JP5483194B2 - Sliding mode control device and vehicle automatic steering control device - Google Patents

Description

  The present invention relates to a sliding mode control device and a vehicle automatic steering control device.

  In recent years, various technologies related to automatic steering control have been proposed and put into practical use, and the main design is PID control or LQI control. In addition, as another feedback control, sliding mode control, which has less influence on disturbance and higher robustness than PID control and LQI control, has attracted attention.

JP 2010-23682 A Japanese Patent No. 3616734

  In sliding mode control that sets the switching hyperplane (switching function) and converges the state quantity of the control target and the time change rate of the state quantity to the switching hyperplane, chattering occurs when the state quantity reaches the switching hyperplane from the initial state. The phenomenon occurs. For example, when the automatic steering control of the vehicle is executed by the sliding mode control, the steering angle of the steering wheel frequently changes due to the chattering phenomenon, and the driver feels uncomfortable.

  As a method for suppressing the chattering phenomenon of sliding mode control, it is known to set a boundary layer on the switching hyperplane.

  However, since the value of the conventional general boundary layer is set as a constant and the determination is made by trial and error, it is difficult to set an optimal boundary layer.

  Accordingly, an object of the present invention is to provide a sliding mode control device capable of accurately reducing chattering while maintaining good convergence to the switching hyperplane.

  In order to solve the above problems, the sliding mode control device of the present invention includes a boundary layer setting unit and a control unit. The boundary layer setting means sets the boundary layer on the switching hyperplane. The control means converges the state quantity of the control target and the time change rate of the state quantity on the switching hyperplane, and also limits the nonlinear input for causing the control target state quantity and the time change rate of the state quantity to reach the switching hyperplane. Control means for reducing in the layer. The boundary layer setting means sets the boundary layer such that the value increases as the difference between the inclination of the convergence locus of the state quantity of the control target in the initial state and the time change rate of the state quantity and the inclination of the switching hyperplane increases.

  In the sliding mode control, when the control target state quantity and the time change rate of the state quantity are converged on the switching hyperplane, the slope of the convergence trajectory of the control target state quantity and the state quantity time change rate in the initial state and the switching hyperplane The chattering phenomenon increases as the difference from the inclination increases.

  In this regard, in the above configuration, the value of the boundary layer is set to be larger as the difference between the inclination of the convergence locus of the state quantity of the control target in the initial state and the time change rate of the state quantity and the inclination of the switching hyperplane is larger. It is possible to accurately reduce the occurrence of chattering while maintaining good convergence to the switching hyperplane.

  An automatic steering control device for a vehicle according to the present invention includes the above-described sliding mode control device, and includes a vehicle position detection unit and a course setting unit. The vehicle position detection means detects the current position of the vehicle. The course setting means sets a target course of the vehicle. The state quantity to be controlled is the distance between the position of the vehicle and the target route. The control means calculates the target steering angle for switching the distance between the vehicle position and the target path and the time change rate of the distance to converge on the hyperplane, and calculates the target steering current corresponding to the calculated target steering angle as a steering actuator. Output to.

  In the above configuration, it is possible to reduce the uncomfortable feeling given to the driver by suppressing the change in the steering angle.

  The sliding mode control device may further include a hyperplane setting unit that sets a switching hyperplane so that a nonlinear relationship is established between the state quantity of the control target and the time change rate of the state quantity. The hyperplane may be defined by a function including a third term term of the state quantity and a third term term of the time change rate of the state quantity.

  In the above configuration, in the switching hyperplane set by the hyperplane setting means, a nonlinear relationship is established between the state quantity of the controlled object and the time rate of change of the state quantity, and the time of the controlled quantity of the state quantity and the state quantity Since the rate of change is not constrained by a fixed linear relationship, it is possible to design a switching hyperplane that can achieve high robustness. Therefore, the above sliding mode control device is applied to an automatic steering control device for a vehicle, and the distance between the vehicle position and the target lane is set as a state quantity, and the time change rate is set as the time change rate of the state quantity to set a hyperplane. In this case, it is possible to design a switching hyperplane that exhibits optimum tracking performance as the road curvature increases without the distance and time change rate being constrained by a fixed linear relationship.

  According to the sliding mode control device of the present invention, the chattering phenomenon can be accurately reduced.

