Controller for plant

【0001】

[Technical field to which the invention belongs]

The present invention relates to a plant control device, in particular, to identify model parameters indicating the characteristics of a controlled object model obtained by modeling a plant to be controlled in real time, and to perform control using the model parameters. ..

【0002】

[Conventional technology]

The plant to be controlled is modeled, the model parameters of the model to be controlled are calculated by the parameter adjustment mechanism, and the air-fuel ratio of the air-fuel mixture supplied to the internal combustion engine is fed back to the target air-fuel ratio by the adaptive controller using this model parameter. An air-fuel ratio control device for controlling has been conventionally known (for example, Japanese Patent Application Laid-Open No. 11-73206).

【0003】

In this control device, the model parameter is calculated by adding the update component calculated according to the identification error of the model parameter to the initial value of the model parameter. The model parameter update component is modified to reduce the effect of historical identification errors, preventing model parameters from drifting under the influence of disturbances and destabilizing control.

【0004】

[Problems to be Solved by the Invention]

In the apparatus described in the above publication, drift of model parameters due to the influence of disturbance is prevented. However, since the model parameters calculated by the parameter adjustment mechanism are used as they are in the adaptive controller, there is room for improvement in maintaining the stability of the adaptive controller.

【0005】

The present invention has been made focusing on this point, and it is possible to further improve the stability of control using the model parameters while identifying the model parameters of the controlled object model that models the controlled plant. An object of the present invention is to provide a control device for a plant that can be used.

【0006】

[Means for solving problems]

In order to achieve the above object, the invention according to claim 1 is identified by an identification means for identifying a model parameter vector of a controlled object model in which a plant is modeled based on the input and output of the plant, and the identification means. In a plant control device provided with a control means for controlling the plant using the model parameter vector, the identification means includes an identification error calculation means for calculating an identification error (ide) of the model parameter vector and the identification. An update vector calculation means for calculating an update vector (dθ) according to an error, an update vector correction means for modifying the update vector so as to reduce the influence of a past value of the identification error, and an update vector correction means.<u style="single">A reference vector calculation means for calculating a reference vector of the model parameter according to a parameter indicating a change in the dynamic characteristics of the plant, and</u>Before<u style="single">Note</u>Quasi-vector (θbas)<u style="single">e)</u>Model parameter vector calculation means for calculating the model parameter vector by adding the modified update vector to, and elements of the model parameter vector calculated by the model parameter vector calculation means (a1, a2, b1, c1). It is characterized by having a limiting means for limiting the value of.

【0007】

According to this configuration, the update vector is calculated according to the identification error of the model parameter vector, the update vector is modified to reduce the influence of the past value of the identification error, and the update vector is modified to the reference vector of the model parameter. By adding, the model parameter vector is calculated, and the value of the element of the model parameter vector is further limited within a predetermined limit range. Therefore, the stability of control can be further improved while preventing the drift of model parameters.<u style="single">Further, since the reference vector of the model parameter is calculated according to the parameter indicating the change in the dynamic characteristics of the plant, an appropriate reference vector corresponding to the change in the dynamic characteristics of the plant can be obtained. As a result, model parameters can be quickly converged, especially when the plant contains non-linear elements.</u>【0008】

The invention according to claim 2 is characterized in that, in the control device of the plant according to claim 1, the update vector calculation means calculates the update vector by using a fixed gain algorithm. According to this configuration, since the update vector is calculated using the fixed gain algorithm, the amount of calculation can be reduced.

【0009】

The invention according to claim 3 is the control device of the plant according to claim 1 or 2, wherein the update vector correction means has a predetermined value greater than 0 and less than 1 in the past value of at least one element of the update vector. It is characterized in that the update vector is modified by multiplying (DELTAi, EPSi).

【0010】

According to this configuration, the update vector is modified by multiplying the past value of at least one element of the update vector by a predetermined value greater than 0 and less than 1, thus reducing the effect of the past value of the identification error. , It is possible to prevent the drift of the model parameter vector.

【0011】

The invention according to claim 4 is the control device of the plant according to claim 3, wherein the update vector correction means is an element (element related to the calculation of b1) of the update vector related to the input of the plant or the said. Elements not related to plant input / output (elements related to the operation of c1) are not multiplied by the predetermined values (DELTAi, EPSi).

【0012】

According to this configuration, the elements of the update vector that are related to the input of the plant or the elements that are not related to the input and output of the plant are not multiplied by a predetermined value that is greater than 0 and less than 1, and therefore a steady deviation occurs due to the modification of the update vector. Can be prevented.

【0015】

Claim<u style="single">5</u>The invention described in the present invention uses an identification means for identifying a model parameter vector of a controlled object model in which a plant is modeled based on the input and output of the plant, and a model parameter vector identified by the identification means.<u style="single">, So that the output of the plant matches the control target value</u>In a plant control device provided with a control means for controlling the plant, the identification means includes an identification error calculating means for calculating an identification error (ide) of the model parameter vector and an identification error calculating means for calculating the identification error (ide).<u style="single">A dead zone setting means for setting a dead zone according to the output or control target value of the plant, and</u>The identification error<u style="single">The dead zone</u>When (-EIDNRLMT ide EIDNRLMT), the model parameter vector is used by the identification error correction means for correcting the identification error in the decreasing direction and the identification error (idenl) corrected by the identification error correction means. It is characterized by having a model parameter vector calculating means for calculating the above, and a limiting means for limiting the value of the element of the model parameter vector calculated by the model parameter vector calculating means within a predetermined limit range.

【0016】

According to this configuration, the identification error of the model parameters<u style="single">Dead zone</u>When it is within, the identification error is corrected in the decreasing direction, the model parameter vector is calculated using the corrected identification error, and the value of the element of the model parameter vector is further limited within a predetermined limit range. Therefore, the stability of control can be further improved while preventing the drift of model parameters.<u style="single">Further, since the dead zone is set according to the output of the plant or the control target value, it is possible to prevent the identification error to be reflected in the value of the model parameter from being ignored as the residual identification error.</u>【0017】

Claim<u style="single">6</u>The invention according to claim 1<u style="single">5</u>Any of<u style="single">Item 1</u>In the plant control device according to the above, the limiting means limits the values of the plurality of elements so that the plurality of elements (a1, a2) of the model parameter vector satisfy a predetermined relationship (FIG. 18). It is characterized by.

【0018】

According to this configuration, the values of the plurality of elements are limited so that the plurality of elements of the model parameter vector satisfy a predetermined relationship, so that the stability of the control using the model parameter vector can be improved. Claim<u style="single">7</u>The invention described in is claimed.<u style="single">5</u>In the plant control device according to the above, the identification error correcting means has the identification error (ide).<u style="single">Dead zone</u>When it is within (-EIDNRLMT ide EIDNRL MT), the identification error (ide) is set to 0.

【0019】

According to this configuration, the identification error is as described above.<u style="single">Dead zone</u>When it is inside, the identification error is set to 0, so that the influence of the identification error that should not be reflected in the value of the model parameter can be eliminated and the drift prevention effect of the model parameter can be enhanced. Claim<u style="single">8</u>The invention according to claim 1<u style="single">7</u>Any of<u style="single">Item 1</u>In the control device of the plant according to the above, the plant includes a throttle valve of an internal combustion engine and a throttle valve drive device including a drive means for driving the throttle valve, and the control means determines the opening degree of the throttle valve. It is characterized in that a parameter for determining a control input to the throttle valve drive device is calculated so as to match the target opening degree.

【0020】

According to this configuration, the throttle valve opening is controlled to match the target opening using the model parameters identified by the identification means, so that the controllability of the throttle valve opening to the target opening is improved. At the same time, the stability of throttle valve opening control can be improved.

【0021】

BEST MODE FOR CARRYING OUT THE INVENTION

Hereinafter, embodiments of the present invention will be described with reference to the drawings. (First Embodiment) FIG. 1 is a diagram showing a configuration of a throttle valve control device according to a first embodiment of the present invention. A throttle valve 3 is provided in the intake passage 2 of the internal combustion engine (hereinafter referred to as engine) 1. The throttle valve 3 has a return spring 4 as a first urging means for urging the throttle valve 3 in the valve closing direction and elasticity as a second urging means for urging the throttle valve 3 in the valve opening direction. The member 5 and the member 5 are attached. Further, the throttle valve 3 is configured to be driven by a motor 6 as a driving means via a gear (not shown). When the driving force of the motor 6 is not applied to the throttle valve 3, the opening TH of the throttle valve 3 is the default opening THDEF (for example, 5 degrees) in which the urging force of the return spring 4 and the urging force of the elastic member 5 are balanced. ) Is held.

【0022】

The motor 6 is connected to an electronic control unit (hereinafter referred to as "ECU") 7, and its operation is controlled by the ECU 7. The throttle valve 3 is provided with a throttle valve opening sensor 8 for detecting the throttle valve opening TH, and the detection signal is supplied to the ECU 7.

【0023】

Further, the ECU 7 is connected to an accelerator sensor 9 that detects the depression amount ACC of the accelerator pedal that detects the required output of the driver of the vehicle equipped with the engine 1, and the detection signal is supplied to the ECU 7. The ECU 7 includes an input circuit to which the detection signals of the throttle valve opening sensor 8 and the accelerator sensor 9 are supplied, an AD conversion circuit that converts the input signal into a digital signal, a central arithmetic unit (CPU) that executes various arithmetic processes, and a CPU. It is equipped with a memory for storing the program to be executed, maps and tables referred to by the program, and an output circuit for supplying a drive current to the motor 6. The ECU 7 determines the target opening THR of the throttle valve 3 according to the accelerator pedal depression amount ACC, and determines the control amount DUT of the motor 6 so that the detected throttle valve opening TH matches the target opening THR. , The electric signal corresponding to the control amount DUT is supplied to the motor 6.

【0024】

In the present embodiment, the throttle valve drive device 10 including the throttle valve 3, the return spring 4, the elastic member 5, and the motor 6 is the control target, and the input to the control target is the duty ratio DUT of the electric signal applied to the motor 6. The output to be controlled is the throttle valve opening TH detected by the throttle valve opening sensor 8.

【0025】

When the response frequency characteristics of the throttle valve drive device 10 are actually measured, the gain characteristics and the phase characteristics shown by the solid lines in FIG. 2 can be obtained. Therefore, the model defined by the following equation (1) was set as the controlled model. The response frequency characteristics of this model are shown by the broken lines in Fig. 2, and it has been confirmed that they are close to the characteristics of the throttle valve drive device 10.<img he="15" id="000002" wi="137" file="2_0003602811.tif" img-format="tif" img-content="drawing" />Here, k is a parameter representing the discretized time, and DTH (k) is the throttle valve opening deviation amount defined by the following equation (2). DTH (k) = TH (k) -THDEF (2) Here, TH is the detected throttle valve opening degree, and THDEF is the default opening degree. In addition, a1, a2, b1, and c1 in Eq. (1) are model parameters that determine the characteristics of the controlled model, and d is the dead time.

【0026】

The model defined by the above equation (1) is a DARX model (delayed autoregressive model with exogeneous input) of a discrete-time system adopted to facilitate the application of adaptive control. In equation (1), in addition to the model parameters a1 and a2 related to the output deviation amount DTH and the model parameter b1 related to the input duty ratio DUT, the model parameter c1 not related to input / output is set. This model parameter c1 is a parameter indicating a deviation of the default opening THDEF and a disturbance applied to the throttle valve drive device. That is, the model parameters a1, a2, by the model parameter identifyr<u style="single">b1</u>At the same time, by identifying the model parameter c1, the default opening deviation and disturbance can be identified.

【0027】

FIG. 3 is a functional block diagram of the throttle valve controller realized by ECU7, which includes an adaptive sliding mode controller 21, a model parameter identifier 22, and a predicted throttle valve after the dead time d has elapsed. The target opening degree of the throttle valve 3 according to the state predictor 23 that calculates the opening deviation amount (hereinafter referred to as "predicted deviation amount") PREDTH (k) (= DTH (k + d)) and the accelerator pedal depression amount ACC. It consists of a target opening setting unit 24 for setting the THR.

