JP2012108730A - Control input setting method of sliding mode control unit - Google Patents

Abstract

In a sliding mode control apparatus, a smoothing function parameter of a switching input can be theoretically determined.
A smoothing function parameter δ, a value σ _th [σ _{e −safety} factor] and a fluctuation of input in which a safety factor is expected for a steady deviation limit σ _e on a coordinate having a steady deviation σ and a switching input u _nl as coordinate axes. by setting the maximum value d _max and scope of the switching input shaft side of the straight line La connecting the origin [0,0] of Pa coordinate point determined by the switching input u _nl is input when the state deviation limit sigma _e It becomes higher than the maximum fluctuation value d _max , and robustness against input fluctuation can be ensured. Thus, the smoothing function parameter δ is theoretically set by setting the smoothing function parameter δ based on the steady-state deviation limit σ _e and the input fluctuation maximum value d _max using the σ− _un n coordinates. Thus, variations in tuning results and tuning man-hours by the operator can be reduced.
[Selection] Figure 4

Description

  The present invention relates to a control input setting method for a sliding mode control device used in a control system such as an actuator.

  Conventionally, in a control system such as an actuator, feedback control such as PID control has been performed as control for tracking a plurality of state quantities (for example, displacement, displacement speed, displacement acceleration, etc.) of a control target to a desired target state quantity. Generally done. However, in conventional control such as PID, it may be difficult to ensure sufficient robustness against disturbances and characteristic changes of the controlled object.

  Therefore, in recent years, sliding mode control, which is one of the current controls, is applied to control of a control system such as an actuator. In the sliding mode control, a switching hyperplane represented by a switching function having a plurality of state quantities to be controlled as variables is designed in advance, and the state quantity to be controlled is converged on the switching hypervariable plane by high gain control ( The value of the switching function is converged to 0), and the state quantity is constrained on the super switching plane by the equivalent control input (see, for example, Patent Documents 1 to 4). According to such sliding mode control, the state quantity of the controlled object can be stably constrained on the switching hyperplane while sufficiently securing robustness against disturbance, characteristic change of the controlled object, and the like.

JP 2001-134302 A JP 2005-293564 A JP 2009-299510 A JP 2010-229973 A

  In a control device that controls a controlled object using sliding mode control, a phenomenon in which the state quantity of the controlled object vibrates at a high frequency near the switching hyperplane, so-called chattering, may occur, and in order to suppress such sliding mode chattering A smoothing function may be used. In this case, in order to suppress chattering while maintaining robustness, tuning (adaptation) of parameters of the switching input gain and the smoothing function is required, but a method for theoretically determining the smoothing function parameter has not been established. . For this reason, the tuning of the smooth function parameters is currently performed by the operator on a trial and error basis based on experience, and the tuning results and the tuning man-hours inevitably vary. In addition, it is necessary to confirm the robustness after tuning by actual machine evaluation.

  The present invention has been made in consideration of such a situation, and in a sliding mode control apparatus for controlling a controlled object by sliding mode servo control, a switching input of two control inputs (equivalent control input, switching input) is provided. It is an object of the present invention to realize a control input setting method capable of theoretically determining the smoothing function parameter.

  The present invention provides a control input (equivalent control input (linear input), switching input (nonlinear input)) for converging the state quantity on a switching hyperplane set by a switching function having the state quantity to be controlled as a variable. A control input setting method applied to a sliding mode control device that calculates and controls a control target by sliding mode servo control (servo control including an error integral value z = ∫ (r−y) dt in a state quantity) is assumed. In such a control input setting method, when chattering in the sliding mode is suppressed using a smooth function, the smooth function parameter is expressed as a steady deviation limit on a coordinate having a steady deviation and a switching input as coordinate axes. A technical feature is that the setting is made in a range on the switching input axis side of a straight line connecting a point determined by the maximum value of input fluctuation and the origin of coordinates.