It is a mimetic diagram of vehicles provided with an automatic steering control device of this embodiment. It is a block block diagram of the automatic steering control apparatus of FIG. It is a schematic diagram which shows a vehicle and a target travel line. It is a flowchart of an automatic steering control process. It is a schematic diagram which shows a switching hyperplane. It is a figure which shows an example of the relationship between the curvature of a target travel line, and the coefficient G. It is a figure which shows the relationship between the inclination of the convergence locus | trajectory of the state quantity of the control target of an initial state and the time change rate of a state quantity, the inclination of a switching hyperplane, and a chattering phenomenon.

  Hereinafter, an embodiment of the present invention in which a sliding mode control device of the present invention is applied to an automatic steering control device of a vehicle will be described with reference to FIGS.

  As shown in FIGS. 1 to 3, the vehicle 1 includes a steering wheel 2, a sensor unit 3, a controller (ECU: Electric Control Unit) 4, a drive current output unit 5, and a steering actuator 6. The sensor unit 3, the controller 4, and the drive current output unit 5 constitute a travel control device. The steering wheel 2 is provided in the cab of the vehicle 1 and receives a steering input from the driver.

  The sensor unit 3 includes a camera 31, a yaw rate sensor 33, a handle angle sensor 34, a vehicle speed sensor 35, and an automatic steering switch 36.

  The camera 31 is a CCD camera, images the road surface ahead of the vehicle 1 in the traveling direction, and outputs the image signal to the controller 4. In the present embodiment, a predetermined target travel line (target course) LT is set between the white lines LL and LR on both the left and right sides of the travel path of the vehicle 1, and travel by automatic steering is performed along the set target travel line LT. Therefore, the camera 31 images the left and right white lines LL and LR ahead of the traveling direction.

  The yaw rate sensor 33 detects the yaw rate φ ′ (rotational angular velocity: rad / s) generated in the vehicle 1 during cornering, with the left rotation as the positive direction, and outputs the detected yaw rate φ ′ to the controller 4. The steering wheel angle sensor 34 detects the steering angle δ (rad) input to the steering wheel 2 and outputs the detected steering angle δ to the controller 4. The vehicle speed sensor 35 detects the vehicle speed V (m / s) of the vehicle 1 and outputs the detected vehicle speed V to the controller 4.

  The automatic steering switch 36 is provided in the driver's cab of the vehicle 1 and outputs a switching signal corresponding to an on / off operation from the driver to the controller 4. The automatic steering switch 36 is set to off at the initial stage when the power is turned on, and is set to on in response to an operation from the driver. When the automatic steering switch 36 is set to ON, the controller 4 executes an automatic steering control process. Hereinafter, a state where the automatic steering switch 36 is set to ON will be described.

  The controller 4 includes a CPU (Central Processing Unit), a ROM (Read Only Memory), and a RAM (Random Access Memory). Various signals are sequentially input from the sensor unit 3 to the controller 4. In the ROM or RAM, an automatic steering control program is stored in advance, and the CPU reads out and executes the automatic steering control program, whereby the target travel line setting unit 41, the target deviation calculation unit 42, the target steering angle calculation unit 43, and It functions as a command current value calculation unit 44.

  The target travel line setting unit 41 sets a target travel line LT on the road surface. The target travel line LT is a course that is a target when the vehicle 1 travels by automatic steering. The target travel line setting unit 41 of the present embodiment analyzes an image signal captured by the camera 31 to detect a range of the width W between the left and right white lines LL and LR on the road surface as the travel lane of the vehicle 1, and the travel lane The position (line) at which the distance from the white line on one side (for example, the white line LL on the left side) is 50% (W / 2) of the width W of the travel lane is set as the target travel line LT.

  The target deviation calculation unit 42 calculates the current relative position (target deviation y) of the vehicle 1 with respect to the target travel line LT set by the target travel line setting unit 41. Specifically, by analyzing the image signal captured by the camera 31 to detect the distance D from the left white line LL to the vehicle 1 and subtracting the set distance W / 2 of the target travel line LT from the distance D, A target deviation y (y = D−W / 2) is calculated.

  In the present embodiment, the imaging signal captured by the camera 31 is used for setting the target travel line LT and calculating the target deviation y. However, even if other information other than the imaging signal of the camera 31 is used. Good. For example, information detected by the GPS may be received, the target travel line LT may be set based on the received information, and the target deviation may be calculated. In addition, an information generator along the target travel line LT may be provided in advance on the road surface, and the position information of the target travel line LT may be acquired from the information generator to detect the target travel line LT and the target deviation y. .

