Method for automatic steering of a motor vehicle

Abstract

The invention relates to a method for the automatic steering of a motor vehicle (10) comprising wheels (11, 12), at least two of which are steered wheels, comprising the steps of: - acquiring parameters relating to the path to be followed by the motor vehicle to avoid an obstacle, and - calculating, by means of a computer (13), a first steering setpoint of the steering actuators of the steered wheels and a second steering setpoint of at least one differential braking actuator of the wheels as a function of said parameters, wherein the first and second steering setpoints are each determined by means of a controller respecting a model of limiting setpoint variations and / or ranges.

Description

Technical Field

[0001] This invention generally relates to the automation of motor vehicle trajectory following.

[0002] This invention has particularly advantageous applications in motor vehicle driving assistance tools.

[0003] More specifically, the present invention relates to a method for automatic guidance of a motor vehicle, enabling the vehicle to follow an obstacle avoidance trajectory.

[0004] The present invention also relates to a motor vehicle equipped with a computer suitable for performing the method. Background Technology

[0005] Considering the safety of motor vehicles, they are currently equipped with driver assistance systems or autonomous driving systems.

[0006] Among these systems, the Automatic Emergency Braking system (more commonly known as AEB) is known, and it is designed to avoid any collisions with obstacles located in the vehicle's path. These systems are designed to detect obstacles on the road and, if an obstacle is detected, intervene in the motor vehicle's conventional braking system.

[0007] However, in some situations, these emergency braking systems cannot prevent a collision or are unusable (for example, if another vehicle is closely following behind the vehicle).

[0008] In response to these situations, automatic avoidance systems (more commonly known as the abbreviation AES, which stands for "automatic avoidance steering" or "automatic emergency steering") have been developed. These systems allow vehicles to avoid obstacles by intervening in the vehicle's steering or differential braking system to deviate from its trajectory.

[0009] An obstacle avoidance method is also known from document WO 2020099098, in which a steering angle and a differential braking system are controlled in combination to manage the avoidance trajectory. The steering angle system is used to ensure good stability at medium speeds, while the braking system is used at high speeds.

[0010] Sometimes, the trajectory imposed on the vehicle by the AES system is at its limit in terms of controllability, which does not allow the driver to take full control of the vehicle again. Summary of the Invention

[0011] To overcome the aforementioned drawbacks of the prior art, the present invention proposes the use of a hybrid controller that intervenes in the steering angle of the steering wheels and the differential braking of the left and right wheels of the vehicle, and is adapted to create a guide setpoint that limits the magnitude and / or rate of steering changes imposed on the motor vehicle.

[0012] More specifically, the present invention provides a method for automatic guidance of motor vehicles, the method comprising the following steps:

[0013] - Obtain parameters related to the trajectory of the motor vehicle avoiding obstacles, and

[0014] - The computer calculates the first guide setpoint of the steering angle actuator of the steering wheel and the second guide setpoint of at least one differential brake actuator of the wheel based on the parameters.

[0015] The first guide setpoint and the second guide setpoint are each determined by means of a controller that satisfies the setpoint amplitude and / or variation limit model.

[0016] Therefore, according to the present invention, a hybrid control law that intervenes in the steering angle and differential braking can be used to make the vehicle follow an obstacle avoidance trajectory. This control law is optimized to ensure good vehicle control performance (i.e., fast enough to ensure obstacle avoidance) and stability and robustness.

[0017] To this end, the present invention proposes imposing constraints on amplitude and rate (i.e., setpoint changes). The invention also preferably proposes to prohibit calculations when proven necessary, i.e., to guide the setpoint to stop control of the brakes or steering angle when conditions require it. This stop can be temporary or permanent until the AES system is reactivated.

[0018] The following are other advantageous and non-limiting features of the method according to the invention, which can be considered individually or in any technically feasible combination:

[0019] - The first guide setpoint and / or the second guide setpoint are determined by means of a controller that satisfies the setpoint amplitude and variation limit model;

[0020] - The first and second guide setpoints are determined based on a coefficient that fixes the contribution of each actuator (when guiding the vehicle to turn), the coefficient being calculated based on the yaw rate of the motor vehicle and the torque applied by the driver to the steering wheel of the motor vehicle.

[0021] The coefficient is also calculated based on a parameter whose value varies depending on whether the first guide setpoint is saturated (i.e., whether it is limited by the controller);

[0022] - Calculate this coefficient such that if the absolute value of the torque applied by the driver to the steering wheel of the motor vehicle exceeds a predetermined threshold, the second guide setpoint is zero;

[0023] - Calculate this coefficient so that when the vehicle's yaw rate is greater than the control threshold and / or when the first guide setpoint is able to stabilize the vehicle on its own, only the steering angle actuator of the steering wheel is used;

[0024] - Calculate this coefficient so that the steering angle actuator of the steering wheel and the differential brake actuator of the wheel can be used in combination in other aspects;

[0025] - Calculate this coefficient and make it change continuously over time;

[0026] - It is specified that when the absolute value of the torque applied by the driver to the steering wheel of the motor vehicle exceeds a predetermined threshold, the determination of the first guide setpoint is (preferably permanently) interrupted;

[0027] - It is stipulated that when the absolute value of the torque applied by the driver to the steering wheel of the motor vehicle exceeds another predetermined threshold, the determination of the second guide setpoint will be interrupted (temporarily or permanently);

[0028] - This enables the controller that can determine the second guide setpoint to satisfy the amplitude limiting model, so that the second guide setpoint remains less than or equal to the limit value;

[0029] - The limits are determined based on the vehicle's speed and the vehicle's yaw rate;

[0030] - In order to calculate the first pilot setpoint, it is stipulated that a first unsaturated setpoint is determined and the first setpoint is derived from it by means of a closed-loop pseudo-controller having a direct chain transfer function, which includes a hyperbolic tangent function of the deviation between the first unsaturated setpoint and the first setpoint.

