DEVICE AND METHOD FOR CONTROLLING THE TRAIL OF A MOTOR VEHICLE

Description

[0001] The present invention relates to the trajectory control of a motor vehicle, and more particularly to a trajectory control device and method.

[0002] Motor vehicles are frequently equipped with trajectory control devices that assist the driver in keeping the vehicle in a traffic lane. They act on the vehicle's direction by modifying the steering angle of the vehicle's steered wheels. A control law is implemented and may be a comfort control law of the lane centering type, also known as "Lane Centering Assist" or by the corresponding acronym "LCA". The control law may also be a safety control law of the lane keeping assistance type, also known as "Lane Keeping Assist" or by the corresponding acronym "LKA".

[0003] In this context, it is generally chosen to implement a smooth control law that avoids jolts. This avoids surprising the driver and improves the comfort of the vehicle's occupants.

[0004] A trajectory control device including a feedback module makes it possible to implement such a control law. The feedback module creates a closed loop and has slow dynamics. To improve the performance level, an anticipatory module can also be added. The anticipatory module adds an open-loop term to the closed-loop term provided by the feedback module.

[0005] For example, we can refer to document FR 3 051 756 which illustrates a real-time vehicle trajectory control device. This device comprises a feedback module generating a closed-loop term and an anticipatory module generating an open-loop term. The closed loop keeps the vehicle in the center of a virtual lane which is always considered straight. The open loop takes into account the curvature and compensates for the effect of the turn on the states and the control.

[0006] Although such a device is generally satisfactory, and in particular never compromises passenger safety, certain vehicle usage situations can cause unpleasant situations for the driver, such as oversteer or abnormal vehicle tracking. For example, if the motor vehicle is overloaded, abnormal tracking of the vehicle from the center of the lane is sometimes observed.

[0007] To overcome this drawback, the closed loop can be made more dynamic so that oversteer or toe-out is corrected by the closed loop. This ensures good safety for the vehicle's occupants. However, the more dynamic closed loop is less comfortable for the driver and vehicle occupants.

[0008] Document US2007 / 213901A1 describes a method for adjusting an anticipator module which notably has the aforementioned drawbacks.

[0009] In view of the above, the invention aims to remedy the aforementioned drawbacks.

[0010] More particularly, the invention aims to improve comfort in a vehicle when implementing trajectory control with a view to maintaining in a traffic lane, in particular when a bend appears.

[0011] To this end, a method is proposed for adjusting an anticipator module equipping a trajectory control device of a motor vehicle, said module using a bicycle model of said vehicle, in which: detecting whether the anticipator module is unsuitable during a turn by taking into account a lateral deviation from an ideal trajectory and / or a contribution from a feedback module of the control device, determining primary parameters, calculating a secondary parameter by an optimization calculation method taking into account the determined primary parameters, and updating a bicycle model of the vehicle by taking into account the calculated secondary parameter.

[0012] Such a method makes it possible to correct the adjustment of the anticipator module. This results in a correction of the trajectory deviation more ensured by the anticipator module and therefore the possibility of reducing the contribution of the feedback module, providing better comfort in the motor vehicle. In particular, the abnormal distance from the center of the track mentioned above comes in particular from the impact of the load on the dynamics of the vehicle. The method according to the invention makes it possible to take this impact into account.

[0013] The anticipator module is detected to be unsuitable if: the lateral deviation from an ideal trajectory is greater than a predefined deviation threshold, and a ratio of a contribution from the feedback module to a contribution from the anticipatory module for the steering control is greater than a predefined ratio threshold.

[0014] This detection allows better identification of conditions requiring correction of the anticipator module setting.

[0015] Preferably, when updating a bicycle model of the vehicle, a corrected bicycle model characteristic data is determined taking into account the secondary parameter, an average is calculated between a current bicycle model characteristic data and the corrected bicycle model characteristic data and the current bicycle model characteristic data is replaced by the calculated average.

[0016] Such an update step makes it possible to increase the robustness of the adjustment process.

[0017] Different variations are possible regarding the secondary parameter.

[0018] According to a first variant, the calculation of a secondary parameter comprises the calculation of a corrected understeer gradient, the corrected understeer gradient preferably being a characteristic data of the bicycle model of the vehicle.

