Prediction module, associated device and method for real-time control path

By introducing prediction modules and variable gain equipment into the real-time path control device, and adjusting the steering commands according to changes in the physical parameters of the vehicle, the problem of poor stability in the curve center in the prior art is solved, and a more robust path control effect is achieved.

CN113677579BActive Publication Date: 2024-11-01AMPERE +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202080027466.4
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Priority Date
2019-04-01
Filing Date
2020-03-27
Publication Date
2024-11-01
Estimated Expiration
2040-03-27

AI Technical Summary

Technical Problem

The existing real-time path control equipment is difficult to maintain the stability of the vehicle in the center of the curve when the vehicle physical parameters change, such as after the loading or tire state changes, resulting in an increase in excessive steering and lateral deviation.

Method used

A prediction module is designed to calculate submodules and variable gain equipment to adjust the steering command according to changes in the physical parameters of the vehicle, reduce the lateral deviation between the center of gravity of the vehicle and the center of the lane, and calculate the pseudo-measurement vector to compensate by using the bicycle model. Curvature of the curve and generate steering commands through the observer to keep the vehicle on a straight path.

Benefits of technology

It effectively reduces the lateral deviation of the vehicle in the curve, improves the stability of turning, avoids the problem of excessive steering, and ensures the stable driving of the vehicle in the center of the lane.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN113677579B_ABST
    Figure CN113677579B_ABST
Patent Text Reader

Abstract

A prediction module (6) for a real-time path control device for a motor vehicle comprises: a calculation submodule (65) for calculating a steering command for compensating for the curvature of a curve in a lane of the vehicle, and a variable gain device (70) connected to the output of the calculation submodule. The gain of the variable gain device is connected to a controller (73) which can adjust the gain value according to the result of comparing the components of the current measurement vector of the state variables of the device with each other and with a detection threshold so as to reduce the lateral deviation between the center of gravity of the vehicle and the center of the lane of the vehicle, and the output of the variable gain device is a steering command for compensating for the curvature of the curve.
Need to check novelty before this filing date? Find Prior Art

Description

[0001] The present invention relates to a path control device for a vehicle.

[0002] More specifically, the present invention relates to a real-time path control device of a vehicle designed to compensate for the curvature of a curve in a lane of the vehicle, a prediction module built into the device, and an implementation method for the device.

[0003] Motor vehicles may be equipped with path control devices designed to help the driver keep the vehicle in a lane, and in particular to keep the vehicle in the center of a lane. Such devices may also be found in autonomous vehicles, where driving is completely delegated to the vehicle. Such devices act on the steering of the vehicle by adjusting the steering angle of the vehicle's steering wheels. The steering angle suitable for the vehicle's situation is determined using a control algorithm contained in the device. The algorithm implements a control law and uses a dynamic model of the vehicle. Such a control law for helping to keep the vehicle in the center of a lane is generally referred to as "lane centering assist" (LCA).

[0004] For comfort, the controls should be smooth and not cause any jolts that would be uncomfortable or unexpected for the driver.

[0005] Application FR 3051756 filed by the same applicant indicates Figures 1 to 5 Presented prior art. This application describes a real-time path control device DISP for a vehicle, which is intended in particular to ensure the comfort of the passengers.

[0006] Figure 1 The real-time path control device DISP is shown applied to a motor-driven vehicle 1 having two steered wheels, preferably the front wheels of the vehicle, controlled by actuators controlled by control signals u.

[0007] In the following, the operator represents the derivative of x.

[0008] The steerable front wheels of vehicle 1 are controlled by an actuator controlled by a control signal U.

[0009] The vehicle 1 also has sensors designed to measure vehicle parameters such as the actual longitudinal velocity V, the steering angle δ of the front wheels, the yaw rate of the vehicle Or a heading angle (e.g., the static heading angle corresponds to the front wheel angle for the closest gear ratio). The vehicle may also be equipped with a RaCam device combining the characteristics of an optical camera and a radar for providing a model of the guidance line of the vehicle's lane in the form of a polynomial y(x), or an optical camera for determining the polynomial y(x).

[0010] The vehicle 1 is equipped with an on-board processor comprising a controller device 2 for generating a control signal Ust to bring the physical state vector ξ of the vehicle into conformity with a set-point state vector ξ*, thereby ensuring that the vehicle 1 follows a desired path.

[0011] The onboard processor also includes an observer 3 for generating in real time an estimated path following state vector ξ of the vehicle 1 moving at a speed V using the command Ust and a current measurement vector η of state variables associated with the path following physical state vector ξ of the vehicle 1.

