (P.sub.3x; P.sub.3y) and has the equation:

P(s)=P.sub.0(1−s).sup.3+3P.sub.1s(1−s).sup.2+3P.sub.2s.sup.2(1−s)+P.sub.3s.sup.3  (5)

where S varies between 0 and 1.

[0077] It is assumed that the abscissa axis X represents the time-dependent variable x representative of the variation in the change in path with respect to time, and the ordinate axis Y represents the lateral position yl of the vehicle 1 with respect to the frame of reference R(O, X, Y) of the lane VC1.

[0078] The vehicle 1 moves between the center C1 of the first lane VC1 corresponding to an initial lateral position yl.sub.ini defined by the coordinates of the control point P.sub.0 located on the center C1 and a final position yl.sub.fin defined by the coordinates of the control point P.sub.3 positioned on the center C2.

[0079] This yields the following equation:

[00005]$\begin{array}{cc}\left(\begin{array}{c}x\\ yl\end{array}\right)=\left(\begin{array}{c}{P}_{0x}\\ {\rho }_{0y}\end{array}\right){\left(1-s\right)}^3+3\left(\begin{array}{c}{p}_{1x}\\ P_{1y}\end{array}\right){s\left(1-s\right)}^2+3\left(\begin{array}{c}{P}_{2x}\\ P_{2y}\end{array}\right)s^2\left(1-s\right)+\left(\begin{array}{c}{f}_{3x}\\ P_{3y}\end{array}\right)s^3& \left(6\right)\end{array}$

[0080] As a variant, the Bézier curve may comprise more than four control points in order to improve the precision with which the path TRAJ is modeled, although this then does increase the computational workload.

[0081] It is assumed in what follows that the path TRAJ of the vehicle moving sideways to the left and the path of the vehicle moving sideways to the right (not depicted) are symmetrical.

[0082] Therefore the coordinates of the control points P.sub.0, P.sub.1, P.sub.2 and P.sub.3 are chosen so that:

P.sub.0x=0  (7)

P.sub.3x−P.sub.2x=P.sub.1x  (8)

P.sub.0y=yl.sub.ini  (9)

P.sub.3y=yl.sub.fin  (10)

[0083] In addition the following definitions are imposed:

P.sub.1y=P.sub.0y  (11)

P.sub.2y=P.sub.3y  (12)

[0084] By differentiating the Bézier curve P the maximum lateral velocity V.sub.lat,max and the maximum lateral acceleration a.sub.lat,max are defined:

[00006]$\begin{array}{cc}{V}_{\mathrm{lat},\max }=\frac{.Math.P_{3y}-P_{0y}.Math.}{P_{2x}}& \left(13\right)\end{array}$$\begin{array}{cc}{a}_{\mathrm{lat},\max }=2\frac{.Math.P_{3y}-P_{0y}.Math.}{3P_{1x}^2}& \left(14\right)\end{array}$

[0085] The defining of a maximum value for the lateral velocity V.sub.lat,max and of a maximum value for the lateral acceleration a.sub.lat,max enables the coordinates P.sub.1x, P.sub.2x and P.sub.3x to be determined.

[0086] The maximum value for the lateral velocity and the maximum value for the lateral acceleration are determined for example empirically during optimization of the path-control device.

[0087] The maximum value for the lateral velocity is for example equal to 0.8 m/s and the maximum value for lateral acceleration is for example equal to 0.5 m/s.sup.2.

[0088] In general, the maximum value for maximum lateral acceleration a.sub.lat,max and the maximum value for lateral velocity V.sub.lat,max are chosen so that the lane-change path is comfortable for the passengers of the vehicle 1.

[0089] In order to reduce the maximum value for maximum lateral acceleration a.sub.lat,max when the vehicle 1 is engaging the second traffic lane VC2, all that is required is to increase the value of the coordinate P.sub.3x.

[0090] By discretizing equation (6) by introducing the variable i that varies from 0 to N, for each instant x.sub.i, equation (6) is equal to:

[00007]$\begin{array}{cc}\left(\begin{array}{c}{x}_i\\ {\mathrm{yl}}_i\end{array}\right)=\left(\begin{array}{c}{P}_{0x}\\ P_{0y}\end{array}\right){\left(1-s_i\right)}^3+3\left(\begin{array}{c}{P}_{1x}\\ P_{1y}\end{array}\right){s_i\left(1-s_i\right)}^2+3\left(\begin{array}{c}{P}_{2x}\\ P_{2y}\end{array}\right)s_i^2\left(1-s_i\right)+\left(\begin{array}{c}{p}_{3x}\\ {\rho }_{3y}\end{array}\right)s_i^3& \left(15\right)\end{array}$

[0091] The module 6 comprises a processing unit UT employing equation (16).