[0031] - To calculate the second setpoint, a second unsaturated setpoint is determined and a second half-saturated setpoint is derived from it by means of a closed-loop pseudo-controller having a direct-chain transfer function that includes a hyperbolic tangent function of the deviation between the second unsaturated setpoint and the second half-saturated setpoint. Then, another closed-loop pseudo-controller is used, which provides the second setpoint as an output and whose direct-chain transfer function includes a hyperbolic tangent function of the deviation between the second half-saturated setpoint and the second setpoint.

[0032] The present invention also provides a method for creating a controller for use in the bootstrapping method as described above, wherein:

[0033] -Acquisition device behavior matrix model

[0034] - Determine at least some coefficients of the matrix of the behavioral matrix model.

[0035] - Two controllers are derived from this, each of which satisfies the requirement to follow the trajectory to be taken, i.e., satisfying the guide setpoint amplitude limit model and / or the guide setpoint change limit model.

[0036] The present invention also provides a motor vehicle including a steering angle actuator for a steering wheel, a differential brake actuator for a wheel, and a computer for guiding the actuator, the computer being programmed to perform the method described above.

[0037] Various features, variations, and embodiments of the present invention can of course be combined with each other in various combinations, as long as they are not incompatible or mutually exclusive. Detailed Implementation

[0038] The following description, with reference to the accompanying drawings given by way of non-limiting example, will clearly explain the content of the invention and how it can be practiced.

[0039] In the attached diagram:

[0040] [ Figure 1 A diagram is a schematic top-down view of a motor vehicle traveling on a road, showing the trajectory the vehicle will take.

[0041] [ Figure 2 ]yes Figure 1 A schematic 3D diagram of a motor vehicle, showing four consecutive positions along a trajectory to avoid obstacles;

[0042] [ Figure 3 [Illustration 1] is a simplified diagram showing the closed-loop transfer function used to implement the steering angle of the vehicle wheels as a saturation function;

[0043] [ Figure 4 [Illustration 1] is a simplified diagram showing the closed-loop transfer function of the saturation function used to implement differential braking of vehicle wheels;

[0044] [ Figure 5 [This is a simplified diagram illustrating the operational steps of determining values ​​that can be used in guiding motor vehicles according to the present invention;]

[0045] [ Figure 6 [ ] is a graph showing the various saturation and suppression functions that can be used within the scope of this process of guiding motor vehicles.

[0046] Figure 1 The vehicle 10 is designated as a motor vehicle, which conventionally includes a chassis defining a passenger compartment, two steerable front wheels 11, and two non-steerable rear wheels 12. As a variant, the two rear wheels may also be steerable, but this would require compliance with the command rules described below.

[0047] The motor vehicle 10 has a conventional steering system that allows intervention in the orientation of the front wheels 11 to enable vehicle steering. Specifically, the conventional steering system includes a steering wheel connected to a tie rod to pivot the front wheels 11. In the considered example, the conventional steering system also includes an actuator 31 (in... Figure 6 (as indicated in the text) The actuator enables intervention in the orientation of the front wheels based on the orientation of the steering wheel and / or based on a request received from the computer 13.

[0048] Additionally, the vehicle has a differential braking system that allows direct intervention on the front wheels 11 (and rear wheels 12) to simultaneously steer and decelerate the vehicle. This differential braking system includes, for example, a guide differential motor or electric motor positioned at the vehicle wheels, or brake calipers that are guided independently of each other. The differential braking system also has at least one actuator 32 (in... Figure 6 (As indicated in the diagram), the actuator is designed to directly intervene in the wheel rotation speed based on a request received from the computer 13. It is assumed here that the differential braking system has multiple actuators 32.

[0049] The computer 13 then directs the auxiliary steering actuator 31 and the differential braking system actuator 32 according to the traffic conditions encountered. For this purpose, the computer has at least one processor, at least one memory, and input and output interfaces.

[0050] The computer 13 is adapted to receive input signals from various sensors via its interface.

[0051] These sensors include, for example, the following:

[0052] - This enables devices that can identify the position of a vehicle relative to its lane, such as a front-facing camera.

[0053] - Enables the detection of obstacles 20 located on the trajectory of the motor vehicle 10. Figure 2 Devices such as RADAR or LIDAR remote detectors,

[0054] - At least one lateral detection device, such as a RADAR or LIDAR, that enables observation of the environment to the sides of the vehicle.

[0055] -A device, such as a gyroscope, that allows determination of the yaw rotation speed (about the vertical axis) of the motor vehicle 10.

[0056] - Sensors for the position and angular velocity of the steering wheel, and

[0057] - A sensor for the torque applied to the steering wheel by the driver.

[0058] The computer 13 is adapted to transmit one setpoint to the auxiliary steering actuator 31 and another setpoint to the actuator 32 of the differential braking system via its interface.

[0059] This also allows vehicles to be forced to follow trajectory T0 to avoid obstacle 20 (see Figure 2 ).

[0060] Computer 13 stores data used within the scope of the following methods using its memory.

[0061] Specifically, the computer stores application software consisting of computer programs including instructions, which the processor executes to enable the computer to perform the methods described below.