【0028】

The adaptive sliding mode controller 21 calculates the duty ratio DUT by the adaptive sliding mode control so that the detected throttle valve opening TH matches the target opening THR, and outputs the calculated duty ratio DUT. By using the adaptive sliding mode controller 21, the follow-up response characteristic of the throttle valve opening TH to the target opening THR can be appropriately changed by using a predetermined parameter (VPOLE), and as a result, the throttle valve 3 can be changed. It is possible to avoid an impact (collision with the throttle fully closed stopper) when moving from the valve open position to the fully closed position, and to change the engine response to the accelerator operation. In addition, it is possible to ensure stability against errors in model parameters.

【0029】

The model parameter identifier 22 is a modified model parameter vector θL (θL).^T= [a1, a2, b1, c1]) is calculated and supplied to the adaptive sliding mode controller 21. More specifically, the model parameter identifier 22 calculates the model parameter vector θ based on the throttle valve opening degree TH and the duty ratio DUT. Further, the modified model parameter vector θL is calculated by performing limit processing on the model parameter vector θ, and the modified model parameter vector θL is supplied to the adaptive sliding mode controller 21. In this way, the optimum model parameters a1, a2, and b1 for making the throttle valve opening TH follow the target opening THR are obtained, and further, the model parameter c1 indicating the disturbance and the deviation of the default opening THDEF is obtained.

【0030】

By using the model parameter identifier 22 that identifies model parameters in real time, adaptation to changes in engine operating conditions, compensation for variations in hardware characteristics, compensation for fluctuations in power supply voltage, and adaptation to changes in hardware characteristics over time. Is possible.

【0031】

The state predictor 23 calculates the throttle valve opening TH (predicted value) after the dead time d, more specifically the predicted deviation amount PREDTH, based on the throttle valve opening TH and the duty ratio DUT, and performs the adaptive sliding mode. Supply to controller 21. By using the predicted deviation amount PREDTH, it is possible to secure the robustness of the control system with respect to the dead time of the controlled object, and to improve the controllability in the vicinity of the default opening THDEF where the dead time is particularly large.

【0032】

Next, the operating principle of the adaptive sliding mode controller 21 will be described. First, the target value DTHR (k) is defined as the deviation amount between the target opening THR (k) and the default opening THDEF by the following equation (3). DTHR (k) = THR (k) -THDEF (3) Here, if the deviation e (k) between the throttle valve opening deviation amount DTH and the target value DTHR is defined by the following equation (4), the adaptive sliding mode controller The switching function value σ (k) of is set as shown in the following equation (5).<img he="30" id="000003" wi="123" file="3_0003602811.tif" img-format="tif" img-content="drawing" />Here, VPOLE is a switching function setting parameter that is set to a value greater than -1 and less than 1.

【0033】

On the phase plane defined with the vertical axis as the deviation e (k) and the horizontal axis as the previous deviation e (k-1), the deviation e (k) satisfying σ (k) = 0 and the previous deviation e (k) Since the combination with -1) becomes a straight line, this straight line is generally called a switching straight line. The sliding mode control focuses on the behavior of the deviation e (k) on this switching straight line, so that the switching function value σ (k) becomes 0, that is, the deviation e (k) and the previous deviation e (k). Control so that the combination of -1) is placed on the switching straight line on the phase plane, and robust against disturbances and modeling errors (difference between the characteristics of the actual plant and the characteristics of the modeled controlled model). The control is realized and the throttle valve opening deviation amount DTH is made to follow the target value DTHR.

【0034】

By changing the value of the switching function setting parameter VPOLE in Eq. (5), as shown in Fig. 4, the damping characteristic of the deviation e (k), that is, the throttle valve opening deviation amount DTH follows the target value DTHR. The characteristics can be changed. Specifically, when VPOLE = -1, the characteristic does not follow at all, and the smaller the absolute value of the switching function setting parameter VPOLE, the faster the following speed can be.

【0035】

The throttle valve control device is required to satisfy the following requirements A1 and A2. A1) Avoid collision with the throttle fully closed stopper when moving the throttle valve 3 to the fully closed position. A2) Non-linear characteristics near the default opening THDEF (the urging force of the return spring 4 and the urging force of the elastic member 5) Improved controllability for changes in elastic characteristics due to equilibrium, backlash of gears interposed between motor 6 and throttle valve 3, and dead zone where throttle valve opening does not change even if duty ratio DUT changes. Therefore, it is necessary to reduce the convergence speed of the deviation e (k) in the vicinity of the fully closed position of the throttle valve and increase the convergence speed in the vicinity of the default opening degree THDEF.

【0036】

According to the sliding mode control, the convergence speed can be easily changed by changing the switching function setting parameter VPOLE. Therefore, in the present embodiment, the amount of change in the throttle valve opening TH and the target value DTHR is DDTHR (= DTTHR (k). )-The switching function setting parameter VPOLE is set according to DTHR (k-1)). Thereby, the above requirements A1 and A2 can be satisfied.

【0037】

As described above, in the sliding mode control, the deviation e (k) is constrained on the switching straight line by constraining the combination of the deviation e (k) and the previous deviation e (k-1) (hereinafter referred to as deviation state quantity). Converges to 0 at the specified convergence speed and robustly against disturbances and modeling errors. Therefore, in the sliding mode control, it is important how to put the deviation state quantity on the switching straight line and constrain it there.

【0038】

From such a viewpoint, the input to the control target (controller output) DUT (k) (also referred to as Usl (k)) reaches the equivalent control input Ueq (k) as shown in the following equation (6). It is constructed as the sum of the rule input Urch (k) and the adaptive rule input Uadp (k).<img he="15" id="000004" wi="123" file="4_0003602811.tif" img-format="tif" img-content="drawing" /> 【0039】

The equivalent control input Ueq (k) is an input for constraining the deviation state quantity on the switching straight line, and the arrival rule input Urch (k) is an input for placing the deviation state quantity on the switching straight line. The rule input Uadp (k) is an input for suppressing the influence of modeling error and disturbance and placing the deviation state quantity on the switching straight line. The calculation method of each input Ueq (k), Urch (k) and Uadp (k) will be described below.

【0040】

Since the equivalent control input Ueq (k) is an input for constraining the deviation state quantity on the switching straight line, the condition to be satisfied is given by the following equation (7). σ (k) = σ (k + 1) (7) Using Eqs. (1) and Eqs. (4) and (5) to obtain the duty ratio DUT (k) that satisfies Eq. (7), the following Eq. (9) ) Is obtained, and this becomes the equivalent control input Ueq (k). Further, the arrival rule input Urch (k) and the adaptive rule input Uadp (k) are defined by the following equations (10) and (11), respectively.

[Number 1]<img he="76" id="000005" wi="153" file="5_0003602811.tif" img-format="tif" img-content="drawing" /> 【0041】

Here, F and G are the reaching rule control gain and the adaptive rule control gain, respectively, and are set as described below. Also, ΔT is the control cycle. The calculation of the above equation (9) requires the throttle valve opening deviation amount DTH (k + d) after the lapse of the dead time d and the corresponding target value DTHR (k + d + 1). Therefore, the predicted deviation amount PREDTH (k) calculated by the state predictor 23 is used as the throttle valve opening deviation amount DTH (k + d) after the lapse of the dead time d, and the target value DTHR (k + d + 1) is used. As a result, the latest target value DTHR will be used.

【0042】

Next, the arrival rule control gain F and the adaptation rule control gain G are determined by the arrival rule input Urch and the adaptation rule input Uadp so that the deviation state quantity is stably placed on the switching straight line. Specifically, assuming a disturbance V (k), the setting conditions for the gains F and G are obtained by finding the conditions for the switching function value σ (k) to be stable with respect to the disturbance V (k). As a result, it was obtained as a stable condition that the combination of gains F and G satisfies the following equations (12) to (14), in other words, it is within the region shown by hatching in FIG.

【0043】

F&gt; 0 (12) G&gt; 0 (13) F &lt;2- (ΔT / 2) G (14) As described above, the equivalent control input Ueq (k) and the reaching rule are obtained by equations (9) to (11). The input Urch (k) and the adaptive law input Uadp (k) can be calculated, and the duty ratio DUT (k) can be calculated as the sum of these inputs.

【0044】

The model parameter identifier 22 calculates the model parameter vector of the controlled object model based on the input (DUT (k)) and the output (TH (k)) of the controlled object as described above. Specifically, the model parameter identifier 22 calculates the model parameter vector θ (k) by the sequential identification algorithm (generalized sequential least squares algorithm) according to the following equation (15). θ (k) = θ (k-1) + KP (k) ide (k) (15) θ (k)^T= [a1', a2', b1', c1'] (16) [0045]

Here, a1', a2', b1'and C1'are model parameters before the limit processing described later is performed. In addition, ide (k) is an identification error defined by the following equations (17), (18) and (19). DTHHAT (k) is an estimated value of throttle valve opening deviation amount DTH (k) calculated using the latest model parameter vector θ (k-1) (hereinafter referred to as estimated throttle valve opening deviation amount). Is. KP (k) is a gain coefficient vector defined by the following equation (20). Further, P (k) in Eq. (20) is a fourth-order square matrix calculated by Eq. (21) below.

[Number 2]<img he="42" id="000006" wi="153" file="6_0003602811.tif" img-format="tif" img-content="drawing" />[Number 3]<img he="47" id="000007" wi="153" file="7_0003602811.tif" img-format="tif" img-content="drawing" /> 【0046】

By setting the coefficients λ1 and λ2 in Eq. (21), the identification algorithm according to Eqs. (15) to (21) becomes one of the following four identification algorithms. λ1 = 1, λ2 = 0 Fixed gain algorithm λ1 = 1, λ2 = 1 Least squares algorithm λ1 = 1, λ2 = λ Gradual gain algorithm (λ is a predetermined value other than 0,1) λ1 = λ, λ2 = 1 Weight Least squares algorithm (λ is a predetermined value other than 0,1) [0047]

On the other hand, in this embodiment, it is required to satisfy the following requirements B1), B2), and B3). B1) Adaptation to changes in quasi-static dynamic characteristics and variations in hardware characteristics "Changes in quasi-static dynamic characteristics" means changes in characteristics with a slow rate of change, such as fluctuations in power supply voltage and aging of hardware. B2) Adaptation to dynamic dynamic characteristic changes Specifically, it means adaptation to dynamic characteristic changes corresponding to changes in throttle valve opening TH. B3) Prevention of drift of model parameters Prevents problems such as an increase in the absolute value of model parameters due to the influence of identification errors caused by non-linear characteristics of controlled objects that should not be reflected in model parameters.

【0048】

First, in order to satisfy the requirements of B1) and B2) above, the weighted least squares algorithm is adopted by setting the coefficients λ1 and λ2 to the predetermined values λ and 0, respectively. Next, the drift of model parameters will be described. As shown in Fig. 6, after the model parameters have converged to some extent, there is a residual identification error caused by non-linear characteristics such as the friction characteristics of the throttle valve, or when a disturbance with a non-zero mean value is constantly applied. , Residual identification error accumulates and causes drift of model parameters.

【0049】

Since such a residual identification error should not be reflected in the value of the model parameter, the dead zone processing is performed using the dead zone function Fnl as shown in FIG. 7 (a). Specifically, the modified identification error idenl (k) is calculated by the following equation (23), and the model parameter vector θ (k) is calculated using this modified identification error idenl (k). That is, the following equation (15a) is used instead of the above equation (15). Thereby, the above requirement B3) can be satisfied. idenl (k) = Fnl (ide (k)) (23) θ (k) = θ (k-1) + KP (k) idenl (k) (15a) [0050]

The dead band function Fnl is not limited to the one shown in FIG. 7 (a), and is, for example, a discontinuous dead band function as shown in FIG. 7 (b) or an incomplete dead band function as shown in FIG. 7 (c). May be used. However, when the incomplete dead zone function is used, drift cannot be completely prevented.

【0051】

Further, the amplitude of the residual identification error changes according to the amount of fluctuation of the throttle valve opening TH. Therefore, in the present embodiment, the dead band width parameter EIDNRLMT, which defines the width of the dead band shown in FIG. 7, is set according to the mean square value DDTHRSQA of the amount of change in the target throttle valve opening THR calculated by the following equation (24). It is set (specifically, it is set so that the dead band width parameter EIDNRLMT increases as the mean square value DDTHRSQA increases). This makes it possible to prevent the identification error that should be reflected in the value of the model parameter from being ignored as the residual identification error. The DDTHR of the equation (24) is the amount of change in the target throttle valve opening THR and is calculated by the following equation (25).

