Method for autonomously controlling a steering control actuator of an automotive device

Abstract

The invention relates to a method for autonomously controlling a steering control actuator of an automotive device (10), comprising steps of: - acquiring a desired trajectory (T0) for the device, - acquiring parameters relating to said trajectory and / or said device, and - calculating an actuator control setpoint as a function of said parameters, by means of a global controller. According to the invention, the global controller is determined on the basis of two elementary controllers, one of which guarantees better trajectory tracking than the other, and a weighting coefficient of at least one of the two elementary controllers.

Description

Description Title of the invention: Method for autonomously controlling a steering control actuator of an automotive device. Technical field of the invention

[0001] The present invention relates generally to the automation of trajectory tracking for automotive devices.

[0002] It finds a particularly advantageous application in the context of driver assistance systems for motor vehicles, but it can also be applied to the fields of aeronautics or robotics.

[0003] It relates more specifically to a method for autonomously controlling a steering control actuator of an automotive device, comprising the following steps: - acquisition of a desired trajectory for the device, - acquisition of parameters relating to said trajectory and / or to a posture of said device relative to said trajectory and / or said device, and of - calculation by a computer of a control setpoint for said actuator as a function of said parameters, by means of a global controller.

[0004] It also concerns a device equipped with a computer adapted to implement this process. State of the art

[0005] In an effort to improve the safety of motor vehicles, they are currently being equipped with driver assistance systems or autonomous driving systems.

[0006] Among these systems, the following systems were developed: - automatic avoidance (better known by the abbreviation AES, from the English "Automatic Evasive Steering" or "Automatic Emergency Steering") which allows to avoid an obstacle by deviating the vehicle from its trajectory, either by acting on the steering of the vehicle, or by acting on the differential braking system of the vehicle; - Lane Keeping Assist (LKA), which keeps the vehicle in its lane by acting on the direction of the vehicle before the vehicle crosses a lane separation marking; - Lane Centering Assist (LCA), which allows the vehicle to be driven along the center line of its lane by acting on the steering of the vehicle; - Semi-Automatic Lane Change (SALC), which allows the vehicle to change lanes automatically after the driver has activated the turn signals, by acting on the steering of the vehicle.

[0007] These systems all have in common that they act on the direction of the vehicle. They use a "controller" to calculate a steering command for the vehicle, which is then sent to the power steering actuator.

[0008] One of the difficulties in developing these systems concerns the management of the interaction between the controller and the driver.

[0009] Typically, a poor balance between the torque exerted by the driver on the steering wheel and the torque exerted by the power steering actuator can create an unexpected feeling or misunderstanding for the driver, which may even lead the driver to make a dangerous decision.

[0010] We then know from document WO2017085377 a solution applicable to the LKA system only, which makes it possible to minimize the torque exerted by the power steering actuator when this torque is in the opposite direction to that exerted by the driver.

[0011] However, this solution is not applicable to the other systems mentioned. Furthermore, it only addresses one specific scenario (when the pairs have opposite directions).

[0012] We therefore wish to find a solution that ensures better management of the interaction between the controller and the driver, and that is applicable to all the aforementioned systems (LCA, LKA, SALC, AES). Presentation of the invention

[0013] In this context, the present invention proposes to use two elementary controllers, one of which guarantees better trajectory tracking and the other better comfort, then to calculate the overall controller according to the situation (in particular according to how the driver acts on the steering wheel), favoring one or the other of the two elementary controllers.

[0014] More particularly, the invention proposes a method as defined in the introduction, in which the global controller used is a function of two elementary controllers, one of which guarantees better trajectory tracking than the other, and of a weighting coefficient of one of the two elementary controllers relative to the other.

[0015] In other words, the global controller is adaptive, taking into account the driver's intention (and here more specifically taking into account the torque exerted by the driver on the steering wheel).

[0016] Thus, thanks to the invention, the arbitration between the torque setpoint calculated by the global controller and the torque applied by the driver is managed by the weighting coefficient, which allows one of the two elementary controllers to be given more or less importance than the other, depending on its value.

[0017] This solution is simple from a computational complexity standpoint. Its stability is particularly easy to verify. Furthermore, it relies on only a limited number of measured data points, such as the torque exerted by the driver on the steering wheel and the vehicle's trajectory relative to the desired trajectory.

[0018] Another advantage of the invention is that it does not affect the architecture already used by the applicant to implement vehicle steering control. The invention simply consists of a modification to the feedback loop, which does not require any changes to the loop's inputs and outputs. It merely allows for modification of the output values.

[0019] Other advantageous and non-limiting features of the process according to the invention, taken individually or in all technically possible combinations, are as follows: - the overall controller is equal to the sum of one of the two elementary controllers weighted by said coefficient, and the other of the two elementary controllers weighted by a complementary value of said coefficient; - the global controller is a convex combination of the two elementary controllers; - the coefficient is determined as a function of a force or torque that an individual applies to a steering system that is adapted to influence the trajectory of said device; - the coefficient is determined by equations which differ depending on whether the said force or the said torque is greater than or less than a predetermined threshold; - the coefficient is determined so as to vary at a rate of increase or decrease always below a predetermined speed threshold; - the coefficient is determined based on at least one first parameter relating to an action that an individual applies to a steering system adapted to influence the trajectory of said device and / or at least one other parameter relating to the posture of said device relative to said trajectory or to a lateral velocity of said device; - the coefficient is determined based on at least one parameter chosen from: a parameter relating to a force or torque that the individual applies to the steering system, n a parameter relating to a time derivative of the force or torque, n a parameter relating to a lateral deviation between the device and said trajectory, and n a parameter relating to a lateral velocity of the device; - if the force or torque is greater than the predetermined threshold, the coefficient is determined as a function of said at least one first parameter and said at least one other parameter; - if the force or torque is less than the predetermined threshold, the coefficient is determined independently of the posture of said device with respect to said trajectory (it is preferably zero or a function of said at least one first parameter); - the device is a motor vehicle which is adapted to drive on roads and which includes at least one steering wheel, in which case said actuator is adapted to control the steering of said at least one steering wheel, and the steering command is a steering angle command of said at least one steering wheel.