[0019] Such a variant is preferable as it requires less memory and computing resources.

[0020] Preferably, the primary parameters comprise a lane curvature, a vehicle speed and a steering wheel angle, the corrected understeer gradient being calculated by minimizing a steering wheel angle deviation, and advantageously by minimizing the function: f = ∑ i ρ i × L tot + ∇ sv × v i 2 × d − SWA mesur é i 2 Or, L all is the wheelbase of the vehicle, d is the gear ratio of the vehicle's steering column, ∇ sv is the understeer gradient and whatever an iteration i, ρ(i) is the curvature of the taxiway during iteration i, v(i) is the vehicle speed during iteration i and SWA measured (i) is the steering wheel angle during iteration i.

[0021] According to a second variant, the calculation of the secondary parameter includes the calculation of a corrected front axle stiffness and / or a corrected rear axle stiffness.

[0022] The method may optionally include identifying other parameters of the bicycle model such as mass, inertia, center of mass position.

[0023] Preferably, the primary parameters include a lateral deviation from an ideal trajectory, a longitudinal speed of the vehicle, a heading angle of the vehicle, a steering angle and a curvature of the traffic lane and wherein a reference trajectory is constructed from the primary parameters, the optimization being implemented by minimizing the deviation between the reference trajectory and a trajectory of the vehicle determined from characteristic data of the current bicycle model of the vehicle.

[0024] Advantageously, the reference trajectory is constructed on the basis of the primary parameters by applying the relation: X ref t k = X route t k − y L t k ⋅ sin ψ rel + ψ route t k ⋅ t k − t k − 1 Y ref t k = Y route t k + y L t k ⋅ cos ψ rel + ψ route t k ⋅ t k − t k − 1 Or tk is the time relative to an iteration k, (X ref (tk ),Y ref (tk )) is the pair of coordinates of the reference trajectory at the instant tk, (X route (tk ),Y route (tk )) is the coordinate pair of the taxiway center trajectory at time tk , and L (tkis the lateral offset between the vehicle's center of gravity and the center of the traffic lane, Read ψ is the relative yaw angle of the vehicle, ψ route (tk is the tangent angle to the taxiway center trajectory at time tk , the taxiway center trajectory coordinate pair and the tangent angle to the taxiway center trajectory being determined by applying the initialization relation: ψ route 0 = X route 0 = Y route 0 = 0 et la relation de récurrence : X route t k = v ⋅ cos ψ route t k ⋅ t k − t k − 1 + X route t k − 1 Y route t k = v ⋅ sin ψ route t k ⋅ t k − t k − 1 + Y route t k − 1 ψ route t k = ρ ⋅ v ⋅ t k − t k − 1 + ψ route t k − 1

[0025] In another implementation, the vehicle's bicycle model is reset each time the vehicle is not used.

[0026] Such a reset is particularly advantageous in that it avoids having an unsuitable anticipator module if the vehicle's operating conditions have changed during the period of non-use of the vehicle.

[0027] According to another aspect, there is provided a computer program comprising code configured to, when executed by a processor or an electronic control unit, implement the method as defined above.

[0028] According to yet another aspect, there is provided a device for adjusting an anticipator module equipping a trajectory control device of a motor vehicle, said module using a bicycle model of said vehicle, comprising a detection module configured to detect whether the anticipator module is unsuitable during a turn by taking into account a lateral deviation from an ideal trajectory and / or a contribution from a feedback module of the control device, a module for determining primary parameters, a calculation module capable of calculating a secondary parameter by an optimization calculation method taking into account the primary parameters determined by the determination module and an update module configured to update a bicycle model of the vehicle by taking into account the secondary parameter calculated by the calculation module.

[0029] Other aims, characteristics and advantages of the invention will appear on reading the following description, given solely by way of non-limiting example, and made with reference to the appended drawings in which: [ Figure 1 ] is a schematic representation of a trajectory control device of a motor vehicle, [ Figure 2 ] is a schematic representation of an adjustment device of the device shown in the figure 1 , [ Figure 3 ] is a diagram of a method according to a first embodiment of the invention, [ Figure 4 ] [ Figure 5 ] are graphs representing the lateral deviation and the contribution of an anticipatory module and a feedback module during the process represented on the figure 3 , [ Figure 6 ] is a representation of the calculation of the secondary parameter during the process shown in the figure 3 , [ Figure 7] is a diagram of a method according to a second embodiment of the invention, [ Figure 8 ] is a representation of the calculation of secondary parameters during the process of the figure 7 .