[0062] Before describing the method, we may introduce the different variables that will be used, some of which are... Figure 1 This was explained in the text.

[0063] The total mass of a motor vehicle will be expressed as “m” and in kg.

[0064] The inertia of a motor vehicle about a vertical axis passing through its center of gravity CG will be represented as "J" or "I". z , and expressed in Nm.

[0065] The distance between the vehicle's center of gravity CG and the front axle will be represented as "l". f , and expressed in meters.

[0066] The distance between the center of gravity CG and the rear axle will be represented as "1". r , and expressed in meters.

[0067] The cornering stiffness coefficient of the front wheels will be expressed as "C". f ", expressed in N / rad.

[0068] The cornering stiffness coefficient of the rear wheel will be expressed as "C". r ", expressed in N / rad.

[0069] These cornering stiffness coefficients of wheels are concepts well-known to those skilled in the art. For example, the cornering stiffness coefficient of the front wheel can therefore be expressed as equation F. f =2.C f .α f , where F f It is the lateral sliding force of the front wheel and α f It's the drift angle of the front wheels.

[0070] The steering angle formed by the front steering wheel and the longitudinal axis A1 of the motor vehicle 10 will be denoted as "δ" and expressed in rad.

[0071] δ, a variable expressed in rad ref The saturation steering angle setpoint will be transmitted to the auxiliary steering actuator.

[0072] δ, a variable expressed in rad K This will represent the unsaturated steering angle setpoint. At this stage, it can only be said that the concept of saturation will be associated with the value or the limit of value variation.

[0073] The yaw rate of the vehicle (about the vertical axis passing through its center of gravity CG) will be denoted as "r" and expressed in rad / s.

[0074] The relative heading angle between the vehicle's longitudinal axis A1 and the tangent to the avoidance trajectory T0 (the vehicle's desired trajectory) will be denoted as "Ψ". L , and expressed in rad.

[0075] At a visual distance "1s" in front of the vehicle, the lateral deviation between the longitudinal axis A1 (passing through the center of gravity CG) of the motor vehicle 10 and the avoidance trajectory T0 will be represented as "y". L , and expressed in meters.

[0076] At a visual distance "1s" in front of the vehicle, the set point for the lateral deviation between the longitudinal axis A1 (passing through the center of gravity CG) of the motor vehicle 10 and the avoidance trajectory T0 will be represented as "y". L-ref , and expressed in meters.

[0077] The trajectory following error will be represented as "e" yL ", expressed in meters. It will be equal to the lateral deviation setpoint y L-ref With lateral deviation y L The difference between them.

[0078] The aforementioned line-of-sight distance "1s" will be measured from the center of gravity CG and expressed in meters.

[0079] The drift angle of the motor vehicle 10 (the angle between the vehicle's velocity vector and its longitudinal axis A1) will be denoted as "β" and expressed in rad.

[0080] The speed of a motor vehicle along its longitudinal axis A1 will be represented as “V” and expressed in m / s.

[0081] The constant "g" will be the acceleration due to gravity, expressed in milliseconds (m / s). -2 Expressed in units.

[0082] The average curvature of the road surface and motor vehicles will be expressed as ρ ref and with m -1 Expressed in units.

[0083] The setpoint of the yaw moment applied by the differential braking device will be denoted as "M". z_ref , and expressed in Nm.

[0084] The constants “ξ” and “ω” will represent the dynamic characteristics of the steering angle of the vehicle’s front wheels.

[0085] In itself, the constant "ω" f "will represent the dynamic characteristics of any bounded disturbance "w" applied to the vehicle.

[0086] Steering angular velocity will represent the steering angular velocity of the front wheels.

[0087] As an assumption, the roads on which motor vehicles are traveling are assumed to be straight and flat. These roads extend along a main axis.

[0088] Before describing the method to be executed by computer 13 to implement the invention, in the first part of this explanation, the constraints selected to ensure the effective guidance of the motor vehicle can be summarized, and then the calculations that enable the implementation of the controller to carry out the invention can be described so as to make it clear where these calculations come from and what they are based on.

[0089] exist Figure 6 In part X1, the saturation steering angle setpoint δ that will be transmitted to the auxiliary steering actuator 31 has been simulated and calculated. ref and the yaw setpoint torque M to be transmitted to the differential brake actuator 32 z_ref The process.

[0090] Furthermore, part X2 is given to illustrate how these set points interact with actuators 31 and 32.

[0091] The second part, X2, shows the constraints applied to the aforementioned setpoints to ensure the driver's stability and controllability of the vehicle when an AES-type driver assistance system is activated.

[0092] Apply to the saturation steering angle setpoint δ ref The first constraint Z5 is speed saturation. More precisely, it is a limit on the vehicle's steering angular velocity. The steering angular velocity threshold used is denoted as v.

[0093] This first constraint provides the current steering angle δ _meas Feedback loop and controller K DAE This allows for the acquisition of temporary engine torque.

[0094] Applied to controller K DAEThe second constraint Z7 on the output is amplitude saturation. More precisely, it is a limitation on the absolute value of the aforementioned temporary engine torque. The engine torque threshold used is denoted as T. EPS_saturation_1 .

[0095] Box Z6 shows the setting point δ for disabling adjustment of the saturation steering angle. ref The first mechanism. When the absolute value of the torque applied to the steering wheel by the driver exceeds the value expressed as... When the threshold is reached, the first mechanism is designed to prevent regulation and thus prevent the calculation of any engine torque to be supplied to the auxiliary steering actuator 31.