  The target steering angle calculation unit 43 constitutes a hyperplane setting unit, a boundary layer setting unit, and a control unit. The target steering angle calculation unit 43 is for causing the vehicle 1 to travel according to the target travel line LT based on the target travel line LT detected by the target travel line setting unit 41 and the target deviation y detected by the target deviation calculation unit 42. The steering angle (target steering angle θtotal for switching the target deviation y and the time differential y ′ of the target deviation to converge on the hyperplane σn) is calculated. In addition, the target steering angle calculation unit 43 is configured so that the value increases as the difference between the inclination of the convergence locus of the target deviation y in the initial state and the time derivative y ′ of the target deviation and the inclination of the switching hyperplane σn increases. φ is set, and the target steering angle θtotal is calculated so that the nonlinear input for switching the target deviation y and the time differential y ′ of the target deviation to reach the hyperplane σn decreases in the boundary layer φ. The calculation process of the target steering angle θtotal and the setting process of the boundary layer φ will be described later.

  The command current value conversion unit 44 constitutes control means, and converts the target steering angle θtotal calculated by the target steering angle calculation unit 43 into a command current value (target steering current) i. The command current value i is a current value for energizing the steering actuator 6 in order to cause the vehicle 1 to travel along the target travel line LT. The command current value conversion unit 44 calculates how much current needs to be supplied to the steering actuator 6 in either the left or right direction in order to drive the steering wheel 2 to the target steering angle θtotal. . The drive current output unit 5 constitutes a control unit, and supplies the steering actuator 6 with the command current value i calculated by the command current value conversion unit 44.

  The steering actuator 6 is driven by the drive current i input from the drive current output unit 5. The steering wheel 2 rotates in conjunction with the driving of the steering actuator 6.

  Next, the target steering angle θtotal calculation process and the boundary layer φ setting process executed by the target steering angle calculation unit 43 will be described. In addition, the definition of the symbol used in the following description is as follows.

I: Yaw moment of inertia of vehicle 1 (kg · m ² )
M: Mass of vehicle 1 (kg)
K _f : Front wheel cornering power (N / rad)
K _r : Rear wheel cornering power (N / rad)
l _f : Distance from the center of gravity of the vehicle to the front wheel axle (m)
l _r : Distance from vehicle center of gravity to rear axle (m)
δ: Steering actual steering wheel angle (rad) detected by the steering wheel angle sensor 34
V: Vehicle speed detected by the vehicle speed sensor 35 (m / s)

  The state equation of the steering system is expressed by equation (1).

  The equivalent control input of the steering mode control is calculated by the equation (2).

  The optimal equivalent control input to the regulator that minimizes the evaluation function J of Equation (2) is represented by Equation (3).

なお、Ｋ１，Ｋ２，Ｋ３，Ｋ４は、式（２）を満たす最適ゲインを示す定数である。

K1, K2, K3, and K4 are constants indicating optimum gains that satisfy Expression (2).

  In the case of the conventional non-linear control input of general sliding mode control, the hyperplane σ shown in Equation (4) is set, and the target steering angle θtotal is calculated by Equation (5). Here, h is a constant indicating the inclination of the switching hyperplane (switching line), and K is a constant indicating the gain.

  However, in the switching hyperplane σ set by the above-described conventional method, the state quantity y and the time change rate y ′ of the state quantity are constrained to a constant linear relationship, and therefore the curvature change of the road (target travel line LT). The robustness with respect to is weak, and the follow-up performance deteriorates as the road curvature increases.

  On the other hand, in this embodiment, as shown in FIG. 5, a hyperplane σp obtained by vertical conversion of the hyperplane σ is obtained (formula (6)), and a new one defined by the two hyperplanes σ and σp is obtained. A coordinate system is set, and a switching hyperplane σn shown in Expression (7) is set in this new coordinate system.

  The switching hyperplane σn is defined by a function including a cubed term of the state quantity (target deviation y) and a cubed term of the time variation rate of the state quantity (time differential y ′ of the target deviation), and the state quantity (target deviation) A nonlinear relationship is established between y) and the time rate of change of the state quantity (time differential y ′ of the target deviation). Further, by changing the value of the coefficient G, the nonlinear switching hyperplane σn can be freely designed.