[0030] It is represented in the form of a block diagram on the figure 1 a trajectory control device 2 intended to be incorporated into a motor vehicle (not shown). The device 2 is an advanced driver assistance system, also known by the English term “advanced driver assistance system” or by the corresponding acronym “ADAS”. More particularly, the device 2 has the function of developing a command for a steering system of the motor vehicle so as to keep the vehicle in the center of a virtual lane.

[0031] For this purpose, the block diagram of the figure 1comprises a first block 4 corresponding to the vehicle in which the device 2 is incorporated. Block 4 is subject to an input being a steering angle request. Block 4 delivers an output being the available measurements, in this case the yaw rate, the yaw angle relative to the road, the lateral deviation and the steering angle. The block diagram comprises a second block 6 corresponding to an observer, a sensor or an estimator. The block diagram comprises a third block 8 corresponding to a corrector, for example constituted by a gain vector.

[0032] The device 2 comprises a first return loop 10 connected downstream of the block 4, comprising the observer 6 and supplied to a subtractor 12. The device 2 comprises a second return loop 14 extending from a connection point between the blocks 8 and 4 to the block 6. The assembly constituted by the loops 10 and 14, the block 6 and the subtractor 12 is designated in the present application by the expression “feedback module” also known by the English term “feedback module”.

[0033] The observer of block 6 implements a state representation based on the bicycle model. In the present application, the expression “bicycle model” designates the bicycle model of the vehicle used by the device 2. The bicycle model is notably used by software means in the context of calculating an anticipation term. The bicycle model is based on a state vector x whose components are the following seven states: ψ̇ : speed of the vehicle's relative heading angle to the roadway, Read ψ : relative heading angle of the vehicle in relation to the roadway, ẏ L ' : lateral speed of the vehicle relative to the roadway, and L : lateral deviation of the vehicle from the roadway, δ̇ : front wheel angle speed, δ : steering angle, in this case front wheel angle, ∫ and L ': integral of the lateral deviation.

[0034] The device 2 further comprises an anticipator module 16. The anticipator module 16 is schematically represented on the block diagram of the figure 1 by two dotted rectangles. Module 16 is also known by the Anglo-Saxon term “feedforward module”.

[0035] The anticipator module 16 comprises a summer 18 located between the connection point of the loop 14 and the block 4. The summer 18 adds an open loop term δ eq .

[0036] The anticipator module 16 comprises a subtractor 20 placed between the connection point of loop 10 and block 6. The subtractor 20 subtracts an open loop vector from loop 10.

[0037] The state representation is represented by the equation below: d dt ψ ψ rel y ˙ L y L δ ˙ δ ∫ − y L ︸ X = − C f I f 2 + C r l r 2 I 2 v C l I f − C r l r I z − C f I f − C r I r I z v 0 0 C f I f I z 0 1 0 0 0 0 0 0 − C f I f − C r I r mv C f + C r m − C f + C r mv 0 0 C f m 0 0 0 1 0 0 0 0 0 0 0 0 − 2 ξω − ω 2 0 0 0 0 0 1 0 0 0 0 0 − 1 0 0 0 ︸ A ψ ψ rel y ˙ L y L δ ˙ δ ∫ − y L ︸ X + 0 0 0 0 ω 2 0 0 ︸ B δ δ roues + 0 − v − v 2 0 0 0 0 ︸ B p ρ ψ ψ rel y L δ ∫ − y L ︸ Y = 1 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 ︸ C ψ ˙ ψ rel y ˙ L y L δ ˙ δ ∫ − y L Or C f is the equivalent stiffness of the front axle, C r is the equivalent stiffness of the rear axle, L f is the distance between the front axle and the center of gravity, L r is the distance between the center of gravity and the rear axle, m is the mass of the vehicle, IZ is the moment of inertia of the vehicle and v is the longitudinal velocity of the vehicle.

[0038] This bicycle model is a simplification of reality. However, the characteristic data of the bicycle model used by device 2 are directly linked to the vehicle configuration, and in particular to configuration parameters such as the distribution of mass on the vehicle, tire pressure, etc.

