CONTROL DEVICE FOR A SELF-PROPELLED MOTOR VEHICLE

Description

[0001] The present invention relates to a control device for an autonomously driving vehicle.

[0002] It applies to the field of automated vehicle driving, which notably includes functions such as lane keeping, speed regulation, and vehicle steering. The objectives of automated vehicle driving are, in particular, to improve the safety and efficiency of travel. It relies especially on active safety systems, incorporating driver assistance systems adapted to modify the vehicle's dynamic behavior in critical situations, such as ESP (Electronic Stability Program), which controls the stability of a vehicle's trajectory, notably through braking actions. For example, when a vehicle enters a curve at too high a longitudinal speed, it can be difficult to maintain the road's curvature, and the vehicle may begin to understeer.The ESP system then intervenes automatically to keep the vehicle on the trajectory desired by the driver. Generally, if the vehicle deviates from the driver's intended trajectory, the ESP system will send engine torque and / or braking torque command signals to correct the vehicle's trajectory.

[0003] For vehicle control systems such as the ESP system, the vehicle's yaw rate is a key parameter to know in order to maintain the vehicle's lateral stability.

[0004] Lateral stability is indeed crucial for passenger safety during extreme lateral maneuvers or when maneuvering in adverse driving conditions, such as on snow or ice, or in the event of a sudden loss of tire pressure or a sudden crosswind. Vehicle stability control systems are therefore used to improve lateral stability in such adverse conditions. To achieve this, as previously mentioned, yaw rate is a key variable that a vehicle stability control system must know. This variable can be measured by dedicated sensors onboard the vehicle. It can also be estimated using other onboard sensors such as lateral accelerometers and wheel speed sensors.The vehicle stability control system includes a state observer that estimates yaw rate information, which is not measured but is necessary for control. The state observer, built on a model of the vehicle's dynamics, receives the vehicle's speed and lateral acceleration to estimate the yaw rate.

[0005] For the algorithms implemented in vehicle stability control systems to function correctly, the input data they receive must be accurate, particularly to prevent system malfunctions and instability. As previously discussed, a number of these data points, including yaw rate, are calculated by observers based on a vehicle-specific dynamic model. This means that these observers must always function correctly; that is, the vehicle's operating range must always fall within the response limits where the vehicle model has been validated. In other words, it is essential to be able to continuously determine whether the vehicle is still operating within the model's validation range and thus whether the control system is still capable of managing the vehicle in automated driving mode.

[0006] Automated driving vehicles therefore require appropriate inputs to be able to properly control the vehicle, but it is also necessary to be able to detect critical situations where the nominal behavior of the control system is no longer assured.

[0007] The patent application, not yet published, filed in the name of the applicant and bearing application number FR1909681, concerns an application for automated driving vehicles, where the vehicle control system is designed to react to unstable situations in which the vehicle's physical limits may be reached. A typical example is a sharp turn where the vehicle may enter at a speed significantly exceeding the physical limits imposed by the grip between the tires and the road surface in the turn. This document describes a system where the vehicle model is capable of predicting the vehicle's future positions in autonomous mode along the entire path ahead, in order to identify in advance future vehicle positions corresponding to situations where the vehicle's driving limits are violated and thus make proactive decisions to prevent these situations.

[0008] However, while this system does allow for proactive reactions to prevent the vehicle from entering uncontrollable situations beyond its driving limits, it is not designed to react in real time to sudden and unpredictable changes in driving conditions that could impair the vehicle's automated driving mode. In particular, in certain driving situations, there may be a mismatch between the vehicle model response used by the vehicle stability control system and the actual vehicle behavior, which can lead to a highly unstable vehicle state.Such situations, where the vehicle may no longer be able to correct its trajectory in automated driving mode, can occur, for example, in the event of sudden and unpredictable changes in the nature of the tire-road contact (e.g., the presence of oil, sand, or gravel on the road that could compromise wheel grip), or in the event of a sudden tire blowout. In these situations, the vehicle model used by the control system may no longer be valid, which represents a potential danger since the entire control logic of the vehicle stability control system relies on the use of this model. The paper "Combining local trajectory planning and tracking control for autonomous ground vehicles navigating along a reference path" (LI XIAOHUI et al.) describes a system with similar drawbacks.JPH11160205A corrects a vehicle model based on the difference between the estimated yaw rate and the measured yaw rate, the estimated yaw rate depending on both the vehicle speed and the steering angle.