  According to the present invention, in a sliding mode servo control system in which the state quantity includes the error integral value z = ∫ (r−y) dt, the smoothing function parameter δ (see formula (10) described later) of the switching input is expressed as a steady-state deviation. And the switching input on the coordinate (see FIG. 4) on the coordinate axis (see FIG. 4), it is set in the range on the switching input axis side from the straight line connecting the point determined by the steady deviation limit and the input fluctuation maximum value and the origin of the coordinate. At the time of the deviation limit, the switching input (arrival input) becomes higher than the input fluctuation maximum value, and robustness against the input fluctuation can be ensured. And by setting the smoothing function parameter to a value as small as possible in the range where the chattering of the sliding mode does not occur in the range on the switching input axis side than the straight line on the coordinates, while ensuring robustness, Chattering can be suppressed.

  The steady deviation limit can be obtained from the control target performance (steady deviation, steady deviation allowable time) given from the control specifications and the like. Further, the input fluctuation maximum value can be set (assumed) from an input disturbance.

  As described above, in the present invention, since the smoothing function parameter of the switching input is set based on the steady deviation limit and the input fluctuation maximum value using coordinates with the steady deviation and the switching input as coordinate axes, chattering is suppressed. The smoothing function parameter to be set can be theoretically set. Thereby, the variation in the tuning result and the tuning man-hour by the operator can be reduced. In addition, the number of tuning steps can be reduced.

In the present invention, for example, as shown in FIG. 4, a value σ _th [σ _{e −safety} factor] in which the safety factor is expected for the steady deviation limit σ _e on the coordinates having the steady deviation σ and the switching input un _nl as coordinate axes. And the smoothing function parameter δ may be set in a range on the switching input axis (vertical axis) side of the straight line La connecting the point Pa determined by the input fluctuation maximum value d _max and the coordinate origin [0, 0]. In this way, robustness can be ensured more reliably.

  According to the present invention, the smoothing function parameter of the switching input of the sliding mode servo control is set based on the steady deviation limit and the input fluctuation maximum value using coordinates with the steady deviation and the switching input as coordinate axes. Thus, it is possible to theoretically set the smoothing function parameter that suppresses chattering, and it is possible to reduce variations in tuning results and tuning man-hours by the operator.

1 is a schematic configuration diagram of a vehicle equipped with an automated manual transmission to which the present invention is applied. It is a block diagram which shows the structure of the hydraulic control circuit which controls the hydraulic pressure of a select actuator, a shift actuator, and a clutch actuator, and the structure of control systems, such as ECU. It is a control block diagram of a servo type sliding mode control system. It is a figure which shows typically the process at the time of tuning a smooth function parameter.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  In this example, an automatic manual transmission in which shift shifting (gear stage switching) is automated is targeted, and the present invention is applied to a sliding mode control for controlling an actuator (including an actuator of an automatic clutch) of the automated manual transmission. An example will be described.

  First, a vehicle equipped with an automated manual transmission will be described with reference to FIG. The vehicle of this example is equipped with an engine 1, an automatic clutch 2, the automated manual transmission 3, a hydraulic control circuit 400, an ECU (Electronic Control Unit) 100, and the like.

  The engine 1 is an internal combustion engine such as a gasoline engine or a diesel engine, and a crankshaft 11 as an output shaft thereof is connected to the automatic clutch 2.

  The automatic clutch 2 includes a dry single-plate friction clutch 20 and a clutch operating device 200. The friction clutch 20 includes a flywheel, a clutch disk, a pressure plate, a diaphragm spring, a clutch cover, and the like that are connected to the crankshaft 11 of the engine 1 (see, for example, JP 2010-203586 A).

  The clutch operating device 200 includes a release bearing, a release fork, a hydraulic clutch actuator 201, and the like. A driving force (forward / reverse movement) of the clutch actuator 201 is transmitted to the diaphragm spring via the release fork and the release bearing. To displace the pressure plate of the friction clutch 20 so that the clutch disk is strongly sandwiched between the pressure plate and the flywheel (clutch engagement state), or the pressure plate is pulled away from the clutch disk. The state (clutch disengagement state) can be set (see, for example, JP 2010-203586 A). The driving of the clutch actuator 201 (hydraulic oil flow rate control) is controlled by the hydraulic control circuit 400 and the ECU 100.

  The automated manual transmission 3 has a configuration similar to that of a general manual transmission such as a parallel gear transmission having six forward speeds and one reverse speed.