[0039] Closed-loop lateral control aims to minimize the state vector x around zero, corresponding to a straight line. When a turn occurs, three states must be corrected: the relative heading angle ψ read , the derivative of relative heading angle ·ψ̇' and the steering angle at the front wheels δ . The angle δ is connected to the steering wheel angle by a second-order dynamic whose gain corresponds to the gear ratio of the steering column.

[0040] In order to improve the performance of the module 16, the device 2 comprises an adjustment device 22 shown in the figure 2 .

[0041] In reference to the figure 2, the device 22 comprises a detection module 24. The module 24 is in information connection with the block 6 of the device 2 so as to quantify the lateral deviation of the vehicle from an ideal trajectory. In the example illustrated, the ideal trajectory is a center line of a traffic lane in which the vehicle is traveling. The module 24 is also in information connection with the feedback module and the anticipator module 16 of the device 2. In doing so, the module 24 is capable of comparing a contribution from the feedback module for the steering control of the vehicle with a contribution from the anticipator module 16 for the steering control of the vehicle. In doing so, the module 24 is provided with the hardware and software means to detect whether the module 16 is unsuitable during a turn.

[0042] The device 22 comprises a determination module 26. The module 26 is in information connection with the module 24. More particularly, the module 26 is configured to be activated if the module 24 has detected that the anticipator module 16 was unsuitable during a turn. The module 26 is in information connection with the block 6 and / or with sensors and / or with estimators equipping the motor vehicle so as to detect a plurality of primary parameters.

[0043] The device 22 comprises a calculation module 28. The module 28 is in information connection with the module 26. More particularly, the module 28 is configured to calculate one or more secondary parameters when the module 26 has determined primary parameters. In this case, the module 28 implements a calculation method by optimization on the basis of the primary parameters determined by the module 26.

[0044] The device 22 further comprises an update module 30. The module 30 is in information connection with the module 28 so as to be able to update the bicycle model of the vehicle used by the device 2 taking into account the secondary parameter(s) calculated by the module 28.

[0045] In reference to the figure 3 , a method according to a first embodiment of the invention is schematically represented. The method is implemented by means of the device 22 shown in the figure 2 . In this mode of implementation, the calculation of the open loop term is determined by applying the equation: δ eq = ρ L tot + ∇ sv ⋅ v 2 Or ρ is a curvature of the traffic lane, L all is the wheelbase of the vehicle, v is the speed of the vehicle, in this case the longitudinal speed, and ∇ sv is an understeer gradient determined by: ∇ sv = − m C f L f − C r L r C f C r L tot

[0046] The method is implemented regularly, for example every 10 ms. The method comprises a first test step E11. During step E11, it is determined whether the vehicle has just started a period of non-use. For example, it can be determined during step E11 that an engine fitted to the motor vehicle has been switched off. If, during step E11, it is detected that the vehicle is entering a period of non-use, a step E12 is applied. Otherwise, a test step E13 is implemented. During step E12, the parameters of the bicycle model are reset to the initialization values.

[0047] Step E13 is implemented during a turn. During this step, it is detected whether the module 16 is unsuitable for the conditions of use of the vehicle in which the device 2 is incorporated. Step E13 is implemented by the module 24. More specifically, during step E13, the lateral deviation from an ideal trajectory is monitored and the respective contributions of the feedback module and the module 16 for the steering control are monitored.

[0048] More precisely, in the illustrated example, it is detected that the module 16 is unsuitable if the absolute value of the maximum lateral deviation exceeds a deviation threshold se equal to 0.2 m and if the absolute value of the ratio between the contribution of the feedback module on the contribution of the module 16 at the time of the maximum lateral deviation is greater than a ratio threshold sr equal to 0.1.

[0049] In reference to the figure 4 , we have schematically represented the lateral deviation ( and L) and on a second graph the contributions of the feedback module ( θ FB ) and module 16 ( θ FF ) for a simulation of a turn at 90 km / h with a curvature of 3x10 -3< m -1< when module 16 is adapted to the vehicle's operating conditions.