[0009] Also, there is a need to improve the robustness of the lateral stability control of a vehicle in automated driving mode, in all driving conditions, including in the aforementioned situations of sudden and unpredictable changes in these driving conditions.

[0010] Another need is to be able to detect in real time situations where control stability can no longer be ensured, due to a mismatch between vehicle behavior and the vehicle model response.

[0011] To this end, the invention relates to a control device for an autonomous or assisted driving motor vehicle, comprising a route planning module capable of storing information relating to a planned trajectory, a longitudinal movement controller of the vehicle capable of generating output control signals to acceleration and braking actuators, and a lateral movement controller of the vehicle capable of generating output control signals to steering actuators, so that the vehicle follows the planned trajectory in an assisted or automated driving mode, the output of said lateral controller being based on a minimization of the error between a desired yaw rate, obtained from the curvature of the road on the planned trajectory, and the current yaw rate of the vehicle.estimated from a model of the vehicle's dynamic behavior capable of providing, as input to said lateral controller, an estimate of the yaw rate based on the measured longitudinal speed of the vehicle and the measured steering angle, characterized in that it comprises a vehicle lateral dynamics compensation module capable of dynamically correcting model parameters in case of discrepancy between the model and the actual dynamic behavior of the vehicle, based on a comparison between the estimated yaw rate value and a yaw rate value measured by an on-board sensor, such that said estimate provided by the model to the lateral controller is corrected instantaneously.

[0012] Advantageously, the compensation module is adapted to estimate a compensated steering angle by means of an adaptive algorithm that takes into account the difference between the estimated value of the yaw angular velocity and the measured value of the yaw angular velocity, the output of the model being corrected on the basis of the compensated steering angle provided as input to the model.

[0013] According to the invention, the device includes a safety module, connected to the output of the vehicle's lateral dynamics compensation module, said safety module being adapted to detect driving degradation states in automated mode based on a discrepancy between the output of the corrected model and the actual dynamic behavior of the vehicle.

[0014] Advantageously, the corrected model is capable of providing an estimate of the linearized lateral forces exerted on the front and rear wheels of the vehicle, said safety module being capable of receiving said linearized lateral forces and comparing said linearized lateral forces with the lateral forces calculated using data provided by sensors on board the vehicle.

[0015] Preferably, the safety module includes a calculation module capable of calculating the lateral forces exerted on the front and rear wheels from values ​​of lateral acceleration of the vehicle, steering angle and yaw angular velocity, measured by the on-board sensors.

[0016] Preferably, the safety module includes a comparator module capable of performing the difference between linearized lateral forces and calculated lateral forces and of activating a safety mode when said difference exceeds a predefined threshold.

[0017] Advantageously, said safety module is capable of generating vehicle safety control signals for the longitudinal and transverse controllers, said safety control signals prevailing over the automated mode control signals when said safety mode is activated.

[0018] Preferably, these safety signals are adapted to control the stopping of the vehicle.

[0019] Advantageously, said safety module is capable of generating an alert signal to a human-machine interface of the vehicle when said safety mode is activated.

[0020] The invention also relates to a motor vehicle characterized in that it comprises a control device as described above.

[0021] Other features and advantages of the present invention will become more apparent upon reading the following description, given by way of illustrative and non-limiting example and with reference to the accompanying drawings: [ Fig.1 ] is a diagram illustrating an architecture of the control device of the invention; [ Fig.2 ] is a graph showing the evolution over time of the yaw angular velocity, as measured by a sensor onboard the vehicle and as calculated using the vehicle model, under driving conditions corresponding to a sloping road; Fig. 3 ] is a graph similar to that of the Fig. 2 , but under different driving conditions, corresponding to a slippery road; Fig. 4 ] is a synoptic diagram illustrating the operation of the vehicle's lateral dynamics compensation module shown in the figure 1 ; [ Fig. 5 ] is a synoptic diagram illustrating the operation of the security module shown in the figure 1 .