[0096] Applying yaw moment setpoint M z_ref The third constraint Z8 is amplitude saturation. More precisely, it is the yaw moment setpoint M. z_ref The value of this limit is intended to prevent excessive differential braking and to ensure satisfactory stability and controllability of the vehicle. The yaw moment threshold used is called the limit M. z_max .

[0097] The limit M for this saturation z_max Preferably, it is variable. It will be calculated based at least on the vehicle's yaw rate r and speed V (in box Z10). Optionally, it also depends on the torque applied to the steering wheel by the driver.

[0098] Apply to amplitude saturation yaw moment setpoint M z_ref The fourth constraint Z9 is velocity saturation. More precisely, it is the constraint on the yaw moment setpoint M. z_ref The rate of change is limited. The threshold for the rate of change of yaw moment used is expressed as... (It also uses simplified symbols) express).

[0099] This fourth constraint is achieved through controller K. 制动 This allows the braking torque to be provided to the actuator 32 of the differential braking system.

[0100] This fourth constraint is considered preferred so that the yaw moment setpoint M can be closed or reopened. z_ref This improves the controllability of the vehicle.

[0101] Box Z11 shows the setting point M used to disable yaw moment adjustment. z_ref The second mechanism. When the absolute value of the torque applied to the steering wheel by the driver exceeds the value expressed as... When the threshold is reached, this second mechanism is designed to prevent regulation and therefore prevent the calculation of any differential engine torque.

[0102] The aforementioned thresholds υ and T EPS_saturation_1 and This was obtained by conducting road tests using a test vehicle of the same model as the vehicle on which the invention will be implemented.

[0103] Specifically, the threshold T EPS_saturation_1 It is obtained for different vehicle speeds V, as it varies according to this parameter. It is obtained through testing with the differential braking function deactivated.

[0104] Similarly, the threshold v and It was obtained through road tests at different vehicle speeds V, because they vary according to this parameter.

[0105] In itself, the limit M z_max It was obtained in a specific way using frame Z10.

[0106] If the absolute value of the vehicle's yaw rate r exceeds the yaw rate threshold (here denoted as r), ctrl If ), then the limit is chosen to be equal to denoted as The variable. Otherwise, the limit M z_max It will be selected as equal to zero.

[0107] This can be written in the form of the following equation, Math1:

[0108]

[0109] It can also be represented graphically, such as Figure 4 Part W1 is shown.

[0110] Yaw angular velocity threshold r ctrl This corresponds to the maximum yaw rate at which the vehicle can still be controlled by the driver at a given speed V.

[0111] Yaw angular velocity threshold r ctrl and variables The method involves determining or calculating through testing on test vehicles, or calculating through testing and then adjusting. This is the first solution applied here.

[0112] Therefore, during the first step, the differential braking function of the test vehicle is deactivated, and then a vehicle controllability test is conducted to determine the thresholds that can be used to constrain Z5 and Z7.

[0113] Next, during the second step, the vehicle's differential braking function is reactivated, and a new controllability test is conducted to determine the maximum steering angle threshold δ at multiple different speeds (e.g., in increments of 5 km / h). max (V). As a variant, these steering angle thresholds can be obtained through calculation.

[0114] Finally, during the third step, the yaw rate threshold r is derived from this. ctrl and variables

[0115] Yaw angular velocity threshold r ctrl The following equation is derived more precisely by modeling the vehicle using a bicycle model (described in more detail below):

[0116]

[0117] In this equation, k is the understeer gradient, which is calculated using the following equation:

[0118]

[0119] Using the same modeling, the variables can be obtained through the following equation.

[0120]

[0121] It can be resolved now. Figure 6 The two inhibition mechanisms are represented by boxes Z6 and Z11. Each of these mechanisms involves a threshold.

[0122] Used to negate the steering angle setpoint δ ref threshold It is determined through a controllability test performed on the vehicle while the differential braking function is deactivated. This threshold varies depending on the vehicle's speed V.

[0123] Used to negate the yaw moment setpoint M z_ref threshold This threshold is determined through a controllability test when the differential braking system is activated. It also varies depending on the vehicle's speed, V.

[0124] It can be noted that these thresholds satisfy the following inequality:

[0125]

[0126] In this inequality, the variable It is based on the threshold T EPS_Saturation_I Therefore, considering the function f corresponding to the auxiliary law of the auxiliary steering actuator 31, the following can be written:

[0127]

[0128] At this stage of the instruction manual, it can be used... Figure 5 This summarizes the process by which the values ​​of the aforementioned thresholds and variables are obtained to ensure vehicle controllability.

[0129] As shown in the figure, during the first step E1, when the differential braking function of the test vehicle is deactivated, a controllability test is performed to obtain the threshold T. EPS_saturation_1 and The value of .

[0130] Next, during the second step E2, the differential braking function is reactivated, and a series of new controllability tests are performed to obtain the maximum steering angle δ. max The value of .

[0131] In this step, a threshold is also calculated. The value of .

[0132] During the third step E3, considering the maximum steering angle δ max Simultaneous calculation of threshold and

[0133] Finally, during step E4, the results of the controllability test are used to refine the threshold. The value of and the threshold are determined. and driver torque The threshold value.

[0134] The various constraints mentioned above can now be set, as well as the methods for selecting between the steering wheel angle and the differential braking system in order to guide the vehicle.

[0135] Therefore, you can refer to Figure 6 The left part X1.

[0136] In this figure, box Z1 corresponds to the box that allows the determination of the trajectory to be followed to avoid obstacle 20. Since the method of determining this trajectory is not part of the subject matter of this invention, it will not be described here. Then, when the AES function is activated, box Z1 allows the determination of the lateral deviation setpoint y. L-ref and relative heading angle Ψ L .