  The value of the coefficient G is set according to the curvature of the target travel line LT. An example of the relationship between the target travel line LT and the coefficient G (setting information of the coefficient G) is shown in FIG. Note that the relationship between the target travel line LT and the coefficient G is determined and set in advance by experience or the like.

  Accordingly, the non-linear input θnl of the target steering angle θtotal is expressed as shown in Equation (8).

  If the nonlinear input θnl in the equation (8) is used as it is, a chattering phenomenon occurs. Therefore, in order to suppress this, the nonlinear input θnl shown in the equation (9) is used instead of the equation (8).

  The sat (σn / φ) of equation (9) is shown in equation (10).

  In Expressions (9) and (10), φ is a boundary layer, and sgn () in Expression (10) is a sign function. As is clear from equation (10), the magnitude (absolute value) of θnl when the magnitude (absolute value) of σn exceeds the magnitude (absolute value) of φ is a constant value, and the magnitude of σn When is less than or equal to φ, the magnitude of θnl decreases from the constant value. That is, the nonlinear input θnl for causing the state quantity y and the time change rate y ′ of the state quantity to reach the switching hyperplane decreases in the boundary layer.

  The value of the boundary layer φ greatly affects the performance of the sliding mode control. For example, when the value of the boundary layer φ is large, the chattering phenomenon is suppressed, but the convergence (restraint) to the switching hyperplane is deteriorated. On the contrary, if the value of the boundary layer φ is small, the convergence to the switching hyperplane is good, but the chattering phenomenon cannot be reduced.

  In the case of the conventional non-linear control input of general sliding mode control, a predetermined constant value (fixed value) is set as the boundary layer φ. Therefore, the boundary layer φ must be set to an optimum value that has good convergence to the switching hyperplane and can reduce the chattering phenomenon, and it is difficult to determine the value of the boundary layer φ. In particular, when a switching hyperplane in which a nonlinear relationship is established between the state quantity y and the time change rate y ′ of the state quantity is set, it is extremely difficult to determine and set the optimum value of the boundary layer φ in advance. Have difficulty.

  Here, focusing on the relationship between the slope of the convergence trajectory (constraint trajectory) of the state quantity y and the time change rate y ′ of the state quantity in the initial state, the slope of the switching hyperplane σn, and the occurrence of the chattering phenomenon. 7, when the difference between the inclination of the convergence locus of the state quantity y in the initial state and the time change rate y ′ of the state quantity and the inclination of the switching hyperplane σn is large (indicated by a solid line as locus 1). When the difference between the inclination of the convergence trajectory and the inclination of the switching hyperplane σn is small (indicated by a broken line as the trajectory 2), the trajectory 1 generates more chattering than the trajectory 2. For this reason, in this embodiment, the boundary layer φ is not a fixed value, but the gradient of the convergence locus (constraint locus) of the state quantity y of the controlled object in the initial state and the time change rate y ′ of the state quantity and the switching hyperplane σn. The value of the boundary layer φ is set so as to increase as the difference from the slope of the angle increases.

式（１１）において、｜σｎ’｜は、初期状態の収束軌跡の傾きであり、｜３Ｇσｐ^２σｐ’｜は、切り換え超平面σｎの傾きである。また、φ０は、所定の定数である。

In equation (11), | σn ′ | is the slope of the convergence locus in the initial state, and | 3Gσp ² σp ′ | is the slope of the switching hyperplane σn. Φ0 is a predetermined constant.

  The target steering angle θtotal is calculated by Expression (12).

  Next, automatic steering control processing executed by the automatic steering control device of this embodiment will be described.

  The automatic steering control process is repeatedly executed every predetermined time. When the automatic steering control process is started, the controller 4 reads necessary parameters (step S1). In this process, the controller 4 reads various signals from the camera 31, the yaw rate sensor 33, the vehicle speed sensor 35, and the like. Also, based on the read parameters, necessary calculations such as time differentiation and integration, and analysis of the imaging signal are performed.

  Next, the target travel line setting unit 41 sets a target travel line LT (step S2), and the target deviation calculation unit 42 analyzes the imaging signal captured by the camera 31 and calculates the target deviation y (step S3). ).

  Next, the target steering angle calculation unit 43 calculates the target steering angle θtotal (step S4). That is, the value of the coefficient G is obtained from the curvature of the target travel line LT set in step S2 and the preset setting information of the coefficient G, and the target deviation y, the time differential y ′ of the target deviation, and the coefficient G are expressed by the formula ( 4) Substituting into Equations (6) and (7) to set the switching hyperplane σn. Then, the boundary layer φ is set from the equation (11), and the target deviation y and the time derivative y ′ of the target deviation are switched from the equations (9), (10), and (12) to converge on the hyperplane σn. The target steering angle θtotal is calculated.