[0022] In reference to the figure 1 The vehicle control system 1 includes an autonomous driving control module 2. This control module 2 comprises a first controller 3, called the longitudinal controller, which receives as input a current vehicle speed value from an on-board vehicle speed sensor 4, and a reference speed value from a route planner module 5. This route planner module stores information relating to a planned route for the vehicle, including reference speed values ​​predicted along the planned route. The longitudinal controller 3 is adapted to generate engine torque and braking control signals 6, 7 to corresponding vehicle acceleration and braking actuators, so as to minimize the speed error between the current value and the reference value.

[0023] The control module 2 also includes a second controller 8, called the lateral controller, which receives as input a desired yaw rate value obtained from information from the route planner module 5 including road curvature, location information from a vehicle location module 9 and a current yaw rate value of the vehicle, corresponding to a calculated value 10 based on a model of the vehicle's dynamic behavior capable of providing an estimate of the yaw rate and which, according to the invention, is dynamically corrected by a vehicle lateral dynamics compensation module 11, designed to compensate for errors in the nominal vehicle model in certain driving situations as will be explained in more detail later.From these different inputs, the lateral controller 8 is adapted to generate steering angle control signals 12 to vehicle steering actuators, acting on the steering angle of the vehicle's steering wheels, in order to minimize the yaw error between the desired yaw angular velocity (obtained from road layout information as a function of curvature) and the current yaw angular velocity from the dynamically corrected vehicle model.

[0024] The action of the compensation module 11 aims to provide the lateral controller 8 with a correct value of the yaw angular velocity, by dynamically compensating the vehicle model to take into account lateral dynamics of the vehicle not modeled in the nominal model, which are caused by unexpected situations suddenly changing the vehicle's driving conditions, such as for example the nature of the wheel-ground contact and which are likely to bring the vehicle into an operating range outside the response limits where the nominal vehicle model has been validated.

[0025] The vehicle's lateral controller 8 can thus be fed by the output of the compensated model, improving the robustness of the control provided under all driving conditions, including these unexpected ones. However, model compensation that provides a vehicle yaw rate corresponding to the actual value is independent of potential saturation of the lateral forces acting on the vehicle, which can still lead to uncontrolled vehicle states. In other words, the vehicle's lateral controller can be fed with correct yaw rate values ​​thanks to the compensation applied to the model, but the vehicle may nevertheless be unable to correct its trajectory in autonomous mode because the vehicle's physical driving limits are exceeded.It is necessary to be able to detect these critical situations where the nominal control of the vehicle's lateral stability is no longer ensured, in order to be able to take measures guaranteeing the safety of the vehicle's passengers in these circumstances.

[0026] Therefore, once the lateral controller 8 is supplied with correct inputs, thanks to the first yaw rate correction step implemented by the model's compensation module 11, a second step, implemented by a safety module 13, consists of detecting whether the vehicle's physical driving limits are exceeded. When the maximum lateral capacities are reached, the vehicle and the model used for control do not behave identically, and a discrepancy can thus be detected between the lateral forces estimated by the model and the lateral forces calculated using data measured by the vehicle's onboard sensors. The safety module 13 therefore receives as input values ​​for the vehicle's lateral acceleration, steering angle, and yaw rate, as measured by onboard sensors 14, 15, and 16, respectively.The safety module 13 is also connected to the compensation module 11. The safety module 13 is designed to detect when the vehicle's physical limits are exceeded by comparing the linearized lateral forces from the corrected model output by the compensation module 11 to the lateral forces calculated using the lateral acceleration, steering angle, and yaw rate values ​​measured by the vehicle's onboard sensors. This comparison allows for the precise identification of lateral force saturation. The safety module 13 is connected to a human-machine interface 17 of the vehicle. If the comparison result exceeds a given threshold, the safety module 13 triggers the generation of an audible and / or visual warning signal at the vehicle's human-machine interface 14. The safety module 13 is also connected to the vehicle's longitudinal controller 3 and lateral controller 8.The safety module 13 is capable of generating a safety control signal for the controllers acting on the vehicle's actuators, intended to place the vehicle in a safe state, for example by stopping the vehicle, to ensure the safety of the passengers if the driver does not respond to the warning signal.