[0012] The observer device 3 is designed to generate an estimated state vector ξ in real time as reliably as possible.

[0013] The on-board processor also comprises a prediction module 4 which adds a second steering command Uff to the first steering command Ust generated by the device 2 according to the curvature γff of the curve to compensate for the curve.

[0014] The second steering command Uff enables the vehicle 1 to negotiate a curve having a curvature γff. The second command is added to the control signal Ust to cause the controller device 2 to adjust the path of the vehicle 1 so that the vehicle 1 follows a straight route.

[0015] Therefore, the control signal U is also the sum of the first command Ust and the second command Ust.

[0016] The real-time path control device of the vehicle 1 includes an actuator, a speed sensor, a sensor for measuring the steering angle δ, a sensor for measuring the yaw angular velocity Sensor and RaCam device, device 2, observer 3 and module 4.

[0017] The physical state vector ξ of the vehicle is:

[0018]

[0019] where Ψ is the relative heading angle between the vehicle axis and the tangent to the reference path, is the lateral velocity of the vehicle deviating from the ideal path of the vehicle, yl is the lateral deviation between the center of gravity of the vehicle and the center of the vehicle's lane and the tangent of the vehicle's forward path, is the change in steering angle, and ∫-yl represents the integral of the lateral position.

[0020] The set point state vector ξ* is:

[0021]

[0022] And estimate the state vector yes:

[0023]

[0024] The vector ξ is unknown because the internal state of vehicle 1 is not fully accessible.

[0025] The device 3 implements a vehicle model known as a bicycle model, which is provided in the form of a matrix as follows:

[0026]

[0027] Among them, δ req is the sum of the control signal Ust and the second steering command Uff, and γff is the curvature radius of the curve.

[0028] Angle δ ref The value of is equal to the command Uff.

[0029] Steering angle δ and reference steering angle δ ref are linked by a second-order transfer function that models the power steering.

[0030] The matrix A includes constant coefficients and variable coefficients throughout a given driving sequence, depending on the following parameters of vehicle 1:

[0031] - Sideways stiffness of the front wheel C f ,

[0032] - Lateral stiffness of rear wheel C r ,

[0033] - Distance from center of gravity to front axle l f ,

[0034] - Distance from center of gravity to rear axle l r ,

[0035] - the total mass of the vehicle m,

[0036] - the speed of the vehicle V,

[0037] - the inertia J of the vehicle about the vertical axis passing through its center of gravity,

[0038] - the damping coefficient ξ of the second-order function, and

[0039] - The specific ripple ω of the second order transfer function.

[0040] Matrix B δ depends on the specific pulsation ω and the speed V, while the matrix B ρ It depends on the speed V.

[0041] The values ​​of the parameters of the vehicle 1 are determined when the vehicle 1 is unloaded and the tire pressures are at optimal values, and these parameters are fixed throughout the life of the vehicle 1 .

[0042] The device 3 receives, via a first input, a current measurement vector η, which is associated with the physical state vector ξ by the following instrumental relation C:

[0043]

[0044] The current measurement vector η has five components:

[0045]

[0046] Yaw rate It is measured by the yaw rate sensor.

[0047] The RaCam device determines the polynomial y(x).

[0048] The observer 3 receives via a second input a command Ust which is a steering angle δ.

[0049] A third input terminal of the observer 3 is connected to a first output terminal S1 of the prediction module 4 .

[0050] The prediction module 4 also has a second output terminal S2 connected to an adder so that the control signal U is the sum of the control signal Ust and the control signal Uff, a first input terminal E1 connected to the vehicle 1 to receive the current measurement vector η, and a second input terminal E2 connected to the vehicle 1 to receive the polynomial y(x).

[0051] The polynomial y(x) is used to determine the relative heading angle Ψ, the lateral deviation yl, and to calculate the integral of the lateral position ∫-yl.

[0052] The prediction module 4 includes a first submodule 41 connected to the second input terminal E2, and a second submodule 42 connected to the first submodule 41, the input terminal E1, and the output terminals S1 and S2.

[0053] The first submodule 41 receives a polynomial y(x) which specifies the geometry of the lane guide line at each point at a distance x in front of the vehicle 1 and determines a curvature γff which is sent to the second submodule 42 .

[0054] The second submodule 42 also receives the current measurement vector η.

[0055] During a curve with curvature Vff, module 4 corrects the relative heading angle Ψ, yaw rate and wheel angle δ.