[0092] The parameter S comprises all of the values S.sub.i, where i varies from 1 to N, N being a whole number.

[0093] The number N is chosen to be large enough to obtain a path for the vehicle 1 from the curve P that is not excessively discretized so as not to slow the control loop Br and detract from passenger comfort, and so that the number is not excessively high, so as to minimize the computation workload of the processing unit UT.

[0094] N is for example equal to 50, the curve P being discretized into 50 points.

[0095] Naturally the value of N may be other than 50.

[0096] The value of N may be chosen so as not to discretize the path excessively, the value of N being for example greater than 30.

[0097] Furthermore, the value of N can be chosen so as not to overload the processing unit UT, the value of N being for example less than 100.

[0098] The module 6 further comprises a time counter CP, the module 6 triggering the counter CP upon the lane change and stopping the counter when the value reached by the counter is greater than or equal to the value of the time-dependent variable x associated with the control point on the Bézier curve P.sub.3 that comprises the greatest value for the time-dependent variable P.sub.3x.

[0099] The counter CP operates with a period Tc.

[0100] The period Tc is chosen according to the operating frequency of the module 6 and of the loop Br.

[0101] The period Tc is for example equal to 10 ms.

[0102] The stopping of the counter CP indicates the end of the lane change.

[0103] A change in path is detected for example when a torque applied to the steering column of the vehicle 1 is above a detection threshold and an electrical impulse notably emitted by a steering change device is detected.

[0104] If the vehicle 1 is not completely aligned on the center C1 of the second lane VC2 at the end of the change in path, the control loop Br aligns the vehicle 1 with the center C1 later.

[0105] FIG. 5 illustrates an exemplary embodiment of the path-control device 5 during the lane change.

[0106] It is assumed that equation (15) and the control points P.sub.0, P.sub.1, P.sub.2 and P.sub.3 are stored in a memory MEM of the module 6.

[0107] The lane change is initiated in a step 10.

[0108] Next, in a step 11, the module 6 triggers the counter CP.

[0109] For each increment Tc of the counter, the module 6 determines the setpoint state vector on the basis of the lateral deviation yl.sub.i, of the time-dependent variable x.sub.i and of the parameter S.sub.i.

[0110] The setpoint state vector ξ* comprises setpoint values comprising for example the setpoint yaw velocity {dot over (Ψ)}.sub.ref, a setpoint heading angle Ψ.sub.ref and a setpoint lateral velocity y{dot over (l)}.sub.ref the setpoint lateral deviation yl.sub.ref.

[0111] The setpoint steering lock angle δref, the variation in steering lock angle δ{dot over (r)}ef and the integral of the setpoint lateral position ∫−ylref are equal to zero.

[0112] In a variant, the setpoint steering lock angle δref, the variation in steering lock angle δ{dot over (r)}ef and the integral of the lateral position ƒ−ylref may be calculated on the basis of the lateral deviation yli.

[0113] Two consecutive increments are separated by the duration Tc.

[0114] The module 6 determines a value for the parameter S.sub.i such that each increment Tc of the time counter Cp is comprised between two consecutive time-dependent variables x.sub.i and x.sub.i+1.

[0115] The module 6 then determines the two consecutive lateral deviations yl.sub.i and yl.sub.i+1 from the value of the parameter S.sub.i.

[0116] The module 6 determines an intermediate lateral deviation yl.sub.1 by interpolating between the two consecutive lateral deviations yl.sub.i and yl.sub.i+1.