[0020] The invention also proposes an automotive device comprising at least one actuator which is adapted to influence the trajectory of said device and a computer to control said actuator, in which the computer is programmed to implement a process as described above.

[0021] Of course, the various features, variants, and embodiments of the invention can be combined in various ways, provided they are not incompatible or mutually exclusive. Detailed description of the invention

[0022] The description that follows, with regard to the attached drawings, given by way of non-limiting examples, will make it clear what the invention consists of and how it can be carried out.

[0023] Regarding the attached drawings:

[0024] [Fig.1] is a schematic top view of a motor vehicle travelling on a road and which is adapted to implement a method according to the invention;

[0025] [Fig.2] is a representation of a reference frame attached to the motor vehicle of [Fig.1] and of various variables used in the process according to the invention;

[0026] [Fig.3] is a schematic top view of the motor vehicle and a trajectory to follow;

[0027] [Fig.4] is a diagram illustrating a closed-loop transfer function used in the method according to the invention to control the motor vehicle of [Fig.1];

[0028] [Fig.5] is a graph illustrating three exponential functions.

[0029] Device

[0030] In [Fig.1], a motor vehicle 10 is shown.

[0031] This is a car, but it could also be a truck, a bus...

[0032] This motor vehicle 10 typically comprises a chassis and bodywork elements that define a passenger compartment, two steerable front wheels 11, and two non-steerable rear wheels 12. Alternatively, these two rear wheels could also be steerable with an adaptation of the control law shown below.

[0033] This motor vehicle 10 includes a conventional steering system for controlling the orientation of the front wheels 11 so as to turn the motor vehicle 10. This conventional steering system includes, in particular, a steering wheel 14 ([Fig.4]) connected to tie rods to rotate the front wheels 11, and an actuator for controlling the orientation of the front wheels according to a request received from a computer 13. This could typically be an electric power steering actuator acting on a steering column or on a rack of the motor vehicle 10.

[0034] Computer 13 is then intended to control this power steering actuator. For this purpose, it includes at least one processor, at least one memory, and various input and output interfaces.

[0035] Thanks to its input interfaces, the calculator 13 is adapted to receive input signals from various sensors.

[0036] Among these sensors, the following are planned, for example: - a device such as a front-facing camera, enabling the position of the motor vehicle 10 to be determined in relation to its lane of travel, - possibly one (or more) lateral devices such as a RADAR or LIDAR remote detector, allowing observation of the environment on the sides of the motor vehicle 10, - a device such as a gyroscope, enabling the determination of the yaw rate (around a vertical axis) of the motor vehicle 10, and - a position (and angular velocity) sensor for the steering wheel.

[0037] Thanks to its output interfaces, the calculator 13 is adapted to transmit a command to the power steering actuator F.

[0038] It thus allows the motor vehicle 10 to be forced to follow a particular trajectory, such as an obstacle avoidance trajectory, or a trajectory to stay in its lane, or a trajectory to center itself on its lane, or a trajectory to change lanes.

[0039] Thanks to its memory, calculator 13 stores data used in the process described below.

[0040] In particular, it stores a computer application, consisting of computer programs including instructions whose execution by the processor allows the computer to implement the process described below.

[0041] Variables

[0042] Before describing this process, we can introduce the different variables that will be used, some of which are illustrated in figures 1 and 2.

[0043] The total mass of the motor vehicle 10 will be denoted "m" and will be expressed in kg.

[0044] The inertia of the motor vehicle 10 about a vertical axis passing through its center of gravity CG will be denoted "J" and will be expressed in Nm

[0045] The distance between the center of gravity CG and the front axle of the motor vehicle 10 will be noted as "l f and will be expressed in meters.

[0046] The distance between the center of gravity (CG) and the rear axle will be noted as "l r and will be expressed in meters.

[0047] The front wheel drift stiffness coefficient will be noted as "C f » and will be expressed in N / rad.

[0048] The rear wheel drift stiffness coefficient will be noted as "C r» and will be expressed in N / rad.

[0049] These wheel drift stiffness coefficients are concepts well known to those skilled in the art. For example, the front wheel drift stiffness coefficient 11 is the one used to write equation F f = 2. C f .has f , with F f the lateral sliding force of the front wheels and a f the drift angle of the front wheels.

[0050] The steering angle (measured or estimated) that the front steering wheels 11 make with the longitudinal axis Al of the motor vehicle 10 will be noted "ô" and will be expressed in rad.

[0051] The variable ô ref , expressed in rad, will denote the saturated steering angle setpoint, as it will be transmitted to the power steering actuator.