[0027] In summary, the vehicle control device 1 is adapted, on the one hand, via the compensation module 11, to provide a correct estimate of the yaw angular velocity and linearized lateral forces from the vehicle model, correcting the inadequacies between the response of the vehicle model and the actual behavior of the vehicle and, on the other hand, via the safety module 13, to detect degraded driving states in the automated mode of the vehicle, corresponding to driving situations that suddenly occur, where the stability of the control can no longer be ensured, so as to apply a preventive safety procedure allowing the vehicle to be made safe, typically stopping the vehicle, which prevails over the nominal commands for driving in automated mode.

[0028] We will now describe in more detail the respective operation of the compensation module 11 and the safety module 13.

[0029] The compensation module 11 is designed to provide, on the one hand, an appropriate yaw rate for the lateral controller 8 and, on the other hand, linearized lateral forces to be compared in the safety module 13. The objective here is to provide correct inputs for the lateral controller and the safety module by correcting the difference between the nominal model and the actual vehicle.

[0030] We will use the bicycle model, which is well-known in itself, to describe the dynamic behavior of the vehicle in yaw and drift. Thus, the yaw angular velocity and the lateral dynamics of the vehicle can be described by the following differential equation: I z ψ ¨ = aF f cos δ − bF r + M z m a y = F f cos δ + F r + g sin θ cos ϕ where a and b are the distances from the vehicle's center of gravity to the front and rear wheel axes, respectively, m is the vehicle's mass, and Iz is the moment of inertia about the vertical axis Z. ψ, θ, and φ are the vehicle's yaw angle, bank angle, and pitch angle, respectively. Mz, ay, and δ are the yaw moment, lateral acceleration, and steering angle imposed by a dedicated actuator, respectively. Ff and Fr are the front and rear lateral forces exerted on the vehicle's front and rear wheels through wheel / ground contact, enabling the vehicle to turn. They depend on the steering angle δ, the slip angle, and the vehicle's speed.

[0031] However, to control lateral movement in autonomous vehicles, the model described above must be simplified because the design of the lateral controller and the prediction of movement use a linear vehicle model and rely solely on signals available from inexpensive sensors. The simplification is achieved by first assuming that the lateral forces exerted on the front and rear wheels are proportional to the lateral slip angle when the vehicle is used within the linear range of forces exerted on the wheels. Furthermore, it is assumed that the smallest steering angle is defined such that: cos(δ) = 1 and sin(δ) = δ.

[0032] Let β be the angle between the total velocity vector of the vehicle, measured around the vehicle's center of gravity, and the longitudinal direction of the vehicle. We have: tan β = v y v x , soit β = tan − 1 v y v x

[0033] Where Vx and Vy are the speeds along the longitudinal and lateral axes of the vehicle.

[0034] As previously mentioned, the vehicle's yaw rate is a key variable for lateral control in autonomous vehicles. Therefore, this variable must be known precisely.

[0035] THE Figures 2 And 3 are graphs showing the evolution over time of the yaw angular velocity, in degrees per second, as measured by a sensor on board the vehicle (Ψ̇ sensor) and as calculated using the model described above (Ψ̇ model), with the measured steering angle δ in two different driving conditions, respectively on a sloping road ( figure 2 ) and on a slippery road. The offset illustrated in the figure 2 The difference between the measured and calculated yaw rate values ​​is due to the fact that the model used does not take into account the effect of the slope or the angles of inclination. On the figure 3 Where the data was acquired on a flat but slippery road, the lag is caused by the saturation of lateral forces. Such a lag is particularly evident around the 45th second on the figure 3 These two examples, which show a discrepancy between the measured and calculated values ​​of the angular yaw rate, clearly illustrate the need to compensate for the vehicle model to take into account all potential driving situations in autonomous vehicles, in order to increase their robustness and operational safety.