  Next, the command current value conversion unit 44 converts the target steering angle θtotal into the command current value i (step S5), and the drive current output unit 5 drives the steering actuator 6 by energization of the command current value i (step S6). ), The automatic steering control process is terminated.

  By repeatedly executing the processing of step S1 to step S6, the target deviation y (state quantity to be controlled) and the time deviation y ′ (time rate of change of the state quantity) of the target deviation are converged on the switching hyperplane σn. The

  According to the present embodiment, the inclination of the convergence locus of the target deviation y (state quantity of the controlled object) in the initial state and the time derivative y ′ (time change rate of the state quantity) of the target deviation and the inclination of the switching hyperplane σn. Since the value of the boundary layer φ is set to be larger as the difference is larger, it is possible to accurately reduce the occurrence of chattering while maintaining good convergence to the switching hyperplane σn.

  In the switching hyperplane σn, a non-linear relationship is established between the target deviation y and the time derivative y ′ of the target deviation, and the target deviation y and the time derivative y ′ of the target deviation are not constrained to a constant linear relationship. Therefore, it is possible to design a switching hyperplane σn having high robustness that exhibits optimum tracking performance as the road curvature increases.

  The above-described embodiment is an example of the present invention. For this reason, the present invention is not limited to the above-described embodiments, and various modifications can be made without departing from the technical idea according to the present invention, even if other than the above-described embodiments. Of course.

  For example, in the above embodiment, the case where the boundary layer φ is set on the switching hyperplane σn in which a nonlinear relationship is established between the state quantity of the control target and the time change rate of the state quantity has been described. Similarly to the present embodiment, the boundary layer φ may be set on the switching hyperplane where a linear relationship is established between the quantity and the time change rate of the state quantity. In the above-described embodiment, the hyperplane σn is switched by a function including the cube term of the target deviation y (state quantity of the control target) and the cube term of the time derivative y ′ (time change rate of the quantity of state) of the target deviation. However, the switching hyperplane may be defined by another function that establishes a non-linear relationship between the target deviation y and the time derivative y ′ of the target deviation. Moreover, although the example of the sliding mode control device applied to the automatic steering control device has been described in the above embodiment, the sliding mode control device of the present invention is widely applicable to various devices other than the automatic steering control device.

  The present invention is applicable to an apparatus that performs sliding mode control.

1: Vehicle 2: Steering wheel 5: Drive current output unit (control means)
6: Steering actuator 42: Target deviation calculator 43: Target steering angle calculator (hyperplane setting means, boundary layer setting means, control means)
44: Automatic steering current calculation unit (control means)
LT: Target travel line (target course)
y: target deviation (distance between vehicle position and target course, state quantity of control target)
y ′: Time differential of target deviation (time change rate of distance between vehicle position and target course, time change rate of state quantity)
σn: switching hyperplane θtotal: target steering angle φ: boundary layer

Claims (3)

Boundary layer setting means for setting a boundary layer on the switching hyperplane;
A non-linear input for converging the state quantity of the controlled object and the time change rate of the state quantity on the switching hyperplane and causing the state quantity of the controlled object and the time change rate of the state quantity to reach the switching hyperplane. Control means for reducing in the boundary layer,
The boundary layer setting means is configured to increase the value as the difference between the inclination of the convergence locus of the state quantity of the control target in the initial state and the time change rate of the state quantity and the inclination of the switching hyperplane increases. A sliding mode control device characterized by setting a layer. The sliding mode control device according to claim 1,
A sliding mode control device comprising hyperplane setting means for setting the switching hyperplane so that a non-linear relationship is established between the state quantity of the control target and the time change rate of the state quantity. An automatic steering control device for a vehicle having the sliding mode control device according to claim 1 or 2,
Vehicle position detection means for detecting the current position of the vehicle;
Route setting means for setting a target route of the vehicle,
The state quantity of the control target is a distance between the position of the vehicle and the target course,
The control means calculates a target steering angle for converging a distance between the position of the vehicle and the target route and a time change rate of the distance to the switching hyperplane, and a target steering corresponding to the calculated target steering angle. An automatic steering control device that outputs current to a steering actuator.

Images