[0052] The variable ô K, expressed in rad, will denote the unsaturated steering angle setpoint. At this stage, we can only specify that the concept of saturation will be linked to steering angle and steering speed limits that would not necessarily be respected with the variable θ K , but which would be with the variable ô ref .

[0053] The variable ô sat , expressed in rad, will denote the semi-saturated steering angle setpoint. It is derived from the unsaturated setpoint ô K and is saturated in steering angle only. The saturated setpoint ô ref will be calculated based on this semi-saturated setpoint ô sat .

[0054] The reference point of the motor vehicle 10 will here have its center of gravity CG as its origin. Its abscissa Xv will be oriented along the longitudinal axis Al of the motor vehicle 10, and its ordinate Yv will be oriented laterally, on the left side of the motor vehicle 10.

[0055] The yaw rate of the motor vehicle 10 (around the vertical axis passing through its center of gravity CG) will be denoted "r" and will be expressed in rad / s.

[0056] The relative heading angle between the longitudinal axis Al of the motor vehicle 10 and the tangent to the predefined trajectory (desired trajectory of the vehicle) will be denoted "Ψ L » and will be expressed in radians.

[0057] The lateral deviation between the longitudinal axis Al of the motor vehicle 10 (passing through the center of gravity CG) and the predefined trajectory, at a sighting distance "1s" located in front of the vehicle, will be noted "y L " and will be expressed in meters."

[0058] The lateral deviation instruction between the longitudinal axis Al of the motor vehicle 10 (passing through the center of gravity CG) and the avoidance trajectory T0, at a sighting distance "1s" located in front of the motor vehicle 10, will be noted "y L _ refand will be expressed in meters.

[0059] The trajectory tracking error will be noted as "e yL and will be expressed in meters. It will be equal to the difference between the lateral deviation setpoint y L _ ref and the lateral gap y L .

[0060] The aforementioned "1s" aiming distance will be measured from the center of gravity CG and will be expressed in meters. It may typically be zero.

[0061] The drift angle of the motor vehicle 10 (angle that the velocity vector of the motor vehicle 10 makes with its longitudinal axis Al) will be noted "β" and will be expressed in rad.

[0062] The speed of the motor vehicle 10 will be noted as "V" and will be expressed in m / s.

[0063] The component of this velocity V along the ordinate Yv will be called the lateral velocity v y and will be expressed in m / s.

[0064] Process

[0065] The method according to the invention is intended to enable the motor vehicle 10 to precisely follow a trajectory, in autonomous mode, when a driving assistance function of the motor vehicle 10 is activated.

[0066] This driver assistance function could be an obstacle avoidance system (AES), or a lane centering system (LCA) or a semi-automatic lane change system (SALC).

[0067] In the remainder of this presentation, we will consider that this is a function of keeping the motor vehicle 10 in its lane of travel (LKA).

[0068] This process is implemented when the LKA function is activated and a trajectory for keeping the motor vehicle 10 in its lane has been calculated.

[0069] As referenced in [Fig.3], this function typically consists of: - detect a risk of the vehicle leaving its lane if it remains on its trajectory, and then, if such a risk appears, - plan a trajectory that avoids the vehicle 10 leaving its lane (this trajectory being referred to hereafter as the reference trajectory T0), - act on the power steering actuator so that the vehicle 10 follows the reference trajectory T0, ^ OOOOOO - stop the piloting of the power steering actuator once the motor vehicle 10 is close to the central axis AO of the traffic lane and its longitudinal axis Al is parallel to the central axis AO.

[0070] It should be noted that the method of activating the LKA function and calculating the reference trajectory T0 is not strictly speaking the subject of the present invention, and will therefore not be described here.

[0071] It should be noted here that the trajectory tracking is intended to be operated autonomously by the computer 13, but that it must also be able to be used as an aid for the driver when the latter is holding the steering wheel but is not exerting the necessary torque on the steering wheel to avoid the obstacle.

[0072] Before describing the process that will be executed by the computer 13 to implement the invention itself, we will be able in the first part of this presentation to briefly describe calculations which allow us to synthesize a global controller K (adapted to calculate a steering instruction for the motor vehicle 10 so that the latter follows the reference trajectory T0 in a stable and efficient manner), in order to understand where these calculations come from and on what springs they rely.

[0073] At this stage, it should be noted that there are many methods for synthesizing a controller, but we will only describe one as an illustrative example.

[0074] According to this method, the dynamic behavior of the motor vehicle 10 is considered to be modeled, for example, by an improved bicycle model.

[0075] Typically, the behavior of the motor vehicle 10 can be modeled using the following Mathl equation. [Math.l] Cy+c; C r L r -Cflf mV mV 2 0 0 0 mV 0 P'0 1 Cfh-CJr C4;+C / lf Cflf r J ■ JV 0 0 0 0 0 J 0 0 1 0 0 0 0 0 g y L + <5 re / + 0 w V ls V 0 0 0 5 0 0 0 0 0 - - CO 2 0 8 0 0 0 0 0 1 0 0 \ P reflecting 0 0 0 0 0 0

[0076] In this equation, the term co fis used to model the dynamics of the curvature of the reference trajectory T0 (which is calculated elsewhere and is therefore already known).

[0077] The term p ref The curvature of the reference trajectory T0 then allows us to take into account the trajectory of the motor vehicle 10 (and therefore the curvature of the road) in the modeling of the dynamic behavior of the motor vehicle 10.