[0036] To compensate for this discrepancy, the vehicle lateral dynamics compensation module 11 is designed to correct the vehicle model in these particular driving situations where a discrepancy appears with the actual behavior of the vehicle and thus provide the lateral controller with a correct value for the yaw angular velocity.

[0037] There figure 4 is a synoptic diagram illustrating the operation of the compensation module 11, allowing to obtain an accurate estimate of the angular yaw rate at the output of the vehicle model M in all driving conditions.

[0038] The vehicle lateral dynamics compensation module 11 includes compensator based on the nominal vehicle model M.

[0039] More specifically, the compensation module 11 uses as input the measured yaw rate Ψ̇ sensor, the vehicle's speed from the vehicle's onboard speed sensor 4, and road profile information (slope, incline) provided by road profile storage means 18 corresponding to the planned trajectory. The yaw rate Ψ̇ model is initially calculated using model M based on the vehicle's longitudinal speed and steering angle data measured, for example, by a wheel steering angle sensor, preferentially taking into account road slope and incline information.

[0040] To do this, the parameters of the vehicle model M include the lateral stiffness of the front wheels Cf, the lateral stiffness of the rear wheels Cr, the longitudinal distance from the center of gravity of the vehicle to the front wheels a, the longitudinal distance from the center of gravity of the vehicle to the rear wheels b, the mass of the vehicle and the yaw moment of inertia of the vehicle Iz.

[0041] Then, the model 110 compensator takes into account the measured yaw rate Ψ̇ sensor and the yaw rate calculated using the model Ψ̇ model to provide, based on a comparison between these respective yaw rates, a compensated steering angle δcomp that minimizes the difference between them. Module 110 implements an adaptive parametric algorithm to obtain the unmodeled vehicle dynamics by calculating the compensated steering angle δcomp as follows: δ comp = C Q ψ ˙ sensor − ψ ˙ model where C(Q) is an adaptive controller based on adaptation parameters Q calculated online using a parametric adaptive algorithm (PAA).

[0042] The compensated steering angle is then provided to the vehicle model M so that the yaw rate calculated using the model Ψ̇ model converges to the yaw rate measured by the sensor Ψ̇. The model output is therefore corrected in real time by the compensation module 11, and the lateral controller 8 thus receives a correct yaw rate input.

[0043] Once the yaw rate has been compensated, the linearized lateral forces exerted on the front and rear wheels of the vehicle are estimated. These linearized lateral forces, respectively front Ff_I and rear Fr_l, are estimated as follows: F f_l = 2 c f δ − β + a ψ ˙ v x F r_l = 2c r b ψ ˙ v x − β

[0044] The M model of the vehicle, closed by the model 110 compensator, therefore allows for a linearized system.

[0045] These linearized lateral forces, thus estimated, are provided by the vehicle's lateral dynamics compensation module 11 and are then used as input to the safety module 13, where they are compared with the lateral forces calculated using the vehicle's onboard sensors. As explained previously, the purpose of this comparison is to detect if the vehicle's physical driving limits are exceeded, in the event of a discrepancy between the vehicle model and its actual behavior.

[0046] There figure 5 is a synoptic diagram illustrating the operation of security module 13, allowing this comparison to be made.

[0047] The safety module 13 includes a calculation module 130 capable of calculating the lateral forces exerted on the front and rear wheels from values ​​of lateral acceleration of the vehicle, steering angle, and yaw rate, as measured by the on-board sensors, respectively the lateral acceleration sensor 14, the steering angle sensor 15, and the yaw rate sensor 16. These lateral forces, respectively front Ff and rear Fr, are estimated as follows: F f = I z ψ ¨ + b m a y L cos δ F r = a m a y − I z ψ ¨ L where ay corresponds to the measured transverse acceleration.

[0048] The lateral forces calculated by the calculation module 130 on the basis of the sensor data and the linearized lateral forces estimated by the compensation module 11, are provided as input to a comparator module 131 of the safety module 13, capable of performing a comparison between these respective calculated and estimated lateral forces.