[0117] The module 6 then determines setpoint values according to the following equations:

[00008]$\begin{array}{cc}\stackrel{.}{y}{l}_{\mathrm{ref}}=\frac{y_{i+1}-y_i}{x_{i+1}-x_i}& \left(16\right)\end{array}$$\begin{array}{cc}{\stackrel{.}{\Psi }}_{ref}=\frac{a_{\mathrm{lat}}}{V}& \left(17\right)\end{array}$

where V is the longitudinal velocity of the vehicle 1 and a.sub.lat is the instantaneous lateral acceleration on the path equal to:

[00009]$\begin{array}{cc}{a}_{\mathrm{lat}}=\frac{y_{i+2}-2y_{i+1}+y_i}{x_{i+1}-x_i}& \left(18\right)\end{array}$$\mathrm{and}$$\begin{array}{cc}{\Psi }_{ref}=\frac{\stackrel{.}{y}{l}_{\mathrm{ref}}}{V}& \left(19\right)\end{array}$

[0118] The setpoint lateral deviation yl.sub.ref is equal to:

[00010]$\begin{array}{cc}\{\begin{array}{c}{\mathrm{yl}}_{ref}=\mathrm{yl}1\mathrm{when}\mathrm{the}\mathrm{vehicle}\mathrm{is}\mathrm{traveling}\mathrm{in}\mathrm{the}\mathrm{first}\mathrm{traffic}\mathrm{lane}\mathrm{VC}1\\ {\mathrm{yl}}_{ref}=\mathrm{yl}1-{\mathrm{yl}}_{\mathrm{fin}}\mathrm{when}\mathrm{the}\mathrm{vehicle}\mathrm{is}\mathrm{traveling}\mathrm{in}\mathrm{the}\mathrm{second}\mathrm{traffic}\mathrm{lane}\mathrm{VC}2\end{array}& \left(20\right)\end{array}$

[0119] In an optional step 12, the module 6 can determine a corrective radius of curvature ρ.sub.SALC on the basis of the instantaneous lateral acceleration a.sub.lat, the lateral velocity {dot over (y)}l and of the longitudinal velocity V of the vehicle, using the following equation:

[00011]$\begin{array}{cc}{\rho }_{\mathrm{SLAC}}=\frac{{\mathrm{Va}}_{\mathrm{lat}}-\stackrel{.}{y_l}{a}_{\mathrm{lat}}}{{\left(V^2+{\stackrel{.}{y}}_l^2\right)}^{\frac{3}{2}}}& \left(21\right)\end{array}$

[0120] The value of the corrective radius of curvature ρ.sub.SALC is transmitted to the predictor module 4 and added to the second command Uff to improve the dynamics of the path following.

[0121] Steps 11 and 12 are repeated until such time as the value reached by the counter CP is greater than or equal to the value P.sub.3x of the control point P.sub.3 (step 13) for each increment of the counter CP.

[0122] As soon as the value of the counter CP becomes greater than or equal to the value P.sub.3x, the next step becomes step 14 of waiting for the next lane change.

[0123] Depending on the data received, notably by the camera, the module 6 may modify the coordinates P.sub.2y and P.sub.3y in order to take into account dynamically the change in the width of the second lane VC2 so that the curve P can be recalculated so as to fit the path of the vehicle 1 to its environment at the time of the lane change.

[0124] FIG. 6 depicts the variation with respect to time in the setpoint lateral position yl.sub.ref (curve CB1), in the lateral position yl of the vehicle 1 (curve CB2) and in the angle of the front wheels δ (curve CB3) upon a lane change to the right and then to the left by the vehicle 1 driving at a velocity V of 90 km/h.

[0125] Modeling the lane-change path using the Bézier curve P enables the geometric properties of the curve P to be easily adapted to take account of the environment in which the vehicle 1 is moving, notably the width of the lane to which the vehicle has moved.

[0126] In addition, calculating the setpoint state vector for the change in path and altering the geometric properties of the curve P to suit do not require significant computation power.

[0127] The initial determination of the geometric properties of the curve P requires only that the maximum lateral velocity and the maximum lateral acceleration of the vehicle and the duration of the lane-change operation be set ahead of time.

[0128] Furthermore, if the lane changes to the right or to the left of the vehicle are symmetrical in their path (same lateral acceleration at the start and the end of the paths) only the lateral velocity and the lateral acceleration need to be determined.

[0129] The parameters for adjustment, which for example comprise the lateral acceleration, the lateral velocity, the duration of the lane-change operation, may be determined on the basis of regulatory constraints and in such a way as to ensure the comfort of the passengers of the vehicle during the change in path.

[0130] Finally, the lane-change path can easily be adapted to the control loop Br that already exists on most vehicles, enabling existing control laws to be reused.