[0137] Box Z2 allows for selection between the steering system and the differential braking system to optimally follow the avoidance trajectory. This allows for the determination of the coefficient α. DB The value of this coefficient represents the ratio of differential braking to the steering angle to be applied. When its value is zero, differential braking is deactivated, and when its value is the maximum (equal to 1), the steering angle is deactivated. This box will be described in detail below.

[0138] Box Z3 corresponds to the mathematical function that enables modeling of the aforementioned constraints Z5 and Z7. It receives the unsaturated steering angle setpoint δ. KAs input.

[0139] Box Z4 corresponds to the mathematical function that enables modeling of the aforementioned constraints Z8 and Z9. It receives the unsaturated yaw moment setpoint M. Kz As input.

[0140] These unsaturated setpoints are represented by K. δ (used to calculate the unsaturated steering angle setpoint δ) K ) and K M (used to calculate the setpoint M of the unsaturated yaw moment) Kz The controller obtains these values. They depend on the coefficient α. DB .

[0141] To understand the calculation of these two boxes, we can assume that the dynamic behavior of the vehicle can be modeled by the following equation Math7.

[0142]

[0143] This model is an improved bicycle model.

[0144] However, it does not thereby limit the steering angle and speed of the vehicle's front wheels 11, or the differential braking torque applied to the vehicle's wheels, or the variation of that braking torque. Nevertheless, these limitations prove particularly important to ensure that the driver can regain control of the vehicle at any time.

[0145] The saturation of the steering angular velocity can be formulated as follows:

[0146]

[0147] In this equation Math 8, the threshold v is, for example, equal to 0.0491 rad / s. If the gear reduction coefficient for steering is equal to 16, then this corresponds to 0.785 rad / s (i.e., 45° / s) at the steering wheel.

[0148] like Figure 3 As shown, the steering angular velocity limiter is noteworthy because it forms a closed-loop pseudo-controller (i.e., a controller that performs simple and finite calculations), which has the following characteristics:

[0149] - The direct chain transfer function equal to the threshold v (to satisfy the conditions specified in Math 8), the product of the 1 / s integrator and the corrector, which is a hyperbolic tangent function of Δ.α.

[0150] - An indirect chain (or "feedback chain") transfer function equal to one.

[0151] Hyperbolic tangent functions are designed to represent a variety of functions that have a form similar to the hyperbolic tangent function, including inverse trigonometric functions (such as arctangent), error functions (usually denoted as erf), Goodman functions (usually denoted as gd), and hyperbolic trigonometric functions (such as hyperbolic tangent).

[0152] Its receiving unsaturated steering angle setpoint 8 ref As input, and transmit the saturation steering angle setpoint δ ref As output.

[0153] In this figure, the coefficient Δ corresponds to the variable δ K With δ ref The deviation between them. The coefficient α is a constant between 0 and infinity, and it is the only parameter that can affect the speed or agility of the steering angle limiter.

[0154] Therefore, this steering angular velocity limiter has the advantage of being easy to adjust because it is sufficient to adjust the coefficient α. This allows for continuous and smooth (infinitely differentiable) commands.

[0155] Therefore, considering the form of the steering angular velocity limiter L2, the following equation Math9 can be written:

[0156]

[0157] Therefore, a pseudo-linear controllability model of the vehicle is obtained.

[0158] More precisely, the following parameter θ can then be introduced. δ :

[0159]

[0160] And then equation Math 10 can be rewritten in the following form:

[0161]

[0162] Equation Math 11 is a characteristic of the state representation, and it shows the setpoint variation constraint model as parameter θ. δ The function is linear.

[0163] Based on this, a controller can then be determined that ensures good adherence to the avoidance trajectory T0, satisfies the setpoint variation constraint model, and conforms to the coefficient α. DB The prescribed conditions.

[0164] Yaw moment setpoint M z_ref The saturation can be formulated as follows:

[0165] [Math 12]

[0166] |M z_rsf |≤M z_max

[0167] Yaw moment setpoint M z_ref The saturation of the rate of change can be formulated as follows:

[0168]

[0169] According to the present invention, the yaw moment setpoint M z_ref The amplitude and rate of change are intended to be limited not by imposing strict thresholds, but by using setpoint amplitude limiters and setpoint change limiters.

[0170] refer to Figure 4 First, it can be described that this will be used to modify the unsaturated yaw moment setpoint M. Kz In order to generate the half-saturated yaw moment setpoint M z_sat Amplitude limiter.

[0171] This amplitude limiter is noteworthy because it forms a closed-loop pseudo-controller, which has the following characteristics:

[0172] - equals the limit M z_max (To satisfy the conditions specified in Equation Math 12) and the direct chain transfer function of the product of the corrector, which is a hyperbolic tangent function of β1.ε1.

[0173] - An indirect chain (or "feedback chain") transfer function equal to one.

[0174] In Figure 4 In the equation, coefficient ε1 corresponds to variable M. z_sat With M Kz The deviation between them. The coefficient β1 is a constant between 0 and infinity, which is the only parameter that can affect the speed or flexibility of the amplitude limiter.

[0175] Using this pseudo-controller allows for precise control of the yaw moment setpoint M. z_ref It imposes time limits and also ensures the continuity of changes in the setpoint.