  The automated manual transmission 3 (hereinafter also referred to as “transmission 3”) is a constantly meshing stepped transmission including a synchromesh mechanism 300, and includes an input shaft 31, an output shaft 32, and the input shaft 31 and output. Six pairs of gears 311, 312,... 316 with different reduction ratios provided on the shaft 32 are provided, and one of the gear pairs 311, 312,. Thus, the six forward speed ratio is set. FIG. 1 schematically shows only three gear pairs 311 (gears 311a and 311b), a gear pair 312 (gears 312a and 312b), and a gear pair 316 (316a and 316b).

  One (input shaft side) gears 311a, 312a,... 316a constituting the gear pair 311, 312,... 316 are supported so as to rotate or idly rotate integrally with the input shaft 31 of the transmission 3, and the other ( The gears 311b, 312b,... 316b on the output shaft side are supported by the output shaft 32 of the transmission 3 so as to integrally rotate or idle. In this example, the input shaft side gears 311a, 312a,... 316a are supported so as to rotate integrally with the input shaft 31, and the output shaft side gears 311b, 312b,. It is supported to idle.

  The gears 311a, 312a,... 316a of these gear pairs 311, 312,... 316 and the gears 311b, 312b,. One of the gear pairs, for example, the gear 311b on the output shaft side of the gear pair 311 is connected to the output shaft 32 by the synchromesh mechanism 300, whereby the gear pair 311 enters the power transmission state, and the gear corresponding to the gear pair. A stage (for example, the first speed (1st)) can be obtained.

  Further, by connecting the gear 316b on the input shaft side of the gear pair 316 to the output shaft 32 by the synchromesh mechanism 300, the gear pair 316 enters a power transmission state, and the gear stage corresponding to the gear pair (for example, the sixth speed (6th )) Can be obtained. Although not shown, for example, a reverse gear pair is provided on the input shaft 31 of the transmission 3, and the reverse gear stage can be set by engaging the idle gear provided on the countershaft with the reverse gear pair. It has become.

  An input shaft 31 of the transmission 3 is connected to the friction clutch 20 (clutch disc). The rotation of the output shaft 32 of the transmission 3 is transmitted to the drive wheels 7 through the differential gear 5 and the axle 6.

  The synchromesh mechanism 300 uses a select actuator 301 and a shift actuator 302 as drive sources. Each drive (hydraulic oil flow rate control) of the select actuator 301 and the shift actuator 302 is controlled by the hydraulic control circuit 400 and the ECU 100.

  As shown in FIG. 2, the hydraulic control circuit 400 includes a select solenoid valve 401, a shift solenoid valve 402, a clutch solenoid valve 403, an oil pump 404, and the like. The select solenoid valve 401, the shift solenoid valve 402, and the clutch solenoid valve 403 are all linear solenoid valves. As the oil pump 404, a mechanical oil pump or an electric oil pump driven by the operation of the engine 1 is used.

  The select solenoid valve 401 is configured to adjust the flow rate of hydraulic oil (hydraulic pressure: stroke of the select actuator 301) supplied to the select actuator 301 of the automated manual transmission 3 in accordance with a command signal (drive current) from the ECU 100. Has been. When the piston rod of the select actuator 301 moves forward or backward in the select direction, the select operation of the synchromesh mechanism 300 of the automated manual transmission 3 is performed (see, for example, Japanese Patent Application Laid-Open No. 2010-203586). The movement amount (select stroke) of the piston rod of the select actuator 301 is detected by a select stroke sensor 501.

  The shift solenoid valve 402 is configured to adjust the flow rate of hydraulic fluid (hydraulic pressure: stroke of the shift actuator 302) supplied to the shift actuator 302 of the automated manual transmission 3 in accordance with a command signal (drive current) from the ECU 100. ing. When the piston rod of the shift actuator 302 moves forward or backward in the shift direction, the shift operation of the synchromesh mechanism 300 of the automated manual transmission 3 is performed (see, for example, JP 2010-203586 A). The shift amount (shift stroke) of the piston rod of the shift actuator 302 is detected by the shift stroke sensor 502.

  The clutch solenoid valve 403 is configured to adjust the flow rate of hydraulic oil (hydraulic pressure: stroke of the clutch actuator 201) supplied to the clutch actuator 201 of the automatic clutch 2 in accordance with a command signal (drive current) from the ECU 100. ing. When the piston rod of the clutch actuator 201 moves forward or backward, the automatic clutch 2 (friction clutch 2) is disconnected or engaged. A movement amount (clutch stroke) of the piston rod of the clutch actuator 201 is detected by a clutch stroke sensor 503.