[0050] There figure 5 illustrates the evolution in an identical turn of the gap and L and contributions θ FB And θ FF with an unsuitable module 16. More specifically, the rigidities of the vehicle's front and rear axles were increased by 30% without changing the characteristic data C f And C r of the bicycle model used by module 16. Such a scenario is for example likely to occur when the user of the motor vehicle changes tires. On the figures 4 And 5 , θ total corresponds to the total contribution for the steering command, i.e. the sum of the contributions θ FF and θ FB .

[0051] It emerges from the graphs of the figure 4 that the gap and L maximum is, in absolute value, of the order of 0.16 m. When the gap and L maximum occurs, the contribution θ FB is less than 1° while the contribution θ FF is 10°. By contrast, it emerges from the figure 5 that the gap and L maximum is of the order of 0.33 m. At this time, the contribution θ FB is of the order of -1.8° and the contribution θ FF is 12°. From these results and by comparison with the thresholds se and sr mentioned above, it appears that module 16 is unsuitable in the case of the figure 5 .

[0052] Module 16 corresponds to an inversion of the model in steady state. If the bicycle model is well identified during a turn, the deviation and L will be small and the contribution θ FF will be predominant over the contribution θ FB .Advantageously, step E13 is only triggered when the lateral acceleration is greater than an acceleration threshold sa . In doing so, the detection of an unsuitable anticipator module is only implemented when the lateral dynamics are sufficiently excited. This results in a further improved relevance of the detection.

[0053] Again in reference to the figure 3 , if during step E13 it was detected that module 16 was unsuitable, a step E14 is implemented. Step E14 is implemented during a following turn. Step E14 is implemented by module 26. During step E14, primary parameters are determined, in this case the curvature ρ lane, longitudinal speed v of the vehicle, steering angle Measured SWA and a time t.

[0054] Next, a step E15 is implemented for calculating a secondary parameter. Step E15 is implemented by module 28. More particularly, the secondary parameter is calculated by an optimization calculation method based on the primary parameters determined during step E14. This optimization consists of finding the understeer gradient ∇ sv such that module 16 provides almost all of the steering control and the contribution θ FB almost zero when cornering.

[0055] There figure 6 schematically illustrates the calculation of the gradient ∇ sv optimal. The diagram of the figure 6 comes from Simulink software.

[0056] The diagram of the figure 6 includes a block 32 for entering the variable to be determined, in this case ∇ sv (Grad_sv).

[0057] Block 32 is connected to a pure gain block 34. The function of block 34 is to multiply the variable to be determined to convert degrees into radians and to apply the gear ratio between the steering angle at the front wheels and the angle at the steering wheel. The result from block 34 is sent to a multiplier block 36.

[0058] A primary parameter input block 44 allows the speed to be entered v (speed) of the vehicle. Block 44 is connected to block 36 so that the result from block 34 is multiplied by the square of the vehicle speed.

[0059] The result from block 36 is supplied to a summing block 38. Block 38 is linked to a block 46 for entering the constant Ltot corresponding to the wheelbase of the vehicle.

[0060] The result of the sum calculated by block 38 is supplied to a multiplier block 40. Block 40 is in communication with a block 48 for entering a primary parameter making it possible to enter the curvature ρ (rho) of the traffic lane.

[0061] The result of the multiplication calculated by block 40 is supplied to a pure gain block 42. Block 42 multiplies this result to convert radians to degrees. The result obtained in degrees is a steering angle SWA_deg in degrees supplied to output block 50.

[0062] To find the gradient ∇ sv appropriate, we fill in the inputs, namely the speed v in block 44 and the curvature ρ in block 48. The algorithm searches for a gradient ∇ sv such that the SWA_deg output obtained in block 50 corresponds in the least squares sense to the angle Measured SWA at the real wheel. Therefore, we seek, depending on the gradient, to minimize the function fdefined as: f = ∑ i ρ i × L tot + ∇ sv × v i 2 × d − SWA mesur é i 2

[0063] In the illustrated example, the optimization is performed on a series of measurements. During a turn, several iterations are implemented. During an iteration i, the primary parameters ρ(i), v(i) And SWA measured (i) are determined. After the last iteration i=n, ​​the function f is minimized in the least squares sense taking into account iterations 1 to n. In this equation, L all is the wheelbase of the vehicle and d is the gear ratio of the vehicle's steering column. An example of an algorithm that can be implemented to determine the gradient ∇ sv The appropriate parameter is the "Lsqnonlin" function of the Matlab software. At the end of step 15, a secondary parameter was determined, in this case an understeer gradient value ∇ sv ,