[0056] The closed loop including the vehicle 1 and the devices 2 and 3 minimizes the state vector ξ to approximately the zero vector corresponding to the straight line, with the setpoint vector ξ* being equal to the zero vector.

[0057] The dynamics of this closed loop change slowly, thus ensuring good comfort.

[0058] The prediction module 4 operates in an open loop to ensure good cornering performance.

[0059] The second submodule 42 calculates the wheel angle δ corresponding to the curvature γff eq and the associated pseudo-calculated measurement vector η eq .

[0060] Wheel angle δ eq Given by the following equation:

[0061]

[0062] And the pseudo-calculated measurement vector η eq yes:

[0063]

[0064] in:

[0065]

[0066] Angle δ eq Sent by the second output terminal S2 of module 4, and measuring vector η eq Sent by the first input terminal S1, so that the pseudo-calculated measurement vector η is subtracted from the measurement vector η eq , so that the observer module 3 only acts on the path deviation on the virtual straight lane.

[0067] However, the prediction module 4 is based on a bicycle model and does not change during the life of the vehicle 1 .

[0068] When the physical parameters of the vehicle 1 change, such as a certain mass being loaded into the trunk, or a tire being deflated or damaged, the parameters of the prediction module 4 are no longer correct, and performance degradation ensues, resulting in a large lateral deviation from the center of the lane when turning.

[0069] Figure 2 and Figure 3 It shows the evolution over time of the steering angle δ and the lateral deviation yl under the action of the real-time path control device when the parameters of the vehicle 1 are at nominal values ​​and the trunk is empty.

[0070] An 18cm deviation was found when entering and exiting the curve.

[0071] Figure 4 and Figure 5 The evolution over time of the steering angle δ and the lateral deviation yl under the action of the real-time path control device is shown when an additional mass of 300 kg is included in the luggage compartment of the vehicle 1 .

[0072] A deviation of 75cm was found when entering and exiting the curve.

[0073] and Figure 2 and Figure 3 Compared to the nominal configuration shown, the vehicle 1 becomes more susceptible to oversteer.

[0074] Therefore, the present invention aims to make the real-time path control device more robust when turning by adjusting the commands from the prediction module according to changes in the physical parameters of the vehicle.

[0075] In view of the above situation, the present invention proposes a prediction module of a device for real-time control of the path of a motor vehicle, the module comprising: a sub-module for calculating a steering command for compensating for the curvature of a curve in the lane of the vehicle, and a variable gain device connected to the output end of the calculation sub-module.

[0076] The variable gain device is connected to a controller which is capable of adjusting the gain value based on the results of comparing components of a current measurement vector of a state variable of the device with each other and with a detection threshold so as to reduce the lateral deviation between the center of gravity of the vehicle and the center of the lane of the vehicle, the output of the variable gain being a steering command for compensating for the curvature of the curve.

[0077] According to one feature, the module also comprises a second calculation submodule connected to the output of the variable gain device and designed to calculate a pseudo-calculated measurement vector using the model of the vehicle.

[0078] Preferably, the model of the vehicle is a bicycle model.

[0079] The present invention also relates to a real-time path control device for a vehicle, which is designed to compensate for the curvature of a bend in the vehicle's lane and comprises a prediction module as previously defined, and an observer which generates in real time an estimated straight lane following state vector for the vehicle, such as to produce a steering command for stabilizing the vehicle's path relative to the straight lane, the observer being connected to the prediction module.

[0080] The invention also relates to a real-time path control method for a motor vehicle, the real-time path control method being designed to compensate for the curvature of a bend in the vehicle's lane.

[0081] When it is detected that the vehicle is oversteering compared to a nominal configuration, a gain of a variable gain device of the prediction module is adjusted to reduce a lateral deviation between a center of gravity of the vehicle and a center of a lane of the vehicle.

[0082] Preferably, the vehicle is considered to be oversteering compared to a nominal configuration when the lateral deviation and the direction of the bend are oriented in the same direction in a reference frame of the vehicle and the lateral deviation is greater than a detection threshold.

[0083] Advantageously, the gain of the variable gain device is reset when the lateral deviation and the direction of the curve are oriented in different directions in a reference system of the vehicle and the lateral deviation is smaller than a detection threshold.