[0078] This model, however, does not in itself allow for limiting the steering angle and turning speed of the vehicle's front wheels 11. Such a limitation is particularly important to ensure that the driver of the motor vehicle 10 is able to regain control of the vehicle at any time.

[0079] Such limitations can be expressed using the following equation, by a saturation in steering speed:

[0080] [Math.2] |<5 re / | <v

[0081] As well as using the following equation, by saturation in steering amplitude:

[0082] [Math.3] |<5 re y | < IJ

[0083] In equation Math 2, the coefficient u is a constant representing the maximum steering speed. This constant is defined either by calculation or based on a series of tests conducted on a test vehicle. For example, it is equal to 0.0491 rad / s, which corresponds to 0.785 rad / s at the steering wheel (i.e., 45° / s) if the steering ratio is equal to 16.

[0084] In equation Math 3, the coefficient q is a constant representing the maximum steering angle. This constant is defined either by calculation or based on a series of tests conducted on a test vehicle. For example, it is equal to 0.0328 rad, which in our example corresponds to 0.524 rad at the steering wheel (i.e., 30°).

[0085] The constraint expressed by the Math 3 equation makes it possible to limit the torque exerted by the power steering actuator so that an average driver can manually compensate for this torque.

[0086] Indeed, the greater the steering angle, the greater the force applied by the power steering actuator. This limitation ensures that the driver can regain control of the vehicle without having to counteract excessive force. This angle will then depend on the force applied by the type of actuator chosen.

[0087] Note that the aforementioned values ​​are given as an example and could alternatively be reduced (for example 25° / s and 20° to ensure greater comfort).

[0088] Here, we want to limit the angle and speed of steering of the steering wheels 11 not by imposing a sudden threshold, but rather by progressively saturating the amplitude of the setpoint and the variation of the setpoint.

[0089] Figure 4 shows the control architecture used to control the power steering actuator so that the motor vehicle 10 follows the reference trajectory T0 while respecting the aforementioned constraints.

[0090] In this figure, the SAT1 block illustrates the amplitude saturation of an unsaturated steering angle setpoint θ K It receives this instruction as input, the calculation method of which will be described below, and it provides as output a semi-saturated steering angle instruction θsat We observe that this block operates in an open loop.

[0091] The SAT2 block set illustrates the speed saturation of the semi-saturated steering angle setpoint θ sat It receives this semi-saturated input and outputs the saturated steering angle input. ref (the one that will be transmitted to the power steering actuator). We observe that this is a closed loop. In this set of SAT2 blocks, corresponding to a "pseudo rate limiter" function, an input summing modulator is therefore provided, which allows the calculation of the difference A between the semi-saturated steering angle setpoint θ sat and the saturated steering angle setting ref at the previous time step. It includes a multiplier block allowing this difference to be multiplied by a parameter X, and a saturation block preventing the derivative of the saturated steering angle setpoint from being exceeded. refaccording to the Math2 equation and an integrator block allowing obtaining the saturated steering angle setpoint ô ref (via a Laplace transform).

[0092] The parameter X represents the dynamics of the SAT2 blocks (for our application, we can consider X=500), and the larger X is, the closer this pseudo rate limiter gets to a "rate-limiter" function.

[0093] In [Fig. 4], the block P sys represents the open-loop system that describes the dynamics of the motor vehicle 10 and the behavior of the power steering actuator.

[0094] We observe that this block receives an input disturbance w and a saturated steering angle setpoint ô ref It provides an output vector y and an error z.

[0095] Here, this error z to be minimized is a function of the trajectory tracking error e yLand the relative heading angle between the longitudinal axis Al of the motor vehicle 10 and the tangent to the reference trajectory T0 (hereafter referred to as heading error) L ), which we know should be minimized. We can then write:

[0096] [Math.4] Z = ^ L +< W L

[0097] In this equation, the term 'a' is a tuning coefficient that allows us to adjust the error we wish to minimize (heading angle error or position tracking error). This choice of output error 'z' ensures both accurate position tracking and accurate heading tracking.

[0098] In [Fig. 4], a global controller K (also called a "corrector") is shown, which calculates the unsaturated steering angle setpoint. K .

[0099] This global controller K is a function of two elementary controllers K confort and K dyna which are synthesized in a similar way and which each allow the calculation of a preliminary unsaturated steering angle setpoint ôK >comfort, ô K . dyna Each of these elementary controllers K confort and K dyna receives as input a state feedback term (the output vector y) which depends on the state of the motor vehicle 10, and a saturation compensation term which depends on the saturated steering angle setpoint ô ref calculated at the previous time step.

[0100] The term saturation compensation makes it possible to strengthen the stability of the controller in non-linear mode, that is to say in cases where the control of the power steering actuator is saturated in amplitude or in speed.

[0101] The unsaturated steering angle setpoint ô Kwill then be calculated based on the preliminary instructions, unsaturated steering angle ô K oh K The proportion in which one or the other of these preliminary instructions will be taken into account in the calculation of the unsaturated steering angle instruction ô K will be determined by an Alpha block.

[0102] In practice, this Alpha block will allow the determination of a weighting coefficient a for one of the preliminary unsaturated steering angle instructions ô K.confort , oh K.dyna compared to the other. This coefficient will be calculated based on the condition of the motor vehicle 10 and the actions of the driver 300 of the motor vehicle 10.