[0176] Considering the form of this amplitude limiter, equation Math14 can be written as:

[0177] [Math 14]

[0178] M z_sat =M z_max .tanh(β1.(M z_sat -M Kz ))

[0179] like Figure 4 As shown, the rate of change limiter of the yaw moment is also noteworthy because it forms a closed-loop pseudo-controller with the following characteristics:

[0180] - equal to a constant (To meet the conditions specified in Equation Math 13), the direct-chain transfer function of the product of the 1 / s integrator and the corrector, which is a hyperbolic tangent function of β²ε².

[0181] - An indirect chain (or "feedback chain") transfer function equal to one.

[0182] Its receiving half-saturated yaw moment setpoint M z_sat As input, and transmit the yaw moment setpoint M z_ref As output.

[0183] In this figure, the coefficient ε2 corresponds to the variable M. z_ref With M z_sat The deviation between them. The coefficient β2 is a constant between 0 and infinity, and it is the only parameter that can affect the speed limiter's speed or flexibility.

[0184] Using this corrector allows for effective time-limiting of yaw moment variations, while also ensuring the continuity of these variations.

[0185] Considering the form of this speed limiter, equation Math15 can be written as:

[0186]

[0187] Therefore, a pseudo-linear controllability model of the vehicle is obtained.

[0188] We can now introduce some variables to simplify the expressions of equations Math 14 and Math 15 so that the entire model can be represented quasi-linearly in the form of state representation (i.e., in the form of LPV (linear parameter variation)).

[0189] Equation Math 14 can be written in the form of equation Math 16:

[0190]

[0191] in,

[0192]

[0193] Then we can introduce the following parameter θ Δ :

[0194]

[0195] Next, equation Math 15 can be rewritten as equation Math 19:

[0196]

[0197] in,

[0198]

[0199]

[0200]

[0201]

[0202] Equation Math 19 is a property of the state representation, and it shows the full yaw moment amplitude and velocity limit model as external parameters. and The function of (parameters that can be calculated when the vehicle is in motion) is quasi-linear.

[0203] The bicycle model in Equation Math 7 can then be enhanced with this state representation to obtain new usable models.

[0204] In this way, the controller can be synthesized using optimization methods such as linear matrix inequalities (in... Figure 6 K in Chinese δ K M (and boxes Z3 and Z4 are represented).

[0205] These controllers can then be implemented in the computer 13 of vehicle 10 in the production series of test vehicles.

[0206] For example, it can be described that the execution of the first controller (makes it possible to obtain the steering angle setpoint δ) ref The synthesis of the first controller (the controller that enables the yaw moment setpoint) will be performed in a similar manner, and therefore will not be described in detail here.

[0207] The enhanced bicycle model of equation Math 7 is written as:

[0208]

[0209] Then we can consider the state vector x, which can be written in the following form:

[0210]

[0211] The goal then is to determine controller K. δThe form is such that the unsaturated steering angle setpoint δ can be calculated based on the state vector x. K Status feedback.

[0212] To understand how to determine a controller K that is suitable in terms of both stability and speed δ Our behavioral model can now be written in a general form:

[0213]

[0214] In this equation, C y A is the identity matrix, A is the dynamic matrix, Bu is the command matrix, and B is the command matrix. w The perturbation matrix can be written in the following form:

[0215]

[0216] In itself, the controller K is defined as static state feedback. δ It can be expressed in the following form:

[0217] [Math 28]

[0218] δ K =K δ .x

[0219] To find the optimal controller K δ Various methods can be used.

[0220] The method used here is the linear matrix inequality. Therefore, it is based on the convex optimization criterion under the constraints of the linear matrix inequality.

[0221] The goal is to more precisely optimize the control system K by changing the selection of poles. δ Defined closed-loop gain.

[0222] Use three matrix inequalities, and they are defined by the following inequalities.

[0223] [Math 29]

[0224] A i Q+B i R+(A i Q+B i R) T +2μQ<0

[0225]

[0226]

[0227] In these inequalities, i equals 1 or 2, and then matrix A...i and B i It can be defined in the following ways:

[0228] [Math 32]

[0229] A1=A(θ δmin )

[0230] A2=A(θ δmax )

[0231] B1 = B u (θ δmin )

[0232] B2 = B u (θ δmax )

[0233] form The matrix is ​​written in the form

[0234] Controller K δ Defined by the following formula:

[0235] [Math 33]

[0236] K δ =RQ -1

[0237] Assume the vehicle's speed is constant (and therefore all matrices of the system are assumed to be constant).

[0238] These three inequalities ensure that the dynamics of the closed loop remain constrained. This is because, with these constraints, the poles of the closed loop are limited by the radius γ, the minimum distance μ to the imaginary axis, and the aperture angle. Within a defined area.

[0239] This method proves effective because it involves determining the steering wheel angle at each instant in a reasonable (and manageable by an average driver) manner and in a way that can be executed by the actuator. These constraints also ensure the stability of the closed loop.

[0240] The goal here is to minimize the radius γ. Once the controller K is obtained... δ The unsaturated steering angle setpoint can be calculated using the following formula:

[0241]

[0242]

[0243] value θ δmin and θ δmax This is introduced into these three matrix inequalities.

[0244] θ δ The value of δ K and δ ref The correlation between the deviations reflects the controller K δ The degree to which the controllability limit set by equation Math 8 is violated.

[0245] By definition, θ δ It lies between 0 (excluding the extreme values) and 1 (including the extreme values). When θ δ When the value equals 1, the unsaturated setpoint δ of the calculated steering wheel angle is... K It conforms well to the controllability limit. When it approaches 0, the calculated unsaturated setpoint δ for the steering wheel angle is... K A value of θ will impose too much steering dynamics, which can create a risk of vehicle instability. δ Taking a value between 0 and 1 does not meet the controllability limit, but there may be no risk of vehicle instability.