[0064] Following step E15, a step E16 of updating the bicycle model is implemented. During step E16, an average is calculated between the gradient ∇ current_sv current and the corrected gradient ∇ sv _corrected calculated during step E15. The result is an average understeer gradient ∇ sv _ average : ∇ sv _ moyen = ∇ sv _ actuel + ∇ sv _ corrig é 2

[0065] Although, in the illustrated example, an average is calculated between the current and corrected gradients, one can of course, without departing from the scope of the invention, envisage another type of calculation, for example a weighted arithmetic average: ∇ sv _ moyen = 0 , 2 × ∇ sv _ actuel + 0 , 8 × ∇ sv _ corrig é

[0066] In the illustrated example, the gradient ∇ svis a characteristic data item of the bicycle model. Thus, during step E16, a characteristic data item of the corrected bicycle model was determined. However, it is of course not outside the scope of the invention to envisage a bicycle model having other characteristic data. According to an alternative example, the gradient∇ sv determined during step E15 can be used to determine a pair of drift stiffnesses ( C f , C r ) of the vehicle based on the equation: ∇ sv = − m C f L f − C r L r C f C r L tot

[0067] The method then comprises a step E17 during which the gradient ∇ sv is replaced _current by the average ∇ sv _ average . Steps E16 and E17 are implemented by module 30. In this way, the bicycle model converges towards a model as close as possible to reality and the contribution θ FFof module 16, which becomes more adapted to the actual configuration of the vehicle, is increasingly predominant during a turn. It is then possible to reduce the dynamics of the feedback module so that comfort is improved for the driver and occupants of the motor vehicle.

[0068] In reference to the figure 7 , a method according to a second embodiment of the invention is shown. Identical elements bear the same references. Step E14 is replaced by a step E24, step E15 is replaced by a step E26 and a step E25 is incorporated between steps E24 and E26. In this embodiment, the calculation of the open loop term is determined by applying the equation: δ eq = ρ L tot − m C f L f − C r L r C f C r L tot ⋅ v 2

[0069] During step E24, the primary parameters determined are the lateral deviation and Lrelative to an ideal trajectory, the speed v, in this case longitudinal, of the vehicle, the relative heading angle Read ψ of the vehicle, the angle at the steering wheel Measured SWA and the curvature ρ of taxiway.

[0070] In step E25, a reference trajectory is constructed from the primary parameters determined in step E24. More specifically, the reference trajectory is determined in a Cartesian reference frame (x,y) by assuming that the position and heading of the vehicle at the start of recording define the origin of the Cartesian reference frame (H1) and by calculating the trajectory using a recurrence formula.

[0071] Due to hypothesis (H1), we have, at t = 0: ψ route 0 = X route 0 = Y route 0 = 0 Or ψ route , X route And Y route are respectively the heading angle, the abscissa and the ordinate of a taxiway center trajectory.

[0072] The derivative of the taxiway center trajectory heading angle with respect to time is written as: ψ ˙ route = ρ ⋅ v

[0073] By integrating this expression with respect to time, we obtain, in discrete form: ψ route t k = ρ ⋅ v ⋅ t k − t k − 1 + ψ route t k − 1

[0074] This variable allows the trajectory of the road to be reconstructed in the absolute reference frame: X route t k = v ⋅ cos ψ route t k ⋅ t k − t k − 1 + X route t k − 1 Y route t k = v ⋅ sin ψ route t k ⋅ t k − t k − 1 + Y route t k − 1

[0075] In this way, the taxiway center trajectory is estimated. The relationship between the taxiway center trajectory heading angle, the relative yaw angle, and the absolute yaw angle is written as follows: ψ rel = ψ absolu − ψ route ⇒ ψ absolu = ψ rel + ψ route

[0076] From this equation and the lateral deviation measurements and L and the relative yaw angle Read ψ measured, we construct the trajectory actually followed by the vehicle around the trajectory of the center of the traffic lane in an absolute reference frame: X ref t k = X route t k − y L t k ⋅ sin ψ rel + ψ route ⋅ t k − t k − 1 Y ref t k = Y route t k + y L t k ⋅ cos ψ rel + ψ route ⋅ t k − t k − 1