[0103] We can then write:

[0104] [Math.5] ~ comfort + (1 - ® ) ^K, dyna

[0105] - Example of a summary

[0106] Preferably, the two elementary controllers K confort and K dyna are synthesized according to strictly identical methods, but on the basis of at least partly different data so that one of the elementary controllers prioritizes the comfort of the driver 300 (in particular his feeling at the wheel) while the other prioritizes the following of the reference trajectory T0 (in particular in terms of speed).

[0107] In this presentation, we will not explain in detail how these elementary controllers are synthesized.

[0108] This summary is indeed well known, for example from document FR3121410A1 or document FR3127186 A 1.

[0109] We can only specify that it consists in particular of solving an optimization problem. Several ways of solving it are possible, and we can give here the main equations of one of these methods.

[0110] The method used in this example is the use of linear matrix inequalities (LMI). It is carried out using convex optimization criteria subject to linear matrix inequality constraints (the linearity of the terms of the matrices used ensures that the mathematical problem can be solved without requiring an excessive computational load). [YES] The objective is more precisely to optimize the gains of the closed loop defined by the elementary controller K confort or K dyna .

[0112] More precisely, if there exist matrices of appropriate dimensions R(p), Q, Li(p), L2(p), Ti(p), T2(p) such that the optimization problem below is feasible, then we obtain an elementary controller K confort or K dyna which satisfies the desired control objectives.

[0113] The matrix inequalities used here are three in number and are defined by the following inequalities, in which F is to be minimized.

[0114] [Math.6] QA T + QA'B' + + AQ + BA X Q + BBR QC* BB Ï T Ï L BT ^-L^p) B w * 0 0 0 -2T, -rX 0 -2T20 * * * * ■

[0115] [Math.7] Z» n< * (Jîj 1

[0116] [Math.8] Q QA^AR^ '2 * m-

[0117] In these matrix inequalities, we can define the following terms:

[0118] [Math.9] B p A = 0.

[0119] [Math.10] B =

[0120] [Math.11] HAS 1 = [0 -z |

[0121] [Math.12] Bi = At

[0122] [Math.13] C^-Cflf Cf mV + in\ mV Cflf-CJ. Crfi+Cflj Cfb JV -V Ap — V s - (I) 2 1 -Wf

[0123] [Math.14] B p - V

[0124] [Math.15] B w = Wf

[0125] In these inequalities, a matrix of the form ry 1 is written under the ly r w- form ry i. -* W-

[0126] The matrix variables R, Q, L|, L2, T T2 are expressed here in the form of matrices of appropriate dimensions.

[0127] The speed of the motor vehicle 10 is assumed to be constant (therefore all matrices in the system are constant).

[0128] Matrix variables are expressed in the following form:

[0129] [Math.16] Q = F 1 R(p) = K(p)F' 1\(p) = U\\p' l T2(p) = uj(.p) Li(p) =QG r i (p') L2(p) = QG T 2(p')

[0130] Each of the elementary controllers K confort or K dyna is then calculated using the following equation:

[0131] [Math.17] Æ ( = RQ'

[0132] With i = "comfort" or "dynamic".

[0133] The three matrix inequalities guarantee the stability of the system with or without perturbation.

[0134] The first of these inequalities also guarantees the performance (in the sense of the Hoo standard) of the closed-loop system when the system is subjected to a disturbance, with and without saturation of the control inputs.

[0135] The two elementary controllers K confort and K dynaThey are synthesized, one to guarantee good comfort and the other to guarantee good trajectory tracking using the same data, with the exception of the co parameter f to which two different values ​​will be given for the synthesis of these two controllers.

[0136] Indeed, the closer the value of this parameter is to zero, the more dynamic the synthesized controller will be (ensuring good trajectory tracking). Conversely, the higher its value, the more the synthesized controller will guarantee good comfort.

[0137] Thus, the previous optimization problem will be solved twice, once with a parameter co f of small value, and another time with a parameter co f of a larger value. The value of this parameter co fThis will typically be between 2.0.01 Hz and 2.0.1 Hz. The exact values ​​used may vary depending on the motor vehicle 10 and will be chosen following a test campaign.

[0138] Once the two elementary controllers K confort and K dyna defined, it will be necessary to verify that the global controller K is stable, regardless of the value of the coefficient a. Note that this global controller K is a convex combination of the two elementary controllers K confort and K dyna Here, it can be written in the following way:

[0139] [Math.18] K global ~ comfort + (1 " ^j^dyna

[0140] To verify that the global controller K is stable, it will be necessary to verify that there exists a (Lyapunov) matrix X such that the following two inequalities are satisfied:

[0141] [Math.19] / (A + BK con j ortJX + X ( A + BK con j or ( ) 0

[0142] [Math.20] [A + BKclyna)

[0143] Once the global K controller is well established, it can be stored in the memory of the vehicle computers.

[0144] - Implementation on the vehicle

[0145] With the calculation assumptions now well established, we can describe the process that will be executed by the computer 13 of the motor vehicle 10 to implement the invention.

[0146] Calculator 13 is programmed here to implement this process recursively, that is, step by step, and in a loop.

[0147] To do this, in a first step, the calculator 13 checks that the function concerned (here the LKA function) is activated and that a reference trajectory T0 has been planned.

[0148] If so, the calculator 13 will then seek to define a control instruction for the power steering actuator, allowing it to follow this reference trajectory T0 as closely as possible.