[Number 4]<img he="39" id="000008" wi="153" file="8_0003602811.tif" img-format="tif" img-content="drawing" /> 【0052】

Here, the throttle valve opening deviation amount DTH is adapted to the target value DTHR. Since the sliding mode controller 21 controls the throttle valve opening deviation amount DTH, similarly, the target value DTHR in the equation (25) is changed to the throttle valve opening deviation amount DTH, and the throttle is throttled. It is also possible to change the dead band width parameter EIDNRLMT by calculating the change amount DDTH of the valve opening deviation amount DTH and using the root mean square value DDTHRSQA obtained by substituting DDTHR for DDTHR in Eq. (24).

【0053】

Further, in order to further enhance the robustness of the control system, it is effective to further stabilize the adaptive sliding mode controller 21. Therefore, in the present embodiment, limit processing is performed on each element a1', a2', b1'and c1'of the model parameter vector θ (k) calculated by the above equation (15), and the modified model parameter vector θL (k) is applied. (θL (k)^T= [a1, a2, b1, c1]) is calculated. Then, the adaptive sliding mode controller 21 executes the sliding mode control using the modified model parameter vector θL (k). The details of the limit processing will be described later with reference to the flowchart.

【0054】

Next, a method of calculating the predicted deviation amount PREDTH by the state predictor 23 will be described. First, the matrices A and B and the vectors X (k) and U (k) are defined by the following equations (26) to (29).

[Number 5]<img he="67" id="000009" wi="153" file="9_0003602811.tif" img-format="tif" img-content="drawing" />The following equation (30) can be obtained by rewriting the equation (1) that defines the controlled model using these matrices A and B and the vectors X (k) and U (k). X (k + 1) = AX (k) + BU (kd) (30) [0055]

When X (k + d) is obtained from Eq. (30), the following Eq. (31) is obtained.

[Number 6]<img he="49" id="000010" wi="153" file="10_0003602811.tif" img-format="tif" img-content="drawing" />Here, if the matrices A'and B'are defined by the following equations (32) and (33) using the model parameters a1', a2', b1'and c1' before the limit processing, the prediction vector XHAT (k + d) ) Is given by the following equation (34).

[Number 7]<img he="81" id="000011" wi="153" file="11_0003602811.tif" img-format="tif" img-content="drawing" /> 【0056】

DTHHAT (k + d), which is an element of the first row of the prediction vector XHAT (k + d), is the predicted deviation amount PREDTH (k) and is given by the following equation (35).<img he="30" id="000012" wi="150" file="12_0003602811.tif" img-format="tif" img-content="drawing" />Here, α1 is matrix A'^d1 row 1 column element, α2 is matrix A'^d1 row 2 column element, βi is matrix A'^diB'one-by-one element, γi is the identity matrix A'^diIt is a 1-by-2 element of B'.

【0057】

The predicted deviation amount PREDTH (k) calculated by the equation (35) is applied to the equation (9), and the target values DTHR (k + d + 1), DTHR (k + d), and DTHR (k +) are further applied. By replacing d-1) with DTHR (k), DTHR (k-1), and DTHR (k-2), respectively, the following equation (9a) can be obtained. The equivalent control input Ueq (k) is calculated by Eq. (9a).

[Number 8]<img he="43" id="000013" wi="153" file="13_0003602811.tif" img-format="tif" img-content="drawing" /> 【0058】

In addition, using the predicted deviation amount PREDTH (k) calculated by equation (35), the predictive switching function value σpre (k) is defined by the following equation (36), and the arrival rule input Urch (k) and adaptive law input Uadp (k) is calculated by the following equations (10a) and (11a), respectively.<img he="15" id="000014" wi="152" file="14_0003602811.tif" img-format="tif" img-content="drawing" />[Number 9]<img he="41" id="000015" wi="153" file="15_0003602811.tif" img-format="tif" img-content="drawing" /> 【0059】

Next, the model parameter c1'is a parameter indicating the deviation and disturbance of the default opening THDEF as described above. Therefore, as shown in FIG. 8, although it fluctuates due to disturbance, the default opening deviation can be regarded as almost constant within a relatively short period of time. Therefore, in the present embodiment, the model parameter c1'is statistically processed, the center value of the fluctuation is calculated as the default opening deviation thdefadp, and used for calculating the throttle valve opening deviation amount DTH and the target value DTHR.

【0060】

The least squares method is generally known as a method of statistical processing, but this statistical processing by the least squares method usually stores all the data within a certain period of time, that is, the identified model parameter c1'in memory. It is executed by performing a batch operation at a certain point. However, this batch calculation method requires a huge amount of memory to store all the data, and further requires an inverse matrix calculation, which causes an increase in the amount of calculation.

【0061】

Therefore, in the present embodiment, the sequential least squares algorithm of adaptive control represented by the above equations (15) to (21) is applied to statistical processing, and the minimum square center value of the model parameter c1 is deviated from the default opening degree. It is calculated as thdefadp.

【0062】

Specifically, θ (k) and θ (k) of the above equations (15) to (21)^TIs replaced with thdefadp, ζ (k) and ζ (k)^TIs replaced with "1", ide (k) is replaced with ec1 (k), KP (k) is replaced with KPTH (k), P (k) is replaced with PTH (k), and λ1 and λ2 Are replaced with λ1'and λ2', respectively, to obtain the following equations (37) to (40).

[Number 10]<img he="59" id="000016" wi="153" file="16_0003602811.tif" img-format="tif" img-content="drawing" /> 【0063】

One of the four algorithms described above can be selected by setting the coefficients λ1'and λ2', but in equation (39), the coefficient λ1'is set to a predetermined value other than 0 or 1, and the coefficient λ2' By setting to 1, we adopted the weighted least squares method.

【0064】

In the operations of the above equations (37) to (40), the only values to be stored are thdefadp (k + 1) and PTH (k + 1), and the inverse matrix operation is unnecessary. Therefore, by adopting the sequential least squares algorithm, it is possible to perform statistical processing of the model parameter c1 by the least squares method while overcoming the drawbacks of the general least squares method.

【0065】

The default opening deviation thdefadp obtained as a result of statistical processing is applied to the above equations (2) and (3), and the throttle valve is replaced by the following equations (41) and (42) instead of the equations (2) and (3). The opening deviation amount DTH (k) and the target value DTHR (k) are calculated. DTH (k) = TH (k) -THDEF + thdefadp (41) DTHR (k) = THR (k)-THDEF + thdefadp (42) [0066]

By using equations (41) and (42), even if the default opening THDEF deviates from the design value due to variations in hardware characteristics or changes over time, the deviation is compensated for and accurate control is performed. be able to.

【0067】

Next, the arithmetic processing in the CPU of the ECU 7 for realizing the functions of the adaptive sliding mode controller 21, the model parameter identifyr 22, and the state predictor 23 described above will be described.

【0068】

FIG. 9 is an overall flowchart of throttle valve opening control, and this process is executed by the CPU of the ECU 7 every predetermined time (for example, 2 msec). In step S11, the state variable setting process shown in FIG. 10 is executed. That is, the calculations of equations (41) and (42) are executed to calculate the throttle valve opening deviation amount DTH (k) and the target value DTHR (k) (FIGS. 10, steps S21 and S22). Note that (k), which indicates the value this time, may be omitted.

【0069】

In step S12, the calculation of the model parameter identifier shown in FIG. 11, that is, the calculation process of the model parameter vector θ (k) according to the above equation (15a) is executed, and further the limit process is executed to execute the modified model parameter vector θL (k). ) Is calculated.

【0070】

In the following step S13, the operation of the state predictor shown in FIG. 21 is executed to calculate the predicted deviation amount PREDTH (k). Next, the arithmetic processing of the control input Usl (k) shown in FIG. 22 is executed using the modified model parameter vector θL (k) calculated in step S12 (step S14). That is, the equivalent control input Ueq, the arrival rule input Urch (k), and the adaptive rule input Uadp (k) are calculated, and the control input Usl (k) (= duty ratio DUT (k)) is calculated as the sum of these inputs. To do.

【0071】

In the following step S16, the stability determination process of the sliding mode controller shown in FIG. 29 is executed. That is, the stability discrimination is performed based on the differential value of the Lyapunov function, and the stability discrimination flag FSMCSTAB is set. This stability determination flag FSMCSTAB indicates that the adaptive sliding mode controller 21 is unstable when set to "1". When the stability determination flag FSMCSTAB is set to "1" and the adaptive sliding mode controller 21 becomes unstable, the switching function setting parameter VPOLE is set to the stabilization predetermined value XPOLESTB (see Fig. 24, steps S231 and S232). At the same time, the equivalent control input Ueq is set to "0", and the control is stabilized by switching to the control using only the arrival rule input Urch and the adaptive rule input Uadp (see FIG. 22, steps S206 and S208). When the adaptive sliding mode controller 21 becomes unstable, the calculation formulas of the arrival rule input Urch and the adaptive rule input Uadp are further changed. That is, the values of the ultimate rule control gain F and the adaptive rule control gain G are changed to the values that stabilize the controller 21, and the ultimate rule input Urch and the adaptive rule input Uadp are calculated without using the model parameter b1 ( See Figures 27 and 28). By the stabilization process as described above, the unstable state of the adaptive sliding mode controller 21 can be terminated at an early stage and returned to the stable state. In step S17, the thdefadp calculation process shown in FIG. 30 is executed to calculate the default opening deviation thdefadp.

【0072】

FIG. 11 is a flowchart of the arithmetic processing of the model parameter identifyr 22. In step S31, the gain coefficient vector KP (k) is calculated by the equation (20), and then the estimated throttle valve opening deviation amount DTHHAT (k) is calculated by the equation (18) (step S32). In step S33, the arithmetic processing of idenl (k) shown in FIG. 12 is executed, and the estimated throttle valve opening deviation amount DTHHAT (k) calculated in step S32 is applied to the equation (17) to apply the identification error ide (k). ), And the dead zone processing using the function shown in Fig. 7 (a) is performed to calculate the modified identification error idenl.

【0073】

In the following step S34, the model parameter vector θ (k) is calculated by the equation (15a), and then the stabilization process of the model parameter vector θ (k) is executed (step S35). That is, the modified model parameter vector θL (k) is calculated by performing the limit processing of each model parameter.

【0074】

FIG. 12 is a flowchart of the idlel (k) arithmetic process executed in step S33 of FIG. In step S51, the identification error ide (k) is calculated by Eq. (17). Next, the value of the counter CNTIDST incremented in step S53 is set to a predetermined value XCNTIDST corresponding to the waste time d of the controlled object (for example, it is set to "3" corresponding to the waste time d = 2). Determine if it is greater than (step S52). Since the initial value of the counter CNTIDST is "0", first proceed to step S53, increment the counter CNTIDST by "1", set the identification error ide (k) to "0" (step S54), and step. Proceed to S55. Immediately after starting the identification of the model parameter vector θ (k), the correct identification error cannot be obtained by the calculation by Eq. (17). Therefore, in steps S52 to S54, the identification error is not used by the calculation result by Eq. (17). I am trying to set ide (k) to "0".

【0075】

If the answer in step S52 is affirmative (YES), the process immediately proceeds to step S55. In step S55, low-pass filtering of the identification error ide (k) is performed. Specifically, when identifying a model parameter of a controlled object having a low-pass characteristic, the identification weight for the identification error ide (k) of the least squares identification algorithm is the frequency characteristic as shown by the solid line L1 in FIG. 13 (a). However, this is a characteristic in which the high frequency component is attenuated as shown by the broken line L2 by low-pass filter processing. This is due to the following reasons.

【0076】

The frequency characteristics of the actual controlled object and the controlled object model that models it are shown by solid lines L3 and L4 in FIG. 13 (b), respectively. That is, when the model parameter is identified by the model parameter identifier 22 for the controlled object having the low-pass characteristic (the characteristic that the high frequency component is attenuated), the identified model parameter is greatly influenced by the high frequency region blocking characteristic. The gain of the controlled model in the low frequency range becomes lower than the actual characteristics. As a result, the correction of the control input by the sliding mode controller 21 becomes an overcorrection.