[0077] In step E26, we seek to find a bicycle model reproducing the reference trajectory when excited with the same longitudinal speed and the same steering angle. This calculation will be explained with reference to the diagram of the figure 8 . The diagram of the figure 8 presents two blocks 52 for entering variables to be determined. More particularly, blocks 52 allow the entry of variables corresponding to the drift rigidity C f on the front axle and the drift rigidity C r on the rear axle. Although, in the example illustrated, the variables to be determined are drift rigidities, one can of course, without departing from the scope of the invention, envisage other variables to be determined, for example the moment of inertia IZ , the distances L f And L r or even the mass m.

[0078] The diagram of the figure 8has four constant 54 input blocks. In the example shown, the constants entered correspond to the distance L f , the distance L r , the wheelbase L all (Ltot) and the mass m (Mtot) of the vehicle.

[0079] The diagram of the figure 8 has two primary parameter input blocks 56 and 58. Block 56 is used to enter the speed v (speed). Block 58 is used to enter the steering wheel angle Measured SWA (SWA_rad) in radians.

[0080] The diagram of the figure 8 comprises a matrix generation block 60 for generating the matrix A of the state representation. The diagram of the figure 8 includes a pure gain multiplication block 62 dividing the angle entered in block 58 by the gear ratio to obtain the steering angle at the front wheels.

[0081] The diagram of the figure 8comprises a matrix multiplier block 64 multiplying the result of block 62 by the matrix generated by block 60. An integrator block 66, a matrix multiplier block 68 and a summer 70 are arranged at the output of block 64.

[0082] An integrator block 72 and a summer 74 downstream of the integrator 66 provide a result to a cosine block 76, the output of which is multiplied at speed v by a multiplication block 78 and integrated by an integrator block 80. In parallel, the aforementioned result is submitted to a sine block 82, the output of which is multiplied at speed v by a multiplication block 84 and integrated by an integrator block 86. Two output blocks 88 and 90 collect the signals X veh And Y veh integrated by blocks 80 and 86, respectively.

[0083] The calculation model corresponding to the figure 8 allows to obtain, from the speed v and the steering wheel angle SWA, a trajectory ( X veh ,Y veh) of vehicle delivered by blocks 88 and 90. By modifying the variables entered in blocks 52, we seek to ensure that the output coordinates delivered by blocks 88 and 90 correspond to the coordinates of the reference trajectory. More specifically, this correspondence is implemented in the least squares sense over the entire chosen recording. In doing so, we seek to minimize the function f : f = ∑ i X ref i − X veh i 2 + Y ref i − Y veh i 2

[0084] Such a minimization can be done using for example the “Lsqnonlin” function of Matlab. In doing so, we obtain a pair of rigidities ( C f_corrected ,C r_corrected ) corrected.

[0085] Steps E16 and E17 of the process of the figure 7 are similar to steps E16 to E17 of the method of the figure 3 , the two secondary parameters C f And C r replacing the secondary parameter ∇ sv .

[0086] More particularly, during step E16 of the process of the figure 7 , we calculate an average C r_mean and an average C r_mean : C f _ moyenne = C f _ actuel + C f _ corrig é 2 C r _ moyenne = C r _ actuel + C r _ corrig é 2 Or C r_current , C r_current are respectively the current rigidities on the front and rear axles, i.e. the rigidities before the implementation of the process.

[0087] During step E17 of the process of the figure 7 , we replace the rigidity C f_current by the average C f_average and we replace the rigidity C r_current by the average C r_mean .

[0088] In this way, a bicycle model was established to increase the contribution θ FF of module 16 during a turn. It should be noted that the corrected stiffness pair is not necessarily the most representative of the motor vehicle. Indeed, it is sufficient to find a combination of values ​​reproducing the behavior of the motor vehicle. In the second implementation mode, several pairs ( C f ,C r ) result in a contribution θ FBof the almost zero feedback module in stabilized turn. The invention makes it possible to converge towards any of these pairs and therefore makes it possible, by mobilizing few computing resources, to increase the contribution θ FF from the anticipatory module to the steering control and therefore improve the comfort of the occupants of the motor vehicle.