[0149] It begins by calculating or measuring parameters relating to the reference trajectory T0 and / or the posture of the motor vehicle 10 in relation to the reference trajectory T0 and / or the motor vehicle 10 itself (and more specifically then relating to the position of its steering system).

[0150] In the example presented above, calculator 13 therefore begins by calculating or measuring the terms of the output vector y, namely: - lateral velocity v y , - the measured steering angle, - the time derivative of the measured steering angle θ, - the yaw rate r, - the relative angle of heading L - trajectory tracking erroryL , and- the curvature p ref of the reference trajectory TO.

[0151] Calculator 13 then uses the elementary controllers K confort and K dyna stored in its memory, as well as the coefficient a (the calculation method of which will be described below), to determine the values ​​of the unsaturated steering angle setpoints ô K and saturated oh ref .

[0152] The steering angle setting is saturated. ref will then be transmitted to the power steering actuator to turn the wheels of the motor vehicle 10 in order to follow the reference trajectory T0.

[0153] Since the process is implemented in a loop, this instruction will naturally adapt to the conditions encountered by the motor vehicle.

[0154] As explained above, the coefficient used will allow us to prioritize either following the reference trajectory T0 or driving comfort.

[0155] - Alpha

[0156] We can then explain how this parameter is determined at each iteration of the loop.

[0157] The coefficient a is a variable between 0 and 1, inclusive.

[0158] By definition, if its value is equal to 0, then the global controller K is equal to the elementary controller K dyna Conversely, if its value is equal to 1, then the global controller K is equal to the elementary controller K confort All intermediate values ​​represent a compromise between these elementary controllers.

[0159] Here, we would prefer to use the elementary K controller. confortWhen the driver actively acts on the steering wheel, the torque exerted by the power steering actuator provides little or no resistance to the torque exerted by the driver on the steering wheel. Conversely, when the driver is not actively steering the vehicle, the elementary controller K is preferred. dyna , in order to have good tracking of the reference trajectory T0.

[0160] However, we do not simply want to use one or the other of these elementary controllers alternately. The global controller K used is instead the result of a compromise between the values ​​from these elementary controllers.

[0161] To obtain the best compromise between comfort and trajectory tracking (i.e., to calculate the coefficient a), at least two parameters can be defined. One of these parameters will take into account the torque exerted by the driver on the steering wheel, while the other will take into account the extent to which the motor vehicle 10 has an appropriate posture, particularly in relation to the reference trajectory T0.

[0162] Here, we will preferentially define four parameters.

[0163] "Parameter a T

[0164] The first parameter has T is relative to the pair r d that the driver applies to the steering wheel. This first parameter has T is calculated here via the following equation:

[0165] [Math.21] has r = min( l,

[0166] In this equation, r maxis a calibration parameter (to be determined by tests on a test vehicle). It defines the torque value r d applied to the steering wheel beyond which the driver is considered to be driving actively, so that we want to maximize driving comfort.

[0167] This first parameter oq is therefore a variable between 0 and 1, inclusive. It is equal to 0 when the couple r d is zero and equal to 1 when the pair r d exceeds the calibration parameter r max Otherwise, it increases linearly with the absolute value of the pair r. d .

[0168] n Parameter a d T / dt

[0169] The second parameter has d T / dt is relative to the derivative of the pair r d that the driver applies to the steering wheel. This second parameter has d T / dt is calculated here via the following equation:

[0170] [Math.22] . / i 'drldf\ \ ^dr / dt ~ 1, g Tmax )

[0171] In this equation, ôr max is a calibration parameter (to be determined by tests on a test vehicle). It defines the rate of change of the torque r d beyond which we consider that the driver is actively driving, so we want to maximize driving comfort.

[0172] This second parameter has d T / dt is therefore a variable between 0 and 1, inclusive. It is equal to 0 when the rate of change of the torque r d is zero and equal to 1 when this speed exceeds the calibration parameter ôr max Otherwise, it increases linearly with the absolute value of the derivative of the pair r d .

[0173] In summary, these first two parameters oq, a d T / dt They vary depending on the driver's actions on the steering wheel. They increase when the driver actively acts on the steering wheel.

[0174] n Parameter oq

[0175] The other two parameters o, oq vary according to the posture of the motor vehicle 10. By posture, we mean position and orientation.

[0176] Thus, the third parameter oq relates to the lateral deviation y L between the longitudinal axis Al of the motor vehicle 10 (passing through the center of gravity CG) and the reference trajectory T0, at the aiming distance "1s" located in front of the vehicle. Here, this aiming distance can be considered equal to 0. This third parameter has y is calculated here via the following equation:

[0177] [Math.23] has y = l-min( 1, +Sa v )

[0178] In this equation, y max and Say are calibration parameters (to be determined by testing on a test vehicle). max defines the value of the lateral deviation y LBeyond which point the difference is considered too large, so much so that trajectory tracking is prioritized over comfort. Its y This, in turn, prevents the parameter from having y is zero when the lateral deviation y L is zero, so that the global controller K remains even in this case a function of the elementary controller K dyna .

[0179] This third parameter has y is therefore a variable between Sa y and 1, inclusive. It is equal to 1-Sa y when the lateral gap y L is zero and equal to 0 when this lateral deviation y L exceeds the calibration parameter y max It decreases linearly with the absolute value of the lateral deviation y L .

[0180] Parameter a v

[0181] The fourth parameter has v is relative to the lateral velocity v yof the motor vehicle 10 (it is therefore not linked to the reference trajectory T0).