【0077】

Therefore, by using low-pass filtering to make the frequency characteristics of the weights of the identification algorithm as shown by the broken line L2 in Fig. 13 (a), the frequency characteristics of the controlled model are shown by the broken line L5 in Fig. 13 (b). The characteristics are as shown, and the characteristics are matched with the actual frequency characteristics, or the gain of the controlled model is modified so that it is slightly higher than the actual gain. As a result, overcorrection by the controller 21 can be prevented, the robustness of the control system can be enhanced, and the control system can be made more stable.

【0078】

In the low-pass filter processing, the past values of the identification error ide (ki) (for example, 10 past values corresponding to i = 1 to 10) are stored in the ring buffer, and these past values are multiplied by a weighting coefficient. Execute by adding. Further, since the identification error ide (k) is calculated using the above equations (17), (18) and (19), the throttle valve opening deviation amount DTH (k) and the estimated throttle valve opening deviation are calculated. Perform the same low-pass filter processing as the quantity DTHHAT (k), or the throttle valve opening deviation deviations DTH (k-1) and DTH (k-2) and the duty ratio DUT (kd-1). The same effect can be obtained by performing the low-pass filter processing of.

【0079】

Returning to FIG. 12, in the following step S56, the dead zone processing shown in FIG. 14 is executed. In step S61 of FIG. 14, for example, n = 5 in the above equation (24), the mean square value DDTHRSQA of the change amount of the target throttle valve opening THR is calculated, and then the EIDNRLMT shown in FIG. 15 is calculated according to the mean square value DDTHRSQA. Search the table and calculate the dead band parameter EIDNRLMT (step S62).

【0080】

In step S63, it is determined whether the identification error ide (k) is larger than the dead band width parameter EIDNRLMT, and when ide (k)&gt; EIDNRLMT, the modified identification error idenl (k) is calculated by the following equation (43). (Step S67). idenl (k) = ide (k)-EIDNRLMT (43) [0081]

If the answer in step S63 is negative (NO), it is further determined whether the identification error ide (k) is less than the value of the dead band width parameter EIDNRLMT with a negative sign (step S64), and ide (k). When &lt;-EIDNRLMT, the modified identification error idenl (k) is calculated by the following equation (44) (step S65). idenl (k) = ide (k) + EIDNRLMT (44) If the identification error ide (k) is within ± EIDNRLMT, the modified identification error idenl (k) is set to 0 (step S66).

【0082】

FIG. 16 is a flowchart of the stabilization process of θ (k) executed in step S35 of FIG. In step S71, initialization is performed by setting the flags FA1STAB, FA2STAB, FB1LMT and FC1LMT used in this process to "0", respectively. Then, in step S72, the limit processing of a1'and a2'shown in FIG. 17 is executed, in step S73, the limit processing of b1'shown in FIG. 19 is executed, and in step S74, the limit processing of c1'shown in FIG. 20 is executed. Execute limit processing.

【0083】

FIG. 17 is a flowchart of the limit processing of a1'and a2' executed in step S72 of FIG. FIG. 18 is a diagram for explaining the process of FIG. 17, which is referred to together with FIG. In FIG. 18, the combination of model parameters a1'and a2'that requires limit processing is indicated by "x", and the range of the combination of stable model parameters a1'and a2'is hatched (hereinafter, "" It is indicated by "stable region"). The process of FIG. 17 is a process of moving the combination of model parameters a1'and a2'out of the stable region into the stable region (position indicated by "").

【0084】

In step S81, it is determined whether or not the model parameter a2'is equal to or greater than the predetermined a2 lower limit value XIDA2L. The predetermined a2 lower limit XIDA2L is set to a negative value greater than "-1". Even if the predetermined a2 lower limit XIDA2L is set to "-1", stable modified model parameters a1 and a2 can be obtained, but the n-th power of the matrix A defined in the above equation (26) becomes unstable ( This means that a1'and a2'do not diverge but oscillate), so they are set to a value greater than "-1".

【0085】

If a2'&lt;XIDA2L in step S81, set the modified model parameter a2 to this lower limit XIDA2L and set the a2 stabilization flag FA2STAB to "1". When the a2 stabilization flag FA2STAB is set to "1", it indicates that the modified model parameter a2 is set to the lower limit XIDA2L. In FIG. 18, the modification of the model parameters by the limit processing P1 in steps S81 and S82 is indicated by an arrow line (line with an arrow) with P1.

【0086】

If the answer in step S81 is YES, i.e. a2' XIDA2L, the modified model parameter a2 is set to the model parameter a2'(step S83). In step S84 and step S85, it is determined whether or not the model parameter a1'is within a range within the predetermined a1 lower limit value XIDA1L and the predetermined a1 upper limit value XIDA1H. The predetermined a1 lower limit value XIDA1L is set to a value greater than or equal to -2 and less than 0, and the predetermined a1 upper limit value XIDA1H is set to, for example, 2.

【0087】

If the answers in steps S84 and S85 are both affirmative (YES), i.e. XIDA1L a1' XIDA1H, the modified model parameter a1 is set to the model parameter a1'(step S88). On the other hand, when a1'&lt;XIDA1L, the modified model parameter a1 is set to the lower limit XIDA1L, and the a1 stabilization flag FA1STAB is set to "1" (steps S84, S86). When a1'&gt; XIDA1H, the modified model parameter a1 is set to the upper limit XIDA1H, and the a1 stabilization flag FA1STAB is set to "1" (steps S85, S87). The a1 stabilization flag FA1STAB, when set to "1", indicates that the modified model parameter a1 has been set to the lower limit XIDA1L or the upper limit XIDA1H. In FIG. 18, the modification of the model parameters by the limit processing P2 in steps S84 to S87 is indicated by the arrow line with P2.

【0088】

In step S90, it is determined whether or not the sum of the absolute value of the modified model parameter a1 and the modified model parameter a2 is equal to or less than the predetermined stability determination value XA2STAB. The predetermined stability determination value XA2STAB is set to a value close to "1" and smaller than "1" (for example, 0.99).

【0089】

The straight lines L1 and L2 shown in FIG. 18 are straight lines satisfying the following equation (45). a2 + | a1 | = XA2STAB (45) Therefore, step S90 determines whether the combination of modified model parameters a1 and a2 is on or below the straight lines L1 and L2 shown in FIG. If the answer in step S90 is affirmative (YES), then the combination of modified model parameters a1 and a2 is within the stable region of FIG. 18, and the process ends immediately.

【0090】

On the other hand, when the answer in step S90 is negative (NO), the modified model parameter a1 is the value obtained by subtracting the predetermined a2 lower limit value XIDA2L from the predetermined stability judgment value XA2STAB (since XIDA2L &lt;0, XA2STAB-XIDA2L&gt; XA2STAB It is determined whether or not the following (establishes) (step S91). When the modified model parameter a1 is (XA2STAB-XIDA2L) or less, the modified model parameter a2 is set to (XA2STAB- | a1 |) and the a2 stabilization flag FA2STAB is set to "1" (step S92). ..

【0091】

If the modified model parameter a1 is greater than (XA2STAB-XIDA2L) in step S91, set the modified model parameter a1 to (XA2STAB-XIDA2L), set the modified model parameter a2 to the predetermined a2 lower limit XIDA2L, and stabilize a1. Set both the flag FA1STAB and the a2 stabilization flag FA2STAB to "1" (step S93).

【0092】

In FIG. 18, the modification of the model parameters by the limit processing P3 in steps S91 and S92 is shown by the arrow line with P3, and the modification of the model parameters by the limit processing P4 in steps S91 and S93 is shown. It is indicated by an arrow with "P4". As described above, by the processing of FIG. 17, the limit processing is executed so that the model parameters a1'and a2'are within the stable region shown in FIG. 18, and the modified model parameters a1 and a2 are calculated.

【0093】

FIG. 19 is a flowchart of the limit processing of b1'executed in step S73 of FIG. In steps S101 and S102, it is determined whether or not the model parameter b1'is within a range within the predetermined b1 lower limit value XIDB1L and the predetermined b1 upper limit value XIDB1H. The predetermined b1 lower limit value XIDB1L is set to a positive predetermined value (for example, 0.1), and the predetermined b1 upper limit value XIDB1H is set to, for example, "1".

【0094】

If the answers in steps S101 and S102 are both affirmative (YES), i.e. XIDB1L b1' XIDB1H, the modified model parameter b1 is set to model parameter b1'(step S105). On the other hand, when b1'&lt;XIDB1L, the modified model parameter b1 is set to the lower limit value XIDB1L, and the b1 limit flag FB1LMT is set to "1" (steps S101 and S104). If b1'&gt; XIDB1H, set the modified model parameter b1 to the upper limit XIDB1H and set the b1 limit flag FB1LMT to "1" (steps S102, S103). The b1 limit flag FB1LMT, when set to "1", indicates that the modified model parameter b1 has been set to the lower limit XIDB1L or the upper limit XIDB1H.

【0095】

FIG. 20 is a flowchart of the limit processing of the model parameter c1'executed in step S74 of FIG. In steps S111 and S112, it is determined whether or not the model parameter c1'is within a range within the predetermined c1 lower limit value XIDC1L and the predetermined c1 upper limit value XIDC1H. The predetermined c1 lower limit value XIDC1L is set to, for example, -60, and the predetermined c1 upper limit value XIDC1H is set to, for example, 60.

【0096】

If the answers in steps S111 and S112 are both affirmative (YES), i.e. XIDC1L c1' XIDC1H, the modified model parameter c1 is set to the model parameter c1'(step S115). On the other hand, when c1'&lt;XIDC1L, the modified model parameter c1 is set to the lower limit value XIDC1L, and the c1 limit flag FC1LMT is set to "1" (steps S111, S114). When c1'&gt; XIDC1H, set the modified model parameter c1 to the upper limit XIDC1H and set the c1 limit flag FC1LMT to "1" (steps S112, S113). The c1 limit flag FC1LMT, when set to "1", indicates that the modified model parameter c1 has been set to the lower limit XIDC1L or the upper limit XIDC1H.

【0097】

FIG. 21 is a flowchart of the arithmetic processing of the state predictor executed in step S13 of FIG. In step S121, the matrix operation is executed to calculate the matrix elements α1, α2, β1 to β2, and γ1 to γd of the above equation (35). In step S122, the predicted deviation amount PREDTH (k) is calculated by Eq. (35).

【0098】

FIG. 22 is a flowchart of a process for calculating the control input Usl (= DUT) to the throttle valve drive device 10 executed in step S14 of FIG. In step S201, the arithmetic processing of the predictive switching function value σpre shown in FIG. 23 is executed, and in step S202, the arithmetic processing of the integrated value of the predictive switching function value σpre shown in FIG. 26 is executed. In step S203, the equivalent control input Ueq is calculated by the above equation (9). In step S204, the arithmetic processing of the arrival rule input Urch shown in FIG. 27 is executed, and in step S205, the arithmetic processing of the adaptive law input Uadp shown in FIG. 28 is executed.

【0099】

In step S206, it is determined whether or not the stability determination flag FSMCSTAB set in the process of FIG. 29 described later is "1". When the stability determination flag FSMCSTAB is set to "1", it indicates that the adaptive sliding mode controller 21 is unstable.

【0100】

When FSMCSTAB = 0 in step S206 and the adaptive sliding mode controller 21 is stable, the control input Usl is calculated by adding the control inputs Ueq, Urch and Uadp calculated in steps S203 to S205 (step). S207).

【0101】

On the other hand, when FSMCSTAB = 1 and the adaptive sliding mode controller 21 is unstable, the sum of the arrival rule input Urch and the adaptive rule input Uadp is calculated as the control input Usl. That is, the equivalent control input Ueq is not used to calculate the control input Usl. This makes it possible to prevent the control system from becoming unstable.

【0102】

In the following steps S209 and S210, it is determined whether or not the calculated control input Usl is within the range of the predetermined upper and lower limit values XUSLH and XUSLL, and if the control input Usl is within the range of the predetermined upper and lower limit values, this is immediately performed. End the process. On the other hand, when the control input Usl is equal to or less than the predetermined lower limit value XUSLL, the control input Usl is set to the predetermined lower limit value XUSLL (steps S209, S212), and when the control input Usl is equal to or more than the predetermined upper limit value XUSLH, the control input Usl is set to the predetermined lower limit value XUSLL. Set Usl to the predetermined upper limit XUSLH (steps S210, S211).

【0103】

FIG. 23 is a flowchart of the arithmetic processing of the predictive switching function value σpre executed in step S201 of FIG. In step S221, the arithmetic processing of the switching function setting parameter VPOLE shown in FIG. 24 is executed, and then the arithmetic of the predictive switching function value σpre (k) is executed by the above equation (36) (step S222).