[0084] Other objects, features and advantages of the invention are set forth in the following description given purely by way of non-limiting example and with reference to the accompanying drawings, in which:

[0085] [ Figure 1 ], as previously mentioned, is a schematic diagram of a real-time path control device for a vehicle according to the prior art,

[0086] [ Figure 2 ]

[0087] [ Figure 3 ], as previously mentioned, shows the evolution over time of the steering angle and the lateral deviation under the action of a control device according to the prior art when the parameters of the vehicle are at nominal values ​​and the trunk is empty,

[0088] [ Figure 4 ]

[0089] [ Figure 5 ] shows the evolution over time of the steering angle and the lateral deviation under the action of a control device according to the prior art when an additional mass is included in the trunk of the vehicle,

[0090] [ Figure 6 ] is a schematic diagram of an embodiment of a real-time path control device for a vehicle according to the present invention,

[0091] [ Figure 7 ] shows an embodiment of a prediction module according to the present invention,

[0092] [ Figure 8 ] is a schematic diagram of a vehicle driving in a lane with a curve.

[0093] [ Fig. 9 ] shows an implementation method of a real-time path control device for a vehicle according to the present invention, and

[0094] [ Fig.10 ]

[0095] [ Fig.11 ] shows the evolution over time of the steering angle and the lateral deviation under the action of the control device according to the invention when an additional mass is included in the luggage compartment of the vehicle.

[0096] Figure 6 1 is a schematic diagram of a real-time path control device 5 of a vehicle 1 according to one aspect of the present invention, the real-time path control device being used to compensate for the curvature of a curve in a lane of the vehicle 1. Figure 1 Identical elements of the device DISP are indicated by the same reference numerals.

[0097] The vehicle 1, the controller device 2 for generating the control signal Ust, and the measurement vector η calculated from the current measurement vector η and the pseudo eq To generate the estimated straight lane following state vector of the vehicle in real time Observer 3.

[0098] The device 5 also includes a prediction module 6, which has a first input terminal 61 connected to the vehicle 1 and receiving the current measurement vector η, a second input terminal 62 connected to the vehicle 1 and receiving the polynomial y(x), a first output terminal 63 connected to an adder so that the control signal U is the sum of the control signal Ust and the control signal Ueff generated by the module 6, and a second output terminal 64 connected to the third input terminal of the observer 3.

[0099] The prediction module 6 is an open loop.

[0100] Figure 7 An embodiment of a prediction module 6 is shown.

[0101] The prediction module 6 comprises a first submodule 41 and a calculation submodule 65 for calculating a steering command, the first submodule being used to determine the curvature γff based on a polynomial y(x) determined by the RaCam device and being connected to an input 62 of the module 6, the polynomial giving the geometry of the lane guide line at each point at a distance x in front of the vehicle 1, the calculation submodule comprising a first input 66 connected to the output of the submodule 41, a second input 67 connected to the input 61 of the module 6 and an output 68 connected to the input 69 of the variable gain device 70.

[0102] The variable gain device 70 comprises a control input 71 connected to an output 72 of a controller 73 .

[0103] The controller 73 also has an input 74 connected to the input 61 of the module 6 .

[0104] The variable gain device 70 has an output 75 which is connected firstly to the first output 63 of the module 6 and secondly to a first input 76 of a second calculation submodule 77 .

[0105] The second calculation submodule 77 also has a second input 78 connected to the first input 61 of the module 6 and an output 79 connected to the second output 64 of the module 6 .

[0106] The calculation submodule 65 for calculating the steering command calculates the steering command for compensating the curvature γff.

[0107] The steering command is equal to the wheel angle δ given by equation (7) eq , and is implemented by a calculation submodule 65 for calculating a steering command.

[0108] The controller 73 controls the variable gain device 70 based on the results of comparing the components of the current measurement vector η of the state variables of the device with each other and with the detection threshold S so as to reduce the lateral deviation yl between the center of gravity of the vehicle 1 and the center of the lane of the vehicle.

[0109] The second submodule 77 determines the pseudo-calculated measurement vector η according to equation (8): eq .

[0110] In a variant, the first curvature determination submodule 41 may be arranged outside the prediction module 6 .

[0111] Figure 8 The vehicle 1 is shown traveling on a lane 80 having a curve with a curvature γff.

[0112] The vehicle 1 comprises a reference system R, the origin of which coincides, for example, with the nominal center of gravity of the vehicle 1 .

[0113] The path of vehicle 1 follows the center guide line 81 of lane 80 .

[0114] At time t1, the vehicle 1 is in a straight portion of the lane 80. The front wheels 82 and rear wheels 83 of the vehicle 1 are aligned, and the lateral deviation yl and the wheel angle δ are substantially zero.

[0115] At time t2 , vehicle 1 is on a curve.

[0116] The device 5 reads the non-zero lateral deviation yl2 between the nominal center of gravity of the vehicle 1 and the central guide line 81 of the lane 80, and generates a compensation for the set point state vector ξ* and the estimated state vector The steering command Ust of the deviation between them is used to eliminate the deviation or move it towards zero.