Claims

1. Method for setting an anticipator module (16) with which a control device (2) controlling the trajectory of a motor vehicle is equipped, said module using a bicycle model of said vehicle, wherein: - a detection is made (E13) as to whether the anticipator module (16) is unsuitable if: - the lateral deviation (yL) with respect to an ideal trajectory is greater than a predefined deviation threshold (se), and - a ratio of a contribution (θFB) of the feedback module to a contribution (θFF) of the anticipator module (16) for steering control is greater than a predefined ratio threshold (sr), - primary parameters are determined (E14, E24), - a secondary parameter is calculated (E15, E26) by an optimization-based calculation method taking account of the determined primary parameters, and - the bicycle model of the vehicle is updated (E16, E17) by taking account of the calculated secondary parameter.

2. Method according to Claim 1, wherein, when a bicycle model of the vehicle is updated, a characteristic datum (∇sv_corrected, Cf_corrected, Cr_corrected) of the corrected bicycle model is determined by taking account of the secondary parameter, an average (∇sv_average, Cf_average, Cr_average) between a characteristic datum (∇sv_current, Cf_current, Cr_current) of the current bicycle model and the characteristic datum (∇sv_current, Cf_corrected, Cr_corrected) of the corrected bicycle model is calculated, and the characteristic datum (∇sv_current, Cf_current, Cr_current) of the current bicycle model is replaced by the calculated average (∇sv_average, Cf_average, Cr_average).

3. Method according to either one of Claims 1 and 2, wherein the calculation of the secondary parameter (E15) comprises the calculation of a corrected understeering gradient (∇sv_corrected), the corrected understeering gradient (∇sv_corrected) preferably being a characteristic datum of the bicycle model of the vehicle.

4. Method according to Claim 3, wherein the primary parameters comprise a traffic lane curvature (ρ), a speed (v) of the vehicle and a steering wheel angle (SWAmeasured), the corrected understeering gradient (∇sv_corrected) being calculated by the minimization of a steering wheel angle deviation, and preferably by the minimization of the function: f = ∑ i ρ i × L tot + ∇ sv × v i 2 × d − SWA measuré i 2 in which Ltot is the wheelbase of the vehicle, d is the gear reduction ratio of the steering column of the vehicle, ∇sv is the understeering gradient and, regardless of an iteration i, ρ(i) is the traffic lane curvature upon the iteration i, v(i) is the speed of the vehicle upon the iteration i and SWAmeasured(i) is the steering wheel angle during the iteration i.

5. Method according to either one of Claims 1 and 2, wherein the calculation of the secondary parameter (E26) comprises the calculation of a corrected front train stiffness (Cf_corrected) and / or of a corrected rear train stiffness (Cr_corrected).

6. Method according to Claim 5, wherein the primary parameters comprise a lateral deviation (yL) with respect to an ideal trajectory, a longitudinal speed v of the vehicle, a heading angle (ψrel) of the vehicle, a steering wheel angle (SWAmeasured) and a traffic lane curvature (ρ) and wherein a reference trajectory (Xref,Yref) is constructed from the primary parameters, the optimization being implemented by the minimization of the deviation between the reference trajectory (Xref,Yref) and a trajectory of the vehicle (Xveh,Yveh) determined from characteristic data of the current bicycle model of the vehicle.

7. Method according to any one of Claims 1 to 6, wherein the bicycle model of the vehicle is reset (E12) on each period of absence of use of the vehicle.

8. Computer program comprising a code configured to, when it is executed by a processor or an electronic control unit, implement the method according to any one of Claims 1 to 7.

9. Setting device (22) for setting an anticipator module (16) with which a control device (2) controlling the trajectory of a motor vehicle is equipped, said module using a bicycle model of said vehicle, comprising a detection module (24) configured to detect whether the anticipator module (16) is unsuitable during a turn by taking account of a lateral deviation (yL) with respect to an ideal trajectory and / or a contribution (θFB) of a feedback module of the control device (2), a module (26) for determining primary parameters, a computation module (28) capable of calculating a secondary parameter by an optimization-based calculation method taking account of the primary parameters determined by the determination module (26) and an updating module (30) configured to update the bicycle model of the vehicle by taking account of the secondary parameter calculated by the computation module (28).