[0182] At this stage, as shown in [Fig.3], we can conventionally define this speed as positive when the motor vehicle 10 deviates from the reference trajectory T0, and negative when it approaches it.

[0183] This fourth parameter has v is then calculated here using the following equation:

[0184] [Math.24] tz r = l-max(0, min( 1, +Scq,) )

[0185] In this equation, v max and His v These are calibration parameters (to be determined through testing on a test vehicle). max defines the positive value of lateral velocity v y beyond which speed is considered very high, making it desirable to maximize trajectory tracking. v This, in turn, prevents the parameter from having vis zero when the lateral velocity v y is zero, so that the global controller K remains even in this case a function of the elementary controller K dyna .

[0186] This fourth parameter has v is therefore a variable between 0 and 1, inclusive. It is equal to 1 - Sa v when the lateral velocity is zero. It is equal to 0 at least when the lateral velocity exceeds the calibration parameter v max It tends towards 1 when the lateral velocity is very negative. Thus, it decreases when the lateral velocity increases.

[0187] In summary, the third and fourth parameters have larger values ​​the more the motor vehicle 10 correctly follows the reference trajectory T0.

[0188] a Couple threshold

[0189] To determine the best compromise (defined by the coefficient a) between comfort and trajectory tracking, we can also define a torque threshold r sma u (to be determined by tests on a test vehicle).

[0190] This threshold is used here to determine which of two predefined methods the coefficient a should be calculated.

[0191] To prevent discontinuous changes in the value of the coefficient a during a change of method, an algorithm for limiting the rate of variation of this coefficient a can also be used.

[0192] We can then describe these two methods as well as the algorithm for limiting the rate of change of this coefficient a.

[0193] Case No. 1

[0194] We can begin by considering the case where the couple r d applied by the driver on the steering wheel (in absolute value) is greater than or equal to the torque threshold

[0195] In this case, the method for calculating the coefficient a takes into account the four parameters mentioned above.

[0196] In the calculation of the coefficient a, the action of the conductor is quantified by taking into account only the larger of the values ​​of the two parameters oq and a dT / dt, so as to maximize driving comfort for all types of driver action (whether exerting a large and stable torque, or exerting a torque that varies rapidly).

[0197] Similarly, in the calculation of the coefficient a, the tracking of the reference trajectory T0 is quantified by taking into account only the smallest of the values ​​of the two parameters a y and a v .

[0198] The coefficient a is thus calculated here using the following mathematical formula:

[0199] [Math.25] has - ( w„.max ( cq.) + ( 1-M'a ).min(a,, cq) ) p

[0200] In this equation, w a and p are calibration parameters.

[0201] The calibration parameter w a allows the driver's action to be weighted in relation to following the reference trajectory T0.

[0202] Since trajectory tracking and driving comfort are opposing objectives, a compromise must be found. This is why this calibration parameter w a which allows this compromise to be made, which is then adjusted taking into account the four parameters mentioned above.

[0203] Thus, the more this calibration parameter w a The closer this parameter is to 1, the greater the influence of the driver's action will have on the final calculation of the coefficient a. Conversely, the closer this calibration parameter w is to 1, the greater the impact will be on the final calculation of the coefficient a. a will be close to 0, the more the tracking error will have a significant impact on the final calculation of the coefficient a.

[0204] The calibration parameter p, on the other hand, allows us to impose an exponential behavior on the coefficient a.

[0205] Thus, the coefficient a will approach 0 more quickly and 1 less quickly if this calibration parameter p is greater than 1. Conversely, the coefficient a will approach 1 more quickly and 0 less quickly if this calibration parameter p is less than 1. To better understand why, the exponential behavior of a function f(x)=x is represented in [Fig. 5]. p depending on whether the calibration parameter p is equal to 0.5 or 1 or 2.

[0206] In summary, to establish a compromise where more importance is given to driving comfort, the w calibration parameter should be chosen. a in a range between 0.5 and 1 and the calibration parameter p in a range between 0 and 1.

[0207] Conversely, to establish a compromise where more importance is desired for trajectory tracking, the calibration parameter w will be chosen. a in an interval between 0 and 0.5 and the calibration parameter p greater than 1.

[0208] In practice, these values ​​are stored in the computer and were chosen during the test campaign, taking into account the feelings of the test drivers.

[0209] Case #2

[0210] We can now consider the case where the couple r d applied by the driver on the steering wheel (in absolute value) is less than the torque threshold r sma n. In this eventuality, it is indeed understood that the driver does not actively act on the steering wheel so that trajectory following can be prioritized.

[0211] In this case, the method for calculating the coefficient a is different from that described above.

[0212] Thus, the coefficient a could simply be chosen to be equal to 0.

[0213] However, here, to achieve a more gradual transition of the coefficient a between 0 and 1 when the driver begins to act on the steering wheel, this coefficient is calculated based on at least one of the aforementioned parameters, and in this case, based on exactly two of these parameters. More precisely, parameters a are excluded. y and a v The reasoning behind not taking these parameters into account is that we do not want to reduce the dynamics of the global controller K if the tracking error is small.

[0214] The coefficient a is thus calculated here using the following mathematical formula:

[0215] [Math.26] a = max(a„ a dr / dt ) p

[0216] In this equation, p is still a calibration parameter allowing us to impose an exponential behavior on the coefficient a.