[0049] More specifically, the comparator module 131 calculates the difference between the linearized lateral forces and the calculated lateral forces, and compares this difference to a predefined threshold ε. When this difference exceeds the predefined threshold, the comparator module infers a degradation in the vehicle's automated driving mode, to the point where control stability can no longer be guaranteed. Exceeding the predefined threshold ε also triggers a safety mode. This safety mode includes, firstly, the generation of a warning signal 132 at the vehicle's human-machine interface 17, intended to inform the driver of the vehicle's uncontrollable state.Exceeding the threshold simultaneously triggers the generation and sending of safety control signals 133 to the controllers acting on the vehicle's actuators, including the steering wheel and brake actuators, in order to bring the vehicle into a safe state, consisting of stopping the vehicle, if the driver does not react.

Claims

1. Control device (1) of a motor vehicle with autonomous or assisted driving, comprising a route planning module (5) capable of storing information relating to a planned trajectory, a controller (3) of the longitudinal movement of the vehicle capable of generating, as output, command signals (6, 7) to acceleration and braking actuators and a controller (8) of the lateral movement of the vehicle capable of generating, as output, command signals (12) to steering actuators, such that the vehicle follows the planned trajectory in an assisted or automated driving mode of the vehicle, the output of said lateral controller (8) being based on a minimization of the error between a desired angular yaw rate, obtained from the curvature of the road on the planned trajectory, and the current angular yaw rate of the vehicle, estimated from a model (M) of the dynamic behaviour of the vehicle capable of supplying, as input to said lateral controller (8), an estimation of the angular yaw rate (Ψmodel) based on the measured longitudinal speed of the vehicle and on the measured steering angle, comprising a module (11) for compensating the lateral dynamics of the vehicle, capable of dynamically correcting parameters of the model (M) in the event of deviation between the model and the real dynamic behaviour of the vehicle, as a function of a comparison between the estimated angular yaw rate value (Ψ̇model) and an angular yaw rate value (Ψ̇sensor) measured by an on-board sensor (16), such that said estimation supplied by the model to the lateral controller is corrected instantaneously, said device being characterized in that it comprises a safety module (13), linked to the output of the module (11) for compensating the lateral dynamics of the vehicle, said safety module (13) being adapted to detect impaired driving states in automated mode on the basis of a difference between the output of the corrected model and the real dynamic behaviour of the vehicle.

2. Device according to Claim 1, characterized in that the compensation module (11) is adapted to estimate a compensated steering angle (δcomp) by means of an adaptive algorithm taking into account the difference between the estimated value of the angular yaw rate (Ψ̇model) and the measured angular yaw rate value (Ψ̇sensor), the output of the model (M) being corrected on the basis of the compensated steering angle supplied as input of the model.

3. Device according to Claim 1 or 2, characterized in that the corrected model is capable of supplying an estimation of the linearized lateral forces (Ff_L, Fr_L) exerted on the front and rear wheels of the vehicle, said safety module (13) being capable of receiving said linearized lateral forces and of comparing said linearized lateral forces with the calculated lateral forces (Ff, Fr) using data supplied by sensors (14, 15, 16) on board the vehicle.

4. Device according to Claim 3, characterized in that the safety module (13) comprises a computation module (130) capable of calculating the lateral forces (Ff, Fr) exerted on the front and rear wheels based on values of lateral acceleration of the vehicle, of steering angle, and of angular yaw rate, measured by the on-board sensors.

5. Device according to Claim 3 or 4, characterized in that said safety module (13) comprises a comparator module (131) capable of establishing the difference between the linearized lateral forces (Ff_L, Fr_L) and the calculated lateral forces (Ff, Fr) and of activating a safety mode when said difference exceeds a predefined threshold (ε).

6. Device according to Claim 5, characterized in that said safety module (13) is capable of generating vehicle securing command signals to the longitudinal and lateral controllers, said securing command signals prevailing over the driving command signals in automated mode when said safety mode is activated.

7. Device according to Claim 6, characterized in that said securing command signals are adapted to command the stopping of the vehicle.

8. Device according to any one of Claims 5 to 7, characterized in that said safety module (13) is capable of generating an alert signal to a human-machine interface (17) of the vehicle when said safety mode is activated.

9. Motor vehicle, characterized in that it comprises a control device (1) according to any one of the preceding claims.