【0104】

In the following steps S223 and S224, it is determined whether or not the calculated predictive switching function value σpre (k) is within the range of the predetermined upper and lower limit values XSGMH and XSGML, and the predicted switching function value σpre (k) is the predetermined upper and lower limit values. If it is within the range of, this process is immediately terminated. On the other hand, when the predictive switching function value σpre (k) is equal to or less than the predetermined lower limit value XSGML, the predictive switching function value σpre (k) is set to the predetermined lower limit value XSGML (steps S223, S225), and the predictive switching function value σpre ( When k) is equal to or greater than the predetermined upper limit value XSGMH, the prediction switching function value σpre (k) is set to the predetermined upper limit value XSGMH (steps S224 and S226).

【0105】

FIG. 24 is a flowchart of the arithmetic processing of the switching function setting parameter VPOLE executed in step S221 of FIG. In step S231, it is determined whether or not the stability determination flag FSMCSTAB is "1", and when FSMCSTAB = 1 and the adaptive sliding mode controller 21 is unstable, the switching function setting parameter VPOLE is stabilized. Set to the specified value XPOLESTB (step S232). The stabilization predetermined value XPOLESTB is set to a value greater than "-1" and very close to "-1" (eg -0.999).

【0106】

When FSMCSTAB = 0 and the adaptive sliding mode controller 21 is stable, the change amount DDTHR (k) of the target value DTHR (k) is calculated by the following equation (46) (step S233). DDTHR (k) = DTHR (k)-DTHR (k-1) (46) In step S234, the VPOLE map is searched according to the throttle valve opening deviation amount DTH and the change amount DDTHR of the target value calculated in step S233, and the switching function setting parameter VPOLE is calculated. As shown in Fig. 25 (a), the VPOLE map increases when the throttle valve opening deviation amount DTH takes a value near 0 (when the throttle valve opening TH takes a value near the default opening THDEF). Values other than around 0 are set to be almost constant with respect to changes in the throttle valve opening deviation amount DTH. The VPOLE map is set so that the VPOLE value increases as the change amount DDTHR of the target value increases, as shown by the solid line in Fig. (B), but the throttle valve opening deviation amount DTH is 0. When the value in the vicinity is taken, as shown by the broken line in the figure, the amount of change DDTHR of the target value is set to increase when the value in the vicinity of 0 is taken.

【0107】

That is, when the target value DTHR of the throttle valve opening has a large change in the decreasing direction, the switching function setting parameter VPOLE is set to a relatively small value. As a result, it is possible to prevent the throttle valve 3 from colliding with the throttle fully closed stopper. Further, in the vicinity of the default opening THDEF, the switching function setting parameter VPOLE is set to a relatively large value, and the controllability in the vicinity of the default opening THDEF can be improved.

【0108】

As shown in FIG. 6C, when the throttle valve opening TH is near the fully closed opening or near the fully open opening, the switching function setting parameter VPOLE may be set to be reduced. As a result, when the throttle valve opening TH is near the fully closed opening or near the fully opened opening, the follow-up speed with respect to the target opening THR becomes slower, and the throttle valve 3 fully closed stopper (functions as a stopper even at the fully opened opening). ) Can be more reliably prevented from colliding with.

【0109】

In the following steps S235 and S236, it is determined whether or not the calculated switching function setting parameter VPOLE is within the range of the predetermined upper and lower limit values XPOLEH and XPOLEL, and when the switching function setting parameter VPOLE is within the range of the predetermined upper and lower limit values. Immediately ends this process. On the other hand, when the switching function setting parameter VPOLE is equal to or less than the predetermined lower limit value XPOLEL, the switching function setting parameter VPOLE is set to the predetermined lower limit value XPOLEL (step S23).<u style="single">6</u>, S23<u style="single">8</u>), When the switching function setting parameter VPOLE is equal to or higher than the predetermined upper limit value XPOLEH, the switching function setting parameter VPOLE is set to the predetermined upper limit value XPOLEH (step S23).<u style="single">5</u>, S23<u style="single">7</u>)。

【0110】

FIG. 26 is a flowchart of a process for calculating the integrated value SUMSIGMA of the prediction switching function value σpre, which is executed in step S202 of FIG. The integrated value SUMSIGMA is used to calculate the adaptive law input Uadp in the process of FIG. 28 described later (see the above equation (11a)).

【0111】

In step S241, the integrated value SUMSIGMA is calculated by the following formula (47). ΔT in the following equation is the execution cycle of the operation. SUMSIGMA (k) = SUMSIGMA (k-1) + σpre × ΔT (47) In the following steps S242 and S243, it is determined whether the calculated integrated value SUMSIGMA is within the range of the predetermined upper and lower limit values XSUMSH and XSUMSL. If the integrated value SUMSIGMA is within the specified upper and lower limit values, this process is immediately terminated. On the other hand, when the integrated value SUMSIGMA is equal to or less than the predetermined lower limit value XSUMSL, the integrated value SUMSIGMA is set to the predetermined lower limit value XSUMSL (steps S242, S244), and when the integrated value SUMSIGMA is equal to or more than the predetermined upper limit value XSUMSH, the integrated value is set. Set SUMSIGMA to the predetermined upper limit XSUMSH (steps S243, S245).

【0112】

FIG. 27 is a flowchart of the arithmetic processing of the arrival rule input Urch executed in step S204 of FIG. In step S261, it is determined whether or not the stability determination flag FSMCSTAB is 1. When the stability determination flag FSMCSTAB is "0" and the adaptive sliding mode controller 21 is stable, the control gain F is set to the predetermined normal gain XKRCH (step S262), and the following equation (48) (the equation (10a)) The arrival rule input Urch is calculated by the same formula as) (step S263). Urch = -F × σpre / b1 (48) [0113]

On the other hand, when the stability determination flag FSMCSTAB is "1" and the adaptive sliding mode controller 21 becomes unstable, the control gain F is set to the predetermined stabilization gain XKRCHSTB (step S264), and the model parameter b1 is not used. The arrival rule input Urch is calculated by the following equation (49) (step S265). Urch = -F × σpre (49) [0114]

In the following steps S266 and S267, it is determined whether or not the calculated arrival rule input Urch is within the range of the predetermined upper and lower limit values XURCHH and XURCHL. This process ends immediately. On the other hand, when the arrival rule input Urch is equal to or less than the predetermined lower limit XURCHL, the arrival rule input Urch is set to the predetermined lower limit XURCHL (steps S266, S268), and when the arrival rule input Urch is equal to or more than the predetermined upper limit XURCHH. , Set the arrival rule input Urch to the predetermined upper limit value XURCHH (steps S267, S269).

【0115】

When the adaptive sliding mode controller 21 becomes unstable in this way, it is adapted by setting the control gain F to the predetermined stabilization gain XKRCHSTB and calculating the arrival rule input Urch without using the model parameter b1.<u style="single">Sliding mode</u>The controller 21 can be returned to a stable state. When the identification by the model parameter identifier 22 becomes unstable, the adaptive sliding mode controller 21 becomes unstable. Therefore, the adaptive sliding mode controller 21 is stabilized by not using the unstable model parameter b1. can do.

【0116】

FIG. 28 is a flowchart of the arithmetic processing of the adaptive law input Uadp executed in step S205 of FIG. In step S271, it is determined whether or not the stability determination flag FSMCSTAB is 1. When the stability determination flag FSMCSTAB is "0" and the adaptive sliding mode controller 21 is stable, the control gain G is set to the predetermined normal gain XKADP (step S272), and the following equation (50) (the equation (11a)) The adaptive law input Uadp is calculated by the equation (corresponding to) (step S273). Uadp = -G × SUMSIGMA / b1 (50) [0117]

On the other hand, when the stability determination flag FSMCSTAB is "1" and the adaptive sliding mode controller 21 becomes unstable, the control gain G is set to the predetermined stabilization gain XKADPSTB (step S274), and the model parameter b1 is not used. The adaptive law input Uadp is calculated by the following equation (51) (step S275). Uadp = -G × SUMSIGMA (51) [0118]

When the adaptive sliding mode controller 21 becomes unstable in this way, it is adapted by setting the control gain G to the predetermined stabilization gain XKADPSTB and calculating the adaptive law input Uadp without using the model parameter b1.<u style="single">Sliding mode</u>The controller 21 can be returned to a stable state.

【0119】

FIG. 29 is a flowchart of the stability determination process of the sliding mode controller executed in step S16 of FIG. In this process, stability discrimination is performed based on the differential term of the Lyapunov function, and the stability discrimination flag FSMCSTAB is set.

【0120】

In step S281, the switching function change amount Dσpre is calculated by the following equation (52), and then the stability discrimination parameter SGMSTAB is calculated by the following equation (53) (step S282). Dσpre = σpre (k) -σpre (k-1) (52) SGMSTAB = Dσpre × σpre (k) (53) In step S283, it is determined whether the stability determination parameter SGMSTAB is equal to or less than the stability determination threshold XSGMSTAB. When SGMSTAB&gt; XSGMSTAB, it is determined that the controller 21 may be unstable, and the instability detection counter CNTSMCST is incremented by "1" (step S285). When SGMSTAB XSGMSTAB, the controller 21 is determined to be stable, and the count value of the instability detection counter CNTSMCST is held without being incremented (step S284).

【0121】

In step S286, it is determined whether or not the value of the instability detection counter CNTSMCST is equal to or less than the predetermined count value XSSTAB. When CNTSMCST XSSTAB, the controller 21 determines that it is stable and sets the first determination flag FSMCSTAB1 to "0" (step S287). On the other hand, when CNTSMCST&gt; XSSTAB, the controller 21 determines that it is unstable and sets the first determination flag FSMCSTAB1 to "1" (step S288). The count value of the instability detection counter CNTSMCST is initialized to "0" when the ignition switch is turned on.

【0122】

In the following step S289, the stability determination period counter CNTJUDST is decremented by "1", and then it is determined whether or not the value of the stability determination period counter CNTJUDST is "0" (step S290). The stability discrimination period counter CNTJUDST is initialized to the predetermined discrimination count value XCJUDST when the ignition switch is turned on. Therefore, the answer in step S290 is negative (NO) at first, and the process immediately proceeds to step S295.

【0123】

After that, when the stability determination period counter CNT JUDST becomes 0, the process proceeds from step S290 to step S291, and it is determined whether or not the first determination flag FSMCSTAB1 is 1. Then, when the first judgment flag FSMCSTAB1 is "0", the second judgment flag FSMCSTAB2 is set to "0" (step S293), and when the first judgment flag FSMCSTAB1 is "1", the second judgment is made. Set the flag FSMCSTAB2 to "1" (step S292).

【0124】

In the following step S294, the value of the stability discrimination period counter CNTZUDST is set to the predetermined discrimination count value XCJUDST, the value of the instability detection counter CNTSMCST is set to 0, and the process proceeds to step S295. In step S295, the stability determination flag FSMCSTAB is set to the logical sum of the first determination flag FSMCSTAB1 and the second determination flag FSMCSTAB2. For the second judgment flag FSMCSTAB2, even if the answer in step S286 becomes affirmative (YES) and the first judgment flag FSMCSTAB1 is set to "0", until the value of the stability judgment period counter CNTJUDST becomes "0", " It is maintained at 1 ". Therefore, the stability determination flag FSMCSTAB is also maintained at "1" until the value of the stability determination period counter CNTJUDST becomes "0".

【0125】

FIG. 30 is a flowchart of the calculation process of the default opening deviation thdefadp executed in step S17 of FIG. In step S251, the gain coefficient KPTH (k) is calculated by the following equation (54). KPTH (k) = PTH (k-1) / (1 + PTH (k-1)) (54) [0126]

Here, PTH (k-1) is a gain parameter calculated in step S253 at the time of the previous execution of this process. In step S252, the model parameters calculated by the model parameter identifyr arithmetic processing shown in FIG.<u style="single">c</u>The gain coefficient KPTH (k) calculated in 1'and step S251 is applied to the following equation (55) to calculate the default opening deviation thdefadp (k).<img he="23" id="000017" wi="148" file="17_0003602811.tif" img-format="tif" img-content="drawing" /> 【0127】

In step S253, the gain parameter PTH (k) is calculated by the following equation (56).<img he="15" id="000018" wi="150" file="18_0003602811.tif" img-format="tif" img-content="drawing" />In the equation (56), λ1'and λ2' are set to predetermined values XDEFADPW and "1" in the equation (39), respectively.