[0217] It will be observed that the couple r d being below the torque threshold r sma n, the parameters a T and a dT / dt (those quantifying the driver's action on the steering wheel) will take reduced values ​​so that the coefficient a will take a low value, guaranteeing good dynamics to the global controller K.

[0218] Speed ​​limit

[0219] As stated above, to avoid any discontinuous change in the value of the coefficient a when changing the calculation method, an algorithm for limiting the rate of variation of this coefficient a can be used.

[0220] Thus, this algorithm can be used to limit the rate of increase of the coefficient a to a speed threshold Sv1, and to limit the rate of decrease of the coefficient a to another speed threshold Sv2. These two thresholds are calibration parameters, the optimal values ​​of which can be found experimentally during the test campaign.

[0221] If the value of the coefficient a increases, we can write:

[0222] [Math.27] I da I 1 Â l ^1

[0223] If the value of the coefficient decreases, we can write:

[0224] [Math.28] I da I / y 1^1 < 5v2

[0225] Typically, this algorithm can be implemented in a similar way to that described above to determine the saturated steering angle setpoint ô ref (at the time, we were talking about a "pseudo rate limiter" function).

[0226] The present invention is in no way limited to the embodiment described and represented, but a person skilled in the art will be able to make any variation in accordance with the invention.

[0227] Thus, the process can be applied to other lateral steering systems of the motor vehicle (LCA, SALC, AES...).

[0228] More generally, it can be applied to other types of fields in which a particular trajectory must be followed, for example in aeronautics or robotics.

[0229] In robotics, it could be envisaged that the individual piloting the device (previously called the driver) would not be on board the device and that he would have a joystick equipped with means of measuring the force applied to him.

[0230] According to a variant of the invention described, more elementary controllers could be used and therefore several weighting coefficients could be used to weight these elementary controllers in the calculation of the overall controller.

[0231] According to another variant, a single method for calculating the coefficient a could have been used, in which case no threshold for the couple r would be used. sma n.

Claims

Demands

1. A method for autonomously controlling a steering control actuator of an automotive device (10), comprising the steps of: - acquiring a desired trajectory (T0) for the device (10), - acquiring parameters (v y , r, L , y L , ô, dô / dt, p ref ) relating to said trajectory (T0) and / or to a posture of said device (10) with respect to said trajectory (T0) and / or said device (10), and - calculation by a computer (13) of a piloting instruction (ô ref ) of said actuator according to said parameters (v y , r, L , y L , ô, dô / dt, p ref ), by means of a global controller (K), characterized in that the global controller (K) is determined as a function of two elementary controllers (K confort K dyna), one of which guarantees better tracking of the trajectory (T0) than the other, and a weighting coefficient (a) for one of the two elementary controllers (K confort K dyna ) compared to the other.

2. A control method according to claim 1, wherein the overall controller (K) is equal to the sum of one of the two elementary controllers (K confort K dyna ) weighted by said coefficient (a), and on the other hand of the two elementary controllers (K confort K dyna ) weighted by a complementary value of said coefficient (a).

3. A control method according to any one of claims 1 and 2, wherein the coefficient (a) is determined as a function of a force or a torque (r d ) that an individual (300) applies to a steering system that is adapted to influence the trajectory of said device (10).

4. A control method according to claim 3, wherein the coefficient (a) is determined by equations that differ depending on whether said force or said torque (r d ) is greater than or less than a predetermined threshold (r sma u).

5. A control method according to claim 4, wherein the coefficient (a) is determined so as to vary at a rate of increase or decrease always less than a predetermined speed threshold (Sv1, Sv2).

6. A control method according to any one of claims 1 to 5, wherein the coefficient (a) is determined as a function of at least one first parameter (oq, a d T / dt ) relating to an action that an individual (300) applies to a steering system that is adapted to influence the trajectory of said device (10) and / or at least one other parameter(a y , has v) relating to the posture of said device (10) in relation to said trajectory (T0) or to a lateral velocity of said device (10).

7. A control method according to claim 6, wherein the coefficient (a) is determined as a function of at least one parameter selected from: - a parameter (a T ) relating to a force or a couple (r d ) that the individual (300) applies to the steering system, - a parameter (a d T / dt ) relating to a time derivative of said force or said couple (r d ), - a parameter (a y ) relating to a lateral deviation (y L ) between the device (10) and said trajectory (T0), and - a parameter (a v ) relative to a lateral velocity (v y ) of the device (10).

8. A piloting method according to claims 4 and 6, wherein: - if the force or the torque (r d) is greater than the predetermined threshold (tsmaii), I e coefficient (a) is determined as a function of said at least one first parameter (a T , has d T / dt ) and said at least one other parameter (a y , has v ), And - if the force or the torque (r d is less than the predetermined threshold (r sma u), the coefficient (a) is determined independently of the posture of said device (10) with respect to said trajectory (T0).

9. A steering method according to any one of claims 1 to 8, wherein the apparatus (10) is a motor vehicle adapted for road use and comprising at least one steering wheel (11), wherein said actuator is adapted to control the steering of said at least one steering wheel (11), and wherein the steering command ( r ef) is a steering angle instruction of said at least one steering wheel (11).

10. Automotive device (10) comprising at least one actuator which is adapted to influence the trajectory of said device (10) and a computer (13) for controlling said actuator, characterized in that the computer (13) is programmed to implement a method according to any one of claims 1 to 9.

Images