  A drive current is continuously supplied during the drive control of the select solenoid valve 401, the shift solenoid valve 402, and the clutch solenoid valve 403. In addition, since the output is substantially continuous even in high-speed current ON-OFF control (duty control) such as PWM, energization control in such a case is also included.

-ECU-
The ECU 100 executes control of the clutch operation device 200 and the shift operation device 300. In addition, the ECU 100 of this example includes a sliding mode controller (SMC) 101, an observer 102, a limiter 103, and the like that constitute the following sliding mode control system (see FIG. 2).

-Servo type sliding mode control system-
Next, a servo type sliding mode control system including the sliding mode controller (SMC) 101 will be described with reference to a control block diagram of FIG. Since the control of the select actuator 301 of the automated manual transmission 3 and the control of the shift actuator 302 are basically the same, in this example, the select actuator 301 will be described as a representative.

  The sliding mode control system shown in FIG. 3 includes the sliding mode controller 101, the observer 102, the limiter 103, and the like. The state quantity of the control target T (for example, the stroke displacement of the select actuator 301, the stroke displacement speed, Stroke displacement acceleration) is controlled.

  The observer 102 estimates the estimated state quantity [x0 ^] (stroke displacement) based on the control input u input to the control target T and the actual stroke y of the control target T (for example, the detection value of the select stroke sensor 501). [X1 ^] (stroke displacement speed), [x2 ^] (stroke displacement acceleration), and estimated stroke [y ^] are estimated. [X0 ^], [x1 ^], [x2 ^], and [y ^] are assumed to be used instead of “hat”, which is a general notation for expressing an estimated value, and is shown in FIG. ) Uses general notation.

The sliding mode controller 101 is a known servo controller to which the sliding mode control theory is applied, and a deviation between the target stroke (target value) r and the estimated stroke [y ^] (actual stroke (output) y), and The estimated state quantities ([x0 ^], [x1 ^], [x2 ^]) are input. The sliding mode controller 101 calculates the control input u based on the deviation between the target stroke r and the estimated stroke [y ^] and the estimated state quantity, and gives it to the controlled object T. The control input u is configured will be described later equivalent control input (linear input) u _eq and the switching input of two control inputs of the (nonlinear input) u _nl.

  Here, in the servo type sliding mode control system of this example, fluctuations in drive current (drive current variation) of the select solenoid valve 401 that controls the hydraulic oil flow rate (hydraulic pressure) of the select actuator 301 are treated as input disturbances. .

  In the servo type sliding mode control system shown in FIG. 3, the limiter 103 is provided in consideration of the input saturation of the control target T.

-Control input setting method-
Next, a control input setting method of the sliding mode controller (SMC) 101 of this example will be described.

<Derivation of model formula for SMC design>
Since responsiveness and positioning accuracy are important, a servo system is constructed. First, in order to design the servo system, a relational expression (equation (1) below) of the integral value (error integral value) z of the difference between the target value r and the output y is defined.

  A type 1 servo system is configured by using an expansion system in which the error integral value z of (1) above is added to the state variables (state equation (2) below).

  The above expression (2) is expanded into the following vector notation state equation (expression (3)).

<Equivalent control input u _eq >
Of the two control inputs of the sliding mode control, an equivalent control input u _eq is derived. For the equivalent control input u _eq , the behavior of the sliding mode state system is determined by a linear input. Here, the switching function is defined as follows (Equation (4)).

  The switching function in sliding mode is

When the above state equation (formula (5)) is modified by the switching function σ (x),

It becomes. From the above equation, the equivalent control input u _eq can be expressed as follows (Equation (6)).

<Design of switching hyperplane S>
The design of the switching hyperplane S uses a design method that utilizes the zeros of the system. Specifically, a method is used in which the feedback gain F _reg of optimum control represented by the following equation (Equation (7)) is selected as the switching hyperplane S.

  However, P is a solution of the Riccati equation (equation (8) below) for an arbitrary Q> 0.

  Here, in the control input setting method of this example, a design method for specifying a stability margin is applied. That is, the switching hyperplane S is set so that the real part of the system zero is not more than [−ε]. Specifically, the A matrix is given as in the following equation (Equation (9)), and the switching hyperplane S is set from the Riccati equation.