[0246] In other words, θ δmin and θ δmax The choice of value affects controller K δ The performance and robustness are directly affected. The interval [θ] δmin θ δmax The larger the value, the better the controller K. δ The smaller the range, the worse the performance of the controller K, but the stronger its robustness. Conversely, the smaller the range, the better the performance of the controller K. δ The better the performance, the worse its robustness.

[0247] Logically speaking, the value θ δmax Selected as equal to 1 (in this case, controller K) δ It works in linear mode because this is usually the case and does not violate controllability constraints.

[0248] On the other hand, the value θ δmin Determining this value requires a trade-off between performance and robustness. Determining this value is equivalent to determining δ. K With δ ref The absolute value of the deviation between them is used to impose a maximum threshold.

[0249] In summary, the controller K is used to calculate the specific motor vehicle model. δ The method lies in determining αDB v, α, θ δmin and θ δmax The value of .

[0250] Next, the method includes determining matrix A. i B iThe coefficients are then determined, and equations Math 29 to Math 31 are solved to derive the controller K. δ The controller ensures good adherence to the avoidance trajectory T0 and meets the steering angle setpoint variation limit model.

[0251] Now we can describe box Z2 in more detail.

[0252] Now we can describe how we obtain the coefficient α. DB The method described in the review explains the ratio of differential braking and steering angle to be applied.

[0253] Therefore, we first choose the initial variable α. DB_raw .

[0254] Only when the absolute value of the torque applied by the driver to the steering wheel is less than or equal to the threshold. That choice was made only when the time was right.

[0255] Specifically, if it exceeds the threshold, the set point will not be calculated.

[0256] When it is less than or equal to the threshold In the case that the following two cumulative conditions are met, the initial variable α DB_raw It is selected as equal to 1. Otherwise, it is selected as equal to zero.

[0257] The first condition is the variable θ δ The absolute value is less than or equal to the variable θ δ min The absolute value of.

[0258] The variable θ δ min It is the variable θ δ The minimum value, at which the steering angle controller K... δ It is capable of stabilizing the vehicle itself. It is a variable because it depends on the vehicle's speed V.

[0259] In other words, the first condition is to check whether the steering angle setpoint is saturated.

[0260] The second condition is that the absolute value of the yaw rate is less than or equal to the yaw rate threshold r. ctrl The absolute value of.

[0261] Once the initial variable α is selected DB_raw The value of α (0 or 1) can be used to calculate the coefficient α using the following equation. DB :

[0262]

[0263] In this equation, the parameter τ DBIt is a time constant, which allows the filtering coefficient α to be used. DB This parameter is designed to prevent any sudden changes in comfort, ensuring a good driving experience. Therefore, its value can be adjusted according to the driver's desired feel.

[0264] The parameter s is a Laplace variable.

[0265] At this stage, the process to be executed by computer 13 of one of the aforementioned production series of motor vehicles can be described in order to implement the present invention.

[0266] Here, computer 13 is programmed to perform the process recursively (i.e. step by step and cyclically).

[0267] Therefore, during the first step, computer 13 checks whether the Automatic Obstacle Avoidance Function (AES) is activated.

[0268] If this is the case, it attempts to detect the presence of potential obstacles in the path of motor vehicle 10. For this purpose, it uses a RADAR or LIDAR remote detector.

[0269] This step is repeated cyclically when there are no obstacles.

[0270] Once an obstacle is detected 20 (see Figure 2 Computer 13 then plans an avoidance trajectory T0 that can avoid the obstacle 20.

[0271] Then, computer 13 will attempt to define guide setpoints for the conventional steering system and differential braking system, which allows for optimal adherence to the avoidance trajectory T0.

[0272] Therefore, it begins by calculating or measuring parameters that specifically characterize the vehicle's dynamic behavior, such as:

[0273] -Measured steering angle δ,

[0274] -The time derivative of the measured steering angle δ

[0275] - The saturation steering angle setpoint δ obtained in the previous time increment ref ,

[0276] -Yaw rate r,

[0277] - Relative heading angle Ψ L ,

[0278] - Lateral deviation setpoint y L-ref Time derivative,

[0279] -Track follows error e yL ,

[0280] -Drift angle β,

[0281] -coefficient α DB .

[0282] like Figure 6 As shown, computer 13 then uses controller K stored in its memory. δ and K M Therefore, these controllers will make it possible to determine the unsaturated steering angle setpoint δ. K Unsaturated yaw angle setpoint M Kz The value of .

[0283] Next, Figure 6 The pseudo-controller represented by boxes Z3 and Z4 will allow the steering angle δ to be derived from it. ref and yaw moment M z_ref The saturation setpoints are then transmitted to actuators 31 and 32 to cause the vehicle to deviate from its initial trajectory.

[0284] In other words, therefore controller K δ The combination of the pseudo-controller represented by box Z3 and the pseudo-controller forms the first master controller, enabling the steering angle setpoint δ to be determined based on the parameters listed above. ref .

[0285] Similarly, controller K M The combination of the pseudo-controller represented by box Z4 and the pseudo-controller forms a second master controller, allowing the yaw moment setpoint M to be determined based on the parameters listed above. z_ref .

[0286] This invention is by no means limited to the described embodiments. Rather, those skilled in the art will know how to add any variations according to the invention.