[0117] The sign of the wheel angle δ2 makes it possible to determine the direction of the curve orientation.

[0118] Fig. 9 An implementation method for a device 5 is shown.

[0119] During step 80 , the controller 73 determines whether the vehicle 1 is oversteering while in a curve.

[0120] If the lateral deviation yl and the direction of the curve are oriented in the same direction in the reference system R of the vehicle 1 and the value of the lateral deviation yl is greater than the detection threshold S, the vehicle 1 is considered to be oversteering compared to nominal.

[0121] exist Figure 8 In the reference system R of the vehicle 1 shown, the orientation angle δ2 and the lateral deviation yl2 are positive.

[0122] Therefore, the lateral deviation yl2 and the direction of the curve are oriented in the same direction.

[0123] Therefore, it is assumed that the lateral deviation yl2 is greater than the detection threshold S.

[0124] The method proceeds to step 81 .

[0125] If either of these two conditions is not met, the method remains at step 80 .

[0126] During a step 81 , the controller 73 sets the gain of the variable gain device 70 to a predetermined value.

[0127] During this step, the gain of the variable gain device 70 is varied from 1 to the predetermined value, for example, 0.75.

[0128] The predetermined value is determined, for example, by empirical testing of the behavior of the vehicle 1 or by numerical simulation of different predetermined values.

[0129] The gain adjustment reduces the value of the lateral deviation yl between the center of gravity of the vehicle 1 and the center of the lane of the vehicle, such as Fig.10 and 11 As shown, these two figures show the evolution over time of the steering angle δ and the lateral deviation yl under the effect of the device 5 when the vehicle's trunk contains an additional mass of 300 kg.

[0130] The lateral deviation yl is reduced to 20 cm.

[0131] Furthermore, during a step 81 , the controller 73 determines whether the vehicle 1 is still oversteering compared to nominal on the bend.

[0132] In step 82 , if the lateral deviation yl and the direction of the curve are oriented in different directions in the reference system R of the vehicle 1 and the lateral deviation yl is less than the detection threshold S, the controller 73 resets the gain of the variable gain device 70 .

Claims

1. A prediction module (6) for a real-time path control device of a motor vehicle, the module comprising: A computing submodule (65) for computing a steering command for compensating for the curvature of a curve in the lane of the vehicle, and a variable gain device (70) connected to the output end of the computing submodule, characterized in that the variable gain device is connected to a controller (73) capable of adjusting the gain value based on the result of comparing the components of a current measurement vector (η) of the state variables of the device with each other and with a detection threshold (S) so as to reduce the lateral deviation (yl) between the center of gravity of the vehicle and the center (81) of the lane of the vehicle, the output of the variable gain being a steering command for compensating for the curvature (γff) of the curve.

2. The module according to claim 1 further comprises a second calculation submodule (77) connected to the output of the variable gain device (70) and designed to calculate a pseudo-calculated measurement vector (η) using the model of the vehicle. eq ).

3. The module according to claim 2, wherein: The model of the vehicle is a bicycle model.

4. A real-time path control device (5) for a vehicle, the real-time path control device being designed to compensate for the curvature (γff) of a curve in a lane (80) of the vehicle (1), and comprising a prediction module (6) as claimed in one of the preceding claims, and an observer (3) which generates in real time an estimated straight lane following state vector ( ), such as to generate a steering command (Ust) for stabilizing the path of the vehicle relative to the straight lane, the observer is connected to the prediction module.

5. A real-time path control method for a motor vehicle, the real-time path control method being designed to compensate for the curvature (γff) of a curve in a lane (80) of the vehicle (1), the method being implemented by a real-time path control device according to claim 4, characterized in that When it is detected that the vehicle is oversteering compared to a nominal configuration, the gain of a variable gain device (70) of a prediction module (6) is adjusted so as to reduce a lateral deviation (yl) between the center of gravity of the vehicle and the center (81) of the vehicle's lane.

6. The method of claim 5, wherein: When the lateral deviation (yl) and the direction of the curve are oriented in the same direction in the vehicle's reference frame (R), and the lateral deviation is greater than a detection threshold (S), the vehicle is considered to be oversteering compared to a nominal configuration.

7. The method of claim 6, wherein: When the lateral deviation (yl) and the direction of the curve are oriented in different directions in the reference system (R) of the vehicle and the lateral deviation is less than a detection threshold (S), the gain of the variable gain device (70) is reset.