【0128】

By the processing of FIG. 30, the model parameter c1'is statistically processed by the sequential weighted least squares method, and the default opening deviation thdefadp is calculated. In this embodiment, a part of the throttle valve drive device 10 and the ECU 7 (the output circuit that supplies the drive current to the motor 6) corresponds to the plant, the process of FIG. 22 corresponds to the control means, and the process of FIG. 11 is identified. Corresponding to the means, the process of FIG. 12 corresponds to the identification error calculating means, the process of FIG. 14 corresponds to the identification error correcting means, step S34 of FIG. 11 corresponds to the model parameter vector calculating means, and the process of FIG. Corresponds to the limiting means.

【0129】

(Second Embodiment) In the first embodiment described above, the controlled target model is defined by using the equation (1) including the dead time d, and the state predictor 23 is used to predict after the dead time d has elapsed. By calculating the deviation amount PREDTH, the controlled target model including the dead time is controlled. Therefore, it is necessary to execute the calculation corresponding to the state predictor 23 on the CPU, and the amount of calculation on the CPU becomes large. Therefore, in the present embodiment, in order to reduce the computing load applied to the CPU, the controlled target model is defined by the following equation (1a) in which the waste time d is set to "0", and the waste time d is set to "0". The modeling error caused by is compensated by the robustness of the adaptive sliding mode control.<img he="15" id="000019" wi="146" file="19_0003602811.tif" img-format="tif" img-content="drawing" /> 【0130】

In order to further reduce the computing load of the CPU, a fixed gain algorithm is adopted as a model parameter identification algorithm. In addition, as a method to prevent drift of model parameters in order to further stabilize the control, in dead zone processing<u style="single">Substitute</u>It uses other methods. Hereinafter, this embodiment will be described in detail, focusing on the differences from the first embodiment. It is the same as the first embodiment except for the points described below.

【0131】

FIG. 31 is a functional block diagram of the throttle valve controller realized by ECU7, which includes an adaptive sliding mode controller 21a, a model parameter identifier 22a, a model parameter scheduler 25, and an accelerator pedal depression amount ACC. It is composed of a target opening setting unit 24 for setting a target opening THR of the throttle valve 3 according to the above.

【0132】

The detected throttle valve opening TH is input to the adaptive sliding mode controller 21a instead of the predicted deviation amount PREDTH, and the duty ratio is controlled by the adaptive sliding mode control so that the throttle valve opening TH matches the target opening THR. DUT is calculated.

【0133】

By using the adaptive sliding mode controller 21a, the same effect as that described in the first embodiment can be obtained, and the robustness of the control system with respect to the dead time of the controlled object can be ensured. Therefore, it is possible to compensate for the modeling error caused by setting the dead time d to "0".

【0134】

The model parameter identifier 22a is a modified model parameter vector θL (θL) by a method different from that of the first embodiment.^T= [a1, a2, b1, c1]) is calculated and supplied to the adaptive sliding mode controller 21a. More specifically, the model parameter identifier 22a corrects the reference model parameter vector θbase supplied from the model parameter scheduler 25 based on the throttle valve opening TH and the duty ratio DUT to correct the model parameter vector θ. calculate. Further, the modified model parameter vector θL is calculated by performing limit processing on the model parameter vector θ, and the modified model parameter vector θL is supplied to the adaptive sliding mode controller 21a. In this way, the optimum model parameters a1, a2, and b1 for making the throttle valve opening TH follow the target opening THR are obtained, and further, the model parameter c1 indicating the disturbance and the deviation of the default opening THDEF is obtained.

【0135】

The model parameter scheduler 25 uses the reference model parameter vector θbase (θbase) based on the throttle valve opening TH.^T= [a1base, a2base, b1base, c1base]) is calculated and supplied to the model parameter identifier 22a.

【0136】

In the present embodiment, since the control target model is defined by the above equation (1a), the adaptive sliding mode controller 21a uses the equivalent control input Ueq, the arrival rule input Urch, and the adaptive rule input Uadp as the above equation (9a), Instead of (10a) and (11a), it is calculated by the following equations (9b), (10b) and (11b).

[Number 11]<img he="76" id="000020" wi="153" file="20_0003602811.tif" img-format="tif" img-content="drawing" /> 【0137】

Equations (9b) to (11b) can be obtained by setting the waste time d of the equations (9) to (11) to "0". The model parameter identifier 22a calculates the model parameter vector of the controlled object model based on the input (DUT (k)) and the output (TH (k)) of the controlled object as described above. Specifically, the model parameter identifier 22a calculates the model parameter vector θ (k) by the following equation (15) (reposted). θ (k) = θ (k-1) + KP (k) ide (k) (15) [0138]

The identification error ide (k) of equation (15) is defined by the following equations (17) (repost), (18) (repost) and (19a). In the equation (19a), the waste time d in the equation (19) is set to 0. The gain coefficient vector KP (k) is defined by the following equation (20) (repost), and the square matrix P (k) of the equation (20) is calculated by the following equation (21) (repost) [Equation 12].<img he="42" id="000021" wi="153" file="21_0003602811.tif" img-format="tif" img-content="drawing" />[Number 13]<img he="47" id="000022" wi="153" file="22_0003602811.tif" img-format="tif" img-content="drawing" /> 【0139】

In the present embodiment, in addition to satisfying the following requirements B1 to B3 as in the first embodiment, it is required to further satisfy the following requirements B4 and B5. B1) Adaptation to changes in quasi-static dynamic characteristics and variations in hardware characteristics "Changes in quasi-static dynamic characteristics" are characteristic changes with a slow rate of change, such as fluctuations in power supply voltage and aging of hardware. B2) Adaptation to dynamic dynamic characteristic changes Specifically, it means adaptation to dynamic characteristic changes corresponding to changes in throttle valve opening TH. B3) Prevention of drift of model parameters Prevents problems such as an increase in the absolute value of model parameters due to the influence of identification errors caused by non-linear characteristics of controlled objects that should not be reflected in model parameters. B4) Matching with the computing power of the ECU Specifically, it is required to further reduce the amount of computing. B5) Stabilization of model parameters (control performance) Specifically, it is required to suppress variations in identified model parameters as much as possible.

【0140】

First, in order to satisfy the requirement B4, the fixed gain algorithm is adopted by setting the coefficients λ1 and λ2 to 1,0, respectively. As a result, since the square matrix P (k) becomes constant, the calculation of the equation (21) can be omitted, and the amount of calculation can be significantly reduced.

【0141】

That is, when the fixed gain algorithm is adopted, the equation (20) is simplified as the following equation (20a). In equation (20a), P is a square matrix with constants as diagonal elements.

[Number 14]<img he="22" id="000023" wi="153" file="23_0003602811.tif" img-format="tif" img-content="drawing" />According to the algorithm simplified in this way, the amount of calculation can be reduced. However, the equation (15) for calculating the model parameter vector θ (k) can be rewritten as the following equation (15b), and since it has an integral structure of the identification error ide (k), the model parameter drifts. easy.<img he="15" id="000024" wi="148" file="24_0003602811.tif" img-format="tif" img-content="drawing" />Here, θ (0) is an initial value vector having the initial value of the model parameter as an element.

【0142】

Therefore, in the present embodiment, in order to prevent such drift of the model parameters, the model parameter vector θ (k) is calculated by the following equation (15c) instead of the above equation (15b).<img he="30" id="000025" wi="148" file="25_0003602811.tif" img-format="tif" img-content="drawing" />Here, DELTA is an oblivion coefficient vector having an oblivion coefficient DELTA i (i = 1 to 4) as an element, as shown by the following equation. DELTA = [DELTA1, DELTA2, DELTA3, DELTA4]

【0143】

The forgetting coefficient DELTAi is set to a value between 0 and 1 (0 &lt;DELTAi &lt;1) and has a function of gradually reducing the influence of past identification errors. However, either the coefficient DELTA3 related to the calculation of the model parameter b1 or the coefficient DELTA4 related to the calculation of the model parameter c1 is set to "1" so that the forgetting coefficient is not substantially multiplied. By setting a part of the elements of the forgetting coefficient vector DELTA to "1" in this way, it is possible to prevent a steady deviation between the target value DTHR and the throttle valve opening deviation amount DTH. If both the coefficients DELTA3 and DELTA4 are set to "1", the drift prevention effect of the model parameters becomes insufficient, so it is desirable to set only one of them to "1".

【0144】

By rewriting equation (15c) into recurrence formula form, the following equations (15d) and (15e) are obtained. The method of calculating the model parameter vector θ (k) using the following equations (15d) and (15e) instead of the equation (15) is hereinafter referred to as the δ modification method, and dθ (d θ (15e) defined by the equation (15e)). k) is called "update vector". θ (k) = θ (0) + dθ (k) (15d) dθ (k) = DELTA × dθ (k-1) + KP (k) ide (k) (15e) [0145]

According to the algorithm using the δ correction method, not only the drift prevention effect that satisfies the requirement B3 but also the stabilization effect of the model parameters that satisfy the requirement B5 can be obtained. That is, the initial value vector θ (0) is always saved, and the update vector dθ (k) is also limited by the action of the forgetting coefficient vector DELTA, so that each model parameter is stable near the initial value. Can be made to.

【0146】

Furthermore, since the model parameters are calculated while adjusting the update vector dθ (k) by identification based on the input / output data of the actual control target, the model parameters suitable for the actual control target can be calculated and the above requirement B1 is also satisfied. .. Next, in order to satisfy the requirement B2, in this embodiment, the model parameter vector θ (k) is calculated by the following equation (15f) using the reference model parameter vector θbase instead of the initial value vector θ (0) in the above equation (15d). Was decided to be calculated. θ (k) = θbase + dθ (k) (15f) [0147]

Since the reference model parameter vector θbase is set by the model parameter scheduler 25 according to the throttle valve opening deviation amount DTH (= TH-THDEF), it is adapted to the change in dynamic characteristics corresponding to the change in the throttle valve opening TH. And the above requirement B2 can be satisfied.

【0148】

As described above, in the present embodiment, the fixed gain algorithm is adopted to reduce the amount of calculation of the ECU (requirement B4), and the algorithm using the δ correction method is adopted to change the quasi-static dynamic characteristics. By adopting the model parameter scheduler 25, it realizes adaptation to hardware characteristic variations (requirement B1), stabilization of model parameters (control performance) (requirement B5), and prevention of model parameter drift (requirement B3). , Adaptation to changes in dynamic characteristics corresponding to changes in throttle valve opening TH (requirement B2) has been realized.

【0149】

Limit processing is applied to each element a1', a2', b1'and c1'of the model parameter vector θ (k) calculated by the equation (15f), and the modified model parameter vector θL (k) (θL (k)).^T= [a1, a2, b1, c1]) is calculated in the same manner as in the first embodiment.

【0150】

In addition, the model parameter c1'is statistically processed, the center value of the fluctuation is calculated as the default opening deviation thdefadp, and the throttle valve opening deviation amount DTH and the target value DTHR are calculated by the following equations (41) (42) (repost). The point of calculation is the same as that of the first embodiment. DTH (k) = TH (k) -THDEF + thdefadp (41) DTHR (k) = THR (k)-THDEF + thdefadp (42) [0151]

Next, the arithmetic processing in the CPU of the ECU 7 for realizing the functions of the adaptive sliding mode controller 21a, the model parameter identifier 22a, and the model parameter scheduler 25 described above will be described.

【0152】

FIG. 32 is an overall flowchart of throttle valve opening control. In this process, step S13 (calculation of the state predictor) of the throttle valve opening control process shown in FIG. 9 is deleted, and steps S12, S14 and S16 are changed to steps S12a, 14a and 16a, respectively.

【0153】

In step S12a, the calculation of the model parameter identifier shown in FIG. 33, that is, the calculation process of the model parameter vector θ (k) according to the above equation (15f) is executed, and further the limit process is executed to execute the modified model parameter vector θL (k). ) Is calculated.