Claims

1. A method for automatically guiding a motor vehicle (10) having wheels (11, 12), at least two of which are steering wheels, the method comprising the steps of: - Obtain the parameters β, r, Ψ related to the trajectory of the motor vehicle (10) avoiding the obstacle (20). L e yL ,δ,δ ref ,in, The parameter β represents the angle between the vehicle's velocity vector and its longitudinal axis, expressed in rad. The parameter r represents the yaw rate of a motor vehicle about a vertical axis passing through its center of gravity, expressed in rad / s. Parameter Ψ L The relative heading angle between the longitudinal axis of the motor vehicle and the tangent to the desired trajectory of the motor vehicle, expressed in rad; Parameter e yL The error in trajectory following is expressed in meters. The parameter δ represents the steering angle between the front steering wheel and the longitudinal axis of the motor vehicle, expressed in rad. Parameter δ ref Indicates the saturation steering angle setpoint, expressed in rad; as well as - By computer (13) based on the parameters β, r, Ψ L e yL ,δ,δ ref Calculate the first guide setpoint (δ) of the steering angle actuator (31) of these steering wheels ref ) and the second guide setpoint (M) of at least one differential brake actuator (32) of these wheels (11, 12) z_ref ), Among them, the first guide setting point (δ) ref ) and the second guide setting point (M) z_ref Each of these is determined by a controller that satisfies the setpoint amplitude and / or variation limit model, and Among them, the first guide setting point (δ) ref ) and the second guide setting point (M) z_ref The coefficient α of the contribution of each actuator (31, 32) is fixed. DB To determine, the coefficient α DB Based on the yaw rate r of the motor vehicle (10), the torque applied by the driver to the steering wheel of the motor vehicle (10), and the parameter θ δ To calculate the parameter θ δ The value is based on the first guide setpoint (δ) ref Whether it is restricted by the controller.

2. The method as described in claim 1, wherein, The first guide setting point (δ) ref ) and / or the second guide setting point (M z_ref It is determined by using a controller that meets the setpoint amplitude and variation limit model.

3. The method as described in claim 1 or 2, wherein, Calculate the coefficient α DB , so that: - If the absolute value of the torque applied by the driver to the steering wheel of the motor vehicle (10) exceeds a predetermined threshold ( Then the second guide setting point (M) z_ref ) is zero, - When the vehicle's yaw rate r is greater than the control threshold (r ctrl ) and / or when the first guide set point (δ) ref When the vehicle (10) is able to stabilize itself, only the steering angle actuators (31) of these steering wheels are used, and - Otherwise, the steering angle actuators (31) of these steering wheels and the differential brake actuators (32) of these wheels (11, 12) are used in combination.

4. The method of claim 3, wherein, Calculate the coefficient α DB This causes it to change continuously over time.

5. The method as described in claim 1 or 2, wherein, The rule states that when the absolute value of the torque applied by the driver to the steering wheel of the motor vehicle (10) exceeds a predetermined threshold ( When interrupted, the first guide setting point (δ) is interrupted. ref The determination of ).

6. The method as described in claim 1 or 2, wherein, The rule states that when the absolute value of the torque applied by the driver to the steering wheel of the motor vehicle (10) exceeds a predetermined threshold ( When interrupting the second guide setting point (M), z_ref The determination of ).

7. The method as described in claim 1 or 2, wherein, This enables the determination of the second guide setpoint (M) z_ref The controller satisfies the amplitude limiting model, such that the second guide setpoint (M) z_ref Keep less than or equal to the limit (M) z_max ), and wherein the limit (M) z_max The value is determined based on the speed (V) and yaw rate r of the motor vehicle (10).

8. The method as claimed in claim 1 or 2, wherein, To calculate the first guide setpoint (δ) ref The first unsaturated setpoint (δ) is specified. K And the first pilot setpoint (δ) is derived from it using a closed-loop pseudo-controller with a direct-chain transfer function. ref The direct-chain transfer function includes the first unsaturated setpoint (δ). K ) and the first guide setting point (δ) ref The hyperbolic tangent function of the deviation (Δ) between ).

9. The method as claimed in claim 1 or 2, wherein, To calculate the second guide setpoint (M) z_ref ),Regulation: - Determine the second unsaturated setpoint (M) Kz And by means of a closed-loop pseudo-controller with a direct-chain transfer function, the second half-saturation setpoint (M) is derived. z-sat The direct-chain transfer function includes the second unsaturated setpoint (M). Kz ) and the second half-saturation setpoint (M) z-sat The hyperbolic tangent function of the deviation (ε1) between ) is then... - Use another closed-loop pseudo-controller that provides the second boot setpoint (M) z_ref The output is the second half-saturation setpoint (M), and its direct-chain transfer function includes this second half-saturation setpoint. z-sat ) and the second guide setting point (M) z_ref The hyperbolic tangent function of the deviation (ε2) between ).

10. A method for creating a controller for use in a bootstrapping method as described in any one of the preceding claims, wherein, Regulation: - Acquisition of the device's behavior matrix model, - Determine the matrix (A) of this behavioral matrix model. i B i At least some coefficients of ) - Two controllers are derived from this, each of which satisfies the requirement to follow the trajectory to be taken, i.e., satisfying the guide setpoint amplitude limit model and / or the guide setpoint change limit model.

11. A motor vehicle (10) comprising a steering angle actuator (31) for a steering wheel, a differential brake actuator (32) for a wheel (11, 12), and a computer (13) for guiding the steering angle actuator (31) and the differential brake actuator (32), characterized in that, The computer (13) is programmed to perform the method as described in any one of claims 1 to 9.

Images