  The above method of obtaining the switching plane S is an example, and the switching plane S may be obtained by other generally known methods.

<Design of switching input _unl >
Of the two control inputs of the sliding mode control, a switching input (arrival input) _unl is derived. Specifically, in order to realize the sliding mode (σ → 0), a Lyapunov function candidate is selected as follows:

  At this time, the control input is

If it satisfies, as the existence condition of sliding mode

Holds. Based on such conditions, the switching input _unl is defined as follows.

Here, k _{smc in the} equation (10) is a switching input gain. If the switching input gain k _smc is selected to be a positive value (k _smc > 0), the differential value of the Lyapunov function is negative. Value, and a stable sliding mode can be realized.

In order to prevent chattering, a smoothing function was made using the smoothing function parameter δ. FIG. 4 shows the smoothed switching input _unl . As shown in FIG. 4, if the smoothing function parameter δ is set to a large value, the switching input _unl is further smoothed to reduce chattering, but the robustness tends to decrease. If the smoothing function parameter δ is set to a small value, robustness can be maintained, but chattering tends to increase.

<Parameter tuning>
- as a guide for determining the switching input set value of the tuning switching input gain (reaching input gain) k _smc, cited be set to a large value switching input gain k _smc than the maximum value d _max of the matching condition is satisfied variations . By performing such setting, robustness can be ensured. In this example, as described above, since the fluctuation of the drive current of the select solenoid valve 401 is treated as an input disturbance, the switching input gain k _smc is set to be larger than the assumed maximum value of the current fluctuation.

It should be noted that the switching hyperplane S may be increased in order to improve control followability. However, if the switching hyperplane S is set to a large value, the switching input gain k _smc also becomes a large value, and chattering is likely to be induced. Therefore, the switching input gain k _smc is set in consideration of this point.

Tuning of Smooth Function Parameter δ First, the condition for determining the smooth function parameter δ, that is, the necessary condition of the smooth function parameter δ for canceling chattering without impairing the robust performance will be described below.

  In the control input setting method of this example, it is considered that there is no change amount of the state quantity for the switching function σ when the steady deviation occurs.

Can lead to. That is, these expressions mean that the value of the switching function σ increases with the error e as a speed.

Now, assuming that const = 0, sliding based on control target performance (steady deviation [for example, 0.5 mm], steady deviation allowable time [for example, 700 ms duration]) given from control specifications (for example, diagnosis conditions), etc. A steady deviation amount (steady deviation limit σ _e ) that may deviate from the mode is calculated. Here, since the magnitude of the switching input u _nl that can include the input disturbance is determined, the smoothing function parameter δ may be set to be larger than the assumed input disturbance (the above-described variation in driving current). . More specifically, a value σ _th [σ _{e −safety} factor] that anticipates a safety factor is used as the steady deviation limit σ _e , and when the value σ _th [σ _{e −safety} factor] is _satisfied , the switching input _unl is input. By setting the smoothing function parameter δ so as to exceed the fluctuation maximum value d _max, robustness against input fluctuation can be ensured.

A specific example of a method for setting the smoothing function parameter δ will be described with reference to FIG. FIG. 4 shows coordinates with the steady deviation σ as the horizontal axis and the switching input un _nl as the vertical axis.

On the coordinates (σ- _unl coordinate) in FIG. 4, a value σ _th [σ _{e −safety} factor] that takes into account the safety factor in the steady deviation limit σ _e (calculated value) and the input maximum fluctuation value σ _max (above The point determined by the input disturbance (assumed value) is Pa, and the smoothing function parameter δ is within the range of the switching input _unl axis (vertical axis) side from the straight line La connecting the point Pa and the origin of coordinates [0, 0]. Set.

By setting the smoothing function parameter δ in this way, the switching input un _nl exceeds the input fluctuation maximum value d _max when σ _th [σ _{e −safety} factor], and robustness against input fluctuation is ensured. can do. Then, by setting the smoothing function parameter δ to a value as small as possible within the range on the switching input _unl axis (vertical axis) side of the coordinate line La shown in FIG. 4 and in the range where chattering in the sliding mode does not occur. Chattering can be suppressed while ensuring robustness.