【0154】

In step S14a, the arithmetic processing of the control input Usl (k) shown in FIG. 36 is executed using the modified model parameter vector θL (k). That is, the equivalent control input Ueq, the ultimate rule input Urch (k), and the adaptive rule input Uadp (k) are calculated by the above equations (9b), (10b), and (11b), and the control input Usl ( Calculate k) (= duty ratio DUT (k)).

【0155】

In step S16a, the stability determination process of the sliding mode controller shown in FIG. 41 is executed. That is, the stability determination of the sliding mode controller is performed using the switching function value σ instead of the prediction switching function value σpre, and the stability determination flag FSMCSTAB is set. The processing when the stability determination flag FSMCSTAB is set to "1" is the same as that of the first embodiment.

【0156】

FIG. 33 is a flowchart of the arithmetic processing of the model parameter identifyr 22a. In this process, steps S31 to S34 of the arithmetic process of the model parameter identifyr shown in FIG. 11 are changed to steps S31a to S34a, respectively, and steps S33b and S33c are added.

【0157】

In step S31a, the gain coefficient vector KP (k) is calculated by the equation (20a), and then the estimated throttle valve opening deviation amount DTHHAT (k) is calculated by the equations (18) and (19a) (step S32a). In step S33a, the arithmetic processing of ide (k) shown in FIG. 35 is executed, and the identification error ide (k) is calculated. In step S33b, the update vector dθ (k) is calculated by the equation (15e), then the θbase table shown in FIG. 34 is searched according to the throttle valve opening deviation amount DTH, and the reference model parameter vector θbase is calculated (step). S33c). Reference model parameters a1base, a2base and b1base are set in the θbase table. When the throttle valve opening deviation amount DTH takes a value near "0" (throttle valve opening TH is near the default opening THDEF), the reference model parameters a1base and b1base decrease, and the reference model parameter a2base increases. It is set to do. The reference model parameter c1base is set to "0".

【0158】

In step S34a, the model parameter vector θ (k) is calculated by the equation (15f), and then the stabilization process of the model parameter vector θ (k) is executed as in the first embodiment (step S35). That is, the modified model parameter vector θL (k) is calculated by performing the limit processing of each model parameter.

【0159】

FIG. 35 is a flowchart of the ide (k) arithmetic processing executed in step S33a of FIG. 33. In this process, step S56 (dead zone process) of the ide (k) arithmetic process in FIG. 12 is deleted, and step S51 is changed to step S51a. That is, in the present embodiment, the dead zone processing is not executed because the drift of the model parameters is prevented by the δ correction method.

【0160】

Further, in step S51a, the estimated throttle valve opening deviation amount DTHHAT (k) is calculated by the equations (18) and (19a), and the identification error ide (k) is calculated using the estimated throttle valve opening deviation amount DTHHAT (k). ) Is calculated. In the present embodiment, since the dead time d of the controlled target model is set to 0, the predetermined value XCNTIDST in step S52 is set to, for example, 2.

【0161】

FIG. 36 is a flowchart of a process for calculating the control input Usl (= DUT) to the throttle valve drive device 10 executed in step S14a of FIG. 32. In this process, steps S201 to S205 of the Usl arithmetic process shown in FIG. 22 are changed to steps S201a to S205a, respectively.

【0162】

In step S201a, the arithmetic processing of the switching function value σ shown in FIG. 37 is executed, and in step S202a, the arithmetic processing of the integrated value of the switching function value σ shown in FIG. 38 is executed. In step S203a, the equivalent control input Ueq is calculated by the above equation (9b). In step S204a, the arithmetic processing of the arrival rule input Urch shown in FIG. 39 is executed, and in step S205a, the arithmetic processing of the adaptive law input Uadp shown in FIG. 40 is executed.

【0163】

FIG. 37 is a flowchart of the arithmetic processing of the switching function value σ executed in step S201a of FIG. In this process, steps S222 to S226 of the arithmetic processing of the predictive switching function value σpre shown in FIG. 23 are changed to steps S222a to 226a, respectively.

【0164】

In step S222a, the switching function value σ (k) is calculated by the above equation (5). Subsequent steps S223a to S226a replace "σpre" in steps S223 to S226 in FIG. 23 with "σ", and perform the same limit processing as the processing in FIG. 23 for the switching function value σ (k). Do.

【0165】

FIG. 38 is a flowchart of a process for calculating the integrated value SUMSIGMAa of the switching function value σ, which is executed in step S202a of FIG. In this process, steps S241 to S245 of the integrated value calculation process of σpre shown in FIG. 26 are changed to steps S241a to S245a, respectively. The integrated value SUMSIGMAa is used to calculate the adaptive law input Uadp in the process of FIG. 40 described later (see the above equation (11b)).

【0166】

In step S241a, the integrated value SUMSIGMAa is calculated by the following formula (47a). SUMSIGMAa (k) = SUMSIGMAa (k-1) + σ × ΔT (47a) In the following steps S242a to S245a, the calculated integrated value SUMSIGMAa is subjected to the same limit processing as in FIG. 26.

【0167】

FIG. 39 is a flowchart of the arithmetic processing of the arrival rule input Urch executed in step S204a of FIG. In this process, steps S263 and S265 of the arrival rule input Urch arithmetic process shown in FIG. 27 are changed to steps S263a and S265a, respectively.

【0168】

That is, in the present embodiment, the switching function value σ is used instead of the predictive switching function value σpre to calculate the arrival rule input Urch when the adaptive sliding mode controller 21a is stable (step S263a), and the adaptive sliding mode controller. Execute the calculation of the arrival rule input Urch when 21a is unstable (step S265a).

【0169】

FIG. 40 is a flowchart of the arithmetic processing of the adaptive law input Uadp executed in step S205a of FIG. In this process, steps S273 and S275 of the adaptive law input Uadp arithmetic process shown in FIG. 28 are changed to steps S273a and S275a, respectively.

【0170】

That is, in the present embodiment, the calculation of the adaptive law input Uadp when the adaptive sliding mode controller 21a is stable (step S273a) and the adaptive sliding mode controller 21a are unstable by using the integrated value SUMSIGUMAa of the switching function value σ. The calculation of the adaptive law input Uadp when is (step S275a) is executed.

【0171】

FIG. 41 is a flowchart of the stability determination process of the sliding mode controller executed in step S16a of FIG. 32. In this process, steps S281 and S281 in FIG. 29 are changed to steps S281a and S282a, respectively.

【0172】

In step S281a, the switching function change amount Dσ is calculated by the following equation (52a), and in step S282a, the stability discrimination parameter SGMSTAB is calculated by the following equation (53a). That is, the stability is determined based on the switching function value σ instead of the predictive switching function value σpre. Dσ = σ (k) -σ (k-1) (52a) SGMSTAB = Dσ × σ (k) (53a) [0173]

In this embodiment, a part of the throttle valve drive device 10 and the ECU 7 (the output circuit that supplies the drive current to the motor 6) corresponds to the plant, the process of FIG. 36 corresponds to the control means, and the process of FIG. 33 is identified. Corresponding to the means, the process of FIG. 35 corresponds to the identification error calculating means, step S33b of FIG. 33 corresponds to the update vector calculating means and the update vector correcting means, and step S34a of FIG. 33 corresponds to the model parameter vector calculating means. However, step S35 in FIG. 33 corresponds to the limiting means.

【0174】

(Third Embodiment) FIG. 42 is a block diagram showing a configuration of a control system according to a third embodiment of the present invention. This control system includes the plant 101 to be controlled, the pH sensor 102 that detects the pH (pH) of the mixed solution that is the output of the plant, and the subtractor 103 that subtracts the first reference value V1BASE from the pH sensor output V1OUT. , The target value generation unit 104 that generates the control target value V1TARGET, the operation amount determination unit 105 that determines the first operation amount U1, the first operation amount U1 and the second reference value V2BASE are added, and the second operation amount It consists of an adder 106 that outputs U2.

【0175】

The subtractor 103, the target value generation unit 104, the manipulated variable determination unit 105, and the adder 106 are specifically composed of an electronic control unit including a CPU, a memory, an input / output circuit, and the like. The plant 101 is composed of a flow rate control valve 111 that controls the flow rate of the alkaline liquid according to the second operation amount U2, and a stirrer 112 that agitates the alkaline liquid and the acidic liquid supplied via the flow rate control valve 111. Become. The plant 101 outputs a mixed solution having a desired pH value by stirring the alkaline solution and the acidic solution.

【0176】

The manipulated variable determination unit 105 includes an identifier 121 that identifies the model parameter vector of the controlled model that models the plant 101, an adaptive sliding mode controller 122, and a predictor 123. The identifyr 121, the adaptive sliding mode controller 122, and the predictor 123 correspond to the model parameter identifyr 22, the adaptive sliding mode controller 21, and the state predictor 23 in the first embodiment, respectively, and have the same functions as these. Have.

【0177】

Hereinafter, the correspondence between the components and parameters in the present embodiment and the components and parameters in the first embodiment will be described. pH sensor 10<u style="single">2</u>Corresponds to the throttle valve opening sensor 8 and the pH sensor 10<u style="single">2</u>The output V1OUT of is corresponding to the throttle valve opening TH. The first target value V1BASE corresponds to the default opening degree THDEF, and in the present embodiment, the pH value corresponds to, for example, neutrality. Therefore, the deviation amount DV1 corresponds to the throttle valve opening deviation amount DTH. Further, the target value generation unit 104 corresponds to the target opening degree setting unit 24, and the control target value V1TARGET corresponds to the target value DTHR of the throttle valve opening deviation amount. In the first embodiment, the function of the subtractor 103 is included in the model parameter identifyr 22 and the state predictor 23.

【0178】

The second reference value V2BASE is added to bias the center value of the first manipulated variable U1 which is the output of the adaptive sliding mode controller 122. In the first embodiment, there is no component corresponding to the adder 106, so the second reference value V2BASE is effectively "0" (ie, U1 = U2 = Usl). In the present embodiment, the second reference value V2BASE is set to a value such that the opening degree of the flow rate control valve 111 is 50%, for example.

【0179】

The flow rate control valve 111 corresponds to a switching element (included in the output circuit of the ECU 7 and not shown or described) whose on / off control is controlled by a pulse signal having a duty ratio of DUT, and the alkaline solution corresponds to the power supply voltage. Further, the output flow rate V2 of the flow control valve 111 corresponds to the drive current of the motor 6, the stirrer 112 corresponds to the valve body of the motor 6 and the throttle valve 3, and the acidic liquid is the intake air applied to the valve body of the throttle valve 3. It corresponds to the negative pressure of the pipe and the urging force of the return spring 4 and the elastic member 5. The pH value V1 of the mixture output from the stirrer 112 corresponds to the actual throttle valve opening.

【0180】

Since there is a correspondence as described above, the plant 101 can be modeled in the same manner as in the first embodiment, and the same control method can be applied. That is, the appraiser 121 calculates the modified model parameter parameter vector θL based on the first manipulated variable U1 and the deviation amount DV1 by the same arithmetic processing as in the first embodiment, and the predictor 123 calculates the modified model parameter parameter vector θL. Based on the quantity U1, the deviation amount DV1, and the modified model parameter vector θL, the predicted deviation amount PREDV1 is calculated by the same arithmetic processing as in the first embodiment, and the adaptive sliding mode controller 122 determines the predicted deviation amount PREDV1 and the modified model. Based on the parameter vector θL, the first manipulated variable U1 is calculated so that the predicted deviation amount PREDV1 matches the control target value V1TARGET by the same arithmetic processing as in the first embodiment. Therefore, by setting a desired relative pH value (amount of deviation from the first reference value V1BASE) as the control target value V1TARGET, the output V1 of the plant can be matched with the desired pH value. In the present embodiment, the identifyr 121 corresponds to the identification means, and includes the identification error calculating means, the identification error correcting means, the model parameter vector calculating means, and the limiting means.

【0181】

(Modified Example of Third Embodiment) FIG. 43 is a diagram showing a modified example of the configuration shown in FIG. 42. In this modification, plant 101a is controlled instead of plant 101 in FIG. The plant 101a provides the plant 101 with a flow rate sensor 113 for detecting the output flow rate V2 of the flow rate control valve 111 and a flow rate control valve 111 so that the flow rate sensor output V2OUT matches the flow rate value corresponding to the second operation amount U2. It is configured by adding a feedback controller 114 to control.