As described above, using the coordinates shown in FIG. 4, the smoothing function parameter based on the value σ _{th in} which the safety factor is expected in the steady deviation limit σ _e (calculated value) and the input fluctuation maximum value d _max (assumed value). By setting δ, it is possible to theoretically (visually) set the smoothing function parameter δ that suppresses the chattering of the sliding mode while ensuring the robustness, so that the tuning result by the operator (individual difference) and The variation in tuning man-hours can be reduced, and the tuning man-hours can be reduced.

  Further, based on FIG. 4, it is possible to confirm at a glance whether or not the tuning result of the smoothing function parameter δ satisfies the robustness. This makes it possible to omit the work of evaluating the robustness after tuning with an actual machine.

In the above example, the smoothing function parameter δ is set to a point Pa determined by a value σ _th [σ _{e −safety} factor] that takes into account the safety factor in the steady-state deviation limit σ _e and the input maximum fluctuation value σ _max and the origin of the coordinates. Although it is set in the range on the switching input _unl axis (vertical axis) side from the straight line La connecting [0, 0], the present invention is not limited to this, and the steady deviation limit σ _e on the coordinates shown in FIG. And the smoothing function parameter δ is set in a range on the switching input _unl axis (vertical axis) side of the straight line La connecting the point Pb determined by the input maximum fluctuation value σ _max and the origin [0, 0] of the coordinates. May be. Even if such a setting method is adopted, the switching input _unl becomes higher than the input fluctuation maximum value d _max when the steady deviation limit σ _e is reached, and robustness against the input fluctuation can be ensured.

  In the above example, the control example of the select actuator 301 (shift actuator 302) of the automatic manual transmission 3 has been described. However, the smoothing function parameter δ is also set by the same method for the control of the clutch actuator 201 of the automatic clutch 2. It may be set.

-Other embodiments-
In the above example, the example in which the actuator of the automated manual transmission is controlled by the sliding mode servo control has been described, but the present invention is not limited to this, as long as the state quantity changes with time, The present invention can also be applied to a case where a state quantity to be controlled is controlled by sliding mode servo control (servo control including an error integral value z = ∫ (ry) dt in the state quantity).

  For example, other transmissions such as an automatic transmission that sets a gear stage using a clutch and brake and a planetary gear device, or a belt-type continuously variable transmission (CVT) that continuously adjusts a transmission gear ratio. The present invention can also be applied to the case where the actuator of the machine is controlled by sliding mode servo control (servo control including an error integral value z = ∫ (ry) dt in the state quantity). Further, for example, the present invention can also be applied to the case where sliding mode servo control is performed on various devices and devices such as an internal combustion engine, a braking device (brake), and an electric motor as control targets.

  The present invention can be used for a control input setting method of a sliding mode control device used in a control system such as an actuator, and more specifically, a method of setting a smoothing function parameter of a switching input (nonlinear input) of sliding mode control. Can be used effectively.

100 ECU
DESCRIPTION OF SYMBOLS 101 Sliding mode controller 102 Observer T Control object 2 Automatic clutch 201 Clutch actuator 3 Automated manual transmission 301 Select actuator 302 Shift actuator 400 Hydraulic control circuit 401 Select solenoid valve 402 Shift solenoid valve 403 Clutch solenoid valve 501 Select stroke sensor 502 Shift stroke sensor 503 Clutch stroke sensor

Claims (2)

A sliding mode control device that calculates a control input for converging the state quantity on a switching hyperplane set by a switching function having a state quantity of a controlled object as a variable, and controls the controlled object by sliding mode servo control. A control input setting method to be applied,
When the smoothing function is used to suppress the chattering in the sliding mode, the smooth function parameters are determined by the steady deviation limit and the input fluctuation maximum value on the coordinates with the steady deviation and the switching input as the coordinate axes, and the origin of the coordinates. A control input setting method for a sliding mode control device, characterized in that setting is made in a range on the switching input shaft side with respect to a straight line connecting the two. In the control input setting method of the sliding mode control device according to claim 1,
The smoothing function parameter is set in a range closer to the switching input shaft than a straight line connecting a point determined by a value that allows for a safety factor in a steady deviation limit on the coordinates and the input fluctuation maximum value and the origin of the coordinates. A control input setting method for a sliding mode control device.

Images