【0182】

In this way, the same modeling and the same control method as in the third embodiment can be applied to the plant including the local feedback loop. In the first embodiment, since the drive circuit of the motor is known, detailed description is not given. However, a current sensor for detecting the output current of the switching element controlled to be turned on / off is provided, and the detected current value ID is provided. However, the feedback control may be performed so as to match the current value IR corresponding to the operation amount Usl, and this modification corresponds to the case where such a circuit configuration is adopted in the first embodiment.

【0183】

(Fourth Embodiment) FIG. 44 is a block diagram showing a configuration of a control system according to a fourth embodiment of the present invention. This control system replaces the manipulated variable determination unit 105 of FIG. 42 with the manipulated variable determination unit 105a, and corresponds to the control system shown as the second embodiment. It is the same as the third embodiment except as described below.

【0184】

The manipulated variable determination unit 105a includes an identifier 121a, an adaptive sliding mode controller 122a, and a parameter scheduler 124. The identifyr 121a, the adaptive sliding mode controller 122a, and the parameter scheduler 124 correspond to the model parameter identifyr 22a, the adaptive sliding mode controller 21a, and the model parameter scheduler 25 in the second embodiment, respectively, and have the same functions as these.

【0185】

That is, the parameter scheduler 124 calculates the reference model parameter vector θbase based on the deviation amount DV1 by the same arithmetic processing as in the second embodiment, and the identifier 121a uses the first operation amount U1, the deviation amount DV1 and the reference. The modified model parameter vector θL is calculated based on the model parameter vector θbase by the same arithmetic processing as in the second embodiment, and the adaptive sliding mode controller 122a is based on the deviation amount DV1 and the modified model parameter vector θL. The first manipulated variable U1 is calculated so that the deviation amount DV1 matches the control target value V1TARGET by the same arithmetic processing as in the above embodiment. Therefore, by setting a desired relative pH value (amount of deviation from the first reference value V1BASE) as the control target value V1TARGET, the output V1 of the plant can be matched with the desired pH value.

【0186】

In the present embodiment, the identifyr 121a corresponds to the identification means, and includes the identification error calculation means, the update vector calculation means, the update vector correction means, the model parameter vector calculation means, and the limiting means.

【0187】

(Modified Example of Fourth Embodiment) FIG. 45 is a diagram showing a modified example of the configuration shown in FIG. 44. In this modification, the plant 101a is controlled instead of the plant 101 in FIG. Plant 101a is identical to plant 101a in FIG. As described above, the same modeling and the same control method as in the fourth embodiment can be applied to the plant including the local feedback loop.

【0188】

(Other Embodiments) As a method for calculating the identification error ide (k) of the model parameter, the ε correction method described below may be adopted instead of the δ correction method. That is, instead of the above equation (15c), the model parameter vector θ (k) may be calculated by the following equation (15g).<img he="38" id="000026" wi="148" file="26_0003602811.tif" img-format="tif" img-content="drawing" />Here, EPS is an oblivion coefficient vector having an oblivion coefficient EPSi (i = 1 to 4) as an element, as shown by the following equation. EPS = [EPS1, EPS2, EPS3, EPS4]

【0189】

The forgetting coefficients EPS1, EPS2 and EPS4 are set to a value between 0 and 1 (0 &lt;EPSi &lt;1) like the forgetting coefficient DELTAi, and have a function of gradually reducing the influence of past identification errors.

【0190】

However, in the case of the ε correction method, the coefficient EPS3 related to the calculation of the model parameter b1 must always be set to "1". This is due to the following reasons. In the case of the ε correction method, when the identification error ide (k) becomes small, all the model parameters become values near zero. However, since the model parameter b1 is applied to the denominators of the equations (9b), (10b), and (11b), when the model parameter b1 approaches "0", the input Usl to the controlled object diverges. ..

【0191】

Equation (15g) differs from equation (15c) in that the initial value vector θ (0) is also multiplied by the forgetting coefficient vector EPS. By rewriting the equation (15g) into the recurrence formula form, the following equation (15h) is obtained. A method of calculating the model parameter vector θ (k) using the following equation (15h) instead of the equation (15) is called an ε correction method. θ (k) = EPS × θ (k-1) + KP (k) ide (k) (15h) [0192]

The ε correction method also reduces the influence of past identification error aids, so that model parameter drift can be prevented. Further, in the second embodiment, the drift of the model parameters is prevented by the δ correction method, but the correction identification error idenl (k) is calculated by the dead zone processing (Fig. 14) as in the first embodiment. , The model parameter vector θ (k) may be calculated using this.

【0193】

Further, in the first embodiment, the δ correction method or the ε correction method may be adopted instead of the dead zone treatment. Further, when the δ modification method is adopted in the first embodiment, the model parameter scheduler is introduced as in the second embodiment, and the update vector is added to the reference model parameter vector θbase calculated by the model parameter scheduler. It is desirable to calculate the model parameter vector θ in the above format.

【0194】

[Effect of the invention]

As described in detail above, according to the invention of claim 1, the update vector is calculated according to the identification error of the model parameter vector, and the update vector is modified so as to reduce the influence of the past value of the identification error. By adding the modified update vector to the reference vector of the model parameter, the model parameter vector is calculated, and the value of the element of the model parameter vector is further limited within a predetermined limit range. Therefore, the stability of control can be further improved while preventing the drift of model parameters.<u style="single">Further, since the reference vector of the model parameter is calculated according to the parameter indicating the change in the dynamic characteristics of the plant, an appropriate reference vector corresponding to the change in the dynamic characteristics of the plant can be obtained. As a result, model parameters can be quickly converged, especially when the plant contains non-linear elements.</u>【0195】

According to the invention of claim 2, since the update vector is calculated by using the fixed gain algorithm, the amount of calculation can be reduced.

【0196】

According to the invention of claim 3, the update vector is corrected by multiplying the past value of at least one element of the update vector by a predetermined value greater than 0 and less than 1, and thus the past of the identification error. The effect of the value is reduced and the model parameter vector can be prevented from drifting.

【0197】

According to the invention of claim 4, the element related to the input of the plant or the element not related to the input / output of the plant of the update vector is not multiplied by a predetermined value larger than 0 and smaller than 1, so that the update vector is modified. It is possible to prevent the occurrence of steady deviation due to.

【0200】

Claim<u style="single">5</u>According to the invention described in, the identification error of the model parameter<u style="single">Dead zone</u>When it is within, the identification error is corrected in the decreasing direction, the model parameter vector is calculated using the corrected identification error, and the value of the element of the model parameter vector is further limited within a predetermined limit range. Therefore, the stability of control can be further improved while preventing the drift of model parameters.<u style="single">Further, since the dead zone is set according to the output of the plant or the control target value, it is possible to prevent the identification error to be reflected in the value of the model parameter from being ignored as the residual identification error.</u>【0201】

Claim<u style="single">6</u>According to the invention described in the above, since the values of the plurality of elements are limited so that the plurality of elements of the model parameter vector satisfy a predetermined relationship, the stability of the control using the model parameter vector can be improved. it can.

【0202】

Claim<u style="single">7</u>According to the invention described in the above, the identification error is described above.<u style="single">Dead zone</u>When it is inside, the identification error is set to 0, so that the influence of the identification error that should not be reflected in the value of the model parameter can be eliminated and the drift prevention effect of the model parameter can be enhanced.

【0203】

Claim<u style="single">8</u>According to the invention described in the above, since the control to match the throttle valve opening degree with the target opening degree is performed by using the model parameter identified by the identification means, the controllability of the throttle valve opening degree to the target opening degree is controlled. At the same time, the stability of the throttle valve opening control can be improved.

[Simple explanation of drawings]

FIG. 1 is a diagram showing a throttle valve drive device for an internal combustion engine and a control device thereof according to an embodiment of the present invention.

FIG. 2 is a diagram showing a frequency characteristic of the throttle valve drive device shown in FIG.

3 is a functional block diagram showing a function realized by the electronic control unit (ECU) of FIG. 1. FIG.

FIG. 4 is a diagram showing a relationship between a control characteristic of a sliding mode controller and a value of a switching function setting parameter (VPOLE).

FIG. 5 is a diagram showing a setting range of control gain (F, G) of the sliding mode controller.

FIG. 6 is a diagram for explaining drift of model parameters.

FIG. 7 is a diagram showing a function for correcting an identification error.

FIG. 8 is a diagram for explaining that the default opening deviation of the throttle valve is reflected in the model parameter (c1').

FIG. 9 is a flowchart of a throttle valve opening degree control process.

FIG. 10 is a flowchart of a process of setting a state variable in the process of FIG.

11 is a flowchart of a process for executing an operation of a model parameter identifyr in the process of FIG. 9. FIG.

12 is a flowchart of a process for executing an identification error (ide) operation in the process of FIG. 11. FIG.

FIG. 13 is a diagram for explaining low-pass filtering of identification error (ide).

FIG. 14 is a flowchart of dead zone processing in the processing of FIG.

FIG. 15 is a diagram showing a table used in the process of FIG.

FIG. 16 is a flowchart of a model parameter vector (θ) stabilization process in the process of FIG.

FIG. 17 is a flowchart of limit processing of model parameters (a1', a2') in the processing of FIG.

FIG. 18 is a diagram for explaining changes in model parameter values due to the processing of FIG.

FIG. 19 is a flowchart of a limit process of a model parameter (b1') in the process of FIG.

FIG. 20 is a flowchart of a limit process of a model parameter (c1') in the process of FIG.

FIG. 21 is a flowchart of a process of executing an operation of a state predictor in the process of FIG.

FIG. 22 is a flowchart of a process for executing a control input (Usl) operation in the process of FIG.

FIG. 23 is a flowchart of a process for executing an operation of a predictive switching function value (σpre) in the process of FIG. 22.

FIG. 24 is a flowchart of a process for executing an operation of a switching function setting parameter (VPOLE) in the process of FIG. 23.

FIG. 25 is a diagram showing a map used in the process of FIG. 24.

FIG. 26 is a flowchart of a process of executing an operation of an integrated value of a predictive switching function value (σpre) in the process of FIG. 22.

FIG. 27 is a flowchart of a process for executing an operation of arrival rule input (Urch) in the process of FIG. 22.

FIG. 28 is a flowchart of a process of executing an operation of an adaptive law input (Uadp) in the process of FIG. 22.

FIG. 29 is a flowchart of a process for executing stability determination of the sliding mode controller in the process of FIG.

FIG. 30 is a flowchart of a process for executing a calculation of a default opening degree deviation (thdefadp) in the process of FIG.

FIG. 31 is a functional block diagram showing a function realized by the electronic control unit (ECU) of FIG. 1 (second embodiment).

FIG. 32 is a flowchart of a throttle valve opening degree control process (second embodiment).

FIG. 33 is a flowchart of a process for executing an operation of a model parameter identifyr in the process of FIG. 32.

FIG. 34 is a diagram showing a table used in the process of FIG. 33.

FIG. 35 is a flowchart of a process for executing an identification error (ide) operation in the process of FIG. 33.

FIG. 36 is a flowchart of a process for executing a control input (Usl) operation in the process of FIG. 32.

FIG. 37 is a flowchart of a process for executing an operation of a switching function value (σ) in the process of FIG. 36.

FIG. 38 is a flowchart of a process of executing an operation of an integrated value of a switching function value (σ) in the process of FIG. 36.

[Fig. 39] Fig. 39 is a flowchart of a process of executing an operation of an arrival rule input (Urch) in the process of FIG. 36.

FIG. 40 is a flowchart of a process for executing an operation of an adaptive law input (Uadp) in the process of FIG. 36.

FIG. 41 is a flowchart of a process for executing stability determination of the sliding mode controller in the process of FIG. 32.

FIG. 42 is a block diagram showing a configuration of a control system according to a third embodiment of the present invention.

FIG. 43 is a block diagram showing a modified example of the configuration shown in FIG. 42.

FIG. 44 is a block diagram showing a configuration of a control system according to a fourth embodiment of the present invention.

FIG. 45 is a block diagram showing a modified example of the configuration shown in FIG. 44.

[Explanation of symbols]

1 Internal combustion engine 3 Throttle valve 7 Electronic control unit 10 Throttle valve drive device 21,21a Adaptive sliding mode controller (control means) 22,22a Model parameter identifyr (identification means, identification error calculation means, update vector calculation means, update vector correction means) , Identification error correction means, model parameter vector calculation means, limiting means) 24 Target opening setting unit 25 Model parameter scheduler